Abstract
Olfactory ensheathing cells (OECs) are a special type of glial cells that have characteristics of both astrocytes and Schwann cells. Evidence suggests that the regenerative capacity of OECs is induced by soluble, secreted factors that influence their microenvironment. These factors may regulate OECs self-renewal and/or induce their capacity to augment spinal cord regeneration. Profiling of plasma membrane and extracellular matrix through a high-throughput expression proteomics approach was undertaken to identify plasma membrane and extracellular matrix proteins of OECs under serum-free conditions. 1D-shotgun proteomics followed with gene ontology (GO) analysis was used to screen proteins from primary culture rat OECs. Four hundred and seventy nonredundant plasma membrane proteins and 168 extracellular matrix proteins were identified, the majority of which were never before reported to be produced by OECs. Furthermore, plasma membrane and extracellular proteins were classified based on their protein–protein interaction predicted by STRING quantitatively integrates interaction data. The proteomic profiling of the OECs plasma membrane proteins and their connection with the secretome in serum-free culture conditions provides new insights into the nature of their in vivo microenvironmental niche. Proteomic analysis for the discovery of clinical biomarkers of OECs mechanism warrants further study.
Keywords
Introduction
In the mammalian adult central nervous system (CNS), recovery from injury is generally poor, and spontaneous nerve regeneration occurs only in restricted regions of the CNS. The damage to nerve fiber pathways results in a devastating loss of function, due to the disconnection of nerve fibers from their targets (22,34,59). An exception to this rule is found in the olfactory system, in which olfactory receptor neurons (ORNs) undergo natural and injury-induced turnover and functional neurogenesis throughout life (25). To accommodate successful adult axon targeting, newly generated ORNs from the peripheral nervous system (PNS) that extend axons to synapse in the olfactory bulb (OB) of the CNS are aided by glia unique to the olfactory system, the olfactory ensheathing cells (OECs) (45).
For instance, OECs have been transplanted into sites of spinal cord injury and other CNS lesions, with the anticipation that OECs may recreate the plastic environment of the olfactory system elsewhere (12,21). Ao et al.'s research shows that the effectiveness of OECs induced neural differentiation of neural stem cells (NSCs) and their culture medium concentrations of total protein secretion was positively correlated. OEC conditioned medium (OCM) can induce NSCs to differentiate into neurons, but the same protein concentration of conditioned medium of astrocytes mainly induced differentiation to glial cells of NSCs (2). The results of other research groups also showed that the OCM can promote embryonic stem cells and bone marrow mesenchymal stem cells to nerve cell differentiation (23,55). Several different rat spinal cord injury (SCI) models have indicated that OECs can mediate regeneration and functional recovery. However, the degrees of functional and anatomical recovery reported by different laboratories have been quite disparate, ranging from little to extensive functional or anatomical restoration (47,51). Nevertheless, experimental human trials using cells reportedly containing OECs are now ongoing in several countries, with little understanding of whether OECs can faithfully facilitate regeneration (27,38).
Regardless of the transplantation or lesion model, a clear understanding of how OECs directly modulate the hostile environment of a CNS lesion will allow us to design more effective approaches for promoting neural regeneration and protection and minimize potentially undesirable secondary sprouting (leading to neuropathic pain or autonomic dysreflexia, for example). It has been shown that OECs can be purified and expanded from the lamina propria of the olfactory mucosa, the peripheral compartment through which ORN axons pass en route to the OB. Moreover, OECs have demonstrated clear differences from stem cells (SCs) in terms of cellular interaction with astrocytes and CNS tissues (5,6,48). To identify potential key molecules in the secretome of OECs that could account for their unique capability to intermingle with scar-associated cells and to stimulate regenerative axonal growth, researchers have confined their studies to those molecules that were exclusively expressed by OECs, in particular, the cell surface and the secreted proteins.
Immunohistochemistry studies have demonstrated that OECs can express different glial markers: p75 neurotrophin receptor, glial fibrillary acidic protein (GFAP), protein S-100, and adhesion molecules. It is also reported that OECs can synthesize several trophic factors, such as nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), neurturin, neurotrophin 4/5 (NT-4/5), and glial growth factor 2 (GGF2) (36,57).
So far, most of the studies aimed at identifying factors released by OECs and involved in neuronal development and regeneration have pursued a traditional approach, involving the characterization of the role of one or few secreted proteins. In this study, using 1D-shotgun proteomics, we not only focused on the OECs secretome but also enriched the plasma membrane of OECs because of their functions as a barrier and communication interface due to the presence of specific integral membrane proteins, which play important biological and pharmacological roles in the exchange of the material and energy between the cell and its environment, cell–cell interactions, and signal transport.
