WO2014041087A1 - Muscle secretome and uses thereof - Google Patents

Abstract

The present invention relates to the muscle secretome and uses thereof. The inventors have shown that differentiating human muscle cells release not only a plethora of soluble secreted proteins through conventional secretory mechanisms but also complex protein/nucleic acid cargos via membrane microvesicle (MVs) shedding. Two types of plasma membrane-derived vesicles were identified: nanovesicles harbouring typical exosomal features and larger heterogeneous MVs frequently referred to as microparticles. The present invention relates to an isolated nanovesicle secreted by a muscle cell, an isolated microparticle secreted by a muscle cell, optionally for use in therapy, a method for delivering a molecule of interest into a target cell, a method for diagnosing or monitoring a muscular disease in a subject, a method for aiding in the evaluation of treatment efficacy in a subject and an antibody microarray.

Description

Muscle secretome and uses thereof

FIELD OF THE INVENTION

The present invention relates to the muscle secretome.

BACKGROUND OF THE INVENTION

Secreted proteins ("secretome") constitute an important class of active molecules regulating physiological and pathological processes. In higher organisms protein secretion is complex and tightly regulated, reflecting the functionality of a cell in a given environment. Besides endocrine organs specialised in the secretion of specific molecules, there is increasing evidence that many other tissues including skeletal muscle secrete factors with local and systemic effects. Skeletal muscle accounts for about 40% of body mass, is responsible for movement and is an important metabolic organ. In adults, it has a very low basal rate of cellular turnover but retains a remarkable capacity to adapt to normal physiological demands during growth and training, and to regenerate in response to injury or disease. Secreted signalling molecules including growth factors and cytokines have been shown to modulate activation, proliferation and differentiation of satellite cells. The source of these factors within regenerating muscle tissue is still under investigation. Numerous lines of evidence suggest that human muscle cells in vitro secrete a compendium of signalling molecules sufficient to drive myotube formation in an autocrine and paracrine fashion. However, muscle regeneration in vivo is complex and requires a coordinated interplay between satellite cells, inflammatory and muscle-resident stromal cells. This type of cross-talk involves necessarily cell to cell signalling, and if many partners involved have been identified, much less is known about the factors secreted by muscle cells influencing the behaviours of other cell types. Despite recent efforts at defining the skeletal muscle secretome as a whole {Bortoluzzi et al. Proteins. 2006 Mar 15;62(3):776-92; Chan et al., J Proteome Res. 2007 Feb;6(2):698-710}, the picture remain largely incomplete. SUMMARY OF THE INVENTION

The inventors have shown that differentiating human muscle cells release not only a plethora of soluble secreted proteins through conventional secretory mechanisms but also complex protein/nucleic acid cargos via membrane microvesicle (MVs) shedding. The soluble "secretome" contains 254 proteins/peptides with signalling potential including 45 previously implicated in myogenesis. Many more have proven roles in modulating a range of other cell types, implying a much broader role for myoblasts in regulating skeletal muscle homeostasis and remodelling in vivo. Deeper analysis of the muscle secretome led to the identification of two types of plasma membrane-derived vesicles: nanovesicles harbouring typical exosomal features and larger heterogeneous MVs frequently referred to as microparticles. The inventors characterized the protein and RNA content of these two types of muscle-derived secreted MVs and demonstrated that they can dock and fuse differentially with adjacent muscle cells. Moreover, the inventors provided evidence for the intercellular exchange of a functional protein cargo, alkaline phosphatase activity, to target cells following microvesicle uptake. Thus, muscle-derived MVs act in vivo as "physiological liposomes" delivering protein/RNA cargo to target cells acting in concert with soluble signalling molecules to modulate complex intercellular signalling networks during muscle regeneration. Thus, the present invention relates to an isolated nanovesicle secreted by a muscle cell. The present invention also relates to an isolated microparticle secreted by a muscle cell.

The invention also relates to the use of said nanovesicle and/or said microparticle as a diagnostic biomarker for muscular diseases.

Said nanovesicle and said microparticle have proteomic profiles which are specific to the muscle. These specific proteomic profiles facilitate their isolation and their identification.

The present invention also relates to the use of said nanovesicle and/or said microparticle for delivering a molecule of interest into a target cell. DETAILED DESCRIPTION OF THE INVENTION

Nanovesicles and microparticles according to the invention

Thus, the present invention relates to two morphologically distinct populations of isolated small bilayer membrane vesicles which are muscle secreted:

- nanovesicles which are cup-shaped vesicles, 69 (±20) nm in diameter and pelleting at 100,000 g;

-a polymorph microparticles often with an electron dense core, (130 (±36) nm in diameter) in diameter, pelleting at 20,000 g.

Their protein cargo was analyzed. A total of 764 and 564 unique proteins were identified from the 20,000g and 100,000g fractions respectively (see Table S7-S8).

Comparison of the 2 muscle-derived microvesicles proteome datasets by peptide counting (cf. Liu, et al., 2004) revealed that nanovesicles and microparticles shared a common set of protein but a significant proportion of their proteome is different reflecting their distinct intracellular origins (Table S9).

Typically the nanovesicles and microparticles according to the invention may be isolated by filtration and ultracentrifugation.

Advantageously, the nanovesicles and microparticles according to the invention may be frozen and stored at -80°C without losing their ability to transfer material to the target cell. Typically, said nanovesicle or said microparticle is secreted by a muscle cell selected from the group consisting of a skeletal muscle cell, a cardiomyocyte, a smooth muscle cell, and a myoblast.

In a particular embodiment, said nanovesicle or said microparticle is secreted by a skeletal muscle cell. In a particular embodiment, said muscle cell overexpresses a molecule of interest.

By "overexpressing", it is meant any means known in the art to enhance the amount of protein expressed by a given cell.

Typically, the overexpression of the molecule of interest may be achieved by transfecting the muscle cell with an expression vector encoding molecule of interest.

In a preferred embodiment, overexpression is obtained by transfection of an exogenous DNA. Suitable transfection methods are classical methods known to the skilled person, such as calcium phosphate transfection, transfection using liposomes (also known as lipofection) or electroporation. It falls within the ability of the skilled person to select the appropriate transfection method for a given muscle cell. The term "overexpression" also covers the overexpression of an endogenous molecule, i.e. a molecule which is naturally expressed by the muscle cell. The overexpression may be achieved by the introduction of additional copies of the gene encoding said molecule or in the stimulation of the expression of the endogenous molecule. By way of example, the muscle cell can be placed under culture conditions known to enhance the expression of said endogenous molecule.

When transient overexpression of the molecule of interest is used, the nanovesicles or micropaticles are typically harvested from the cell supernatant 48-72 hours post- transfection.

In an embodiment of the invention, the muscle cell overexpresses 2 to 5 different molecules of interest

Typically, molecules of interest may be selected from the group consisting of peptides, proteins, mRNA, miRNA, viral vectors... Nanovesicles and/or microparticles according to the invention as delivery vehicles

The present invention also relates to an in vitro method for delivering a molecule of interest into a target cell by contacting said target cell with a nanovesicles and/or a microparticle according to the invention comprising said molecule of interest.

Further, the inventors have demonstrated that nanovesicles and/or microparticles of the invention can be used to efficiently deliver said molecule of interest to a target cell.

Advantageously, the inventors have shown that the molecule of interest retains its functionality once it has been transferred into the target cell.

The present invention also relates to an in vivo method for delivering a molecule of interest into a target cell by contacting said target cell with a nanovesicles and/or a microparticle of the invention comprising said molecule of interest.

Examples of target cells are involved in the muscle regeneration such as myoblasts, myotubes, muscle fibres, inflammatory cells or neighbouring fibroblasts.

The nanovesicles and/or a microparticle of the invention may also be administered subcutaneously, and thereby target dermal fibroblasts.

Typically several (e.g., 2, 3, 4, 5, 6...) different molecules of interest may be delivered into a target cell by contacting said target cell with several different nanovesicles and/or microparticles of the invention comprising said several different molecules of interest.

Typically, the contacting of said several different nanovesicles and/or microparticles may be simultaneous or sequential. As opposed to gene delivery achieved by multiple viral vectors, protein delivery with nanovesicles and/or microparticles according to the invention is an original method to introduce rapidly a function in a targeted cell, without involvement of the transcription machinery or any viral integration processes, which represent a very serious oncogenic risk for clinical use. The nanovesicles and/or microparticles according to the invention could be used in virtually any cell type including resting or fully differentiated cells, without any tumorigenic risk. They can deliver proteins or RNA with an effect limited in time by the half-life of the molecule. Such vectors do not exist at present.

Contrary to classical gene transfer techniques such as gene transfection, the method of the invention does not rely of transcription and translation within the target cell, does not perturb the cell metabolism. Thus, it is believed that the method according to the invention does not activate any interferon response and is therefore a more specific method for delivering a molecule of interest to a target cell.

Due to the low amounts of material delivered by the nanovesicles and/or microparticles according to the invention and to their non-genetic nature, the nanovesicles and/or microparticles appear to be useful for applications where low and transient presence of molecules may lead to striking biological effects.

In a preferred embodiment, a nanovesicle or a microparticle according to the invention is virus free.

Alternatively, nanovesicles or microparticles could also be used as packaging for viral particles, since they could physiologically contain AAV or HIV particles. The advantage of such a role would be to target the viral particles towards the cells targeted by the nanovesicles or microparticles. These nanovesicles or microparticles could be targeted towards a limited number of cell types by including in their membrane proteins that will recognize specific receptors on specific cell types.

The person skilled in the art will readily select the appropriate pairs of molecule of interest/ target cell, according to each specific goal. In one aspect the invention relates to the in vitro use of the nanovesicles and/or microparticles according to invention.

In one aspect the invention relates to the in vivo use of the nanovesicles and/or microparticles according to invention.

In one aspect, the invention relates to the use of a nanovesicle and/or a microparticle according to the invention for non-therapeutic applications.

Typically, the nanovesicles and/or microparticles according to the invention can be used for introducing a molecule of interest into a target cell in vitro in order to study the physiological effect of said molecule of interest. There are many possible applications of the invention as tool for basic science investigation.

Another example is the delivery of a cellular protein regulating cell expansion differentiation or death, in an in vitro cellular model.

The invention also relates to a method for inducing or potentiating cell differentiation by delivery of transcription factors

The invention also relates to a nanovesicle and/or a microparticle according to the invention for use in therapy.

In particular, the invention also relates to a nanovesicle and/or a microparticle according to the invention for use in the treatment and/or the prevention of sarcopenia, or of muscular dystrophies such as dysferlinopathies, (Limb-Girdle muscular dystrophy (LGMD)).

In an embodiment, the nanovesicle and/or the microparticle of the invention is the active ingredient in a pharmaceutically acceptable formulation suitable for administration to the subject. Generally this comprises a pharmaceutically acceptable carrier for the active ingredient. The specific carrier will depend upon a number of factors (e.g., the route of administration). The "pharmaceutically acceptable carrier" means any pharmaceutically acceptable means to mix and/or deliver the targeted delivery composition to a subject. This includes a pharmaceutically acceptable material, composition or vehicle, such as a liquid or solid filler, diluent, excipient, solvent or encapsulating material, involved in carrying or transporting the subject agents from one organ, or portion of the body, to another organ, or portion of the body. Each carrier must be "acceptable" in the sense of being compatible with the other ingredients of the formulation and is compatible with administration to a subject, for example a human.

Administration to the subject can be either systemic or localized. This includes, without limitation, dispensing, delivering or applying the active ingredient (e.g. in a pharmaceutical formulation) to the subject by any suitable route for delivery of the active compound to the desired location in the subject, including delivery by intramuscular injection, subcutaneous/intradermal injection, intravenous injection, transdermal delivery.

For performing said methods the skilled person may use the teaching of WO2009/100029 or WO2007/126386 which describe the use of exosomes as delivery vehicles.

Muscle nanovesicles and muscle microparticles according to the invention as diagnostic tools

The present invention relates to a method for diagnosing or monitoring a muscular disease in a subject, comprising the step of detecting the presence or absence of one or more biomarkers within a nanovesicle and/or a microparticle according to the invention obtained from a biological sample of said subject, wherein said one or more biomarkers are associated with said muscular disease. The method may further comprise the step of comparing the result of the detection step to a control (e.g., comparing the amount of one or more biomarkers detected in the sample to one or more control levels), wherein the subject is diagnosed as having the disease if there is a measurable difference in the result of the detection step as compared to a control. Another aspect of the invention is a method for aiding in the evaluation of treatment efficacy in a subject suffering from a muscular disease, comprising the step of detecting the presence or absence of one or more biomarkers within a nanovesicle and/or a microparticle according to the invention obtained from a biological sample of said subject, wherein the biomarker is associated with the treatment efficacy.

The method may further comprise the step of providing a series of biological samples over a period of time from the subject. Additionally, the method may further comprise the step or steps of determining any measurable change in the results of the detection step (e.g., the amount of one or more detected biomarkers) in each of the biological samples from the series to thereby evaluate treatment efficacy.

Typically, a method according to the invention may comprise the step or steps of isolating these muscle nanovesicles and/or microparticles using specific surface antigens such as the ones described in Table S12 by means of specific antibodies (FACS or MACS).

Said subject is a mammal, a human or nonhuman primate, a dog, a cat, a horse, a cow, other farm animals, or a rodent (e.g. mice, rats, guinea pig. etc.) In a preferred embodiment said subject is a human.

Typically the biological sample obtained from the subject is a sample of bodily fluid. Examples of suitable body fluids are blood and urine.

Alternatively the biological sample obtained from the subject may be a culture of muscle cells isolated from the subject.

Typically said muscular disease may be any pathological conditions resulting from an impairment of the muscle regeneration, such as fibrosis, adipogenesis or chronic inflammation. Typically said muscular disease is selected from the group consisting of neuromuscular diseases and sarcopenia.

The person skilled in the art will readily select the appropriate one or more biomarkers according to each specific muscular disease (cf. Table S13 for a list of well known biomarkers according to each specific muscular disease).

Typically said one or more biomarkers associated with a muscular disease are:

i) a species of nucleic acid or protein;

ii) a level of expression of one or more nucleic acids, peptides or proteins;

iii) a nucleic acid, peptide or protein variant; or

iv) a combination of any of the foregoing.

Typically said nucleic acid may be a mRNA or a miRNA.

Typically said peptides or proteins may be one of the proteins listed in Table SI 3.

The present invention relates to methods for detecting, diagnosing, monitoring, treating or evaluating a muscular disease in a subject comprising the steps of, isolating a nanovesicle and/or a microparticle according to the invention from a biological sample obtained from a subject, and analyzing one or more biomarkers contained within the nanovesicle and/or the microparticle. The one or more biomarkers are analyzed qualitatively and/or quantitatively, and the results are compared to results expected or obtained for one or more other subjects who have or do not have the muscular disease. The presence of a difference in the nucleic acid, peptide and protein content of nanovesicle and/or the microparticle of the subject, as compared to that of one or more other individuals, can indicate the presence or absence of, the progression of the muscular disease, or the susceptibility to a muscular disease in the subject.

