A detection kit for predicting the response of ulcerative colitis patients to vedolizumab treatment, a use of a reagent in the preparation of the kit and an integrated detection system

Abstract

The present application belongs to the field of medical detection, and particularly relates to a detection kit for predicting the response of an ulcerative colitis patient to vedolizumab treatment, the use of reagents in the preparation of the kit and an integrated detection system. The kit comprises reagents for detecting the expression levels of genes TMEM273, HSF1, SLC25A17 and STT3A in a sample of a subject; the expression levels of the four genes are used to predict and evaluate whether the patient is responsive to vedolizumab. The present application has high specificity for vedolizumab targeting alpha4beta7, but no predictive efficiency for etrolizumab targeting beta7, proving that it is a non-pandemic inflammation indicator with high target accuracy. The present application can realize the breakthrough of "liquid biopsy": the histological model shows significant predictive efficiency in an independent peripheral blood cohort, breaking the technical bottleneck that "local intestinal immune genes cannot be detected by blood sampling", greatly reducing the detection threshold and patient pain.

Description

Technical Field

[0001] This invention belongs to the field of biomedical testing, specifically relating to a test kit for predicting the response of ulcerative colitis patients to vedelizumab treatment, the use of a reagent in the preparation of the kit, and an integrated detection system. Background Technology

[0002] Vedolizumab is a humanized monoclonal antibody primarily used to treat moderate to severe ulcerative colitis (UC) and Crohn's disease (CD). It helps relieve symptoms and maintain disease remission by specifically inhibiting the intestinal inflammatory response. Vedolizumab targets α4β7 integrin, a protein that mediates the migration of lymphocytes to the intestine. By blocking the binding of α4β7 to adhesion molecules on the surface of intestinal vascular endothelial cells, it reduces local intestinal inflammation, thereby controlling disease progression. Vedolizumab is a prescription drug administered intravenously. The initial treatment typically involves once every three weeks (weeks 0, 2, and 6), followed by maintenance therapy every eight weeks. Specific dosage and duration of treatment should be adjusted by a physician based on the patient's condition.

[0003] Currently, the main methods for assessing the condition of patients with ulcerative colitis and guiding medication use in clinical practice include: 1. Clinical and endoscopic scoring systems: such as the Mayo Criterion. This method relies heavily on the doctor's subjective judgment and endoscopic observation, and requires invasive colonoscopy; 2. Routine non-specific biochemical indicators: such as C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and fecal calprotectin; 3. Existing molecular diagnostic techniques: mostly histological-level pan-inflammatory gene sets, used to assess the severity of the disease.

[0004] The shortcomings of existing technologies include the following: 1. Strong lag and invasiveness: Endoscopic examination is an invasive procedure, causing significant patient discomfort; biochemical indicators require "post-treatment monitoring" several weeks after medication to determine drug effectiveness, lacking "pre-treatment prediction" capabilities; 2. Lack of targeted drug specificity: Existing models cannot accurately map the molecular pathways of specific drugs. For example, they cannot distinguish the microscopic differences between vedelizumab (blocking α4β7) and itrazumab (blocking β7 and αEβ7) in downstream immune cascade reactions; 3. Difficulty in translating peripheral blood (liquid biopsy): Because intestinal local immune signals are easily diluted by systemic blood circulation, most gene markers currently mined from intestinal mucosal tissue have extremely poor predictive efficacy in blood and cannot be translated into convenient blood test kits.

[0005] Therefore, there is a need in the art for a test kit that can be used to predict the response of ulcerative colitis patients to vedelizumab treatment via blood testing, the use of a reagent in the preparation of the kit, and an integrated detection system. Summary of the Invention

[0006] This invention first provides a test kit for predicting the response of vedelizumab to ulcerative colitis patients, comprising reagents for detecting the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A in subject samples; the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A are used to predict and assess whether the patient responds to vedelizumab.

[0007] In this invention, the detection kit is named VeRGeS-UC.

[0008] In one specific implementation, the subject sample is selected from any one or more of intestinal mucosal biopsy tissue, peripheral blood, serum, plasma, or peripheral blood mononuclear cells.

[0009] In this invention, the blood sample used for examination is a non-invasive test, while the intestinal mucosal biopsy tissue sample used is a minimally invasive test.

[0010] In one specific embodiment, the reagent includes primer pairs and / or probes for detecting the expression levels of the genes TMEM273, HSF1, SLC25A17, and STT3A by reverse transcription real-time quantitative PCR, i.e., RT-qPCR.

