System for evaluating clinical prognosis of low-grade glioma based on expression of multiple glycosylation genes and application of system

Abstract

The invention belongs to the technical field of low-grade glioma, and provides a system for evaluating clinical prognosis of low-grade glioma based on expression of a plurality of glycosylation genes and application of the system. The plurality of glycosylation genes comprise ASPM (Aspartate Sequence Protein Matrix), CHI3L1 (CHI3L1), LILRA4 (Lymphocyte Leukemia Region Amplification 4), MSN (Matrix Sequence Natrix), OCIA (Oligomeric Cyclovibrio The risk score feature model for evaluating the clinical prognosis of the low-grade glioma is RiskScore = (0.382 * A) + (0.017 * C) + (0.068 * L) + (0.258 * M) + (0.116 * O) + (-0.137 * P) + (0.109 * S) + (0.073 * T); in the formula, RiskScore is a risk score, and A, C, L, M, O, P, S and T are expression quantities of the glycosylation genes respectively; a new view angle is provided for clarification of a potential mechanism of LGG prognosis, and a direction is provided for future individualized cancer immunotherapy.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-grade gliomas, and particularly relates to a system for evaluating the clinical prognosis of low-grade gliomas based on the expression of multiple glycosylation genes and its application. Background Art

[0002] The treatment of low-grade glioma (LGG) has always been a hot topic in the treatment of intracranial tumors. Glycosylation is a common protein post-translational modification that plays an important role in tumor transformation. Many studies have shown that there is a close association between glycosylation modification and tumor progression. However, the expression status of glycosylation-related genes in LGG and their potential role in the LGG microenvironment have not been widely studied.

[0003] Gliomas are the most common tumors in the human central nervous system (CNS)[1] and are recognized as a cancerous disease that begins in the CNS. They are known for their rapid growth and aggressive tendencies[2]. WHO grade II gliomas, such as diffuse astrocytomas and oligodendrogliomas, are classified as low-grade gliomas[3,4]. These tumors usually have a long and indolent course. However, studies have shown that LGGs may undergo malignant differentiation, leading to severe disability and death[5], thereby reducing the patient's quality of life and chances of survival[6].

[0004] Although conventional surgical resection, radiotherapy combined with immunotherapy, and novel electric field therapy have improved the prognosis of glioma patients, the overall prognosis is still poor due to the heterogeneity of LGG and epigenetic mutations of molecules within the tumor. Therefore, there is an urgent need to fully understand the molecular mechanisms of glioma occurrence and development, discover new biomarkers, and improve and validate innovative predictive indicators to accurately assess the prognosis of LGG patients [7].

[0005] Glycosylation is an enzymatic process that attaches sugars to proteins, lipids, and other glycans. This major post-translational modification (PTM) occurs in the endoplasmic reticulum and Golgi compartments of all cells and is mediated by the coordinated action of different glycosyltransferases and glycosidases [8]. Glycosylation-mediated PTMs play a crucial role in regulating fundamental processes such as cell division, differentiation, immune responses, and cell-cell interactions. Alterations in the N-linked or O-linked glycosylation patterns of regulatory proteins, such as transcription factors or cell receptors, can lead to a variety of diseases, including cancer. These alterations contribute to the micro- and macro-heterogeneity of tumor cells [9].

[0006] Studies have shown that glycosylation-related genes are closely associated with the prognosis of patients with breast cancer

[10] , ovarian cancer

[11] , liver cancer

[12] , cervical cancer

[13] , and pancreatic cancer

[14] . Compared with non-malignant cells, tumor cells often show abnormal glycosylation patterns. Abnormally expressed glycosylation-related genes have been shown to be effective biomarkers for a variety of tumors

[15] .

[0007] Therefore, in-depth study of the pathogenesis of glycosylation in LGG and identification of effective biomarkers are crucial to optimize the diagnosis, prevention, and management of LGG. Summary of the invention

[0008] The present invention provides a system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylation genes and its application.

[0009] The present invention is realized by the following technical scheme: a system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylation genes, wherein the multiple glycosylation genes include: ASPM, CHI3L1, LILRA4, MSN, OCIA, PTGER4, SERPING1 and TNFRSF12A; the risk score characteristic model for evaluating the clinical prognosis of low-grade glioma is: RiskScore=(0.382×A)+(0.017×C)+(0.068×L)+(0.258×M)+(0.116×O)+ (-0.137×P)+(0.109×S)+(0.073×T); where RiskScore is the risk score, A is the expression level of ASPM; C is the expression level of CHI3L1; L is the expression level of ILRA4; M is the expression level of MSN; O is the expression level of OCIA; P is the expression level of PTGER4; S is the expression level of SERPING1; T is the expression level of TNFRSF12A; where RiskScore is greater than 2.665581 and is considered high-risk, while RiskScore less than 2.665581 is considered low-risk.

[0010] Compared with normal tissues, the ASPM, LILRA4, MSN, PTGER4, and TNFRSF12A were significantly upregulated in low-grade glioma LGG, while the expressions of CHI3L1, OCIAD2, and SERPING1 were reduced; in astrocytes, the expressions of ASPM, CHI3L1, and TNFRSF12A were higher than those in other cell populations; in microglial cell populations, the expressions of LILRA4, MSN, PTGER4, and SERPING1 were higher than those in other cell populations; and the expression of OCIAD2 in T cells was higher than that in other cell populations.

[0011] Analysis of the TCGA-LGG cohort combined with GTEx normal brain tissue data revealed significant expression differences in prognostic signature genes. ASPM expression was positively correlated with astrocyte proportion (cor=0.61, p<0.05), OCIAD2 expression was negatively correlated with oligodendrocyte proportion (cor=-0.59, p<0.05), and SERPING1 expression was positively correlated with T cell proportion (cor=0.57, p<0.05).

[0012] The present invention also provides the use of the system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes in the preparation of a reagent or system for evaluating the clinical prognosis of low-grade glioma.

[0013] The present invention also provides the use of the system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylation genes in the preparation of a reagent or system for predicting or evaluating the potential effectiveness of sensitivity to chemotherapeutic drugs.

[0014] Furthermore, the system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylation genes is used in the preparation of reagents or systems for identifying the short-term effectiveness of chemotherapy drugs in patients.

