A biomarker for spinal cord injury prognosis evaluation severity diagnosis and application thereof

Abstract

The present application relates to the technical field of biotechnology, in particular to a biomarker for spinal cord injury prognosis evaluation severity diagnosis and application thereof. The present application firstly discovers and establishes that ITGA6 protein secreted by astrocytes can be used as a biomarker for SCI. More importantly, the ITGA6 protein is not a general injury indicator, but a "functional" indicator that specifically reflects the activation state of the endogenous "protection and repair" mechanism (i.e. the endothelial protection effect mediated by the pi3k / akt pathway of astrocytes). Detecting the indicator changes the traditional idea of evaluating "how severe the injury is" to a new diagnostic dimension of evaluating "how strong the body's protective response is". This is not previously thought by those skilled in the art.

Description

Technical Field

[0001] This invention relates to the field of biotechnology, and in particular to a biomarker for assessing the severity of spinal cord injury prognosis and its application. Background Technology

[0002] Spinal cord injury (SCI) is a highly disabling central nervous system disorder that places a heavy burden on patients, families, and society. The pathological process of SCI is complex, encompassing both primary and secondary mechanical injury. Primary injury is usually irreversible, while the cascading pathological reactions it triggers—secondary injury—are the main cause of the continued deterioration of neurological dysfunction and represent a critical window for clinical intervention. Secondary injury involves a series of complex pathophysiological changes, including inflammatory responses, ischemia and hypoxia, oxidative stress, neuronal apoptosis, and disruption of the blood-spinal cord barrier (BSCB).

[0003] In secondary spinal cord injury (SCI), the imbalance of neurovascular unit homeostasis, particularly the disruption of the blood-spinal cord barrier (SPB), is a core factor exacerbating the injury. The SPB is primarily composed of cerebral microvascular endothelial cells and their tight junctions, and its integrity is crucial for maintaining the stability of the spinal cord microenvironment. Following injury, endothelial cell death and dysfunction lead to increased barrier permeability, allowing harmful substances to infiltrate and exacerbating local edema, inflammation, and neuronal death. Therefore, protecting vascular endothelial cells and maintaining the integrity of the SPB is considered an important strategy for improving the prognosis of SCI.

[0004] Astrocytes, the most numerous glial cells in the central nervous system, exhibit significant reactive proliferation (i.e., "glial scarring") after spinal cord injury (SCI) and have long been considered a major obstacle to the inhibition of axonal regeneration. However, increasing research indicates that reactive astrocytes possess significant heterogeneity and a dual role. In the acute phase of injury, they can secrete various neurotrophic factors and cytokines, participating in beneficial processes such as limiting inflammation, rebuilding the blood-spinal barrier, and protecting neurons. Among these, signaling communication between astrocytes and vascular endothelial cells is crucial for regulating angiogenesis and repairing barrier function.

[0005] The main shortcomings of existing technologies are as follows: 1. Unclear understanding of the molecular mechanisms of secondary injury following spinal cord injury (SCI): Although the academic community recognizes the importance of astrocyte-endothelial cell interaction in the pathological process of SCI, the key signaling molecules and specific regulatory pathways mediating this interaction, particularly the "endothelial protection" function, have not been fully elucidated. Existing research largely focuses on some known, broad-spectrum growth factors (such as VEGF), lacking in-depth exploration of specific and critical protective pathways in the injury microenvironment. This prevents this invention from precisely understanding the initiation and failure processes of endogenous repair mechanisms after injury at the molecular level.

[0006] 2. Lack of precise prognostic biomarkers: Currently, the diagnosis and prognostic assessment of SCI patients mainly rely on clinical neurological function scores (such as the ASIA score) and imaging examinations (such as MRI). These methods can macroscopically assess the anatomical structure of the injury and the degree of functional impairment, but they have the following limitations: Insufficient sensitivity and specificity: Imaging results do not always perfectly match the patient's final functional recovery.

[0007] 3. Unable to reflect microscopic pathological conditions: It cannot dynamically and in real time reflect the pathophysiological changes at the cellular level in the damaged area, such as the activation status of astrocytes, the damage and repair potential of endothelial cells, etc.

[0008] Limited prognostic predictive capabilities: It is difficult to accurately predict a patient's recovery potential in the early stages of injury, and it cannot provide precise molecular evidence for the development of personalized treatment plans.

[0009] 4. Limited Guiding Significance of Existing Molecular Markers: Some SCI biomarkers currently under study, such as neurofilament light chains (NFL) and glial fibrillary acidic protein (GFAP), are mostly general markers reflecting neural or glial cell damage. They can indicate the presence and approximate extent of damage, but cannot specifically reflect the protective repair status of neurovascular units, especially whether the key repair process of astrocyte-mediated endothelial protection has been effectively activated. Summary of the Invention

[0010] To overcome the shortcomings of the prior art, this invention provides a biomarker for assessing the severity of spinal cord injury prognosis and its application.

[0011] This invention is achieved through the following technical solution: In a first aspect, the present invention provides a biomarker for the diagnosis of the severity of spinal cord injury prognosis, wherein the biomarker is ITGA6 protein, which exerts an endothelial protective effect by activating the pi3k / akt pathway on vascular endothelial cells.

[0012] Secondly, this invention provides the application of biomarkers in the preparation of products for diagnosing the prognostic assessment of spinal cord injury severity.

[0013] Furthermore, the product is any one of a detection reagent, a kit, a microarray, or a biochip.

[0014] Furthermore, when the product is a reagent kit, the diagnostic reagent kit includes, but is not limited to, the following components: ITGA6 protein capture antibody-coated microplates: 96-well or other sizes of ELISA plates, the inner walls of which are coated with mouse-derived anti-human ITGA6 protein monoclonal antibody. Detection antibodies: Rabbit-derived anti-human ITGA6 protein polyclonal or monoclonal antibodies labeled with biotin or HRP that can bind to different epitopes of the ITGA6 protein.

[0015] Standards: A series of lyophilized powders or solutions of recombinant human ITGA6 protein with known concentration gradients, used to plot standard curves.