[0011] In one specific embodiment, the primer pair includes: specific upstream and downstream primers for gene TMEM273, SEQ ID NO:1 and SEQ ID NO:2, respectively; specific upstream and downstream primers for gene HSF1, SEQ ID NO:3 and SEQ ID NO:4, respectively; specific upstream and downstream primers for gene SLC25A17, SEQ ID NO:5 and SEQ ID NO:6, respectively; and specific upstream and downstream primers for gene STT3A, SEQ ID NO:7 and SEQ ID NO:8, respectively; SEQ ID NO:1: TGGCTGGTGCTATGTGGTTAC; SEQ ID NO:2: GCAAGCAGTCAGAAGGGAAC; SEQ ID NO:3: ATGGAAGCGTGCGTGTC TAC; SEQ ID NO:4: TCGGGACGCATAGAGGAA; SEQ ID NO:5: GCCGGGAGTTTCTACTCATT; SEQ ID NO:6: GGCATACCCGTATTCTT GACC; SEQ ID NO:7: GCAGACAGACGAAACAGGGA; SEQ ID NO:8; NO:8: GGCAGTAAGAA TCCCAGCAG.

[0012] In one specific embodiment, the detection kit further includes reagents required for RT-qPCR detection, which include reverse transcriptase, DNA polymerase, dNTPs, fluorescent dyes or fluorescently labeled probes, and PCR reaction buffer.

[0013] In one specific embodiment, the reagents include library construction reagents for detecting the expression levels of the genes TMEM273, HSF1, SLC25A17, and STT3A by RNA sequencing, i.e., RNA-seq.

[0014] In one specific embodiment, the kit further includes a reagent for detecting the expression level of an internal reference gene, wherein the internal reference gene is selected from one or more of GAPDH, ACTB, 18S rRNA, and β-actin, and is used to correct the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A.

[0015] The present invention also provides the use of a reagent in the preparation of a kit, specifically the use of a reagent for detecting the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A in the preparation of a kit for predicting the efficacy of vedelizumab in patients with ulcerative colitis. The kit is used to predict and evaluate the treatment response of vedelizumab by detecting the expression levels of the four genes in samples from patients with ulcerative colitis.

[0016] In one specific implementation, the prediction and assessment are based on a comprehensive score calculated from the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A. When the expression level of a specific gene combination is higher or lower than a preset cutoff value, the patient is determined to be responsive or non-responsive to vedelizumab.

[0017] The present invention also provides an integrated detection system for predicting the efficacy of vedelizumab in patients with ulcerative colitis, comprising: a reagent compartment for carrying reagents in the aforementioned detection kit, an amplification detection module for performing RT-qPCR or RNA-seq, and a data processing module for acquiring fluorescence signals and outputting efficacy prediction results according to a preset algorithm.

[0018] This invention offers at least the following beneficial effects: It exhibits high specificity (84.6%) for vedelizumab targeting α4β7, while showing no predictive efficacy for etrazumab targeting β7 (negative control), demonstrating its non-pan-inflammatory marker and high targeting precision. This invention also achieves a breakthrough in "liquid biopsy": the histological model demonstrates significant predictive efficacy in an independent peripheral blood cohort, overcoming the technical bottleneck that "intestinal local immune genes cannot be detected through blood sampling," greatly reducing the detection threshold and patient discomfort. Attached Figure Description

[0019] Figure 1 ROC curves for predicting the response of vedelizumab in UC patients using the VeRGeS-UC score in the training set (E-MTAB-7914).

[0020] Figure 2 Decision curve analysis (DCA) was performed to predict the response of vedelizumab to UC patients using the VeRGeS-UC score in the training set (E-MTAB-7914).

[0021] Figure 3 ROC curves for predicting the response of UC patients to ectuzumab in the validation set (GSE72819) using the VeRGeS-UC score.

[0022] Figure 4 ROC curves were used to predict the response of vedelizumab to the VeRGeS-UC score in an additional validation set (GSE191328) using peripheral blood samples from UC patients.