[0015] This invention constructs and validates a glycosylation-related risk signature based on eight glycosylation-related genes (ASPM, CHI3L1, LILRA4, MSN, OCIAD2, PTGER4, SERPING1, and TNFRSF12A) as a novel prognostic biomarker for LGG patients, and may predict ICB immunotherapy response. In addition, we also explored the relationship between risk signatures and immune signatures and drug sensitivity. The predictive value, immunotherapy effect, and chemotherapy effect of the prognostic model for LGG patients were evaluated, providing a basis for the precise treatment of LGG patients. The glycosylation-related risk signature constructed and validated in this invention for predicting OS in LGG patients is associated with changes in the LGG tumor immune microenvironment and may predict immunotherapy response. Our study provides a novel and comprehensive perspective for clarifying the potential mechanisms of LGG prognosis and provides a direction for future personalized cancer immunotherapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1Glycosylation pathway analysis of single-cell RNA sequencing of glioma and its biological significance; Figure: A: UMAP diagram shows the distribution of four major cell types in the comprehensive data set; B: Dot plot of marker gene expression levels in each cell type; C: Expression and distribution of N-glycosylation and O-glycosylation pathway activity scores in cell types; D: Volcano map of differentially expressed genes (DEGs) between microglia and other subpopulations (|logFC|>0.585, P.adj<0.05); E: Gene Ontology (GO) analysis (with Log10 (p-adjust) as the horizontal coordinate); F: KEGG analysis (with Log10 (p-adjust) as the horizontal coordinate); Figure 2 Distribution of markers and glycosylation scores in single-cell RNA sequencing of gliomas; Figure: A. Heat map showing the expression levels of marker genes in oligodendrocytes, T cells, astrocytes, and microglia; B. UMAP diagram showing the distribution of N-glycosylation and O-glycosylation pathway activity scores in cell types using different scoring methods (AUCell, UCell, and SSGSEA); Figure 3 Analysis of N-glycosylation pathway in TCGA-LGG cohort and its biological significance; Figure: A: Kaplan-Meier curve analysis of overall survival difference between the two groups; B: Volcano plot of differentially expressed genes (DEGs) between high and low N-glycosylation groups (|logFC|>0.585, P.adj<0.05); C: Gene Ontology (GO) analysis (with Log10 (p-adjust) as horizontal coordinate); D: KEGG analysis (with Log10 (p-adjust) as horizontal coordinate); E: Heat map of differences in immune scores and immune microenvironment between high and low N-glycosylation groups; Figure 4 Analysis of O-glycosylation pathway in TCGA-LGG cohort and its biological significance; Figure: A: Kaplan-Meier curve analysis of overall survival difference between the two groups; B: Volcano plot of differentially expressed genes (DEGs) between high and low O-glycosylation groups (|logFC|>0.585, P.adj<0.05); C: Gene Ontology (GO) analysis (with Log10 (p-adjust) as horizontal coordinate); D: KEGG analysis (with Log10 (p-adjust) as horizontal coordinate); E: Heat map of differences in immune scores and immune microenvironment between high and low O-glycosylation groups; Figure 5For the establishment and validation of the risk score; in the figure: A: The Venn diagram shows the differentially expressed genes (DEGs) between microglia and other cell types in the GSE117891 dataset, the differentially expressed genes in the high-low N-glycosylation groups in the TCGA-LGG cohort, and the intersection genes between the differentially expressed genes in the high-low O-glycosylation groups; B: The contribution coefficient of each component gene in the risk score. C: A heat map showing the expression levels of eight genes in different risk score groups; D: The distribution of risk scores of TCGA-LGG patients, sorted from lowest to highest. The survival status of each LGG patient was classified according to the risk score; EG: The Kaplan-Meier curves in the three cohorts (TCGA-LGG, CGGA-LGG693, CGGA-LGG325) show the differences in overall survival between different risk score groups; HJ: The ROC curve describes the predictive performance of the risk score for the 1-, 3-, and 5-year overall survival of LGG patients in the three cohorts; Figure 6 The establishment and validation of nomo; Figure: A: Univariate Cox regression analysis of clinical characteristics and risk scores; B: Multivariate Cox regression analysis of clinical characteristics and risk scores; C: Nomogram combining age, gender and risk score for 1-, 3- and 5-year survival prediction; DF: ROC curves describe the predictive performance of the nomogram for 1-, 3- and 5-year overall survival of LGG patients in three cohorts (TCGA-LGG, CGGA-LGG325, CGGA-LGG693); Figure 7 The establishment and validation of risk score and nomo; Figure: AB: After univariate COX regression analysis, LASSO regression analysis identified 8 prognostic DEGs; CE: Calibration curves of risk scores of three cohorts (TCGA-LGG, CGGA-LGG693, CGGA-LGG325), showing the accuracy of predicting survival in different cohorts; FG: Distribution of risk scores of LGG patients in CGGA-LGG693 and CGGA-LGG325 cohorts, sorted from lowest to highest, and the survival status of each LGG patient was classified according to the risk score; HJ: Calibration curves of nomo model of three cohorts (TCGA-LGG, CGGA-LGG693, CGGA-LGG325); Figure 8Differential expression of prognostic signature genes and immune microenvironment; Figure: A: TCGA combined with GTEx database shows the differential expression of 8 prognostic signature genes in normal brain tissue and LGG tissue; B: Expression and distribution of 8 prognostic signature genes in cell types in the GSE117891 single cell dataset; C. Dot plot showing the expression levels of 8 prognostic signature genes in different cell populations. D. Bar chart showing the percentage of each cell population in each sample in the GSE117891 single cell dataset. E. Heat map showing the Person correlation between cell proportions and gene expression, the numbers in the blocks represent the correlation coefficients, and the asterisks represent the p-values. F. Heat map showing the differences in immune scores and immune microenvironments between high-risk and low-risk score subgroups in the TCGA-LGG cohort; Fig. 9 Analysis of immunotherapy response and drug sensitivity based on risk score; Figure: A: TIDE score, Dysfunction score, Exclusion score and MSI in different risk groups; B: The proportion of immunotherapy responders in different risk groups in the TCGA-LGG cohort; C: The proportion of immunotherapy responders in different risk groups in the pediatric peptide vaccine immunotherapy cohort; D: The proportion of immunotherapy responders in different risk groups in the GBM PD-L1 immunotherapy cohort; E: ROC curve of risk score prediction for immune response in the pediatric peptide vaccine immunotherapy cohort; F: ROC curve of risk score prediction for immune response in the GBM PD-L1 immunotherapy cohort; GI: Sensitivity analysis of risk score and anti-tumor drugs; Fig.10 This is the survival analysis of the drug-using group and the non-drug-using group of the TCGA cohort; in the figure: A: Kaplan-Meier survival curve of all LGG patients; B: Kaplan-Meier survival curve of LGG short-term survival patients; C: Kaplan-Meier survival curve of LGG long-term survival patients; D: Kaplan-Meier survival curve of all patients in the LGG high-risk group; E: Kaplan-Meier survival curve of short-term survival patients in the LGG high-risk group; F: Kaplan-Meier survival curve of long-term survival patients in the LGG high-risk group; G: Kaplan-Meier survival curve of all patients in the LGG low-risk group; H: Kaplan-Meier survival curve of short-term survival patients in the LGG low-risk group; I: Kaplan-Meier survival curve of long-term survival patients in the LGG low-risk group; Fig.11This is the survival analysis of the medication group and the non-medication group of the CGGA325 cohort; in the figure: A: Kaplan-Meier survival curve of all LGG patients; B: Kaplan-Meier survival curve of LGG short-term survival patients; C: Kaplan-Meier survival curve of LGG long-term survival patients; D: Kaplan-Meier survival curve of all patients in the LGG high-risk group; E: Kaplan-Meier survival curve of short-term survival patients in the LGG high-risk group; F: Kaplan-Meier survival curve of long-term survival patients in the LGG high-risk group; G: Kaplan-Meier survival curve of all patients in the LGG low-risk group; H: Kaplan-Meier survival curve of long-term survival patients in the LGG low-risk group; Fig.12 This is the survival analysis of the medication group and the non-medication group of the CGGA693 cohort; in the figure: A: Kaplan-Meier survival curve of all LGG patients; B: Kaplan-Meier survival curve of LGG short-term survival patients; C: Kaplan-Meier survival curve of LGG long-term survival patients; D: Kaplan-Meier survival curve of all patients in the LGG high-risk group; E: Kaplan-Meier survival curve of short-term survival patients in the LGG high-risk group; F: Kaplan-Meier survival curve of long-term survival patients in the LGG high-risk group; G: Kaplan-Meier survival curve of all patients in the LGG low-risk group; H: Kaplan-Meier survival curve of short-term survival patients in the LGG low-risk group; I: Kaplan-Meier survival curve of long-term survival patients in the LGG low-risk group. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the field to which the present invention belongs, and the materials publicly cited herein and those cited by them will be incorporated by reference.

[0018] The equivalent techniques of the specific embodiments described that can be understood by routine experiments by those skilled in the art are all included in this application. The experimental methods in the following examples are all conventional methods unless otherwise specified. The instruments and equipment used in the following examples are all conventional laboratory instruments and equipment unless otherwise specified; the experimental materials used in the following examples are all purchased from conventional biochemical reagent stores unless otherwise specified.

