Use of detection reagents in detection of zmynd8 fusion gene positive sgbm

Abstract

The application discloses application of a detection reagent in detection of ZM fusion gene (PTPRZ1-MET fusion gene) positive secondary glioblastoma (sGBM), and belongs to the technical field of biological medicine, so as to solve the problems of the existing ZM fusion detection method, such as complex operation, poor specificity, and unsuitability for routine clinical detection. Based on high-throughput RNA sequencing and machine learning algorithm (XGBoost) analysis, it is found for the first time that the combined expression profile of three genes, MET, PCDHGA3 and FAM3C, can significantly distinguish ZM fusion positive and negative samples, and the diagnostic accuracy (AUC) is as high as 99.8%, with high sensitivity and specificity. The detection product of the application is suitable for paraffin-embedded (FFPE) pathological tissue samples and frozen samples, and can complete the detection in combination with mature platforms such as multiplex immunohistochemistry or high-throughput RNA sequencing, and has good sample compatibility and operation feasibility.

Description

Technical Field

[0001] This invention belongs to the field of biomedical technology, and in particular relates to the application of detection reagents in the detection of ZM fusion gene positive sGBM. Background Technology

[0002] Glioblastoma (GBM) is one of the most aggressive and deadliest malignant tumors of the central nervous system, with an extremely poor prognosis. Despite current standard treatments including surgical resection, radiotherapy, and temozolomide chemotherapy, the median survival remains less than two years. Secondary glioblastoma (sGBM) is a subtype that progresses from low-grade astrocytoma, commonly seen in young patients, and is significantly different from primary GBM (pGBM) at the molecular level. In recent years, the development of large-scale high-throughput sequencing technology has driven in-depth research into the molecular subtyping of GBM and potential driver genes.

[0003] In existing studies, the PTPRZ1-MET (ZM) fusion gene has been identified as one of the representative driving events in sGBM, promoting cell proliferation, invasion, and treatment resistance by activating the downstream MET pathway, and significantly associated with poor patient prognosis. Therefore, accurate detection of ZM fusions is of great significance for precision clinical diagnosis and treatment. However, there are no specific antibodies for ZM fusion detection, and current detection methods mainly rely on RT-PCR + Sanger sequencing and fluorescence in situ hybridization (FISH).

[0004] Reverse transcription PCR (RT-PCR) combined with Sanger sequencing is cumbersome, relies on high-quality RNA, requires precise primer design for fusion breakpoints, and cannot cover unknown or newly discovered fusion types. It is unsuitable for high-throughput clinical screening due to its high technical threshold and cost. Fluorescence in situ hybridization (FISH) requires precise experimental procedures and pathological field selection, is susceptible to spatial heterogeneity of samples leading to false negatives, has expensive probes, and a long detection cycle, making it unsuitable for large-scale rapid clinical diagnosis. It may not effectively identify genes with closely spaced fusion sites (such as PTPRZ1 and MET), and has limited ability to identify low-frequency, early-disseminated fusion cells. Therefore, there is currently a lack of a highly sensitive, specific, low-cost, and applicable auxiliary diagnostic method for ZM fusions in routine clinical tissue samples. Summary of the Invention

[0005] To address the aforementioned problems, the purpose of this invention is to provide a detection reagent for detecting ZM fusion gene-positive sGBM. Based on high-throughput RNA sequencing and machine learning algorithm (XGBoost) analysis, this invention is the first to discover that the co-expression profiles of three genes—MET, PCDHGA3, and FAM3C—can significantly distinguish between ZM fusion-positive and negative samples, with a diagnostic accuracy (AUC) as high as 99.8%, far superior to single-gene detection, exhibiting high sensitivity and specificity. Further verification through multiplex immunofluorescence staining shows that the co-expression of these three genes at the protein level is also highly specific, accurately identifying ZM fusion-positive sGBM tumor tissue, providing multi-level verification evidence from the transcriptional to the protein level.

[0006] To achieve the above objectives, this application provides the following technical solution: The first aspect of the present invention provides the use of reagents for the simultaneous detection of MET, PCDHGA3 and FAM3C in the preparation of a detection product for PTPRZ1-MET fusion gene-positive secondary glioblastoma.

[0007] In a preferred embodiment of this aspect, the detection product is a reagent for detecting the expression levels of the MET, PCDHGA3, and FAM3C genes, and the reagent is used to perform an RNA sequencing method.

[0008] In some preferred examples in this regard, when the fusion probability value of the expression levels of MET, PCDHGA3, and FAM3C genes is ≥0.1, the PTPRZ1-MET fusion gene fusion status is positive; otherwise, the PTPRZ1-MET fusion gene fusion status is negative.