[0023] Figure 5 The image shows the four-gene RT-qPCR amplification curves of VeRGeS-UC based on a real-world cohort baseline. Detailed Implementation

[0024] This invention provides a minimally invasive and non-invasive dual-use VeRGeS-UC kit for predicting the efficacy of vedelizumab in ulcerative colitis. Its core technology involves detecting the expression levels of four specific genes—TMEM273, HSF1, SLC25A17, and STT3A—in patient samples and calculating predicted risk values ​​using a specific mathematical model. The patient samples include intestinal mucosal tissue or peripheral blood. This VeRGeS-UC kit detects these four specific genes in combination, and this combination specifically maps to the α4β7 pathway-mediated immune microenvironment remodeling.

[0025] Specifically, total RNA was extracted from patients' baseline (pre-medication) samples, and the standardized expression levels of the four genes mentioned above were determined by high-throughput RNA sequencing (RNA-seq). The values, after Log2 transformation and scale standardization, were then used to calculate the risk score of VeRGeS-UC using the following formula: VeRGeS-UC=2.2454-0.5144×TMEM273+0.0468×HSF1-0.1649×SLC25A17-0.2099×STT3A.

[0026] Judgment criteria: The system's built-in optimal cutoff value (cutoff value) is 0.449 (sensitivity 50%, specificity 84.6%). When a patient's risk score is lower than the set cutoff value, they are judged as "sensitive / responsive" to vedelizumab treatment; when it is higher than the cutoff value, they are judged as "insensitive / non-responsive".

[0027] When using the test kit described in this invention, patients who are determined to be "sensitive / responsive" to vedelizumab treatment are recommended to continue vedelizumab treatment, while patients who are determined to be "insensitive / non-responsive" to vedelizumab treatment are not recommended to continue vedelizumab treatment.

[0028] Non-invasive cross-tissue application: The technical solution of this invention is not only applicable to colonoscopy biopsy tissue, but also groundbreakingly applicable to peripheral blood samples (whole blood), overcoming the problem of local signal dilution in the blood, and achieving the goal of predicting the efficacy in advance by only drawing baseline venous blood from the patient.

[0029] Example 1

[0030] This embodiment illustrates the derivation and construction process of the four-gene joint prediction formula in this invention. This embodiment uses raw RNA-seq quantitative data: that is, the most basic gene / transcriptograph expression level count results obtained from RNA-seq experiments, which has not undergone much artificial processing or standardization.

[0031] The specific scientific construction steps of the VeRGeS-UC (TMEM273, HSF1, SLC25A17, STT3A) prediction formula for the efficacy of vedelizumab proposed in this invention are as follows.

[0032] The first step involved rigorous screening and phenotypic definition of the clinical cohort. First, the inventors obtained raw RNA-seq quantitative data (n=314) from a large-scale inflammatory bowel disease patient sample dataset (E-MTAB-7914, sourced from the ArrayExpress platform). Using an automated batch processing program, the TPM (total molecular weight per million transcripts) expression levels of tens of thousands of genes / transcripts in each sample were precisely extracted. The TPM metric eliminates the influence of gene length and sequencing depth, ensuring absolute comparability of expression levels between different samples. Subsequently, the data from all samples were seamlessly stitched together to construct a global high-throughput expression matrix containing the complete gene set and the entire cohort of samples.

[0033] The second step is the standardization, noise reduction, and grouping of clinical phenotypes. To eliminate the heterogeneity of records in real-world clinical data, this invention performs rigorous standardization, cleaning, and deduplication of the original clinical metadata. 1) Patient screening: Only patients with ulcerative colitis treated with vedelizumab at baseline (week 0) are screened, n=159, including 81 responders and 78 non-responders; 2) Binarization of clinical outcomes: A strict "endoscopic remission" is used as the gold standard for efficacy. Patients who achieve endoscopic remission are defined as the "responder group," and those who do not are defined as the "non-responder group." Through the above steps, this invention excludes samples with ambiguous phenotypic records and specifically identifies a high-quality sample cohort with clear vedelizumab treatment outcomes.

[0034] The third step involves double-blind alignment of clinical phenotype and molecular data. Before statistical mining, this invention uses unique sample identifiers to perform absolute intersection matching between the selected clinical phenotype data and the column names of the expression matrix. This step eliminates the risk of misalignment between clinical information and gene expression data, ensuring that every calculation in subsequent statistical analysis accurately corresponds to a specific patient individual.