Transplantation of OECs into the lesion site of the central nervous system is known to enhance regeneration of transected axons and functional recovery. Despite the advances in our understanding of the functional properties of OECs, there is still little known about the molecular mechanisms through which OECs mediate axonal repair (30). Therefore, this work also want to search for important biomarkers involved in cell signal pathway and connect with the plasma membrane, the release of which, by OECs, was the messenger contacts internal and external cellular environment.
Materials and Methods
Materials
Trypsin (proteomics sequencing grade), dithiothreitol (DTT), iodoacetamide (IAA), trifluoroacetic acid (TFA), phenylmethylsulfonyl fluoride (PMSF), sodium deoxycholate (SDC), protease inhibitor cocktail, R-cyano-4-hydroxycinnamic acid (HCCA), 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES), and sucrose were purchased from Sigma-Aldrich (St. Louis, MO, USA). Acrylamide, bisacrylamide, thiourea, urea, glycine, Tris, and sodium dodecyl sulfate (SDS) were from Amresco (Solon, OH, USA). HPLC-grade formic acid, acetone, methanol (MeOH), and acetonitrile (ACN) were from Fisher Scientific Canada (Edmonton, Canada). TPCK-modified porcine trypsin was purchased from Promega (Madison, WI, USA). Water was obtained with a Milli-Q Plus purification system (Millipore Filter Corporation, Bedford, MA, USA). Amicon Ultra-4 Centrifugal Filter Devices with low-binding Ultracel membranes: 3000 NMWL was from Millipore (Millipore Filter Corporation). Horseradish peroxidase-conjugated antibody anti-rabbit was procured from Cell Signaling (Danvers, MA, USA). All culture media supplies and secondary fluorescent antibodies were purchased from Invitrogen (Carlsbad, CA, USA). All other regents were domestic products of highest grade available.
Primary Culture and Purification of OECs
All experimental procedures in this study were conducted according to institutional guidelines for laboratory animal care. All Wistar rats were obtained from the laboratory center of the Medical College, Middle South China University. For each of three separate experiments, olfactory bulbs from ten 5-week-old Wistar rats were collected in cool sterile D-Hanks balanced salt solution. Primary OECs were purified and cultured according to the methods described by Chuah et al. (17). Briefly, the meningeal membranes was stripped off with fine forceps and the outer nerve layers of the olfactory bulb were dissected, then the pooled tissue was digested at 37°C in 0.25% trypsin and 0.03% collagenase for 15 min. This step was repeated twice with a fresh solution. Trypsinization was stopped by adding DMEM supplemented with 10% fetal bovine serum (FBS). The digested tissue was mechanically dissociated by trituration and filtrated through an 80-mm nylon mesh followed by centrifugation at 1000 x g for 10 min. Cells were resuspended and plated in flasks fed with fresh complete medium DMEM supplemented with 10% FBS, 2 mM l-glutamine, penicillin (50 mg/ml) and streptomycin (50 U/ml). After 24 h, the culture was treated with 10 μM cytosine arabinoside for 48 h to minimize the population of fibroblasts. To eliminate fibroblasts that survived the antimitotic agent, OECs cultures were processed to an additional step of passaging cells from one flask to a new one. This step reduces contaminating cells because they adhered more readily to plastic than OECs. In the last passage OECs were placed on the poly-l-lysine-treated flasks and cultured in complete medium supplemented with 20 mg/ml bovine pituitary extract (Sigma) at a final density of 5 × 104 cells/ml. When the cells reached roughly 90% confluence, the supernatant was removed and fresh medium (DMEM not containing FBS) was added.
OEC Characterization Using Flow Cytometry
OECs were cultured to 80% confluence, rinsed twice with phosphate buffered saline (PBS), and detached using 3 ml of trypsin-EDTA (0.25% in 1 mmol L−1 EDTA; Gibco). Trypsin activity was stopped using 9 ml of DMEM/F-12 supplemented with 10% FBS. Cells were centrifuged and the cell pellet was resuspended in FACS wash (1x PBS, 0.2% w/v bovine serum albumin, and 0.05% sodium azide). The cell suspension was diluted to 1 × 105 cell aliquots per 100 μl, incubated with FITC-conjugated anti-P75 (low-affinity nerve growth factor receptor) primary antibody and incubated at 37°C in darkness for 30 min. Subsequently, cells were centrifuged and the pellets resuspended in cold FACS wash; the washing procedure was repeated to ensure removal of unbound antibody. OECs were subjected to flow cytometry analysis using a Becton Dickinson FACScan bench-top flow cytometer (San Jose, CA, USA), equipped with a 488-nm solid-state laser. Bivariate forward and side-angle scatter gating were used to identify homogenous populations while excluding cellular debris and dead cells. Alexa 488 fluorescence was visualized using 530/30-bandpass filters, respectively (20,000 events were collected). Photomultiplier voltages were adjusted to ensure that autofluorescence associated with nonapoptotic samples described a Gaussian distribution within the first two log-decades on univariate histograms. Data analysis was performed using Cytomation Summit v3.1 software (Cytomation; Fort Collins, CO, USA).
