RNA expression profiles from normal tissues were sourced from the Genotype-Tissue Expression (GTEx) database. The fragments per kilobase of transcript per million mapped reads (FPKM) values in the expression dataset were transformed using the log2(x+0.001) method to facilitate analysis. Furthermore, RNA sequencing data, along with clinical characteristics of osteosarcoma patients, were obtained from the TARGET database. Afterward, the TARGET and GTEx datasets were combined, and batch effects were addressed using the sva package. The FPKM values were utilized as indicators of gene expression levels. In addition this study was approved by the Ethics Committee of the Second Affiliated Hospital of Nanchang University. Samples from 32 patients diagnosed with OS were collected in the Second Affiliated Hospital of Nanchang University.

A comparative analysis was conducted between the 40 T-cell exhaustion-related genes (TEXRGs) and the genes identified within the osteosarcoma transcriptome dataset to isolate TEXRGs associated with osteosarcoma. 40 T-cell exhaustion-associated genes were obtained from the article by Zhang Z et al. (34). The differential expression analysis was carried out using the limma package in R, applied to the TARGET-OS and GTEx datasets, to identify differentially expressed genes (DEGs). The criteria for selection included an FDR threshold of less than 0.05 and an absolute log₂ fold change (|log₂FC|) greater than 2. Subsequently, an intersection of TEXRGs and DEGs was performed to identify the differentially expressed TEXRGs relevant to osteosarcoma.

In this study, differentially expressed TEXRGs were analyzed by Kyoto Encyclopedia of Genomes (KEGG) and GO enrichment analysis using the R software packages clusterProfiler, org.Hs.eg.db, enrichplot, GOplot, and ggplot2 with a significance threshold of p < 0.05 (34–39).

The limma package in R was employed to perform a co-expression analysis of osteosarcoma-associated differentially expressed TEXRGs and lncRNAs (TEXRLs) within the osteosarcoma transcriptomic dataset. This analysis aimed to identify TEXRLs relevant to osteosarcoma, utilizing screening criteria of |Pearson correlation coefficient| > 0.4 and p < 0.001.

To extract TEXRLs significantly associated with osteosarcoma prognosis, the survival package in R was applied, conducting univariate COX regression analysis with a significance threshold of p < 0.05, thereby calculating the hazard ratio (HR) values. Significant variables (p < 0.05) were then analyzed by the least absolute shrinkage and selection operator (LASSO) using the ‘glmnet’ software package. Candidate genes were then identified based on the optimal penalty parameter λ determined by the 1-SE (standard error) criterion (40–43), which helped reduce the risk of overfitting while determining the optimal number of TEXRLs for inclusion in the model development. The dataset was divided into training and testing cohorts to assess the model’s accuracy. A risk prognosis model was constructed for the entire dataset, as well as separately for the training and testing groups. The expression levels of the osteosarcoma prognosis TEXRLs were multiplied by their respective regression coefficients and summed to calculate the sample risk score. Subsequently, the samples from the overall dataset, training, and testing groups were stratified into high- and low-risk categories based on the median risk score. Riskscore =(-2.300*AC090559.2)+(-2.574*AL031775.1)+(2.146*LINC01060)+(-6.282*LINC02777)+(-2.543*PSMB8-AS1)+(-1.268*AC135178.4).

The analysis of risk curves, survival outcomes, receiver operating characteristic (ROC) curves, and independent prognostic evaluations was performed on the risk prognostic models across the entire sample, as well as the training and testing cohorts. The statistical software R was employed to produce both the survival status map and the risk heatmap associated with the prognostic model, facilitating the assessment of differences in patient survival times and overall survival prognoses across high- and low-risk groups. To construct survival curves, the survival and survminer packages in R were utilized, while the survival, survminer, and timeROC packages were applied to generate ROC curves. Furthermore, the survival package in R was leveraged to carry out independent prognostic evaluations through univariate and multivariate Cox regression analyses, aiming to ascertain whether the risk score could function as an independent prognostic factor (44, 45).

The analysis of the tumor microenvironment concerning the osteosarcoma transcriptome data was performed using the limma, CIBERSORT, and estimate packages in R, which generated immune scores, stromal scores, and comprehensive scores for each patient diagnosed with osteosarcoma. Furthermore, the limma and ggpubr packages in R were applied to evaluate the variations in immune, stromal, and overall scores across the risk prognostic model within the entire cohort of samples.