[0019] 1. Experimental Methods 1. Data summary: Transcriptome data and clinical information of low-grade glioma (LGG) were obtained from The Cancer Genome Atlas (TCGA) database (https: / / portal.gdc.cancer.gov / ) using the "TCGAbiolinks"

[16] R package, including data of 504 LGG patients with survival information. At the same time, LGG validation data were obtained from the CGGA-LGG693 and CGGA-LGG325 datasets in the Chinese Glioma Genome Atlas (CGGA) database (http: / / www.cgga.org.cn), and patients classified as WHO grade II were selected, of which the CGGA-LGG693 cohort contained 172 patients and the CGGA-LGG325 cohort contained 98 patients. In addition, single-cell transcriptome data (GSE117891)

[17] were obtained from the Gene Expression Omnibus (GEO, https: / / www.ncbi.nlm.nih.gov / geo / ), covering samples from 73 regions of 13 glioma patients and 1 patient with brain metastasis. Gene expression data from 1,152 normal brain tissues were obtained from the Genotype-Tissue Expression (GTEx) project (https: / / commonfund.nih.gov / GTEx). Data from the LGG pediatric peptide vaccine immunotherapy cohort (12 patients) and the GBM (glioblastoma) anti-PD-1 immunotherapy cohort (34 patients) were obtained from the TIGER database (http: / / tiger.canceromics.org / ).

[0020] 2. Single-cell RNA sequencing analysis: The LGG single-cell RNA sequencing (scRNA-seq) data were analyzed using the R package “Seurat”

[18] . First, quality control was performed, including filtering out low-quality genes detected in less than 3 cells, removing low-quality cells with less than 300 identified genes, and excluding cells with mitochondrial gene content greater than 15%, ribosomal gene content less than 3%, and hemoglobin gene content greater than 0.1%. The MALAT1 housekeeping gene and mitochondrial genes were also filtered out, and doublets were removed. For the retained cells, the gene expression matrix was normalized using the NormalizeData function in Seurat, and centering and scaling were applied using the ScaleData function. The top 2000 highly variable genes were extracted, and then principal component analysis (PCA) was performed, and clustering was performed using the top 20 principal components. Cell populations were visualized using uniform manifold approximation and projection (UMAP). Cluster analysis was performed using the FindClusters function with a resolution of 0.3 to identify different cell clusters.

[0021] To determine the cell types, specific cell markers obtained from the official website of CellMarker 2.0 (http: / / 117.50.127.228 / CellMarker / ) were used to identify the cell types. The FindAllMarkers function in the Seurat package was used to identify differentially expressed genes between different cell populations.

[0022] 3. Single-cell glycosylation gene set enrichment score: To evaluate the enrichment of the glycosylation gene set in each cell, the irGSEA.score function in the “irGSEA“

[19] package was used. The KEGG glycosylation gene set (C2 classification), including “KEGG_N_GLYCAN_BIOSYNTHESIS” and “KEGG_O_glycan_biosynthesis”, was downloaded from the MSigDB database, and several enrichment analysis algorithms were applied to the normalized RNA sequencing data, including AUCell, UCell, and ssGSEA.

[0023] 4. Glycosylation pathway activity analysis: GSVA

[20] was used to calculate the KEGG glycosylation pathway enrichment score for each patient in the TCGA-LGG cohort. Patients were divided into high N-glycosylation group and low N-glycosylation group based on the median N-glycosylation score; similarly, patients were divided into high O-glycosylation group and low O-glycosylation group based on the median O-glycosylation score. Differential expression analysis was performed using "Deseq2"

[21] with the screening criteria of P.adjust < 0.05 and log2FC absolute value > 0.585 to identify differentially expressed genes. These genes were used to study the molecular features associated with high and low activity of the N-glycosylation and O-glycosylation pathways.

[0024] 5. Functional enrichment analysis: Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) functional enrichment analysis was performed on differentially expressed genes using the R package "clusterProfiler"

[22] . Only results with p.adjust < 0.05 were considered significant enrichment.

[0025] 6. Immune infiltration analysis: The changes in immune score, stromal score, ESTIMATE score, and tumor purity in the samples were evaluated using the ESTIMATE

[23] algorithm (Estimating stromal and immune cells in malignant tumors based on expression data). In addition, the relative abundance of different cell types was determined using the CIBERSORT

[24] and ssGSEA (single sample gene set enrichment analysis) methods using the "CIBERSORT"

[24] and "GSVA"

[20] R packages, respectively.

[0026] 7. Prognostic glycosylation-based signature construction: To establish the prognostic signature based on glycosylation, the R package "glmnet"

[25] was used to perform univariate Cox regression and least absolute shrinkage and selection operator (LASSO) regression, and 10-fold cross-validation was performed. In the LASSO regression, "lambda.min" was selected to prevent overfitting. The risk score of each LGG patient, i.e., RiskScore, was calculated using the following formula: RiskScore=(0.382×A)+(0.017×C)+(0.068×L)+(0.258×M)+(0.116×O)+(-0.137×P)+(0.109×S)+(0.073×T); where RiskScore is the risk score, A is the expression level of ASPM; C is the expression level of CHI3L1; L is the expression level of ILRA4; M is the expression level of MSN; O is the expression level of OCIA; P is the expression level of PTGER4; S is the expression level of SERPING1; and T is the expression level of TNFRSF12A.

[0027] 8. Performance evaluation and validation: Kaplan-Meier analysis was used to compare the overall survival of the high RiskScore and low RiskScore groups based on the median value. The receiver operating characteristic (ROC) curve generated by the R package "timeROC"

[26] was used to evaluate the predictive accuracy of the risk score for 1-year, 3-year, and 5-year survival.

[0028] 9. Construction of prognostic nomogram: The R package “rms” was used to construct a prognostic nomogram containing RiskScore and other clinical features. The performance of the nomogram was evaluated by calibration curve and ROC curve.

[0029] 10. Analysis of prognostic characteristic genes: The expression data of the TCGA-LGG cohort were combined with the expression data of normal brain tissue in the GTEx database, and the Wilcoxon test was performed using the R package "rstatix". Box plots were then generated to visualize gene expression levels. In addition, the proportion of each cell type in each sample in the single-cell data was calculated, and the average gene expression in each cell type group was determined. The genes with the highest average expression were selected, and the correlation between the average gene expression and the cell content in each sample was calculated using the cor.test function provided by R. The cor.test function uses the Pearson correlation coefficient to evaluate the linear relationship between gene expression and cell content.

[0030] 11. Predicting immunotherapy responsiveness: The Tumor Immune Dysfunction and Exclusion (TIDE) algorithm is used to predict the potential response of LGG patients to treatment with immune checkpoint inhibitors (ICIs)

[27] . A higher TIDE score indicates a lower likelihood of benefiting from immunotherapy and an increased risk of immune escape. The dysfunction score indicates the level of immune cell dysfunction infiltrating the tumor, while the rejection score reflects the degree of immune cell rejection within the tumor microenvironment.

[0031] 12. Drug sensitivity analysis: To predict the sensitivity of LGG patients in the high and low RiskScore groups to different drugs, the R package “oncoPredict”

[28] was used in combination with the expression matrix and drug treatment information in the Genomics of Drug Sensitivity to Cancer (GDSC) database as a training set to evaluate the chemotherapy sensitivity of LGG patients. The half-inhibitory concentration (IC50 value) was used as the main indicator to evaluate the sensitivity of LGG cancer cells to chemotherapy drugs, and the difference in IC50 between the high and low RiskScore groups was compared by Wilcoxon test. p < 0.05 was considered statistically significant.

[0032] 12. Statistical analysis: All statistical analyses were performed using R software (version 4.3.0). All R packages used in the study are open access and free to use. Statistical significance was determined by p < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001).

[0033] 2. Experimental Results 1. Annotation of cell types and glycosylation scores: A total of 5,368 cells and 22,559 genes were screened from the GSE117891 dataset. By collecting common cell marker genes of brain tissue from the CellMarker 2.0 website and combining them with classic cell markers, four cell types were finally annotated: oligodendrocytes (MBP, MOP, PLP1, MAG), T cells (CD3D, CD3E, CD8A), astrocytes (GFAP, AQP4, SOX9, CLU), and microglia (TMEM119, CX3CR1, P2RY12). Oligodendrocytes accounted for 744 cells, about 13.7%, T cells accounted for 379 cells, about 7.1%, astrocytes accounted for 2993 cells, about 55.8%, and microglia accounted for 1252 cells, about 23.4%. Use UMAP and dot plots to visualize these cell types ( Figure 1 A, see Figure 1 B and Figure 2 A).

