Biomarker for detecting glioma microenvironment and application thereof

Abstract

The present invention provides a biomarker for the detection of glioma microenvironment (TME), comprising a serum marker selected from any one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14) and potential transforming growth factor beta binding protein 2 (LTBP2), and a pathological marker selected from any one or more of Annexin A1 (ANXA1), Annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14) and potential transforming growth factor beta binding protein 2 (LTBP2); the pathological marker is selected from one or more of periostin (POSTN), lactoferrin (LTF) and a phospholipase A2 group IIA (PLA2G2A), and the pathological marker is selected from one or more of periostin (POSTN), lactoferrin (LTF) and phospholipase A2 group IIA (PLA2G2A). The invention also provides an application of the biomarker in diagnosis of glioma. The invention provides an innovative idea and a biomarker for non-invasive diagnosis of glioma, and can be used for prognosis and malignant diagnosis of glioma.

Description

Technical Field

[0001] This invention relates to the field of biodiagnostic technology, and more specifically to biomarkers for detecting the glioma microenvironment (TME) and their uses. Background Technology

[0002] Gliomas are the most common primary brain tumors and are usually classified into four grades (1, 2, 3, and 4) based on histopathology. Grade 4 gliomas, also known as glioblastomas (GBM), are the most malignant gliomas. In 2021, the WHO classification system redefined adult diffuse gliomas into three molecular subtypes: astrocytoma (AC), characterized by isocitrate dehydrogenase gene mutation (IDHMU) with chromosome 1p / 19q not deleted (MU-Non); oligodendroglioma (OD), characterized by IDH-tagged MU with chromosome 1p / 19q deletion (MU-Cod); and GBM, characterized by IDH wild-type (IDHWT). [1] Despite significant progress in glioma biology, treatment options for high-grade gliomas remain limited, with high recurrence rates and extremely poor prognoses. The median survival for patients is approximately 14 months, and the 5-year survival rate is less than 5%. [2, 3] Current standard treatments, including surgical resection, radiotherapy, temozolomide chemotherapy, and emerging immunotherapy, have failed to achieve ideal results. [4] Other strategies, such as small molecule inhibitors, anti-angiogenic therapy, immune checkpoint inhibitors (ICB), antibody-drug conjugates, and oncolytic viruses, have also failed to show significant efficacy in clinical practice, mainly due to the heterogeneity of gliomas. [5, 6] The Cancer Genome Atlas (TCGA) revealed significant heterogeneity of gliomas both between and within tumors. [7] Currently, three main heterogeneous manifestations of gliomas are recognized: (1) genetic heterogeneity—a glioma may contain multiple genetically different subclones; (2) epigenetic heterogeneity—glioma cells mimic the hierarchical structure of developing cells and are in a transcriptional state determined by multiple epigenetic mechanisms; (3) tumor microenvironment (TME) heterogeneity—due to differences in stromal cells, immune cells, and molecules surrounding the tumor, gliomas exhibit diverse malignant phenotypes.

[0003] The tumor metastasis environment (TME) is a complex cellular and molecular environment that sustains glioma survival and progression and is considered an important “soil” supporting glioma malignancy. [8, 9] Increasing evidence suggests that the TME plays a key role in promoting glioma progression, treatment resistance, and recurrence. [10-12] The TME of gliomas consists of microglia, astrocytes, neurons, immune cells, pericytes, and endothelial cells, which interact with tumor cells to drive tumor malignancy.

[12] Compared with other solid tumors, gliomas have fewer infiltrating tumor lymphocytes (TILs), mainly regulatory T cells, while cytotoxic CD8 cytotoxic T lymphocytes (CTLs) are scarce and in a depleted state. [13, 14] Since CD8 CTLs play a key role in ICB-induced tumor killing, their lack makes GBM an immunologically “cold tumor” and thus poorly responsive to ICB, in stark contrast to “hot tumors” rich in active CTLs. [15-18] Despite the central role of the tumor mesenchyme (TME), current glioma classification relies primarily on the intrinsic genetic characteristics of the tumor, [1-3] which leaves a significant gap in understanding TME heterogeneity and hinders the development of effective TME-targeted therapies. A comprehensive characterization of TME regulators is crucial for reprogramming “cold” TMEs into “hot” TMEs and enhancing the efficacy of intracellular chemotherapy (ICB). Consistent with this, previous studies in this invention have shown that simultaneous blocking of CXCL8 and PD-1 can synergistically remodel the TME of GBMs from cold to hot and significantly inhibit tumor growth in vivo.

[19]

[0004] To date, most TME-related studies have focused on single components; however, since TMEs are inherently networked, a systematic analysis of their protein-protein interaction networks is needed.

[0005] References

[0006] 1.Louis, DN et al. The 2021 WHO Classification of Tumors of theCentral Nervous System: a summary. Neuro Oncol 23, 1231-1251 (2021).

[0007] 2.Berger, T. R., Wen, P. Y., Lang-Orsini, M. & Chukwueke, U. N. WorldHealth Organization 2021 Classification of Central Nervous System Tumors andImplications for Therapy for Adult-Type Gliomas: A Review. JAMA Oncol 8,1493-1501 (2022).

[0008] 3.Ostrom, Q. T. et al. CBTRUS Statistical Report: Primary Brain andOther Central Nervous System Tumors Diagnosed in the United States in 2012-2016. Neuro Oncol 21, v1-v100 (2019).

[0009] 4.Ito, H., Nakashima, H. & Chiocca, E. A. Molecular responses toimmune checkpoint blockade in glioblastoma. Nat Med 25, 359-361 (2019).

[0010] 5.Cloughesy, T. F. et al. Durable complete responses in somerecurrent high-grade glioma patients treated with Toca 511 + Toca FC. NeuroOncol 20, 1383-1392 (2018).

[0011] 6.Desjardins, A. et al. Recurrent Glioblastoma Treated withRecombinant Poliovirus. N Engl J Med 379, 150-161 (2018).

[0012] 7.Nehama, D., Woodell, A. S., Maingi, S. M., Hingtgen, S. D. & Dotti,G. Cell-based therapies for glioblastoma: Promising tools against tumorheterogeneity. Neuro Oncol 25, 1551-1562 (2023).

[0013] 8.Barker, H. E., Paget, J. T., Khan, A. A. & Harrington, K. J. Thetumour microenvironment after radiotherapy: mechanisms of resistance andrecurrence. Nat Rev Cancer 15, 409-425 (2015).

[0014] 9.Nagarsheth, N., Wicha, M. S. & Zou, W. Chemokines in the cancermicroenvironment and their relevance in cancer immunotherapy. Nat Rev Immunol17, 559-572 (2017).

[0015] 10.Friebel, E. et al. Single-Cell Mapping of Human Brain CancerReveals Tumor-Specific Instruction of Tissue-Invading Leukocytes. Cell 181,1626-1642 e1620 (2020).

[0016] 11.Klemm, F. et al. Interrogation of the Microenvironmental Landscapein Brain Tumors Reveals Disease-Specific Alterations of Immune Cells. Cell181, 1643-1660 e1617 (2020).

[0017] 12.Andersen, B. M. et al. Glial and myeloid heterogeneity in thebrain tumour microenvironment. Nat Rev Cancer 21, 786-802 (2021).

[0018] 13.Chen, D. S. & Mellman, I. Elements of cancer immunity and thecancer-immune set point. Nature 541, 321-330 (2017).

[0019] 14.Keskin, D. B. et al. Neoantigen vaccine generates intratumoral Tcell responses in phase Ib glioblastoma trial. Nature 565, 234-239 (2019).

[0020] 15.Arrieta, V. A. et al. Immune checkpoint blockade in glioblastoma:from tumor heterogeneity to personalized treatment. J Clin Invest 133 (2023).

[0021] 16.Reardon, D. A. et al. Effect of Nivolumab vs Bevacizumab inPatients With Recurrent Glioblastoma: The CheckMate 143 Phase 3 RandomizedClinical Trial. JAMA Oncol 6, 1003-1010 (2020).

[0022] 17.Medikonda, R., Dunn, G., Rahman, M., Fecci, P. & Lim, M. A reviewof glioblastoma immunotherapy. J Neurooncol 151, 41-53 (2021).

