The Clinical Motivation Metastasis is the most lethal stage of melanoma progression, and the majority of patients diagnosed at advanced stages have poor overall survival. Despite increasing incidence, the molecular mechanisms driving melanoma metastasis remain incompletely understood. Identifying the genes and pathways responsible for metastatic transition is critical for discovering both prognostic markers - to predict which patients will progress - and therapeutic targets to interrupt that progression.
The Bioinformatics Workflow This 2019 study from Shanghai Skin Disease Hospital applied an integrated bioinformatics pipeline to the GEO dataset GSE8401, which contains Affymetrix microarray gene expression data from 52 metastatic melanoma biopsy specimens and 31 primary melanoma specimens. The pipeline combined differential gene expression analysis, Gene Ontology (GO) and KEGG pathway enrichment, protein-protein interaction (PPI) network construction, and TCGA mutation analysis - using each layer to cross-validate and prioritize candidate genes.
Study Scope and Outputs Of 22,283 probes evaluated, 1,187 differentially expressed genes (DEGs) were identified - 505 upregulated and 682 downregulated in metastatic versus primary melanoma. From these, nine key candidate genes (PIK3R3, CENPM, AURKA, LAMA1, PCNA, ADCY1, BUB1, NDC80, and PRKCA) were identified based on high network connectivity and multi-pathway involvement. Mutation analysis from TCGA identified BAP1 as the most significant mutated gene, and its network partners (ASXL1, PSMD3, PSMD11, UBC) were also significantly associated with melanoma survival.
Data Sources and Preprocessing The GSE8401 dataset was preprocessed using the R affy package (v1.50.0), including background correction, normalization, and expression calculation. Probes not matching gene symbols were excluded, and the average of multiple probes mapping to the same gene was used. Differential expression was assessed with the limma package using a threshold of P less than 0.05 and absolute log2 fold-change greater than 1. TCGA mutation data (TCGA_SKCM project) provided 80 melanoma cases with mutational profiles for cross-validation.
GO, KEGG, and PPI Analysis Gene Ontology enrichment (molecular function, cellular component, biological process) and KEGG pathway analysis were performed using DAVID with P less than 0.05 as the significance threshold. For PPI network construction, DEGs were mapped to the STRING database (combined confidence score above 0.4) and visualized in Cytoscape. Hub genes were identified by filtering for degree greater than 30 in the PPI network (highly connected nodes). Cross-talk genes were those involved in at least three KEGG signaling pathways - a more stringent filter that identifies genes sitting at the intersection of multiple disease-relevant pathways.
Survival and Expression Analysis Overall survival curves for key candidate genes were generated using GEPIA, which uses TCGA data to split patients into high-expression and low-expression groups and computes hazard ratios with log-rank p-values. Expression differences between primary and metastatic melanoma were analyzed using the UALCAN portal. BAP1 mutation analysis and its effect on survival in the 40 metastatic melanoma cases within TCGA was performed via cBioPortal, with survival compared between 14 BAP1-mutated cases and 26 wild-type cases.
Top GO Terms GO enrichment of the 1,187 DEGs revealed that for the Cellular Component category, 'extracellular exosome' and 'extracellular space' processes were the most significantly enriched - pointing to a role for secreted vesicles and extracellular communication in the metastatic phenotype. For Biological Process, 'epidermis development', 'keratinocyte differentiation', and 'cell adhesion' were most enriched, reflecting the epithelial-to-mesenchymal transition-like processes that allow melanoma cells to detach from primary tumors and invade surrounding tissue. Molecular Function was dominated by 'structural molecule activity'.
KEGG Pathway Enrichment The top enriched KEGG pathways among DEGs were amoebiasis, ECM-receptor interaction, and focal adhesion. The focal adhesion pathway governs the mechanical connections between the extracellular matrix and the intracellular cytoskeleton - a critical regulator of cell migration, invasion, and anchorage-independent growth, all hallmarks of metastatic behavior. ECM-receptor interaction reflects the ability of metastatic cells to remodel their extracellular environment to facilitate invasion and colonization of distant sites.
PPI Network and Hub Gene Identification Of the 1,187 DEGs, 447 were successfully mapped onto the STRING PPI network, forming a complex interaction landscape with 740 genes excluded for lack of documented interactions. The degree greater than 30 filter identified nine highly connected hub genes: PIK3R3, CENPM, AURKA, LAMA1, PCNA, ADCY1, BUB1, NDC80, and PRKCA. These hub genes are not merely differentially expressed - they sit at network intersection points where perturbation is most likely to have widespread downstream effects on metastatic signaling.
