Prognostic value of SOX9 in cervical cancer: Bioinformatics and experimental approaches

Abstract

Among gynecological cancers, cervical cancer is a common malignancy and remains the leading cause of cancer-related death for women. However, the exact molecular pathogenesis of cervical cancer is not known. Hence, understanding the molecular mechanisms underlying cervical cancer pathogenesis will aid in the development of effective treatment modalities. In this research, we attempted to discern candidate biomarkers for cervical cancer by using multiple bioinformatics approaches. First, we performed differential expression analysis based on cervical squamous cell carcinoma and endocervical adenocarcinoma data from The Cancer Genome Atlas database, then used differentially expressed genes for weighted gene co-expression network construction to find the most relevant gene module for cervical cancer. Next, the Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses were performed on the module genes, followed by using protein-protein interaction network analysis and Cytoscape to find the key gene. Finally, we validated the key gene by using multiple online sites and experimental methods. Through weighted gene co-expression network analysis, we found the turquoise module was the highest correlated module with cervical cancer diagnosis. The biological process of the module genes focused on cell proliferation, cell adhesion, and protein binding processes, while the Kyoto Encyclopedia of Genes and Genomes pathway of the module significantly enriched pathways related to cancer and cell circle. Among the module genes, SOX9 was identified as the hub gene, and its expression was associated with cervical cancer prognosis. We found the expression of SOX9 correlates with cancer-associated fibroblast immune infiltration in immune cells by Timer2.0. Furthermore, cancer-associated fibroblast infiltration is linked to cervical cancer patients' prognosis. Compared to those in normal adjacent, immunohistochemical and real-time quantitative polymerase chain reaction (qPCR) showed that the protein and mRNA expression of SOX9 in cervical cancer were higher. Therefore, the SOX9 gene acts as an oncogene in cervical cancer, interactive with immune infiltration of cancer-associated fibroblasts, thereby affecting the prognosis of patients with cervical cancer.

Keywords: bioinformatics; cancer-related fibroblasts; cervical cancer; differentially expressed genes (DEGs); hub gene; weighted gene co-expression network analysis (WGCNA).

FIGURE 1

Flowchart for this study.

FIGURE 2

Determination of soft thresholds and testing of scale-free networks. (A) Correlation between log (K) and log [P (k)] corresponding to different soft thresholds. The higher the coefficient, the more the network conforms to the distribution of the scale-free network. (B) Mean value of gene neighbor coefficients in the gene network corresponding to different soft thresholds, which reflects the average connectivity level of the network. (C) Distribution of the connectivity of each node in the network. (D) Scatter plot of log (K) vs. log [P (k)] and the linear regression results show that the correlation coefficient is 0.92.

FIGURE 3

Classification of gene clustering trees by dynamic tree cut. A total of five gene modules were obtained, with different colors indicating different gene modules. The gray color indicates the genes that do not belong to any known module.

FIGURE4

(A) Relationship between the module traits and the module significance. (B) The turquoise module has the highest correlation with the diagnosis level (p = 0.82, p = 0.65 × 3e-38).

FIGURE 5

GO and KEGG pathway analysis. (A) Biological process of GO, (B) cellular components of GO, (C) molecular function of GO, and (D) KEGG pathway of the turquoise module significantly enriches pathways related to cancer and cell circle.

FIGURE 6

Association between SOX9 gene expression and overall survival and disease-free survival of cervical cancer patients in TCGA. (A) Low SOX9 gene expression patient had a high overall survival rate (p = 0.038). (B) Low SOX9 gene expression patient had a high disease-free survival rate (p = 0.049). Low expression of SOX9 may lead to higher survival rates and disease-free survival.

FIGURE 7

Box plot of SOX9 expression in TCGA tumors. The mRNA expression of SOX9 is higher than that of normal tissues in most TCGA tumors.

FIGURE 8

Correlation between SOX9 gene expression and CAF immune infiltration. Expression of SOX9 positively correlates with tumor-associated fibroblasts.

FIGURE 9

Association between CAF and overall survival in Timer2. (A) Low CAF had high overall survival rate than high CAF. (B) Low CAF + low SOX9 expression group had high overall survival rate than the high CAF + high SOX9 expression group.

FIGURE 10

SOX9 immunohistochemical images of normal cervical and cervical cancer tissues on the HPA database and patient samples. (A) normal cervix (HPA), (B) cervical cancer (HPA), (C) normal cervical glands, (D) cervical adenocarcinoma, (E) normal cervical epithelium, and (F) cervical squamous carcinoma.

FIGURE 11

Quantitative RT-PCR for cancer and adjacent normal cervix tissue. Quantitative RT-PCR assay showed significantly increased SOX9 mRNA level in cervical cancer tissues compared with adjacent normal cervix tissues (p = 0.031).