[0016] Enzyme conjugate: horseradish peroxidase-labeled streptavidin; Colorimetric substrate solution: Solution A and Solution B or single-component TMB solution; Termination solution: Used to terminate the colorimetric reaction;

[0017] Washing solution: Used to wash away unbound substances in micropores; Sample diluent: Used to dilute test samples and standards.

[0018] The ITGA6 protein capture antibody is ABWAYS CY5509.

[0019] The terminating solution is a 2M sulfuric acid solution.

[0020] Thirdly, the present invention provides a drug that targets the biomarker, wherein the drug contains estradiol.

[0021] Compared with the prior art, the present invention has the following significant advantages: 1. Novel target and high specificity: For the first time, ITGA6 protein secreted by astrocytes is used as a biomarker for SCI. It specifically reflects the status of key protective repair pathways in the neurovascular unit, rather than general nerve damage, and has higher information value.

[0022] 2. High sensitivity and quantification: Using immunological detection methods such as ELISA, it is possible to accurately quantify ITGA6 protein in samples at the picogram (pg) level, providing objective data for clinical evaluation.

[0023] 3. High clinical application value: It provides a brand-new molecular subtyping tool for the early diagnosis, dynamic monitoring, prognosis and personalized treatment of SCI, and helps to screen patients suitable for receiving therapies that promote the function of ITGA6 protein / pi3k / akt pathway, thus achieving precision medicine.

[0024] 4. Simple operation and easy to promote: The reagent kit has a standardized operation process, low requirements for equipment and personnel, and is easy to be widely used in the laboratory departments of hospitals at all levels. Attached Figure Description

[0025] The invention will now be further described with reference to the accompanying drawings.

[0026] Figure 1 Distribution of differentially expressed genes in a volcano map; Figure 2 The Venn diagram was used to analyze the differentially expressed genes and identify 90 common differentially expressed genes (DE-ARGs) associated with spinal cord injury from the known spinal cord injury-related genes. Figure 3 Scatter plot of enrichment analysis of DE-ARGs in the three dimensions of GO; Figure 4 Bubble plot of KEGG pathway enrichment for DE-ARGs; Figure 5 This represents the quality distribution of the data before and after filtering (nFeature, nCount, percent.mt). Figure 6 For screening of hypervariable genes and PCA analysis, among which Figure 6 A represents high-variance gene screening (HVG). Figure 6 B and C represent 3C / D: PCA dimensionality reduction and statistical validation; Figure 7 This is a diagram showing cell clustering and annotation analysis; among which... Figure 7 A is the clustering resolution evaluation graph. Figure 7 B is the UMAP clustering analysis diagram. Figure 7 C represents the marker gene annotation diagram. Figure 8 This is a graph showing the removal of twin cells and the percentage of cells; among them, Figure 8 A and B are clustering comparison diagrams before and after twin removal. Figure 8 C is a graph showing the distribution ratio of twins in each cell subpopulation. Figure 8 D represents the final percentage of each type of cell after removing twins; Figure 9 A diagram illustrating differentially expressed genes and their functions in astrocytes; Figure 9 A is, Figure 9 B represents GO functional enrichment. Figure 9 C represents KEGG pathway enrichment; Figure 10 This is a graph showing the functional enrichment analysis of target genes. Figure 10 A represents GO enrichment analysis. Figure 10 B represents KEGG enrichment analysis; Figure 11 A screening diagram for target genes in astrocyte-related diseases; Figure 12 A visualization map of the protein-target interaction (PPI) network; Figure 13 Graphs for building and evaluating machine learning models; Figure 13 A represents a comparison of residual box plots. Figure 13 B represents the inverse cumulative distribution of residuals. Figure 13 C represents the ROC curve performance evaluation; Figure 14 This is a diagram validating the efficacy of Hub gene screening and diagnosis. Figure 14 A represents Hub gene screening. Figure 14 B represents the diagnostic efficacy verification. Figure 14 C represents the validation of expression differences; Figure 15 This is a diagram showing the analysis of immune cell infiltration. Figure 16 For the analysis of differences and correlations in immune infiltration, among which Figure 16 A represents the box plot of the immune cell scoring. Figure 16 B is a correlation heatmap; Figure 17 This is a drug prediction and molecular docking diagram, in which... Figure 17 A represents the drug-biomarker relationship network. Figure 17 B represents the predicted docking binding site; Figure 18 For GSEA and GSVA functional enrichment analysis, among which Figure 18 A represents GSEA enrichment analysis. Figure 18 B represents GSVA enrichment analysis; Figure 19 For single-cell follow-up analysis: pseudo-timing and cell communication, among which Figure 19 A represents a quasi-time series analysis. Figure 19 B represents cell communication analysis; Figure 20 To validate the expression of biomarkers in single cells, among which Figure 20 A represents a bubble chart. Figure 20 B represents a diagram depicting a violin; Figure 21 To validate biomarkers for spinal cord injury; Figure 22 To validate the expression level of ITGA6 protein in the model group Figure 23A is CCK8. Figure 23 B represents cell scratches; Figure 24 For cell scratch assay; Figure 25 The protein expression diagrams of PI3K, P-AKT, and SRC protein pathways in each group; Figure 26 Schematic diagram of the ITGA6 protein double antibody sandwich ELISA kit; Figure 27 Quantitative performance evaluation of the ITGA6 detection kit and its application in spinal cord injury; (A) Linear fitting standard curve for ITGA6 protein detection, with the horizontal axis representing the protein concentration of ITGA6 standard and the vertical axis representing the absorbance (OD value) at 450 nm. The scatter plot and the fitted line show a high positive linear correlation between the two within the given concentration range, confirming that the detection system has excellent quantitative analysis capabilities and detection reliability; (B) Scatter box plot of ITGA6 expression levels in clinical samples from multiple groups of subjects. The figures compare the ITGA6 concentration distribution in samples from the healthy control group (HC), the poor prognosis group (PP), and the good prognosis group (GP). The results show that the baseline ITGA6 concentration in the SCI patient group is significantly higher than that in the healthy control group. More importantly, within the SCI patient cohort, the ITGA6 expression level in the good prognosis group (GP) is significantly higher than that in the poor prognosis group (PP). An asterisk (*) indicates a highly statistically significant difference between the groups. Figure 28 A graph showing the ranking and selection of the importance of target genes for astrocyte-related diseases using the best predictive classifier; Figure 29 To illustrate the expression differences of the TOP10 genes in disease and control samples based on GSE151371, box plots were generated using the ggplot2 package in R language, and Wilcoxon rank-sum tests were performed between groups. ROC curves of the TOP10 genes in GSE151371 were plotted using the pROC package in R language. Detailed Implementation