Cell Cycle and Mitotic Checkpoint Genes BUB1, AURKA, NDC80, CENPM, and PCNA are all involved in cell cycle regulation and mitotic fidelity. BUB1 (a discriminator between melanoma and benign nevi in prior studies) is downstream of SIRT1 - a known melanoma oncogene - and its high expression promotes metastasis and poor prognosis. AURKA induces mitotic spindle formation, promotes cell multiplication and migration, is overexpressed in malignant melanoma, and is driven by FOXM1 and MAPK/ERK signaling. NDC80 is a kinetochore component that is highly expressed in pancreatic, hepatocellular, gastric, colorectal, and bladder cancers, where its overexpression promotes proliferation, migration, and invasion.
Signaling and Adhesion Genes PIK3R3, PRKCA, ADCY1, and LAMA1 are all involved in signaling and extracellular matrix interactions. PIK3R3 activates the PI3K/AKT/mTOR signaling axis, serving a crucial role in tumor survival, proliferation, and motility; its overexpression has been linked to metastasis in colorectal and lung cancers. PRKCA (protein kinase C alpha) controls melanoma cell growth - its activation promotes tumor invasiveness, and siRNA-mediated knockdown significantly suppresses invasion in melanoma models. ADCY1 participates in the melanogenesis pathway and several cancer-related KEGG pathways. LAMA1 is a laminin subunit involved in the extracellular matrix structure critical for cell attachment and migration.
Survival Significance Survival analysis using GEPIA (TCGA data) confirmed that aberrant expression of all nine hub genes - PIK3R3, CENPM, AURKA, LAMA1, PCNA, ADCY1, BUB1, NDC80, and PRKCA - was significantly associated with poor overall survival in melanoma patients. PRKCA, BUB1, and LAMA1 also showed significantly different expression levels between primary and metastatic melanoma in the UALCAN portal analysis, confirming their functional relevance to the metastatic transition specifically.
BAP1 as a High-Frequency Melanoma Mutation Mutation analysis of 80 TCGA melanoma cases identified GNAQ, GNA11, BAP1, and SF3B1 as having the highest mutation frequencies. Among these, only BAP1 mutation showed statistically significant association with overall survival. BAP1 (BRCA1-associated protein 1) is a tumor suppressor gene whose somatic mutations occur across multiple malignancies, and inactivating BAP1 mutations are specifically associated with high metastatic risk in uveal melanoma. In the present study, patients with low BAP1 expression had significantly poorer overall survival.
BAP1-Associated PPI Network A BAP1-specific PPI network was constructed to identify genes whose expression is correlated with BAP1 function. Of the BAP1-associated genes analyzed, survival analysis for ASXL1, PSMD3, PSMD11, and UBC were all statistically significant. High expression of ASXL1, PSMD3, and PSMD11 was associated with poor overall survival in melanoma. ASXL1 is an epigenetic regulator that forms a complex with BAP1, and ASXL1 mutations are frequent in myeloid malignancies. PSMD3 and PSMD11 are proteasome subunits - their overexpression may reflect enhanced protein degradation capacity that supports tumor cell survival and proliferation.
BAP1 Mutation in Metastatic Cases A focused survival analysis in 40 metastatic melanoma cases (14 with BAP1 mutations, 26 without) confirmed that BAP1 mutation was associated with poor prognosis specifically in the metastatic setting. BAP1 expression was also significantly lower in metastatic melanoma compared with primary melanoma in the UALCAN expression analysis. Together, these data position BAP1 as both a prognostic marker for metastatic melanoma and a potential therapeutic target - its loss may compromise epigenetic tumor suppression mechanisms and enable the gene expression reprogramming required for metastatic progression.
Single Dataset Limitation The study used a single GEO dataset (GSE8401) as the primary data source - no suitable dataset for meta-analysis or external validation was available at the time of publication. This limits the robustness of DEG identification since microarray results can be platform- and batch-specific. The original dataset collected specimens between 1992 and 2001, raising questions about the contemporary relevance of the gene signatures given intervening changes in treatment standards and patient selection.
Absence of Functional Validation The study is entirely bioinformatics-based; no wet lab experiments were conducted to functionally validate the identified candidate genes in melanoma cell lines or animal models. While several findings were supported by prior publications (e.g., BUB1 and AURKA in melanoma), genes such as CENPM and UBC had no prior cancer-specific literature, and their functional roles in melanoma remain uncharacterized. Knockdown or overexpression experiments in melanoma cell lines would be required to confirm causal roles in metastatic behavior.
Future Directions Future studies should validate the candidate gene expression in independent clinical samples from multiple institutions and validate their prognostic value in patients treated with current standard-of-care therapies including immune checkpoint inhibitors and BRAF/MEK targeted therapy. BAP1 and its associated genes warrant particular priority for functional investigation, as their role in the BAP1-ASXL1 epigenetic complex provides a potentially druggable mechanism. As more comprehensive TCGA melanoma data becomes available, expanded survival analysis including treatment-stratified cohorts will provide more clinically actionable insights.