Network Analysis for the Identification of Differentially Expressed Hub Genes Using Myogenin Knock-down Muscle Satellite Cells

Muscle, a multinucleate syncytium formed by the fusion of mononuclear myoblasts, arises from quiescent progenitors (satellite cells) via activation of muscle-specific transcription factors (MyoD, Myf5, myogenin: MYOG, and MRF4). Subsequent to a decline in Pax7, induction in the expression of MYOG is a hallmark of myoblasts that have entered the differentiation phase following cell cycle withdrawal. It is evident that MYOG function cannot be compensated by any other myogenic regulatory factors (MRFs). Despite a plethora of information available regarding MYOG, the mechanism by which MYOG regulates muscle cell differentiation has not yet been identified. Using an RNA-Seq approach, analysis of MYOG knock-down muscle satellite cells (MSCs) have shown that genes associated with cell cycle and division, DNA replication, and phosphate metabolism are differentially expressed. By constructing an interaction network of differentially expressed genes (DEGs) using GeneMANIA, cadherin-associated protein (CTNNA2) was identified as the main hub gene in the network with highest node degree. Four functional clusters (modules or communities) were identified in the network and the functional enrichment analysis revealed that genes included in these clusters significantly contribute to skeletal muscle development. To confirm this finding, in vitro studies revealed increased expression of CTNNA2 in MSCs on day 12 compared to day 10. Expression of CTNNA2 was decreased in MYOG knock-down cells. However, knocking down CTNNA2, which leads to increased expression of extracellular matrix (ECM) genes (type I collagen α1 and type I collagen α2) along with myostatin (MSTN), was not found significantly affecting the expression of MYOG in C2C12 cells. We therefore propose that MYOG exerts its regulatory effects by acting upstream of CTNNA2, which in turn regulates the differentiation of C2C12 cells via interaction with ECM genes. Taken together, these findings highlight a new mechanism by which MYOG interacts with CTNNA2 in order to promote myoblast differentiation.


Introduction
Skeletal muscle, one of the most highly organized structures in the body, acts as a source of power for locomotion and other daily activities essential for survival. In vertebrates, development of skeletal muscle that commences at the embryonic stage ends only after postnatal growth during which an organism attains its fully developed size [1]. Skeletal muscle is unique in that this tissue arises from the fusion of mononuclear myoblasts accompanied by the expression of several myogenic regulatory factors (MRFs) following cell cycle exit to ensure the coordinated response to neural input [2][3][4][5][6][7]. Among basic helix-loop-helix (bHLH) and MADS-box families of MRFs (MYOD, MYF5, MRF4, and myogenin: MYOG) that play a critical role in myogenic differentiation, MYOD and MYF5 specifically play redundant roles during myoblast proliferation [8]. MYOG is responsible for terminal differentiation and cannot be compensated by other MRFs [9][10][11][12]. MYF5, MYOD, and MRF4 also spur the expression of genes that are essential for muscle satellite cells (MSCs) proliferation [8,13,14].
MSC progeny can be distinguished from their quiescent progenitors based on distinctive gene expression patterns. In adults, MSCs cycle through the steps of embryonic myogenesis to either add to or replace current muscle fibers [15][16][17][18][19]. Unlike the enigmatic status of genes that perform important functions in bovines, expression of a large number of genes (particularly those corresponding to different transcription factors) has been observed in mouse MSCs [20]. Therefore, it is important to delineate the expression profile of genes with unknown function in bovine-derived MSCs. Our interests in obtaining the regulatory profile of genes with important functions in mouse MSCs led us to perform the current investigation with bovine MSCs to have a clear understanding of bovine muscle development. By employing microarray, expressed sequence tag (EST) followed by RNA-Seq techniques to MSCs satellite cell analysis, we were able to delineate the regulatory network of genes corresponding to different transcription factors and certain prominent members of the extracellular protein family, involved in controlling myoblast differentiation [20,21]. We were able to elucidate and assign specific roles to certain genes, such as transthyretin, that are novel with respect to their involvement in myogenesis [22]. While investigating the gene expression profile of MYOG knock-down (MYOG kd ) in bovine MSCs using RNA-Seq, we observed differential expression patterns of many genes, particularly those involved in biological pathways such as cell proliferation and DNA replication (up-regulated) or phosphate metabolic processes (down-regulated) [23].
A large number of high-throughput studies have been performed to explore the functional roles of genes found to have altered expression patterns during skeletal muscle differentiation [5,[24][25][26]. Microarray [12,27,28] along with RNA-Seq [29] studies have improved our knowledge of myogenesis by identifying diverse types of target genes encoding myogenic transcription factors or novel myogenic regulatory factors which govern the fusion of myoblasts into myotubes that were otherwise hard to detect with conventional methods. In spite of these improvements, the roles that these genes play in skeletal muscle development as well as the understanding of underlying molecular mechanisms are still poorly understood. Therefore, we performed the current study to investigate the molecular mechanism underlying muscle cell differentiation by identifying genes involved in myogenesis. We created an interaction network of differentially expressed genes (DEGs) as well as additional related genes predicted by Gene-MANIA, identified hub genes in the network based on their node degree distribution, and performed a functional study using gene knock-down. Furthermore, we identified functional modules in this network of DEGs, and an enrichment analysis of these modules was carried out by the Database for Annotation, Visualization and Integrated Discovery (DAVID) functional analysis tool. Our findings offer new insights into the role of hub genes in skeletal muscle development that will help to develop strategies for improving the meat quality and in combating muscle diseases.

