Osteosarcoma is the most common primary malignant bone tumor in orthopedics, with two peaks in adolescents and the elderly. The peak age of adolescent-onset is about 15 years old, mainly primary OS, and the second peak age is about 75 years old, mainly secondary OS [21]. OS mainly occurs in the epiphyseal region of the long diaphysis, where blood transport is abundant [1]. In the early stage of osteosarcoma, blood transport to the lung is most common and develops rapidly, which greatly reduces the survival of patients with osteosarcoma [22]. With the increase and improvement of treatment methods, the survival rate of patients with osteosarcoma has been greatly improved, but how to further improve the prognosis of patients is a major clinical challenge, especially for patients with pulmonary metastatic osteosarcoma, which occurs earlier and has a worse prognosis [23]. Hence, on the one hand, we should be looking for more effective treatments; on the other hand, we should also develop some new ideas. We can use high-throughput sequencing technology and existing sequencing results to screen genes, and then predict the prognosis of patients so that we can carry out more personalized treatment for patients.

For nearly 20 years, we have considered the Rho family as an anti-tumor target, especially for RAS-driven tumors. Rho family proteins transition between active GTP-binding states and inactive GDP-binding states, which are regulated by the ARHGAP family and can increase the intrinsic GTPase activity of Rho GTPase to convert it to inactive GDP-binding states [11]. Nowadays, it is believed that most ARHGAP genes have multiple functional domains except the RHOGAP functional domain, which integrates signal factors in many signaling pathways and may mediate the interaction between the Rho family and other signaling pathways. However, the number of ARHGAP proteins is much larger than that of their substrate Rho proteins, and ARHGAP proteins have diverse biological functions. Therefore, the deep regulatory mechanism of ARHGAP on the Rho family is still far from clear. The ARHGAP family is involved in many biological activities, such as exocytosis, endocytosis, cytokinesis, cell differentiation, cell migration, neuronal morphogenesis, angiogenesis, and tumor suppression [11]. In recent years, the relationship between ARHGAP family genes and tumor development, invasion, and metastasis has attracted more and more attention [24]. In this study, we screened the ARHGAP family genes and finally screened 5 genes to construct a risk model for evaluating the prognosis of patients with osteosarcoma, and to provide certain ideas and help for clinical treatment.

In this study, we first collected gene expression information and clinical information of patients with osteosarcoma in the TARGET cohort and the GSE39055 cohort and then removed patients without follow-up information and survival status to prepare the data for our next analysis. For better analysis, we then obtained the genetic information of the ARHGAP family from the online website, and uniformly named it in the two datasets. We then performed univariate COX analysis in the TCGA cohort and screened out 9 genes associated with the prognosis of patients with osteosarcoma. For subsequent analysis and verification, we randomly divided patients in the TCGA cohort into two groups at a ratio of 1: 1, namely the training cohort and the testing cohort. We performed LASSO regression analysis and multivariate COX regression analysis on the patients in the training cohort and finally obtained five genes: ARHGAP1, ARHGAP8, ARHGAP10, ARHGAP25, and ARHGAP28, and then we constructed a risk prediction model for patients with osteosarcoma based on these five genes. We calculated each patient's risk score and divided them into high- and low-risk groups, and then analyzed the relationship between survival status and survival time and risk score in the training cohort, testing cohort, entire internal cohort, and GSE39055 cohort respectively, finally, Kaplan–Meier survival analysis and ROC curve were performed for verification. The results show that the high-risk group had a lower survival time than the low-risk group. This suggests that our risk model can well predict the prognosis of patients with osteosarcoma. We can tell that these genes are protective factors for patients with osteosarcoma. Among them, ARHGAP1, ARHGAP8, and ARHGAP10 play different roles in different tumors or different pathways. ARHGAP1 was the first gene discovered in this family, and its content in cervical cancer cells and Ewing Sarcoma (ES) cells was lower than that in the matching normal tissue, which proved that it could inhibit the cell vitality, cell migration, and invasion of these two cancer cells in a time-dependent manner to a certain extent [25, 26]. However, in breast cancer (BC), ARHGAP1 is a carcinogenic factor, and its expression level in BC samples is higher than that in normal tissues, its overexpression can promote the proliferation and invasion of BC cells while inhibiting its expression can significantly inhibit the growth of tumors [27,28,29]. ARHGAP1 may also regulate the bone microenvironment by inhibiting the RhoA/ROCK pathway, which stimulates osteogenic differentiation of mesenchymal stem cells [30]. However, the role of ARHGAP1 in osteosarcoma has not been reported, which can be further studied in the future. Similarly, ARHGAP8 is overexpressed in most colorectal cancers compared to normal tissues, but we observe relatively low expression in Bladder cancer, suggesting that ARHGAP8 may play different roles in different tumors, but its role in osteosarcoma is unknown [31, 32]. ARHGAP10 is well known as a tumor suppressor and has been demonstrated in a variety of cancers, such as Uterine leiomyomas (ULs), prostate cancer, ovarian cancer (OC), lung cancer, colon carcinoma (CRC) and BC [33,34,35,36,37,38]. Cdc42, a key protein that cancer cells need to metastasize, helps them spread through the bloodstream to other parts of the body. In ovarian cancer, RHGAP10 inhibits Cdc42 activity in cells, in turn, it can inhibit the growth and invasion of tumors, thus playing a role in cancer suppression [37]. However, ARHGAP10 manifested as an oncogene in gastric tumors and non-small cell lung cancer (NSCLC) [39,40,41]. The expression level of ARHGAP10 in NSCLC is higher than that in normal tissues. When its expression is decreased, the expression of GLUT1 is also decreased, which inhibits the glucose metabolism process of cells and thus the progression of cancer [40]. Therefore, the role of ARHGAP10 in different tumors may be related to its participation in different pathways or different regulatory molecules upstream and downstream of the same pathway. However, its specific role in osteosarcoma is still unclear, so we can conduct further research. ARHGAP25 has also been widely studied as a tumor suppressor gene in cancer, including Pancreatic adenocarcinoma (PAAD), NSCLC, Lung cancer, and CRC [14, 42,43,44,45,46]. Epithelial-mesenchymal transition (EMT) is a common mechanism of tumor metastasis, which can reduce the adhesion between cells so that tumor cells can be separated from the original site to metastasize [47]. The Wnt/β-catenin pathway can increase the viability and invasion ability of cancer cells by activating EMT [48]. However, ARHGAP25 exerts its anticancer effects by negatively regulating EMT and Wnt/β-catenin pathways [44]. Of course, ARHGAP25 may regulate different pathways in different tumors to play a role in cancer inhibition. However, whether ARHGAP25 can inhibit the metastasis of osteosarcoma has not been studied, and it can become the object of our subsequent research [49]. Finally, the expression level of ARHGAP28 in osteosarcoma is significantly related to the prognosis and survival time of patients, but it has not been studied in osteosarcoma. Therefore, we speculate that ARHGAP28 is a tumor suppressor gene for osteosarcoma, and we plan to further study it in the next step.