[0035] Step 4: Statistical testing of differential expression and screening of candidate targets. For the aligned data from the vedelizumab response and non-response groups, this invention designed a highly robust non-parametric statistical testing procedure. 1) Low-expression gene filtering: Background noise genes that are almost not expressed in the global sample set are removed, greatly reducing the statistical burden of subsequent multiple hypothesis testing. 2) Calculation of fold change in expression: The average TPM value of the response and non-response groups is calculated. To avoid mathematical calculation abnormalities caused by extremely low expression levels (denominator of 0 or infinity), this invention introduces a smoothing factor (+0.01) when calculating the Log2 fold change, making the assessment of fold change more robust and realistic. 3) Significance test: Considering that gene expression levels in patients in the real world often exhibit a skewed distribution rather than an ideal normal distribution, this invention uses the Wilcoxon rank-sum test (also known as the Mann-Whitney U test). This nonparametric test method is insensitive to outliers and can robustly identify genes that have a truly statistically significant difference between the responder and non-responder groups.

[0036] Using the aforementioned multi-level quality control statistical algorithm, this invention successfully calculated the P-values ​​and Log2FC values ​​of tens of thousands of transcripts between two groups. After removing invalid data, the transcripts were rigorously sorted according to statistical significance (P-value from smallest to largest), thereby precisely locating a group of core candidate genes, such as TMEM273, HSF1, SLC25A17, and STT3A, which showed statistically and biologically significant differences. These genes constitute a high-quality "feature pool" for the subsequent construction of the efficacy prediction model in this invention.

[0037] The fifth step involves dimensionality reduction and redundancy removal based on LASSO and multivariate regression. Although the previous differential analysis identified a large number of statistically significant genes, these genes exhibit severe "co-expression (collinearity)"; that is, multiple genes reflect the same change in the same biological pathway, and increasing the number does not add additional information. This invention uses a penalty term in the LASSO regression algorithm and a multivariate logistic regression model to forcibly compress redundant features, narrowing the candidate pool from hundreds to a single-digit number of independent genes with complementary predictive value—a "core driver gene pool."

[0038] Step 6: Exhaustive combination and "model efficacy saturation curve" test. To find the optimal number of genes for clinical translation, the inventors conducted iterative exhaustive search and cross-validation from N=1 to N=10 for the core candidate gene library, evaluating the area under the prediction curve (AUC) and Youden index of models with different gene numbers in independent cohort sets. 1) When the number of genes is 2 or 3 (model underfit): Although the detection cost is the lowest, the sensitivity or specificity of the model is significantly lacking due to insufficient coverage of immune / metabolic biological pathways, and the overall prediction accuracy cannot reach the threshold for clinical decision support; 2) When the number of genes is 4 (efficacy mutation point / optimal balance point): The four genes TMEM273, HSF1, SLC25A17, and STT3A form an excellent complementary network. At this point, the model's predictive efficacy reaches its "maximum marginal benefit," accurately identifying and excluding patients who do not respond to vedelizumab, thereby optimizing clinical medication decisions and reducing ineffective medical costs; 3) When the number of genes increases to 5 or more (model overfitting and diminishing marginal utility): Experiments show that adding a 5th or 6th candidate gene to the four-gene model results in a negligible improvement in the overall accuracy of the model (AUC increases by less than 0.01), and even a decrease in generalization ability in some external test sets due to "overfitting" to real noise. Meanwhile, in practical clinical applications (such as using RT-qPCR or targeted sequencing), an increase in the number of genes means a linear increase in detection costs, time costs, and quality control difficulty.

[0039] Step 7: Quantitative and Standardized Processing of Gene Expression Data. For the patient samples included in the analysis, we extracted the TPM values ​​for these four specific target genes. To eliminate systematic errors from sequencing depth and make the data more consistent with a normal distribution, we performed a logarithmic transformation of all extracted TPM values ​​using log2(TPM+1). This standardized preprocessing step is an internationally accepted practice in transcriptomics analysis, effectively reducing the interference of extreme outliers on the final model fit and improving the robustness of the model.

[0040] Step 8: Fitting the multivariate logistic regression model. After obtaining standardized expression data and clear clinical outcomes, this study employed the classic multivariate logistic regression algorithm for mathematical modeling. We used the logarithmic expression levels of four genes as continuous independent variables and the patient's clinical outcome (0 or 1) as a binary dependent variable. Logistic regression can objectively assess and quantify the independent contribution of these four genes in predicting the efficacy of vedelizumab and automatically calculate the best-fit weights.