For performing said methods the skilled person may use the teaching of WO2009/100029, WO2009/015357 or WO2009/021322, which all describe the use of cancer exosomes as diagnostic markers. Nanovesicles and microparticles of the invention have proteomic profiles which are specific to the muscle. In particular, the specific surface molecules carried by the nanovesicles and microparticles of the invention may be used to identify, isolate and/or purify the nanovesicles and microparticles.

Typical specific surface molecules carried by the nanovesicles are listed in Table SI 2. Typical specific surface molecules carried by the microparticles are listed in Table S12. Suitable purification methods include, but are not limited to immunoprecipitation, affinity chromatography, FACS and magnetic beads coated with specific antibodies or aptamers.

Muscle specific antibody microarray

The present invention also relates to an antibody microarray which could be used to detect specifically variations in the muscle secretome.

Typically said antibody microarray comprises a solid support with a plurality of antibodies immobilized on the solid support, wherein said plurality of antibodies is able to bind to at least 30, 40, 60, 70, 80, 90, 100, 110, 120, 130, or 140 different proteins listed in one of the tables Tl, T2, and T3.

Typically said antibody microarray comprises a solid support with a plurality of antibodies immobilized on the solid support, wherein said plurality of antibodies is able to bind to at least 30, 100, 200, 300, 400, or 443 different proteins listed in T4. In particular, said plurality of antibodies is able to bind to at least 45, 148, 143 and 443 different proteins selected from tables Tl, T2, T3 and T4 respectively. Table Tl may be used to detect modifications in muscle development and formation. Table T2 may be used to detect molecules involved in the muscle extracellular matrix as well as the detection and quantification of fibrosis in disease situations. Table T3 may be used to measure deregulation in muscle homeostasis. Table T4 may be used to assess the whole muscle secretome.

In particular, said plurality of antibodies is able to bind to all the proteins listed one of the tables Tl , T2, T3 or T4.

Antibody microarray is a well known technology, for producing an antibody microarray of the invention, the skilled person may use for example the teaching of WO 2007/130549 or WO0127611. In the following, the invention will be illustrated by means of the following examples and figures.

FIGURE LEGENDS

Figure 1 : Characterization of the secretome of differentiating human myoblasts

(a) Myoblast differentiation was assessed by immunocytochemistry using anti-myosin heavy chain (MF20, green); nuclei were counterstained with Hoechst (blue). Scale bar, 100 μπι. By 72 hours (T=72H), large branched multinucleated (MF20)-positive myotubes were formed. The efficiency of myoblast differentiation was confirmed by Western Blot analysis of the myogenic factors MyoD and Myogenin. Emerin, a ubiquitous nuclear envelope protein, was used as loading control, (b) The mRNA expression profile from muscle cultures after 72h of differentiation was analysed using expression microarrays. After stringent computational sorting an in silico virtual secretome was generated, containing 393 gene products. This virtual secretome was compared with the "experimental secretome" of differentiating human myoblasts obtained by analysing the culture medium by combined proteomics strategies. Among the 965 proteins thus identified, computational analysis classified 257 proteins (27%) as classical or leaderless secreted molecules. The remaining unclassified proteins (708; 73%) were of various intracellular origins, (c) To understand how "soluble" secreted proteins might be related to specialized cellular functions and/or processes, we performed detailed functional analysis in silico. (d) The functional analysis in (d) revealed selective enrichment of 45 soluble secreted proteins crucial for myogenesis and myoblast differentiation. Values represent the -Log (p-value). The p-value was calculated after Right-tailed Fisher's exact test; only significantly overrepresented categories (p<0.05) are shown.

Figure 2: Potential extracellular pathways and/or networks involved in muscle differentiation.

A protein-protein interaction network was generated using Ingenuity Pathway Analysis (IPA) on 45 extracellular proteins crucial for myogenesis and myoblast differentiation following serum starvation (see Table S6 for protein descriptions, HUGO gene symbols and methods). The intensity of "light grey" and "dark grey" indicates the degree of down or upregulation respectively (based on expression data from MAP assay or immunoblot). A solid line indicates a direct interaction while a dashed line indicates an indirect interaction.

Figure 3: Differentiated human muscle cells secrete 2 distinct types of microvesicles. (a) Electron micrographs of the 20,000 and 100,000 g pellets from the muscle secretome showing polymorph vesicles (microparticles) and smaller vesicles of the characteristic size and shape of exosomes (scale bar: 200 nm). (b) 393 (41%) of the intracellular muscle-derived proteins identified in the secretome were also identified in the proteome of muscle-derived microvesicles (MVs). Among unclassified proteins, 46 (4%) were predicted as putatively soluble/leaderless secreted (SecretomeP). The dashed line represents 434 proteins previously identified in MVs produced by other cells; (c, d) Characterization of the proteomes of the two MV subtypes (c) 901 non-redundant proteins were identified by LC-MS/MS. Inset Venn diagram shows proteins identified in exosome and microparticles and the overlap between fractions. To further assess selective enrichment, the relative abundance of proteins in the two MV fractions was evaluated by MS/MS spectral count; fold change comparison of the 2 proteomes are shown, (d) The two protein profiles have distinctive distributions across cellular compartments, reflecting their intracellular origins and cargos (Top panel). Data were presented as a percentage of total proteins in each sub-fraction. In addition, to investigate the functional differences between the two types of vesicles (Bottom panel), we assessed exosomal and MPs proteins for overrepresented molecular and cellular functions. Values represent the -Log (p-value); only significantly overrepresented categories (p<0.05) are shown. Uncoloured histograms illustrate analyses performed on the entire proteome of exosome or MPs whereas coloured histograms are based on the comparison of enriched fractions, (e) Schematic representation of exosome formation and release showing a protein subset enriched in muscle-derived exosomes. It includes proteins involved in multivesicular endosomes (MVE) biogenesis, highlighted in dark grey, as well as tetraspanin surface proteins, features of endocytic organelles. These data further confirm that differentiating human myoblasts secrete bona fide exosomes. (f) Ultrastructural analysis (EM on ultrathin cryosections of differentiating myoblast cultures) revealed the presence of MVE. MVE were observed in close vicinity to the cell surface (Scale bars Ιμπι; Inset lOOnm).

Figure 4: Identification and characterization of the muscle-derived microvesicles RNA cargo.

(a) Bioanalysis of nucleic acids isolated from MVs and myotubes revealed two prominent peaks corresponding to the 18S and 28S ribosomal RNA. When intact exosomes were subjected to RNase treatment the RNA showed no difference in quality when compared to untreated exosomes. When the same procedure was applied to isolated MPs, the overall RNA plot showed reduced peak height. RNase treatment on isolated RNA confirmed complete digestion, (b, c) Illumina microarrays were used to identify mRNA from muscle-derived exosomes, microparticles and myotubes. This analysis confirmed the presence of 185 core transcripts in exosomes and 4431 transcripts in MPs (2% and 46% of the number detected in the donor cells, respectively), (b) Scatterplots of microvesicular and cellular mRNA expression levels. The x- and y-axes represent probe intensity (log2). Gene expression signals in independent preparations were closely correlated (left); very poor correlations were found of averaged RNA profiles compared between sample types (right). The most abundant transcripts in both subtypes of MVs were distinct from the abundant transcripts in myotubes. (c) Hierarchical clustering of microarray data showed that the mRNA cargos of exosomes and MPs are more closely related to each other than to the expression pattern of myotubes. (d) Computational analysis clustered exosomal transcripts into 7 functional groups (Zinc finger C2H2-type, Integral to membrane Ionic channel, Transcription, Protein Secretion, Ion Transport, Signal Transduction. Abbreviations correspond to HUGO gene symbols.

Figure 5: Muscle-derived microvesicles can dock, fuse and deliver functional protein to target cells.

(a) To investigate whether myoblast-derived MVs could enter target cells, exosomes and microparticles labelled with the lipid fluorochrome PKH67 were incubated with differentiating muscle cultures overnight. Treated cells were assayed for PKH67 (green) and the expression of the plasma membrane protein NCAM1 (CD56) (Red) by confocal microscopy. Nuclei were visualized with Hoechst (blue). Both types of vesicles were internalized as shown by the presence of labelled lipid inside the cytosol of target cells. High resolution images are shown on the right panels. Scale bars, 10 μπι. (b) Representative frames from a time-lapse recording of cultures (72h after switch to serum-free medium) exposed to MPs or exosomes labelled with PKH67 for 25h. Labelled MVs were internalized within muscle cells with different kinetics. Paired fluorescence (top) and phase-contrast (bottom) images are shown, (c) MVs as well as their donor myotubes express alkaline phosphatase (purple). Human dermal fibroblasts were incubated with exosomes or MPs for 48 hours, resulting in the appearance of alkaline phosphatase activity in the cytosol. Scale bar, 50 μπι.

Figure 6. Experimental strategy for defining the "secretome" of differentiating human myoblasts: Interrelating proteomic and genomics data

(a) We attempted to characterize the complete potential secretome by defining the mRNA profile of differentiating primary myotubes and applying stringent computational sieving to identify soluble secreted proteins. We analysed mRNA from muscle cultures after 72hr of differentiation using HumanWG-6 genome-wide expression arrays (Illumina). The array experiment was performed on three independent RNA preparations. The raw data from these 3 experiments were quantile normalized and analyzed using GeneSpring software (Agilent Technologies). The mean signal values (AVG_Signal) and the present/marginal/absent calls (Detection Pval) were computed for all 48,803 probes. Data were filtered to remove genes annotated as "non-coding" or "uncharacterized cDNA clones" and/or with Detection Pval > 0.01 and AVG_Signal < 140 [x 2 the background signal on the arrays]; 9212 probe sets remained. Starting with this database of gene products we applied stringent computational sieving to generate a comprehensive list of potentially secreted gene products (detailed workflow provided in Fig. 7). Out of the 9212 gene transcripts, 393 were identified as strong candidates for being secreted into the extracellular space via either classical or unconventional pathways which allowed us to generate an in silico virtual secretome (results in Table 1). (b) For a more direct detection and quantification of proteins secreted (secretome) by muscle cells, we have used a triple proteomic strategy: i) multiplexed immunoassay (MAP); ii) 2D gel electrophoresis coupled to mass spectrometry (2DE-MS); iii) gel-free tandem mass spectrometry (LC-MS/MS). Combining the results from these different approaches resulted in identification of 965 proteins after redundancy reduction. Each of the proteomic strategies employed contributed to this total with no single strategy being able to identify all of the proteins. The schematic illustrates the analytical synergy of the 3 approaches and shows relative numbers of proteins identified by each. While classical 2DE allowed detection of highly expressed proteins and assessed the general extent of potential post-translational modifications, LC-MS/MS is an extremely powerful tool to identify a large number of proteins simultaneously over a wide dynamic range. However low abundance and low molecular weight proteins, such as cytokines and growth factors, overlooked by both 2DE and LC-MS/MS are detected by antibody array. The 965 identified proteins were further submitted to a similar computational method as for our transcriptome dataset for predicting secretory mechanisms.

Figure 7. Computational workflow for predicting secretory pathways

In order to identify the gene products/proteins most likely to be secreted via the classical or ER/Golgi-dependent secretory machinery, sequential computational sieving steps were performed, using signal peptide prediction, recognition of trans-membrane domains, and protein annotation analysis as previously described with modifications (Bortoluzzi, Scannapieco et al. 2006). Predictions of signal peptides and signal peptidase cleavage sites were made by SignalP 3.0 (Emanuelsson, Brunak et al. 2007). Once these were selected, a filter was applied to remove those that contained more than one transmembrane (TM) segment in their protein sequences predicted by TMHMM 2.0 (Krogh, Larsson et al. 2001). As a refinement, Phobius predictor was used to sort protein sequences with only one TM region from those whose single TM domain overlapped with the signal peptide (Kail, Krogh et al. 2004). Subsequently, putative classically secreted proteins (Signal Peptide positive/Transmembrane domain negative) were scanned for the presence of an ER retention signal. Those containing in their sequence the extended KDEL motif were discarded (PROSITE PS00014/ER_TARGET) (Hulo, Bairoch et al. 2008). Finally, for each remaining protein, the annotation information of the corresponding entry in the GO annotation (Ashburner, Ball et al. 2000), Uniprot (http ://www.uniprot. org) and Ingenuity Knowledge base (Ingenuity Systems®, www.ingenuity.com ) was taken into account for identification of those proteins, selected by previous sieving steps, with known intracellular roles and/or localizations. If no signal peptide was detected by SignallP, SecretomeP 2.0 was used to produce predictions of non-classical i.e. not signal peptide triggered protein secretion (Bendtsen, Jensen et al. 2004). When the Neural Network score exceeded 0.6, proteins were classified as putatively secreted via the non classical pathway. This sequence-based method remains predictive, therefore, prior to the final secretion prediction, the gene products/proteins were further examined for extracellular localisation by taking into account the annotation information from various sources (as described above).

Figure 8. Analysis of conditioned media during human primary myoblast differentiation in vitro using antibody-based assays.

(a) Conditioned media (CM) from differentiating cultures were screened for the expression of known cytokines and growth factors by antibody arrays (Human MAP assay, Rules based Medicine). Among 89 factors screened, 38 markers were detected in the myoblast secretome. The heat map shows their dynamics of secretion during the first hours of differentiation (0, 24, 48 and 72h). High levels of secretion are displayed in dark grey; whereas low levels are displayed in light grey. Alpha fetoprotein (AFP), Tissue factor (F3), Glutathione S-transferase Alpha 1 (GSTA1), Mucin-16 (MUC16) and TNFa (TNF) remained at steady levels at all time-points. However, almost all the other markers showed levels which increased progressively from 24 to 72 hours with the important exception of IGF-I (IGF1) which decreased steadily from initially high levels, (b) The MAP assay panel as originally designed for biomarker profiling in the blood is missing several crucial proteins involved in myoblast differentiation such as IGF-II (IGF2) and galectin-1 (LGALS1). Therefore, the time courses of secretion of IGF-II (IGF2) and galectin-1 (LGALS1) as well as two other secreted proteins (MIF and SPARC) were examined by Western blot analysis (15 μΐ concentrated CM per lane). The relative amount of each secreted molecules in CM is shown below (Values are means densitometric arbitrary units (AU) + SD (n=3)). From 24 to 72h, IGF2, LGALS1 and MIF accumulated gradually in the CM of human myoblast differentiating cultures, whereas SPARC showed the reverse secretion pattern with greatest level at 24h but remained detectable at 72h. LGALS1 and SPARC were detected with the 2 others proteomics screens performed on CM collected after 72h of differentiation, whereas MIF was among the few proteins only detected by 2DE and IGF2 identified by HPLC- ESI-MS/MS with only one unique high score peptide.

