Title: Laminin-5 activates extracellular matrix production and osteogenic gene focusing in human mesenchymal stem cells

We recently reported that laminin-5, expressed by human mesenchymal stem cells (hMSC), stimulates osteogenic gene expression in these cells in the absence of any other osteogenic stimulus. Here we employ two dimensional liquid chromatography and tandem mass spectrometry, along with the Database for Annotation, Visualization and Integrated Discovery (DAVID), to obtain a more comprehensive profile of the protein (and hence gene) expression changes occurring during laminin-5-induced osteogenesis of hMSC. Specifically, we compare the protein expression profiles of undifferentiated hMSC, hMSC cultured on laminin-5 (Ln-5 hMSC), and fully differentiated human osteoblasts (hOST). We find a marked reduction in the number of proteins (e.g., those involved with calcium signaling and cellular metabolism) expressed in Ln-5 hMSC compared to hMSC, consistent with our previous finding that hOST express far fewer proteins than do their hMSC progenitors, a pattern we call “osteogenic gene focusing.” This focused set, which closely resembles that expressed by mature hOST, includes osteogenic extracellular matrix proteins (collagen, vitronectin) and their integrin receptors, calcium signaling proteins, and enzymes involved in lipid metabolism. These results provide direct evidence that laminin-5 alone stimulates global changes in gene/protein expression in hMSC that lead to commitment of these cells to the osteogenic phenotype, and that this commitment correlates with extracellular matrix production. Introduction Human mesenchymal stem cells (hMSC) are a population of multipotent cells found within the bone marrow and periosteum (Barry et al., 2004). Their ability to differentiate into at least three, and possibly as many as seven, different cell types (Pittenger et al., 2004) makes them attractive tools for tissue engineering and cellular models of development. The molecular and biochemical mechanisms governing hMSC differentiation and commitment are not well understood, especially in response to extracellular matrix (ECM) protein binding. The laminin (Ln) family of ECM proteins are ubiquitously expressed but are especially abundant in the basement membrane of many epithelial and endothelial tissues, where they mediate cell attachment, migration, and tissue organization in conjunction with other ECM proteins (Malinda et al., 1996). Each laminin molecule is a heterotrimer, composed of an α-, β-, and γsubunit. The subunits share homology with one another and form an asymmetric cross-like structure with one long and three short arms joined by disulfide bonds (Colognato et al., 2000). The Ln-5 isoform is composed of α3, β3, and γ2 subunits. Expression of the γ2 subunit has only been found in Ln-5, while the α3 subunit is found in both Ln-6 and Ln-7. The role of Ln family members in osteogenic differentiation is not known (Roche et al., 1999), though expression of the γ2 chain has been previously detected in bone marrow (Siler et al., 2002). Though Ln-5 is typically only found in tissues derived from endoderm and ectoderm, we have recently discovered expression and roles for Ln-5 in mesoderm tissues; in controlling the growth and migration of vascular smooth muscle cells (Kingsley et al., 2001;Kingsley et al., 2002a;Kingsley et al., 2002b) and in promoting an osteogenic phenotype in hMSC (Klees et al., 2005). These observations raise the question as to how Ln-5 exerts these effects in mesodermal tissues, especially with respect to changes in cellular phenotype. Recent advances in proteomic approaches have greatly improved the ability to find new markers for cellular differentiation. In particular, two-dimensional liquid chromatography tandem mass spectrometry (2D LC-MS/MS) is a powerful approach for identifying protein constituents in cell populations. For example, mass spectrometry reveals differences in tyrosine phosphorylation of hMSC proteins in response to epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and transforming growth factor β1 (TGFβ1) (Wang et al., 2004;Kratchmarova et al., 2005), and suggests that phosphatidylinositol 3-kinase is a possible control point in the osteogenic differentiation process. Mass spectroscopic profiling of proteins expressed in ECMstimulated hMSC has thus far not been reported. We recently used advanced proteomics and mass spectrometry to compare the protein expression profiles of undifferentiated