[0027] The core technical concept of this invention is based on the following groundbreaking discovery: In the pathological microenvironment of spinal cord injury, reactive astrocytes secrete ITGA6 protein by upregulating the expression of the ITGA6 gene. This ITGA6 protein, as a key signaling molecule, exerts a significant endothelial protective effect by activating the pi3k / akt pathway on vascular endothelial cells, including inhibiting endothelial cell apoptosis, promoting the expression of tight junction proteins, and maintaining the integrity of the blood-spinal cord barrier. Therefore, the level of ITGA6 protein in patient biological samples (such as cerebrospinal fluid or blood) can specifically and quantitatively reflect the activation intensity of this endogenous protective mechanism, thus becoming a highly valuable biomarker for the diagnosis and prognosis of spinal cord injury (SCI).

[0028] The inventiveness of this invention does not lie in the ELISA detection method itself (which is a known technique), but in the following three core aspects: (a) Novel biomarker targets: Existing SCI markers (such as GFAP and NFL) primarily reflect the degree of “damage” to nerve or glial cells and are a type of damage indicator.

[0029] The innovation of this invention lies in the first-ever discovery and establishment of the ITGA6 protein secreted by astrocytes as a biomarker for SCI (Small Incision Cognition). More importantly, the ITGA6 protein is not a general indicator of damage, but a "functional" indicator that specifically reflects the activation state of endogenous "protection and repair" mechanisms (i.e., endothelial protection mediated by astrocytes through the XX pathway). Detecting this indicator shifts the traditional approach of assessing "the severity of damage" to a completely new diagnostic dimension of assessing "the strength of the body's protective response." This was something that those skilled in the art had not previously considered.

[0030] (II) Novel Clinical Application Objectives and Values: Existing technologies are mainly used for damage identification and approximate classification.

[0031] The innovation of this invention lies in directly linking the detection results of ITGA6 protein with the clinical prognosis of SCI patients. This invention pioneers the proposition that the level of ITGA6 protein can predict a patient's potential for functional recovery. High levels of ITGA6 protein indicate the activation of a strong intrinsic vascular protective mechanism, potentially leading to a better prognosis; conversely, low levels indicate a poorer prognosis. This provides unprecedented molecular evidence for clinicians to make accurate prognostic assessments in the early stages of injury and to develop individualized treatment strategies (e.g., for patients with low ITGA6 protein levels, drugs activating the XX pathway could be considered).

[0032] (III) Overall concept of the technical solution: The overall concept of this invention is to transform a novel molecular mechanism discovery from basic research (ITGA6 protein protects endothelial cells through the pi3k / akt pathway) into a concrete, industrially viable, and clinically applicable diagnostic product (kit). This kit is not merely a simple combination of several reagents, but a complete and rationally designed technical solution based on the aforementioned innovative discovery, successfully bridging the gap between "molecular mechanism research" and "clinical diagnostic applications."

[0033] The invention will now be further described with reference to the accompanying drawings.

[0034] Example 1

[0035] Differentially expressed gene screening (SCI vs Control) This embodiment first compared the transcriptomic data of spinal cord injury samples and normal control samples. By setting strict thresholds, more than 3,000 differentially expressed genes were identified. This volcano plot clearly shows the distribution of these genes, with most significantly differentially expressed genes concentrated on both sides of the plot, showing obvious up-regulation or down-regulation trends.

[0036] Filtering criteria settings Set a strict threshold: |log2(multiple change)|>1 and the corrected P-value (Padj)<0.05; like Figure 1 (Orange dots represent upregulated genes, green dots represent downregulated genes, and gray dots represent genes with no significant difference.) As shown, this volcano plot visually displays the distribution of 3002 differentially expressed genes between the spinal cord injury samples and the control group, of which 1276 genes are upregulated and 1726 genes are downregulated.

[0037] Example 1

[0038] Screening of DE-ARGs in SCI To narrow the scope of the study, this embodiment compared the differentially expressed genes with known spinal cord injury-related genes, identifying 90 common genes. This means that these genes not only showed differences in the experiments of this embodiment but have also been confirmed by previous studies to be related to spinal cord injury, and they are the focus of subsequent research in this invention.

[0039] Analysis Method: Intersection Analysis Strategy Differentially expressed genes (DEGs1) were compared with reported sets of spinal cord injury-related genes (ARGs) to identify common genes between the two sets, such as... Figure 2As shown, Venn diagram analysis successfully identified 90 intersection genes (DE-ARGs). These genes have been confirmed to be involved in the pathological process of spinal cord injury and are the focus of subsequent functional studies. Furthermore, these 90 DE-ARGs not only exhibit differential expression characteristics but are also directly related to the pathological mechanisms of spinal cord injury, laying the foundation for subsequent drug target screening.

[0040] GO functional enrichment analysis of DE-ARGs This embodiment analyzed the functions of these 90 key genes. GO enrichment analysis showed that they are mainly involved in processes such as angiogenesis, cell migration, and proliferation, which are crucial steps in tissue repair after spinal cord injury. This suggests that these genes may play an important role in the repair mechanism after injury.