Figure 9. Putative Extracellular Matrix (ECM) Network.

This network was constructed with 91 "soluble" secreted proteins annotated previously as components of the ECM using Ingenuity Pathway Analysis (IPA 5.0, http://www.ingenuity.com/). Through meticulous review of the literature related to skeletal muscle ECM (Sanes 1994; Grounds 2008) and annotation information derived from various sources (GO annotation, Uniprot and Ingenuity Knowledge base), we carefully constructed a simplified protein network based on known physical protein- protein interactions, modelling the ECM secreted by differentiating human myoblasts. Four distinct interconnected sub-networks could be visualised in the ECM "interact ome": "fibrillar matrix", "basement membrane", "elastic fibers" and "proteases and inhibitors" (Figure 5B), mimicking the highly organised interstitial connective tissue surrounding individual muscle fibers in vivo. Protein abbreviations correspond to HUGO gene symbols and are reported in Table S5 with detailed information of the associated proteins. The intensity of light grey and dark grey indicates the degree of down or upregulation respectively (if expression data was available from MAP assay or immunoblot). A solid line indicates a direct interaction while a dashed line indicates an indirect interaction.

Figure 10. Functional classification of the human myoblast "soluble" secretome.

203/257 "soluble" secreted proteins were successfully assigned into one or more biological functions as determined by IPA (a) Eighteen molecular/cellular functions were over-represented. The x axis shows the negative log of p value and the numbers on the bars indicate number of proteins, (b) Among those over-represented functions, the top 5 could be related to the development and function of various physiological systems (e.g. general cellular and metabolic processes, immune response and hematopoiesis, vascularization, skeletal system and connective tissue, innervation, skeletal muscle and myogenesis). See Table S5 for a detailed listing of identified "soluble" secreted muscle protein species and their corresponding functional annotation.

Figure 11. Muscle-derived microvesicles characterization

(a) Western Blot analysis of microvesicles isolated from CM of human differentiating myoblasts. Whole-cell lysates (CP), culture medium (CM) and Microparticles (MPs) or Exosomes (Exo) preparations (10 μg proteins per lane) were separated on SDS-PAGE and electroblotted. Blots were probed with antibodies against CD81, CD82, CD63, CD9, CLIO, LGALS1, HSPA8 (Hsp70). Protein abundance as detected by Western Blot correlated with exosome and MP MS/MS spectral count data, (b) lodixanol gradient analysis of the 100,000 g pellet (exosomal fraction). Following separation of the pellet on an lodixanol gradient, twelve successive fractions of increasing densities were collected. The fractions containing exosomes were identified by immunoblotting for the expression of the indicated tetraspanin, molecular signature of secreted exosomes It can be seen that muscle derived exosomes float at a density of 1.11-1.14 g/ml as shown by the presence of CD9, CD81 and CD63. EXAMPLES

Abstract

Coordinated interplay between multiple cell types is critical for efficient muscle regeneration. This cross-talk involves direct cell-to-cell interaction and secreted signalling molecules acting in an auto/paracrine fashion. Here we show that differentiating human muscle cells in vitro release not only soluble secreted proteins through conventional secretory mechanisms but also complex protein/nucleic acid cargos via membrane microvesicle shedding. The soluble "secretome" contains 254 proteins/peptides with signalling potential including 45 previously implicated in myogenesis, while others are known to modulate various cell types, implying a much broader role for myoblasts in muscle remodelling in vivo. We also identified and characterised two types of plasma membrane-derived vesicles: nano vesicles harbouring typical exosomal features and larger, morphologically distinct, microvesicles. We show that these vesicles differ in their protein and RNA content, differentially dock and fuse with adjacent cells, and demonstrate delivery of a functional protein cargo to target cells following microvesicle uptake. Thus, the intercellular signalling networks invoked during in vivo muscle regeneration may employ soluble signalling molecules acting in concert with muscle-derived microvesicles delivering protein/RNA cargo to target cells Results/Discussion

Human satellite cells were isolated and expanded in vitro (Edom, Mouly et al. 1994). Cells differentiated to form multinucleated myotubes after 72 hours in serum-free culture conditions (Fig. la). The mRNA expression profile of differentiated myotubes was used to construct a "virtual secretome" (Fig. 6-7). After stringent computational sieving, out of the 9212 gene transcripts detected in the cells, 393 were identified as strong candidates to be secreted via either the "classical" ER/Golgi-dependent secretory machinery (Signal Peptide positive/Transmembrane domain negative/ ER retention signal negative) or via alternative signal peptide-independent mechanisms (unconventional/non-classical secretion) (Nickel and Rabouille 2009) (Table SI). This list of candidates represents the complete secreting potential of the cells, but not necessarily what is actually secreted into the extracellular space. To identify the proteins actively secreted during myotube formation and hence define the human muscle secretome, we analysed the conditioned medium of differentiated cultures by combining three powerful and complementary proteomic strategies (Fig. 6): multi-analyte Luminex based immunoassay (MAP; Fig. 8 and Table S2), 2D gel electrophoresis (2DGE) and gel-free tandem mass spectrometry (HPLC-ESI-MS/MS). Nine hundred and sixty five non-redundant proteins were identified, out of which 569 (59%) were newly described (Table S5), whereas 396 (41%) had been detected in previous rodent secretome studies (Chan, McDermott et al. 2007; Yoon, Yea et al. 2008; Henningsen, Rigbolt et al. 2010). These proteins were investigated through the same computational workflow to predict their status as soluble secreted proteins (Fig. 7). Only 257 of the 965 contained motifs indicative of secretion via either classical or unconventional pathways (Fig. lb and Table S6), out of which 207 corresponded to those predicted by the in silico analysis of transcripts. A large number of these secreted proteins (91) were structural and regulatory components of the extracellular matrix (ECM; Fig. 9), suggesting that human primary myoblasts participate to the formation and remodelling of their own extracellular matrix during differentiation. Based on ontology descriptions and statistical enrichment analyses (Ingenuity Knowledge base), the 257 soluble secreted proteins are involved in diverse molecular/cellular processes. Notably, 107 proteins had reported roles in skeletal muscle homeostasis, myogenesis or disease (Table S6), with 45 being crucial for myogenesis (Fig. lc,d). To determine how these 45 secreted proteins might synergise and contribute to myoblast differentiation, a putative muscle differentiation interactome model was generated after seeding with key muscle markers such as the myogenic regulatory factors Myogenin (MYOG) and MYOD1. Figure 2 shows these proteins coalesce into networks based on "differentiation", "matrix remodelling & migration" and "fusion", all events essential to myodifferentiation. Important factors include IL6 which mediates hypertrophic muscle growth by controlling satellite cell recruitment and fusion (Serrano, Baeza-Raja et al. 2008), LGALSl (galectin-1) enhancing myoblast fusion (Watt, Jones et al. 2004), MMP2 essential for matrix remodelling during muscle growth and regeneration (Yagami-Hiromasa, Sato et al. 1995) and members of the TGF- β family and insulin-like growth factors (TGFB1, Myostatin (MSTN), IGF1, IGF2) which will either inhibit or promote muscle differentiation and hypertrophy (Langley, Thomas et al. 2002; Jacquemin, Butler-Browne et al. 2007). The remaining 150 secretory proteins have not been described in a muscle context and either had unassigned functions (54) or were associated with housekeeping (43) or specific processes such as the immune response, vascularisation, connective tissue and innervation (54) (Fig. Id; Fig 10, Table S6). Examples include VEGF-C and Placenta growth factor (PGF) which act on blood vessels (Roy, Bhardwaj et al. 2006), GDF15 on inflammatory cells (Bootcov, Bauskin et al. 1997), and CTGF on neighbouring fibroblasts (Leask and Abraham 2004). Therefore these secreted proteins potentially interact with neighbouring muscle and non-muscle cells playing a crucial role in orchestrating muscle differentiation and remodelling.

As observed in figure lc, in addition to the 257 extracellular secreted proteins, 708 intracellular proteins from various origins (including membrane, cytosol, Golgi/ER, mitochondria, nucleus) were found in the secretome of differentiating myoblasts. We hypothesized that intracellular proteins are transported outside the cell as components of secreted microvesicles (Cocucci, Racchetti et al. 2009; Thery, Ostrowski et al. 2009). Data mining revealed that 434 of our muscle secretome proteins (46%) have been previously found in microvesicles released from a broad range of cell types (references in Table S5). So far, two major mechanisms of microvesicle secretion have been described, each leading to the release of distinct cargo-loaded vesicles into the extracellular space: (i) microparticles (or shedding vesicles) generated from the direct budding of the plasma membrane; and (ii) exosomes, nanovesicles of endocytic origin released into the extracellular environment upon fusion of multivesicular endosomes with the plasma membrane. To further investigate the hypothesis that differentiating muscle cultures produce protein-laden microvesicles, culture supernatants were subjected to differential centrifugation (Miguet, Pacaud et al. 2006; Thery, Ostrowski et al. 2009). Two distinct populations of small bilayer membrane vesicles were isolated (Fig. 3a). The first were cup-shaped vesicles, 69 (±20) nm in diameter, sedimenting at 100,000 g, selectively enriched for tetraspanin surface proteins (CD81, CD82, CD63 and CD9), Hsp70 (HSPA8) , with a buoyant density between 1.11 and 1.14 g/ml, all features characteristic of exosomes (Fig. l la,b). The second population were polymorph microvesicles with electron dense cores, 80 to 290 nm in diameter, which pelleted at 20,000 g and were tetraspanin-negative except for CD81, but enriched in CLIC1 and galectin-1 (LGALS1) (Fig. 11a), and will hereafter be called microparticles (MPs). After differential purification, the specific protein cargo of these two types of vesicles was analyzed using the same mass spectrometry-driven proteomics strategy employed for secretome mapping (i.e. HPLC-ESI-MS/MS). A total of 764 unique proteins were identified from MPs and 564 from exosomes (Fig. 3c and Table S7-S8). This proteomic analysis confirms that 393 (41%) of the proteins within the muscle secretome are transported in association with secreted microvesicles (Fig. 3b and Table S5). Comparison of the exosome and microparticle proteome datasets by peptide counting (Liu, Sadygov et al. 2004) revealed that, although they share a common set of proteins, a significant proportion (65%) differs reflecting their distinct intracellular origins and different cargos (Fig. 3c; Table S9). Objective computational sieving and database mining (Ingenuity and UniProtKB) shows that MPs are preferentially enriched in proteins normally found within the ER, mitochondria, ER-Golgi, cytoskeleton, cytosol (e.g. ER resident proteins, (CALU, CALR, HSP90B1, HSPA5), TCP-1 chaperonins (TCP1, CCT2), actin and tubulins (ACTB, TUBA1B), ribosomal subunits (RPL4, RPL10, RPS5, RPS17), translation initiation factors (EIF3A, EIF5), poly(A) binding protein (PABPC1), and proteasome sub-components (PSMA1, PSMB1)). Conversely, proteins from the plasma membrane, sub-plasma membrane, endosome and lysosome are enriched within exosomes (e.g. Integrins (ITGA4, ITGA6, ITGA7), MHC molecules (HLA-A HLA-B) and tetraspanins (CD 9, CD63, CD81, CD82), flotillin-1 (FLOT1), Alix (PDCD6IP), TSGlOl, lysosome-associated protein (LAMP2), Fig. 3d, top panel and Table S9). These findings are consistent with the endosomal origin of exosomes and their involvement in receptor- mediated endocytosis (Pan and Johnstone 1983). More detailed assessment of the specific protein cargo of muscle-derived exosomes indicated the presence of signature proteins previously described for exosomes (Mathivanan and Simpson 2009; Thery, Ostrowski et al. 2009) (highlighted in Table S8-S9) such as the protein complex ESCRT machinery, which is required for both formation of multivesicular endosomes (MVE) (Fig. 3e) and the recruitment of their endosome- derived cargo proteins. Ultrastructural examination via transmission electron microscopy of differentiating cultures shows that MVE are present within myotubes from which microvesicles are derived (Fig. 3f, top panel), and are observed in close vicinity to the cell surface, suggesting their direct fusion with the plasmalemma prior to release in the extracellular environment (Fig. 3f, bottom panel). As mentioned above, exosomes show enrichment of integral membrane proteins and subplasmalemmal proteins when compared to MPs (Fig. 3d top panel). Analysis of the molecular and cellular functions of these proteins established an involvement in cell-to-cell signalling (Fig. 3d, bottom panel). Conversely, proteins identified in MPs principally involve RNA- post- translational modification, amino acid metabolism, protein synthesis, folding, post- translational modification and trafficking as well as molecular transport and protein degradation. The distinct functions of the proteins contained in each type of vesicle reinforce their potentially specific role.

There is increasing evidence that secreted membrane vesicles also carry selected RNAs, raising the possibility of transfer of genetic material to neighbouring cells (Ratajczak, Miekus et al. 2006; Valadi, Ekstrom et al. 2007; Skog, Wurdinger et al. 2008). Both exosomes and MPs from myotubes contained RNAs as evidenced by the presence of 18S and 28S ribosomal RNA peaks. These RNAs were afforded protection from RNase action by the microvesicle membrane, particularly in exosomes, while RNA adhering to the microvesicle exterior was completely destroyed (Fig. 4 a). Given the different subcellular origins and protein cargoes of exosomes and MPs, we predicted the RNA cargo would also be different. Hierarchical clustering of microarray data indicated that the transcriptome of myotubes, MPs and exosomes were indeed separable into distinct groups (Fig. 4c). Further analysis identified 185 core transcripts in exosomes, whereas MPs contained 4431 transcripts. The abundance of 185 exosome transcripts correlated with that in myotubes, suggesting they are a subset of myotube transcripts - although these were not simply the most abundant myotube transcripts. In MPs, a group of 101 transcripts showed enrichment in MPs relative to myotubes (more than 3 fold; p<0.01), whilst 2366 were less abundant than in myotubes (Fig. 4b). Given that the microparticle preparation was potentially contaminated with cellular RNA (Fig. 4a,b), we restricted subsequent bioinformatic analyses to the exosome transcripts. Clustering of functional annotation indicated that exosomal mRNA transcripts were selectively enriched in two main categories namely, transcripts encoding membrane proteins, predominantly receptors and ionic channels, and regulators of transcription, in particular C2H2-type zinc-finger transcription factors (Fig. 4d). Therefore, if muscle exosomes were to be taken up by recipient cells, and their RNA cargo translated, they would likely induce dramatic changes in intracellular signalling cascades. Microvesicles have been implicated in paracrine and autocrine signalling in many cell types (Ratajczak, Miekus et al. 2006; Valadi, Ekstrom et al. 2007; Al-Nedawi, Meehan et al. 2008) and may prove to be as important as the classical soluble secretome in cell- to-cell signalling. We therefore tested whether myotubes were capable of taking up exosomes and MPs. Purified exosomes and MPs labelled with the lipid fluorochrome PKH67 were incubated overnight with differentiated human cultures. We observed that both types of vesicles can dock, fuse and are internalized into target cells, as evidenced by the detection of labelled lipid components inside these cells by confocal microscopy (Fig. 5a). MPs were internalized into both myotubes and monucleated cells whereas exosomes were mainly found in mononucleated cells. To determine whether this result could be explained by a difference in internalization kinetics, live imaging of the uptake was carried out. As shown in figure 5b, after 5 hours of incubation, exosomes were already fused whereas MPs were detected intracellulary only after 10 hours incubation. Exosomes were also internalized preferentially into mononucleated cells since a signal could be detected in myotubes only after 25 hours of incubation. In contrast, MPs were internalized by both myotubes and mononucleated cells. To determine whether microvesicle uptake is associated with transfer of a functional cargo into recipient cells, we investigated the activity of alkaline phosphatase, one of the proteins identified by our proteomic screen in both exosomes and MPs, in dermofibroblasts with no endogenous activity. Alkaline phosphatase activity was detected after 48 hrs of incubation with purified exosomes or MPs (Fig. 5c). The expression of muscle-derived microvesicle delivered alkaline phosphatase by fibroblasts is powerful evidence that the protein cargo is functional after entry into the cytoplasm of recipient cells.

