MASP1 as a favorable prognostic biomarker in pediatric osteosarcoma: an integrated analysis of machine learning, bioinformatics, and validation experiments

Introduction

Osteosarcoma (OS) is a rare bone-forming tumor that primarily affects children and adolescents. It has a worldwide annual incidence of approximately 1–3 cases per million individuals (1,2). Despite accounting for only about 5% of cancers in children and adolescents, OS contributes significantly to pediatric cancer-related deaths. Since the 1970s, the survival rates of non-metastatic OS have dramatically improved, but this has not been the case for metastatic OS (3).

Before the establishment of surgery with the neo- and adjuvant chemotherapy, OS was largely treated by surgery alone. However, even with complete local control and wide margins, 80–90% of OS patients develop fatal pulmonary metastases within 1 year of diagnosis (4). This shows the high malignancy of OS and suggests that undetectable micrometastases may be present at the time of diagnosis. Therefore, in addition to surgical treatment, systemic therapy is required.

Following the advent of chemotherapy in the second half of the 20th century, neoadjuvant and adjuvant therapies were applied to OS, increasing the survival rate of OS patients to 70% (5). However, this has not significantly improved the prognosis of patients with large metastases at diagnosis or those with recurrent disease, whose survival rates remain at only 20% (6). Unlike many other pediatric cancers, intensive therapy has failed to significantly improve the survival rates of OS patients over the past decades. Thus, there is an urgent need to identify new early diagnostic and treatment response biomarkers that could also serve as potential therapeutic targets for OS.

Heterogeneity, both within tumors and among individuals, is a major characteristic of OS tumors, and is also one of the reasons for its inconsistent therapeutic outcomes. The common genomic initiating biological processes (BPs) driving the development of OS remain unknown. The complexity of the OS somatic genome is the main cause of intratumoral heterogeneity, which is characterized by chromosomal aneuploidy, gene mutations and/or copy number alterations, intratumoral heterogeneity, and genomic instability (7). This instability is marked by large-scale chromosomal rearrangements and the presence of localized hypermutation patterns known as “kataegis” (8). A small subset of genes (TP53, RB, MDM2, ATRX, and DLG2) have been found to be recurrently mutated in OS (9). Recently, a subset of OS genes with genomic alterations in DNA repair pathways were identified (10). These alterations in DNA repair pathway genes share similarities with BRCA1/2-deficient tumors (11). Certain hereditary syndromes, such as Li-Fraumeni, Rothmund-Thomson, Werner, Bloom, and familial retinoblastoma, are also associated with an increased risk of OS (8). However, in the majority of cases (95%), OS is sporadic (8).

In summary, due to the undefined oncogenic events associated with the high cellular heterogeneity of tumor cells, the development of molecularly targeted therapies, especially those for OS, is extremely challenging. OS grows in the bone microenvironment, a highly specific, complex, and dynamic environment composed of various cells, including osteoblasts, stromal cells, vascular cells, immune cells, and a mineralized extracellular matrix (ECM). Under physiological conditions, the coordinated and fine-tuned activities of osteoblasts, vascular cells, and stromal cells ensure bone homeostasis. According to Paget’s theory (12), tumor cells proliferate rapidly in this microenvironment and manage to regulate bone physiological pathways to their advantage, thereby gaining a survival and growth advantage. The cross-talk between OS and the bone microenvironment involves multiple environmental signals, including various cytokines and chemokines (13). In addition to the high heterogeneity of tumor cells, and the complex regulation of the tumor genome and microenvironment, chemoresistance is also a significant challenge in the treatment of OS. Even patients who initially respond well to treatment typically require very high doses of combination chemotherapy. At least 30% of OS patients cannot be cured using current surgical and chemotherapy combinations (14).

Targeted therapies and immunotherapies may improve future OS treatments. Additionally, the identification of resistance mechanisms could lead to the development of strategies to overcome resistance during ongoing therapy, thereby sensitizing OS patients to chemotherapy. Moreover, in recent years, immunotherapy has achieved good therapeutic effects in various solid tumors. Due to the low immunogenicity and immunosuppressive microenvironment of OS, immunotherapy has not yet been approved in OS patients. Thus, the present study sought to identify immune related molecular markers with potential prognostic and therapeutic value in OS via a transcriptomic analysis of genomic alterations, machine-learning models, an analysis of immune cell infiltration in the immune microenvironment, and an analysis of the effects of different immune cells on OS prognosis and sensitivity to conventional drugs. We present this article in accordance with the TRIPOD reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-262/rc).