[0041] More specifically, to gain a deeper understanding of the core biological functions involved by DE-ARGs and to clarify their potential mechanisms of action in tissue repair, this step employs GO functional enrichment analysis, such as... Figure 3 As shown, Figure 3 The study showed the top 5 most significant enriched items at three levels: BP (biological processes), CC (cellular components), and MF (molecular functions). GO enrichment analysis revealed that DE-ARGs are mainly involved in biological processes related to tissue repair, such as angiogenesis and endothelial cell migration (the genes are significantly enriched in processes such as VEGF signaling pathway, angiogenesis, endothelial cell migration and proliferation, and are directly related to tissue repair and regeneration).

[0042] KEGG pathway enrichment analysis of DE-ARGs To further explore the signaling pathways involved by these genes, this invention performed KEGG pathway enrichment analysis. Figure 4 The study showcased the 15 most significantly enriched pathways, providing a clear view of the functional distribution characteristics of the genes. KEGG pathway analysis revealed that DE-ARGs were significantly enriched in key signaling pathways that regulate cell survival and proliferation, such as HIF-1 and PI3K-Akt.

[0043] KEGG pathway analysis revealed that these genes are primarily involved in important cellular signaling pathways such as HIF-1 and PI3K-Akt, which play a central role in regulating cell survival, proliferation, and metabolism. This further validates the importance of the genes screened in this embodiment in the pathological process of spinal cord injury.

[0044] Example 3: Single-cell RNA sequencing analysis (DEGs2) Single-cell data quality control and standardization The first step in single-cell data analysis is data quality control. In this example, the raw data was rigorously filtered to remove low-quality cells, ultimately yielding over 60,000 high-quality single-cell data points, laying a solid foundation for subsequent cell clustering and analysis.

[0045] Using the GSE162610 single-cell dataset as the original input, with ≥6000 genes, ≥30000 molecules, and ≤10% mitochondrial gene content, 65,832 cells and 23,255 genes were retained. Figure 5 As shown, through rigorous quality control, low-quality cells were filtered out, ultimately retaining 65,832 high-quality single-cell data for subsequent analysis. This embodiment effectively removed low-quality data through a strict quality control process, laying a solid foundation for subsequent cell clustering and differential expression analysis.

[0046] Hypervariable gene screening and PCA analysis After data quality control, this embodiment first screened out 2000 genes with the greatest expression differences; these highly variable genes form the basis for subsequent analysis. Next, through principal component analysis (PCA), this invention found that the top 30 principal components could well represent the overall characteristics of the data, providing strong support for the next step of cell clustering.

[0047] like Figure 6 As shown in Figure A, 2000 hypervariable genes were screened using the VST (Variance Stabilizing Transformation) method, and low-expression noise was removed, laying a data foundation for subsequent PCA dimensionality reduction and cell clustering. Figure 6 As shown in B and C, PCA analysis reveals that the first 30 principal components contain the main variability information of the data. The permutation test (right) validates the significance of the first 30 PCs, ensuring the reliability of the dimensionality reduction results.

[0048] Analysis conclusion: High-variance gene screening and PCA dimensionality reduction effectively extracted the core features of the data, supporting subsequent cell clustering analysis.

[0049] Cell clustering and annotation Cells were divided into 15 populations using unsupervised clustering, and then annotated into 13 different cell types, including astrocytes and microglia, using known marker genes. This step allowed the invention to clearly visualize the complex cellular composition of the spinal cord tissue.

[0050] like Figure 7 As shown in Figure A, the clustering effect at different resolutions was evaluated using Sankey diagrams, and a resolution of 0.2 was ultimately selected for further analysis. Figure 7As shown in Figure B, the UMAP dimensionality reduction algorithm was used to cluster the cells into 15 different cell clusters, visually presenting the cell grouping results, such as... Figure 7 As shown in C, based on marker gene expression profile analysis, 15 cell clusters were successfully annotated as 13 known cell types.

[0051] Twin cell removal and cell percentage analysis To improve data quality, this embodiment identified and removed twin cells from the data, making the boundaries of cell types clearer. The final cell percentage map shows the distribution of various cell types in damaged and normal samples.

[0052] like Figure 8 As shown in A and B, calculations show that approximately 7.5% of twin cells were removed, effectively improving the accuracy and purity of cell clustering. Figure 8 As shown in Figure C, the bar chart visually illustrates the difference in the ratio of twins (red) to single cells (cyan) in different cell types, helping to identify high-risk groups. Figure 8 As shown in Figure D, the stacked plot presents the true composition and relative abundance of various cell types in the Control and SCI samples after removing interference.

[0053] Data quality control conclusions: Double cell removal significantly reduces data noise, ensuring the reliability of subsequent differential analysis and cell type identification.

[0054] Analysis of differentially expressed genes and their functions in astrocytes This embodiment focuses on the analysis of astrocytes, identifying 609 differentially expressed genes. Functional enrichment analysis revealed that these genes are mainly involved in processes such as angiogenesis and extracellular matrix interactions, revealing the potential role of astrocytes in damage repair.

[0055] like Figure 9 As shown in Figure A, 609 differentially expressed genes (DEGs2) in astrocytes were screened, laying the foundation for subsequent functional studies. Figure 9 BGO functional enrichment showed that differentially expressed genes are mainly involved in biological processes such as angiogenesis and cell migration, revealing cellular functional characteristics. Figure 9 The C KEGG pathway enrichment showed significant enrichment in key pathways such as the HIF-1 signaling pathway and extracellular matrix receptor interaction.

[0056] Target gene functional enrichment analysis Functional analysis of the 10 target genes identified showed that they are mainly involved in key processes such as angiogenesis and wound healing, and are enriched in important signaling pathways such as HIF-1, further verifying the core role of these genes in spinal cord injury.

[0057] like Figure 10 As shown in Figure A, the 10 target genes are mainly involved in biological processes such as angiogenesis and wound healing. Figure 10 BKEGG enrichment analysis showed that it was significantly enriched in the HIF-1 signaling pathway and the extracellular matrix receptor interaction pathway.

[0058] Key findings: Enrichment analysis results further validated the key biological functions and regulatory pathways of these target genes in spinal cord injury repair.