[0034] Based on the gene expression levels of the "KEGG_N_GLYCAN_BIOSYNTHESIS" and "KEGG_O_GLYCAN_BIOSYNTHESIS" pathways in the KEGG pathway database, the glycosylation score of each cell type was calculated using three algorithms: AUCell, UCell, and ssGSEA. The results showed that for both N-glycosylation and O-glycosylation, microglia showed the highest glycosylation score in all three algorithms, significantly higher than other cell types ( Figure 1 C, Figure 2 B). This suggests that glycosylation may play an important role in the function of microglia.

[0035] 2. Analysis of differentially expressed genes in microglia: To understand the unique molecular characteristics of microglia, 1,684 differentially expressed genes (DEGs) were identified compared with other cell types and visualized using volcano plots ( Figure 1 D). GO and KEGG enrichment results ( Figure 1 E, F) showed that these DEGs were enriched in pathways related to cytokine regulation, immune response activation, ATP synthesis coupled to electron transport, and other related signaling pathways.

[0036] 3. Glycosylation analysis of the TCGA-LGG cohort: Kaplan-Meier survival analysis revealed significant survival differences based on N-glycosylation pathway activity (P<0.001, Figure 3 A), indicating that patients with lower N-glycosylation pathway activity had better prognosis. 3,019 DEGs were identified between high and low N-glycosylation pathway activity subgroups ( Figure 3 B, and functional enrichment analysis using GO and KEGG highlighted potential changes in immune signaling pathways ( Figure 3 C, D). In addition, the immune microenvironment of each patient was evaluated. The results showed that patients with high N-glycosylation pathway activity showed elevated stromal scores, immune scores, and estimated scores, as well as reduced tumor purity, indicating increased tumor heterogeneity and a more complex immune microenvironment ( Figure 3 E).

[0037] There was no significant difference in prognosis between the high and low O-glycosylation groups ( Figure 4 A). Similarly, enrichment analysis of 3,164 DEGs between high and low O-glycosylation pathway activity groups ( Figure 4 B) shows that these genes are enriched in signaling pathways related to chromosome segregation, synaptic signaling regulation, and cell cycle ( Figure 4 C, D). Compared with the low activity group, the high O-glycosylation pathway activity group showed higher matrix score, immune score, and estimated score, indicating a similar complex immune microenvironment ( Figure 4 E). These findings highlight the potential role of N- and O-glycosylation levels in influencing immune cell types and responses in the LGG tumor microenvironment.

[0038] 4. Establishment and validation of the prognostic signature model based on glycosylation: After recognizing the impact of glycosylation on the prognosis and immune microenvironment of LGG patients, the prognostic signature was further explored to understand the potential complexity of LGG. The differentially expressed genes (DEGs) of microglia and other cell populations in GSE117891, the DEGs of high and low N-glycosylation groups in the TCGA-LGG cohort, and the DEGs of high and low O-glycosylation groups were screened, and a Venn diagram was drawn and the intersection was taken, and 125 DEGs related to glycosylation changes were obtained ( Figure 5 A). Through univariate Cox regression analysis, 105 genes were confirmed to be associated with prognosis. After LASSO regression, 8 key genes were identified ( Figure 7 AB), constructed a prognostic feature model called RiskScore ( Figure 5 B), including ASPM, CHI3L1, LILRA4, MSN, OCIAD2, PTGER4, SERPING1 and TNFRSF12A.

[0039] The formula is as follows: RiskScore=(0.382×A)+(0.017×C)+(0.068×L)+(0.258×M)+(0.116×O)+(-0.137×P)+(0.109×S)+(0.073×T); where RiskScore is the risk score, A is the expression level of ASPM; C is the expression level of CHI3L1; L is the expression level of ILRA4; M is the expression level of MSN; O is the expression level of OCIA; P is the expression level of PTGER4; S is the expression level of SERPING1; and T is the expression level of TNFRSF12A. The RiskScore of each patient in the TCGA-LGG cohort was calculated, and the patients were divided into high RiskScore group and low RiskScore group according to the median RiskScore. The survival status diagram of the patients was shown ( Figure 5 D) and the expression levels of 8 genes in RiskScore ( Figure 5 C). The results showed that patients with higher RiskScore values ​​had worse survival outcomes, and these patients had higher expression levels of ASPM, CHI3L1, LILRA4, MSN, OCIAD2, PTGER4, SERPING1, and TNFRSF12A. Kaplan-Meier analysis showed that patients in the low RiskScore group had significantly better overall survival (OS) (P<0.001, Figure 5 E).

[0040] To evaluate the predictive ability of the prognostic signature, receiver operating characteristic (ROC) curves for 1-, 3-, and 5-year OS were generated ( Figure 5 H), the area under the curve (AUC) was 0.882, 0.859 and 0.785, respectively, indicating that the model has good predictive performance. In addition, the prognostic signature was validated in two independent external validation sets (CGGA_693, CGGA_325), producing satisfactory results, with 5-year AUCs of 0.669 and 0.634 ( Figure 5 FG, 5I-J).

[0041] To improve clinical utility, a nomogram was created by combining RiskScore and clinical parameters. Univariate and multivariate Cox regression analyses evaluated the impact of clinical characteristics and RiskScore on LGG survival, and the results showed that age and RiskScore were important predictors of survival ( Figure 6 AB). Based on these predictors, a nomogram was designed to estimate the 1-, 3-, and 5-year survival probabilities of LGG patients ( Figure 6C). The ROC curve showed that the AUC of the nomogram was 0.906 at 1 year, 0.898 at 2 years, and 0.819 at 3 years ( Figure 6 D), calibration curve demonstrates the predictive accuracy of the nomogram ( Figure 7 H). In addition, the predictive performance of the nomogram was validated in two independent external validation sets ( Figure 6 EF, Figure 7 IJ). Collectively, these results suggest that the RiskScore-based nomogram, rooted in the concept of glycosylation, has substantial clinical predictive value for LGG.

[0042] 5. Differences in prognostic signature gene expression and immune landscape: Analysis of the TCGA-LGG cohort combined with GTEx normal brain tissue data revealed significant expression differences in prognostic signature genes. Compared with normal tissues, ASPM, LILRA4, MSN, PTGER4, and TNFRSF12A were significantly upregulated in LGG patients, while CHI3L1, OCIAD2, and SERPING1 showed decreased expression ( Figure 8 A). We further examined the distribution of these eight genes in the GSE117891 dataset ( Figure 8 B, C). The results showed that ASPM, CHI3L1, and TNFRSF12A were more highly expressed in astrocytes, while LILRA4, MSN, PTGER4, and SERPING1 were more highly expressed in microglia. OCIAD2 was more highly expressed in T cells. The proportion of each cell type in each sample was calculated, and the average expression of the gene in the group of cells with the highest expression was determined. Then, the correlation between gene expression and cell proportion was analyzed ( Figure 8 D, E). The results showed that ASPM expression was positively correlated with the proportion of astrocytes (cor=0.61, p<0.05), OCIAD2 expression was negatively correlated with the proportion of oligodendrocytes (cor=-0.59, p<0.05), and SERPING1 expression was positively correlated with the proportion of T cells (cor=0.57, p<0.05).

[0043] The differences in the immune microenvironment between patients with different risk score groups in the TCGA-LGG cohort were also investigated. The results showed that the glycosylation score was positively correlated with the risk score. Compared with patients in the low-risk score group, patients in the high-risk score group had higher stromal, immune, and estimated scores, but showed lower tumor purity ( Figure 8 F). This suggests a more complex immune microenvironment characterized by increased tumor heterogeneity.