Methods

Downloading and preprocessing of raw data

The following three datasets were downloaded from the Gene Expression Omnibus (GEO) database: GSE42352 (comprising 3 normal samples, and 84 OS samples), GSE126209 (comprising 11 normal samples, and 12 OS samples), and GSE12865 (comprising two normal samples, and 12 OS samples). The sva package in R software was used for the batch correction and the merging of the datasets for the differential expression analysis. Additionally, the GSE16091, GSE21257, and GSE39058 datasets were downloaded, which contained the data of 34, 53, and 37 OS patients, respectively, along with their transcriptomic data and overall survival information. These datasets were merged using the sva package for the survival analysis. We also downloaded the complete clinical and transcriptomic data of 88 OS patients from the TARGET database (https://www.cancer.gov/ccg/access-data). Finally, an immune-related gene list comprising 1793 genes was downloaded from the ImmPort database (https://immport.org/shared/home).

Acquisition of immune-related DEGs in OS

GSE42352, GSE126209, and GSE12865 underwent batch correction and merging to obtain a differential expression dataset. A principal component analysis (PCA) was conducted using the PCA package to verify the gene expression levels before and after batch correction. The differential expression analysis was carried out using the limma package based on the following criteria: a log₂ fold change (|log₂FC|) >1 and a false discovery rate (FDR) <0.05. The VennDiagram package was used to intersect the differentially expressed genes (DEGs) and immune-related genes to identify the immune-related DEGs in OS.

Functional enrichment analysis of immune-related DEGs in OS

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analyses of the immune-related DEGs were performed using the org.Hs.eg.db and clusterProfiler packages. The results were visualized using ggplot2 and ggpubr. Only the top 10 GO functions and KEGG pathways with a q value <0.05 were displayed.

Model construction using machine-learning methods

Machine-learning techniques, including eXtreme Gradient Boosting (XGB), random forest (RF), generalized linear model (GLM), and SVM, were applied to construct diagnostic models for OS using the caret, randomForest, xgboost, and kernlab packages. The model optimization was based on the residual reverse cumulative distribution, residual root mean square minimum values, and the areas under the curve (AUCs) of the receiver operating characteristic (ROC) curves for the four machine-learning methods. The model gene with the smallest inverse cumulative residual distribution and the smallest root mean square residual, as well as the largest area under the ROC curve, was selected for subsequent analyses. Bar charts generated using the rms package were used to identify key genes in the predictive models. In this study, we constructed machine-learning models based on the immune-related DEGs from the differential expression cohort of the GEO.

Identification of key genes in OS and their expression-outcome correlations

The GSE16091, GSE21257, and GSE39058 datasets were merged to obtain a survival cohort comprising the transcriptomic data and overall survival clinical information of 124 OS patients (Table 1). A univariate Cox regression survival analysis was performed to explore the correlation between the expression of the immune-related DEGs and OS. The genes identified by the optimal machine-learning model and those with P values <0.05 in the univariate Cox analysis were intersected to identify the key genes of OS. Boxplots were used to compare the expression differences of the key genes between the normal bone tissues and OS bone tumor tissues. The ROC curves were plotted using the pROC package to evaluate the diagnostic accuracy of the expression levels of the key genes in differentiating between normal and OS tissues. Additionally, a survival analysis of the expression of the key genes and overall survival was conducted using the survival datasets from the GEO and TARGET databases.

Construction of the ceRNA network for key genes

Potential target microRNAs (miRNAs) for the key genes were predicted using the following four databases: miRDB, miRanda, miRWalk, and TargetScan. Only miRNAs identified in all four databases were considered candidate miRNAs. Subsequently, target miRNA-related long non-coding RNAs (lncRNAs) were predicted using the SpongeScan database. A messenger RNA-miRNA-lncRNA ceRNA network was constructed using Cytoscape (version 3.10.1).

Analysis of immune cell infiltration in OS

The batch correction and the merging of the GEO survival cohort and TARGET cohort were performed using the sva package. The CIBERSORT algorithm was applied to calculate the relative abundance of 22 immune cell types for each OS sample (the total immune cell content across the 22 types summed to 100%). A Spearman’s correlation analysis was then conducted to assess the relationship between the expression of the key genes and immune cell content. Additionally, the association between immune cell content and patient prognosis was analyzed.