Preparation of Crude OECs Plasma Membrane (OPM) Fraction
OECs were homogenized in 4 ml of 250 mM sucrose, 15 mM Tris, pH 7.4, 0.1 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail with a Tissue Tearor (IKA products, T8 ultra-turrax, Germany) at 4°C. The homogenate was centrifuged at 600 x g for 10 min at 4°C to remove nuclei and unbroken cells. The pellet was homogenized and centrifuged again under the same conditions. The resulting supernatant was transferred to another tube and centrifuged at 100,000 x g for 1 h at 4°C (Beckman, Ti 70 rotor) to obtain a plasma membrane-contained fraction, and resuspended in 0.2 M potassium phosphate (pH 7.8) for further analysis. Total proteins were quantified using the Bio-Rad RC-DC protein assay kit.
Preparation of OECs Conditioned Medium (OCM) Fraction
At the end of the primary culture of OECs, the conditioned medium was pooled (after 24-h incubation) purified through a series of centrifugation steps (200 x g for 5 min; 1000 x g for 10 min), and filtered through a 0.45-μm syringe filter to remove detached cells and cellular debris (16,26). The medium was concentrated using Amicon Ultra-4 Centrifugal Filter Devices with low-binding Ultracel membranes: 3000 NMWL (7500 x g for 20 min, twice). Total proteins were quantified using the Bio-Rad RC-DC protein assay kit.
SDS-PAGE and Western Blotting
SDS-PAGE protein gels were performed according to standard procedures. OPM and OCM were redissolved by adding a trace of bromophenol blue and heated at 100°C for 5 min. Samples were subjected to SDS-PAGE using 11.5% separation gel and 4.8% stacking gel. After electrophoresis, the gel was stained using blue silver. To prepare samples for Western blotting, OPM and OCM were lysed (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40 w/v, and 5 mM EDTA) and loaded at 50 μg/lane. The proteins in the gel were transferred to a PVDF membrane. The membrane was blocked with 5% nonfat dry milk in TBST (150 mM NaCl, 0.1% Tween-20, 25 mM Tris, pH 7.5) for 1 h at room temperature and then incubated with the appropriate antibody in the same solution for 1 h at room temperature. After washing with TBST, the membrane was incubated with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. The membrane was washed with TBST again and the bolt was developed using the Western lightning chemiluminescence reagent. Quantification of the band intensity was measured by densitometry analysis of the images using Quantity One 4.6.2 software (Bio-Rad).
MS Analysis and Identification of Proteins
The in-gel digestion was done according to our previous work (13,14,61). The final solution was pooled and lyophilized in a SpeedVac (Thermo Savant, NY) and dissolved in 30 μl of 0.1% formic acid (FA), then analyzed by a on-line Agilent 1200 capillary system (Agilent, Waldbronn, Germany) coupled to a high-capacity ion trap mass spectrometer (HCT Ultra™, Bruker Daltonics, Bremen, Germany) with an electrospray ionization source. Samples were desalted and concentrated at a flow rate of 20 μl/min on a short C18 precolumn Zorbax SB (500 μm i.d., 3.5 cm; Agilent) connected in the front of an analytical capillary column. For the separation with the C18 PepMap™ column (180 μm i.d., 15 cm; LC-Packings, Sunnyvale, CA, USA), a flow rate of 3 μl/min was used generated by a cap-flow splitter cartridge (3/500) and an initial pump flow rate of 500 μl/min. Each sample injection volume in all experiments was 30 μl. For the chromatography, the following solvents were used: solvent A (98% H2O, 1.9% acetonitrile, 0.1% formic acid) and solvent B (95% acetonitrile, 4.9% H2O, 0.1% formic acid). The eluting gradients used for separation of tryptic digests was 5% to 35% B in 147 min, 30% to 80% B in 15 min, followed by 80% B for 10 min and then by 5% B in 10 min. The mass spectrometer was operated in positive ion mode at a 4000 V capillary voyage. The nebulizer pressure was 10 psi. The flow rate of drying gas was 5 μl/min. The temperature of drying gas was 300°C. The capillary voltage was 4000 V. The full MS scan mode was standard enhanced (m/z 350 to 1600). The five most abundant ions detected in each MS scan were selected for collision-induced dissociation (MS/MS) with 1.0 V collision energy. The peptides were analyzed using the data-dependent MS/MS mode over the m/z range 100–2000. System control and data collection were done by EsquireControl™ software version 6.0 (Bruker Daltonics).