[0023] 18.McGranahan, T., Therkelsen, K. E., Ahmad, S. & Nagpal, S. CurrentState of Immunotherapy for Treatment of Glioblastoma. Curr Treat OptionsOncol 20, 24 (2019).

[0024] 19.Liu, H. et al. Neutralizing IL-8 potentiates immune checkpointblockade efficacy for glioma. Cancer Cell 41, 693-710 e698 (2023).

[0025] 20.Yoshihara, K. et al. Inferring tumour purity and stromal andimmune cell admixture from expression data. Nat Commun 4, 2612 (2013).

[0026] 21.Newman, A. M. et al. Robust enumeration of cell subsets fromtissue expression profiles. Nat Methods 12, 453-457 (2015).

[0027] 22.Sherman, B. T. et al. DAVID: a web server for functionalenrichment analysis and functional annotation of gene lists (2021 update).Nucleic Acids Res 50, W216-W221 (2022).

[0028] 23.Barbie, D. A. et al. Systematic RNA interference reveals thatoncogenic KRAS-driven cancers require TBK1. Nature 462, 108-112 (2009).

[0029] 24. Ravi, VM et al. Spatially resolved multi-omics deciphersbidirectional tumor-host interdependence in glioblastoma. Cancer Cell 40, 639-655 e613 (2022).

[0030] 25.Mootha, VK et al. PGC-1alpha-responsive genes involved inoxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 34, 267-273 (2003). Summary of the Invention

[0031] This invention targets highly malignant tumor microenvironment (TME) in IDH-labeled WT gliomas, proposing that POSTN, LTF, and PLA2G2A can be used as pathological biomarkers, while ANXA1 can be used as a serum biomarker, highlighting the potential of incorporating TME origin characteristics into diagnostic strategies.

[0032] This invention first provides biomarkers for detecting the glioma microenvironment (TME), comprising serum biomarkers and pathological biomarkers. The serum biomarkers are selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2). The pathological biomarkers are selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

[0033] The present invention also provides the use of the above-mentioned biomarkers in the preparation of kits for the diagnosis of gliomas.

[0034] In one embodiment of the invention, it comprises an antibody that specifically recognizes a serum biomarker; said serum biomarker is selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2).

[0035] In one embodiment of the invention, it includes using the antibody to detect serum biomarkers in a serum sample via ELISA technology.

[0036] In one embodiment of the invention, it comprises an antibody that specifically recognizes a pathological marker selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

[0037] In one embodiment of the invention, the sample for pathological marker detection is selected from tissue samples of human glioma patients, preferably tumor tissue from patients with isocitrate dehydrogenase (IDH) wild-type (IDH-WT) and / or IDH mutant (IDH-MU).

[0038] In one embodiment of the present invention, the prognosis and / or malignant progression of glioma are determined based on the detection results of biomarkers.

[0039] The present invention further provides a diagnostic kit for glioma, comprising antibodies for recognizing biomarkers selected from serum biomarkers and / or pathological biomarkers;

[0040] The serum biomarker is selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2);

[0041] The pathological markers are selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

[0042] The beneficial effects of the above-described technical solution of the present invention are as follows:

[0043] This invention reveals the clinical relevance of TME cell populations and identifies different subsets of pericytes, lymphocytes, and myeloid cells. Finally, in the highly malignant TME of IDH-labeled WT gliomas, this invention proposes POSTN, LTF, and PLA2G2A as pathological biomarkers, and ANXA1 as a serum biomarker. This presents an innovative approach for glioma prognosis or malignant progression diagnosis based on TME origin characteristics, providing reliable biomarkers for non-invasive diagnosis of glioma prognosis or malignant progression. Attached Figure Description

[0044] Figure 1This image presents a map of the compositional characteristics and clinical significance of the tumor microenvironment (TME) in different glioma subtypes. A is a heatmap showing the specific enrichment patterns of different glioma subtypes in terms of matrix, immunity, and total estimated score in the TCGA_GBMLGG dataset. B shows Kaplan-Meier survival analysis indicating significant survival differences in the TCGA_GBMLGG dataset after stratification by matrix, immunity, estimated score, and tumor purity threshold. C shows the significant cellular heterogeneity of the TME among different glioma subtypes in the TCGA_GBMLGG dataset. D identifies 11 major cell populations using UMAP projection of single-cell data: tumor cells (Tum), proliferating tumor cells (Pro). T cells (tum), neurons (Neu), pericytes (Peri), endothelial cells (Endo), T cells (T), macrophages (MΦ), oligodendrocytes (Oligo), monocytes (MO), dendritic cells (DC), and microglia (MG); E shows integrated scRNA data revealing TME cellular heterogeneity in different glioma subtypes; F shows representative immunohistochemical (IHC) images: MΦ-M2 (CD163), CD4 T cells (CD4), endothelial cells (CD31), and pericytes (PDGFRB) from different glioma subtypes. Corresponding statistical bar charts are provided alongside. Scale bar: 10 μm.

[0045] Figure 2 shows the ESTIMATE score atlas of glioma subtypes based on transcriptomic and proteomic data. A is a scatter plot comparing stromal score, immune score, total estimated score, and tumor purity in different pathological subtypes and grades of gliomas in the TCGA_GBMLGG dataset; B is the Kaplan-Meier survival analysis of different pathological subtypes of gliomas in the TCGA_GBMLGG cohort, stratified according to stromal score, immune score, total estimated score, and tumor purity threshold; C is the flowchart of LC-MS / MS detection of fresh frozen glioma tissue; D is the principal component analysis (PCA) of proteomic features of different pathological subtypes of gliomas; and E is the ESTIMATE score of different pathological subtypes of gliomas based on proteomic data.

[0046] Figure 3 shows the analytical atlas of TME cells in different pathological subtypes of glioma; where A is the Kaplan-Meier survival analysis of TME cell composition in the TCGA_GBMLGG cohort based on the CIBERSORT algorithm; BD are the sample number, total number of cells, and number of TME cells in the integrated analysis of single-cell transcriptome data, respectively; E is a statistical bar chart of immunohistochemical (IHC) results of neurons (TUBB3), oligodendrocytes (Oligo2), microglia (TMEM119), MΦ-M1 (CD861), and CD8 T cells (CD8) in different glioma subtypes; F is the protein level atlas of cell type-specific markers based on glioma proteome data.

[0047] Figure 4 Molecular characterization maps of TME in IDHWT and IDHMU gliomas are shown. A represents the transcriptional maps of IDHWT-179 and IDHMU-68 genes, visualized through unsupervised hierarchical clustering in different glioma subtypes. B shows the biological processes (blue) of malignant TME-related pathways (red) compared to non-malignant TME enrichment (blue) in functional enrichment analysis. C shows the top-ranked genes upregulated by IDHWT compared to IDHMU in four glioma databases (TCGA_GBMLGG, CGGA_Primary, Gravendeel, and Kamoun). D shows representative IHC images of PLA2G2A, POSTN, and LTF in IDHWT and IDHMU samples. Corresponding statistical bar charts are provided alongside. Scale bar: low magnification. Figure 10 μm, inset 40 μm; E shows 127 proteins from the WT-179 gene validated by label-free quantitative LC-MS / MS and functionally annotated by GOBP; F shows 46 proteins from the MU-68 gene validated by label-free quantitative LC-MS / MS.

[0048] Figure 5The distribution maps of WT-179 and MU-68 characteristic genes in different glioma subtypes are shown. A and B represent the ssGSEA scores of WT-179 and MU-68 characteristics in glioma pathological features, respectively. C shows H&E staining images of glioma sections and AUCell scores of WT-179 and MU-68 characteristics based on spatial transcriptome data. D shows the prognostic stratification of IDHWT (left) and IDHMU (right) samples based on gene characteristic scores, using the TCGA_GBMLGG dataset. E and F represent the scRNA-seq spatial mappings of WT-179 and MU-68 characteristics in different TME cell types, respectively. G and H represent the UMAP projections of 11 major cell populations in IDHWT and IDHMU samples after integrating scRNA data. I and J represent the UMAP maps of cell distribution of WT-179 and MU-68 characteristics in IDHWT and IDHMU gliomas based on ssGSEA scores.