Datasets and network analysis
For this study, 230 up-and 223 down-regulated genes identified with bovine MYOG kd RNA-Seq data were analyzed [23]. Functional interactions between these DEGs were predicted by the GeneMANIA webserver [30]. In addition to the DEGs, 50 additional genes were used to create the interaction network using the Gene Ontology (GO) term "biological process" and Homo sapiens as a source species. Three interaction networks were created using up-regulated genes, down-regulated genes, and a combination of the up-and down-regulated genes. Relationships of genes in the network in terms of coexpression, physical and genetic interactions, pathways, colocalization, protein domain similarity, and predicted interactions were evaluated.

Detection of hub genes
In scale-free biological networks [31], nodes having large number of interacting partners represent hubs in the network. Hubs were detected by calculating the node degree distribution [32] using the Network Analyzer (http://apps.cytoscape.org/apps/networkanalyzer) plugin of Cytoscape.

Community analysis and comparison of different clustering methods
The three interaction networks were visualized with Cytoscape 2.8.2 [33]. We also compared three widely used network clustering algorithms including MCL [34,35], MCODE [36], and the greedy algorithm (GLay) [37] using clusterMaker [38] and GLay plugins [39]. To explore the enriched biological functions within each cluster, the detected clusters were subjected to functional enrichment analyses by the DAVID tool (http://david.abcc.ncifcrf.gov/home.jsp). For the enrichment analysis, only communities with at least 10 nodes were evaluated.

Bovine MSCs culture
Bovine skeletal muscles were collected from the hind leg of 24-to 26-month old Korean cattle [40] with a body weight of 550-600 ㎏. The animals were handled according to a protocol approved by the Animal Care and Concern Committee of the National Institute of Animal Science (Republic of Korea). The collected muscle tissue was minced and digested with trypsin-EDTA (Gibco, CA, USA), centrifuged at 90 × g for 3 min, and the upper phase was passed through a 40-㎛ cell strainer (Falcon, NY, USA). Following centrifugation of the filtrate at 2,500 rpm, the resulting pellet was collected, washed with Dulbecco's modified Eagle's medium (DMEM; HyClone Laboratories, UT, USA), and cultured at 37°C under 5% CO 2 in DMEM supplemented with 10% fetal bovine serum (FBS; HyClone Laboratories) and 1% penicillin/ streptomycin (P/S). The culture medium was changed every other day. To induce differentiation, the cells were allowed to grow without reducing serum (DMEM with 10% FBS and 1% P/ S) for 10, 12, 14, and 16 day. Bovine MSCs used in this study were provided by the Bovine Genome Resources Bank of Yeungnam University (Gyeongsan, Republic of Korea).

shRNA construction and MYOG knock-down in bovine MSCs
Bovine MYOG shRNA was designed using nucleotide sequence information obtained from National Center for Biotechnology Information (NCBI; AB257560) and cloned with a pRNAT-U6.2/Lenti vector (GeneScript, NJ, USA). MYOG shRNA or scrambled vector (MYOG wt ) were transfected in 293 FT cells using CaCl 2 to generate viral particles. After 2 days of transfection, the supernatant containing viral particles expressing shRNA specific for bovine MYOG or the scrambled vector was collected and used to transfect MSCs (Day 8). The transfected cells were selected with 50 ㎍/㎖ of G418 (Calbiochem, CA, USA). The selected cells were allowed to differentiate and harvested at Day 21. The following oligonucleotide was used to generate MYOG shRNA: 5 0 -GGATCCCGCGCAGACTCAAGCCGCCGGTGTTCAAGA GACACCTTCTTGAGTCTGCGCTTTTCCAACTCHGAG-3 0 (sense). 293 FT cells were procured from Korea University Dr. Chun's lab, Republic of Korea.

