Osteosarcoma is the most common human primary malignant bone tumor. It usually occurs in children and adolescents with a high incidence, rapid progression, and great metastatic potential. Over the past decades, despite the rate of local tumor control has been improved with advances in the treatment strategies, the long-term survival rate for patients remains unchanged [1, 2]. So as to improve the overall survival rate of OS patients, a better understanding of the underlying molecular mechanism of OS is needed. Previous studied have demonstrated several target biomarkers of OS based on both gene expression profiling and micro-RNA profiling of the primary tumor [21]. Wang and Baumhoer et al. had analyzed the OS sample by RT-PCR or immunohistochemistry and observed that insulin-like growth factor-1 receptor (IGF-1R) and cysteine-rich intestinal protein 1(CRIP1) were independent prognostic markers for OS and the high expression of these two markers was correlated with progression and metastasis of OS [22, 23]. A study performed by Duan et al. evaluated the role of miRNA in the progression of OS by RT-PCR and reported that three miRNAs, miR-127-3p, miR-199a-3p and miR-376c, were significantly downregulated in OS cell lines compared to osteoblasts, whereas miR-191 and miR-151-3p were increased in OS cell lines [24]. However, there’s still a lack of effective prognostic biomarkers for patients with OS. WGCNA, an effective method to detect the complicated relationships between different genes and phenotypes, provides insights into signaling networks which might be associated with clinical traits of interest. The analysis of WGCNA allows for identification of biological-relevant modules and hub genes, which may eventually serve as biomarkers for detection or treatment [9]. By constructing WGCNA to detect gene expression of OS, Fan et al. identified matrix metallopeptidase 3 (MMP3) and vascular endothelial growth factor B (VEGFB) as critical genes for OS development [25]. Likewise, Wang et al. reported the involvement of Matrix extracellular phosphoglycoprotein (MEPE) and Hemoglobin A2 (HBA2) in OS [26]. However, the precise mechanism of OS remains to be elucidated and more candidate biomarkers or therapeutic target needs to be discovered. Here, we utilized WGCNA to investigate co-expression modules related to OS and predict candidate gene sets that underlies a given biological process.
A total of 17 co-expression modules were first constructed by the 9854 genes from 47 human OS samples by WGCNA method, followed by module-trait correlations. As a result, we identified the green module that related to clinical traits (patient’s OS location). Then the functional enrichment analysis was conducted in green module. Consistent with the previous study by Xiong and colleague, our result showed that the green module was mainly enriched in pathway associated with skeletal system development, bone development and extracellular structure organization [6]. Since the development of OS was accompanied by the abnormal expression of oncogenes and anti-oncogene, the enriched pathway that the green module involved in may regulate genes participated in the development of the skeletal system. Heng et al. assessed the DEGs and microRNAs in OS samples and found the downregulated DEGs COL4A1 and COL5A1 were enriched in extracellular structure organization as well as extracellular collagen biosynthesis pathway [27]. Increased extracellular collagen degradation might promote the development of OS and metastasis to lungs, which indicated that genes involved in collagen synthesis might have a negative effect on extracellular matrix, which may be responsible for OS tumor invasion in an undirect manner.
In addition, we constructed PPI network and identified hub genes in OS-related regulatory network. Here we used a degree to evaluate the importance of genes in the regulatory network. The genes with the highest degrees were identified as hub genes and the top 10 hub genes were COL6A1, MMP13, DCN, FAM20C, FBN1, GNG2, PCOLCE, PCOLCE2, RUNX2 and VCAN. Furthermore, we validated 5 genes, MMP13, DCN, GNG2, PCOLCE, and RUNX2, had significantly high expression profile in the validation data set.
Several of the genes have been reported associated with OS. Matrix metalloproteinase-13(MMP13) has been identified as a critical modulator involved in tumor osteolysis. Ma et al. identified MMP-13 as the target gene of proteolytic enzyme that regulated tumor-induced osteolysis [28]. Hirahata and colleagues showed that high MMP13 expression was of great importance for the survival of OS cells [29]. Decorin (DCN), Procollagen C-proteinase enhancer protein (PCOLCE) and Runx2 were found to play a vital role in lung metastasis of OS [30, 31]. Runx2 is a kind of protein that play major roles in bone formation. A study by Won evaluated the expression of Runx2 by immunohistochemical staining of OS tissues and concluded that high expression of Runx2 was significantly correlated to OS metastasis. Overexpression of Runx2 was also responsible for poor response to chemotherapy relative to good responders [32]. No reports have revealed that G Protein Subunit Gamma 2 (GNG2) was related to OS and further study was needed.