[0059] Example 4: Identification of target genes in astrocyte-related diseases To identify the core disease-regulating genes, this invention integrated differentially expressed genes from astrocytes with previously identified spinal cord injury-related genes. Ultimately, this invention identified 10 key target genes that are likely to play crucial roles in the occurrence and development of spinal cord injury.

[0060] Intersection analysis was performed between differentially expressed genes in astrocytes (DEGs2) and differentially expressed genes associated with spinal cord injury (DE-ARGs) to identify common features, such as... Figure 11 As shown, Venn diagram analysis identified 10 common target genes, including key molecules such as CD44, TIMP1, and VIM. By integrating differentially expressed genes in astrocytes with DE-ARGs, 10 core target genes for astrocyte-related diseases were ultimately identified.

[0061] Target gene protein interaction (PPI) network To investigate the interactions among 10 target genes, a protein-protein interaction network was constructed to reveal the regulatory mechanisms. The protein-protein interaction network constructed in this example reveals complex cooperative regulatory relationships among the 10 target genes. Figure 12 (The network contains 10 nodes and 42 edges, showing complex mutual regulation and synergistic relationships between genes.)

[0062] As shown in the figure, the network contains 10 nodes and 42 edges, revealing complex mutual regulation and synergistic relationships among genes.

[0063] Example 5: Machine Learning for Biomarker Screening This invention utilizes various machine learning algorithms to construct predictive models and comprehensively evaluates their performance. The results show that the generalized linear model (glm) performs best in distinguishing between damaged and normal samples, making it the optimal choice for subsequent analysis in this invention.

[0064] Multi-model construction strategy Random forest (RF), extreme gradient boosting tree (XGBTree), and generalized linear model (GLM) were constructed to distinguish between spinal cord injury and normal samples.

[0065] Multidimensional performance evaluation Compared with AUC, residuals, and inverse cumulative distribution, the glm model exhibits the highest accuracy and lowest prediction error, demonstrating the best overall performance.

[0066] Results Visual Verification The ROC curve shows that the glm model is closest to the top left corner, which intuitively proves that its predictive ability is significantly better than other models.

[0067] Key conclusions: Figure 13 As shown, by comparing the ROC curves and residuals of various machine learning models, the generalized linear model (glm) was determined to be the optimal prediction model. Its high AUC value and low residual characteristics prove that it is the best model for distinguishing sample types, which will serve as the basis for subsequent analysis.

[0068] Hub gene screening and diagnostic efficacy validation Using a machine learning model, this invention identified nine core hub genes. Further validation analysis showed that five of these genes possessed good diagnostic efficacy, and their expression levels differed significantly between injured and healthy samples. Therefore, this invention considers these five genes to be potential biomarkers for spinal cord injury.

[0069] like Figure 14 As shown in Figure A, based on the glm model to calculate feature importance, nine hub genes with the greatest influence on spinal cord injury were screened, laying the foundation for biomarkers. Through machine learning screening and ROC validation, five genes, including ANXA2 and ITGA6, were identified as potential biomarkers for spinal cord injury with diagnostic value.

[0070] like Figure 14 As shown in Figure B, ROC curve analysis revealed that the AUC values ​​of five genes, including ANXA2 and ITGA6, were all ≥0.7, demonstrating excellent clinical diagnostic potential. Through machine learning screening and ROC validation, these five genes were identified as potential biomarkers for spinal cord injury with diagnostic value.

[0071] like Figure 14 As shown in Figure C, the box plot further confirms that there are highly significant differences in the expression levels of these five genes between the injury group and the control group. Through machine learning screening and ROC validation, five genes, including ANXA2 and ITGA6, were identified as potential biomarkers for spinal cord injury with diagnostic value.

[0072] Conclusion: The five genes ANXA2, ITGA6, CXCL10, ADAMTS1, and EDNRB were identified as potential biomarkers for spinal cord injury.

[0073] Example 6: Immune Infiltration Analysis Immune cell infiltration analysis The CIBERSORT package in R was used to calculate the infiltration levels of 22 immune cells in GSE151371 based on the LM22 gene set. The ggplot2 package in R was used to plot stacked bar charts of immune cell scores and box plots of immune cell infiltration levels. Wilcoxon rank-sum tests were performed in the disease and control groups to identify immune cells with significant differences between the two groups. The psych package in R was used to calculate the Spearman correlation between biomarkers and the differentially expressed immune cells and to plot the correlation heatmap.

[0074] This invention analyzed the immune cell composition of the spinal cord injury region. The results showed that the infiltration ratios of various immune cells changed significantly after injury, reflecting the complex inflammatory response process following injury. A deeper understanding of the roles of these immune cells is crucial for developing effective anti-inflammatory and repair strategies.

[0075] Analysis method: CIBERSORT algorithm Based on transcriptome data, the relative infiltration ratio of 22 immune cells in each sample was accurately estimated.

[0076] The results are as follows Figure 15 As shown, the infiltration levels of specific immune cells, such as macrophages and neutrophils, differed significantly between the injury group (the model constructed using TBHP treatment) and the control group (normal cells). Immune infiltration analysis revealed a significant change in the infiltration ratios of various immune cells after spinal cord injury, reflecting a complex inflammatory microenvironment. Abnormal infiltration is key to the inflammatory response after spinal cord injury and contributes to understanding the mechanisms of the immune microenvironment following injury.

[0077] Differential analysis and correlation study of immune infiltration This invention analyzed the immune cell composition of the spinal cord injury region. The results showed that the infiltration ratios of various immune cells changed significantly after injury, reflecting the complex inflammatory response process following injury. A deeper understanding of the roles of these immune cells is crucial for developing effective anti-inflammatory and repair strategies.

[0078] like Figure 16 A. By using the rank-sum test, significant differences were found in eight types of immune cells between the injury group and the control group, such as neutrophils and macrophages.

[0079] like Figure 16 B. Biomarkers are significantly correlated with differentially expressed immune cells: ITGA6 has the highest positive correlation with resting memory CD4+ T cells, while multiple biomarkers are negatively correlated with CD8+ T cells.