[0041] Step 9: Mathematical Extraction and Establishment of the Prediction Formula. After model fitting, we extracted the constant term and regression coefficients corresponding to each gene in the logistic regression model. These regression coefficients, in a biological sense, represent the positive or negative influence weights of each target on the drug response. We linearly combined the above intercepts with weighted gene expression levels to finally establish the risk scoring equation of this invention: VeRGeS-UC=2.2454-0.5144×TMEM273+0.0468×HSF1-0.1649×SLC25A17-0.2099×STT3A.

[0042] Step 10: Scientific Validation and Cutoff Value Determination of the Model. To demonstrate the scientific validity of the formula, we evaluated the model's discrimination using the Receiver Operating Characteristic (ROC) curve (AUC = 0.687), see [link to relevant documentation]. Figure 1 Furthermore, decision curve analysis (DCA) confirmed that the formula has actual clinical net benefits, see [link to relevant documentation]. Figure 2 Meanwhile, by calculating and maximizing the Youden index, the optimal prediction probability cutoff value for distinguishing between effective and ineffective predictions was scientifically defined (diagnostic cutoff value = 0.449, specificity 0.846, sensitivity 0.500).

[0043] Figure 1 and Figure 2 All values ​​represent information from the training set in this invention. Training set: 159 patients, including 81 responders and 78 non-responders. Diagnostic cutoff value = 0.449, specificity 0.846, sensitivity 0.500. This invention provides a highly specific, simplified gene diagnostic model that can accurately identify and exclude patients who do not respond to vedelizumab, thereby optimizing clinical medication decisions and reducing ineffective medical costs. Table 1 shows examples of TPM levels in patient genes in the training set.

[0044] Table 1

[0045] Figure 1 ROC curves were used to predict the response of vedelizumab in patients with ulcerative colitis using the VeRGeS-UC score in the training set. Figure 1The study demonstrated the model's excellent predictive performance at the tissue level, proving that the VeRGeS-UC model can distinguish between vedelizumab responders and non-responders.

[0046] Figure 2 Decision curve analysis (DCA) was performed to predict the response of vedelizumab to patients with ulcerative colitis using the VeRGeS-UC score in the training set. In the DCA plot, the horizontal axis represents the acceptable "cutoff value probability" for both physicians and patients, i.e., the psychological expectation of being willing to accept the risk of false positives in order to achieve therapeutic efficacy; the vertical axis represents the "net clinical benefit", i.e., the net benefit of true positives minus the harm of false positives.

[0047] Figure 2 The diagram contains three key decision-making strategy curves: The first is a horizontal line with a vertical axis of zero, representing the baseline scenario assuming "no vedelizumab treatment for all patients," where the net clinical benefit is constant at zero. The second is a downward-sloping black dashed arc, representing the traditional strategy of blindly administering vedelizumab to all patients based on clinical experience without the guidance of this invention. As shown, when the cutoff probability exceeds approximately 38%, the net benefit of the "vedelizumab treatment for all patients" strategy rapidly drops below zero, becoming a negative benefit, due to the large number of non-responders suffering ineffective treatment, accompanied by side effects and high costs. The third line is a solid curve, representing the strategy of using the four-gene prediction model of this invention to guide clinical drug decisions. Figure 2 As shown, within a very broad cutoff probability range (approximately 20% to 70%) that is consistent with real-world clinical practice, the curve (solid line) of the four-gene model of this invention is consistently and significantly above the strategies of "treating all patients with vedelizumab (black dashed arc)" and "not treating all patients with vedelizumab (horizontal line)". Figure 2 This strongly demonstrates that using the four-gene combination method of this invention to screen target patients produces a significant positive net clinical benefit compared to the current clinical approaches of "treating all patients with vedelizumab" or "not treating any patients with vedelizumab." This invention not only accurately identifies patients who will benefit from treatment but also effectively avoids overtreatment of non-responders, possessing significant clinical translational value and health economic benefits.

[0048] To verify the specificity of this invention for vedelizumab, the ROC curve was also validated in the vedelizumab (anti-β7 single target) treatment cohort (GSE72819, this dataset comes from the GEO platform). Figure 3 This cohort included 70 patients with ulcerative colitis treated with itrazumab, of whom 12 achieved endoscopic remission and 58 did not. Figure 3The results showed that the model's AUC value for predicting the efficacy of ezetizumab was only 0.517. This negative result strongly confirms that the VeRGeS-UC of this invention is a specific predictive biomarker targeting the α4β7 dual-target pathway, rather than a broad-spectrum biomarker for mucosal healing or general inflammation, and possesses high target specificity.