[0049] Figure 6 shows the TME characteristics and functional association map dependent on IDH mutations; where A and B are Venn plots, showing 179 IDHWT upregulated genes and 68 IDHMU upregulated genes identified by consensus in the TCGA_GBMLGG, CGGA_Primary, Gravendeel, and Kamoun databases, respectively; C is DAVID enrichment analysis showing that the WT-179 gene is associated with collagen tissue and immune response; D is Kaplan-Meier survival analysis of PLA2G2A, POSTN, and LTF in the TCGA_GBMLGG dataset; E shows the Kaplan-Meier survival analysis of WT-179 and MU-68 feature scores in the TCGA_GBMLGG dataset; F shows the Kaplan-Meier survival analysis of MU-68 feature scores for the IDHMU sample in the TCGA_GBMLGG dataset; G shows the correlation heatmap between WT-179 and MU-68 scores and ESTIMATE scores; H shows the GSEA analysis plot, which indicates that WT-179 feature scores and corresponding protein scores are enriched in the Verhaak mesenchymal type, while MU-68 feature scores and corresponding protein scores are enriched in the Verhaak preneural type.

[0050] Figure 7Analysis of pericytes in different pathological subtypes of glioma; where A is a t-SNE view of pericyte clusters; B is the average proportion of pericyte subsets in the IDHWT (WT), MU-Non (M), and MU-Cod (MC) glioma subtypes; C is a heatmap showing the preference of different pericyte subsets in the WT, MN, and MC glioma subtypes (Ro / e score); D is the Kaplan-Meier survival curves stratified by the abundance of different pericyte subsets (ECM, Proliferation, PTPRZ1, PGF, and Transport subsets) in the TCGA_GBMLGG dataset. Significance was determined by log-rank test; E is the t-SNE score of pericyte subsets based on different gene characteristics; F is the distribution of pericyte subsets along the pseudo-time trajectory in IDHWT, MU-Non and MU-Cod gliomas; G is the pseudo-time-related changes of pericyte subsets based on different gene characteristics; H is the pseudo-time trajectory of WT-179, MN-30 and MC-83 characteristics in pericyte subsets.

[0051] Figure 8 shows the characterization of pericyte subpopulations in gliomas of different pathological subtypes. A is a dot plot showing the expression of marker genes in different pericyte clusters; (B) is a t-SNE view of pericyte clusters in different glioma subtypes; C is a stacked bar chart showing the proportion of pericyte clusters in IDHMU and IDHWT gliomas estimated using proteomic data; D is a box plot showing the AUCell scores of the WT-179, MU-68, MN-30, and MC-83 features of different pericyte subpopulations in IDHWT, IDHMU, MU-Non, and MU-Cod gliomas; E and G are Monocle2-based pericyte subpopulation trajectory analyses, corresponding to IDHWT (E), MU-Non (F), and MU-Cod gliomas (G), respectively; H is a heatmap showing the dynamic expression patterns of the top 100 genes in the pericyte subpopulations along the time-fitted trajectory, and the GOBP enrichment results of the corresponding gene clusters.

[0052] Figure 9A) Analytical atlas of myeloid cells in different pathological subtypes of glioma; A) t-SNE view of myeloid cell clusters; B) Mean proportion of myeloid cell subsets in different glioma subtypes; C) Heatmap showing the preference of different myeloid cell subsets in different subtypes (Ro / e score); D) Kaplan-Meier survival curves stratified by abundance of different myeloid cell subsets in the TCGA_GBMLGG dataset. Significance was determined by log-rank test; E) t-SNE scores of myeloid cell subsets based on different gene characteristics; F) Distribution of MO and MΦ subsets along the pseudochronic trajectory in IDHWT, MU-Non, and MU-Cod gliomas; G) Pseudochronic correlation changes of MO and MΦ subsets based on different gene characteristics; H) Pseudochronic trajectory of MO and MΦ subsets colored by subtype characteristic gene scores; I) Distribution of MG subset along the pseudochronic trajectory; J) Pseudochronic correlation changes of MG subsets based on different gene characteristics; K) Pseudochronic trajectory of MG subsets colored by subtype characteristic gene scores.

[0053] Figure 10 This study analyzes the characteristics of myeloid cell subpopulations in different histological subtypes of gliomas. A shows a heatmap of marker gene expression in different myeloid cell clusters; B shows a stacked bar chart of the proportion of myeloid cell clusters in IDHMU and IDHWT gliomas estimated based on proteomics data; C shows Kaplan-Meier survival analysis using the TCGA_GBMLGG database, stratified according to the abundance of different myeloid cell subpopulations. Statistical significance was determined using the log-rank test; D shows trajectory analysis of the MO and MΦ subpopulations using Monocle2. The right-hand heatmap shows the dynamic expression patterns of the top 100 genes ranked along pseudo-time and the corresponding enriched GOBP annotations for gene clusters; E shows trajectory analysis of the MG subpopulation using Monocle2. The right-hand heatmap shows the dynamic expression patterns of the top 100 genes ranked along pseudo-time and the corresponding enriched GOBP annotations for gene clusters.

[0054] Figure 11This section presents an analysis of lymphocytes in different pathological subtypes of gliomas. A shows a t-SNE view of lymphocyte clusters; B shows the average proportion of lymphocyte subsets in different glioma subtypes; C is a heatmap showing the preference of different lymphocyte subsets in each subtype (Ro / e score); D is the Kaplan-Meier survival curve stratified by the abundance of different lymphocyte subsets in the TCGA_GBMLGG cohort, with significance determined by the log-rank test; E shows the t-SNE scores of lymphocyte subsets based on different gene characteristics; F shows the distribution of CD4 T cell subsets along a pseudochronological trajectory in IDHWT, MU-Non, and MU-Cod gliomas; G shows the pseudochronological correlation changes of CD4 T cell subsets based on different gene characteristics; H shows the pseudochronological trajectory of CD4 T cell subsets colored by subtype characteristic gene scores; I shows the distribution of CD8 T cell subsets along a pseudochronological trajectory in IDHWT, MU-Non, and MU-Cod gliomas; J shows the distribution of CD8 T cell subsets along a pseudochronological trajectory based on different gene characteristics. Timing-related changes in T cell subsets; K represents the timing trajectory of CD8 T cell subsets colored according to subtype characteristic gene scores.

[0055] Figure 12 shows the atlas of lymphocyte subset characteristics in different histological subtypes of glioma. A is a dot plot of marker gene expression in different lymphocyte clusters; B is a stacked bar chart of the proportion of lymphocyte clusters in IDHMU and IDHWT gliomas estimated based on proteomics data; C is a Kaplan-Meier survival analysis plot stratified according to the abundance of different lymphocyte subsets in the TCGA_GBMLGG database, with statistical significance determined by the log-rank test; D is a trajectory analysis atlas of CD4 T cell subsets using Monocle2, with the right-hand heatmap showing the dynamic expression patterns of the top 100 genes ordered along pseudo-time and the corresponding enriched GOBP annotations of gene clusters; E is a trajectory analysis atlas of CD8 T cell subsets using Monocle2, with the right-hand heatmap showing the dynamic expression patterns of the top 100 genes ordered along pseudo-time and the corresponding enriched GOBP annotations of gene clusters.

[0056] Figure 13This is a map showing the correlation between circulating TME biomarkers in non-invasive diagnosis and prognosis. A is a volcano plot showing elevated levels of ANXA1, ANXA2, MMP14, and LTBP2 in the serum of IDHwt and IDHmu gliomas, serving as serum biomarkers. B is a functional enrichment map of elevated genes in IDHWT vs. IDHMU gliomas, showing association with collagen remodeling and angiogenesis. C is a Kaplan-Meier curve, indicating that high levels of biomarker expression are associated with poorer survival. D is an ELISA result map validating the expression of ANXA1, ANXA2, MMP14, and LTBP2 in the serum of healthy donors and patients with IDHwt or IDHmu gliomas. E is a schematic diagram of TME subtypes: PM-TME (elevated pericyte-derived factors, enriched collagen and angiogenesis) and NM-TME (elevated neuronal-derived factors, maintaining neural homeostasis).

