Melanoma is associated with many factors, is more common in light-skinned races, and has a family history of occurrence. Although melanoma treatment has improved from before, the prognosis for melanoma patients is low due to the lack of precise molecular markers. Therefore, there is an urgent need to identify better and more accurate biomarkers to utilize in the prognosis, diagnosis, and treatment of melanoma. In our study, we used integrated bioinformatics to analyze a total of 435 important genes with co-expression trends identified in the UCSC Xena,GTEx database and the GSE3189 database. These genes were subjected to GO enrichment analysis based on the R package (clusterProfiler), mainly enriched in neutrophil activation involved in immune response, stem cell divisio, and melanosome. As early as 2011, it was noted that there might be a close relationship between neutrophil activation and cancer[20]. Moreover, the relationship between the stem cell divisio and melanosome and cancer has also been reported in the literature[21, 22]. Similarly, these genes were subjected to KEGG pathway enrichment analysis based on the R package (clusterProfiler), and the enrichment results showed that these genes were associated with various cancer pathways, including melanoma, endometrial cancer, prostate cancer, etc.; they were also enriched in transcriptional dysregulation pathways in cancer. In humans, dysregulation of genes such as cofactors and chromatin can lead to many diseases[23]. These genes are even enriched in the melanoma pathway, suggesting that these genes are strongly associated with melanoma. Besides, we screened for the top 10 hub genes associated with melanoma based on how the MCC score was calculated for the CytoHubba plugin in Cytoscape. We also found by analyzing the survival of melanoma patients corresponding to high and low expression of these genes that five of the top 10 hub genes were strongly associated with survival, and all showed that high expression of the genes was associated with a low prognosis in melanoma patients. Finally, we performed immunohistochemical analysis using the HPA database and showed that all four genes showed increased expression in melanoma tumor tissues, whereas their expression was not evident in normal tissues.
FOXM1, also known as Forkhead Box M1, is a gene that encodes a protein that is a transcriptional activator involved in cell proliferation. FOXM1 acts downstream of the PI3K-AKT pathway, the Ras-ERK pathway, the JNK/p38MAPK signaling cascade, and is essential for cell proliferation, differentiation, senescence, DNA damage and repair, and control of the cell cycle[24]. It has also been reported that FOXM1 was overexpressed in a variety of human cancers and that the oncogenic potential of this gene is based on its ability to reactivate target genes involved in different stages of cancer development[25]. It has been shown that the positive feedback of FOXM1 promotes the growth and invasion of gastric cancer and that FOXM1 promotes gastric cancer progression by interacting with PVT1[26]. FOXM1 has also been reported in non-serious epithelial ovarian carcinoma: FOXM1 was upregulated in all epithelial ovarian cancers[27]. In addition, it has been shown that the FOXM1-PSMB4 axis can play a catalytic role in the proliferation and development of cervical cancer[28]. More surprisingly, FOXM1 plays a vital role in many other cancers[29–32]. In our study, FOXM1 was upregulated in tumor tissues compared to normal tissues, suggesting a significant correlation with melanoma. Previous studies have shown that higher levels of FOXM1 in tumor tissues have been strongly associated with poor prognosis in melanoma patients, consistent with our study[33–35]. Kinesin Family Member 20A (KIF20A) is also a protein-encoding gene, and what is known about the diseases associated with this gene is mainly familial restriction Familial Isolated Restrictive Cardiomyopathy and Charcot- Marie-Tooth Disease, Type 4C. Research has also been conducted on the role of this gene in cancer. It has been reported that patients with bladder cancer with high expression of KIF20A have poorer tumor stages and that KIF20A promotes metastasis and proliferation of bladder cancer cells[36]. Also, it has been shown that skin tumor thickness in KIF20A-positive patients with primary melanoma is significantly greater than skin tumor thickness in patients negative for this gene and that KIF20A-positive patients are more likely to relapse earlier[37]. It is well known that while tumor recurrence has a very significant relationship with patient prognosis, this indirectly suggests that KIF20A is associated with survival in melanoma patients, which is consistent with the results of our study. The TPX2 Microtubule Nucleation Factor (TPX2) is a protein-coding gene. The main diseases known to be associated with the TPX2 gene include Capillary Leak Syndrome and Colorectal Cancer. It has been shown that activation of TPX2 expression increases the invasion and proliferation of cervical cancer, promoting cancer development[38]. A study of TPX2 in esophageal cancer showed that the 5-year survival rate of esophageal cancer patients with concomitant high TPX2 expression levels was significantly lower than that of esophageal cancer patients with low TPX2 expression levels[39]. Interestingly, in our study, patients with high TPX2 expression of melanoma had a relatively shorter overall survival than patients with low expression. Cell Division Cycle 20 (CDC20) is also a protein-coding gene. The main diseases known to be associated with this gene are Ceroid Lipofuscinosis, Neuronal, 2. Back in 2015, there were reports that CDC20 could be used as a novel cancer treatment modality[40]. In hepatocellular carcinoma, the upregulation of CDC20 expression predicted a decline in overall survival and disease-free survival[41]. In our study, melanoma patients with high expression of CDC20 had a lower survival time than patients with low expression. In summary, the expression of the four hub genes we studied were all strongly associated with cancer, and in our study, high levels of expression of these genes were accompanied by shorter survival times for melanoma patients.
Our study, like others, has limitations regarding the different tumor types. Although we identified potential prognostic genes between melanoma and normal tissue using three different sources of databases with two different bioinformatics analyses, it was less accurate for each of the different subtypes of melanoma patients. Furthermore, EXO1 failed to find corresponding evidence in the HPA database to identify protein expression of genes between melanoma and normal tissues. The Hub genes associated with survival that can influence melanoma patients’ prognosis should be further validated by a series of experiments on molecular mechanisms.
In conclusion, by combining the WGCNA analysis method with differentially expressed gene analysis, our study identified the genes FOXM1, KIF20A, TPX2, and CDC20, which are highly correlated with survival melanoma patients and have the potential to serve as a prognostic biomarker in melanoma.