This study utilized the GSVA, limma, and GSEABase packages in R to calculate enrichment scores pertaining to immune cell types and immune functionality based on the osteosarcoma transcriptome data. Additionally, the limma, reshape2, and ggpubr packages in R were employed to investigate the differences in immune cell populations and immune function across the risk prognosis model applied to the entire sample cohort.

Single-cell RNA sequencing (scRNA-seq) data were taken from the GSE162454 dataset in the GEO database, containing a total of 6 osteosarcoma samples. Standardized data were pre-processed by the ‘seurat’ software package (version 4.0). Strict quality control measures were used to exclude cells with less than 300 or more than 2000 genes detected. To mitigate batch effects, the ‘CAA’ software package was used. After filtering, a total of 46,544 cells were available for subsequent analyses. Subsequently, primary cell cluster analysis was performed using the FindClusters function of the Seurat package (resolution = 0.15), and the visual clustering results were presented through performing uniform manifold approximation and projection (UMAP) dimension reduction analysis (46). For other specific parameters, please refer to the TISCH2 database (http://tisch.comp-genomics.org/).

Overexpression plasmids for h-AL031775.1 were constructed by Hans Group Holdings Co., Ltd.h-AL031775.1 siRNA (CCCTTTACATTTCCCACTT) and NC siRNA (AAGTCGGGTC

AAGAGAAGC) were purchased from Guangzhou RiboBio Co., Ltd.

Human osteoblast cell line hFOB 1.19 from Procell Life Co., Ltd. was cultured at 33.5°C~34°C, and the medium used was special medium for osteoblasts from Procell Life Co (Wuhan, China). The osteosarcoma cell lines MG63, U2OS, and 143B were obtained from the National Collection of Authenticated Cell Cultures (Shanghai, China). Osteosarcoma cells were cultured in complete DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂. Osteoblast cell lines were maintained in the same medium but cultured at 34°C under identical atmospheric conditions.

Total cellular RNA was isolated using the Trizol method and then reverse transcribed into complementary DNA (cDNA) with the aid of a TAKARA commercial kit (Cat. No. RR047A). This cDNA served as the template for real-time quantitative polymerase chain reaction (RT-qPCR), utilizing TAKARA’s qPCR mix (Cat. No. RR420A). Gene expression levels were quantified using the 2^(-ΔΔCt) comparative method for relative quantification. Primer sequences are shown in Supplementary Table S1 . Primer sequences are synthesized by Guangzhou RiboBio Co., Ltd. (China).

The proliferative potential of osteosarcoma cells was evaluated through EdU incorporation and CCK-8 assays. For the EdU assay, cells were seeded into 96-well plates at a density of 2×10⁴ cells per well. After 8-hour incubation at 37°C, cells were labeled, fixed, and stained in accordance with the YF^®594 Click-iT EDU staining kit protocol (UE, Shanghai, China), then imaged using a fluorescence microscope.

In the CCK-8 assay, 5000 cells per well were seeded in a 96-well plate, and at designated intervals (0 h, 24 h, 48 h, and 72 h), 10% CCK-8 solution (TransGen, Beijing, China) was added. Following a 2-hour incubation at 37°C, absorbance at 450 nm was measured via spectrophotometry to assess cell proliferation rates.

Osteosarcoma cells transfected with plasmids or siRNAs, along with their respective controls, were seeded into six-well plates. Upon reaching 90% confluency, a sterile pipette tip was used to create a scratch in the cell monolayer. Cells were then cultured in serum-free medium for 12 hours. The scratch area was photographed at 0 hours and after 12 hours using an inverted microscope to analyze cellular migration.

To evaluate migratory potential, Transwell assays were performed using 24-well chambers. Osteosarcoma cells were seeded in the upper chamber at a density of 2×10⁴ cells in 200 µL of serum-free medium, while the lower chamber was filled with 600 µL of complete medium supplemented with 20% FBS as a chemoattractant. After an incubation period of 48–72 hours, cells that had migrated through the membrane into the lower chamber were fixed in 4% paraformaldehyde and stained with crystal violet for quantification.

Apoptosis was assessed using a suspension of 10^4 cells per milliliter, seeded in a 6-well plate. Upon reaching 60-70% confluence, cells were harvested, washed with chilled phosphate-buffered saline (PBS) without Ca²⁺ and Mg²⁺, and centrifuged. Cells were then stained with Annexin V and Propidium Iodide (PI) in the dark for 15 minutes. Following the addition of 400 µL of binding buffer, apoptotic status was analyzed by flow cytometry to quantify apoptotic populations.