[0057] Figure 14 This is a graph for validating circulating TME biomarkers and analyzing transcript-protein correlations. A shows statistical information of glioma patients included in this study who underwent serum LC-MS / MS analysis; B is a principal component analysis (PCA) graph of proteomics features, distinguishing between IDHwt and IDHmu serum samples; C is a Venn plot showing serum proteomics results and characteristic genes WT-179, MU-68, MN-30, and MC-83; D is a comparison of ESTIMATE scores for IDHWT and IDHMU gliomas based on serum proteomics data; E is a box plot of ANXA1, ANXA2, MMP14, and LTBP2 transcript levels for IDHWT and IDHMU gliomas in the TCGA_GBMLGG database; and F is a Pearson correlation analysis graph showing serum protein levels and WT-179 gene scores. Detailed Implementation

[0058] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0059] Unless otherwise specified, all reagents used in this embodiment are of analytical grade, and the progress of all chemical reactions is detected by thin-layer chromatography.

[0060] Example 1: Glioma TME exhibits IDH subtype-specific cellular and molecular heterogeneity, which is correlated with clinical outcomes and prognostic characteristics.

[0061] Compared to IDH-mutant (IDH-tag MU) gliomas, IDH-wild-type (IDH-tag WT) gliomas typically exhibit higher malignancy. However, current diagnostic criteria primarily emphasize tumor cell characteristics, while the understanding of TME composition and its clinical relevance remains insufficient. To address this issue, this invention employs multiple integrative methods to systematically characterize the TME heterogeneity of different histological subtypes of gliomas. This invention integrates 46 glioma samples (including 36 glioblastomas (GBM), 4 astrocytomas, and 6 oligodendrogliomas), covering self-tested samples and publicly available data. Single-cell transcriptome data analysis was performed using the Seurat software workflow (v5.1.0). During the quality control phase, cells expressing fewer than 3 genes, cells expressing fewer than 200 genes or more than 9,000 genes, and cells with a mitochondrial gene proportion exceeding 20% ​​were excluded. Potential doublets were identified and removed using the DoubletFinder tool. Subsequently, the Harmony algorithm was used to correct for batch effects between samples. After obtaining high-quality data, principal component analysis (PCA) was performed on the top 3,000 hypervariable genes using the RunPCA function in Seurat, and the global distribution of the major cell populations was visualized using UMAP. Simultaneously, the t-SNE method was used to analyze the fine structure within subpopulations. Cell type marker genes were identified using the FindAllMarkers function. Finally, the inferCNV method was used to infer copy number variations (CNVs) to distinguish between malignant and non-malignant cells.

[0062] This invention first used the ESTIMATE scoring system (R package estimate, v1.0.13)

[20] to evaluate the clinical relevance of TME features in the TCGA_GBMLGG dataset. In 20 IDH-labeled WT gliomas, this invention observed that their stromal, immune, and ESTIMATE scores were significantly elevated, while the tumor purity was lower than that of IDH-labeled MU gliomas ( Figure 2 (A) suggests that TME plays a crucial role in the malignant progression of gliomas. In OD and AC, different grades did not significantly affect these scores ( Figure 2 (A). Hierarchical cluster analysis shows that IDH status is the dominant factor determining TME characteristics ( Figure 1 (A). Survival analysis showed that higher stromal, immune, and ESTIMATE scores were associated with worse overall survival, while higher tumor purity predicted better prognosis. Figure 1(B) Stratified analysis further revealed that these correlations were particularly significant in IDH-labeled WT gliomas, where all three scores predicted adverse outcomes; while in IDH-labeled MU gliomas, only the stromal score had prognostic value (B). Figure 2 (B)

[0063] To assess the translational potential of TME-related molecular features, this invention performed LC-MS / MS proteomics analysis on freshly cryopreserved glioma tissues (n=18). Figure 2 (C) A total of 18 freshly frozen glioma tissues were collected (9 IDH wild-type, 9 IDH mutant, 4 IDU mutant astrocytomas, and 5 oligodendrogliomas). 50 μg of protein extract from each sample was used for reduction, alkylation, and trypsin digestion. The specific steps were as follows: reduction was performed by treatment with 10 mM dithiothreitol (DTT) at 56 ℃ for 30 min; alkylation was then performed by reaction with 20 mM iodoacetamide (IAA) under light-protected conditions for 30 min. Subsequently, sequencing-grade trypsin (Promega) was added at an enzyme-to-substrate ratio of 1:50, and digestion was carried out at 37 ℃ for 16 h. The resulting peptides were then... C18 Solid Phase Extraction Column (Waters Sep-Pak®) After desalination, freeze-dry for later use.

[0064] Chromatographic separation in Vanquish Horizon UHPLC System (Thermo Scientific) The chromatography was performed using a C18 reversed-phase column (ACQUITY UPLC BEH, 1.7 μm, 2.1 × 150 mm, Waters). Mobile phase A was 0.1% formic acid aqueous solution, and mobile phase B was 0.1% formic acid acetonitrile solution. A linear gradient elution program was used: phase B increased from 5% to 35% over 60 min, the flow rate was 300 nL / min, and the column temperature was maintained at 40℃.

[0065] Mass spectrometry analysis was performed on a Q Exactive™ HF-X Hybrid Quadrupole–Orbitrap™ mass spectrometer (Thermo Scientific) in positive ion mode. The full scan range was 350–1600 m / z with a resolution of 60,000. Data-dependent acquisition (DDA) was then performed on the top 20 precursor ions using high-energy collisional dissociation (HCD) with a normalized collision energy of 28% and a dynamic exclusion time of 30 s.

[0066] Raw data were analyzed using MaxQuant (v2.1.0) software, searching the UniProt human proteome database. Parameters were set as follows: false positive rate (FDR) 1%. Cysteine ​​carboxymethylation Set as a fixed modifier. Methionine oxidation was set as a variable modification; Label-Free Quantification (LFQ) was enabled. The model was used. Statistical analysis was performed using Perseus (v1.6.15.0) and GraphPad Prism (v9.0). Quality control steps included retention time calibration and internal standard detection (Pierce™ Retention Time Calibration Kit). Principal component analysis (PCA) showed a clear separation between IDH-labeled WT and IDH-labeled MU tumors. Figure 2 (Middle D), supporting the robustness of the data. Proteomics results further validated that IDH-labeled WT gliomas had higher stromal, immunological, and ESTIMATE scores, while having lower tumor purity ( Figure 2 (E). Subsequently, this invention used CIBERSORT

[21] to analyze the immune cell composition in glioma TME. The results showed that IDH-labeled WT gliomas were enriched with M2 macrophages (MΦ-M2), M0 macrophages (MΦ-M0) and CD4 T cells, while monocyte (MO), B cell and plasma cell infiltration were reduced ( Figure 1 (C) Kaplan-Meier survival analysis showed that high levels of MΦ-M2, MΦ-M0, CD4 T cells, NK cells, and neutrophils predicted poorer survival, while increases in MO, B cells, and plasma cells were associated with better prognosis. Figure 3 (A)

[0067] To construct a comprehensive single-cell atlas of glioma TME, this invention performed scRNA-seq on 3 OD samples and snRNA-seq on 6 GBM samples, and integrated published datasets, including OMIX920 (4 GBM, 3 AC), GSE182109 (22 GBM, 3 OD, 1 AC), and GSE84465 (4 GBM), totaling 46 gliomas (36 GBM, 4 AC, 6 OD). Figure 3(Middle B). Analysis of 46 glioma samples was performed using the Seurat package (v5.1.0). During quality control, genes expressed in fewer than 3 cells, cells expressing fewer than 200 genes or more than 9,000 genes, and cells with a mitochondrial gene proportion exceeding 20% ​​were excluded. Potential doublets were identified and removed using the DoubletFinder tool. Subsequently, the Harmony algorithm was used to correct for batch effects between samples. After obtaining high-quality data, principal component analysis (PCA) was performed on the top 3,000 hypervariable genes using the RunPCA function in Seurat, and UMAP was used to visualize the global distribution of the major cell populations. The t-SNE method was used to analyze the fine structure within subpopulations. Cell type marker genes were identified using the FindAllMarkers function. Finally, the inferCNV method was used to infer copy number variation (CNV) to distinguish malignant and non-malignant cells. After quality control, this invention analyzed 192,083 GBM cells (including 86,125 TME cells), 27,843 AC cells (including 20,085 TME cells), and 55,017 OD cells (including 30,648 TME cells). Figure 3 (C and D). UMAP analysis divided gliomas into 11 distinct cell clusters, including tumor cells, proliferating tumor cells, macrophages, monocytes, microglia, dendritic cells, T cells, endothelial cells, pericytes, neurons, and oligodendrocytes (Oligo). Figure 1 In the middle D and E). Immunohistochemical (IHC) analysis showed that, compared with AC and OD, GBM had a significant increase in MΦ-M2, CD4 T cells, endothelial cells, and pericytes. Figure 1 (Middle F). Other TME cell types did not show significant variation across different histological subtypes ( Figure 3 (Middle E). LC-MS / MS further validated these results, showing that IDH-labeled WT gliomas showed higher expression of pericyte, MΦ-M2, endothelial cell, and CD4 T cell markers, while neuronal and Oligo markers were expressed at lower levels. Figure 3 (F). In summary, the research of this invention demonstrates that glioma TME exhibits subtype-specific cellular and molecular heterogeneity, with IDH status being a decisive factor. The composition of TME is closely related to clinical outcomes, making TME-derived cellular and proteomic features potential prognostic biomarkers and providing new targets for glioma stratification and treatment.