Database Searching
Raw spectrum data were processed and Mascot compatible files were created using DataAnalysis™ 3.4 software (Bruker Daltonics) with the following parameters: compounds threshold 10,000, maximum number of compounds 100,000, and retention time windows 1.0 min. Searches were performed using Mascot™ software 2.1 (Matrixscience, London, UK) and the international protein index (IPI) rat database (version 3.48, 40041 sequences; 20428766 residues, www.ebi.ac.uk) was used for peptide and protein identification. Search parameters were set as follows: enzyme, trypsin; allowance of up to one missed cleavage peptide; mass tolerance, 2.0 Da and MS/MS mass tolerance, 0.8 Da; fixed modification parameter, carbamidomethylation (C); variable modification parameters, oxidation (at Met); auto hits allowed (only significant hits were reported); results format as peptide summary report. We confirmed candidate proteins according to the probability-based Mowse scores. Proteins were generally identified on the basis of two or more peptides whose ions scores exceeded the threshold, p < 0.05, which indicated identification at the 95% confidence level. Most candidate proteins with Mascot scores above the threshold were identified based on at least two identified peptides. For proteins identified by only one peptide with Mascot score exceeding the threshold, their MS/MS spectrum was systematically checked manually. For a protein to be confirmed, (i) the masses of all the major peaks (typically more than seven peaks) in the MS/MS spectrum had to match those of the theoretically calculated fragment ions; (ii) the assignment had to be based on successive four or more bor y-series ions; (iii) the molecular weight of the matched protein was in reasonable agreement with the gel migration data (18,24). We kept to the principle of using the minimum set of protein sequences to account for all observed peptides. A Perl script was written in house to parse significant hits from Mascot output files (html files) into tab-delimited data files suitable for subsequent data analysis. An automated sequence retrieval script was written in Perl using the Bioperl libraries to generate FASTA formatted protein sequence from IPI databases for proteins identified by each MS experiment. The molecular mass values, pI values, and percentage of the protein covered by the matched peptides were retrieved from Mascot output files. The average hydrophobicity for identified proteins was calculated using the ProtParam software available at http://us.expasy-.org by submitting each FASTA file in batch. The proteins exhibiting positive grand average of hydropathicity (GRAVY) values were recognized as hydrophobic and those with negative values were deemed hydrophilic (32).
Gene Ontology Analysis and SignalP Prediction
To evaluate the specificity of our purification methods and examine the primary protein composition of OPM and OCM, all the identified proteins were analyzed with bioinformatics tools. As detailed above, mass spetrometry results were processed through EsquireControl, DataAnalysis, and Mascot tools to generate a list of detected proteins identified by their IPI numbers, which are unique for all proteins identified to date (both through protein sequencing and genome prediction). Information for each protein was gathered utilizing the IPI database of the European Bioinformatics Institute (http://www.ebi.ac.uk/IPI/IPIhelp.html) and the corresponding Entrez Gene database entry (http://www.ncbi.nlm.nih.gov/sites/entrez?db=gene). The theoretical molecular weight and isoelectric point (pI) of identified proteins were retrieved from Mascot output files. Mapping of transmembrane domains (TMD) for the identified proteins was conducted using the TMHMM 2.0 program based on transmembrane hidden Markov model (http://www.cbs.dtu.dk/services/TMHMM) by submitting the FASTA files. The subcellular location and function of the identified proteins were elucidated by gene ontology (GO) component and function terms, respectively, text-based annotation files of which were available for download from GO database ftp site: ftp://ftp.geneontology.org/pub/go/(3). The cellular localization for each protein was based upon GO compartment classifications. The IPI number for each identified protein was correlated to the respective GO protein information, which was used to assign protein localization categories. GO associates gene products to cellular components, molecular functions, and biological processes; these associations were assigned based upon experimental evidence or a suite of documented inferences based upon many criteria including direct assays, sequence or structural similarities, genomic context, and reviewed computational analysis. We searched the GO compartment localization information for each of the proteins identified in OCM and the SignalP 3.0 Server (8,43) was used to predict those unknown proteins that cannot be found in the location information in GO database (http://www.cbs.dtu.dk/services/SignalP-3.0/).
