PDAC remains a treatment-refractory malignancy with poor prognosis. The discovery of some promising molecular targets for the treatment of PDAC provided a new direction for the diagnosis and treatment of PDAC. For example, Jin et al. reported that, PES1 might be a promoting factor of tumor growth and a PDAC prognosis associated protein [23]. Goehrig et al. found that, βig-h3 stromal-derived protein is a main factor driving the immune paracrine interaction mechanism of PDAC, which might be novel target in PDAC treatment[24]. However, compared with other malignant tumors, effective diagnosis and treatment targets of pancreatic cancer are still quite lacking. It is urgent to develop screening procedures for early detection and more efficacious treatment strategies for PDAC[25]. Bioinformatics provides an opportunity to quest for the genetic alterations in pancreatic cancer, and proved a useful way to identify new biomarkers in pancreatic cancer. To verify the results from bioinformatics further, we tested the expression of S100As in pancreatic cancer at the cellular and tissue levels by real-time PCR and immunohistochemistry. Unfortunately, pancreatic cancer tissues from PDAC patients undergoing surgery are scarce due to the few cases of pancreatic cancer surgery. Because the sample size is too small, the heterogeneity between patients obscures some of the original expression levels. Here, we investigated the mRNA expression and prognostic values of S100A family in PDAC for the first time. We expect that our findings could help us to understand the mechanism of PDCA, improve the clinical treatment design, and improve the reliability of prognostic assessment.
The S100 protein family belong to a family of calcium binding proteins (CaBPs) and contain 25 known members, which have a high degree of sequence and structural similarity. The S100A subfamily is among the most distinctive of EF-hand CaBPs and only found in vertebrates. Studies have shown that multiple members of the S100As are abnormally expressed in tumors and are widely involved in processes related to tumorigenesis and progression [26, 27]. However, the function of most S100A has not been fully characterized, especially in pancreatic cancer, the research on S100A is still very limited.
Among the S100As, S100A4 is the most intensively studied. S100A4 is a member of the S100As, which contains two EF-hand calcium-binding motifs. Among its related pathways are Ca2+, cAMP and Lipid Signaling. Fei et al. reported that, the expression of S100A4 was positively correlated with the malignancy degree and the level in lymph node metastasis of human colorectal cancer[28]. Recently, a study about hepatocellular carcinogenesis indicated that, high expression of S100A4 in the fibrotic region of was liver tumor tissue observed, and the expression of S100A4 was associated with advanced disease severity[29]. It was also reported that, S100A4 was one of the key factors in EMT mediated by Shh-Gli1 signaling pathway in PDAC[30]. In present study, Oncomine, GEPIA 2 and UALCAN datasets revealed that the expression of S100A4 was higher in PDAC than in normal pancreatic tissues. We also noticed that, S100A4 was up-regulated in all four pancreatic cancer cells relative to hTERT-HPNE cells. In addition, BBCancer (http://bbcancer.renlab.org/), a database that provides potential clinical significance of RNAs as biomarkers for early cancer detection, was used to analyze the value of S100A in the early diagnosis of pancreatic cancer. We can see from Fig. 13, S100A4/6/8/9/10/11/12/16 were up regulated in extracellular vesicles of pancreatic cancer, which might have diagnostic value.
As an important paralog of S100A4, S100A2 is reported to be abnormally highly expressed in various tumors, such as colorectal cancer[31] and lung cancer[32]. In our study, we indicated that the expression of S100A2 in pancreatic cancer tissues was higher than that in normal pancreatic tissues. It has also been suggested as a negative prognostic biomarker in pancreatic cancer[33]. However, the prognostic role of S100A2 in pancreatic cancer remains to be studied. GO annotations of S100A2 include calcium ion binding and identical protein binding. Among its related pathways are p53 pathway.
S100A6 is also an up-regulated S100As in pancreatic cancer confirmed by bioinformatics and in vitro experiments. Results of BBCancer also showed that, S100A6 was up-regulated in extracellular vesicles of PDAC. Diseases associated with S100A6 include Pilomatrixoma and Retinitis Pigmentosa. The latest research also showed that S100A6 is involved in the process of neurodegenerative diseases[34]. GO annotations of S100A6 were calcium ion binding and calcium-dependent protein binding. Among its related pathways are prostaglandin synthesis and regulation and DNA damage. Further research is needed to explain the role of S100A6 in the development of pancreatic cancer.
Diseases associated with S100A10 include Trachea Leiomyoma and Barrett's Adenocarcinoma. There is evidence that S100A10 has potential value as a biomarker that represents the high-grade cell state of breast cancer[35]. Among its related pathways are Prostaglandin Synthesis and Regulation and Response to elevated platelet cytosolic Ca2+. GO annotations related to this gene include calcium ion binding and lipid binding. An important paralog of this gene are S100A1and S100A14. In our study, both bioinformatics analysis and experimental verification indicated that S100A10 was up-regulated in pancreatic cancer (vs normal controls/paraneoplastic tissue). Low expression of S100A10 was associated significantly with better survival in pancreatic cancer by univariate analysis (P < 0.05). The above results reveal the potential value of S100A10 as a diagnosis and treatment of pancreatic cancer.
Both S100A14 and S100A16 were up-regulated in pancreatic cancer, which was also verified in our experiments. S100A14 has a potential to be clinically useful as prognostic biomarker in several cancer types. For example, Meta-analysis form Hu et al. suggested that S100A14 overexpression might be a predictive biomarker for poor prognosis in patients with breast cancer and ovarian cancer[36]. However, the role of S100A14 in pancreatic cancer is rarely reported. Survival analysis showed that, low expression of S100A14 was associated with significantly better survival. S100A16 was discovered late and its research is not deep enough. In this study, we found that S100A16 is upregulated in PDAC, and its low expression was closely related to a better survival.
Given that the signaling pathways involved in calcium ions have been shown to change the sensitivity of chemotherapy drugs [37]. We therefore also analyzed the drug relations of S100As by Drugbank. As listed in Table 2, Drugs targeted at S100As include Olopatadine, Trifluoperazine and calcium citrate. These drugs are mainly used as inactivators, regulators, corrosion inhibitors and ligands to play the role of drugs. Therefore, S100As have certain potential value in the development of pancreatic cancer chemotherapy resistance and new drug development.
DNA methylation generally occurs at the CpG site, and the conversion of cytosine to 5-methylcytosine is catalyzed by DNA methyltransferase in human genes. Approximately 80% of CpG sites are methylated, and the level of DNA methylation is inversely proportional to transcriptional activity and affects genome stability. Compared with the methylation of the gene-coding region, the methylation of the promoter region has a greater effect on gene expression activity. In our study, we found that the methylation level of the promoter region was negatively correlated with its expression, which may be a reason for the abnormal expression of S100As in pancreatic cancer.
Naturally, there are also many limitations in our research. First, although high mRNA expressions of S100A2/4/6/10/14/16 were independent prognostic factors for shorter OS of pancreatic cancer patients, the online databases used in this study lacked proteome-level data. Further studies consist of larger sample sizes (more clinical samples and different pancreatic cancer cell lines) are required to validate the findings above and to explore the clinical application of S100A2/4/6/10/14/16 in the treatment of PDAC. In addition, we did not explore the potential molecular mechanism of S100As in the occurrence and development of pancreatic cancer. Future research worth to investigate the detailed mechanism between S100As and PDAC.