[0044] 6. Response to immunotherapy in different risk score groups: The Tumor Immune Dysfunction and Exclusion (TIDE) algorithm evaluates the potential for tumor immune escape by analyzing the gene expression profile of tumor samples. The analysis showed that in the TCGA cohort, the high-risk score group exhibited higher TIDE, exclusion, and dysfunction scores, as well as lower MSI (microsatellite instability) scores. This suggests that patients in the high-risk score group are more likely to experience immune escape and may respond worse to immune checkpoint inhibitor (ICI) treatment ( Fig. 9 A). In contrast, the low-risk score group predicted a higher proportion of patients who were likely to respond positively to immunotherapy ( Fig. 9 B). These findings suggest that patients with low risk scores may derive greater benefit from immunotherapy than those with high risk scores.

[0045] In the LGG pediatric peptide vaccine immunotherapy cohort, the high-risk score group predicted a higher proportion of patients who would respond to immunotherapy ( Fig. 9 C). In contrast, in the GBM anti-PD-1 immunotherapy cohort, the low-risk score group indicated a higher proportion of patients who responded to immunotherapy ( Fig. 9 D). The ROC of the risk score for predicting immunotherapy response was 0.77 and 0.626 in the two cohorts, respectively ( Fig. 9 E, F).

[0046] In addition, the relationship between risk score and drug treatment effectiveness in LGG treatment was investigated. The results showed that high risk score values ​​were associated with lower IC50 of temozolomide and gemcitabine and higher IC50 of gefitinib ( Fig. 9 GI, P < 0.05). Therefore, it is suggested that the risk score can be used as a valuable predictor of drug sensitivity in LGG patients.

[0047] 7. Differences in prognosis of drug treatment in different risk score groups: In the TCGA-LGG cohort, we analyzed the survival of the drug group and the non-drug group. The results showed that at the 2-year survival time point, the short-term prognosis (<2 years) and long-term prognosis (>=2 years) of the drug group and the non-drug group showed different trends ( Fig.10 AC). Specifically, among all LGG patients, the short-term prognosis was better in the medication group, while the long-term prognosis was better in the non-medication group.

[0048] When further analyzing the high and low RiskScore groups, this phenomenon was particularly significant in the high RiskScore group. In the high RiskScore group, patients who received drug treatment had a better prognosis in the short term, but a worse prognosis in the long term ( Fig.10DF). In contrast, in the low RiskScore group, the long-term and short-term prognosis of patients had no significant relationship with whether they received drug treatment ( Fig.10 GI).

[0049] In the independent validation cohort CGGA-325, the results were consistent with those of the TCGA-LGG cohort, further supporting the above conclusions ( Fig.11 ). These findings suggest that RiskScore may not only serve as a prognostic indicator, but may also help identify which patients may benefit from drug treatment in the short term, while requiring caution in the long term. However, no consistent results were observed in the CGGA-693 cohort, and further studies are needed to validate these findings ( Fig.12 ).

[0050] Despite the low malignancy of LGG, its treatment still faces many challenges. The slow growth and unclear borders of LGG often make complete resection of the tumor difficult. In addition, radiotherapy and chemotherapy may affect LGG differently, and some patients develop resistance to these treatments, leading to unsatisfactory results. Even with combined surgical resection, radiotherapy and chemotherapy, some patients still experience poor prognosis and high tumor recurrence rates. In addition, LGG patients often face a decline in cognitive function and quality of life during long-term treatment, which makes the management of treatment side effects an important issue

[29] . Given the complexity and individual differences in LGG treatment, it is particularly important to develop a reliable prognostic model.

[0051] In this context, glycosylation, a common post-translational modification of proteins, has become a focus of research. Glycosylation plays a crucial role in helping immune cells to properly localize and migrate

[30] . However, abnormal glycosylation modification is closely related to tumorigenesis, proliferation, invasion, metastasis, and immune escape

[31]

[32] . Genetic, epigenetic, metabolic, inflammatory, and environmental mechanisms can lead to glycosylation modifications, driving several biological processes in cancer

[33] . The first report of abnormal glycosylation modification in tumors can be traced back to 50 years ago

[34] . Due to the sensitivity to glycosylation, even mild pathogenic changes or metabolic stress can lead to glycosylation dysfunction, resulting in abnormal glycans and glycoproteins

[15] . Understanding the causes and consequences of glycosylation changes associated with neoplastic diseases will provide valuable insights into tumor development

[31] .

[0052] Risk models constructed from glycosylation-related genes have been shown to be closely associated with the overall survival and tumor microenvironment of patients with prostate cancer

[35] , lung adenocarcinoma

[36] , renal cell carcinoma

[37] , and bladder cancer

[38] . Risk models of N-glycosylation-related genes can effectively predict the prognosis of patients with hepatocellular carcinoma and the immune status of the tumor microenvironment

[39] . Despite this, the characterization of glycosylation-related gene sets in gliomas has not been comprehensively analyzed, and the relationship between glycosylation-related genes and glioma prognosis has rarely been studied.

[0053] It is worth noting that a recent study showed that a prognostic model constructed by analyzing differential glycosylation-related regulatory genes between glioma and normal brain tissue can accurately predict the prognosis of glioma patients. This is helpful to study the occurrence and progression of glioma and identify new targets for glioma diagnosis and treatment

[40] . However, since the samples were not properly stratified according to the different grades of glioma in the study, this may affect the accurate revelation of the gene expression differences of tumors of different grades and their correlation with prognosis.

[0054] In this study, we constructed and validated a glycosylation-related risk signature based on eight glycosylation-related genes (ASPM, CHI3L1, LILRA4, MSN, OCIAD2, PTGER4, SERPING1, and TNFRSF12A) as a novel prognostic biomarker for LGG patients and potentially predictive of ICB immunotherapy response. In addition, we explored the relationship between the risk signature and immune signature and drug sensitivity.

[0055] Starting from the single-cell level, four cell types were identified in the LGG scRNA-seq data, and the glycosylation pathway was found to be active in the microglial population. Combined with TCGA-LGG data, we constructed and validated a glycosylation-related risk signature based on eight glycosylation-related genes (ASPM, CHI3L1, LILRA4, MSN, OCIAD2, PTGER4, SERPING1, and TNFRSF12A). Among the eight glycosylation-related genes we studied, ASPM (abnormal spindle-like microcephaly-associated protein) is a spindle pole / intermediate protein that regulates mitosis and cytokinesis

[41] . ASPM is abnormally expressed in various tumors, such as glioblastoma

[42] , endometrial adenocarcinoma

[43] , pancreatic cancer

[44] , prostate adenocarcinoma

[45] , and ovarian cancer

[46]

[47] , and is associated with tumor prognosis. CHI3L1 (chitinase-3-like protein 1) is a member of the glycoside hydrolase family

[48] . Elevated serum CHI3L1 levels are associated with disease severity in a variety of human tumors, including breast cancer, colon cancer, prostate cancer, ovarian cancer, brain cancer, thyroid cancer, lung cancer, and liver cancer, leading to poor prognosis and shorter survival

[49] . In glioma, CHI3L1 reprograms the tumor microenvironment (TME) by promoting NF-κB pathway activation, regulating tumor malignancy and local invasiveness, making it a potential therapeutic target for glioma

[50]

[51] . Dendritic cells (DCs) are divided into myeloid dendritic cells (mDCs) and plasmacytoid dendritic cells (pDCs)

[52] , of which pDCs specifically express orphan receptor immunoglobulin-like transcript 7 (ILT7, also known as LILRA4 and CD85g)

[53] . Typically, ILT7 / ILT7L signaling generates negative immune response feedback after viral infection

[54] . Intervention of the ILT7L / ILT7 system may enhance anti-tumor and anti-viral immunity

[52] . The moesin protein encoded by the MSN gene is part of the ezrin-radixin-moesin (ERM) protein family

[55] . Moesin is upregulated in various human cancers, including breast cancer, prostate cancer, pancreatic cancer, lung cancer, and melanoma

[56] . Studies have shown that MSN may be a new therapeutic target for colorectal cancer

[57] . OCIAD2 is part of the ovarian cancer immune response antigen (OCIA) domain family and promotes tumor metastasis by enhancing STAT3 activation and cell migration