Analysis of the expression of key genes and drug sensitivity

The pRRophetic R package, which predicts drug sensitivity based on gene expression data and drug chemical characteristics, was employed to analyze drug sensitivity. The merged gene expression matrix from the GEO survival cohort and TARGET cohort were used for this analysis. The cohort was divided into high- and low-expression groups based on the median expression of the key genes, and the drug sensitivity differences between the two groups were analyzed. The analysis focused on common chemotherapy and targeted drugs used in OS.

RNA isolation and reverse transcription-polymerase chain reaction (RT-qPCR)

OS and normal specimens for RT-qPCR were provided by Baise People’s Hospital. The RNA extraction and complementary DNA synthesis were carried out using Trizol reagent (TaKaRa, Otsu, Japan) and the PrimeScript RT reagent kit with gDNA Eraser (Perfect Real Time, TaKaRa, Japan). Real-time fluorescence quantitative polymerase chain reaction (qPCR) was conducted using SYBR^® Fast qPCR Mix (TAKARA), and the relative expression levels of the genes were normalized to β-actin using the 2^-ΔΔCt method. The forward and reverse primer sequences for the target gene MASP1 were: CACCGTGGAGCTAAACAATATGT and GCTTGATCCGAAACCCATCTG.

Statistical analysis

All the statistical analyses were performed using R software (version 4.3.1). A Kaplan-Meier survival analysis was conducted to assess the differential survival of OS patients, and a log-rank P value <0.05 was considered statistically significant. The difference analysis and correlation analysis were performed via Wilcoxon and Spearman correlation, respectively. A two-sided P value <0.05 was considered statistically significant.

Ethical statements

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was reviewed by the Ethics Committee of the Baise People’s Hospital (Ethics No. 2024-E488-01). The requirement of informed consent was waived due to retrospective nature of the study.

Results

Immune-related DEGs in OS

After the batch correction and merging of the datasets (GSE42352, GSE126209, and GSE12865), the PCA confirmed the successful removal of the batch effects (Figure 1A,1B). The differential expression analysis identified a total of 1370 DEGs for OS, of which, 748 were upregulated and 622 were downregulated (Figure 1C,1D). The immune-related genes and DEGs were intersected using the VennDiagram package, and a set of 174 immune-related DEGs significantly associated with OS were identified (Figure 1E).

Figure 1 Immune-related DEGs in OS. (A) PCA before batch correction. (B) PCA after batch correction, confirming the removal of the batch effects. (C) Heatmap displaying the top 50 DEGs. (D) Volcano plot showing 748 upregulated and 622 downregulated DEGs in OS. (E) Venn diagram identifying 174 immune-related DEGs in OS. DEGs, differentially expressed genes; FC, fold change; OS, osteosarcoma; PCA, principal component analysis.

Functional enrichment analysis of immune-related DEGs

A functional enrichment analysis of the 174 immune-related DEGs was performed using the GO and KEGG pathways. The top 10 BPs of the GO terms were the cytokine-mediated signaling pathway, leukocyte proliferation, response to molecule of bacterial origin, response to lipopolysaccharide, peptidyl-tyrosine modification, peptidyl-tyrosine phosphorylation, mononuclear cell proliferation, cell chemotaxis, regulation of peptidyl-tyrosine phosphorylation, and myeloid leukocyte migration; The top 10 cellular components (CCs) of the GO terms were collagen-containing ECM, vesicle lumen, cell-substrate junction, focal adhesion, cytoplasmic vesicle lumen, secretory granule lumen, external side of plasma membrane, membrane raft, endoplasmic reticulum lumen, and lumenal side of membrane; The top 10 molecular functions (MFs) of the GO terms were receptor ligand activity, cytokine receptor binding, cytokine activity, growth factor activity, glycosaminoglycan binding, cytokine binding, growth factor receptor binding, growth factor binding, hormone activity, and major histocompatibility complex (MHC) class II protein complex binding (Figure 2A,2B).

Figure 2 Functional enrichment analysis of the immune-related DEGs. (A) Bar plot of the GO functional enrichment analysis. (B) Circle diagram of the BPs, CCs, and MFs of the GO terms. (C) Bubble diagram of the KEGG signaling pathways. (D) Circle diagram of the KEGG signaling pathways. BP, biological process; CC, cellular component; DEGs, differentially expressed genes; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

The results of the KEGG functional enrichment analysis revealed that the top 10 enriched signaling pathways were the cytokine-cytokine receptor interaction, PI3K-Akt signaling pathway, MAPK signaling pathway, human cytomegalovirus infection, Influenza A, the Rap1 signaling pathway, Kaposi sarcoma-associated herpesvirus infection, lipid and atherosclerosis, natural killer (NK) cell-mediated cytotoxicity, and viral protein interaction with cytokine and cytokine receptor (Figure 2C,2D).