[0009] In a preferred embodiment of this aspect, the detection product is a reagent for detecting the expression levels of MET, PCDHGA3, and FAM3C proteins, and the reagent is used to perform a multicolor immunohistochemical staining method.

[0010] In some preferred examples of this, when the expression levels of MET, PCDHGA3, and FAM3C proteins are all positive, the PTPRZ1-MET fusion gene fusion status is positive; otherwise, the PTPRZ1-MET fusion gene fusion status is negative.

[0011] In some preferred embodiments of this aspect, the sample used for the reagent is a tumor tissue sample of the subject being tested.

[0012] A second aspect of the present invention provides a method for detecting PTPRZ1-MET fusion gene-positive secondary glioblastoma for non-diagnostic purposes, the method comprising: detecting the expression levels of MET, PCDHGA3 and FAM3C genes or the expression levels of MET, PCDHGA3 and FAM3C proteins in tumor tissue samples in vitro. The expression levels of MET, PCDHGA3, and FAM3C genes need to be input into the XGBoost model to output the identification results of the ZM fusion gene fusion status.

[0013] A fourth aspect of the present invention provides a detection system for PTPRZ1-MET fusion gene-positive secondary glioblastoma, the system comprising: a sample processing module for detecting the expression levels of MET, PCDHGA3 and FAM3C genes in tumor tissue samples using RNA sequencing methods; The input module allows you to input the expression levels of the MET, PCDHGA3, and FAM3C genes. The computing module includes a memory and a processor; wherein the memory stores a computer program, and the processor is configured to execute the computer program stored in the memory to implement an algorithm for the following discrimination: when the fusion probability value of the expression levels of MET, PCDHGA3, and FAM3C genes in the tumor tissue sample to be tested is ≥0.1, the PTPRZ1-MET fusion gene fusion status is positive; otherwise, the PTPRZ1-MET fusion gene fusion status is negative. The output module outputs the fusion status of the PTPRZ1-MET fusion gene in the tumor tissue sample.

[0014] Compared with the prior art, the present invention has the following advantages: This invention is the first to discover that the combined expression profiles of the three genes MET, PCDHGA3, and FAM3C can significantly distinguish between ZM fusion-positive and negative samples, with a diagnostic accuracy (AUC) of up to 99.8%, far superior to single-gene detection, and possessing high sensitivity and specificity.

[0015] The combined detection method of MET, PCDHGA3 and FAM3C genes in this invention has the advantages of simple operation, high repeatability and wide adaptability. It is especially suitable for tissue detection of FFPE samples and frozen samples, and is easy to promote in large-scale clinical practice. Attached Figure Description

[0016] Figure 1 The differentially expressed genes in the ZM+ group and the ZM- group are shown below: A. Volcano plot of all differentially expressed genes, with 359 genes significantly upregulated in the ZM+ group; B. Heatmap of the top 20 genes significantly upregulated in the ZM+ group.

[0017] Figure 2The results of screening for three genes, MET, PCDHGA3, and FAM3C, are shown below: A. The importance ranking of feature genes in the XGBoost model; B. The ROC curve of the three-gene joint discrimination model; C. The confusion matrix of the three-gene model on the validation set, with the vertical axis representing the true results and the horizontal axis representing the model prediction results.

[0018] Figure 3 Multicolor histochemical staining results of paraffin sections from patients in the ZM+ and ZM- groups. Detailed Implementation

[0019] The embodiments of the present invention will be described in detail below with reference to examples. However, those skilled in the art will understand that the following examples are for illustrative purposes only and should not be considered as limiting the scope of the invention. Unless otherwise specified in the examples, conventional conditions or conditions recommended by the manufacturer are followed. Reagents or instruments whose manufacturers are not specified are all commercially available conventional products.

[0020] The first aspect of this invention provides the application of reagents for detecting MET, PCDHGA3, and FAM3C in the preparation of a detection product for PTPRZ1-MET fusion gene-positive secondary glioblastoma.

[0021] According to the present invention, the ZM fusion gene is defined as the PTPRZ1-MET fusion gene. The detection product is a detection reagent suitable for RNA sequencing or multicolor immunohistochemistry methods, such as a specific antibody that can be used for multiplex immunohistochemical staining.

[0022] Example 1: Screening for ZM fusion-positive markers MET, PCDHGA3, and FAM3C in secondary glioblastoma. Three-gene combination screening strategy: This invention uses high-throughput transcriptome sequencing data combined with machine learning methods to systematically screen for three-gene combinations (MET, PCDHGA3, FAM3C) that can specifically identify PTPRZ1-MET (ZM) fusion-positive secondary glioblastoma (sGBM).