The Ethics Committee of Nanchang University approved all experimental procedures involving animals, and all procedures followed the ARRIVE guidelines. Male nude mice, 4 weeks of age, were obtained from Jicui Pharmachem Biotechnology Co Ltd (Jiangsu, China). The skin of nude mice was disinfected in a sterile environment, and the insulin injection needle was used to inject two to three million cells into the axilla of nude mice, and the injection site was disinfected by slowly withdrawing the needle. The health status and tumor growth of the nude mice were monitored daily after the operation.

Human peripheral blood mononuclear cells were extracted using the Human Peripheral Blood Mononuclear Cell Extraction Kit (Solarbio, Beijing). After extraction, erythrocytes were lysed on ice using Lysate Red (Tiangen, Beijing), and then activated for 48h in a 24-well plate incubated with CD3/CD28 cytokines (Thermofisher, USA) on the previous day, and then activated using 1640 complete medium containing IL-2 cytokines (Thermofisher, USA) medium containing IL-2 cytokines (Gibco, USA) was used for incubation. The optimal killing ratios of 1:1, 1:2, 1:4, and 1:8 were used for the detection of tumor cells and T cells, respectively, and the killing ability of T cells was determined by using CCK-8 to detect the OD value of tumor cells after co-cultivation in 96-well plates.

Add 5×10⁶ T cells into the flow tube, add 0.5% BSA to wash the cells, add CD3, CD4, CD8, LAG3, CTLA-4, PD1 antibody (Biolegend, Beijing), avoid light for 30min at room temperature, 0.5% BSA to wash the cells once, centrifuge and add 200ul PBS to use flow cytometry for detection.

Statistical analyses were performed using R software (version 4.2.3). To analyze the data expressing differences, we used the t-test and ANOVA. ROC curves and area under the ROC curve (AUC) were calculated using MedCalc for Windows version 19.3.0 (MedCalc Software, Ostend, Belgium). Data are presented as mean ± SD n=3. ns was considered not statistically significant, and less than 0.001 (***), 0.01 (**), and 0.05 (*) were considered statistically significant.

All analyses were performed using Flowjo, GraphPad Software, and Xiantao Academic.

Analysis of the TARGET dataset, which included 88 osteosarcoma patients, and the GTEx dataset, comprising 90 control samples, identified 9,981 differentially expressed genes (DEGs), including 4,979 upregulated and 5,002 downregulated genes. These findings were visualized using heatmaps generated with R programming ( Figure 1A ).

To further investigate T-cell exhaustion in osteosarcoma, we referred to a previously published study (34), which defined a T-cell exhaustion-related gene set consisting of 40 genes. By intersecting this gene set with the osteosarcoma-specific DEGs, we identified 24 T-cell exhaustion-related genes (TEXRGs) associated with osteosarcoma, including 18 upregulated and 6 downregulated genes ( Figures 1B, C ).

Notably, 11 of these TEXRGs were significantly enriched in pathways associated with the regulation of adaptive immune responses, particularly those involving the somatic recombination of immune receptors from immunoglobulin superfamily domains and lymphocyte-mediated immunity. KEGG and GO analysis further revealed significant enrichment in the T cell receptor binding, JAK-STAT signaling pathway and PD-L1 expression and PD-1 checkpoint pathway in cancer, highlighting the potential role of these 11 TEXRGs in tumor regulation ( Figures 1D, E ; Supplementary Figure 5C ).

Co-expression analysis of these 24 TEXRGs identified 549 lncRNAs associated with T-cell exhaustion. Through Cox regression analysis, we determined that 37 of these lncRNAs were significantly correlated with patient prognosis. These included 12 high-risk T-cell exhaustion-related lncRNAs (AC012442.2, AC083843.3, GTF2IP1, LINC01517, AC004862.1, LINC02167, OLMALINC, LINC00837, LINC01060, RP11-69E11.4, SATB2-AS1, and AC005100.1) and 25 low-risk T-cell exhaustion-related lncRNAs (LINC02777, AC114956.1, AL031775.1, NCF1B, EDIL3-DT, AC110995.1, AC009054.2, POLR2J4, AL359091.3, Z69706.1, NBR2, CARD8-AS1, AJ011932.1, SRP14-AS1, AC090559.1, KANSL1-AS1, AL161785.1, SMIM25, GAPLINC, AC010542.6, AC135178.4, PSMB8-AS1, HAS2-AS1, AC090152.1, and MSC-AS1) ( Figure 1F ; Supplementary Figure 5A ).