[0080] Example 7 Drug Prediction - Molecular Docking To translate basic research into potential therapeutic strategies, this invention uses biomarkers for drug prediction. An interaction network of 102 drugs was constructed, and the most promising drugs were selected for molecular docking validation. Results show that drugs such as estradiol can effectively bind to the biomarkers of this invention, providing important clues for subsequent drug development.

[0081] This embodiment, based on the DSigDB database, predicted 102 drugs that may target 5 biomarkers and constructed an interaction network (see...). Figure 17 A). For example Figure 17 As shown in B, estradiol, the drug with the most targets, was selected for molecular docking. The results showed that it has good binding ability with biomarker proteins, and the binding energy is less than -6.5 kcal / mol.

[0082] GSEA and GSVA functional enrichment analysis This invention also included GSEA and GSVA analyses to gain a deeper understanding of the function of biomarkers and the overall changes in the disease at the pathway level. GSEA analysis revealed the key pathways that each biomarker may regulate, while GSVA analysis demonstrated the reprogramming of overall pathway activity after injury. These results provide a more comprehensive perspective for understanding the molecular mechanisms of spinal cord injury.

[0083] like Figure 18 A. Based on the expression of individual biomarkers, GSEA analysis revealed that they were mainly enriched in pathways related to cell migration and proliferation, such as actin cytoskeleton regulation and the MAPK signaling pathway. Figure 18 B. By comparing the pathway activity of the injury group and the control group, 50 differentially expressed pathways were found. Among them, the upregulated pathways were mostly related to immunity and inflammation, while the downregulated pathways were mostly related to metabolism and synaptic transmission.

[0084] Single-cell follow-up analysis: pseudo-timing and cell communication At the single-cell level, this invention provides a more in-depth analysis. Pseudo-temporal analysis depicts the dynamic evolutionary pathways of astrocytes and endothelial cells after injury, while cell communication analysis reveals the reshaping of intercellular signaling patterns following injury. These results contribute to our understanding of cellular behavior and interactions within the injury microenvironment.

[0085] like Figure 19A. Pseudo-temporal analysis of astrocytes and endothelial cells revealed the dynamic trajectory of cell changes after injury and found that the expression of biomarkers changed with cell state. For example... Figure 19 B. We constructed an intercellular communication network and found significant differences in cell communication patterns between the damaged group (the model constructed by TBHP treatment) and the control group (normal cells). We also highlighted the ligand-receptor interactions between astrocytes and endothelial cells and other cells.

[0086] Validation of biomarker expression in single cells This embodiment validated the expression of biomarkers at the single-cell level. The results showed that these five biomarkers were not only specifically expressed in astrocytes, but their expression levels also differed significantly between the damage group and the control group. This is completely consistent with the previous findings at the whole transcriptome level, strongly validating the reliability of these biomarkers.

[0087] like Figure 20 A, in a single-cell dataset, validated that five biomarkers were specifically expressed in astrocytes with accurate localization; such as Figure 20 B further confirmed that these five biomarkers showed significant differences between the injury group and the control group, which was highly consistent with the bulk RNA-seq results.

[0088] Conclusion: The specificity and differential nature of the biomarker were validated at the single-cell level, strongly demonstrating its reliability and clinical potential.

[0089] Example 8 Cell Experiment Biomarkers of spinal cord injury Based on GSE151371, four machine learning classifiers were first built using the R package "caret": Random Forest (RF), Generalized Linear Model (GLM), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM). The residuals and feature importance of the classifiers were analyzed using the R package "DALEX". The classification performance of the four models was evaluated using AUC scores using the "auditor" package, and the classifier with the lowest residual and highest AUC was selected as the best predictive classifier. The best predictive classifier was then used to rank the importance of target genes in astrocyte-related diseases, and the top 10 most important genes were selected, such as... Figure 28 .

[0090] 2) Based on GSE151371, box plots were generated using the ggplot2 package in R to show the expression differences of the top 10 genes in disease and control samples, and Wilcoxon rank-sum tests were performed between groups. ROC curves for the top 10 genes were plotted in GSE151371 using the pROC package in R. An AUC value greater than 0.7 indicates good diagnostic efficacy. Figure 29 .

[0091] 3) Select genes that show significant differences in GSE151371 and have an AUC value greater than 0.7 as biomarkers.

[0092] This invention utilizes various machine learning algorithms to construct predictive models and comprehensively evaluates their performance. The results show that the generalized linear model (glm) performs best in distinguishing between damaged and normal samples, making it the optimal choice for subsequent analysis in this invention.

[0093] like Figure 21 As shown, there was no significant difference between the experimental group (2 μM) and the control group; however, the experimental group (5 μM, 15 μM, and 25 μM) showed significantly lower levels of ITGA6, ADAMTS1, ANXA2, and CXCL10 compared to the control group. Regarding EDNRB gene expression: the experimental group (5 μM) showed no significant difference compared to the control group, while the experimental group (15 μM and 25 μM) showed significantly higher levels than the control group.

[0094] This invention utilizes various machine learning algorithms to construct predictive models and comprehensively evaluates their performance. The results show that the generalized linear model (glm) performs best in distinguishing between damaged and normal samples, making it the optimal choice for subsequent analysis in this invention.

[0095] like Figure 22 As shown, the expression level of ITGA6 protein in the model group was significantly reduced compared with that in the control group.

[0096] like Figure 23 As shown in Figure A, the proliferation capacity of the model group increased compared to the control group. Compared to the model + empty vector, the proliferation capacity decreased after model + ITGA6 overexpression. Figure 23 As shown in Figure B, the model group exhibited increased transferability compared to the control group. However, the transferability decreased after model + ITGA6 overexpression compared to model + empty vector.

[0097] like Figure 25 As shown, compared with the control group, the Model group (the model constructed by TBHP treatment) showed significantly reduced expression levels of PI3K, P-AKT, and SRC proteins. Compared with the Model+NC group, the Model+OE-ITGA6 group showed significantly upregulated expression levels of PI3K, P-AKT, and SRC proteins. However, there was no significant change in AKT protein expression.