C2C12 cell culture
Mouse C2C12 cells (kindly provided by the Korean Cell Line Bank, Republic of Korea) were cultured in DMEM (HyClone Laboratories) supplemented with 10% FBS (HyClone Laboratories) and 1% P/S (Invitrogen, CA, USA) at 37°C in 5% CO 2 . For differentiation, cells grown to 70% confluence were incubated with differentiation media (DMEM with 2% FBS), and cultured for 0, 2, 4, and 6 days during which time the medium was changed every 2 day.

MYOG and CTNNA2 knock-down in C2C12 cells
MYOG expression was knocked down using shRNA. C2C12 cells grown to 30% confluence were transfected with 1 ng of vector containing MYOG shRNA or the scrambled vector using transfection reagent (Santa Cruz Biotechnology, CA, USA). Transfected cells treated with 2 ㎍/㎖/ puromycin (Santa Cruz Biotechnology) for selection were grown to 70% confluence before switching to differentiation media. MYOG knock-down cells were cultured in differentiation media. Detailed information describing the shRNA sequences is provided in A) in S1 Table. For CTNNA2 knock-down using siRNA, C2C12 cells grown to 50% confluence were transfected with 100 μM of control or CTNNA2-specific siRNA (Thermo, Waltham, CO, USA) using Lipofectamine (Invitrogen) in transfection medium (Gibco, Grand Island, NY, USA). The transfected cells were grown to 70% confluence before switching to differentiation media. CTNNA2 knock-down cells were cultured in differentiation media. Detailed information describing the siRNA sequences is provided in B) in S1 Table. RNA extraction and real-time RT-PCR Cells were harvested with TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Total RNA was extracted and stored in diethylpyrocarbonate (Sigma Aldrich)-treated H 2 O at -80°C until it was used for further experiments. Complementary DNA (cDNA) was synthesized from 1 ㎍ of total RNA using reverse transcriptase (Invitrogen, Carlsbad, CA). Reverse transcription was then performed at 42°C for 50 min and 72°C for 15 min. Subsequently, 2 ㎕ of cDNA product and 10 pmol each of gene-specific primer were used for PCR with a 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA). A Power SYBR Green PCR Master Mix (Applied Biosystems) was used as the fluorescence source. Primers were designed with Primer 3 software (http://frodo.wi.mit.edu) using sequence information obtained from the NCBI. The primer sequences are provided in S2 Table. Immunocytochemistry C2C12 cells grown in a covered glass-bottom dish (SPL, Yeoju, Republic of Korea) were treated with differentiation medium and stained for CTNNA2 protein following 0 and 6 days of incubation. Briefly, the cells were rinsed with PBS and fixed with 4% formaldehyde. Following permeabilization with 0.2% Triton X-100 (Sigma-Aldrich, MO, USA) for 5 min, the cells were incubated overnight with anti-CTNNA2 antibody (1:50; Abcam, MA, USA) at 4°C in a humid environment. Secondary antibody (1:100; Alexa Fluor 488 goat anti-rabbit SFX kit; Invitrogen) was then applied for 1 h at room temperature. Next, the samples were rinsed with PBS after which the nuclei were counterstained with 4' 6'-diamino-2-phenylindole (DAPI; Sigma-Aldrich). Finally, pictures were taken using a fluorescent microscope equipped with a digital camera (Nikon, NY, USA).

Fusion index determination
Cell nuclei were stained with Giemsa G250 (Sigma-Aldrich) and pictures were captured randomly in three different fields using digital camera equipped with microscope. The number of nuclei in myotubes and total number of nuclei in the cells were also counted in each field. The fusion index was calculated as the percentage of nuclei incorporated into myotubes versus the total number of nuclei.