[0049] Figure 3 To validate the ROC curve of the VeRGeS-UC score in predicting the response of ulcerative colitis patients to itrazumab (GSE72819), the figure shows that the AUC value of VeRGeS-UC for predicting itrazumab efficacy was only 0.503. This negative result indicates that the VeRGeS-UC model is a specific predictive biomarker targeting the α4β7 dual-target pathway, rather than a broad-spectrum biomarker for mucosal healing or general inflammation, and possesses high target specificity.

[0050] Furthermore, the applicant also conducted external validation of the invention in a dataset of peripheral blood samples from patients with inflammatory bowel disease treated with vedelizumab (GSE191328, also from the GEO platform). This cohort included baseline data from 17 patients, with 9 responders and 8 non-responders, and peripheral blood samples. The formula trained on intestinal mucosal tissue was directly substituted into the baseline peripheral blood samples from patients receiving vedelizumab treatment, achieving a predictive power of AUC=0.653. Figure 4 This proves that the present invention can predict therapeutic efficacy through non-invasive blood sampling, and has extremely high clinical translational value.

[0051] Figure 4 In another validation set, the VeRGeS-UC score was used to predict the ROC curve of vedelizumab response in peripheral blood samples (GSE191328) of patients with ulcerative colitis. This invention directly substituted the VeRGeS-UC formula trained in intestinal mucosal tissue into the baseline peripheral blood samples of patients receiving vedelizumab treatment in the validation set (GSE191328), achieving a predictive power of AUC=0.653. This demonstrates that this invention can predict efficacy through non-invasive blood sampling and has high clinical translational value.

[0052] Figure 4 The AUC of the training set for blood tests on patients was 0.653. For gene testing, an AUC greater than 0.6 is considered effective. Especially for non-invasive blood tests on patients, the kit described in this invention achieves an AUC greater than 0.6, which is an unprecedented breakthrough.

[0053] Example 2

[0054] This embodiment validates the RT-qPCR prediction model based on the baseline VeRGeS-UC of a real-world cohort and establishes a simplified scoring system. Considering the convenience and cost of this patent in clinical applications, this embodiment no longer uses raw RNA-seq quantitative data, but instead uses RT-qPCR technology to detect the relative concentration of RNA, i.e., its cycle number (Ct value). Because this embodiment uses RT-qPCR technology, the formula (risk scoring equation) in Example 1 is no longer applicable and needs to be simplified to the simplified scoring system in this embodiment. The content includes the following steps.

[0055] Step 1: Retrospective Collection and Inclusion Criteria of Clinical Samples. To further validate the predictive efficacy of the above-mentioned four biomarker gene combination VeRGeS-UC in real-world clinical practice and to establish a convenient detection method for clinical translation, this invention conducted a retrospective cohort study. The inventors screened patients with moderate to severe ulcerative colitis who had previously received vedelizumab treatment by reviewing the hospital's electronic medical record system. Based on the clinical follow-up outcome (i.e., whether endoscopic mucosal healing was achieved) at week 14 after vedelizumab treatment, 15 patients confirmed as "treatment response" and 15 patients confirmed as "non-response" were strictly selected, totaling 30 patients. Patient clinical information is as follows: the mean ± standard deviation of age in the response group was 47 ± 14.97, and the mean ± standard deviation of age in the non-response group was 44.67 ± 14.99; the gender composition of the response group was 9 males and 6 females, and the gender composition of the non-response group was also 9 males and 6 females. Subsequently, the inventors extracted frozen intestinal mucosal biopsy tissue and peripheral blood cDNA samples from the aforementioned 30 patients before administration of vedelizumab (baseline, Week 0) from the hospital's biobank.

[0056] Step 2: RT-qPCR detection of target gene expression levels. Using the primer combination provided in this invention, the expression levels of four target genes and internal reference genes (β-actin, GAPDH) were detected by RT-qPCR technology. A total of 60 samples were collected from intestinal mucosa and peripheral blood of 30 patients, with three replicates for each sample. After the reaction, the cycle cutoff value (Ct value) of each gene was obtained. The relative expression level of the target gene relative to the internal reference gene was calculated using the ΔCt method.