The plasma membrane proteins and extracellular proteins were associated with their biological processes and molecular functions using GO annotations through the GOTree Machine's algorithm (http://bioinfo.vanderbilt.edu/gotm/) (60). Additional analysis to assess the over-representation of proteins within each functional category was also carried out. Functional categories were considered overrepresented when the number of proteins in a given category were enriched relative to the same category in the whole rat proteome using an Expression Analysis Systematic Explorer algorithm that uses a variant of the Bonferroni probability test to score the overrepresented categories (p < 0.0001).
Network Analysis for the OPM and OCM
To quantify protein interaction properties of the secretome, we used the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database of physical and functional interactions (http://string.embl.de/) (28). The IPI identifiers of our OCM and OPM study were mapped to the Ensemble protein identifiers (638 in total), and then searched against the STRING database for protein-protein interactions. Only interactions between the proteins belonging to the extracellular matrix and plasma membrane dataset were selected. STRING defines a metric called “confidence score” to define interaction confidence; we fetched all interactions for our dataset, which had a confidence score ≥0.7 (high confidence).
Statistics Analysis
Values were expressed as means ± SEM of at least three separate experiments. Statistical analysis was performed using Student's t-test. All statistical tests were two-sided, and differences were considered significant when p < 0.05.
Results
OECs Characterization by Flow Cytometry
To ensure that the correct population of cells had been harvested from rat olfactory bulb, cells were stained for the expression of low-affinity nerve growth factor receptor (P75) and analyzed by flow cytometry. The gated population of viable OECs (circled in Fig. 1A) was shown to have positive expression of P75 (Fig. 1B). We conducted three parallel experiments at the same time and the gate value of OECs is 94%, 89%, and 97%.
Characterization of OECs isolated from 5-week-old Wistar rats using flow cytometry. OECs were stained with Alexa 488-conjugated anti-p75 antibody. (A) Bivariate forward scatter (size) versus side scatter (granularity). The elliptical region R1 shows the gated population of cells selected for sorting in (B). (B) The histogram shows p75 expression by OECs (A1, A2, A3) compared to autofluorescence of unlabeled cells (B1, B2).
Separation and Identification of OPM and OCM Proteins
The conditioned media from three separate primary rat OECs preparations and the same amount of plasma membrane were analyzed using a 1D-shotgun proteomics protocol with a HCT-Ultra ion trap as detailed in Materials and Methods. Due to its convenience and further purification of some dyestuff combined with proteins, SDS-PAGE was used to separate the isolated secreted proteins and membrane proteins. The overall separation profile is shown in Figure 2A. The staining patterns of the OCM and OPM samples were apparently different in the intensities of a number of bands. In addition, we have examined the distribution of β-tubulin, an abundant cytosolic protein in the two fractions. As shown in Figure 2B, β-tubulin was clearly detected in the total cell lysates (lane a) but not in the culture supernatants (lane b). This observation implies that recovery of proteins in the culture supernatants is not due to cell death. Each proteomic sample was analyzed in quadruplicate and the results were processed using a sequential analysis as described in the proteomic data processing section. Stringent criteria were used to identify each protein as present in OPM and OCM: each identified protein had to contain both a high protein probability score and a minimum of two unique peptides. Shotgun proteomics with these rigorous conditions identified a total of 470 different nonredundant proteins in OPM and 168 in OCM. The entire list of identified proteins is provided in homepage of our lab http://protchem.hunnu.edu.cn/Labweb/Resource/Resource_index.htm.
The SDS-PAGE analysis of the proteins isolated from OCM and OPM. (A) Proteins (50 μg) in both lanes were separated by SDS-PAGE. (B) Proteins of the cell extracts (lane a) and culture supernatants (lane b) from OECs were analyzed by Western blot with anti-β-tubulin antibody.
Characterization of the Identified Proteins in OPM and OCM
Of the 470 identified OPM proteins, 305(65%) were in the range of 10–60 kDa, which is the molecular mass distribution typically seen with 2DE-based methods, and 68 (14.6%) had a mass >100 kDa. The identified proteins were distributed across a wide pI range (4.3 to 11.78), and 8.1% of these proteins had pI >10 and 9.3% of these proteins had pI <5. Of the total identified proteins, 223 (47%) were unambiguously confirmed as PM proteins, 75 (33.6%) of which were predicted to have one or more predicted TMDs. We also analyzed the identified PM proteins on the basis of their calculated GRAVY value and predicted transmembrane domains. The GRAVY values of identified OPM proteins ranged from −1.131 to 0.954. These data indicate that, with our combined analytical strategies, the proteins with various physicochemical characteristics can be efficiently identified (Fig. 3).
Physicochemical characterization of proteins identified in OPM and OCM preparation. (Top) Calculated molecular weight; (bottom) calculated isoelectric point. The informations are from the GO database and Swiss-Prot and TrEMBL.