Western blot analysis
After washing with ice-cold PBS, cells were lysed with RIPA buffer (Thermo) containing protease inhibitor cocktail (Thermo). Total protein was isolated by centrifugation of the lysate at 12,000 rpm for 10 min at 4°C after which the protein concentration was determined by the Bradford method. The total protein (40 ㎍) was separated in 8% or 10% SDS-polyacrylamide gels by electrophoresis and transferred to PVDF membranes (Millipore, MA, USA). The blots were subsequently blocked with 3% skim milk (BD, MD, USA) or BSA (Sigma-Aldrich) in TBST for 1 h and then incubated overnight at 4°C with antibodies against CTNNA2 (1:500, Santa Cruz Biotechnology), type I collagen α1 (COL1 α 1; 1:2500, Abcam), Myostatin (MSTN; 1:400, Santa Cruz Biotechnology), or β-actin (1:2000, Santa Cruz Biotechnology) diluted with 1% skim milk or BSA in TBS. After washing with TBST, the blots were incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (goat anti rabbit or mouse, Santa Cruz Biotechnology) for 1 h at room temperature. After washing with TBST, antibody binding was detected with Super Signal West Pico Chemiluminescent Substrate (Thermo).

Statistical analysis
Mean values for normalized expression were compared using Tukey's studentized range (HSD) to identify significant differences in gene expression. P-values < 0.05 were considered statistically significant. Real-time RT-PCR data were analyzed with a one-way ANOVA using PROC GLM in the SAS package ver. 9.0 (SAS Institute, Cary, NC, USA).

DEGs and network construction
A total of 230 up-regulated and 223 down-regulated genes identified by comparing MYOG wt and MYOG kd primary bovine MSCs were studied [23]. The GeneMANIA webserver was used to predict interactions between the DEGs and 50 additional related genes in the network using the GO term "biological process" and source organism Homo sapiens as additional parameters. Out of the 230 up-and 223 down-regulated genes, GeneMANIA was able to recognize 217 and 215 genes, respectively. Using these DEGs and the 50 related genes identified by GeneMANIA, three interaction networks were created. The first interaction network (Network 1) was created using 215 down-regulated genes and showed enrichment of processes related to the TGF-β signaling pathway (S3 Table). In this network, red squares represented down-regulated genes and nodes shown in cyan were GeneMANIA-predicted genes ( Fig 1A). The second network (Network 2) contained 217 up-regulated genes (nodes in green) and GeneMANIA-predicted genes (nodes in cyan; Fig 1B). Processes related to the cell cycle were the main processes in this network (S4 Table). The final network (Network 3) consisted of 432 genes (both up-and downregulated) as well as GeneMANIA-predicted genes ( Fig 1C). The three interaction networks predicted by GeneMANIA were exported to Cytoscape for filtering (to remove duplicate edges and self-loops), and topological properties such as node degree were assessed. After removing the duplicate edges and self-loops, Network 1 contained 264 nodes and 4545 edges whereas Network 2 had 264 nodes and 4309 edges. Similarly, the filtered Network 3 (the main network in this study) had 481 nodes and 11374 edges, and functions related to the cell cycle were overrepresented ( Table 1).

Identification of hub genes
In order to identify potential hub genes, the node degree of each node in all three networks was calculated with the Network-Analyzer plugin of Cytoscape. Genes with the highest node degree were considered hub genes in each of these networks. CTNNA2 was the hub gene in Network 1 and Network 3 with node degrees of 89 and 136, respectively. In Network 2, CDC20 was the top hub with a node degree of 83. In Table 2, all genes that had a node degree 100 in Network 3 are listed. Since CTNNA2 was identified as the top hub gene in two out of three interaction networks (Networks 1 and 3), in vitro studies were designed to explore the functional significance of this gene in skeletal muscle development using primary bovine MSCs. The top hub gene CTNNA2 and its interacting partners are presented in Fig 2A In Fig 2B, it can be observed that these interacting partners are involved in collagen catabolic processes and skeletal muscle development (up-regulated genes), or processes related to cell growth and morphogenesis (down-regulated genes). Similarly, the GeneMANIA-predicted first neighbors of CTNNA2 were involved in either DNA replication or DNA metabolic processes ( Table 3).