[0057] Among them, the full name of gene TMEM273 is transmembrane protein 273, and its upstream and downstream primers are SEQ ID NO:1 and SEQ ID NO:2, respectively. The full name of gene HSF1 is heat shock transcription factor 1, and its upstream and downstream primers are SEQ ID NO:3 and SEQ ID NO:4, respectively. The full name of gene SLC25A17 is solute carrier family 25 member 17, and its upstream and downstream primers are SEQ ID NO:5 and SEQ ID NO:6, respectively. The full name of gene STT3A is STT3 oligosaccharide transferase complex catalytic subunit A, and its upstream and downstream primers are SEQ ID NO:7 and SEQ ID NO:8, respectively. SEQ ID NO:1: TGGCTGGTGCTATGTGGTTAC. SEQ ID NO:2: GCAAGCAGTCAGAAGG GAAC. SEQ ID NO:3: ATGGAAGCGTGCGTGTCTAC. SEQ ID NO:4: TCGGGACGC ATAGAGGAA. SEQ ID NO:5: GCCCGGGAGTTTCTACTCATT. SEQ ID NO:6: GGCATACCCGTATTCTTGACC. SEQ ID NO:7: GCAGACAGACGAAACAGGGA. SEQ ID NO:8:GGCAGTAAGAATCCCAGCAG.

[0058] Figure 5 This figure shows the amplification curves of four genes using the baseline VeRGeS-UC method based on a real-world cohort. The figure includes the results of 60 test samples from 30 patients, with three replicates per sample. The multiple S-shaped curves represent the amplification dynamics of the target genes in different samples during PCR cycles. The horizontal line in the figure (the horizontal line with a vertical axis of 0.105899) represents the set fluorescence signal threshold. The x-axis value corresponding to the intersection of each amplification curve with this fluorescence threshold line is the cycle threshold (Ct value) of the reaction system. The Ct value is used for subsequent relative quantitative analysis and model calculations of the expression levels of the four genes.

[0059] Step 3: Construction of a simplified clinical predictive scoring system (0-4 points). Given the difficulty in widely adopting high-throughput sequencing computational models in routine clinical testing, this invention constructs a simplified clinical scoring system based on the RT-qPCR results of the aforementioned 30 patient samples, without relying on complex computer algorithms. To evaluate the efficacy of the four genes in independently predicting treatment response, the inventors performed ROC curve analysis on the ΔCt values ​​of the 30 patients (30 colonic mucosal tissue samples and 30 blood samples). The results showed that all four target genes exhibited extremely excellent independent diagnostic efficacy. By calculating the maximum Yangen index, the optimal cutoff values ​​for predicting "response" and "no response" for each gene were determined. Specific statistical parameters are shown in Tables 2 and 3 below. Table 2 shows the univariate ROC analysis results of the four target genes in the patient's colonic mucosal tissue samples. Table 3 shows the univariate ROC analysis results of the four target genes in the patient's peripheral blood samples.

[0060] Table 2

[0061] Table 3

[0062] Step 4: Construction of a simplified multi-gene scoring system. To facilitate clinical application and overcome the complexity of continuous variables in clinical decision-making, this invention constructs a "0-4 point integer scoring system" based on the cutoff values ​​determined in Tables 2 and 3. The specific scoring rules are as follows: For colon tissue samples: 1) If the ΔCT value of the TMEM273 gene is <8.3289, score 1 point; otherwise, score 0 points; 2) If the ΔCT value of the HSF1 gene is >7.5031, score 1 point; otherwise, score 0 points; 3) If the ΔCT value of the SLC25A17 gene is <6.8778, score 1 point; otherwise, score 0 points; 4) If the ΔCT value of the STT3A gene is <7.5610, score 1 point; otherwise, score 0 points. For peripheral blood samples: 1) If the ΔCT value of the TMEM273 gene is <8.0077, score 1 point; otherwise, score 0 points. 2) If the ΔCT value of the HSF1 gene is >7.8050, score 1 point; otherwise, score 0 points. 3) If the ΔCT value of the SLC25A17 gene is <7.1951, score 1 point; otherwise, score 0 points. 4) If the ΔCT value of the STT3A gene is <7.7621, score 1 point; otherwise, score 0 points. The scores for these four genes are summed to obtain each patient's "baseline predicted total score" (range 0-4 points). A higher score indicates that the gene expression profile is more consistent with the trend of treatment response.