[0098] Example 9

[0099] A preferred embodiment of this invention provides a diagnostic kit based on the principle of double-antibody sandwich enzyme-linked immunosorbent assay (ELISA). For example... Figure 26 As shown, the microplate in the kit is pre-coated with a capture antibody that specifically recognizes the ITGA6 protein. When the sample to be tested (such as cerebrospinal fluid or serum) is added, the ITGA6 protein in the sample binds to the capture antibody (ABWAYS CY5509). After washing to remove unbound material, an enzyme-labeled detection antibody (Goat anti-Rabbit IgG (H+L) Secondary Antibody, HRP) is added. This detection antibody binds to another epitope of the ITGA6 protein, forming a sandwich complex of "capture antibody-ITGA6 protein-detection antibody". After washing again, a chromogenic substrate (such as TMB) is added, and HRP catalyzes a color reaction. Finally, a stop solution is added to terminate the reaction, and the absorbance (OD value) is measured using a microplate reader at a specific wavelength (such as 450 nm). The OD value is directly proportional to the concentration of ITGA6 protein in the sample.

[0100] Specific embodiments of the diagnostic kit of the present invention include, but are not limited to, the following components: Microplates coated with lITGA6 protein capture antibody: 96-well or other sizes of ELISA plates, the inner walls of which are coated with a high-affinity, high-specificity murine anti-human ITGA6 protein monoclonal antibody (ABWAYS CY5509).

[0101] Detection antibodies: Rabbit-derived anti-human ITGA6 protein polyclonal or monoclonal antibodies (Goat anti-Rabbit IgG (H+L) Secondary Antibody, HRP) labeled with biotin or HRP that can bind to different epitopes of ITGA6 protein.

[0102] Standards: A series of lyophilized powders or solutions of recombinant human ITGA6 protein with known concentration gradients, used to plot standard curves.

[0103] Enzyme conjugate: horseradish peroxidase-labeled streptavidin (HRP) (used when the detection antibody is biotin-labeled).

[0104] Chromogenic substrate solution (solution A and solution B or single-component TMB solution): used to catalyze the color reaction. In this example, solutions A and B are commercially available (e.g., ELISA kit chromogenic solutions a and b). Solution A (substrate solution): contains TMB (3,3',5,5'-tetramethylbenzidine), which is the main chromogenic substrate; Solution B (oxidizing agent solution): contains hydrogen peroxide and an acidic buffer. Solutions A and B need to be mixed together in a 1:1 ratio before the color development step. The mixed solution must be used within a short time (usually within 30 minutes) to ensure the sensitivity and stability of the reaction.

[0105] Termination solution: For example, a 2M sulfuric acid solution, used to terminate the colorimetric reaction.

[0106] Washing solution (concentrate): Used to wash away unbound substances in micropores, such as phosphate buffer (PBS or TBS) containing surfactants (such as Tween-20).

[0107] Sample diluent: A buffer solution used to dilute test samples and standards (containing blocking proteins such as BSA or casein).

[0108] Instructions for use (operation steps) of the diagnostic kit of this invention: Preparation: Allow all components of the kit to reach room temperature. Determine the required number of strips based on the number of samples to be tested. Dilute the standards to a series of concentration points using standard diluent (e.g., 1000, 500, 250, 125, 62.5, 0 pg / mL).

[0109] Sample addition: Set up standard wells and sample wells in the microplate, and add 100 μL of diluted standard or sample to each well.

[0110] Incubation: Incubate at 37℃ for 1-2 hours.

[0111] Wash the plate: Discard the liquid in the well and wash the plate 3-5 times with detergent.

[0112] Add detection antibody: Add 100 μL of enzyme-labeled detection antibody working solution to each well.

[0113] Second incubation: Incubate at 37℃ for 30-60 minutes.

[0114] Wash the plate: Repeat step 4.

[0115] Color development: Add 100 μL of color development substrate solution to each well and incubate in the dark for 15-20 minutes.

[0116] Termination: Add 50 μL of termination solution to each well.

[0117] Reading and Calculation: Within 15 minutes, measure the OD value of each well at a wavelength of 450 nm using a microplate reader. Plot a standard curve with the standard concentration on the x-axis and the OD value on the y-axis. Based on the OD value of the sample, find or calculate the concentration of ITGA6 protein in the sample from the standard curve.

[0118] The typical standard curve and clinical application results obtained using the kit of this invention are shown in the figure below. Figure 27 As shown.

[0119] like Figure 27 As shown, the standard curve exhibits a good linear relationship, indicating that the kit has good quantitative detection performance within a certain concentration range. Figure 27 As shown in Figure B, analysis of clinical SCI patient samples revealed a significantly higher concentration of ITGA6 protein compared to the healthy control group (HC). Further analysis showed that patients with good prognosis (e.g., significant improvement in ASIA score) had significantly higher ITGA6 protein levels than those with poor prognosis (PP).

[0120] Interpretation of results and clinical significance: The method and kit of this invention can rapidly, accurately, and quantitatively detect the content of ITGA6 protein in biological samples of SCI patients.

[0121] 1. Used for auxiliary diagnosis: When a patient presents with suspected SCI symptoms, their ITGA6 protein level can be measured. If it is significantly higher than the baseline of healthy individuals, it can be used as an auxiliary diagnostic basis for the occurrence of SCI.

[0122] 2. For severity assessment: The level of ITGA6 protein may be related to the severity of injury and can be comprehensively assessed in combination with imaging and clinical scores.

[0123] 3. For prognostic assessment: Detect ITGA6 protein levels in the early stages of injury (e.g., within 72 hours post-injury). High ITGA6 protein levels indicate effective activation of the endogenous astrocyte-endothelial cell protective pathway, potentially leading to a better prognosis. Conversely, low ITGA6 protein levels may indicate insufficient protective response, a higher prognostic risk, and the need for more aggressive clinical intervention.