Community analysis
Since Network 3 was largest of the three networks and contained both up-and down-regulated genes, we selected this network to identify functional modules using greedy algorithm (GLay). A total of four functional clusters were detected with only the first three clusters having more than ten genes; these were further subjected to functional analysis to identify enriched GO terms. Fig 3A (Cluster 1-4) shows the functional modules detected by GLay. A DAVID functional analysis tool was employed to categorize the genes in each cluster and observe the overrepresented GO terms in all three modules (Table 4). Overall, 233 enriched GO terms were identified in all three clusters. One hundred and thirty-six statistically significant (pvalue 0.05) GO terms were overrepresented in cluster 1, which was the largest detected cluster. Among the 10 most significant enriched factors, processes related to homeostasis were more prominent in cluster 1 (Table 4A and S5 Table). Other overrepresented processes in this cluster include leukocyte migration, response to wounding, and cell migration and motion. A total of 93 GO terms that were enriched in cluster 2 included processes related to the cell cycle and DNA replication (Table 4B and S6 Table). Cluster 3 was the smallest of all the modules detected. It consisted of only four statistically significant phosphorous metabolism-related GO terms such as protein amino acid phosphorylation and the phosphorous metabolic process (Table 4C). No enriched category was observed in cluster 4 as this cluster was omitted from the functional analysis because it only contained two genes.
In addition to GLay, we also compared the cluster analysis results with two other wellknown clustering programs: MCODE and MCL. MCODE detected 12 clusters, but only nine that contained 10 nodes were selected for enrichment analysis (Fig 3B Cluster 1-8 along with remaining one). These nine modules were associated with 93 statistically significant GO terms. Clusters 1, 8, and 9 did not have any enriched categories. Cluster 2 was enriched in leukocyte migration and enzyme-linked receptor protein signaling pathway, whereas cluster 3 showed overrepresentation of biosynthetic processes such as those involved in nitrogen compound and cofactor biosynthesis. Cluster 4, the largest cluster detected by MCODE, included processes related to the cell cycle and cell division. Cluster 5 was mainly associated with regulation processes that govern nucleocytoplasmic and intracellular transport, response to stimuli, immune effector activity, and neuron differentiation. In cluster 6, the enriched terms included cell motility, cell motion, and phosphorous metabolic process whereas cluster 7 was associated  with responses to wounding and inflammation in addition to the phosphorous metabolic process (A) in S7 Table). On the other hand, a total of seven clusters were detected by MCL with only one large cluster having more than 10 nodes (Fig 3C). This cluster had 102 significantly enriched GO terms with the most enriched representing processes related to the cell cycle and DNA replication (B) in S7 Table). For comparison, the numbers of statistically significant GO terms reported by three different algorithms are summarized in

CTNNA2 expression during differentiation and effects of MYOG knockdown in bovine MSCs
Based on our initial network analysis in which CTNNA2 was identified as the top hub gene with the maximum interacting partners (genes), an in vitro study was carried out to explore the functional significance of this gene in skeletal muscle development using primary bovine MSCs. After isolating the MSCs from bovine hind leg skeletal muscle, MSCs stained with Pax7 were used to determine cell purity. Around 85% of the total cell population was found to express Pax7. To elucidate the function of CTNNA2 during bovine MSC differentiation, changes in mRNA expression over time were evaluated. CTNNA2 expression was increased on Day 12 (Differentiation) compared to Day 10 (Proliferation) (Fig 5A). Correlation of CTNNA2 and MYOG expression was validated through RNA-Seq analysis .
To establish the relationship between MYOG and CTNNA2, MYOG shRNA was transfected in MSCs. Expression of MYOG and CTNNA2 was decreased in MYOG kd cells compared to MYOG wt cells (Fig 5B). Results of these knock-down studies indicated that CTNNA2 expression is involved in myogenesis and is regulated by MYOG.