[0068] Example 2: IDH mutation status defined the pro-malignant gene network of TME-related genes in IDH-tagged WT gliomas, while IDH-tagged MU gliomas exhibited non-malignant characteristics.

[0069] Heterogeneous cellular components in glioma mesenchymal stem cells (TMEs) perform diverse functions through complex protein networks utilizing paracrine, autocrine, and adjacent signaling pathways. However, the specific protein profiles generated by these different cellular components and their combined impact on the cellular and molecular structure of TMEs have not been fully assessed. Given the significant heterogeneity of TME characteristics across different histological subtypes of gliomas, this invention systematically investigates the molecular composition of TME diversity in different glioma subtypes. To establish a comprehensive panel of TME-related proteins, this invention integrates several TME-related gene clusters from the UniProt database and incorporates entries from GeneOntology (GO) for "Extracellular Regions" (GO:0005576), "Extracellular Space" (GO:0005615), "Extracellular Matrix" (GO:0031012), and keywords for "Secretion" (KW_0964), "Cytokines" (KW_0202), and "Growth Factors" (KW_0339). Through this integration method, 3808 non-redundant protein-coding genes (not shown) were finally obtained, representing TME-related genes.

[0070] To identify TME genes associated with IDH mutation status, this invention systematically analyzed gene expression patterns in IDH-labeled WT and IDH-labeled MU tumors across all histopathological grades (grades 2, 3, and 4) using two independent cohorts: TCGA_GBMLGG and CGGA_Primary. This analysis included cross-grade comparisons and grade-stratified analysis. Venn plot analysis revealed a set of conserved gene tags that were consistently upregulated in IDH-labeled WT gliomas and remained relatively low in IDH-labeled MU gliomas. Further validation was performed by stratifying by IDH status using the Gravendeel and Kamoun datasets, followed by differential expression analysis of IDH-labeled WT and IDH-labeled MU subtypes. Under the criteria of P < 0.05 and FoldChange > 1, this invention identified 179 consensus genes upregulated in IDH-labeled WT gliomas (WT-179) and 68 genes elevated in IDH-labeled MU gliomas (MU-68). Figure 6 (A, B). Hierarchical cluster analysis showed that the WT-179 and MU-68 genes exhibited different expression patterns in their respective histological subtypes. Figure 4A). Functional enrichment analysis using DAVID (Database for Annotation, Visualization, and Integrated Discovery)

[22] showed that the WT-179 gene was significantly associated with collagen fiber tissue (GO:0030199), innate immune response (GO:0045087), extracellular matrix tissue (GO:0030198), angiogenesis (GO:0001525), ECM-receptor interaction (KEGG:hsa04512) and PI3K-AKT signaling pathway (KEGG:hsa04151) ( Figure 4 China B and Figure 6 (C). Conversely, the MU-68 gene tends to be enriched in neuronal and synaptic regulatory pathways (C). Figure 4 (B)

[0071] By integrating four glioma databases, this invention found that PLA2G2A, POSTN, and LTF ranked highly in the WT-179 gene panel. Figure 4 (C). This invention further validated IHC protein levels using 18 IDH-labeled MU and 20 IDH-labeled WT clinical glioma specimens. Figure 4 The statistical evaluation of IHC results by the Chinese Institute for Drugs and Research (DI) reinforced the effectiveness of the gene screening strategy, showing that the protein expression of PLA2G2A, POSTN, and LTF in IDH-tagged WT samples was significantly higher than that in IDH-tagged MU samples. Figure 4 (D). Furthermore, all three genes can predict poor survival in glioma patients in the TCGA_GBMLGG database. Figure 6 (D). To further verify whether these gene products could be detected, this invention analyzed proteomic data from fresh glioma tissue. The results showed that the protein products of 127 genes in the WT-179 gene were detectable, and their expression trends were consistent with the transcriptional level, and were higher in IDH-tagged WT gliomas than in IDH-tagged MU samples (D). Figure 4 (E). Functional enrichment analysis of these 127 proteins showed consistency with the original WT-179 gene tag. Figure 4 PLA2G2A, POSTN, and LTF were also detected as upregulated proteins in IDH-tagged WT samples. Figure 4 In IDH-tagged MU gliomas, 46 out of 68 MU-68 genes showed consistency at the protein-transcriptional level. Figure 4 (F). In summary, these data define subtype-related marker genes with detectable protein levels in glioma samples.

[0072] Example 3: Single-cell RNA analysis revealed the contribution of different cells to the TME characteristics of glioma histological subtypes.

[0073] The present invention uses the ssGSEA algorithm (GSVA R package (v1.52.3)) to calculate the scores of the WT-179 gene set and the MU-68 gene set respectively.

[23]

[0074] The results showed that the WT-179 gene score of the IDH-tagged WT samples was significantly higher than that of the IDH-tagged MU samples, while the MU-68 gene score showed the opposite trend. Figure 5 (A and B). In IDH-labeled WT gliomas, samples histologically diagnosed with GBM had higher WT-179 gene scores, while MU-68 gene scores were lower than those of gliomas not diagnosed with GBM. Figure 5 (A and B). However, within the same histological glioma subtype, the two gene scores did not show significant changes across different grades of samples (A and B). Figure 5 A and B). Spatial transcriptome dataset analysis

[24] also supports this result, that is, the IDH-labeled WT sample WT-179 has a higher score, while the IDH-labeled MU sample MU-68 has a higher score ( Figure 5 (C) Furthermore, higher WT-179 gene scores were associated with poorer survival, while higher MU-68 gene scores predicted better survival, a finding validated in multiple glioma databases. Figure 6 (E). After isolating IDH-tagged WT and IDH-tagged MU gliomas, this invention noted that a higher WT-179 gene score also predicted poor survival in glioma patients. Figure 5 (Middle D); while the predictive significance of the MU-68 gene score is only evident in IDH-labeled MU gliomas ( Figure 6 (F). The WT-179 gene score was strongly positively correlated with the matrix score, immune score, and ESTIMATE score, but negatively correlated with tumor purity. Figure 6 (G). Conversely, the MU-68 gene score and the WT-179 score showed an inverse relationship (G). Figure 6 G). Gene set enrichment analysis (GSEA)

[25] further showed that elevated scores of WT-179 gene or WT-127 protein significantly enriched Verhaak glioblastoma stromal features, while elevated scores of MU-68 gene or MU-46 protein preferentially enriched Verhaak preneuronal features ( Figure 6 (H).