Using LASSO regression analysis, we further identified six key T-cell exhaustion-related lncRNAs as the optimal feature set ( Figure 1G ). Importantly, these six lncRNAs (LINC01060, AC090559.1, AC135178.4, AL031775.1, LINC02777, and PSMB8-AS1) successfully constructed a robust risk prediction model for osteosarcoma ( Figure 2G ). Using the prognostic model formula, we calculated risk scores for each sample, stratifying the cohort into red high-risk (N = 42) and blue low-risk (N = 46) groups. The training cohort was further divided into red high-risk (N = 22) and blue low-risk (N = 22) groups, while the testing cohort consisted of red high-risk (N = 20) and blue low-risk (N = 24) groups.

Survival status maps for the entire cohort, as well as the training and testing groups, indicated a marked increase in mortality rates as patients transitioned from blue low-risk to red high-risk categories ( Figures 1H , 2A, H ). Risk heatmaps revealed that the expression levels of the high-risk TEXRL LINC01060 consistently increased from low-risk to red high-risk groups, while blue low-risk TEXRLs, including AC090559.1, AC135178.4, AL031775.1, LINC02777, and PSMB8-AS1, exhibited a progressive decline ( Figures 1I , 2B, I ).

Survival analyses demonstrated significant differences in outcomes between high-risk and low-risk patients for the complete sample cohort [p < 0.0001, HR=0.140(0.065-0.299)], training cohort [p < 0.0001, HR=0.074(0.024-0.228)], and testing cohort [p =0.0054, HR=0.227(0.080-0.646)] ( Figures 1J , 2C, J ). There was also a significant difference in prognosis between high- and low-risk patients in our own cohort sample ( Supplementary Figure 5B ). The receiver operating characteristic (ROC) curve for the entire sample cohort showed areas under the curve (AUC) of 0.821 at 1 year, 0.861 at 3 years, and 0.814 at 5 years ( Figure 1K ). The training cohort’s ROC curve yielded higher AUCs: 0.966 at 1 year, 0.993 at 3 years, and 0.994 at 5 years ( Figure 2D ). The testing cohort demonstrated lower AUCs: 0.667 at 1 year, 0.741 at 3 years, and 0.694 at 5 years ( Figure 2K ). Additionally, the ROC curves for metastasis (AUC = 0.905) and risk score (AUC = 0.821) in the complete cohort indicated strong predictive capacity ( Figure 1L ). In the training cohort, metastasis (AUC = 0.856) and risk score (AUC = 0.966) also showed high predictive values ( Figure 2E ). The test cohort exhibited AUCs of 0.956 for metastasis, 0.667 for risk score, and 0.603 for age ( Figure 2L ).

Univariate independent prognostic analysis revealed that both the risk score (p < 0.001, HR = 1.000) and tumor metastasis (p < 0.001, HR = 4.770) are independent prognostic factors ( Figure 1M ). Multivariate analysis confirmed the risk score (p = 0.006, HR = 1.000) and tumor metastasis (p < 0.001, HR = 4.660) as independent prognostic indicators ( Figure 1N ). In the training cohort, univariate analysis identified the risk score (p = 0.001, HR = 1.000) and tumor metastasis (p = 0.012, HR = 3.964) as significant prognostic determinants ( Figure 2F ). Multivariate analysis further established the risk score (p = 0.025, HR = 1.000) as an independent prognostic factor ( Figure 2G ). In the test cohort, univariate analysis indicated that tumor metastasis (p < 0.001, HR = 8.395) serves as an independent prognostic factor ( Figure 2M ). Moreover, multivariate analysis confirmed tumor metastasis (p < 0.001, HR = 10.075) as a significant independent prognostic indicator ( Figure 2N ). Collectively, our risk prognostic model suggests that both the risk score and tumor metastasis may serve as independent high-risk prognostic factors for osteosarcoma patients.

