Development and validation of a gene-based prognostic model for bone metastasis in breast cancer

Background: Bone is the most common site of metastasis in breast cancer, resulting in a worse quality of life and prognosis for breast cancer patients. Due to the limited efficacy of conventional therapies in breast cancer patients with bone metastases, there is a need for innovative molecular targets in clinical practice. Currently, the progress in developing effective medicines for bone metastases in breast cancer patients is being impeded by the limitations of our understanding of the molecular pathways involved. Methods: We have generated a prognostic bone metastasis signature based on bone metastasis-associated genes identified from Gene Expression Omnibus (GEO). This study utilized bone metastasis-associated genes in order to predict the overall prognosis of individuals with breast cancer. A comprehensive transcriptomic analysis was conducted to assess the clinical importance and immunological characteristics of the signature. Results: Using the gene expression data from the GSE14776 dataset, we identified DEGs significantly associated with bone metastasis. These DEGs can be used to discriminate between disseminated tumor cells (DTCs) and metastatic tumor cells (MTCs). Through training set transcriptome data analysis, a prognostic-associated bone metastasis signature (PABMS) was identified consisting of four genes including SBK1, ZER, CXCL1, and L1CAM). This signature is strongly correlated with the breast cancer patient's overall survival durations. The discriminating ability of the four-gene bone metastasis-associated signature was further confirmed by analyzing protein values in a separate and unrelated group of individuals. The findings demonstrated a distinct correlation between a high-risk score and an extremely unfavorable patient prognosis. It was further determined that the prognostic bone metastasis signature serve as a prognostic predictor that was independent of other variables, using multivariate Cox regression analysis as well as stratification analysis, Additionally, we observed a strong correlation between the risk score and the inflammatory response, the antigen-presenting process, as well as the expression levels of other inhibitory molecules or checkpoints including PD-1, FGL-1, B7-H4, and galectin-9. These results suggest that patients at high-risk experience immune response inhibition. Conclusion: This study aims to create a practical and achievable gene signature by analyzing bone metastases-mediated differentially expressed genes. This gene signature has the potential to be a highly effective tool for predicting patient outcomes and improving predictive diagnostics in cancer treatment. Furthermore, it could aid in therapeutic decision-making by helping to guide personalized treatment strategies.