hMSC, hMSC induced to differentiate into osteoblasts using osteogenic stimulating (OS) media, and fully differentiated osteoblasts (Salasznyk et al., 2005a). We found substantial changes in clusters of functionally related hMSC proteins in response to OS stimulation, and identified a set of related proteins that discriminated osteoblasts and OS-treated hMSC from undifferentiated hMSC. Here, we employ a similar strategy to identify proteins expressed exclusively in Ln-5 treated hMSC (Ln-5 hMSC), and observe that Ln-5 acts to focus protein expression in hMSC to more closely resemble differentiated osteoblasts, consistent with our observation that Ln-5 activates osteogenic gene expression (Klees et. al, 2005). Results To identify the proteins expressed during Ln-5 induced osteogenic differentiation of hMSC and compare them to the protein expression profiles of undifferentiated hMSC and physiologically differentiated hOST, we performed 2D LC-MS/MS on whole-cell lysates of these cell populations. Of the 765 different proteins (≥200 pmol) identified by 2D LC-MS/MS in all three cell populations, 121 were found only in hMSC; 151 were found only in Ln-5 hMSC, and 91 were unique only to mature hOST (Figure 1). These uniquely expressed proteins accounted for 16%, 20%, and 12% of the total proteins identified in each cell population, respectively. In addition, 128 (17%) were shared by hMSC and hOST, 37 (5%) were found in both hMSC and Ln-5 hMSC, and 16 (2%) were present in both Ln-5 hMSC and hOST. Finally, 106 (14%) were found in all three. These results suggest that, in general, Ln-5 hMSC represent a distinct, intermediate stage between hMSC and hOST, consistent with our previous finding that OS-treated hMSC share partial expression profiles with both ends of the osteogenic differentiation spectrum (Salasznyk et. al, 2005a). To evaluate the functional significance of these differences, we accessed the GO (Gene Ontology) Chart featured by DAVID to generate the Biological Process and Molecular Function protein distributions shown in Figure 2. A detailed summary of the findings for each class of proteins is presented here, although readers are encouraged to visit http://www.rpi.edu/~bennek/TissueEng/ln5.html for the complete listing of all reports and text files, including, when applicable: each protein’s accession identification number, molecular weight, GeneCard and Entrez Gene links, links to possible function in stem cells and bone, along with a summary of the protein and its function. Data sorting 945 different proteins were identified by 2D LC MS/MS; 765 of these were identified with accession numbers (in Ensemble, Refseq, or Trembl format) and a protein name. To determine protein functional relationships within and between each data set, we used the Database for Annotation, Visualization and Integrated Discovery (DAVID) (located at http://apps1.niaid.nih.gov/david/). DAVID is a Web-based, client/server application that allows users to access a relational database of functional annotation derived primarily from LocusLink at NCBI. LocusLink has been superceded by Entrez Gene at NCBI. Through DAVID, proteins identified with a LocusLink number or equivalently an Entrez Gene GeneID can be grouped by Gene Ontologies and annotated with gene names and aliases as well as functional summaries. Because DAVID does not accept the accession numbers provided by the Protein Server program, each protein was also matched to its GeneID number (http://www.ncbi.nih.gov/entrez/query.fcgi?db=gene), if available, by accessing the European Bioinformatics Institute (http://www.ebi.ac.uk/IPI/IPIhelp.html), downloading the ipi.HUMAN.xrefs.gz IPI dataset (ftp://ftp.ebi.ac.uk/pub/databases/IPI/current/), and matching each accession number to its corresponding GeneID number. Proteins that were not assigned a GeneID using this method were searched by hand using the online search function (IPI Quick Search) and the search feature located at Entrez Gene by entering available accession numbers and the protein name. The proteomics data were parsed so that seven data sets were identified: Ln-5 (proteins appearing in the Ln-5 hMSC but not in hMSC or hOST cells), hMSC (proteins appearing in the hMSC cells but not in Ln-5 hMSC or hOST cells), hOST (proteins appearing in the hOST cells but not in hMSC or Ln-5 hMSC), Ln-5-hMSC (proteins appearing in the Ln-5 hMSC and hMSC cells but not in hOST cells), Ln-5-hOST (proteins appearing in the Ln-5 hMSC and hOST cells but not