CIBERSORT analysis of immune cell populations revealed a significant increase in naive B cells, naive CD4 T cells, and macrophages in the high-risk cohort ( Figure 3A ). Conversely, activated memory CD4 T cells, monocytes, and M2 macrophages were notably decreased in high-risk individuals ( Figure 3A ). Additionally, ssGSEA analysis indicated a marked downregulation of various immune cell types in the high-risk group, including activated dendritic cells (aDC), cytotoxic cells, dendritic cells (DC), immature dendritic cells (iDC), macrophages, mast cells, neutrophils, NK CD56dim cells, plasmacytoid dendritic cells (pDC), T cells, T follicular helper cells (TFH), Th1 cells, and regulatory T cells (T-Reg) ( Figure 3B ). B cells were significantly upregulated in this cohort ( Figure 3B ).

Comparative analysis of immune functionality showed significant downregulation of several factors in the high-risk cohort, including antigen-presenting cell (APC) co-inhibition, APC co-stimulation, chemokine receptors (CCR), immune checkpoint molecules, cytolytic activity, human leukocyte antigen (HLA), inflammation-promoting agents, major histocompatibility complex (MHC) class I, para-inflammation, T cell co-inhibition, T cell co-stimulation, and Type I interferon (IFN) response ( Figure 3C ).

Examination of the tumor microenvironment revealed notable differences in stromal cell scores (p < 0.001) ( Figure 3D ), immune cell scores (p < 0.001) ( Figure 3E ), and ESTIMATE scores (p < 0.001) ( Figure 3F ) when comparing red high-risk and blue low-risk groups, with the low-risk group exhibiting higher scores across these metrics ( Figures 3D–F ).

Kaplan-Meier survival analysis for AC090559.1, AL031775.1, LINC01060, and LINC02777 indicated significant differences in survival outcomes between patients with blue high and red low expression levels of these lncRNAs ( Figures 4A–F ).

The ROC curve for AC090559.1 demonstrated a commendable AUC of 0.802 at 1 year, 0.693 at 3 years, and 0.607 at 5 years ( Figure 4G ). In contrast, AC136178.4 displayed lower AUC values: 0.680 at 1 year, 0.593 at 3 years, and 0.579 at 5 years ( Figure 4H ). AL031775.1 showed a better performance with AUCs of 0.671 at 1 year, 0.735 at 3 years, and 0.712 at 5 years ( Figure 4I ). LINC01060 also revealed notable AUC values of 0.522 at 1 year, 0.681 at 3 years, and 0.678 at 5 years ( Figure 4J ). LINC02777 exhibited favorable AUCs of 0.676 at 1 year, 0.709 at 3 years, and 0.663 at 5 years ( Figure 4K ). Lastly, PSMB8-AS1 indicated comparatively lower AUCs of 0.698 at 1 year, 0.655 at 3 years, and 0.521 at 5 years ( Figure 4L ).

Differential gene expression of six tumor-expressed long non-coding RNAs (TEXRLs) was assessed in 32 osteosarcoma tissue samples, with corresponding paracancerous tissues serving as controls. The analysis revealed that LINC01060, AL031775.1, LINC02777, and PSMB8-AS1 were significantly downregulated in osteosarcoma tissues compared to the control group, while AC090559.1 was significantly upregulated, and AC135178.4 showed no statistically significant difference ( Supplementary Figure 1A ). Of the four long non-coding RNAs exhibiting downregulation (AL031775.1, LINC01060, LINC02777, and PSMB8-AS1), AL031775.1 demonstrated the most significant statistical difference in RT-qPCR results compared to the control group. Additionally, survival analysis and AUC curve results further supported the superior predictive performance of AL031775.1 over LINC01060, LINC02777, and PSMB8-AS1 ( Figure 4 ). Therefore, we selected AL031775.1 as the candidate gene for further investigation.

The expression of AL031775.1 was evaluated in three osteosarcoma cell lines, with the hFOB 1.19 cell line used as a normal control. This analysis confirmed significant downregulation of AL031775.1 in the MG63, U2OS, and 143B cell lines ( Supplementary Figure 1B ). Subsequently, the expression of AL031775.1 was inhibited using siRNA transfection ( Figures 5A, C ), and we assessed its effects on osteosarcoma cell proliferation, migration, and apoptosis. CCK-8 assay and EdU staining in the MG63 osteosarcoma cell line showed that the reduction of AL031775.1 expression significantly promoted the proliferation of osteosarcoma cells ( Figures 5B, E ), and the same results were observed in the U2OS osteosarcoma cell line ( Figures 5D, G ), Transwell assay showed that down-regulation of AL031775.1 contributed to the enhanced migratory capacity of the MG63 osteosarcoma cell line ( Figure 5F ) and the U2OS osteosarcoma cell line ( Figure 5H ), which was confirmed by a wound healing assay showing enhanced migratory activity of both the MG63 and the U2OS osteosarcoma cell lines ( Figures 5I, J ). Annexin V/PI staining analysis showed that the low expression of AL031775.1 significantly reduced apoptosis in MG63 and U2OS osteosarcoma cells ( Figures 5K, L ).