Model construction using machine learning

Four machine-learning algorithms (XGB, RF, GLM, and SVM) were applied to construct diagnostic models for OS. Model optimization, which was based on the residual reverse cumulative distribution, residual root mean square minimum values, and the AUCs of the ROC curves, showed that the SVM outperformed the other models, achieving the minimum values for the residual root mean square (Figure 3A) and residual reverse cumulative (Figure 3B), and the highest AUC (Figure 3C). The key predictive genes were identified via a feature importance analysis of the SVM, and a bar chart visualization revealed 10 critical immune-related DEGs with high predictive value (Figure 3D).

Figure 3 Model construction using machine learning. (A) Residual root mean square boxplot comparing model performance (XGB, RF, GLM, and SVM). (B) Residual reverse cumulative distribution for model performance (XGB, RF, GLM, and SVM). (C) ROC curves for model performance (XGB, RF, GLM, and SVM). (D) Bar chart showing the top 10 predictive immune-related DEGs from the XGB, RF, GLM, and SVM models. DEGs, differentially expressed genes; GLM, generalized linear model; RF, random forest; ROC, receiver operating characteristic; SVM, support vector machine; XGB, eXtreme Gradient Boosting.

Key genes in OS and their expression-outcome correlation

The survival analysis of the immune-related DEGs was performed on a merged survival cohort of GSE16091, GSE21257, and GSE39058, comprising the data of 124 OS patients. The univariate Cox regression revealed that seven genes were significantly associated with the prognosis of the OS patients (P<0.05) (Figure 4A). The intersection of the genes from the univariate Cox model and the optimal SVM model resulted in the identification of a core gene, MASP1 (Figure 4B). As the boxplot in Figure 4C shows, the expression of MASP1 was significantly more reduced in the OS tumor tissues than the normal bone tissues. The ROC curve analysis further confirmed the diagnostic value of MASP1, which had an AUC of 0.903 [95% confidence interval (CI): 0.769–0.993] (Figure 4D). The prognostic analysis revealed that higher MASP1 expression is associated with a better prognosis in OS patients, and this finding was consistent across both the GEO and TARGET datasets (Figure 4E,4F).

Figure 4 Key genes in OS and their expression-outcome correlation. (A) Univariate Cox analysis identifying seven genes associated with OS prognosis. (B) Intersection of the key genes from univariate Cox and SVM models, identifying MASP1. (C) Boxplot showing reduced MASP1 expression in OS tumors. (D) ROC curve for MASP1 (AUC: 0.903). (E,F) Kaplan-Meier survival analysis showing that high MASP1 expression was associated with a better prognosis in the GEO and TARGET databases. AUC, area under the curve; CI, confidence interval; GEO, Gene Expression Omnibus; OS, osteosarcoma; ROC, receiver operating characteristic; SVM, support vector machine.

ceRNA network for key genes

A competing endogenous RNA (ceRNA) network for the key gene was constructed using Cytoscape. The network comprised one key gene, 10 miRNAs, and 59 lncRNAs, which were predicted to interact with each other (Figure 5). The target miRNAs were identified from four databases (miRDB, miRanda, miRWalk, and TargetScan), and only those present in all four databases were considered to further analysis. The ceRNA network highlighted the potential regulatory interactions between the immune-related genes and their upstream regulators.

Figure 5 ceRNA network for the key genes. ceRNA network for MASP1, which included one key gene, 10 miRNAs, and 59 lncRNAs. ceRNA, competing endogenous RNA; lncRNA, long non-coding RNA; miRNA, microRNA.

Immune cell infiltration in OS

The immune infiltration analysis revealed that the expression level of MASP1 was positively correlated with the infiltration of memory B cells, regulatory T cells, CD8 T cells, activated mast cells, and resting NK cells, but negatively correlated with the infiltration of resting memory CD4 T cells (Figure 6A). The univariate Cox analysis showed that among these six immune cell types, only the infiltration level of the resting memory CD4 T cells was significantly associated with the prognosis of the OS patients (P<0.05) (Figure 6B). The Spearman’s correlation analysis showed that the high expression of MASP1 was weak correlated with a reduced abundance of resting memory CD4 T cells (r=–0.14, P=0.04) (Figure 6C). The high infiltration level of the resting memory CD4 T cells was found to be associated with a poor prognosis in the OS patients (Figure 6D).