in hMSC cells), hMSChOST (proteins appearing in the hOST and hMSC cells but not in Ln-5 treated cells), and ALL (proteins appearing in hOST, Ln-5 hMSC, and hMSC cells). The GeneID of proteins in the seven data sets were categorized in Biological Process and Molecular Function gene ontology categories using the GO (Gene Ontology) Chart feature offered by DAVID 1.0 (http://www.geneontology.org/GO.nodes.html). The GO categories were determined for each data set and then unioned to form a complete list. DAVID was set at intermediate coverage and specificity [Level 3] with a minimum of 4 GeneIDs. This setting was found to best balance coverage and specificity, but not all of the proteins were assigned to a GO category. The 650 proteins with known GeneIDs assigned to GO Categories are known as the CLASSIFIED proteins. The 73 proteins with GeneIDs and no assigned GO category are known as UNCLASSIFIED. The 42 that were not able to be linked to GeneIDs are known as the set of UNCLASSIFIABLE proteins. To facilitate investigation of the GO categories, each GeneID was assigned a gene symbol, a summary, and a list of gene ontologies using DAVID 2.0’s Annotation tool. A relational database management system (DBMS) was used to store, link, and analyze the annotated proteins, annotated GeneIDs, and gene ontology categories. The DBMS used was Microsoft Access 2000, and the database itself consists of 13 tables (comprising over 16000 records), 67 queries, 12 reports, two macros, and two programming modules. Reports were written and generated, using the DBMS facilities, to produce the statistics reported in this paper. In addition, the database was programmed to produce HTML reports in a format suitable for public use, as well as to generate other reports and information used in our analysis. The database was used to categorize the proteins (and their associated gene symbols, GeneIDs, external links, and functional summaries) into the three groups (CLASSIFIED, etc.) described above, and to generate one report per group. The web-based reports are dynamic in the sense that links are provided for each protein that automatically launch a PubMed query (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi) based on the protein name and key words or phrases such as “bone” and “stem cell”, among others. There are also links, for each protein, to its GeneCard and Entrez Gene entries. The report allows the biologist to rapidly examine all proteins that belong to a specific GO category, observe how they were distributed in the seven data sets, and analyze the functionality of the proteins. All of the web reports and text files are available at http://www.rpi.edu/~bennek/TissueEng/ln5.html. Westernblotting and RT-PCR Cells were extracted with RIPA buffer, suspended in 4X Laemmli buffer, and subsequently denatured at 100°C for 5 min. Denatured extracts were resolved by 8% SDS-PAGE and transferred to nitrocellulose. Membranes were blocked with 5% nonfat dried milk and 0.2% Tween-20 for 1 h and then probed with indicated primary antibodies (1:500) overnight at 4°C. After three washes with phosphate-buffered saline (PBS) containing 0.2% Tween-20, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary immunoglobulin G (IgG) (1:10,000) for 1 h, after which they were washed three times and detected by enhanced chemiluminescence. RNA was isolated after 16 days in culture for all conditions using the RNeasy mini kit (Qiagen, Valencia, CA). RT-PCR was performed with the OneStep RT-PCR Kit (MCLAB) and a 96 well thermal cycler (MJ Research, Waltham, MA) using the following primers designed by the Lasergene v5.0 program (DNASTAR, Madison, WI): CD-81 forward 5’-GCCCCCGCGCCCCTTTCTTC-3’ reverse 5’GGATTCCTGGATGGCCCCGTAGCA-3’; β-tubulin forward 5’CCGGGAGGCAGATGGTAGTGACAG-3’ reverse 5’GAGCCGTGGGGTGGGAATGAGC-3’; fibronectin forward 5’TCTGTAGGCCGTTGGAAGGAAG-3’ reverse 5’AGGCGCTGTTGTTTGTGAAGTAGA-3’. 100 ng of template RNA was used per reaction. The reverse transcription step ran for 30 min at 50 ̊C, followed by PCR activation for 15 min at 95 ̊C. Thirty amplification cycles were run, consisting of 1 min denaturation at 94 ̊C, 1 min of annealing at 60 ̊C, and 1 min of extension at 72 ̊C. Final extension was allowed to run 10 min at 72 ̊C. Reaction products were separated by gel electrophoresis using a 1% agarose gel. Bands were visualized by UV illumination of ethidium-bromide-stained gels and captured using a ChemiImager 4400 Gel imaging system (Alpha Innotech, San Leandro, CA).