Conclusion In conclusion, muscle regeneration is a highly synchronized process that involves a multitude of coordinated cellular responses, such as inflammation, neo-vascularisation, muscle differentiation, innervation, and requires precise cell-to-cell signalling. Deregulation of any component of the process will lead to impaired regeneration. We propose that human myoblasts use a combination of soluble secreted proteins as well as secreted microvesicles to regulate the behaviour of neighbouring cells during muscle regeneration and orchestrate organogenesis. The secreted proteins encompass classical signalling factors known to act via cell surface receptors, whilst microvesicles allow transport of molecules that can act directly intracellularly once these vesicles are internalized. We report here for the first time the existence of 2 types of vesicles released by differentiating myoblasts: microparticles and exosomes. While they share some elements in common, their specific protein and RNA cargo suggests new specific roles for these particles. The intercellular microvesicle-mediated transfer of active alkaline phosphatase is powerful evidence that the protein cargo is functional. In addition, we hypothesise that these two types of microvesicles signal differentially to target cells since they use distinct kinetics to penetrate these cells and upload their specific cargos. Horizontal transfer of RNA and protein delivery by release and uptake of microvesicles is becoming well established with evidence that transferred RNA is functional and translated (Ratajczak, Miekus et al. 2006; Valadi, Ekstrom et al. 2007; Skog, Wurdinger et al. 2008). Therefore, it seems highly probable that the combination of the injected microvesicle protein cargo and the translation of their RNA cargo into functional proteins could be sufficient to alter significantly the phenotype of a recipient cell. We hypothesise that perturbations in gene expression downstream of the impact of soluble agonist binding to cell surface receptors may be enhanced by proteins and RNAs derived from the internalisation of microvesicles acting directly within the recipient cell's cytoplasm. If true this opens up an entirely new field for system biology applied to muscle development and repair and the potential to introduce new proteins into cells.

METHODS

Human skeletal muscle culture. Human satellite cells were isolated as described previously (Edom, Mouly et al. 1994) in accordance with French legislation on ethics. Cells were expanded in growth medium (Ham's F10, 20% foetal calf serum (FCS) and 5μg/ml gentamycin (Invitrogen, Paisley, UK) in 5% C0₂, at 37°C. In all experiments, myogenicity was greater than 90% as assessed by the expression of desmin localized by immunostaining (D33; DAKO, Glostrup, DK). After six washes to remove contaminating serum proteins, confluent cultures were switched to a serum-free Dulbecco's modified Eagle's medium (DMEM, Invitrogen) which triggered differentiation. Cell viability in serum-free medium was > 95% as assessed by trypan blue exclusion test (Sigma-Aldrich). The differentiation potential was determined by immunostaining with myosin heavy-chain (MF20). Briefly, cells were fixed in 4% (w/v) paraformaldehyde (PAF) and then incubated with an anti-desmin antibody (D33; DAKO, Glostrup, DK) and/or anti-MF20 (DSHB, Iowa City, IA, USA). Specific antibody binding was revealed using Alexa Fluor® 488 (Invitrogen) or Cy™3 (Jackson Immunoresearch, West Grove, PA, USA) directly coupled to the secondary antibody. Nuclei were counterstained with Hoechst 33258 (Sigma-Aldrich, St. Louis, MO). Images were acquired using an Olympus BX60 fluorescence microscope (Olympus Optical, Hamburg, Germany).

Collection of conditioned media and isolation of muscle-derived vesicles.

Conditioned media (CM) were collected after 24h, 48h, and 72h incubation in serum free medium. The CM was cleared by centrifugation (10 min at 300 g followed by 20 min at 2000 g) and spin concentrated using an Amicon Ultra-4 5kD cut-off spin Column (Millipore, Watford, UK). For LC-MS/MS analysis and 2D-PAGE, concentrated CM samples from 72h differentiated cultures were cleaned up for salt and other contaminant by precipitation using acetone/ethanol. Secreted microvesicles (MVs) were recovered from cleared supernatants by differential centrifugation as previously described with slight modifications (Miguet, Pacaud et al. 2006; Thery, Amigorena et al. 2006). Briefly, microparticles were sedimented by centrifugation at 20,000 g for 70 min at 4°C.The remaining supernatant was further ultracentrifuged at 100,000 g for 70 min at 4°C to pellet the exosomes. Finally, purified vesicles were washed twice in PBS and fixed in 2% PAF prior to electron microscopy or resuspended in PBS. For the density gradient experiment, isolated exosomes were floated in a iodixanol gradient as previously described (Mathivanan, Lim et al. 2010).

Western blot analysis.

Protein extracts were fractioned by SDS PAGE using 4-12% acrylamide Bis-Tris precast gels with a MES buffer system according to the manufacturer's protocol (NuPAGE Novex Bis-Tris gel; Invitrogen) before transfer to PVDF membranes. Membranes were blocked and then incubated with the following primary antibodies: anti- Myogenin (F5D; DSHB); anti-MyoD (5.8A; BD Pharmingen, Franklin Lakes, NJ, USA); anti-Emerin (Novocastra, Ltd., Newcastle, UK), anti-LGALSl (Fitzgerald, Concord, MA, USA), anti-IGF2 (Abeam, Cambridge, UK); anti-MIF (R&D Systems, Inc., Minneapolis, MN, USA); anti-SPARC (Haematologic Technologies, Inc, Essex Junction, VT, USA); CD81 (BD Pharmingen), CD82 (Santa Cruz Biotechnology), CD63 (BD Pharmingen), CD9 (Santa Cruz Biotechnology, Inc., Heidelberg, Germany), anti- CLIC1 (Abnova Corp., Heidelberg, Germany), anti-HSPA8 (Stressgen Biotechnologies Corp., San Diego, CA, USA). Binding of a secondary antibody coupled to HRP was revealed using a chemiluminescence detection reagent (Pierce Biotechnology, Rockford, IL, USA). Signals were detected on film and quantified by densitometry using Quantity One software v. 4.6.5 (Bio-Rad, Hercules, CA, USA). Human multi-analyte profile - Luminex assays. Unconcentrated CMs from differentiating myoblasts were screened for cytokines and growth factors using Luminex multi-analyte profiling (MAP) technology (Oliver, Kettman et al. 1998). The assays (Human MAP® v. 1.6) were performed by Rules Based Medicine (Houston, TX, USA). Heat-maps for visualization of expression data were produced using FIRe v. 2.2 (Garcion, Baltensperger et al. 2006).

Gel-free nano-flow LC-MS/MS Analysis. Precipitated CM and microvesicle pellets were dissolved in urea-buffer and proteins were reduced, alkylated and subsequently digested with endoproteinase Lys-C (Wako, Neuss, Germany) followed by digestion with trypsin (Promega, Southampton, UK). The resulting peptide mixture was fractionated using an ICAT® Cation Exchange Buffer Pack (Applied Biosystems, CA, USA) following the manufacturer's instructions.Bound peptides were eluted by washing the column progressively with elution buffers of increasing ionic strength. Fractions were dried and re-dissolved in 0.1% trifluoroacetic acid. Prior to mass spectrometry, individual SCX-fractions were desalted using modified StageTips (Gobom, Nordhoff et al. 1999; Jensen and Larsen 2007). The desalted fractions were injected into a LTQ- Orbitrap XL (Thermo Electron, Bremen, Germany) using an EASY-nLC (Proxeon A/S, Odense, Denmark). The analysis was repeated three times. The LTQ-Orbitrap raw files were processed for peak list generation using DTASuperCharge (http://msquant.sourceforge.net/). The resulting mgf-files were merged into one file and submitted to X!Tandem TORNADO (2008.02.01.3) (Global Proteome Machine Database; http://human.thegpm.org/tandem/thegpm tandem.html) (Craig, Cortens et al. 2004) for database searching against the Swiss Prot database restricted to the human taxonomy (Swiss Prot 55, Taxonomy: Homo sapiens (human) (17759 sequences)). For bulk protein identifications, a minimum of 2 unique peptides (with significant peptide score log(e)<-2) was required. Identifications based on 1 single unique peptide presenting an acceptable peptide score log(e)<-2 were manually validated according to the rules for validation of peptide candidates of doubly charged ions (Chen, Kwon et al. 2005). Search results were added to gpmDB so they could be readily accessible using look up model function and the following model numbers GPM33000032463, GPM33080000001 and GPM33080000002 for the protein profiling of CM, microparticles and exosomes respectively (.http://gpmdb.thegpm.org/). Comparison of the proteomic datasets of exosomes and MPs was performed using BioVenn (Hulsen, de Vlieg et al. 2008). For a relative estimation of protein abundance (RPA) in exosomes and MPs, Scaffold (2.06.05) (Proteome Software, Portland, Oregon, USA) was used to calculate and normalize spectral counts (the number of assigned MS/MS spectra for each identified protein). A protein was considered as selectively enriched in one microvesicle population compared to the other if it (i) was exclusively detected in one sample type or (ii) had a≥ 3 -fold increase in relative protein abundance (RPA) as assessed by the number of redundant peptides (normalized spectral counts).

2-dimensional gel electrophoresis (2DGE) & LC-MS/MS Analysis of muscle cell Conditioned Media. After protein precipitation, concentrated CM pellets were resuspended in re-hydration solution and separated by 2D-PAGE as previously described (Le Bihan, Hou et al. 2006). Proteins were digested in-gel and analyzed by LC-MS/MS using an LCQ ion-trap mass spectrometer (Thermo Electron) as described previously (Le Bihan, Hou et al. 2006). After conversion of the LCQ raw files into mascot generic files (.mgf) using Proteome Discoverer version 1.1 (Thermo), database searching (Swiss Prot 55, Taxonomy: Homo sapiens (human) (17759 sequences)) was performed using X!Tandem TORNADO (2008.02.01.3). For protein identification, a minimum of 2 unique matching peptides (with significant peptide score log(e)<-l) was required.

Computational sieving for predicting secretory pathways and subcellular localisation. Transcripts expressed in 72h differentiated cultures or proteins identified in the CM were submitted to stringent computational analysis for predicting secretory mechanisms. The general workflow is presented in Fig. 7 For the proteins identified in muscle-derived MVs, a different combination of internet-accessible tools was used to predict reliable subcellular localization: SignalP 3.0 (Emanuelsson, Brunak et al. 2007), Phobius (Kail, Krogh et al. 2004), Big-Pi 3.0 (Eisenhaber, Eisenhaber et al. 2003), ER_retention signals (PROSITE PS00014), TargetPl.l (Emanuelsson, Nielsen et al. 2000), WoLF Psort protein localization predictor (Horton, Park et al. 2007). In order to define a classification based on the subcellular compartment in which the protein was primarily found, we also incorporated subcellular location annotations from the comments section in Uniprot (http://www.uniprot.org), GO (Ashburner, Ball et al. 2000 land Ingenuity (www.ingenuity.com.). For both type of analyses, the BioMart (www.biomart.org) (Smedley, Haider et al. 2009) web interface as well as Protein Center 1.4 (Proxeon, Odense, Denmark) software were used to facilitate database queries and annotations. "Membrane vesicle-associated secretome" database. For the proteins that could not be categorized in the classical and nonclassical secretion pathway, we hypothesised that they could be transported outside the cell in microvesicles such as exosomes or microparticles. A database of the "membrane vesicle-associated secretome" was created by combining 30 previously published proteomic catalogues of extracellular microvesicles from various cell types and species (human, rat, mouse). To compare more easily these various experiments, Protein Center 1.4 (Proxeon, Odense, Denmark) was used to retrieve the HUGO gene symbols of each protein described in these studies. Consequently, a list of 2129 Unique gene products related to extracellular microvesicles was obtained. Our unclassified secretome dataset was screened against this "membrane vesicle-associated secretome" database.

Bioinformatics methods for analysis of specialized functions or processes. To further evaluate the specific functional profiles of the various "soluble" secreted or membrane vesicle-associated proteins, we performed detailed functional analysis using Ingenuity Pathway Analysis System (IPA 5.0, http://www.ingenuity.com/ ; Ingenuity Systems®, Redwood, CA). The functional analysis identified the molecular/cellular functions that were most significantly represented in our three datasets. Fischer's exact test was used to calculate a p-value determining the probability that each biological function assigned to the data set is due to chance alone. To complement this functional classification, our dataset of "soluble" secreted proteins was compared against a list of gene products associated with skeletal muscle function and development as annotated in the Ingenuity knowledge base (1125 HUGO gene identifiers). They were also examined against a list of classical ECM molecules generated from Uniprot & GO annotation (Uniprot KW: 272 = ECM, 84 = Basement Membrane, 654 = Proteoglycans, 430 = Lectins, Serpins) (GO ID: GO:0031012 = extracellular matrix, GO:0005201 = extracellular matrix structural constituent, GO:0043259 = Basal Laminae).

Pathway/network Prediction by Ingenuity Pathway Analysis (IPA). In order to construct protein-protein interaction networks relevant to myogenesis from the panel of soluble secreted proteins found in our study we selected, from the functional profile obtained, a group of extracellular proteins implicated previously in myogenesis and muscle differentiation (45 molecules). To this list we added 5 intracellular molecules known to be induced in myoblast cultures upon serum starvation: MYOG = Myogenin, MYOD1, MYH3 = Myosin Heavy Chain, DES = Desmin, ITGB1 = IntegrinBl. A second group of "soluble" secreted molecules was considered for network generation in the context of muscle differentiation: the 91 molecules annotated previously as ECM components. These 2 lists were submitted independently to IPA: using the network/my pathways tool, customized pathways for our selected targets were created by importing network data related to these molecules from the Ingenuity Pathway Knowledge Base. The graphical representation obtained illustrates the molecular relationships between these different gene products. Nodes are displayed using various shapes that represent the functional class of the gene product.