Bioinformatics Analysis of the Identified OPM and OCM Proteins
In addition to the above-mentioned 103 (22%) PM proteins, 41 (9%) of the 470 identified proteins were categorized as membrane/integral to membrane proteins. Other identified proteins with a subcellular location annotation were annotated as PM-associated (9%), mitochondrial (12%), cytoplasmic (12%), endoplasm reticulum/golgi (9%), and nuclear (2%). Besides, 118 proteins (25%) did not have a subcellular location annotation and were categorized as unknown proteins (Fig. 4A). It is worthy of mentioning that, in fact, the number of PM proteins identified in the present study should be greater than 103, because some of proteins annotated in general terms as “membrane/integral to membrane proteins” were also PM proteins.
The primary subcellular localization of the identified plasma membrane and extracellular matrix proteins. The location information is from the GO database and Swiss-Prot and TrEMBL. The unknown protein that cannot locate from both databases is predicted by SignalP tool.
Of the 168 identified proteins from OCM, 51 (30.1%) were recognized as proteins with GO extracellular localizations or located by Swissprot as secreted protein; 53 (31.4%) were proteins that were predicted from the rat genome but have not yet undergone further characterization or GO component assignment. So we try the SignalP tool to predict those unknown proteins and found that 25 (47.2% of unknown and 14.7% of the fully identified proteins) were proteins that have corresponding amino acid to be a part of a signal peptide (Fig. 4B). The remaining parts of the identified proteins included DNA binding, nucleic acid binding (0.6%), myosin (4.5%), ribosome (2.6%), cytoskeleton, microtubule (7.7%), nucleus (5.8%), microsome (0.6%), endoplasmic reticulum (1.3%), cytoplasm (2.6%), mitochondrion (1.9%), and membrane associated (2.6%).
The hydrophobicity property of proteins is often expressed as the GRAVY index and the ProtParam tool at ExPASy was used to calculate the GRAVY value of each of the proteins. The proteins exhibiting positive GRAVY values were recognized as hydrophobic and those with negative values were deemed hydrophilic. According to the identified extracellular proteins in our process, the GRAVY values varied from −1.0692 to 0.6741 and 149 (88.7%) of them are negative. It shows that most of our identified proteins are hydrophilic proteins, which could have chance to be secreted to the extracellular matrix and function as growth factor or cell signaling.
Signal Pathway Analysis Between OCM and OPM
In the global STRING-generated protein–protein network that include the plasma membrane proteins and the extracellular matrix proteins, several complexes and cellular functions formed prominent, tightly connected clusters as assessed by means of molecular complex detection. The protein–protein interaction network of our acetylating dataset was extracted from STRING. The four most highly ranked and tightly connected network clusters as obtained by MCODE analysis (7) are color coded and later rendered as separate modules. The resulting interactome had 451 nodes and 1,999 interactions, which we call the OECs interaction network (OIN) (Fig. 5).
Interaction networks of membrane proteins and extracellular matrix proteins in different cellular functions from STRING analysis. The protein–protein interation map was draw using the Cytoscape Ver 2.6. Individual networks generated for each specific functions can be seen in Table 1.
Discussion
The finding that OECs are a source of multiple trophic factors became very important because they play a decisive role in CNS regeneration (9,33). Although not fully understood, the role of OECs in neuronal regeneration and endogenous CNS repair functions is now largely acknowledged. Since the original work by Richardson et al. in 1980 (46), an ever growing number of researches are focusing on the OECs through immunoassay, gene chip, and, most recently, proteomics process. In vitro coculturing of adult OECs with embryonic chick ganglia demonstrated increased neurite outgrowth, and transplantation of OECs into the injured CNS facilitates sparing/regeneration of axons, which can lead to improved function (36,49). It was able to promote axonal regeneration and functional recovery when transplanted 45 days after complete transection of the thoracic spinal cord in adult rats (37). Furthermore, some studies showed that olfactory mucosa autograft transplantation into the human injured spinal cord was considered as feasible, relatively safe, and potentially beneficial (35). In the proteomic study of olfactory ensheathing cell- and Schwann cell-expressed proteins by Boyd et al. (10), 40 soluble proteins were identified through a 2D-DIGE procedure followed with MALDI-TOF. They identified calponin, an actin binding protein, as the first definitive phenotypic marker that distinguishes between OECs and SCs in vitro and in vivo. Other investigators also reported a functional proteomic approach, isotope-coded affinity tagging and mass spectrometry, to identify active components of the OECs secretome that stimulate outgrowth. SPARC (secreted protein acidic rich in cysteine) was identified as a OECs-derived matricellular protein that can indirectly enhance the ability of Schwann cells to stimulate dorsal root ganglion outgrowth in vitro (4). While several OEC-released factors have been described, a global characterization and classification of them is missing.