[0075] This invention also utilized integrated scRNA-seq data for cell localization analysis to assess the contribution of different cell types to the molecular characteristics of TME. The WT-179 tag gene was highly expressed primarily in pericytes, tumor cells, and macrophages. Figure 5In contrast, the MU-68 tag gene is specifically enriched in neurons and oligodendrocytes (E); Figure 5 (F). Subsequently, this invention analyzed the cell populations of IDH-tagged WT and IDH-tagged MU gliomas, finding that IDH-tagged WT gliomas showed increased numbers of pericytes, macrophages, monocytes, and T cells, while oligodendrocytes were reduced (F). Figure 5 (G and H). Using the WT-179 tag gene, this invention scores each cell using the AUCell method to quantify its tag gene enrichment level. In the UMAP map, cells are colored according to their AUC scores; the more yellow the color, the higher the enrichment level of the WT-179 tag gene. Results showed that the WT-179 tag gene is mainly enriched in specific tumor subgroups and glioma-associated pericytes (G and H). Figure 5 In the middle I), macrophages and endothelial cells also showed enrichment of tag genes ( Figure 5 (I). Using the same strategy, the MU-68 tag gene was analyzed, and the results showed that it was mainly enriched in tumor subsets and oligodendrocytes, while being almost unexpressed in other cell populations. Figure 5 (J). Notably, the score distribution in the UMAP plot is highly consistent with the cellular distribution of the tag genes in the heatmap. Therefore, these data highlight the increased contribution of TME cells in IDH-tagged WT gliomas, while also demonstrating the crucial role of tumor cells in the TME structure of IDH-tagged MU gliomas.

[0076] Example 4: Pericyte subset heterogeneity in gliomas links IDH mutation status with patient prognosis and malignant progression.

[0077] Single-cell analysis revealed that pericytes play a significant role in malignant progression of glioma tissue metastasis (TME). Using integrated scRNA-seq data from three glioma histological subtypes, this invention extracted pericytes and performed re-cluster analysis.

[0078] Pericyte subpopulations were extracted and clustered again, and the Harmony algorithm was used to correct for batch effects between samples. Principal component analysis (PCA) was performed on the first 3,000 hypervariable genes using the RunPCA function in Seurat, and the internal heterogeneity of the cell subpopulations was analyzed using the t-distributed stochastic neighbor embedding (t-SNE) method.

[0079] Based on specific marker gene expression ( Figure 8 In the middle A), pericytes were annotated into five functionally distinct subsets: ECM, PGF, PTPRZ1, Transport, and Proliferation. Figure 7China A, Figure 8 (Middle B). During the transition from IDHWT to MU-Non and MU-Cod gliomas, the pericyte composition changes significantly: the ECM and Proliferation subsets gradually decrease, while the Transport subset gradually increases and becomes dominant in MU-Cod gliomas. Figure 7 (B) Ro / e analysis further confirmed that the Transport subgroup was enriched in MU-Cod, the PTPRZ1 subgroup was enriched in MU-Non, and the ECM / Proliferation subgroup was enriched in IDHWT glioma ( Figure 7 (C). Bayes Prism deconvolution analysis using our laboratory's proteomic data also supports the increase in ECM pericytes and the decrease in Transport pericytes in IDHWT gliomas, compared to IDHMU gliomas ( Figure 8 (C). Using bulk RNA-seq data from TCGA_GBMLGG to estimate the proportions of each subpopulation, it was found that they were associated with survival: the ECM, Proliferation, and PTPRZ1 subpopulations were associated with poorer prognosis, while the PGF and Transport subpopulations predicted better survival. Figure 7 (D). AUCell scores showed that the WT-179 tag was most enriched in ECM pericytes, and also significantly enriched in the PGF subset, while the MN-30 score was low in all subsets. The MU-68 and MC-83 genes were lowly expressed in all subsets. Figure 7 ECM pericytes showed higher WT-179 scores than IDHMU in IDHWT gliomas, but lower MU-68 and MC-83 scores. Among all subpopulations, ECM pericytes showed the highest WT-179, MU-68, and MC-83 scores, while PGF pericytes showed the highest MN-30 score. Figure 8 (D). Pseudo-temporal trajectory analysis using Monocle2 showed that the trajectories originated from PTPRZ1 and PGF pericytes. In IDHWT gliomas, the trajectories differentiated towards the ECM and concentrated in PGF pericytes; in MU-Non gliomas, the trajectories were divided into ECM- and Transport-dominated branches; in MU-Cod gliomas, the trajectories were simpler, mainly consisting of Transport pericytes. Figure 7 China F and Figure 8 Transcriptional analysis along the pseudo-time sequence showed that IDHWT pericytes initially enriched in ion-responsive pathways, followed by cell adhesion, matrix tissue, and angiogenesis; MU-Non pericytes were initially enriched in angiogenesis, cell adhesion, and IGF receptor signaling; MU-Cod pericytes showed decreased translation and matrix tissue, but enhanced adhesion inhibition and decreased smooth muscle proliferation. Figure 8(H). These data indicate that pericytes possess subtype-specific transcriptional signatures. Tag gene scoring trajectory analysis further confirmed the dynamic changes of WT-179, MN-30, and MC-83 in pericyte subsets (H). Figure 7 (Internal GH): In IDHWT, WT-179 is elevated in ECM pericytes but decreased in PGF; in MU-Non, MN-30 decreases with transport development but increases after ECM formation; in MU-Cod, pericytes contribute the least to MC-83. Overall, IDHWT gliomas show greater heterogeneity in pericyte composition, while IDHMU gliomas are enriched in transport pericytes, suggesting that this subtype has enhanced transport or barrier-related functions.

[0080] Example 5: Specific subpopulation distribution of myeloid cells in glioma histological subtypes and their impact on clinical prognosis.

[0081] Further analysis based on specific marker genes identified fourteen myeloid cell subsets:

[0082] Myeloid subpopulations were extracted and clustered again, and the Harmony algorithm was used to correct for batch effects between samples. Principal component analysis (PCA) was performed on the first 3,000 hypervariable genes using the RunPCA function in Seurat, and the internal heterogeneity of cell subpopulations was analyzed using the t-distributed stochastic neighbor embedding (t-SNE) method, including microglia (MG, Activation, Homeostatic, IFN, Inflammation, Lipid), monocytes (MO, Inflammation, Resting), macrophages (MΦ, Hypoxic, Inflammation, Metal, Resident), dendritic cells (DC, cDC2), neutrophils, and proliferating cells (…). Figure 9 China A and Figure 10 (A). IDHWT gliomas are significantly enriched in the MG-Activation, MG-Hypoxic, MG-Inflammation, and MG-Resident subsets, while IDHMU gliomas have a relatively simple myeloid composition. In IDHMU gliomas, MU-Non gliomas are characterized by MG-Inflammation enrichment, while MU-Cod gliomas show higher MG-Homeostatic and neutrophil levels. Figure 9 (B). Further deconvolution of proteomic data based on BayesPrism data revealed that the proportion of MΦ in IDHWT gliomas was higher than that in IDHMU gliomas ( Figure 10 (B), Ro / e analysis confirmed the distribution preference of myeloid subtypes in different histological subtypes of glioma ( Figure 9 (C). Survival analysis showed subtype-specific prognostic value: MG-IFN and MG-Lipid were associated with poor prognosis, while MG-Activation and MG-Homeostatic predicted better survival. Figure 9 (Middle D). In the MΦ subpopulation, Inflammation, Hypoxic, and Resident were associated with poor survival, while the Metal subpopulation predicted improved survival ( Figure 9 (Middle D). Similarly, MO-Inflammation was associated with shortened survival, MO-Resting with prolonged survival, and DC-cDC2 with better prognosis ( Figure 10 (C). Regarding transcriptional tags, MΦ-Hypoxic, MΦ-Inflammation, and MO-Inflammation mainly contribute to the WT-179 score, while most MG subsets are associated with the MN-30 score; myeloid cells contribute limitedly to the MU-68 and MC-83 tags. Figure 9 (E).

[0083] Quasi-time trajectory analysis of MΦ and MO:

[0084] The monocle (v2.32.0) R package was used to perform pseudo-time trajectory analysis on myeloid cell subpopulations to explore the cell state transition process. The differentially expressed genes (DEGs) on the pseudo-time trajectory were identified using the differentialGeneTest function, and gene clusters and expression heatmaps sorted by pseudo-time were generated using the plot_pseudotime_heatmap function. The top 100 differentially expressed genes in the pseudo-time trajectory were selected, clustered according to expression dynamics, and GO Biological Process (GO BP) enrichment analysis was performed using the clusterProfiler (v4.12.6) R package.

