Construction of Bone Metastasis-Specific Regulation Network Based on Prognostic Stemness-Related Signatures in Breast Invasive Carcinoma

Background/Aims: Bone is the most common metastatic site of Breast invasive carcinoma (BRCA). In this study, the bone metastasis-specific regulation network of BRCA was constructed based on prognostic stemness-related signatures (PSRSs) and their upstream transcription factors (TFs) and downstream pathways. Methods： Clinical information and RNA-seq data of 1080 primary BRCA samples including 1048 samples without bone metastasis and 32 samples with bone metastasis were downloaded from The Cancer Genome Atlas (TCGA). The edgeR method was performed to identify differential expressed genes (DEGs). Next, mRNA stemness index (mRNAsi) was calculated by one-class logistic regression (OCLR). To analyze DEGs by classification, similar genes were integrated into a module by weighted gene co-expression network analysis (WGCNA). Then, univariate and multivariate Cox proportional hazard regression were applied to find the PSRSs. Furthermore, PSRSs, 318 TFs obtained from Cistrome database and 50 hallmark pathways quantified by GSVA were integrated into co-expression analysis. The significant co-expression patterns were used to construct the bone metastasis-specific regulation network. Finally, spatial single-cell RNA-seq and chromatin immunoprecipitation sequence (ChIP-seq) data and multi-omics databases were applied to validate the key scientific hypothesis in the regulation network. Additionally, Connectivity Map (CMap) were utilized to select the potential inhibitors of bone metastasis-specific regulation network in BRCA. Results： Based on edgeR and WGCNA method, 43 PSRSs were identified. In the bone metastasis-specific regulation network, MAF positively regulated CD248 (R = 0.435, P < 0.001), and hallmark apical junction was the potential pathway of CD248 (R = 0.353, P < 0.001). This regulatory pattern was supported by spatial single-cell RNA sequence and ChIP-seq data and multi-omics online databases. Additionally, alexidine was identified as the possible inhibitor targeting bone metastasis of BRCA by CMap analysis. Conclusion： PSRSs played important roles in bone metastasis of BRCA, and the prognostic model based on PSRSs showed the good performance. Especially, we proposed that CD248 was the most significant PSRSs, which was positively regulated by MAF, influencing bone metastasis via apical junction pathway. And this axis might be inhibited by alexidine, providing a potential treatment strategy for bone metastasis of BRCA.