Background and purpose: p21-activated kinase 5 (PAK5) is a recently identified member of PAKs that regulate many intracellular processes such as cytoskeleton remodeling, cell proliferation, cell differentiation, gene transcription and cell apoptosis. Recently, studies found that PAK5 was overexpressed in some cancer such as gastric and colon cancer. However, the expression status and biological function of PAK5 in osteosarcoma are not clearly known. The objective of this study was to investigate the expression of PAK5 in osteosarcoma tissue and their relationships with the prognosis of osteosarcoma. Methods: The expression of PAK5 was detected by using immunohistochemical method in 92 specimens of human osteosarcoma tissues and 33 cases of osteoclastoma tissue, respectively. Results: The positive rate of PAK5 was 71.7% (66/92) in all the 92 cases of osteosarcoma. PAK5 expressions were not related to clinical variables such as gender, age, tumor location, tumor size, histological type and local recurrence, but signiifcantly related to Enneking grade, tumor cell necrosis rate and lung metastasis, and the high expression of PAK5 may reduce the efifciency of chemotherapy. Survival analysis indicated that high expression of PAK5 correlated with poor prognosis of patients with osteosarcoma. Univariate survival analysis showed that the signiifcant prognostic factors were tumor size, Enneking grade, local recurrence, lung metastasis and expression levels of PAK5. COX multivariate regression identified that the PAK5 expression levels (P=0.001) and lung metastasis (P=0.015) were independent prognostic factors of patients with osteosarcoma. Conclusion:The positive expressions of PAK5 closely correlate with Enneking grade, tumor cell necrosis rate and lung metastasis. Detection of PAK5 may be used as a molecular marker for prognosis of osteosarcoma. The high expression of PAK5 may reduce the efifciency of chemotherapy.

Objective:To screen T-cell exhaustion-related signature genes as the prognostic marker for osteosarcoma and establish a prognostic model for osteosarcoma patients based on weighted gene co-expression network analysis(WGCNA)and Least absolute shrinkage and selection operator(LASSO)-COX regression analysis. Methods:GSE21257 dataset was downloaded from Gene Expression Omnibus(GEO)database for the establishment of the prognostic model for osteosarcoma.4 T-cell exhaustion-related gene sets were downloaded from The Molecular Signatures Database(MisgDB)and their enrichment scores in GSE21257 samples were calculated by single sample gene set enrichment analysis(ssGSEA).WGCNA was carried out to screen the gene module that is highly associated with T-cell exhaustion based on ssGSEA results followed by GO(Gene Ontology)and KEGG(Kyoto Encyclopedia of Genes and Genomes)analysis of the biological processes and signaling transduction pathways that those genes are involved in.The signature genes that are highly associated with the prognosis of osteosarcoma patients were obtained through LASSO-COX regression and a prognostic model was established based on these signature genes.Osteosarcoma-related expression profile data from the GSE21257 and TAEGET datasets on XENA were downloaded from the Gene Expression Omnibus.Clinical information for the training and validation sets was obtained.T-cell exhaustion-related genes were screened using a weighted correlation network analysis.Realtime fluorescence quantitative PCR,COX regression analysis,external dataset and nomogram were used to evaluate the reliability and accuracy of the prognostic model.A immunotherapy-related dataset was used to assess the efficacy of this prognostic model for the prediction of patients'responses to immunotherapy. Results:Analysis results based on the ssGSEA scores showed that T-cell exhaustion-related genes were related to the metastasis and age of osteosarcoma patients.Many T-cell exhaustion-related genes were found to be differentially expressed in metastatic and non-metastatic osteosarcoma patients.1 256 T-cell exhaustion-related genes were identified through WGCNA and these candidate markers were mainly distributed in structures like secretory granule membranes and endocytic vesicles and were involved in T-cell activation.COX regression analysis screened 68 significant prognostic markers out of the 1 256 genes,and 12 signature genes were further confirmed with LASSO-COX regression analysis.A prognostic model was established based on the 12 signature genes.Results of real-time fluorescence quantitative PCR showed a similar trend in the expression of most of the signature genes in different osteosarcoma cell lines.COX regression analysis of the internal and external datasets verified that the risk score calculated with the prognostic model was an independent prognostic factor for osteosarcoma patients,and high-risk score was associated with poor prognosis of the patients.Receiver operating characteristic(ROC)curves demonstrated excellent prognostic efficacy of the model.Nomogram analysis verified the prognostic model is highly accurate and reliable in predicting the prognosis of osteosarcoma patients.Analysis using the immunotherapy-related dataset indicated that this prognostic model could also be used to predict patients'responses to immunotherapy. Conclusion:The 12 signature gene(CD300LB,TRO,SNX3,VENTX,PPM1M,DOT1L,CDC37,NAT9,TRMT1,PPP1R3C,CHTF18 and NSUN5)-based prognostic model can effectively predict the prognosis and responses to immune check-point inhibitors for osteosarcoma patients,which may provide evidence for the prediction of prognosis as well as the selection of immunotherapy plans in clinical practice.