We then constructed a nomogram based on the TCGA cohort to incorporate age, gender, and risk scores and simulated 2-, 3-, and 5-year time-dependent AUC curves for osteosarcoma patients in the TCGA and GSE39055 cohorts, respectively. The results showed that the risk score was an independent predictor of the prognosis of patients with osteosarcoma, and the simulation results of the AUC curve were good, which could better prove the accuracy and applicability of the model. To find out which functional pathway the molecular differences between the high- and low-risk groups are enriched in, to better screen the differential genes that can be used as targets for clinical diagnosis and treatment, we screened the differential genes between the high- and low-risk groups and conducted functional enrichment analysis. KEGG results show that it mainly enriched them in Protein digestion and absorption and Cytokine-cytokine receptor interaction, which also confirms our above analysis. They interact with different upstream and downstream molecules in different tumors to show different functions, and the process is very complex [11].

Nowadays, immunotherapy has been emphasized in the treatment of patients with osteosarcoma [50]. Therefore, we further studied whether they relate the risk model to the immune microenvironment, which can provide some ideas for the immunotherapy of osteosarcoma. GSEA analysis showed that the low-risk group had higher enrichment of immune function than the high-risk group [51], such as the B cell receptor signaling pathway, natural killer cell-mediated cytotoxicity, T cell receptor signaling pathway, and antigen processing and presentation. They have reported that Macrophages M2 can promote the generation of tumors [52,53,54]. In our study, the content of Macrophages M2 is relatively high in osteosarcoma. ssGSEA showed that the content of immune cells in the high-risk group was much lower than that in the low-risk group, indicating that the occurrence and development of osteosarcoma are closely related to the immune environment, which will provide new ideas for us to find new therapeutic targets and methods for osteosarcoma in the future.

Since the role of ARHGAP28 in osteosarcoma remains unclear, we confirmed the role of ARHGAP28 through in vitro and in vivo biological experiments. Overexpression of ARHGAP28 had significant effects on the viability, proliferation, migration, and invasion of OS cells. We found that overexpression of ARHGAP28 can inhibit the proliferation, migration, and invasion of osteosarcoma cells. In vivo experiments have shown that overexpression of ARHGAP28 can inhibit tumor growth in mice, and IHC has shown that the reduced level of Ki-67 in the ARHGAP28 overexpression group can inhibit the proliferation of tumor cells. In summary, ARHGAP28 may play a positive role in inhibiting the growth and progression of osteosarcoma.

However, inevitably, our research also has some shortcomings. First, we only used an external GSE39055 cohort for verification, which may have some discrepancies in some data sets. Second, the expression levels of ARHGAP1, ARHGAP8, and ARHGAP10 in our model showed the same trend with the prognosis and survival time of patients with osteosarcoma, but there was no significant correlation. Whether the model constructed by combining these five genes is also applicable to other cohorts needs further verification. Third, we lack clinical samples to verify the accuracy of the model we constructed, so we can only test our hypothesis with cell experiments. Finally, we did not investigate ARHGAP28 further, such as its relationship to human immunity.

The study categorized OS invalids into risk groups based on the ARHGAP family. The high-OS group displayed abnormal immune function, such as the B cell receptor signaling pathway, natural killer cell-mediated cytotoxicity, T cell receptor signaling pathway, and antigen processing and presentation. The results show that ARHGAP family genes are likely to play a role in the immune function of the human body, inhibiting the occurrence and progression of tumors, and these gene targets may also be promising personalized drug targets.

In summary, we constructed a five-gene (ARHGAP1, ARHGAP8, ARHGAP10, ARHGAP25, and ARHGAP28) risk prognostic model based on the ARHGAP family. It can predict the prognosis of patients with osteosarcoma, and verify its accuracy and universality. Finally, we also analyzed the relationship between it and the immune system of patients, which provided ideas and directions for our follow-up research and the management and treatment of clinical patients.