[0023] The specific screening steps are as follows: 1. Sample grouping and data source: A total of 159 clearly genotyped sGBM samples were included, including 15 ZM fusion positive (ZM+) samples and 144 ZM fusion negative (ZM−) samples. Gene expression profiles of all samples were obtained by next-generation sequencing (RNA-Seq). The data were subjected to quality control, transcript alignment and expression level normalization (such as TPM or DESeq2 normalization).

[0024] 2. Differential Expression Analysis: Using the DESeq2 package, statistical tests were performed on the expression data between ZM+ and ZM− samples. The selection criteria were: |log2Fold Change|>1 and FDR<0.05, initially identifying 359 differentially expressed genes. The top 40 significantly differentially expressed genes were used as the feature input set for subsequent modeling, such as... Figure 1 As shown in Figure A, this is a volcano plot of all differentially expressed genes, with the ZM+ group showing a significant upregulation of 359 genes. Figure 1 The heatmap shows that the top 20 genes with significantly upregulated genes are in group B of the ZM+ group.

[0025] 3. Machine learning model construction and feature selection: The XGBoost (eXtreme Gradient Boosting) classifier was used, with the top 40 differential gene expression matrices as input features and ZM fusion states (ZM+ / −) as labels. The XGBoost classifier was trained using 5-fold cross-validation, and the model was trained using the caret package. The final model was then exported to evaluate classification performance.

[0026] During the machine learning model construction process, hyperparameters were optimized using 5-fold cross-validation, ultimately generating the optimal model parameter combination: nrounds=100, max_depth=3, eta=0.1, gamma=0, colsample_bytree=0.8, min_child_weight=1, subsample=0.8. Based on the feature importance score based on "gain" output by the final model, all feature genes were ranked, and the three core genes with the strongest discriminative power were selected.

[0027] Using the RNA expression levels of the three genes MET, PCDHGA3, and FAM3C as input features, the model learns the expression differences between ZM fusion-positive and negative samples through a training set, outputting a fusion-positive probability value (0~1). In model performance evaluation, when the output probability is ≥0.1, the sensitivity is 100% and the specificity is 97.2%. Therefore, this invention uses 0.1 as the optimal judgment threshold: if the model prediction value of the input sample is ≥0.1, it is judged as ZM fusion-positive; otherwise, it is negative. This model has high accuracy and clinical feasibility, and is suitable for assisting in the identification of fusion status after quantification of expression through RNA sequencing or immunostaining.

[0028] Feature Importance Assessment and Three-Gene Screening: Extracting the gain feature importance score of the XGBoost classifier, such as... Figure 2 As shown in Figure A, based on the gene's contribution to classification, MET, PCDHGA3, and FAM3C were ultimately identified as the three most representative characteristic genes. Figure 2 As shown in Figure B, the combined expression of the three genes exhibited extremely high diagnostic performance in distinguishing between ZM+ and ZM− samples, with an AUC of 99.8% on the ROC curve. Figure 2 As shown in Figure C, the confusion matrix also indicates high classification accuracy and few classification errors.

[0029] A detection system for PTPRZ1-MET fusion gene-positive secondary glioblastoma, identified using MET, PCDHGA3, and FAM3C as the three most representative characteristic genes, is described above. The system includes: a sample processing module for detecting the expression levels of MET, PCDHGA3, and FAM3C genes in tumor tissue samples from patients using RNA sequencing; an input module for inputting the expression levels of MET, PCDHGA3, and FAM3C genes; a calculation module including a memory and a processor; wherein the memory stores a computer program, and the processor is configured to execute the computer program stored in the memory to implement an algorithm that performs the following judgment: when the fusion probability value of the MET, PCDHGA3, and FAM3C gene expression levels in the tumor tissue sample is ≥0.1, the PTPRZ1-MET fusion gene fusion status is positive; otherwise, the PTPRZ1-MET fusion gene fusion status is negative; and an output module for outputting the PTPRZ1-MET fusion gene fusion status of the tumor tissue sample.

[0030] Clinical sample validation: The protein expression of the above three genes was validated in independent FFPE tumor tissue samples by multiple immunohistochemical staining; the experimental results showed that MET, PCDHGA3 and FAM3C were highly expressed and co-localized in ZM+ tumors, while they were expressed at very low levels or absent in ZM− samples.

[0031] Example 2: Clinical Sample Validation Step 1: Sample Preparation Obtain tumor tissue samples with a clear provenance and a clinicopathological diagnosis of suspected secondary glioblastoma (sGBM). Acceptable sample types include: surgically excised frozen fresh tumor tissue; and formalin-fixed paraffin-embedded (FFPE) tissue samples.