CTNNA2 expression during differentiation and effects of MYOG, CTNNA2 knock-down in C2C12 cells
To confirm the data obtained from the bovine MSC studies, the expression of CTNNA2 in C2C12 cells was assessed. mRNA expression was measured in cells cultured with 2% FBS for 0, 2, 4, and 6 days. It was found that mRNA levels had gradually increased by day 6 compared to day 0 (Fig 5C).
Protein localization studies using immunocytochemistry revealed that CTNNA2 expression was highly restricted to the cytoplasm. Expression had increased on day 6 compared to day 0 ( Fig 5D). To elucidate the correlation between CTNNA2 and MYOG expression in C2C12 cells, MYOG knock-down was performed with shRNA. Results of the experiment revealed that expression of both MYOG and CTNNA2 was down-regulated in MYOG kd cells compared to MYOG wt cells. Data from the knock-down studies demonstrated that expression of CTNNA2 is regulated by MYOG in C2C12 cells. Myotube formation and ECM gene expression in C2C12 cells with knocked down CTNNA2 expression C2C12 cells transfected with CTNNA2 siRNA (CTNNA2 kd ) were cultured with 2% FBS for 6 day. Cell morphology, tube formation, and fusion index values were similar for both cell (normal and knock-down) lines (Fig 6A and 6B). CTNNA2 and MYOG mRNA expression was also analyzed by real-time RT-PCR. CTNNA2 expression was decreased substantially whereas MYOG expression was not significantly altered by knocking down CTNNA2 (CTNNA2 kd ; Fig  6C). Simultaneously, ECM is considered essential for myotube formation and the expression of muscle-specific gene products [41]. The expression of ECM genes (COL1α1, COL1α2, and MSTN) was increased during myogenesis (unpublished data). To establish the relationship between CTNNA2 and the ECM during myogenesis, mRNA and protein expression of COL1α1, COL1α2, and MSTN was measured in CTNNA2 kd cells. Both mRNA and protein expression of all three genes was significantly increased in CTNNA2 kd cells compared to CTNNA2 wt cells (Fig 6D and 6E). These results indicated that CTNNA2 helps maintain balance between the expression of MRF and ECM gene expression during muscle differentiation.