Figure 6 Immune cell infiltration in OS. (A) Correlation between MASP1 expression and immune cell infiltration. (B) Univariate Cox analysis of immune cell types and OS prognosis. (C) Spearman’s correlation between MASP1 expression and resting memory CD4 T cells. (D) Survival analysis showing a poor prognosis with high resting memory CD4 T cells infiltration. OS, osteosarcoma.

Key gene expression and drug sensitivity

The drug sensitivity analysis showed that a high expression of MASP1 increased the sensitivity of OS patients to commonly used chemotherapy and targeted drugs, such as doxorubicin (Figure 7A), vinblastine (Figure 7B), gemcitabine (Figure 7C), and sorafenib (Figure 7D). However, the expression level of MASP1 did not affect the sensitivity of the OS patients to cisplatin (Figure 7E) and methotrexate (Figure 7F). The expression of MASP1 was significantly decreased in the OS tissue samples (Figure 8). The expression trend observed in the prognostic genes in the tissues was consistent with the open database results.

Figure 7 Key gene expression and drug sensitivity. Drug sensitivity analysis showing that high MASP1 expression increased sensitivity to (A) doxorubicin, (B) vinblastine, (C) gemcitabine, and (D) sorafenib, but not to (E) cisplatin and (F) methotrexate.

Figure 8 qPCR validation of the expression levels of the key genes. OS, osteosarcoma; qPCR, quantitative polymerase chain reaction.

Discussion

Based on the highly heterogeneous nature of OS, we conducted a comprehensive analysis using transcriptomics and machine learning, and identified MASP1 as a novel molecular marker for a favorable prognosis and chemosensitivity in OS. Our findings align with those of numerous studies that have reported that OS is a highly heterogeneous pediatric malignancy, for which genomic heterogeneity is a key characteristic. Through the analysis of publicly available sequencing data, we identified 748 upregulated and 622 downregulated genes in the OS patients compared to the healthy controls. Given the emerging role of immunotherapy in the treatment of OS (15,16), we further intersected these DEGs with immune-related gene sets, and identified 174 immune-related DEGs significantly associated with OS. These results highlight the potential value of immune regulation and immunotherapy in the treatment of OS.

A functional analysis of these 174 immune-related genes revealed their involvement in BPs such as cytokine-mediated signaling pathways, leukocyte proliferation, collagen-containing ECM, and cytokine receptor binding. These findings are consistent with those of previous studies, which have shown the critical role of cytokines in OS progression (17-21). For example, the interleukin (IL)-17-IL-17RA axis has been shown to be essential to OS progression in mice (21). Additionally, OS cells secrete CXCL14 to activate integrin α11β1 on fibroblasts, facilitating pre-metastatic niche formation in the lungs (20). The findings of these studies underscore the importance of cytokines in OS development, and align with the BPs of the immune-related DEGs identified in our study.

The ECM is undeniably crucial in OS (22,23). A study has shown that membrane-anchored and tumor-targeted IL-12 T cell therapy can disrupt cancer-associated fibroblasts and ECM in heterogeneous OS xenograft models (17). Other research has shown that modulating ECM production and oxygen diffusion can influence the chemotherapeutic response in OS (24). These findings suggest that the BPs regulated by these genes play significant roles in the development and progression of OS.

Most single-gene alterations do not play a decisive role in tumor development but rather affect tumor biology through their involvement in molecular signaling pathways, thereby profoundly affecting tumor occurrence, progression, and treatment outcomes. Therefore, we also conducted a functional enrichment analysis of the 174 genes to identify the potential molecular signaling pathways significantly affecting OS. Our results identified several key pathways, including the PI3K-Akt signaling pathway, MAPK signaling pathway, Rap1 signaling pathway, and NK cell-mediated cytotoxicity. The importance of these pathways in OS has been shown in multiple studies (25-31). For example, USP3 has been shown to promote OS progression by deubiquitinating EPHA2 and activating the PI3K/AKT pathway (27). RILP has been shown to inhibit OS progression through the Grb10-mediated PI3K/AKT/mTOR pathway (29). Tomivosertib (eFT508), a MNK1/2 inhibitor, has been shown to overcome chemoresistance in preclinical models of OS by targeting MAPK-interacting kinases (25). Additionally, NK cell immunotherapy has been explored as a potential treatment for OS, and a research has shown its efficacy in targeting OS cells and lung metastases (26). These findings suggest that the molecular pathways involved in the 174 genes not only play crucial roles in the biology of OS but also hold potential as novel therapeutic targets or strategies.