[0032] Sample processing: For frozen tissue samples, store in liquid nitrogen and operate in a dry ice environment before sampling; after grinding with liquid nitrogen, add TRIzol or an equivalent lysis reagent for RNA extraction. For FFPE paraffin sections: the recommended section thickness is 4-10 μm, placed on slides free of RNase contamination; dewax with xylene (usually twice, 10 minutes each time), then hydrate sequentially with different gradients of ethanol to pure water, and perform antigen retrieval by heating with sodium citrate buffer.

[0033] Step 2: Combined detection of MET, PCDHGA3, and FAM3C genes Multiplex immunohistochemical staining: MET, PCDHGA3, and FAM3C were labeled with three different fluorescent primary antibodies; their co-expression and co-localization were observed using a laser confocal imaging system. RNA-seq sequencing: The expression levels of MET, PCDHGA3, and FAM3C were detected.

[0034] Step 3: Interpretation criteria: Based on the results of the Xgboost model or multicolor immunohistochemical staining (e.g., Figure 3 As shown, multicolor histochemical staining of paraffin sections from ZM+ and ZM- patients verified the high expression and co-localization of MET, PCDHGA3, and FAM3C in ZM+ tumor tissues, while almost no significant expression was observed in ZM− tissues. If all three genes were significantly expressed (with a set threshold), the tumor was judged as ZM fusion positive; otherwise, it was judged as fusion negative.

[0035] Example 3 Clinical Trial Multiplex immunohistochemistry (IHC) was performed on paraffin-embedded tumor tissue sections fixed in paraformaldehyde from 50 sGBM patients (20 ZM+ and 30 ZM-). The ZM fusion status of sGBM was confirmed by Sanger sequencing. A multicolor immunohistochemical staining kit was used, and three fluorescently labeled primary antibodies were used to label MET, PCDHGA3, and FAM3C, respectively. Fluorescence signals were observed under a fluorescence microscope, revealing significantly different expression patterns between the two groups. High expression and co-localization of the three genes were observed in ZM+ tumor tissues, while almost no expression was observed in ZM− tissues.

[0036] The method of combined detection of MET, PCDHGA3, and FAM3C genes in this invention has the following advantages: (1) Significantly improves the accuracy and specificity of detection. This invention uses machine learning (XGBoost) to screen three genes (MET, PCDHGA3, and FAM3C) that have significant co-expression characteristics in ZM fusion-positive samples, with a combined diagnostic AUC of up to 99.8%, which is significantly better than traditional single-gene indicators or pathological features. This method can accurately identify the biological status of tumors related to ZM fusion, reduce false negative and false positive rates, and improve clinical diagnostic confidence.

[0037] (2) Effectively replaces existing high-threshold detection technologies Compared to the commonly used methods of FISH, RT-PCR, or next-generation sequencing for detecting fusion genes, this invention does not require direct detection of fusion breakpoints, thus avoiding the limitation of detection sensitivity due to tumor heterogeneity; FISH experiments are cumbersome and easily affected by the field of view of the slide; and the design complexity caused by the diversity of fusion breakpoints. This method is suitable for clinical molecular testing laboratories in a rapid, high-throughput, and low-cost manner.

[0038] (3) Provide auxiliary support for accurate classification and treatment Given that ZM fusion-positive sGBM tumors are typically more malignant and have a worse prognosis, the three-gene combined detection scheme of this invention can be used to assist in tumor subtyping and risk stratification. At the same time, ZM fusion is also becoming a key population for targeted therapy (such as PLB-1001) research, and this invention can serve as a companion diagnostic tool, providing a foundation for precision medicine.

[0039] (4) Possesses clinical potential for expanded applications Besides sGBM, this trigene combination may also exhibit expression characteristics in other tumor subtypes or fusion-related diseases, and has the potential to expand into cross-indication disease detection and mechanism research.

[0040] It is understood that the above specific description of the present invention is only for illustrating the present invention and is not limited to the technical solutions described in the embodiments of the present invention. Those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention to achieve the same technical effect; as long as the use needs are met, they are all within the protection scope of the present invention.

Claims

1. Application of reagents for simultaneous detection of MET, PCDHGA3 and FAM3C in the preparation of PTPRZ1-MET fusion gene positive secondary glioblastoma detection products.

2. The application according to claim 1, characterized in that, The detection product is a reagent for detecting the expression levels of MET, PCDHGA3, and FAM3C genes, and the reagent is used to perform RNA sequencing methods.

3. The application according to claim 1, characterized in that, The detection product is a reagent for detecting the expression levels of MET, PCDHGA3, and FAM3C proteins, and the reagent is used to perform a multicolor immunohistochemical method.

4. The application according to claim 3, characterized in that, When the expression levels of MET, PCDHGA3, and FAM3C proteins are all positive, the PTPRZ1-MET fusion gene fusion status is positive; otherwise, the PTPRZ1-MET fusion gene fusion status is negative.

Images