Electron microscopy. Human myotubes plated on plastic (Thermanox, Nalge Nunc, Rochester, NY , USA) coverslips were fixed in 2.5% glutaraldehyde in 0.1M phosphate buffer, pH 7.4 and further post-fixed in 2% Os0₄. They were gradually dehydrated in acetone including a 2% uranyl en-bloc staining step in 70% acetone, and embedded in Epon resin (EMS, Fort Washington, PA, USA). Ultrathin sections were counterstained with uranyl and lead citrate. MVs were processed as described in Thery et al., 2006 (Thery, Amigorena et al. 2006). Observations were made using a CM120 transmission electron microscope (Philips) at 80 kV and images recorded with a Morada digital camera (Olympus Soft Imaging Solutions GmbH, Miinster, Germany).

Isolation and detection of RNA from muscle-derived vesicles. Isolated MVs were resuspended in PBS and incubated at 37°C for 15 minutes with either 100μg/ml RNase (Fermentas, Glen Burnie, MD, USA) or nuclease free water. RNA was extracted using the RNeasy® Micro kit (Qiagen Ltd., Crawley, UK) with on-column DNase according to manufacturer's instructions. Isolated RNA were quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE, USA) and integrity assessed on an Agilent 2100 Bioanalyser (Agilent Technologies, Inc., Santa Clara, CA, USA). RNA aliquots were subsequently incubated at 37°C for 15 minutes with either 100μg/ml RNase (Fermentas, Glen Burnie, MD, USA) or nuclease free water alongside aliquots of cell-derived RNA acting as a positive control; the integrity of treated samples was reassessed using the Agilent 2100 Bioanalyser. Microarray analysis. Total RNA from 3 independent 72h differentiated cell cultures and MVs recovered from duplicate cultures at 72h differentiation were isolated using the RNeasy® kit (Qiagen Ltd.,Crawley, UK) according to manufacturer's instructions. Isolated RNA were quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE, USA) and cellular RNA quality checked on an Agilent 2100 Bioanalyser (Agilent Technologies, Inc., Santa Clara, CA, USA). Samples were labelled using the Illumina TotalPrep RNA Amplification kit (Applied Biosystems, Foster City, CA, USA) then hybridised to Illumina human whole genome BeadChips (cell samples: HumanWG-6 v3; MVs: HumanHT12 v3; Illumina, Inc., San Diego, CA, USA ) according to manufacturer's instructions. The raw data were pre-processed using GenomeStudio (Illumina, Inc, San Diego, CA, USA) to determine mean signal values (AVG_Signal) and detection calls (Detection Pval). Details regarding the generation of the in silico/wktual secretome from the transcriptome of 72h differentiating cultures are provided in Fig. 6 and 7. To assess overrepresentation of functionally related gene and terms in MVs transcripts, functional annotation clustering was performed using DAVID (David Bioinformatics Resources 6.7; http://david.abcc.ncifcrf.gov/) (Huang da, Sherman et al. 2009)). For this enrichment analysis, the databases interrogated were Uniprot, GO, KEGG, Interpro and PFAM. MV transcripts were compared against a population background of all transcripts expressed in 72h myotubes. Internalization of microvesicles by differentiating myoblast culture- PKH67 labelling. Purified vesicles were labelled with PKH67 green fluorescent labelling kit (Sigma-Aldrich) according to the instruction manual. After clean-up through 30 % sucrose cushion (Thery, Amigorena et al. 2006), the labelled MVs (exosomes and microparticles) were added to differentiating myoblast cultures. Treated cells were incubated at 37 °C overnight. To evaluate the uptake of microvesicles by differentiating cells, after 3 washes in PBS, cells were fixed in 4% formaldehyde and assayed for PKH67 and the expression of the plasma membrane protein NCAM1 (CD56) by confocal microscopy (Leica SPE; Leica Microsystems GmbH, Wetzlar, Germany). Nuclei were vizualised with Hoechst 33258 (Sigma-Aldrich). For live-cell imaging, cell cultures were monitored using a Nikon Ti microscope ((Nikon Instruments Inc., Amstelveen, The Netherlands) equipped with an incubator to maintain cultures at 37 °C and 5% C02 (Okolab, Ottaviano, IT), a CoolSNAP HQ 2 camera (Roper Scientific, Tucson, AZ, USA) and an XY motorized stage (Nikon), driven by Metamorph software (Molecular Devices, Sunnyvale, CA, USA). Images were acquired over a 2 day period after the addition of PKH67 labelled MVs.

Protein transfer experiment: alkaline phosphatase assay. Differentiating human myoblasts and dermofibroblasts (isolated from the skin of a 19 year old donor) were incubated with purified exosomes and microparticles for 48h at 37°C. After 3 washes in PBS, cells were fixed in 100% cold methanol and alkaline phosphatase activity revealed using the SIGMAFast™ BCIP^®/NBT ALP kit according to manufacturer instructions. Coverslips were scanned and images acquired using a Leica DMR microscope (Leica Microsystems GmbH).

Statistical analysis. Statistical analyses were performed using SigmaStat software (v. 3.10) (Systat Software, Inc., San Jose, CA,USA). Results are expressed as the means +/-S.D. of at least three experiments. To determine significance between two groups, comparisons were made using the Student's unpaired t-test. Analyses of multiple groups were performed using one-way ANOVA followed by a Tukey post hoc test. For all tests, a P value <0.05 was considered as significant (*P<0.05, and **P<0.01). References

Throughout this application, various references describe the state of the art to which this invention pertains. The disclosures of these references are hereby incorporated by reference into the present disclosure.

Al-Nedawi, K., B. Meehan, et al. (2008). "Intercellular transfer of the oncogenic receptor EGFRvIII by microvesicles derived from tumour cells." Nat Cell Biol 10(5): 619-624.

Ashburner, M., C. A. Ball, et al. (2000). "Gene ontology: tool for the unification of biology. The Gene Ontology Consortium." Nat Genet 25(1): 25-29.

Bendtsen, J. D., L. J. Jensen, et al. (2004). "Feature-based prediction of non-classical and leaderless protein secretion." Protein Eng Pes Sel 17(4): 349-356.

Bootcov, M. R., A. R. Bauskin, et al. (1997). "MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily." Proc Natl Acad Sci U S A 94(21): 11514-11519.

Bortoluzzi, S., P. Scannapieco, et al. (2006). "Computational reconstruction of the human skeletal muscle secretome." Proteins 62(3): 776-792.

Chan, X. C, J. C. McDermott, et al. (2007). "Identification of Secreted Proteins during Skeletal Muscle Development." J Proteome Res 6(2): 698-710.

Chen, Y., S. W. Kwon, et al. (2005). "Integrated approach for manual evaluation of peptides identified by searching protein sequence databases with tandem mass spectra." J Proteome Res 4(3): 998-1005.

Cocucci, E., G. Racchetti, et al. (2009). "Shedding microvesicles: artefacts no more." Trends Cell Biol.

Craig, R., J. P. Cortens, et al. (2004). "Open source system for analyzing, validating, and storing protein identification data." J Proteome Res 3(6): 1234-1242.

Edom, F., V. Mouly, et al. (1994). "Clones of human satellite cells can express in vitro both fast and slow myosin heavy chains." Dev Biol 164(1): 219-229.

Eisenhaber, F., B. Eisenhaber, et al. (2003). "Prediction of lipid posttranslational modifications and localization signals from protein sequences: big-Pi, NMT and PTS1." Nucleic Acids Res 31(13): 3631-3634.

Emanuelsson, O., S. Brunak, et al. (2007). "Locating proteins in the cell using TargetP, SignalP and related tools." Nat Protoc 2(4): 953-971.

Emanuelsson, O., H. Nielsen, et al. (2000). "Predicting subcellular localization of proteins based on their N-terminal amino acid sequence." J Mol Biol 300(4): 1005-1016. Garcion, C, R. Baltensperger, et al. (2006). "FiRe and microarrays: a fast answer to burning questions." Trends Plant Sci 11(7): 320-322.

Gobom, J., E. Nordhoff, et al. (1999). "Sample purification and preparation technique based on nano-scale reversed-phase columns for the sensitive analysis of complex peptide mixtures by matrix-assisted laser desorption/ionization mass spectrometry." J Mass Spectrom 34(2): 105-116.

Grounds, M. D. (2008). Complexity of Extracellular Matrix and Skeletal Muscle Regeneration. Skeletal Muscle Repair and Regeneration. S. Schiaffino and T. A. Partridge. Dordrecht, The Netherlands, Springer. 3: 269-302.

Henningsen, J., K. T. Rigbolt, et al. (2010). "Dynamics of the skeletal muscle secretome during myoblast differentiation." Mol Cell Proteomics.

Horton, P., K. J. Park, et al. (2007). "WoLF PSORT: protein localization predictor." Nucleic Acids Res 35(Web Server issue): W585-587.

Huang da, W., B. T. Sherman, et al. (2009). "Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources." Nat Protoc 4(1): 44-57.

Hulo, N., A. Bairoch, et al. (2008). "The 20 years of PROSITE." Nucleic Acids Res 36(Database issue): D245-249.

Hulsen, T., J. de Vlieg, et al. (2008). "Bio Venn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams." BMC Genomics 9: 488.

Jacquemin, V., G. S. Butler-Browne, et al. (2007). "IL-13 mediates the recruitment of reserve cells for fusion during IGF- 1 -induced hypertrophy of human myotubes." J Cell Sci 120(Pt 4): 670-681.

Jensen, S. S. and M. R. Larsen (2007). "Evaluation of the impact of some experimental procedures on different phosphopeptide enrichment techniques." Rapid Commun Mass Spectrom 21(22): 3635-3645.

Kail, L., A. Krogh, et al. (2004). "A combined transmembrane topology and signal peptide prediction method." J Mol Biol 338(5): 1027-1036.

Krogh, A., B. Larsson, et al. (2001). "Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes." J Mol Biol 305(3): 567-580.

Langley, B., M. Thomas, et al. (2002). "Myostatin inhibits myoblast differentiation by down-regulating MyoD expression." J Biol Chem 277(51): 49831-49840.

Le Bihan, M. C, Y. Hou, et al. (2006). "Proteomic analysis of fast and slow muscles from normal and kyphoscoliotic mice using protein arrays, 2-DE and MS." Proteomics 6(16): 4646-4661.

Leask, A. and D. J. Abraham (2004). "TGF-beta signaling and the fibrotic response." FASEB J 18(7): 816-827.

Liu, H., R. G. Sadygov, et al. (2004). "A model for random sampling and estimation of relative protein abundance in shotgun proteomics." Anal Chem 76(14): 4193-4201.

Mathivanan, S., J. W. Lim, et al. (2010). "Proteomics analysis of A33 immunoaffinity- purified exosomes released from the human colon tumor cell line LIM1215 reveals a tissue-specific protein signature." Mol Cell Proteomics 9(2): 197-208.

Mathivanan, S. and R. J. Simpson (2009). "ExoCarta: A compendium of exosomal proteins and RNA. " Proteomics. Miguet, L., K. Pacaud, et al. (2006). "Proteomic analysis of malignant lymphocyte membrane microparticles using double ionization coverage optimization." Proteomics 6(1): 153-171.

Nickel, W. and C. Rabouille (2009). "Mechanisms of regulated unconventional protein secretion." Nat Rev Mol Cell Biol 10(2): 148-155.

Oliver, K. G., J. R. Kettman, et al. (1998). "Multiplexed analysis of human cytokines by use of the FlowMetrix system." Clin Chem 44(9): 2057-2060.

Pan, B. T. and R. M. Johnstone (1983). "Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: selective externalization of the receptor." Cell 33(3): 967-978.

Ratajczak, J., K. Miekus, et al. (2006). "Embryonic stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of mRNA and protein delivery." Leukemia 20(5): 847-856.

Roy, H., S. Bhardwaj, et al. (2006). "Biology of vascular endothelial growth factors." FEBS Lett 580(12): 2879-2887.

Sanes, J. R. (1994). The extracellular matrix. Myology Second Edition. A. G. Engel and C. Franzini- Armstrong, McGraw-Hill. 1: 242-260.

Serrano, A. L., B. Baeza-Raja, et al. (2008). "Interleukin-6 is an essential regulator of satellite cell-mediated skeletal muscle hypertrophy." Cell Metab 7(1): 33-44.

Skog, J., T. Wurdinger, et al. (2008). "Glioblastoma microvesicles transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers." Nat Cell Biol 10(12): 1470-1476.

Smedley, D., S. Haider, et al. (2009). "BioMart-biological queries made easy." BMC Genomics 10: 22.

Thery, C, S. Amigorena, et al. (2006). "Isolation and Characterization of Exosomes from Cell Culture Supernatants and Biological Fluids." Current Protocols in Cell Biology: 3.22.21-23.22.29.

Thery, C, M. Ostrowski, et al. (2009). "Membrane vesicles as conveyors of immune responses." Nat Rev Immunol.

Valadi, H., K. Ekstrom, et al. (2007). "Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells." Nat Cell Biol 9(6): 654-659.

Watt, D. J., G. E. Jones, et al. (2004). "The involvement of galectin-1 in skeletal muscle determination, differentiation and regeneration." Glycoconjugate Journal 19: 615-619. Yagami-Hiromasa, T., T. Sato, et al. (1995). "A metalloprotease-disintegrin participating in myoblast fusion." Nature 377(6550): 652-656.

Yoon, J. H., K. Yea, et al. (2008). "Comparative proteomic analysis of the insulin- induced L6 myotube secretome. " Proteomics.