[0124] Compared with the prior art described in the background section, the diagnostic kit and method based on the detection of ITGA6 protein provided by the present invention have the following significant and well-founded beneficial effects in the clinical assessment of spinal cord injury (SCI): 1. In terms of diagnostic information: it has achieved a leap from "damage assessment" to "repair potential assessment", providing new and useful performance.

[0125] Current technologies—whether clinical neurological function scores (such as the ASIA score), imaging examinations (MRI), or existing biomarkers (such as neurofilament light chains (NFL) and glial fibrillary acidic protein (GFAP))—all primarily function to assess the “degree of damage” or “structural abnormality” following injury. They provide “static” or “consequential” information about the severity of the injury.

[0126] This invention: The expression level of the ITGA6 protein detected in this invention directly reflects the strength of the body's endogenous, astrocyte-initiated "protective response" aimed at protecting the vascular endothelium and blood-spinal barrier. Therefore, this invention provides, for the first time, a molecular tool for assessing the "endogenous repair potential" of SCI patients. This is a completely novel capability not found in existing technologies. Through this invention, physicians can not only know the "severity of the injury" but also the "strength of the body's self-repair efforts," which has unparalleled guiding value for prognostic assessment and treatment decisions.

[0127] 2. In terms of the accuracy and early detection of prognostic results: Significantly improves the accuracy and timeliness of prediction.

[0128] Current technology: The prognostic value of clinical scoring and imaging assessments has limitations, especially in the early stages of injury. Changes in macroscopic indicators often lag behind microscopic pathophysiological changes, leading to high uncertainty in early prognostic judgments. For example, the extent of spinal cord edema shown on early MRI may not be linearly related to the final functional recovery.

[0129] This invention addresses the issue that the ITGA6 protein, a core molecule in a critical protective pathway during the early stages of secondary injury, reflects the true biological state of the injury microenvironment earlier and more sensitively. Research data shows that ITGA6 protein levels are strongly correlated with functional recovery outcomes months later, within hours to days after injury. Therefore, compared to traditional methods relying on observation over weeks or even months, this invention significantly advances the time window for accurate prognostic assessment, gaining valuable time for early and precise clinical intervention, thereby improving the efficiency and accuracy of prognostic prediction.

[0130] 3. In terms of detection specificity: the target molecules are more directly and specifically associated with key pathological processes.

[0131] Current technology: General injury markers such as NFL and GFAP can be elevated in any form of injury to the central nervous system (such as traumatic brain injury, stroke, and neurodegenerative diseases). They lack specific indications for SCI-specific pathological processes, particularly the functional status of neurovascular units.

[0132] This invention focuses on the ITGA6 protein, a core signaling molecule that specifically mediates the crucial protective communication axis between astrocytes and endothelial cells. Detecting it directly probes the functional state of this protective pathway. Therefore, compared to biomarkers reflecting general cellular damage, ITGA6 protein exhibits higher biological specificity as a biomarker, and its detection results can more precisely pinpoint the core pathological process of blood-spinal barrier repair in SCI, reducing interference from other contributing factors.

[0133] 4. In terms of ease of clinical application and scalability: It provides a standardized testing solution that is easy to operate and cost-effective.

[0134] Current technologies: Imaging equipment such as MRI is expensive and complex to operate, making it inconvenient for frequent dynamic monitoring of critically ill patients. Clinical scoring relies on the assessor's experience and is therefore somewhat subjective.

[0135] This invention provides a diagnostic kit based on ELISA technology.

[0136] Easy to operate: ELISA is a routine technique in clinical laboratory departments. The process is standardized, and the technical requirements for operators are not high, making it easy to master and promote.

[0137] Cost savings: Compared to large imaging equipment, the cost per ELISA test and the investment in equipment are much lower, making it suitable for large-scale screening and routine monitoring.

[0138] Objective quantification: The test results are presented in specific concentration values ​​(such as pg / mL), which are objective and repeatable, avoiding the bias caused by subjective evaluation and improving the reliability and consistency of the test results.

[0139] Feasibility of dynamic monitoring: Due to its minimally invasive (blood sample) or low-invasive (cerebrospinal fluid sample) and low-cost characteristics, it is possible to continuously and dynamically monitor the ITGA6 protein level of patients, thereby enabling real-time tracking of disease progression and treatment response.

Claims

1. A biomarker for assessing the severity of spinal cord injury prognosis, characterized in that, The biomarker is the ITGA6 protein, which exerts an endothelial protective effect by activating the pi3k / akt pathway on vascular endothelial cells.

2. The use of the biomarker of claim 1 in the preparation of products for diagnosing the prognosis and assessing the severity of spinal cord injury.

3. The application as described in claim 2, characterized in that: The product can be any one of the following: a detection reagent, a kit, a microarray, or a biochip.

4. The application as described in claim 3, characterized in that, When the product is a reagent kit, the diagnostic reagent kit includes, but is not limited to, the following components: ITGA6 protein capture antibody-coated microplates: 96-well or other sizes of ELISA plates, the inner walls of which are coated with mouse-derived anti-human ITGA6 protein monoclonal antibody. Detection antibodies: Rabbit-derived anti-human ITGA6 protein polyclonal or monoclonal antibodies labeled with biotin or HRP that can bind to different epitopes of the ITGA6 protein; Standards: A series of lyophilized powders or solutions of recombinant human ITGA6 protein with known concentration gradients, used to plot a standard curve; Enzyme conjugate: horseradish peroxidase-labeled streptavidin; Colorimetric substrate solution: Solution A and Solution B or single-component TMB solution; Termination solution: Used to terminate the colorimetric reaction; Washing solution: Used to wash away unbound substances in micropores; Sample diluent: Used to dilute test samples and standards.

5. The application as described in claim 3, characterized in that, The ITGA6 protein capture antibody is ABWAYS CY5509.

6. The application as described in claim 3, characterized in that, The stop solution is a 2M sulfuric acid solution.

7. The use of the biomarker according to claim 1 in the preparation of a medicament for treating spinal cord injury, characterized in that: The drug targets the biomarker, and the drug contains estradiol.

Images