[58] . OCIAD2 is associated with the prognosis of bladder cancer patients and shows potential in immunotherapy

[59] . The methylation status of OCIAD2 may be a useful prognostic marker for patients with hepatoblastoma

[60] and lung adenocarcinoma

[61] . PTGER4 is a major prostaglandin E2 (PGE2) receptor, and its genetic variation and expression levels can affect gastric cancer

[62] .SERPING1 encodes a highly glycosylated plasma protein that is involved in the regulation of the complement cascade and immune responses

[63] . Studies have shown that SERPING1 can serve as a new biomarker for the diagnosis and prognosis of prostate cancer

[64] and is associated with early detection of breast cancer bone metastasis

[65] . Fibroblast growth factor-inducible 14 (Fn14; TNFRSF12A) is a cell surface receptor for TNF-like weak inducer of apoptosis (TWEAK) of the tumor necrosis factor (TNF) family. TNFRSF12A expression is generally low in normal tissues but is significantly increased after tissue damage and in many solid tumor types

[66] , including gliomas, breast cancer, esophageal adenocarcinoma, pancreatic cancer, and hepatocellular carcinoma. Overexpression of TNFRSF12A is associated with poor prognosis in these tumors

[67] .

[0056] The resulting risk score serves as a new prognostic biomarker for LGG patients and can predict ICB immunotherapy response. TIDE analysis showed that patients with low risk scores had a higher immunotherapy response rate compared with those with high risk scores. In the external immunotherapy cohort, the results of the GBM anti-PD-1 immunotherapy cohort supported this finding, while the LGG pediatric peptide vaccine immunotherapy cohort had a higher proportion of patients predicted to respond to immunotherapy in the high risk score group due to the small sample size. The data from these two cohorts showed that the risk score can be used to predict immunotherapy response. In addition, we also explored the relationship between risk characteristics, immune characteristics, and drug sensitivity. Our results showed that patients in the high risk score group showed higher glycosylation pathway activity and more complex immune microenvironment compared with the low risk score group, and the high risk score was associated with lower IC50 of temozolomide and gemcitabine and higher IC50 of gefitinib. Innovatively, we found that LGG patients who received drug treatment tended to have a better short-term prognosis but a worse long-term prognosis. In contrast, LGG patients who did not receive drug treatment had a good long-term prognosis but a poor short-term prognosis. This prognostic difference was particularly significant in the high-risk score group. This finding provides new insights into the treatment strategy for LGG.

[0057] By evaluating the predictive value of these genes in the prognostic model of LGG patients, we revealed their potential roles in the effectiveness of immunotherapy and drug treatment, providing a basis for the precision treatment of LGG patients.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

[0059] References: [1]Wen PY, Packer RJ. The 2021 WHO Classification of Tumors of theCentral Nervous System: clinical implications. Neuro Oncol. 2021;23(8):1215-1217. doi:10.1093 / neuonc / noab120 [2]Ostrom QT, Gittleman H, Farah P, et al. CBTRUS statistical report: Primary brain and central nervous system tumors diagnosed in the United States in 2006-2010 [published correction appears in Neuro Oncol. 2014 May;16(5):760]. Neuro Oncol. 2013;15 Suppl 2(Suppl 2):ii1-ii56. doi:10.1093 / neuonc / not151 [3]Louis DN, Perry A, Reifenberger G, et al. The 2016 World HealthOrganization Classification of Tumors of the Central Nervous System: asummary. Acta Neuropathol. 2016;131(6):803-820. doi:10.1007 / s00401-016-1545-1 [4]Morshed RA, Young JS, Hervey-Jumper SL, Berger MS. The managementof low-grade gliomas in adults. J Neurosurg Sci. 2019;63(4):450-457. doi:10.23736 / S0390-5616.19.04701-5 [5]Hayhurst C. Contemporary management of low--grade glioma: aparadigm shift in neuro-oncology. Pract Neurol. 2017;17(3):183-190. doi:10.1136 / practneurol-2017-001604 [6]Youssef G, Miller JJ. Lower Grade Gliomas. Curr Neurol NeurosciRep. 2020;20(7):21. Published 2020 May 22. doi:10.1007 / s11910-020-01040-8 [7]Murphy ES, Leyrer CM, Parsons M, et al. Risk Factors for MalignantTransformation of Low-Grade Glioma. Int J Radiat Oncol Biol Phys. 2018;100(4):965-971. doi:10.1016 / j.ijrobp.2017.12.258 [8]Schjoldager KT, Narimatsu Y, Joshi HJ, Clausen H. Global view ofhuman protein glycosylation pathways and functions. Nat Rev Mol Cell Biol.2020;21(12):729-749. doi:10.1038 / s41580-020-00294-x [9]Bangarh R, Khatana C, Kaur S, et al. Aberrant proteinglycosylation: Implications on diagnosis and Immunotherapy. Biotechnol Adv.2023;66:108149. doi:10.1016 / j.biotechadv.2023.108149

[10] Okazaki M, Mogushi K, Denda-Nagai K, et al. Biological andClinicopathological Implications of Beta-3-N-acetylglucosaminyltransferase 8in Triple-negative Breast Cancer. Anticancer Res. 2021;41(2):845-858. doi:10.21873 / anticanres.14837

[11] Xu X, Wu Y, Jia G, Zhu Q, Li D, Xie K. A signature based onglycosyltransferase genes provides a promising tool for the prediction ofprognosis and immunotherapy responsiveness in ovarian cancer. J Ovarian Res.2023;16(1):5. Published 2023 Jan 7. doi:10.1186 / s13048-022-01088-9

[12] Xu X, Wu Y, Jia G, Zhu Q, Li D, Xie K. A signature based onglycosyltransferase genes provides a promising tool for the prediction ofprognosis and immunotherapy responsiveness in ovarian cancer. J Ovarian Res.2023;16(1):5. Published 2023 Jan 7. doi:10.1186 / s13048-022-01088-9

[13] Danolic D, Heffer M, Wagner J, et al. Role of gangliosidebiosynthesis genetic polymorphism in cervical cancer development. J ObstetGynaecol. 2020;40(8):1127-1132. doi:10.1080 / 01443615.2019.1692801

[14] Liu X, Shi J, Tian L, et al. Comprehensive prognostic and immuneanalysis of a glycosylation related risk model in pancreatic cancer. BMCCancer. 2023;23(1):1229. Published 2023 Dec 14. doi:10.1186 / s12885-023-11725-1

[15] Silsirivanit A. Glycosylation markers in cancer. Adv Clin Chem.2019;89:189-213. doi:10.1016 / bs.acc.2018.12.005

[16] Colaprico A, Silva TC, Olsen C, et al. TCGAbiolinks: an R / Bioconductor package for integrative analysis of TCGA data. Nucleic AcidsRes. 2016;44(8):e71. doi:10.1093 / nar / gkv1507

[17] Yu K, Hu Y, Wu F, et al. Surveying brain tumor heterogeneity bysingle-cell RNA-sequencing of multi-sector biopsies. Natl Sci Rev. 2020;7(8):1306-1318. doi:10.1093 / nsr / nwaa099

[18] Hao Y, Hao S, Andersen-Nissen E, et al. Integrated analysis ofmultimodal single-cell data. Cell. 2021;184(13):3573-3587.e29. doi:10.1016 / j.cell.2021.04.048

[19] Fan C, Chen F, Chen Y, et al. irGSEA: the integration of single-cell rank-based gene set enrichment analysis. Brief Bioinform. 2024;25(4):bbae243. doi:10.1093 / bib / bbae243

[20] Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variationanalysis for microarray and RNA-seq data. BMC Bioinformatics. 2013;14:7.Published 2013 Jan 16. doi:10.1186 / 1471-2105-14-7

[21] Love MI, Huber W, Anders S. Moderated estimation of fold changeand dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550.doi:10.1186 / s13059-014-0550-8

[22] Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package forcomparing biological themes among gene clusters. OMICS. 2012;16(5):284-287.doi:10.1089 / omi.2011.0118

[23] Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumourpurity and stromal and immune cell admixture from expression data. NatCommun. 2013;4:2612. doi:10.1038 / ncomms3612

[24] Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. ProfilingTumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018;1711:243-259. doi:10.1007 / 978-1-4939-7493-1_12

[25] Friedman J, Hastie T, Tibshirani R. Regularization Paths forGeneralized Linear Models via Coordinate Descent. J Stat Softw. 2010;33(1):1-22.

