Frontiers in Oncology · 2021 · 12 citations · 35 references
Life SciencesBioinformaticsOncologyCarcinogenesisGene ExpressionMolecular OncologyBone Morphogenetic Protein SignalingCancer BiomarkersCancer Survival AnalysisCancer TranscriptomicsDifferential Gene Expression AnalysisMolecular Tumor ProfilingBone Morphogenetic Protein ReceptorsCox RegressionDifferential Co-expression AnalysisDifferential Gene ExpressionGene Co-expression Network AnalysisGene Network AnalysisLung AdenocarcinomaPrognostic Expression Signatures
Background Bone morphogenetic proteins (BMPs) regulate tumor progression via binding to their receptors (BMPRs). However, the expression and clinical significance of BMPs/BMPRs in lung adenocarcinoma remain unclear due to a lack of systematic studies. Methods This study screened differentially expressed BMPs/BMPRs (deBMPs/BMPRs) in a training dataset combining TCGA-LUAD and GTEx-LUNG and verified them in four GEO datasets. Their prognostic value was evaluated via univariate and multivariate Cox regression analyses. LASSO was performed to construct an initial risk model. Subsequently, after weighted gene co-expression network analysis (WGCNA), differential expression analysis, and univariate Cox regression analysis, hub genes co-expressed with differentially expressed BMPs/BMPRs were filtered out to improve the risk model and explore potential mechanisms. The improved risk model was re-established via LASSO combining hub genes with differentially expressed BMPs/BMPRs as the core. In the testing cohort including 93 lung adenocarcinoma patients, immunohistochemistry (IHC) was performed to verify BMP5 protein expression and its association with prognosis. Results BMP2, BMP5, BMP6, GDF10, and ACVRL1 were verified as downregulated in lung adenocarcinoma. Survival analysis identified BMP5 as an independent protective prognostic factor. We also found that BMP5 was significantly correlated with EGFR expression and mutations, suggesting that BMP5 may play a role in targeted therapy. The initial risk model containing only BMP5 showed a significant correlation (HR: 1.71, 95% CI: 1.28−2.28, p : 3e-04) but low prognostic accuracy (AUC of 1-year survival: 0.6, 3-year survival: 0.6, 5-year survival: 0.63). Seventy-nine hub genes co-expressed with BMP5 were identified, and their functions were enriched in cell migration and tumor metastasis. The re-established risk model showed greater prognostic correlation (HR: 2.58, 95% CI: 1.92–3.46, p : 0) and value (AUC of 1-year survival: 0.72, 3-year survival: 0.69, and 5-year survival: 0.68). IHC results revealed that BMP5 protein was also downregulated in lung adenocarcinoma and higher expression was markedly associated with better prognosis (HR: 0.44, 95% CI: 0.23–0.85, p : 0.0145). Conclusion BMP5 is a potential crucial target for lung adenocarcinoma treatment based on significant differential expression and superior prognostic value.
35
clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters
Guangchuang Yu, Li-Gen Wang, Yanyan Han et al. · OMICS A Journal of Integrative Biology · 2012 · 38.3K citations
WGCNA: an R package for weighted correlation network analysis
Peter Langfelder, Steve Horvath · BMC Bioinformatics · 2008 · 29.1K citations · Full text
Rebecca L. Siegel, Kimberly D. Miller, Ahmedin Jemal · CA A Cancer Journal for Clinicians · 2019 · 21K citations · Full text
The Genotype-Tissue Expression (GTEx) pilot analysis: Multitissue gene regulation in humans
Kristin G. Ardlie, David S. DeLuca, Ayellet V. Segrè et al. · Science · 2015 · 5.7K citations · Full text
Bone: Formation by Autoinduction
Marshall R. Urist · Science · 1965 · 5.6K citations