References used to build the "membrane vesicle-associated secretome" database Admyre, C, Johansson, S. M., Qazi, K. R., Filen, J. J., Lahesmaa, R., Norman, M., Neve, E. P., Scheynius, A., and Gabrielsson, S. (2007) J. Immunol. 179(3), 1969-1978

Aoki, N., Jin-no, S., Nakagawa, Y., Asai, N., Arakawa, E., Tamura, N., Tamura, T., and Matsuda, T. (2007) Endocrinology 148(8), 3850-3862

Banfi, C, Brioschi, M., Wait, R., Begum, S., Gianazza, E., Pirillo, A., Mussoni, L., and Tremoli, E. (2005) Proteomics 5(17), 4443-4455

Bard, M. P., Hegmans, J. P., Hemmes, A., Luider, T. M., Willemsen, R., Severijnen, L. A., van Meerbeeck, J. P., Burgers, S. A., Hoogsteden, H. C., and Lambrecht, B. N. (2004) Am JRespir Cell Mol Biol 31(1), 114-121

Blanchard, N., Lankar, D., Faure, F., Regnault, A., Dumont, C., Raposo, G., and Hivroz, C. (2002) J. Immunol. 168(7), 3235-3241

Choi, D. S., Lee, J. M., Park, G. W., Lim, H. W., Bang, J. Y., Kim, Y. K., Kwon, K. H., Kwon, H. J., Kim, K. P., and Gho, Y. S. (2007) J Proteome Res 6(12), 4646-4655

Escola, J. M., Kleijmeer, M. J., Stoorvogel, W., Griffith, J. M., Yoshie, O., and Geuze, H. J. (1998) J Biol Chem 273(32), 20121-20127

Faure, J., Lachenal, G., Court, M., Hirrlinger, J., Chatellard-Causse, C, Blot, B., Grange, J., Schoehn, G., Goldberg, Y., Boyer, V., Kirchhoff, F., Raposo, G., Garin, J., and Sadoul, R. (2006) Mol Cell Neurosci 31(4), 642-648

Fevrier, B., Vilette, D., Archer, F., Loew, D., Faigle, W., Vidal, M., Laude, H., and Raposo, G. (2004) Proc Natl Acad Sci U S A 101(26), 9683-9688

Garcia, B. A., Smalley, D. M., Cho, H., Shabanowitz, J., Ley, K., and Hunt, D. F. (2005) J Proteome Res 4(5), 1516-1521

Hegmans, J. P., Bard, M. P., Hemmes, A., Luider, T. M., Kleijmeer, M. J., Prins, J. B., Zitvogel, L., Burgers, S. A., Hoogsteden, H. C, and Lambrecht, B. N. (2004) Am. J. Pathol. 164(5), 1807-1815

Ji, H., Erfani, N., Tauro, B. J., Kapp, E. A., Zhu, H. J., Moritz, R. L., Lim, J. W., and Simpson, R. J. (2008) Electrophoresis 29(12), 2660-2671

Jin, M., Drwal, G., Bourgeois, T., Saltz, J., and Wu, H. M. (2005) Proteomics 5(7), 1940-1952

Kramer- Albers, E.-M., Bretz, N., Tenzer, S., Winterstein, C, Mobius, W., Berger, H., Nave, K.-A., Schild, H., and Trotter, J. (2007) Proteomics - Clinical Applications 1(11), 1446-1461

Mears, R., Craven, R. A., Hanrahan, S., Totty, N., Upton, C, Young, S. L., Patel, P., Selby, P. J., and Banks, R. E. (2004) Proteomics 4(12), 4019-4031

Miguet, L., Pacaud, K., Felden, C, Hugel, B., Martinez, M. C, Freyssinet, J. M., Herbrecht, R., Potier, N., van Dorsselaer, A., and Mauvieux, L. (2006) Proteomics 6(1), 153-171

Peterson, D. B., Sander, T., Kaul, S., Wakim, B. T., Halligan, B., Twigger, S., Pritchard, K. A., Jr., Oldham, K. T., and Ou, J. S. (2008) Proteomics 8(12), 2430-2446 Pisitkun, T., Shen, R. F., and Knepper, M. A. (2004) Proc Natl Acad Sci U S A 101(36), 13368-13373

Potolicchio, I., Carven, G. J., Xu, X., Stipp, C., Riese, R. J., Stern, L. J., and Santambrogio, L. (2005) J. Immunol. 175(4), 2237-2243

Ristorcelli, E., Beraud, E., Verrando, P., Villard, C., Lafitte, D., Sbarra, V., Lombardo, D., and Verine, A. (2008) FASEB J. 22(9), 3358-3369

Sander, T. L., Ou, J. S., Densmore, J. C., Kaul, S., Matus, I., Twigger, S., Halligan, B., Greene, A. S., Pritchard, K. A., Jr., and Oldham, K. T. (2008) Shock 29(4), 504-511

Segura, E., Nicco, C., Lombard, B., Veron, P., Raposo, G., Batteux, F., Amigorena, S., and Thery, C. (2005) Blood 106(1), 216-223

Skokos, D., Le Panse, S., Villa, I., Rousselle, J. C., Peronet, R., David, B., Namane, A., and Mecheri, S. (2001) J. Immunol. 166(2), 868-876

Smalley, D. M., Root, K. E., Cho, H., Ross, M. M, and Ley, K. (2007) Thromb Haemost 97(1), 67-80

Smalley, D. M., Sheman, N. E., Nelson, K., and Theodorescu, D. (2008) J Proteome Res 7(5), 2088-2096

Thery, C., Boussac, M., Veron, P., Ricciardi-Castagnoli, P., Raposo, G., Garin, J., and Amigorena, S. (2001) J. Immunol. 166(12), 7309-7318

Utleg, A. G., Yi, E. C., Xie, T., Shannon, P., White, J. T., Goodlett, D. R., Hood, L., and Lin, B. (2003) Prostate 56(2), 150-161

Valadi, H., Ekstrom, K., Bossios, A., Sjostrand, M., Lee, J. J., and Lotvall, J. O. (2007) Nat Cell Biol 9(6), 654-659

van Niel, G., Raposo, G., Candalh, C., Boussac, M., Hershberg, R., Cerf-Bensussan, N., and Heyman, M. (2001) Gastroenterology 121(2), 337-349

Wubbolts, R., Leckie, R. S., Veenhuizen, P. T., Schwarzmann, G., Mobius, W., Hoernschemeyer, J., Slot, J. W., Geuze, H. J., and Stoorvogel, W. (2003) J Biol Chem 278(13), 10963-10972

Table SI; The in silico/ virtual secretome derived from the gene expression profile of 72h

differentiated cultures, 393 transcripts were identified as strong candidates for being secreted into the extracellular space via either classical or unconventional pathways.

LOXL1 ILMN..137660 PREDICTED: Homo sapiens hsyl o-idase-lfte 1 (LOXL1). mRNA. 1124.8 0.0077! +SP!

CDS9 IL N. .1905 Homo sapiens CD59 molecule, complement regulatory protein (COS9), transcript variant 995.3 0.00001 ♦SP[

2. mRNA. j

CFH ILMN 83Ϊ7 Homo sapiens complement factor H (CFH), transcript variant 1 , mR A. 988.1 0.0002: ♦SP!

TNFRSF25 ILMN .14916 Homo sapiens tumor necrosis factor receptor superfamily, member 25 (TNFR5F25). 938.2 0.0002 ♦SP|

transcript variant 10. mRNA.

ST3GAL1 ILMN. .1974 Homo sapiens ST3 beta-galactoside atpha-2,3-siatyttransfetase 1 (ST3GAL1), transcript 858.9 0.00042 ♦SP!

variant 1, mRNA. j

COL12A1 ILMN. .17802 Homo sapiens collagen, type XII, alpha 1 (COL12A1 ), transcript variant long, mRNA. 845.2 0.003! ♦SPj

CTSC ILMN. .14007 Homo sapiens cathepain C {CTSC). transcript variant 1, mRNA. 758.9 00002!

B P1 ILMN .6483 Homo sapiens bone morphogenatic protein 1 (BMP1), transcript variant 8MP1-3, mRNA 738.1 0.0004; ♦SPj

QPCT ILMN. 6S10 Homo sapiens gtutaminyt-pepUde cyclotransferase (QPCT), mRNA. 727.S 0.00001

FRAS1 ILMN 5292 Homo sapiens Fraser syndrome 1 (FRAS1), transcript variant 2, mRNA. 716.3 0.001; ♦SPj

ECM2 ILMN. .7147 • Homo sapiens extracellular matrix protein 2, female organ and adipocyte specific 675.S 000001 ♦SPi

(ECM2), mRNA.

FUCA2 ILMN. .13837 Homo sapiens fucosidase, alpha-L- 2. plasma (FUCA2), mRNA. 6542 0.00025 ♦SPj

C1R ILMN. .137316 PREDICTED: Homo sapiens complement component 1, r subcomponent, transcript 650.2 0.009: +SP!

variants (C1R). mRNA. j

ENTRD6 ILMN .17684 Homo sapiens ecfonudeoeide triphosphate diphosphohydrolase 6 (putative function) 613.7 (-SP)j

(ENTP06). mRNA.

COL4A5 ILMN. .13514 Homo sapiens eoBagen. typo IV. alpha 5 (COL4A5). transcript variant 1 , mRNA.. 606.4 0.00007 +SP!

CRTAP ILMN. 2952 Homo sapiens cartilage associated protein (CRTAP), mRNA. 5628 0.00241 ♦SPJ

COL18A1 ILMN. .6591 Homo sapiens collagen, type XVIII, alpha 1 (COL18A1), transcript variants, mRNA.. 535.4 0.00367 ♦SPi

RNASE4 ILMN. .21567 Homo sapiens ribonu lease, RNase A family, 4 (RNASE4). transcript variant 3. mRNA. 516.3 0.00023 ♦SPj

LTBP2 ILMN. .139258 Homo sapiens latent transforming growth factor beta binding protein 3 (LTBP3), mRNA. 515 0.00855

ANG ILMN. .13716 Homo sapiens angiogenic ribonuclease, RNase A family, 5 (ANG), transcript variant 2, 477.1 0.00001 V mRNA.

XYLT1 ILMN. .71794 Homo sapiens xytosyltransferase 1 (XYLT1), mRNA. 471.5 0.00672

FAM20C ILMN .36787 Homo sapiens family with sequence similarity 20, member C (FAM20C), mRNA. 4292 0.000 ;

COL11A1 ILM .162953 Homo sapiens collagen, type XI, alpha 1 (COL11A1). transcript variant B. mRNA. 421.7 0.0042!

MATN2 ILMN .23856 Homo sapiens matriltn 2 (MATN2), transcript variant 1 , mRNA. 417.7 0.00013 ♦SP|

IL6ST ILMN. .2496 Homo sapiens interteukin 6 signal transducer (gp1 0, oncostatin M receptor) (IL6ST), 416.2 O.OO001

transcript variant 1, mRNA. Ϊ

PLTP ILMN. .12725 Homo sapiens phospholipid transfer protein (PLTP). transcript variant 2, mR A. 402.2 00004!

POL3S ILMN. .38726 Homo sapiens polyserase 3 (POL3S). mRNA. 354.2 0.0013;

TNFSF15 ILMN. .2068 Homo sapiens tumor necrosis factor (Ugand) superfamily, member 15 (TNFSF1S). 351.6 0.0020! ♦SP|

mRNA.

PENK^" ILMN. .9859 Homo sapiens proenkephalin (PEN ), mRNA. 349 0,00132

IL17B ILMN. .101 9 Homo sapiens interieu in 1 B (IL17B), mRNA. 334.5 0.0013; ♦SPj

COL5A3 ILMN. .16905 Homo sapiens collagen, type V, alpha 3 (COL5A3), mRNA. 333.3 0.0013; ♦SPi <·

PLA2G3 ILMN. .27165 Homo sapiens phosphoiipase A2. group lit (PLA2G3), mRNA. 3305 0.00132

WISP1 ILMN. .9186 Homo sapiens WNT1 inducibfe signaling pathway protein 1 (YVISP1), transcript variant 1 314.4 0.0012!

mRNA.

OLR1 ILMN. .17381 Homo sapiens oxidized low density lipoprotein (lectuv!ike) receptor 1 (OLR1), mRNA. 310.4 0.00264 (-SP

COL15A1 ILMN .15732 Homo sapiens collagen, type XV. alpha 1 (COL15A1), mRNA. 308.8 0.00264 ♦SP!

LCAT ILMN. .29730 Homo sapiens lecithin-cholesterol acyltransferase (LCAT), mRNA. 308.7 0.00264 ♦SPI

STCt ILMN. .16225 Homo sapiens stanntocakin 1 (STC1). mRNA. 307.4 0.00264 ♦SP|

EGFLAM ILMN. .21200 Homo sapiens EGF-like, fibronectin type III and laminin G domains (EGFLAM). 306.9 0.0066;

transcript variant 1, mRNA.

TGFB2 ILMN. .3549 Homo sapiens transforming growth factor, beta 2 (TGFB2). mRNA. 304.5 0.00264

SPP1 ILMN. .9394 Homo sapiens secreted phosphoprotein 1 (SPP1 ), transcript variant 2 (Osteopontin) , 302.8 0.00001 ♦SP!

mRNA.

CRLF1 ILMN. .4898 Homo sapiens cytokine receptor-like factor 1 (CRLF1). mRNA. 299.8 0.00264

COL22A1 ILMN. .18096 Homo sapiens collagen, type XXrt, alpha 1 (COL22A1 ), mRNA. 296.5 0.00264

TNFSF12 ILMN. .25202 Homo sapiens tumor necrosis factor (figand) superfamily, member 12 (TNFSF12), 2949 0.0048! ♦SPj transcript variant 2. mRNA.

ST3GAL4 ILMN. .19390 Homo sapiens ST3 beta-galactoside alpha-2,3-sialyllransferase 4 (ST3GAL4), mRNA. 2924 0.0039!

ST3GAL3 ILMN. .13266 Homo sapiens ST3 bela-galactoside alpha-2,3-S!alyttransferase 3 (ST3GAL3), transcript 291.6 0.0000! ♦SP|

variant 8, mRNA. J

2P3 ILMN. .9563 Homo sapiens rona peltucida glycoprotein 3 (sperm receptor) (ZP3), mRNA. 230.1 0.00106

SE A7A ILMN. .17564 Homo sapiens semaphorin 7A, GPI membrane anchor (John Milton Hagen blood group) 287.9 0.003.9! ♦SP|

(SEMA7A), mRNA. !

KAZALD1 ILMN. 281 9 Homo sapiens azal-type serine peptidase inhibitor domain 1 (KAZALD1). mRNA. 2666 0.0039!

GDF5 ILMN. .27925 Homo sapiens growth differentiation factor 5 (GDF5), mRNA. 280.4 0.00395 ♦SP!

CRISPLD2 ILMN. .7921 Homo sapiens cyateine-rich secretory protein LCCL domain containing 2 (CRISPLD2), 271 0.00395

mRNA.

AOAMTSL1 ILMN. 4508 Homo sapiens ADAMTS-Iike 1 (ADAMTSL1), transcript variant 3, mRNA. 2Θ8 0.0000! ♦SP<

GPC4 ILMN. .11696 Homo sapiens gtypican 4 (GPC4J, mRNA. 264.3 0.0039! ♦SP

ARSA ILMN. .17279 Homo sapiens arytsulfatase A (ARSA). mRNA. 259.2 +SP|

LAMC2 ILMN. .28991 Homo sapiens laminin, gamma 2 (LAMC2L transcript variant 2, mRNA. 257 3 0.00001

AFP ILMN. .19039 Homo sapiens alpha-fetoprotein (AFP), mRNA. 250.4 0.00527

CD109 ILMN. .17848 Homo sapiens CO109 molecule (CD109). mRNA. 2496 0.00527 ♦SPi

PLA2G12A ILMN. .4944 Homo sapiens phosphoiipase A2, group XIIA (PLA2G12A), mRNA. 247.9 000527 LTBP4 IL N_174435 Homo sapiens latent transforming growth factor beta binding protein 4 (LTBP4 J, 247.S 0.00334 ♦SPi transcript variant 2, mRNA.