Discussion
By definition, MSCs under the basal lamina are stem cells in muscle with the ability to selfrenew and differentiate into myoblasts. Once satellite cells differentiate into myoblasts, they fuse into myofibers and contribute to myofiber growth. To obtain insight into the transcriptome of primary bovine MSCs, we recently performed an analysis with MYOG kd samples using the RNA-Seq approach [23]. Generally, a typical high throughput analysis targets the identification of DEGs while a pathway analysis followed by experimental validation of a few selected genes (based on fold change in expression) is performed to measure functional enrichment.
Knowing that DEGs may be equally or more significant for elucidating the process of skeletal muscle development, we created interaction networks of DEGs identified from MYOG kd samples in order to obtain insight into the relationships with respect to one another. In the present study, we constructed three interaction networks (using up-regulated or down-regulated genes and a combination of all genes) from the list of DEGs for MYOG kd primary bovine MSCs. As part of our previous study, we found that genes involved in cell division, DNA replication, and phosphate metabolic processes were differentially expressed [23]. The node degree of each gene in all three networks was calculated. It was observed that CTNNA2 had the highest node degree in Networks 1 and 3, whereas CDC20 was the gene with highest node degree in Network 2. Since CTNNA2 was found to have the highest node degree in two out of three networks, this gene was considered a hub gene in the present study. Therefore, its functional significance in skeletal muscle development was further explored. The CTNNA2 gene encodes catenin alpha-2 or alpha N-catenin in humans and mice, but is recorded as an uncharacterized protein in the UniProt for bovine (UniProt IDs: P26232; Q61301). CTNNA2 protein acts as a linker between cadherin adhesion receptors and the cytoskeleton to control cell-cell adhesion and differentiation in the nervous system. This protein also regulates morphological plasticity of synapses as well as cerebellar and hippocampal lamination during development [42][43][44]. CTNNA2 is also part of a cell surface complex that includes Ig/FNIII proteins CDO and BOC, N-and M-cadherin (promyogenic cell adhesion molecules), and βand α-catenin (cadherin-associated proteins) that may direct several aspects of myogenesis [45][46][47][48].
Recently, we identified CTNNA2 as a DEG that is down-regulated by at least 4-fold in MYOG kd primary bovine MSCs using RNA-Seq analysis [23]. In Fig 2, it can be observed that CTNNA2 CTNNA2 (a cell adhesion protein) has diverse interacting partners including other cell adhesion molecules such as PLXNC1, CDK5R1, CDH18, COL15A1, and JAM2 (down-regulated) as well as IBSP, CLDN6, PCDHB11, and SELPLG (up-regulated). These cell adhesion molecules along with other genes as seen in Fig 2are involved in various biological processes such as collagen catabolism, skeletal muscle development, cell growth, and morphogenesis ( Table 3). Moreover, we also observe cross-talk between these processes, thereby suggesting that one gene may be involved in various myogenic pathways (Fig 2). CTNNA2 expression was also increased during myogenesis. CTNNA2 and MYOG expression was analyzed in CTNNA2 and MYOG knock-down cells. CTNNA2 expression was decreased in MYOG kd . However, expression of MYOG was not significantly altered in CTNNA2 knock-down cells. Results of this experiment indicated that MYOG may be upstream of CTNNA2.
To further explore the role of CTNNA2 in myogenesis, we evaluated genes (MYOG, COL1α1, COL1α2, and MSTN) that are known to play substantial roles in skeletal muscle development [49][50][51]. A substantial increase in the expression of these genes (Fig 6) in CTNNA2 kd cells suggested that CTNNA2 holds a significant role in collagen catabolism, cell adhesion, and myogenesis. Collagen, proteoglycans, and adhesive glycoproteins are key components of the ECM, and are involved in ECM-receptor interactions as well as focal adhesion [50]. Explicit interactions among cells and ECM facilitated by transmembrane molecules or other cell surface-associated factors may directly or indirectly regulate cellular activities such as adhesion and migration [50].
Identifying the structure and function of biological networks is crucial for the exploration of biological phenomena. The interest in exploiting the use of network based studies to address key biological issues is ever increasing [52][53][54]. In this work, four functional modules or clusters were identified in Network 3 using the GLay plugin for Cytoscape. Additionally, the functional enrichment of each module in the network was explored using the DAVID functional annotation tool. Our analysis demonstrated that among the 10 most significant enrichments, cluster 1 homeostasis was highly overrepresented. Some of the genes grouped under homeostasis by the DAVID functional analysis encode channel proteins such as ryanodine receptor 1 (RYR1) and sodium channel protein type 1 subunit alpha (SCN1A). RYR1 is an indispensable factor for maintaining calcium homeostasis in mammalian skeletal muscle. Inactivation of this protein is lethal at birth in mice and mutations of the RYR1 gene in humans are associated various muscle disorders [51]. Similarly, voltage-gated sodium channels are involved in the early increase and consequent transmission of action potential in skeletal muscle [55]. Cluster 2 was enriched in processes related to the cell cycle and DNA replication. It included genes that encode various cell division homologue proteins such as cell division cycle 45 (CDC45), CDC20, and CDC6. The third cluster in the network was enriched in phosphorous metabolic processes. The community structure of our network concurs with results of our recent work on MYOG kd primary bovine MSCs [23]. Myogenesis and associated genes evaluated by CTNNA2 knock-down. C2C12 cells were transfected with CTNNA2-specific siRNA and incubated with 2% FBS for 6 d. A) Morphology of the CTNNA2 wt and CTNNA2 kd cells was similar. B) The fusion index was determined and the values were similar for CTNNA2 wt and CTNNA2 kd cells. C) mRNA expression was analyzed by real-time RT-PCR. CTNNA2 expression was decreased in CTNNA2 knock-down cells. However, MYOG expression was similar in the CTNNA2 wt and CTNNA2 kd cells. D) Expression of ECM genes was analyzed by real-time RT-PCR. COL1α1, COL1α2, and MSTN gene expression was increased in CTNNA2 kd cells. In particular, MSTN expression was the highest among the four genes. E) Protein expression of CTNNA2, COL1α1, and MSTN in CTNNA2 kd was analyzed by Western blotting. CTNNA2 and ECM protein expression levels were similar to the mRNA expression patterns. The p-value indicates the statistical significance (p<0.05, mean ± S.D., n = 3). Molecular target discovery and targeted therapeutics have become indispensable ever since the advancement of bioinformatics [56]. Although the present study highlights the importance of in silico approaches for assigning novel role to genes such as CTNNA2 in myogenesis, at the same time our findings open up several avenues that emphasize the importance of predicting how genes interact with each other to regulate homeostasis. It is also apparent that in silicopredicted results are highly correlated with experimental data. Altogether, understanding the structure and function of DEGs is essential not only for elucidating their roles during proliferation and differentiation, but can also be useful for predicting the binding specificities of genes to counteract several muscular dystrophies by designing potent inhibitors against target genes.
Supporting Information S1