Overexpression vector plasmid transfection was utilized to promote AL031775.1 expression in MG63 and U2OS osteosarcoma cell lines ( Figures 6A, C ), and we subsequently assessed its effects on osteosarcoma cell proliferation, migration, and apoptosis. CCK-8 assay and EdU staining in the MG63 osteosarcoma cell line showed that increased AL031775.1 expression significantly inhibited the proliferation of osteosarcoma cells ( Figures 6B, E ), and the same results were observed in the U2OS osteosarcoma cell line ( Figures 6D, G ). Transwell assay showed that upregulation of AL031775.1 inhibited the migratory ability of the MG63 osteosarcoma cell line ( Figure 6F ) and the U2OS osteosarcoma cell line ( Figure 6H ), which was confirmed by wound healing assay showing decreased migratory activity of the MG63 and U2OS osteosarcoma cell lines(( Figures 6I, J ). Annexin V/PI staining analysis showed that the low expression of AL031775.1 significantly increased apoptosis in MG63 and U2OS osteosarcoma cells ( Figures 6K, L ) Similarly, a decrease in cell proliferation and migration was observed after overexpression of AL031775.1 in the 143b osteosarcoma cell line, along with an increase in apoptosis ( Figures 7A–F ).

To investigate the role of AL031775.1 in osteosarcoma progression, we employed a subcutaneous transplantation model in nude mice using the 143B osteosarcoma cell line. A virus-encapsulated AL031775.1 overexpression plasmid and an empty vector control plasmid were used to establish stable cell transplants. Following transduction, twelve nude mice were divided into two groups for subcutaneous tumor formation. Tumor growth differences between the groups became apparent after one week, and real-time growth curves were recorded ( Figure 7I ). Tumors in the AL031775.1 overexpression group exhibited significant growth inhibition, with reduced fluorescence signals observed in vivo after 24 days, indicating diminished tumor activity ( Figure 7G, H ). Additionally, RT-qPCR analysis of tumor tissues confirmed successful overexpression of AL031775.1 ( Figure 7J ).

In order to gain a comprehensive understanding of the distribution of model genes in the osteosarcoma tumour microenvironment (TME), we performed an in-depth analysis of scRNA-seq data obtained from osteosarcoma patients. After stringent quality control measures, we identified a total of 46,544 cells in the osteosarcoma samples, laying the foundation for subsequent analyses. By evaluating the expression of characteristic genes, we delineated 8 major clusters in the osteosarcoma TME ( Supplementary Figures S4A, B ). The resulting t-SNE plots showed the annotations of these 8 different cell clusters, which included various cell types such as CD4Tconv, CD8Tex, Endothelial, Fibroblasts, Malignant, Mono/Macro, Osteoblasts, Plasmocytes, as Supplementary Figure S3B shown. In addition, the authors analyzed the percentage of different cell types among all cells and the proportion of cell type distribution in different patient samples ( Supplementary Figures S4C, D ). Correspondingly, UMAP plots and violin plots ( Supplementary Figures S4E, F ) highlight that our risk gene AL031775.1 is highly expressed in a wide range of immune cells from CD4Tconv, CD8Tex, Mono/Macro and Plasmocytes.

After isolating human peripheral blood T lymphocytes, T cells were extracted and co-cultured with 143B osteosarcoma cells at varying ratios (143B cell ratios of 1:1, 1:2, 1:4, and 1:8). Following 48 hours of co-culture, the optical density (OD) values of the residual tumor cells were measured using the CCK-8 assay to assess T cell cytotoxicity. The optimal killing efficiency was observed at a 1:4 ratio ( Supplementary Figure 2 ). Downregulation of AL031775.1 in T cells resulted in a significant decrease in T cell proliferation, an increase in apoptosis, and a reduced capacity for tumor cell lysis, accompanied by elevated expression of immune checkpoints LAG3, CTLA4, and PD1 ( Figures 8A–C, G, H ). Conversely, upregulation of AL031775.1 in T cells leads to opposite effects ( Figures 8D–F, I, J ).