NEGRI ILMN_20523 Homo sapiens growth regulator 1 (NEGRI), mRNA. 247 0.00527 ♦SPj

IL28B rLMN_4533 Homo sapiens interieukin 288 (interferon, lambda 3) (IL28B). mRNA. 243 0.00527

CXCLS ILMN 21398 Homo sapiens chemokine (C-X-C motif) ligand 5 (CXCLS), mRNA. 239.9 O.OOOOi ♦spj

GHR ILMNJI2966 Homo sapiens growth hormone receptor (GHR). mRNA. 238.7 0.00527 ♦spj

ADAMTS2 (L N_20S66 Homo sapiens ADAM merattopeptidase. with thrombospondin type 1 motif, 2 233.9 o.ooooe

(AOAMTS2). transcript variant 2, mRNA. 1

WNT10A ILMN_12046 Homo sapiens wingless-type MMTV integration site family, 10A (WNT10A), 232.8 0.00527 +SP| mRNA.

FAM24B ILMN.25126 Homo sapiens family with sequence similarity 24, member B (FAM24B). mRNA. 228.9 o.ooooe

ITLG ILMNJ72829 Homo sapiens KIT ligand (KITLG), transcript variant b, mRNA. 228.5 00040* ♦ SP|

C6orf15 ILMN_8387 Homo sapiens chromosome 6 open reading frame 15 (C6orf15), mRNA. 228.1 0.00527 ♦spj

THSD4 ILMN_25S98 Homo sapiens fhrombospondin, type 1, domain containing 4 (THS04), mRNA. 227.9 0.00527 ♦SPj

FAS ILMN_9068 Homo sapiens Fas (TNF superfamify, member 6) (FAS), transcript variant 3, 2273 000007 ♦SPi mRNA.

C5 H.MN 28154 Homo sapiens complement component 5 (C5). mRNA. 224.1 0.00627 *SPj

C3 ILMN_S682 Homo sapiens complement component 3 (C3), mRNA. 213.7 o.ooos: ♦SPi V

FGL2 ILMNJ9861 Homo sapiens fibrinogen-like 2 (FGL2), mRNA. 212.3 0.00527

IL7R ILMN_10S13 Homo sapiens interieukin 7 receptor (IL7R), mRNA. 210 0.00006

BDNF ILMN_26926 Homo sapiens brain-derived neurotrophic factor (BDNF), transcript variant 3. mRNA. 206.7 0.00651 ♦spj V

MMP14 ILMN_75I1 Homo sapiens matrix metallopepridase 14 (membrane (M P14), mRNA. 202.6 0.00527 ♦SP|

DEFB103A ILMN_16S573 Homo sapiens defensin. beta 103A (OEFB103A), mRNA. 201.7 O.00S27 ♦spj εοΝΐ ILMN_28724 Homo sapiens endotheliri 1 (EON1), mRNA. 201.4 0.00527 ♦spj

CCL26 ILMN 6946 Homo sapiens chemokine (C-C motif) ligand 26 (CCL26). mRNA. 200.6 0.00527 ♦spj

FLT3LG ILMN.4754 Homo sapiens fms-reiafed tyrosine 3 figand (FLT3LG), mRNA. 195 0.00527 +SPj

LIFR IL NJ5930 Homo sapiens leukemia inhibitory factor receptor alpha (LIFR), mRNA. 190.3 0.00527 (-SP),

COL10A1 ILMN_12987 Homo sapiens collagen, type X. alpha 1 (COL10A1), mRNA. 189.7 O.00527 ♦spj

HAPLN3 IL N_19816 Homo sapiens hyafuronan and proteoglycan link protein 3 (HAPL 3), mRNA. 188.3 0.00527

AOAMTS7 ILMN_2S45 Homo sapiens ADAM metallopeptidase with thrombospondtn type I motrf, 7 185.4 0.00527 +SP|

(ADAMTS7), mRNA.

RELN ILMN_92S2 Homo sapiens (RELN), transcript variant 1, mRNA. 184.6 O.OOOi

IL34 ILM S634 Homo sapiens intertoukin 34 (IL34), mRNA. 184.4 0.00527 ^■tspj

CEL ILMN 26946 Homo caifcovyl ester lipase (bile salt-stimulared lipase) (CEL), mRNA. 182 0.00527 ♦SP!

SERPINtl ILMNJ 83655 Homo sapiens serpin peptidase inhibitor, clade 1 (neuroserpin). 1 (SERPINI1), 181.8 0.00012 +SPi mRNA.

CCDC126 ILMN 24365 Homo sapiens coHed-coil domain containing 126 (CCDC126). mRNA. 179.2 O.00527 +SPj

TCN2 IL N 6136 Homo sapiens transcobalamin II; macrocytic (TCN2), mRNA. 1769 0.00527 +SPi

SFRP4 IL NJ3024 Homo sapiens secreted frizzled related protein 4 (SFRP4). mRNA. 176.5 O.0O463 <spj

C0L14A1 ILMN_3668 Homo sapiens collagen, type XIV, alpha 1 (COL14A1). mRNA. 1 5.7 □ 00527 •i

WNT3 ILMN_1260 Homo sapiens wingless-type MMTV integration site family, member 3 (WNT3). mRNA. 1745 000527 *spj

APOC1 ILMN_14337 Homo sapiens apoltpoprotein C-t (APOC1), mRNA. 172.5 000527

SGD3 ILMN 20909 Homo sapiens superoxide dismutase 3, extracellular (SOD3), mRNA. 172.5 0.00527 ♦spj

PRR4 ILMN_16830 Homo sapiens proline rich 4 (lacrimal) (PRR4). transcript variant 1 , mRNA. 171.4 0.00527 +SPj

PGF ILMN_27436 Homo sapiens placental growth factor (PGF), mRNA. 168.3 0.00527 ♦SPi V

IFI30 ILMN.23180 Homo sapiens interferon, gamma-inducible protein 30 (IFI30). mRNA. 1668 000527 ♦spj •1

BGLAP ILMNJ 7038 Homo sapiens bone gamma-carboxygrufamale (gla) protein (osteocalcin) (8GLAP), 1663 0.00527

mRNA..

LA A1 IL N_138581 PREDICTED". Homo sapiens alpha 1 (LAMA1), mRNA. 166.1 O.OOOOE ♦spj

LFNG lLMN_t 66415 Homo sapiens LFNG O-fucosylpeptide 3"beta-N-ace!ylglua^mriiylrransferass (LFNG). 164.9 0.0095S •1 transcript variant 2, mRNA.

CXCL16 ILMM_2980S Homo sapiens chemokine (C-X-C motif) figand 16 (CXCL16). mRNA. 164.2 0.0008:

IL26 ILMN_7281 Homo sapiens interieukin 26 (IL25). mRNA. 163.7 0.006SS +SP|

ANGPTL4 ILMN_31891 Homo sapiens angiopoietin 4 (ANGPTL4), transcript variant 3. mRNA. 162.2 0.0002S ♦spj

VWF ILMN_136180 Homo sapiens von Wii!ebrand factor (VWF), mRNA. 161.2 O.QOSSi ♦SPj

GPX3 ILMN .137905 Homo sapiens glutathione peroxidase 3 (GPX3), mRNA. 159.1 0.0065S ♦SP.

GNRH1 ILMN_28035 Homo sapiens gonadotropin-reteasing hormone t -releasing hormone) 158.3 0.Q06SS

(GNRH1), variant 1. mRNA.

MUC20 ILMN_18898 Homo sapiens mucin 20, cell surface associated (MUC20). transcript variant mRN A. 158.2 0.006SS

I 1

CST9 IL N_4872 Homo sapiens cystatin 9 (testatin) (CST9), mRNA. 158.1 0.0066S +SPj

COL21A1 ILMN_138693 Homo sapiens collagen, type XXI, alpha 1 (COL21 At). mRNA. 158 0.0066! +SPi

SERPING1 ILMN_14941 Homo sapiens peptidase inhibitor, c!ade G (C1 inhibitor), member 1 (SERPING1), 156.1 0.00541

transcript variant 2, mRNA.

FMOD ILMN 29801 Homo sapiens frbromodutin (FMOD), mRNA. 153.2 0.00791 +SPi

NPPA ILMN 8111 Homo sapiens natriuretic precursor A (NPPA). mRNA. 1 8.5 0.0092: +spj

NID1 ILMN 3650 Homo sapiens nidogen 1 (NIDI), mRNA. 143.3 0.0092: y

CYTLI ILMN_7848 Homo sapiens cytokine-tike 1 (CYTLI), mRNA. 141.3 0.00921 ♦spj \'

LGALS1 IL N_138248 Homo sapiens lectin ga!actoside 1 (LGALS1), mRNA. 28547.1 0 (-SP)j

SOD1 ILMNJ4302 Homo sapiens superoxide dismutase 1, soluble (SOD1), mRNA. 14588.3 0 (-SP).

ANXA2 ll_MN_9658 Homo sapiens A2 (ANXA2), transcript variant 1, mRNA. 146182 0 (-SP)!

MIF ILMN_2668e Homo sapiens macrophage migration inhibitory factor )iting factor) 110 9.4

(MIF), mRNA. 1

IL18 ILMN_9174 Homo interieukin 18 finterferon-gamma-inducing factor) (IL18), mRNA. 10895.8 0 (-sp)j

RECTIFIED SHEET U TFRC IL NJ29Q9 Homo sapiens transferrin receptor (p90, CD71) (TFRC), mRNA. 9632.1 C (^■SPji

F13A1 ILMN.9687 Homo sapiens coagulation factor XtH, A1 polypeptide (F13A1), mRNA. 6064.9 0 (-SP)j

FTH1 ILMNJ 1546 Homo sapiens ferritin, heavy polypeptide t (FTH1 ), mRNA. 5403.3 0 (-SP)]

ADAMTS1 IL N_11081 Homo sapiens ADAM metaltopeptidase with thrombospondin type 1 motif, 1 4270.4 0 +spj

(ADAMTS1), mRNA

CDH13 ILMN_26240 Homo sapiens cadherin 13. H-cadherin (heart) (CDH13), mRNA. 3635.6 c ♦SP| V

HDGF ILMNJ6816 Homo sapiens hepatoma-derived growth factor (high-mobflHy group protein 1 -tike) 2668.1 0 (-SP)j

(HDGF). mRNA

GPI ILMN 7872 Homa sapiens glucose phosphate isomerase (GPI), mRNA 2377.3 0 <-sp)|

ISG15 ILMN_6174 Homo sapiens ISG15 ubiquKin-ffke modifier (ISG15). mRNA 2274.5 0 (-SP)j V

IL1B ILMN_27277 Homo sapiens interleukm 1, beta (IL1B). mRNA. 2265.7 c (-SP)j AP2K2 ILMN 24956 Homo sapiens mtogen-activated protein kinase kinase 2 (MAP2K2), mRNA. 2215.9 0 (-SP)!

LGALS3 ILMNJ4333 Homo sapiens lectin, galactoside-binding, soluble, 3 (gatectin 3) (LGALS3), mRNA. 1718.8 0 (-SP)I

IL32 ILMNJ 7936 Homo sapiens irrterteutrin 32 (IL32), transcript variant 4, mRNA. 1564.4 0 (-SP)!

CANT1 ILMN 21488 Homo sapiens calcium activated nucleotidase 1 (CANT1), mRNA. 1372.6 t +SPJ <l

FGF2 1LMNJS7999 Homo sapiens fibroblast growth factor 2 (basic) (FGF2), mRNA 1295.2 c (-SP)I Ί

MFAP1 ILMN 20656 Homo sapiens microftbrfllar-assoctated protein 1 (MFAP1), mRNA. 1045.7 t (-SP)j

RGMB ILMNJ3578 Homo sapiens RGM domain family, member 8 (RGMB), transcript variant 2. mRNA 1044.5 c +SPj

WNT5A ILMNJ 4624 Homo sapiens wingless-type MMTV integration site family, member 5A (WNT5A), 775.5 0 (-SP)j mRNA

A0P HL2 ILMN_2074 Homo sapiens ADP-ribosyihydrolase like 2 (AOPRHL2), mRNA 651.1 0 +SP| V

QS0X2 ILMN 21606 Homo sapiens quiescin Q6 sulfhydryt oxidase 2 (QSOX2), mRNA. 601.6 0 +spj ^■4

VASH1 ILMNJ905 Homo sapiens vasohibin 1 (VASH 1 ), mRNA. 601.2 0 (-SP)j

ILIA ILMN .25320 Homo sapiens interieukm 1 , alpha (1L1 A), mR . 5943 0 (-SP

KIAAQ564 ILMNJ6676 Homo sapiens KIAA0564 (KIAA0564), transcript variant 2, mRNA. 239.9 0 (-SP)j

AGGF1 ILMNJ 76602 Homo sapiens angiogenic factor with G patch and FHA domains 1 (AGGF 1 ), mRNA. 427.4 o.oooo; (-sp.il

LA01 ILMN 7567 Homo sapiens ladinin 1 (LAD1), mRNA 367 0.001 £ (-SP)j

PODN ILMN.28316 Homo sapiens podocan (POON), mRNA 351 0.00132 (-SP)I

LRCH3 ILMNJ4018 Homo sapiens leudne-rich repeats and calponin homology (CH) domain containing 3 318 0.00132 (-SP)j

(LRCH3). mRNA.

PDOCt ILMN_26S71 Homo sapiens Parkinson disease 7 domain containing 1 (PDDC1). mRNA. 280.7 000001 (-SP)I

SERPINS2 IL NJ4466 Homo sapiens serpin peptidase inhibitor, dade B (ovalbumin), member 2 (SERPINB2). 259.1 0.00004 (-SP)j mRNA

SA D1 ILMNJ 2112 Homo sapiens sterile alpha motif domain containing 1 (SAMD'I), mRNA. 252.3 0.00527 (^•SP)j

VEGFA ILMN_5181 Homo sapiens vascular endothelial growth factor A (VEGFA), transcript variant 3, 238.2 0.00759 ♦spj mRNA.

GFOD2 ILMNJ 6965 Homo sapiens glucose-fructose oxidoreductase domain containing 2 (GFOD2), mRNA. 219.8 0.00527 (-sp)l

EN 0X2 ILMNJ 0105 Homo sapiens ecto-NOX disulfide-thiol exchanger 2 (ENOX2), transcript variant 1. 200.1 0.00086 (-SP)! mRNA

IL15 ILMNJ 6803 Homo sapiens iriterteukiri 15 (IL15), transcript variant 1, mRNA. 186.8 000137 (-spj!