[0085] Furthermore, the dynamic relationship between pseudotime and characteristic gene sets (179, 30, and 83 genes) was further evaluated, covering the changing trends of microglia, macrophages, and monocytes in three glioma subtypes (IDH-WT, IDH-MU-Noncodel, IDH-MU-Codel, and IDH-MU), revealing their dynamic reprogramming.

[0086] In IDHWT gliomas, MΦ-Hypoxic and MO-Inflammation predominate in the early stages, differentiating along two trajectories: one maintaining the original composition, and the other showing an increase in MΦ-Resident and MΦ-Inflammation. Figure 9 China F and Figure 10 (D). In MU-Non gliomas, the trajectory originates from MO-Inflammation, MΦ-Resident, and MΦ-Inflammation, with one branch enriched in MΦ-Resident and the other developing towards the MΦ-Hypoxic and MΦ-Metal subsets. Conversely, MU-Cod gliomas originate from MO-Resting and MO-Inflammation, developing towards pure MΦ-Resident or mixed MΦ-Resident and MΦ-Hypoxic subsets, although the estimated cell number is lower. Functional enrichment of MΦ and MO showed that hypoxia, anti-apoptotic signaling, and chemotaxis were weakened in IDHWT gliomas; chemotaxis and angiogenesis were weakened in MU-Non gliomas, while lipid metabolism was enhanced; immunomodulation was suppressed in MU-Cod gliomas, but extracellular response pathways were activated ( Figure 10 (Middle D). Along these trajectories, the WT-179 score of IDHWT tumors decreased, suggesting a reduced contribution of MO-Inflammation to TME, while the MN-30 score of MU-Non gliomas remained stable, and the MC-83 score of MU-Cod gliomas gradually increased ( Figure 9 (G and H). MG subgroup trajectory analysis revealed more complex variations ( Figure 9 China I and Figure 10(E). In IDHWT gliomas, MG-Lipid, MG-IFN, and MG-Homeostatic predominate early, followed by a decrease in MG-Lipid and MG-IFN, while an increase in MG-Homeostatic and MG-Activation, eventually leading to MG-Activation dominating. In MU-Non gliomas, MG-Homeostatic and MG-Activation initially dominate, followed by an increase in MG-Activation and MG-Lipid, eventually shifting to MG-Inflammation and MG-IFN dominance. Similarly, MU-Cod gliomas initiate with MG-Homeostatic and MG-Activation, undergo a transition with increases in MG-IFN, MG-Activation, and MG-Lipid, ultimately converging into the MG-Inflammation and MG-Lipid subsets. MG functional enrichment showed enhanced antigen processing and presentation in IDHWT gliomas, but weakened innate immune responses; weakened chemotaxis and angiogenesis in MU-Non gliomas, but increased lipid metabolism; and suppressed immunomodulation in MU-Cod gliomas, but activated extracellular response pathways. Figure 10 (Middle E). Along the MG trajectory, the WT-179 score of IDHWT gliomas gradually decreased, while the MN-30 and MC-83 scores of MU-Non and MU-Cod remained stable. Figure 9 (J and K). Overall, IDHWT and IDHMU gliomas exhibit different myeloid composition, developmental dynamics, functional status, transcriptional programs, and prognostic associations. Specific myeloid subgroups have opposite prognostic effects, and pseudo-time series analysis reveals differences in differentiation and evolutionary trajectories in TMEs of different glioma subtypes.

[0087] Example 6: TILs exhibit specific composition, developmental trajectory, and prognostic role in glioma subtypes.

[0088] Using integrated scRNA-seq data from three glioma histological subtypes, this invention identifies 11 distinct lymphocyte subsets based on specific marker genes. Figure 12 (A)

[0089] Lymphoid subpopulations were extracted and clustered again, and the Harmony algorithm was used to correct for batch effects between samples. Principal component analysis (PCA) was performed on the first 3,000 hypervariable genes using the RunPCA function in Seurat. The internal heterogeneity of cell subpopulations, including CD4 T cells (Tn-LEF1, Tcm-GPR183, Treg-FOXP3), CD8 T cells (proliferating, Pre-Tex-HAVCR2, IFN, Teff-CX3CR1, Tem-GZMK), NK cells (XCL1, FCGR3A), and B cells (…), was analyzed using t-distributed stochastic neighbor embedding (t-SNE). Figure 11 (A)

[0090] Subtype-specific enrichment patterns were observed: CD4 T cell infiltration was significantly increased in IDHWT gliomas, predominantly Tcm-GPR183, while CD8-TEM-GZMK levels were also elevated. Figure 11 MU-Non gliomas showed a significantly higher proportion of B cells, while MU-Cod gliomas were enriched in the NK-FCGR3A and CD8-Teff-CX3CR1 subsets. Figure 11 (B). Ro / e analysis further validated these distributional preferences ( Figure 11 (C). To cross-validate these findings, this invention applied the BayesPrism deconvolution algorithm to proteomic data and compared IDHWT and IDHMU gliomas. Although the cell proportions inferred from protein data did not perfectly align with scRNA results, IDHWT gliomas still showed an increase in CD4 T, CD8-IFN, and CD8-Pre-Tex-HAVCR2 cells, while CD8-Teff-CX3CR1 cells continued to decrease (C). Figure 12 (B) Kaplan-Meier survival analysis showed that all CD4 subsets were associated with shortened survival. Among the CD8 subsets, proliferative, Pre-Tex-HAVCR2, and IFN cells predicted poor prognosis, while the Tem-GZMK and Teff-CX3CR1 subsets were associated with improved survival. Figure 11 (D). Furthermore, elevated levels of NK and B cell infiltration predict prolonged survival in glioma patients ( Figure 12 (C). Analysis combining these results with subtype-specific gene tags showed that lymphocytes contributed limitedly to the WT-179, MU-68, and MC-83 tags, but significantly to the expression of the MN-30 tag gene. Figure 11 (E).

[0091] CD4 T cell pseudo-temporal trajectory analysis showed that Tn-LEF1 cells represented the naive origin, suggesting glioma-associated naive CD4 T cell infiltration. In IDHWT and MU-Non gliomas, the trajectory branched into the Tcm-GPR183 and Treg-FOXP3 lineages; while in MU-Cod gliomas, one branch developed towards Tcm-GPR183, and the other branch terminated into a mixed lineage (…). Figure 11 China F and Figure 12 (Middle D). Analysis also showed an increased proportion of Treg-FOXP3 in IDHWT gliomas, but almost none in MU-Cod gliomas. Functional enrichment showed enhanced Tcm-GPR183 cell proliferation in IDHWT gliomas, while Treg-FOXP3 cells were associated with cell motility ( Figure 12 In MU-Non gliomas, chemotaxis and T cell regulation processes were enriched, but cell motility was not significant; in MU-Cod gliomas, terminally differentiated CD4 cells were enriched in cytotoxicity and killing functions, suggesting that CD4 T cells may have an inhibitory role in this subtype. Figure 12 (D). Trajectory analysis of tag scores showed that in IDHWT gliomas, the WT-179 score increased from Tn-LEF1 to Tcm-GPR183, but decreased slightly in Treg-FOXP3 cells; in MU-Non gliomas, the MN-30 score remained relatively stable during CD4 differentiation, while in MU-Cod gliomas, the MC-83 score decreased in Tcm-GPR183 compared to Tn-LEF1, supporting the limited contribution of CD4 T cells to the MC-83 tag. Figure 11 G and H). CD8 T cell trajectories show more complex patterns ( Figure 11 China I and Figure 12 (E). In IDHWT gliomas, the trajectory involved a mixture of Pre-Tex-HAVCR2, IFN, Teff-CX3CR1, and Tem-GZMK subsets, ultimately converging into proliferating cells; in MU-Non gliomas, the trajectory initiated from IFN and Tem-GZMK cells, branching into Teff-CX3CR1 and proliferating lineages, with no Pre-Tex-HAVCR2 cells observed; in MU-Cod gliomas, the trajectory began from Tem-GZMK cells, branching into one branch containing IFN and Teff-CX3CR1 cells, and the other terminating in proliferating cells. Transcriptome analysis showed upregulation of genes related to cell proliferation and cytotoxicity in IDHWT gliomas ( Figure 12 In the two subtypes of IDHMU, genes related to cell killing were significantly upregulated along the pseudo-temporal sequence (E); Figure 12(E). Analysis of tag gene scores in CD8 pseudo-time series showed that proliferating and Pre-Tex-HAVCR2 cells contributed significantly to IDHWT-specific TME characteristics, while in MU-Non gliomas, the 30-gene score remained stable across all CD8 subsets; MU-Cod gliomas showed a prominent contribution from initiating Tem-GZMK cells to MU-Cod-specific TME tags. Figure 11 (J and K). Overall, these results depict the heterogeneous composition, developmental trajectory, functional status, and prognostic role of tumor-infiltrating lymphocytes in different glioma subtypes. Lymphocyte subsets exhibited subtype-specific enrichment, lineage dynamics, and functional diversity, and contributed differently to glioma transcriptional tags, highlighting the complex interaction between the immune microenvironment and tumor phenotype in glioma progression.