The diagnosis of OS is often delayed due to the initial presentation of clinical symptoms followed by the imaging detection of lesions, resulting in a low number of early diagnoses. This delay in diagnosis is one of the reasons for the poor prognosis of OS patients. Prognostic assessment in OS primarily relies on clinical disease staging. In our study, we identified seven genes significantly associated with OS prognosis through a univariate Cox regression analysis. The intersection of the genes from the univariate Cox model and the optimal SVM model resulted in the identification of a core gene, MASP1. The prognostic analysis showed that higher MASP1 expression was associated with a better prognosis in the OS patients. This finding was consistent across both the GEO and TARGET datasets. Further, MASP1 was found to have an AUC of 0.903 (95% CI: 0.769–0.993), which confirmed its diagnostic value.

To date, research on MASP1 in the oncology field has been limited. Yu et al. conducted a pan-cancer analysis of MASP1 using bioinformatics and experimental approaches, and found that high MASP1 expression in hepatocellular carcinoma cell lines inhibits the malignant biological behavior of liver cancer cells (32). Similarly, Zhang et al. reported low MASP1 expression in gastric adenocarcinoma patients (33). Given the importance of treatment resistance in OS, we further analyzed the effect of MASP1 on the sensitivity of OS to commonly used therapeutic drugs. The results showed that high MASP1 expression increased the sensitivity of OS patients to chemotherapy and targeted drugs such as doxorubicin, vinblastine, gemcitabine, and sorafenib, while MASP1 expression did not affect their sensitivity to cisplatin and methotrexate, which are the cornerstone drugs in the standard OS treatment regimen. This suggests that MASP1 does not interfere with the efficacy of these foundational drugs but may affect OS prognosis through other BPs.

We also examined the non-coding RNAs regulating MASP1, and identified several miRNAs (e.g., hsa-let-7f-2-3p, hsa-miR-1207-5p, hsa-miR-130a-5p, hsa-miR-130b-5p, hsa-miR-1976, hsa-miR-30b-3p, hsa-miR-574-5p, hsa-miR-576-5p, hsa-miR-873-5p, and hsa-miR-93-3p) and lncRNAs (e.g., C10orf91, RP5-894D12.5, RP11-102K13.5, CTC-459F4.1, RP5-892K4.1, LL22NC03-27C5.1, RP11-618K13.2, RP13-580B18.4, MUC2, and RP11-333E1.2). Previous research has shown than hsa-miR-574-5p negatively regulates MACC-1 expression to inhibit liver metastasis in colorectal cancer (34), while the high expression of C10orf91 and LINC01224 has been found to be associated with a poor prognosis in hepatocellular carcinoma (35). However, the specific regulatory mechanisms of these lncRNAs targeting MASP1 and their effects on different tumor types require further experimental validation.

Cytokines and the tumor microenvironment (TME) play crucial regulatory roles in OS development and treatment resistance. As key components of the TME, immune cells directly influence the biological behavior of tumor cells. Given the emerging interest in immunotherapy for OS, we analyzed the correlation between MASP1 expression and immune cell infiltration in the TME. The results showed that MASP1 expression was positively correlated with the infiltration of B cells (memory), CD8⁺ T cells, and resting NK cells, but negatively correlated with the infiltration of resting CD4⁺ memory T cells. The univariate Cox analysis revealed that only the infiltration level of resting CD4⁺ memory T cells was significantly associated with the prognosis of OS patients (P<0.05). The Spearman correlation analysis showed that high MASP1 expression was correlated with the reduced infiltration of resting CD4⁺ memory T cells (r=–0.14, P=0.04). The high infiltration of resting CD4⁺ memory T cells was associated with a poor prognosis in the OS patients. These findings suggest that MASP1 may affect OS prognosis by modulating immune cell infiltration in the TME. This conclusion is consistent with the findings of Yu et al., who systematically examined the effect of different immune cells on OS (16). Additionally, Li et al. found that the interaction between osteoclasts and regulatory CD4⁺ T cells significantly alters the TME, and is associated with a poor prognosis in OS (36). More research needs to be conducted on MASP1 to further explore its BPs in OS.