FGF9 IL NJ771 Homo sapiens fibroblast growth factor 9 (g!ia-activaiing factor) (FGF9), mRNA 158.3 0.00527 (-SP)j

APOM ILMNJ9748 Homo sapiens apoNpoprotein M (APOM), mRNA 154.7 0.0065S (-SP)I

CTF1 ILMNJ 3986 Homo sapiens cardiotrophin 1 (CTF!), mRNA 61.9 0.00659 (-SP)j

WNT2B ILMNJ8414 Homo sapiens wingless-type MMTV integration site family, member 2B (WNT2B). 158 0.00011 (-SP)j transcript variant WNT-2B2, mRNA.

.

V

V

Q9UBP4 Oickkopf-related protein 3 precursor

-y T ~r ~r ~τ ·?- "y -yy ~r T ~y~y~y -y -y

-y -y-y > ^{■ ■} ~y - -

-y ~y · "y ^■V -y^■ ~y -y -y ~y ~f ~7 -y-y-y-y -y-y-y-y

~r -y >

-y-y - -y -y -y -y -y -y · ^*y · -y ~y -y -y^■ -y ·> > -y -y-y-y -y -y

) tf o o to

ω σι to m innron«o<o fflrto

^« « g g] n (v n — o—i —— p| o s v B _{ft r}

« (j> o ¾| p)

? * ? ^rt ? ^rt I S s 8 a I S5i8il8lIii 5l 5l 5I z? »i agS a 10

KS S & SSS EES S 6 S S S £ o S o jl s8

σ a ».5

r a. a. r : σοSσ8Εf Sσ σ

8 S38 $ 8

V

V

V V

V

V

V

V

V V

v¹

V

V

V

V

V

V.

V

V

V

V

V

V

V

>/ ^■

V

V

V

>,

V

V

-1

V

^■7· ^"? >

O a n u) m ¾ n ^ n

S ω ¾ R S ¾

V

>/

>/

>/ l

>/ V

V

V

V

V

-1

V

V

V

V

47.0 434 >/ ^■J

10 0 Λ

10 0

. -4

^■4

V

^■4

V ^■4

^■4

^■

. -4

V

V

^■4

. -4

^■4

^■4 y

.V

1 v'

^■4

4

4

4

4

4

4 4

4

4

4

4 4

4

4

4

4

4 4

4

4 4

4

4 4

4

4 4 4

4 4

4

4

4 4

4 4 4

4

4

4

4 4

4

v' t

v'

V

V

V

Unclassified SEMA6B Q9H3T3 Semaphorirv6B precursor 95.2 888 '

Unclassified SEPHS1 P49903 Selenide, water dikinase 1 13.8 126 \'

Unclassified SERBP1 Q8NC51 Plasminogen activator inhibitor 1 RNA-bhding protein 44.9 408

Unclassified SERPINB1 P30740 Leukocyte elastase inhibitor 42.7 379

Unclassified SERPINB6 P35237 Serpin B6 11.7 376

Unclassified ;SF1 Q15637 Splicing factor 1 68.3 639

Unclassified_: SF3A1 Q 15459 Splicing factor 3 subunit 1 88.» 793 V

Unclassified SF3B1 075533 Splicing factor 3B subunit 1 145.7 1304

Unclassified SF3B2 Q13435 Splicing factor 3B subunit 2 97.6 872

Unclassified. S6 196 Q9H5 3 Protein kinase-like protein Sg l96 40.0 350 V

Unclassified SGTA 043765 Small glutarra'ne-rich tefratricopeptkte repeat-containing protein alpha 34.0 313

Unclassified SIL1 Q9H173 hfcjcleotide exchange factor SIL1 precursor 52.1 461

, Unclassified SOD2 P04179 SuPXxide dismutase JMrt], mitochondrial precursor 24.7 222

Unclassified SSRP1 QQ8945 FACT complex subunit SSRP1 81,0 709

Unclassified ^■SUGT1 Q9Y2Z0 Suppressor of G2 allele of S P1 homolog 41.0 365

Unclassified SUMF1 Q8NBK3 Sulfatase-modifying factor 1 precursor 40.5 374

Unclassified SUS05 060279 Sushi domain-containing protein 5 precuraor 88,2 629 ^■4

Unclassified SWAP70 Q9UH65 Switch-associated protein 70 69.0 585 V

Unclassified' "SYNP02L Q9H987 Synaptopodin 2-iike protein 102.4 977

Unclassified SYT12 Q81V01 Threoriyl-tRNA synthetase, cytoplasmic 83.4 421

Unclassified TAL0O1 P37837 Transaldolase 22.4 199 i

Unclassified TBCA 075347 Tubulin-specific chaPXne A 12.7 108 V

Unclassified TFG Q92734 Protein TFG 43.4 400

Unclassified ;TGOLN2 043493 Trans-Golgi network integral membrane protein 2 precursor 51.1 480 ΛΙ

. Unclassified TH0P1 P52888 Thimet oligopeptidase 45.6 415 '

Unclassified TMEM132A Q24JP5 Transmembrane protein 132A precursor 110.0 1023

Unclassified . TMPO P42166 Lamina-associated polypeptide 2 isoform alpha 83.1 739

Unclassified TNFRSF14 Q929S6 Tumor necrosis factor receptor superfamiiy member 14 precursor 30.4 283

Unclassified TPM2 P07951 Tropomyosin beta chain 32.8 284

Unclassified TPM3 P06753 Tropomyosin alpha-3 chain 32.8 284

Unclassifie TPP1 014773 Tripeptidyl-peptidase 1 precursor 61.2 563

Unclassified ROVE2 P10155 60 kDa SS-A Ro ribonudeoprotein 60.6 538

UnCfassjfied TSNAX Q99596 Translin-associated protein X 33.1 290

.Unclassified TXNL1 043396 Thioredoxin-like protein 1 32.2 289 \'

^'■: Unclassified U2AF2 P26368 Splicing factor U2AF 65 kDa subunit 53.5 475

Unclassified ■UBAP2L Q14157 Ubiqvitin-associated protein 2-like 114.5 1087

Unclassified UBQLN1 Q9UMX0 Ubiquilin-1 62.5 589 i

Unclassified UCHL3 P15374 Ubiquitin carboxyl-terminal hydrolase Isozyme L3 26.2 230

Unclassified UFM1 P61960 Uoiquitin-f ld modifier 1 precursor 47.7 414 ν^'

. Urelasafied ^■USP14 P54578 Ubiquitin carboxyl-terminal hydrolase 14 59.1 528

Unclassified ; uxsi Q8NBZ7 UDP-glucuronic acid decarboxylase 1 47.5 420

Unclassified VAPA Q9P0LO Vesicle-associated membrane protein-associated protein A 27.9 249 V

Unclassified XP PEP1 Q9NQW7 Xaa-Pro aminopeptidase 1 69.9 623

Unclassified XRCC5 P13010 ATP-de pendent DNA heiicase 2 subunit 2 82.7 732

(UOIJ3J03S SS9[I9pB9j) /Cl9UiqOBlU it.OJ9.I09S JBOISSBp U0U/lBU0I}U9AUO0

-oou sp ΒΙΛ jaodxg puB vtrainqoBui JOJ9J09S ;u9pu9d9p-i8iO{)/¾g JO jBoissB gqj ΒΙΛ j^jodxg i(sui9jo^jd 9nfaiu i ςζ) su jo^jj p9j9J09S siqnjos io sio aij^g _Ha3I3¾DHS,.⁹¾1 ^: S *ΙΦ1

L9

9£6890/eiOZd3/I3d -.80IJ-0/H0Z ΟΛ\ 68

69

Tab

)iiiFuri ^ ^{'" '}i^""a^"a Ϊ€ ^'S^' i s g ^{" "}i¥§^{"' "}i^' fil :

I

I Ξ mm

n a s i! li e

82 .; ^■

able S6: The Proteome of muscle-derived Exosomes, 564 validated proteins

I■

■■■ ■ ■ I■ m I■

m ■

S~S?SJii;;:;iS?S°S5SS2S;S55?Se^,»**^," " ""'"'^{, ,}" "" ""^, '^,'°^,0'""" *""»" ''"'"^,l*'" "*"^,*^,'

94

W

95

P P Π Π Π P

□□ ^' : ; n u t n J □

□ n□ □

■□□ cz:F D n i. τ:ζ . ΕΓΊ O r. ^~z r. □

^{■ '} H ^' [

S S 333333 " 2 " 2 " 22 ~ ~ £ S- ~- 5 2- 2- 222- 22- 2- 222- 2- 2- 2- 2- 2- ' ' 2- 2 5 ' 52- 22- 2- 2- 2- 2- 2· 2- 2- " " " 232 Ιϊ 2 CJ 2 C? 2

□ LJ

Π

∑j ^· [^" □ c- □ c ^" _. π π r..: :: --^~]

□ n□ n. _{c r} □ I

□

C J□ π □□ ri r. ^*

_ CZZT 13 D E^"L r~ □ C3 C Π □ Π □

)

55325 ί 5p555 n- =-5ί s 5 55555255523555552353555553255555?? 5555 ? ? 555????

liilliiigiiiiigSillili gliillilililliilli ^sIi1iiiiibiii§iiiiii

i ||l s2 3 SS | = j iiliiiHiiilHS iHigilii

99

?· ?- ? ?^■ ?^■ ? ? ? ? ? ? ? ? ? ? ? ?^■ ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ??????;·? ? ? ? ? ? ? ? ? ? ? ? ?^■ ? ?■ ? ? ? ? ? ? ?

103

able Tl: Antibody array 1 = 45 molecules essential for muscle formation

Table T3: Antibody array 3 = 143 molecules with reported roles in skeletal muscle homeostasis, muscle formation and regeneration

Table T4: The Human muscle "soluble" Secretome : a catalogue of 443 gene products soluble "secreted candidates"

Accession 14·

HUGO Gene Symbol {Swiss Prot) lllwnina VI 10 Protein Home

1Lr N_8Q23 Alp a-2-macrogtobuIh precursor 7 1LMN_922 ADAM 9 precursor ^■4 ILMNjl 1061 AOAMTS-1 precursor

ILM .20566 ADAMTS-2 precursor ₍ . . .

1LMN_22t>8l ^* ADAMT5-5 precursor

tt_MN _4508 ADAMTS-Gfce protein 1 precurcor

ILMN.,29514 ADM precursor

ILMN_2074 PotyfADP-ribose) gtycorr/drotase ARH3 V ILMN 19039 Aipf! a-fetoprotefn precursor

Agrtn precursor

Angiotensfnogen precurcor

ILMN_ 4097 Alfcofine pttosptiatnse, Ussue-nonepecfrtc Isozyme precursor

ILMN_13716 Angiogerrfn precursor

ILMN_S106 Anglopoiet!n-1 precurcor

Angtopolefin-feloted protein 1 precursor V

ILMN_3tB91 Angiopotefirvrelotect protein 4 precuroor

ILM _20088 AmJnopeptitfose N V ILMN 9658 Annexi A2 V

Apoltpopfotctn A-l precursor V Apo!lpoproLe!n cVIOO precursor

ILMN_I1S2S ^* Apoltpoprotein E precuraor V V

ILMN.23272 Amyloid beta A4 protein precursor

1LMN_28622 Protean ARWET precuraor

,LMN_17279 Arytsurfatase A precurcor V

1LMN_19648 8eta-2-fnicrogiobu&n precursor

protetn 5 precursor V

V

V

Complement C4-A precursor V

11«Ν^24697 Ca!umenfn precursor

ILM _2148B Sotubte colcium-Qcttvoted nucleotidase 1 V

- - CoilSd-^oft 3drn0<r conh3tnmfl protem 126 precursor' ' -

ILMN _28779 Co2ed-coQ domain- containing protetn 80 precuraor

Ectaxin precursor

ILMN _^25185 C-C motif chemo!dne 2 precursor

C-C motif cnenwklne 3 precuroor

C-C motif cnemoklne 5 precursor

p¾i09 C0109 antioen precurso V

Tumor necrosis factor receptor superfamiry mernber S precurcor V

ILMN_26240 Cadnertn-13 precursor V ILMN_12248 Cotnplement factor Θ precursor V ILMN 6337 CcmpJement factor H precursor V

C ftfnase-3-ifke protetn 1 precursor V

Ctiitinase domain-containing protem 1 precursor

I Li/ _ 29894 C-type lectin domain family 11 member A precursor

ILMN 1287 Tetranectin precursor

Ctusterm precursor V

ILMN_162953 Cotao^ a1p†u <XI) cnaih precursor■

* ILMN_ 7802 CoSagen alpna-l(XII) chain precursor

Claims

1. An isolated nanovesicle secreted by a muscle cell.

2. An isolated microparticle secreted by a muscle cell.

3. The nanovesicle according to claim 1 or the microparticle according to claim 2, wherein said muscle cell is selected from the group consisting of a skeletal muscle cell, a cardiomyocyte, a smooth muscle cell, and a myoblast.

4. The nanovesicle according to claim 1 or the microparticle according to claim 2, wherein said muscle cell is a skeletal muscle cell.

5. The nanovesicle or the microparticle according to any of claims 1-4, wherein said muscle cell overexpresses a molecule of interest.

6. The nanovesicle or the microparticle according to any of claims 1-5, for use in therapy.

7. The nanovesicle or the microparticle according to claim 6 or the microparticle according to any of claims 2-5, for use in the treatment and/or the prevention of sarcopenia or a muscular dystrophy.

8. A method for delivering a molecule of interest into a target cell by contacting said target cell with a nanovesicles and/or a microparticle according claim 5 comprising said molecule of interest.

9. The method according to claim 8, wherein said target cell is involved in the muscle regeneration.

10. A method for diagnosing or monitoring a muscular disease in a subject, comprising the step of detecting the presence or absence of one or more biomarkers within a nanovesicle and/or a microparticle according to any of claims 1-4 obtained from a biological sample of said subject, wherein said one or more biomarkers are associated with said muscular disease.

11. A method for aiding in the evaluation of treatment efficacy in a subject suffering from a muscular disease, comprising the step of detecting the presence or absence of one or more biomarkers within a nanovesicle and/or a microparticle according to any of claims 1-4 obtained from a biological sample of said subject, wherein the biomarker is associated with the treatment efficacy.

12. The method according to claim 10 or 11, wherein said biological sample obtained from the subject is a sample of bodily fluid r a culture of muscle cells isolated from the subject.

13. The method according to any of claims 10-12, wherein said muscular disease is a pathological condition resulting from an impairment of the muscle regeneration, such as fibrosis, adipogenesis or chronic inflammation.

14. The method according to any of claims 10-13, wherein the presence or absence of one or more biomarkers is detected within a nanovesicle and within a microparticle according to any of claims 1-4.

15. An antibody microarray comprises a solid support with a plurality of antibodies immobilized on the solid support, wherein said plurality of antibodies is able to bind to at least 30 different proteins listed in tables T1, T2, T3 or T4.