Abstract
Background/Aims: As a malignant and melanocytic tumor, cutaneous melanoma is the devastating skin tumor with high rates of recurrence and metastasis. Bone is the common metastatic location, and bone metastasis may result in pathologic fracture, neurologic damage, and severe bone pain. Although metastatic melanoma was reported to get benefits from immunotherapy, molecular mechanisms and immune microenviroment underlying the melanoma bone metastasis and prognostic factors are still unknown. Methods: Gene expression profiling of 112 samples, including 104 primary melanomas and 8 bone metastatic melanomas from The Cancer Genome Atlas database, was assayed to construct a ceRNA network associated with bone metastases. Besides, we detected the fraction of 22 immune cell types in melanoma via the algorithm of "cell type identification by estimating relative subsets of RNA transcripts (CIBERSORT)." Based on the significant ceRNAs or immune cells, we constructed nomograms to predict the prognosis of patients with melanoma. Ultimately, correlation analysis was implemented to discover the relationship between the significant ceRNA and immune cells to reveal the potential signaling pathways. Results: We constructed a ceRNA network based on the interaction among 8 pairs of long noncoding RNA-microRNA and 15 pairs of microRNA-mRNA. CIBERSORT and ceRNA integration analysis discovered that AL118506.1 has both significant prognostic value (P = 0.002) and high correlation with T follicular helper cells (P = 0.033). Meanwhile, T cells CD8 and macrophages M2 were negatively correlated (P < 0.001). Moreover, we constructed two satisfactory nomograms (area under curve of 3-year survival: 0.899; 5-year survival: 0.885; and concordance index: 0.780) with significant ceRNAs or immune cells, to predict the prognosis of patients. Conclusions: In this study, we suggest that bone metastasis in melanoma might be related to AL118506.1 and its role in regulating thrombospondin 2 and T follicular helper cells. Two nomograms were constructed to predict the prognosis of patients with melanoma and demonstrated their value in improving the personalized management.