This study was the first to identify MASP1 as a highly predictive prognostic biomarker of OS, which has significant implications for treatment sensitivity. However, the limitations of the present study include a lack of any detailed analysis of the molecular pathways and the BP functions and the regulators such as miRNAs and lncRNAs involved in MASP1. Further, its expression levels in OS patients and healthy controls were only preliminarily validated. Future research should seek to elucidate the molecular mechanisms underlying the role of MASP1 in OS and its potential as a therapeutic target.

Conclusions

We developed an efficient and accurate prognostic model to help clinicians assess the prognosis of OS patients and to predict the sensitivity of some therapeutic agents. Notably, this study found that MASP1 exerted protective effects in OS patients. This study was the first to examine the role of MASP1 in the prognosis of OS patients. MASP1 affects the prognosis of OS patients by regulating the T cells CD4 memory infiltrating the TME of OS.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-262/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-262/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-262/prf

Funding: This work was supported by the Baise City Financial Science and Technology Plan (No. 20211816).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-262/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was reviewed by the Ethics Committee of the Baise People’s Hospital (Ethics No. 2024-E488-01). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The requirement of informed consent was waived due to retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Kansara M, Teng MW, Smyth MJ, et al. Translational biology of osteosarcoma. Nat Rev Cancer 2014;14:722-35. [Crossref] [PubMed]
2. Ottaviani G, Jaffe N. The epidemiology of osteosarcoma. Cancer Treat Res 2009;152:3-13. [Crossref] [PubMed]
3. Gibson TM, Mostoufi-Moab S, Stratton KL, et al. Temporal patterns in the risk of chronic health conditions in survivors of childhood cancer diagnosed 1970-99: a report from the Childhood Cancer Survivor Study cohort. Lancet Oncol 2018;19:1590-601. [Crossref] [PubMed]
4. Chou AJ, Gorlick R. Chemotherapy resistance in osteosarcoma: current challenges and future directions. Expert Rev Anticancer Ther 2006;6:1075-85. [Crossref] [PubMed]
5. Link MP, Goorin AM, Miser AW, et al. The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. N Engl J Med 1986;314:1600-6. [Crossref] [PubMed]
6. Harrison DJ, Geller DS, Gill JD, et al. Current and future therapeutic approaches for osteosarcoma. Expert Rev Anticancer Ther 2018;18:39-50. [Crossref] [PubMed]
7. Liang J, Chen J, Hua S, et al. Bioinformatics analysis of the key genes in osteosarcoma metastasis and immune invasion. Transl Pediatr 2022;11:1656-70. [Crossref] [PubMed]
8. Gianferante DM, Mirabello L, Savage SA. Germline and somatic genetics of osteosarcoma - connecting aetiology, biology and therapy. Nat Rev Endocrinol 2017;13:480-91. [Crossref] [PubMed]
9. Chen X, Bahrami A, Pappo A, et al. Recurrent somatic structural variations contribute to tumorigenesis in pediatric osteosarcoma. Cell Rep 2014;7:104-12. [Crossref] [PubMed]
10. Sadoughi F, Maleki Dana P, Asemi Z, et al. DNA damage response and repair in osteosarcoma: Defects, regulation and therapeutic implications. DNA Repair (Amst) 2021;102:103105. [Crossref] [PubMed]
11. Oshima M. Asahi Nuclear Magnetic Resonance-CT. Nihon Rinsho 1983;41:1636-8.
12. Paget S. The distribution of secondary growths in cancer of the breast. 1889. Cancer Metastasis Rev 1989;8:98-101.
13. Alfranca A, Martinez-Cruzado L, Tornin J, et al. Bone microenvironment signals in osteosarcoma development. Cell Mol Life Sci 2015;72:3097-113. [Crossref] [PubMed]
14. Friebele JC, Peck J, Pan X, et al. Osteosarcoma: A Meta-Analysis and Review of the Literature. Am J Orthop (Belle Mead NJ) 2015;44:547-53.