[26] Blanche P, Dartigues JF, Jacqmin-Gadda H. Estimating andcomparing time-dependent areas under receiver operating characteristic curvesfor censored event times with competing risks. Stat Med. 2013;32(30):5381-5397. doi:10.1002 / sim.5958

[27] Jiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction andexclusion predict cancer immunotherapy response. Nat Med. 2018;24(10):1550-1558. doi:10.1038 / s41591-018-0136-1

[28] Maeser D, Gruener RF, Huang RS. oncoPredict: an R package forpredicting in vivo or cancer patient drug response and biomarkers from cellline screening data. Brief Bioinform. 2021;22(6):bbab260. doi:10.1093 / bib / bbab260

[29] Forst DA, Nahed BV, Loeffler JS, Batchelor TT. Low-grade gliomas.Oncologist. 2014;19(4):403-413. doi:10.1634 / theoncologist.2013-0345

[30] Liu YS, Guo XY, Hirata T, et al. N-Glycan-dependent proteinfolding and endoplasmic reticulum retention regulate GPI-anchor processing. JCell Biol. 2018;217(2):585-599. doi:10.1083 / jcb.201706135

[31] Stowell SR, Ju T, Cummings RD. Protein glycosylation in cancer.Annu Rev Pathol. 2015;10:473-510. doi:10.1146 / annurev-pathol-012414-040438

[32] Mereiter S, Balmaña M, Campos D, Gomes J, Reis CA. Glycosylationin the Era of Cancer-Targeted Therapy: Where Are We Heading?. Cancer Cell.2019;36(1):6-16. doi:10.1016 / j.ccell.2019.06.006

[33] Pinho SS, Reis CA. Glycosylation in cancer: mechanisms andclinical implications. Nat Rev Cancer. 2015;15(9):540-555. doi:10.1038 / nrc3982

[34] Meezan E, Wu HC, Black PH, Robbins PW. Comparative studies on thecarbohydrate-containing membrane components of normal and virus-transformedmouse fibroblasts. II. Separation of glycoproteins and glycopeptides bysephadex chromatography. Biochemistry. 1969;8(6):2518-2524. doi:10.1021 / bi00834a039

[35] Liu X, Shi J, Tian L, Xiao B, Zhang K, Zhu Y, Zhang Y, Jiang K,Zhu Y, Yuan H. Comprehensive prognostic and immune analysis of aglycosylation related risk model in pancreatic cancer. BMC Cancer. 2023 Dec14;23(1):1229. doi: 10.1186 / s12885-023-11725-1. PMID: 38097951; PMCID:PMC10720206.

[36] Liu XH, Yang YM, Zhang MM, Fu HY. Five glycosylation-related genesignatures predict the prognostic risks of lung adenocarcinoma. Gen PhysiolBiophys. 2023;42(6):479-493. doi:10.4149 / gpb_2023025

[37] Shen C, Zheng B, Chen Z, Zhang W, Chen X, Xu S, Ji J, Fang X, ShiC. Identification of prognostic models for glycosylation-related subtypes andtumor microenvironment infiltration characteristics in clear cell renal cellcancer. Heliyon. 2024 Mar 13;10(6):e27710. doi: 10.1016 / j.heliyon.2024.e27710. PMID: 38515689; PMCID: PMC10955297.

[38] Liu J, He Y, Zhou W, Tang Z, Xiao Z. A glycosylation risk scorecomprehensively assists the treatment of bladder neoplasm in the real-worldcohort, including the tumor microenvironment, molecular and clinicalprognosis. Front Pharmacol. 2023 Sep 25;14:1280428. doi: 10.3389 / fphar.2023.1280428. PMID: 37818187; PMCID: PMC10560734.

[39] Lin S, Cao Y, ZhuK, Yang C, Zhu X, Zhang H, Zhang R.Identification of a Novel Prognostic Signature Based on N-LinkedGlycosylation and Its Correlation with Immunotherapy Response inHepatocellular Carcinoma. J Hepatocell Carcinoma. 2023 Oct 9;10:1749-1765.doi: 10.2147 / JHC.S417407. PMID: 37841372; PMCID: PMC10575065.

[40] Sun YF, Zhang LC, Niu RZ, et al. Predictive potentials ofglycosylation-related genes in glioma prognosis and their correlation withimmune infiltration. Sci Rep. 2024;14(1):4478. Published 2024 Feb 23. doi:10.1038 / s41598-024-51973-0

[41] Kouprina N, Pavlicek A, Collins NK, et al. The microcephaly ASPMgene is expressed in proliferating tissues and encodes for a mitotic spindleprotein. Hum Mol Genet. 2005;14(15):2155-2165. doi:10.1093 / hmg / ddi220

[42] Chen X, Huang L, Yang Y, et al. ASPM promotes glioblastoma growthby regulating G1 restriction point progression and Wnt-β-catenin signaling.Aging (Albany NY). 2020;12(1):224-241. doi:10.18632 / aging.102612

[43] Zhou JW, Wang H, Sun W, Han NN, Chen L. ASPM is a predictor ofoverall survival and has therapeutic potential in endometrial cancer. Am JTransl Res. 2020;12(5):1942-1953. Published 2020 May 15.

[44] Hsu CC, Liao WY, Chan TS, et al. The differential distributionsof ASPM isoforms and their roles in Wnt signaling, cell cycle progression,and pancreatic cancer prognosis. J Pathol. 2019;249(4):498-508. doi:10.1002 / path.5341

[45] Pai VC, Hsu CC, Chan TS, et al. ASPM promotes prostate cancerstemness and progression by augmenting Wnt-Dvl-3-β-catenin signaling[published correction appears in Oncogene. 2019 Feb;38(8):1354. doi: 10.1038 / s41388-018-0561-0]. Oncogene. 2019;38(8):1340-1353. doi:10.1038 / s41388-018-0497-4

[46] Hao W, Zhao H, Li Z, et al. Identification of potential markersfor differentiating epithelial ovarian cancer from ovarian low malignantpotential tumors through integrated bioinformatics analysis. J Ovarian Res.2021;14(1):46. Published 2021 Mar 16. doi:10.1186 / s13048-021-00794-0

[47] Ye Q, Lei L, Aili AX. Identification of potential targets forovarian cancer treatment by systematic bioinformatics analysis. Eur JGynaecol Oncol. 2015;36(3):283-289.