We decided to focus on proteins actively secreted by OECs because they are known to express a variety of growth factors and cell adhesion molecules and promote axonal regrowth and functional recovery in spinal cord injury in animal models and patients. The characterization of mediators involved in this function is a primary interest of our laboratory. We also start the transplantation of OECs to enhance axonal conduction and partially recovery of injured spinal cord function. It is possible that some of the proteins identified by our proteomics protocol were released by cells other than OECs present in the cultures. However, this is somehow unlikely considering that, in order to carry out the described analyses, we pulled and concentrated media derived from over 10 dishes of confluent cells, the vast majority of which (90% or more) were OECs. It is therefore very unlikely that the released proteins in our experiment came from a much smaller population of cells or did not have the function we are interested in.
In comparison with the former proteome research on OEC function, our study identified 168 proteins in OCM, and represents to date the most complete list of proteins secreted by OECs, although it still does not include all the proteins and peptides released by these cells. Most notably, this analysis did not identify some growth factors known to be released by OECs. Most of the growth factors are small proteins or peptides that are released transiently and in low amounts after the activation of specific receptors and are usually promptly degraded; their low molecular weight, low abundance, and their transient presence in the extracellular environment may explain why they were not detected in our analysis. In spite of this limitation, our proteomics represents, to date, one of the best available tools of cell secretome. In addition, the technique employed in this analysis is more sensitive than the two-dimensional gel approach in detecting proteins present at low levels, though the 2D-DIGE is still a great protocol in differential proteomics.
Network analysis of conditioned medium and plasma membrane proteins from OECs carried out by cellular process revealed that several complexes and cellular functions formed prominent, tightly connected clusters as assessed by means of molecular complex detection (see Table 1). We focused on proteins in these clusters that are located in the extracellular matrix and several of them are involved in the process of cell differentiation (metallopeptidase inhibitor 1, clusterin, dihydropyrimidinase-related protein 2, phosphatidylethanolamine-binding protein 1, Ras-related C3 botulinum toxin substrate 1), growth regulation (CD81 antigen, protein kinase C-binding protein NELL2, attractin), and enzyme activity regulator (metallopeptidase inhibitor 1, isoform 1 of fibronectin, serpinf1, phosphatidylethanolamine-binding protein 1). Among them, the apolipoprotein E and insulin-like growth factor-binding protein 2 are involved in more than one process.
Selected Extracellular Proteins Relative in OCM and OPM Network Analysis Cluster
The IPI number of the identified protein.
The highest Mascot score for the identified protein.
The unique peptide number for the identified protein.
The hydrophobicity property of proteins. Calculated by the ProtParam tool at ExPASy.
Apolipoprotein E (ApoE) belongs to the apolipoprotein A1/A4/E family. The different alleles of the ApoE gene have been reported to influence recovery after traumatic brain injury (TBI) in both human patients and animal models, with the e4 allele typically conferring poorer prognosis for recovery (19). It interacting with LDL receptors is profoundly involved in lipid transport of CNS for tissue repair in the peri-ischemic area after brain ischemia (29). Recently, a report indicated ApoE plays a significant role in lipid metabolism and has been implicated in the growth and repair of injured neurons (39). Insulin-like growth factor-binding protein 2 (IGF-BP2) is involved in the process of regulation of cell growth. The insulin-like growth factor (IGF) system is a key regulator of cell growth, survival, and differentiation. Insulin-like growth factors (IGFs) protect neurons, are important for oligodendrocyte survival and myelin production, and stimulate the proliferation of astrocytes (15). IGF-binding proteins prolong the half-life of the IGFs and have been shown to either inhibit or stimulate the growth-promoting effects of the IGFs on cell culture. They alter the interaction of IGFs with their cell surface receptors (11). Increased microglia IGF-BP2 expression correlates with the timing of a number of posttraumatic events within the CNS, suggesting that IGF-BP2 may have a role as neuroprotection for surviving neurons and signal for local neuronal sprouting, as well as a role in reactive astrogliosis. In the injured CNS, increased IGF-BP2 expression may act to maintain or transport IGF-1 or IGF-2, as well as modulate the local autocrine and paracrine actions of the IGFs (20).