[0092] Example 7: Serum proteomic characteristics reflect IDHWT TME status and predict glioma prognosis

[0093] This invention performed LC-MS / MS-based serum proteomic analysis on glioma patients (n=14) (a total of 14 glioma patients had serum samples, including 4 IDH wild-type and 10 IDH mutant patients). Figure 14 Principal component analysis (PCA) of the proteomic profile showed a significant separation between IDHWT GBM and IDHMU LGG (AC and OD). Figure 14 (B) indicates the detectability of TME subtype-specific tag genes in liquid biopsy.

[0094] Quantitative proteomics:

[0095] 50 μg of protein extract from each sample was used for reduction, alkylation, and trypsin digestion. The specific steps were as follows: first, reduction was performed by treatment with 10 mM dithiothreitol (DTT) at 56℃ for 30 min; then, alkylation was performed by reaction with 20 mM iodoacetamide (IAA) under light-protected conditions for 30 min. Subsequently, sequencing-grade trypsin (Promega) was added at an enzyme-to-substrate ratio of 1:50, and digestion was carried out at 37℃ for 16 h. The resulting peptides were then... C18 Solid Phase Extraction Column (WatersSep-Pak®) After desalting, freeze-dry for later use. Chromatographic separation is performed in... Vanquish Horizon UHPLC System (Thermo Scientific) Mass spectrometry was performed using a C18 reversed-phase column (ACQUITY UPLC BEH, 1.7 μm, 2.1 × 150 mm, Waters). Mobile phase A was 0.1% formic acid aqueous solution, and mobile phase B was 0.1% formic acid acetonitrile solution. A linear gradient elution program was used: phase B increased from 5% to 35% over 60 min, with a flow rate of 300 nL / min and the column temperature maintained at 40 °C. Mass spectrometry analysis was performed on a Q Exactive™ HF-X Hybrid Quadrupole–Orbitrap™ mass spectrometer (Thermo Scientific) in positive ion mode. The full scan range was 350–1600 m / z, with a resolution of 60,000. Data-dependent acquisition (DDA) was then performed on the top 20 precursor ions using high-energy collisional dissociation (HCD) with a normalized collision energy of 28% and a dynamic exclusion time of 30 s. Raw data were analyzed using MaxQuant (v2.1.0) software, searching the UniProt human proteome database. Parameters were set as follows: false positive rate (FDR) 1%. Cysteine ​​carboxymethylation Set as a fixed modifier. Methionine oxidation was set as a variable modification; Label-Free Quantification (LFQ) was enabled. The statistical analysis was performed using Perseus (v1.6.15.0) and GraphPad Prism (v9.0). Quality control procedures included retention time calibration and internal standard detection (Pierce™ Retention Time Calibration Kit).

[0096] The results showed that 81 out of 179 WT-179 tag genes (45.2%) were detectable in serum, significantly higher than the detection rates of the IDHMU (19.1%), Nonco (26.7%), and Codel (19.3%) subtypes. Figure 14 (C). However, compared with the proteome of glioma samples, the serum proteome failed to reflect the difference in ESTIMATE scores between IDHWT and IDHMU gliomas (C). Figure 14 (D). Validation of IDHWT-related proteins showed that four candidate proteins were significantly upregulated in IDHWT serum compared to IDHMU serum (P<0.01, ≥2-fold). Figure 13(A). Pathway enrichment analysis showed that these 81 genes, especially these four validation proteins, functionally clustered in the collagen remodeling and angiogenesis pathways. Figure 13 The presence of B indicates biological similarity. Analysis of the TCGA_GBMLGG database shows that these four genes are expressed higher in IDHWT than in the two IDHMU subtypes ( Figure 14 (E). Kaplan-Meier survival analysis showed its prognostic value, with patients with high expression having significantly worse survival (E). Figure 13 (C). Significant positive correlations were found between the four genes and the WT-179 gene score. Figure 14 (Middle F).

[0097] Example 8

[0098] Serum proteomic analysis of IDHwt on TDHmu revealed that ANXA1, ANXA2, MMP14, and LTBP2 were significantly upregulated in IDHwt compared to the serum of IDHmu glioma patients (P<0.01, ≥2-fold), and can serve as serum biomarkers. Figure 13 (A). Analysis of the GBMLGG database grouped ANXA1, ANXA2, MMP14, and LTBP2 expression levels by median. Results showed that patients with high expression had significantly worse survival. ELISA was used to verify the expression of ANXA1, ANXA2, MMP14, and LTBP2 in the serum of healthy donors and patients with IDHwt or IDHmu gliomas. Results showed that ANXA1, ANXA2, MMP14, and LTBP2 were all lowest expressed in healthy donors, followed by IDHMU, with the highest expression in IDHWT (A). Figure 13 (D). Analysis using the TCGA_GBMLGG database showed that the expression levels of ANXA1, ANXA2, MMP14, and LTBP2 in IDHWT were higher than those of the other two IDHMU isotypes. Figure 14 (E).

[0099] This invention performed ELISA detection on these biomarkers:

[0100] The protein levels of ANXA1, ANXA2, LTBP2, and MMP14 in human plasma were detected using ELISA kits. The experimental procedure was strictly performed according to the kit instructions, including sample loading, incubation, washing, substrate addition for color development, and reaction termination. Finally, the absorbance was measured on a microplate reader to calculate the protein concentration. All ELISA antibodies used were purchased from MLbio (Shanghai, China) and included: ANXA1 (catalog number: ml038038), ANXA2 (catalog number: ml038035), MMP14 (catalog number: ml026258), and LTBP2 (catalog number: ml028628).

[0101] Orthogonal validation results showed that these proteins were significantly elevated in glioma serum compared to normal serum, and even higher in IDHWT serum than in IDHMU serum. Figure 13 The results (D) were highly consistent with LC-MS / MS quantitative results. Overall, this study establishes circulating protein characteristics as a non-invasive diagnostic biomarker reflecting the properties of IDHWT TME, with potential clinical application value for molecular subtyping and prognostic prediction.

[0102] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A biomarker for detecting the glioma microenvironment (TME), comprising a serum biomarker and a pathological biomarker, wherein the serum biomarker is selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2); and the pathological biomarker is selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

2. Use of the biomarker as described in claim 1 in the preparation of a kit for the diagnosis of glioma.

3. The use as described in claim 2, comprising an antibody that specifically recognizes a serum biomarker; said serum biomarker is selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2).

4. The use as described in claim 3, comprising using the antibody to detect serum biomarkers in serum samples via ELISA technology.

5. The use as described in any one of claims 2-4, comprising an antibody that specifically recognizes a pathological marker selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

6. The use as described in claim 5, wherein, The samples used for pathological marker detection are tissue samples from human glioma patients, preferably tumor tissue from patients with wild-type isocitrate dehydrogenase (IDH) (IDH-WT) and / or IDH mutant (IDH-MU).

7. The use as described in any one of claims 2-6, for determining the prognosis and / or malignant progression of glioma based on the detection results of biomarkers.

8. A diagnostic kit for glioma, comprising an antibody for recognizing a biomarker selected from serum biomarkers and / or pathological biomarkers; The serum biomarker is selected from one or more of annexin A1 (ANXA1), annexin A2 (ANXA2), matrix metalloproteinase 14 (MMP14), and potential transforming growth factor β-binding protein 2 (LTBP2); The pathological markers are selected from one or more of periostin (POSTN), lactoferrin (LTF), and phospholipase A2 group IIA (PLA2G2A).

Images