15. Chen C, Xie L, Ren T, et al. Immunotherapy for osteosarcoma: Fundamental mechanism, rationale, and recent breakthroughs. Cancer Lett 2021;500:1-10. [Crossref] [PubMed]
16. Yu S, Yao X. Advances on immunotherapy for osteosarcoma. Mol Cancer 2024;23:192. [Crossref] [PubMed]
17. Hu J, Lazar AJ, Ingram D, et al. Cell membrane-anchored and tumor-targeted IL-12 T-cell therapy destroys cancer-associated fibroblasts and disrupts extracellular matrix in heterogenous osteosarcoma xenograft models. J Immunother Cancer 2024;12:e006991. [Crossref] [PubMed]
18. Irac SE, Oksa A, Jackson K, et al. Cytokine Expression in Canine Lymphoma, Osteosarcoma, Mammary Gland Tumour and Melanoma: Comparative Aspects. Vet Sci 2019;6:37. [Crossref] [PubMed]
19. Lokau J, Schoeder V, Garbers C. he role of interleukin-11 in osteosarcoma. Pathologe 2020;41:163-7. [Crossref] [PubMed]
20. Xu Y, Deng C, Chen H, et al. Osteosarcoma Cells Secrete CXCL14 That Activates Integrin α11β1 on Fibroblasts to Form a Lung Metastatic Niche. Cancer Res 2024;84:994-1012. [Crossref] [PubMed]
21. Yoshimura N, Kariya R, Shimada M, et al. The IL-17-IL-17RA axis is required to promote osteosarcoma progression in mice. Sci Rep 2023;13:21572. [Crossref] [PubMed]
22. Cui J, Dean D, Hornicek FJ, et al. The role of extracelluar matrix in osteosarcoma progression and metastasis. J Exp Clin Cancer Res 2020;39:178. [Crossref] [PubMed]
23. Yang Q, Liu J, Wu B, et al. Role of extracellular vesicles in osteosarcoma. Int J Med Sci 2022;19:1216-26. [Crossref] [PubMed]
24. Sagheb IS, Coonan TP, Randall RL, et al. Extracellular matrix production and oxygen diffusion regulate chemotherapeutic response in osteosarcoma spheroids. Cancer Med 2024;13:e70239. [Crossref] [PubMed]
25. Huang B, Jin P, Yi K, et al. MAPK-interacting kinases inhibition by eFT508 overcomes chemoresistance in preclinical model of osteosarcoma. Hum Exp Toxicol 2023;42:9603271231158047. [Crossref] [PubMed]
26. Kiany S, Huang G, Kleinerman ES. Effect of entinostat on NK cell-mediated cytotoxicity against osteosarcoma cells and osteosarcoma lung metastasis. Oncoimmunology 2017;6:e1333214. [Crossref] [PubMed]
27. Li A, Wang S, Nie J, et al. USP3 promotes osteosarcoma progression via deubiquitinating EPHA2 and activating the PI3K/AKT signaling pathway. Cell Death Dis 2024;15:235. [Crossref] [PubMed]
28. Tullius BP, Setty BA, Lee DA. Natural Killer Cell Immunotherapy for Osteosarcoma. Adv Exp Med Biol 2020;1257:141-54. [Crossref] [PubMed]
29. Wei Z, Xia K, Zheng D, et al. RILP inhibits tumor progression in osteosarcoma via Grb10-mediated inhibition of the PI3K/AKT/mTOR pathway. Mol Med 2023;29:133. [Crossref] [PubMed]
30. Zhang J, Yu XH, Yan YG, et al. PI3K/Akt signaling in osteosarcoma. Clin Chim Acta 2015;444:182-92. [Crossref] [PubMed]
31. Zheng GZ, Zhang QH, Chang B, et al. Dioscin induces osteosarcoma cell apoptosis by upregulating ROS-mediated P38 MAPK signaling. Drug Dev Res 2023;84:25-35. [Crossref] [PubMed]
32. Yu H, Wang C, Ke S, et al. An integrative pan-cancer analysis of MASP1 and the potential clinical implications for the tumor immune microenvironment. Int J Biol Macromol 2024;280:135834. [Crossref] [PubMed]
33. Zhang CH, Xu ZQ, Peng LL, et al. Comprehensive analysis of the immune and prognostic implication of MASP1 in stomach adenocarcinoma. Eur Rev Med Pharmacol Sci 2022;26:6725-41. [Crossref] [PubMed]
34. Cui Z, Tang J, Chen J, et al. Hsa-miR-574-5p negatively regulates MACC-1 expression to suppress colorectal cancer liver metastasis. Cancer Cell Int 2014;14:47. [Crossref] [PubMed]
35. Gong A, Luo X, Tan Y, et al. High expression of C10orf91 and LINC01224 in hepatocellular carcinoma and poor prognosis. Am J Transl Res 2022;14:2567-79.
36. Li F, Tang H, Luo X, et al. Interaction gene set between osteoclasts and regulatory CD4(+) T cells can accurately predict the prognosis of patients with osteosarcoma. Cancer Sci 2023;114:3014-26. [Crossref] [PubMed]

(English Language Editor: L. Huleatt)