[48] Chiang YC, Lin HW, Chang CF, et al. Overexpression of CHI3L1 isassociated with chemoresistance and poor outcome of epithelial ovariancarcinoma. Oncotarget. 2015;6(37):39740-39755. doi:10.18632 / oncotarget.5469

[49] Coffman FD. Chitinase 3-Like-1 (CHI3L1): a putative diseasemarker at the interface of proteomics and glycomics. Crit Rev Clin Lab Sci.2008;45(6):531-562. doi:10.1080 / 10408360802334743

[50] Zhao T, Zeng J, Xu Y, et al. Chitinase-3 like-protein-1 promotesglioma progression via the NF-κB signaling pathway and tumor microenvironmentreprogramming. Theranostics. 2022;12(16):6989-7008. Published 2022 Oct 3.doi:10.7150 / thno.75069

[51] Ku BM, Lee YK, Ryu J, et al. CHI3L1 (YKL-40) is expressed inhuman gliomas and regulates the invasion, growth and survival of gliomacells. Int J Cancer. 2011;128(6):1316-1326. doi:10.1002 / ijc.25466

[52] Palma G, De Laurenzi V, De Marco M, et al. Plasmacytoidsdendritic cells are a therapeutic target in anticancer immunity. BiochimBiophys Acta. 2012;1826(2):407-414. doi:10.1016 / j.bbcan.2012.04.007

[53] Cao W, Bover L, Cho M, et al. Regulation of TLR7 / 9 responses inplasmacytoid dendritic cells by BST2 and ILT7 receptor interaction. J ExpMed. 2009;206(7):1603-1614. doi:10.1084 / jem.20090547

[54] Tsukamoto N, Okada S, Onami Y, Sasaki Y, Umezawa K, Kawakami Y.Impairment of plasmacytoid dendritic cells for IFN production by the ligandfor immunoglobulin-like transcript 7 expressed on human cancer cells. ClinCancer Res. 2009;15(18):5733-5743. doi:10.1158 / 1078-0432.CCR-09-0171

[55] Clucas J, Valderrama F. ERM proteins in cancer progression[published correction appears in J Cell Sci. 2015 Mar 15;128(6):1253. doi:10.1242 / jcs.170027]. J Cell Sci. 2014;127(Pt 2):267-275. doi:10.1242 / jcs.133108

[56] Ansa-Addo EA, Zhang Y, Yang Y, Hussey GS, Howley BV, Salem M,Riesenberg B, Sun S, Rockey DC, Karvar S, Howe PH, Liu B, Li Z. Membrane-organizing protein moesin controls Treg differentiation and antitumorimmunity via TGF-β signaling. J Clin Invest. 2017 Apr 3;127(4):1321-1337.doi: 10.1172 / JCI89281. Epub 2017 Mar 13. PMID: 28287407; PMCID: PMC5373867.

[57] Huang CY, Wei PL, Batzorig U, Makondi PT, Lee CC, Chang YJ.Identification of Moesin (MSN) as a Potential Therapeutic Target forColorectal Cancer via the β-Catenin-RUNX2 Axis. Int J Mol Sci. 2023;24(13):10951. Published 2023 Jun 30. doi:10.3390 / ijms241310951

[58] Zhu A, Pei D, Zong Y, Fan Y, Wei S, Xing Z, Song S, Wang X, GaoX. Comprehensive analysis to identify a novel diagnostic marker of lungadenocarcinoma and its immune infiltration landscape. Front Oncol. 2023 Jun20;13:1199608. doi: 10.3389 / fonc.2023.1199608. PMID: 37409245; PMCID:PMC10319060.

[59] Yao Z, Zhang H, Zhang X, Zhang Z, Jie J, Xie K, Li F, Tan W.Identification of tumor microenvironment-related signature for predictingprognosis and immunotherapy response in patients with bladder cancer. FrontGenet. 2022 Sep 6;13:923768. doi: 10.3389 / fgene.2022.923768. PMID: 36147509;PMCID: PMC9485450.

[60] Honda S, Minato M, Suzuki H, Fujiyoshi M, Miyagi H, Haruta M,Kaneko Y, Hataka KC, Hiyama E, Kamijo T, Okada T, Taketomi A. Clinicalprognostic value of DNA methylation in hepatoblastoma: Four novel tumorsuppressor candidates. Cancer Sci. 2016 Jun;107(6):812-9. doi: 10.1111 / cas.12928. Epub 2016 Apr 27. PMID: 26991471; PMCID: PMC4968605.

[61] Maki M, JeongMin H, Nakagawa T, et al. Aberrant OCIAD2demethylation in lung adenocarcinoma is associated with outcome. Pathol Int.2022;72(10):496–505. doi:10.1111 / pin.13262

[62] Yu S, Tu R, Chen Z, Song J, Li P, Hu F, Yuan G, Zhang R, Li Y.Association of PTGER4 and PRKAA1 genetic polymorphisms with gastric cancer.BMC Med Genomics. 2023 Sep 5;16(1):209. doi: 10.1186 / s12920-023-01645-1.PMID: 37670284; PMCID: PMC10478487.

[63] Ponard D, Gaboriaud C, Charignon D, et al. SERPING1 mutation update: Mutation spectrum and C1 Inhibitor phenotypes. Hum Mutat. 2020;41(1):38-5 doi:10.1002 / hum.23917

[64] Peng S, Du T, Wu W, et al. Decreased expression of serineprotease inhibitor family G1 (SERPING1) in prostate cancer can helpdistinguish high-risk prostate cancer and predicts malignant progression.Urol Oncol. 2018;36(8):366.e1-366.e9. doi:10.1016 / j.urolonc.2018.05.021

[65] Zhao Z, Yang H, Ji G, Su S, Fan Y, Wang M, Gu S. Identificationof hub genes for early detection of bone metastasis in breast cancer. FrontEndocrinol (Lausanne). 2022 Sep 29;13:1018639. doi: 10.3389 / fendo.2022.1018639. PMID: 36246872; PMCID: PMC9556899.

[66] Perez JG, Tran NL, Rosenblum MG, Schneider CS, Connolly NP, KimAJ, Woodworth GF, Winkles JA. The TWEAK receptor Fn14 is a potential cellsurface portal for targeted delivery of glioblastoma therapeutics. Oncogene.2016 Apr 28;35(17):2145-55. doi: 10.1038 / onc.2015.310. Epub 2015 Aug 24.PMID: 26300004; PMCID: PMC4850525.

[67] Wang Y, Zhang S, Xie X, Chen Z, Wu L, Yu Z, Guo X, Chen G.Association of TNFRSF12A Methylation With Prognosis in HepatocellularCarcinoma With History of Alcohol Consumption. Front Genet. 2020 Jan 9;10:1299. doi: 10.3389 / fgene.2019.01299. PMID: 31998364; PMCID: PMC6964049。

Claims

1. A system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes, characterized by: The multiple glycosylation genes include: ASPM, CHI3L1, LILRA4, MSN, OCIA, PTGER4, SERPING1 and TNFRSF12A; the risk score characteristic model for evaluating the clinical prognosis of low-grade glioma is: RiskScore=(0.382×A)+(0.017×C)+(0.068×L)+(0.258×M)+(0.116×O)+(-0.137×P)+(0.109×S)+(0.073×T); wherein: RiskScore is the risk score, A is the expression level of ASPM; C is the expression level of CHI3L1; L is the expression level of ILRA4; M is the expression level of MSN; O is the expression level of OCIA; P is the expression level of PTGER4; S is the expression level of SERPING1; and T is the expression level of TNFRSF12A.

2. The system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes according to claim 1, characterized in that: The RiskScore ≥ 2.665581 indicates high risk, and the RiskScore < 2.665581 indicates low risk.

3. The system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes according to claim 1, characterized in that: Compared with normal tissues, the ASPM, LILRA4, MSN, PTGER4, and TNFRSF12A were significantly upregulated in low-grade glioma LGG, while the expressions of CHI3L1, OCIAD2, and SERPING1 were decreased; in astrocytes, the expressions of ASPM, CHI3L1, and TNFRSF12A were higher than those in other cell populations; in microglial populations, the expressions of LILRA4, MSN, PTGER4, and SERPING1 were higher than those in other cell populations; and the expression of OCIAD2 in T cells was higher than that in other cell populations; ASPM expression was positively correlated with the proportion of astrocytes, cor=0.61, p<0.05; OCIAD2 expression was negatively correlated with the proportion of oligodendrocytes, cor=-0.59, p<0.05; SERPING1 expression was positively correlated with the proportion of T cells, cor=0.57, p<0.

05.

4. Use of the system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes according to claim 1 in the preparation of a reagent or system for evaluating the clinical prognosis of low-grade glioma.

5. Use of the system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylated genes according to claim 1 in the preparation of a reagent or system for predicting or evaluating the potential effectiveness of sensitivity to chemotherapeutic drugs.

6. The use according to claim 5, characterized in that: The system for evaluating the clinical prognosis of low-grade glioma based on the expression of multiple glycosylation genes is used in the preparation of reagents or systems for identifying the short-term effectiveness of chemotherapy drugs in patients.

Images