Metallopeptidase inhibitor 1 (Timp-1) released by OECs promotes the neurite differentiation and acts as an inhibitor of the enzyme activity. It complexes with metalloproteinase (such as collagenases) and irreversibly inactivates them. In the process of hepatic fibrosis, fibroblasts and myofibroblasts are the major cells that express Timps and, the more serious the hepatic fibrosis is in the injured liver, the higher the level of Timp-1 and Timp-2 gene expression (42). Clusterin (Clu) is a multivalent glycoprotein with ubiquitous tissue distribution. It may play a role in apoptosis in the seminal vesicle and other organs. The association of clusterin with elastic fibers could reflect an extracellular chaperone function by either protecting elastic fibers or shielding abnormal elastic material (1). Clusterin production and secretion, particularly by astrocytes, could be neuroprotective, perhaps via its putative actions such as inhibition of complement activation and cytolysis, effects on chemotaxis and apoptosis, and actions as an antistress protein (56).
Dihydropyrimidinase-related protein 2 (Dpysl2) belongs to the DHOase family. It plays a role in axon guidance, neuronal growth cone collapse, and cell migration. Former research showed it is expressed immediately after neuronal birth and is dramatically downregulated in the adult and necessary for signaling by class 3 semaphorins (62). Its association with neurofibromin and CRMP-2 is essential for neuronal cell differentiation. Because both neurofibromin and CRMP proteins are involved in proliferation and axonal morphogenesis, these results suggest that the Dpysl2 interaction with CRMPs contributes to the function of neurofibromin in tumorigenesis and neuronal morphogenesis (44). Ras-related C3 botulinum toxin substrate 1(Rac1) belongs to the small GTPase super family. It is a plasma membrane-associated small GTPase that cycles between active GTP-bound and inactive GDP-bound states. In its active state, it binds to a variety of effector proteins to regulate cellular responses such as secretory processes, phagocytosis of apoptotic cells, epithelial cell polarization, and growth factor-induced formation of membrane ruffles. Data suggest that activation of Rac1 is differentially regulated in specific neuroblastoma cellular regions, perhaps contributing to the complexity of Rho GTPasemediated axon growth (52).
CD81 antigen (Cd81) is temporally upregulated after neuron injury. It plays a role in glial scar formation in response to astrocyte injury and may be involved in regulation of cell interactions and cell growth (53). Protein kinase C-binding protein NELL2 (Nell2) is a homotrimeric glycoprotein that is involved in the growth and differentiation of neural cells. Nell2 may play an important role in the development of the CNS as well as maintenance of neural functions, by regulating the intracellular machinery involving Ca2+ signaling, synaptic transport, and/or release of vesicle (31). Attractin (Atrn) is more widely expressed throughout the CNS than previously reported, and expression of Atrn by various cell types suggests that Atrn may serve multiple functions in the CNS (41). Lack of attractin gene expression induced neurodegeneration by a decrease in active extracellular signal-regulated kinase through an intracellular signaling via oxidative stress (40).
Isoform 1 of fibronectin (Fn1) binds cell surfaces and various compounds including collagen, fibrin, heparin, DNA, and actin. Fibronectin is involved in cell adhesion, cell motility, opsonization, wound healing, and maintenance of cell shape. These data indicate that fibronectin polymerization is a critical regulator of extracellular matrix organization and stability. Recent research suggests that spinal fibronectin is elevated after the peripheral nerve injury and it may be involved in the upregulation of the P2X4 receptor in microglia, which leads to the induction of neuropathic pain (54). Alpha-2 antiplasmin (serpinf1, PEDF) belongs to the serpin family. In vivo evidence suggests that the induced in vivo expression of PEDF is effective in protecting CNS neurons from ischemic insult and may have a role as an inducible endogenous neurotrophic factor in the CNS (50). PEDF is a multipotent factor, capable of affecting not only neurons, but also neonatal astrocytes, and suggests that it may act as a neuroimmune modulator in the developmental brain (58).
Here we show the protein profile of OPM and OCM, which to our knowledge identified the largest number of proteins in the OECs proteome. Most of the OCM have been involved in neuronal regeneration and significantly overexpressed in the related molecular function pathway on the plasma membrane. Therefore, the proteomic profiling of the OECs proteome in serum-free culture conditions provides new insights into the nature of their in vivo microenvironmental niche. Although proteins identified in vitro may not necessarily be a bona fide representation of the in vivo microenvironment, the nature of the secretome profile indicates that the extraction and culture of OECs from their natural in vivo microenvironment may to some extent simulate olfactory neurogenesis in vitro. Our ongoing research now focuses on the expression of these molecules in vivo and the role that these molecules may play during development and regeneration of the primary olfactory system. It is hoped that these experiments will increase our understanding of OEC biology and provide further insight into novel molecules that produced by these cells, which could play a crucial role during CNS regeneration.
Footnotes
Acknowledgments
The authors thank all the members of our group for kindly help and suggestions. This work was supported by a grant from National 973 Project of China (2007CB516809, 2007CB914203), and National Natural Science Foundation of China (30770437).