[0063] Step 5: Determination of Clinical Decision Cutoff Values ​​and Model Validation. The 0-4 scoring system constructed above was used to score the 30 clinical samples in this embodiment. Statistical results showed that the scoring model could significantly distinguish between the two groups of patients: the scores of the response group were highly concentrated in the 3 and 4 score ranges; while the scores of the non-respondent group were mainly distributed in the low score range. To determine the optimal clinical decision cutoff value, the inventors conducted a statistical diagnostic efficacy evaluation of the various cutoff values ​​of the discrete integer scoring system. The results showed that when "total score ≥ 3 points" was used as the positive cutoff value for determining patient response, the system achieved a sensitivity of 80.0% (12 / 15) and a specificity of 86.7% (13 / 15) for both colonic mucosal tissue samples and peripheral blood samples, with a Youden index of 0.667. This cutoff value not only mathematically achieved the best balance between high sensitivity and high specificity, but also perfectly met the practical needs of clinical medication in balancing "no missed diagnoses (high sensitivity)" and "avoiding ineffective overtreatment (high specificity)."

[0064] This embodiment confirms that by specifically detecting and discretely assigning values ​​to the expression levels of four genes—HSF1, TMEM273, SLC25A17, and STT3A—in colonic mucosal tissue and / or peripheral blood samples, a baseline predicted total score ≥3 points can serve as a reliable and objective basis for strongly recommending vedelizumab treatment in clinical practice. A total score ≤2 points indicates an extremely low response rate, and clinicians are advised to consider other alternative treatment options, such as the use of JAK inhibitors.

[0065] In summary, this invention provides a diagnostic kit (VeRGeS-UC) for predicting the efficacy of vedelizumab (targeting α4β7 integrin) in patients with ulcerative colitis. This invention primarily addresses the issues of "blindly trying drugs" and "invasive testing" faced by ulcerative colitis patients receiving biologic therapy, specifically solving the following four core technical challenges: 1. Addressing the lack of accurate predictive biomarkers for vedelizumab efficacy: Currently, there is a lack of objective molecular biomarkers to predict the efficacy of vedelizumab before clinical decision-making. 2. Addressing the lack of "drug target specificity" in existing pan-inflammatory biomarkers: Existing biomarkers cannot distinguish whether a patient is sensitive to anti-TNF-α drugs, anti-β7 single-target drugs (such as itrazizumab), or α4β7 dual-target antagonists (such as vedelizumab). 3. Addressing the problem of traditional testing heavily relying on "invasive colonoscopy biopsy": Existing gene testing is mostly based on intestinal mucosal tissue, requiring invasive colonoscopy to obtain samples, resulting in poor patient compliance and infrequent monitoring. This invention aims to overcome the barriers between tissues and peripheral blood to achieve non-invasive blood sampling prediction. 4. It addresses the problems of wasted medical resources and delayed optimal treatment: Vedelizumab has a slow onset of action and is expensive. Accurately excluding patients with primary non-responders can prevent patients from missing other treatment windows.

[0066] This invention belongs to the field of medical testing, specifically relating to a test kit for predicting the response of vedelizumab treatment in patients with ulcerative colitis, the use of reagents in the preparation of the kit, and an integrated detection system. The kit includes reagents for detecting the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A in subject samples; the expression levels of these four genes are used to predict and assess whether the patient responds to vedelizumab. This invention exhibits high specificity for vedelizumab targeting α4β7, but no predictive efficacy for etrazumab targeting β7, demonstrating its non-pan-inflammatory markers and high targeting precision. This invention achieves a breakthrough in "liquid biopsy": the histological model demonstrates significant predictive efficacy in an independent peripheral blood cohort, overcoming the technical bottleneck that "intestinal local immune genes cannot be detected through blood sampling." Clinical trials have shown that the 0-4 point simplified scoring system constructed based on peripheral blood samples in this invention has a sensitivity of 80.0% and a specificity of 86.7% in predicting the response to vedelizumab treatment when the total score is ≥3 points, which greatly reduces the detection threshold and patient suffering.

[0067] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A diagnostic kit for predicting the response of vedelizumab treatment in patients with ulcerative colitis, characterized in that, The kit includes reagents for detecting the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A in subject samples; the expression levels of these genes are used to predict and assess whether the patient responds to vedelizumab.

2. The detection kit according to claim 1, characterized in that, The subject samples were selected from any one or more of the following: intestinal mucosal biopsy tissue, peripheral blood, serum, plasma, or peripheral blood mononuclear cells.

3. The detection kit according to claim 1, characterized in that, The reagents include primer pairs and / or probes for detecting the expression levels of the genes TMEM273, HSF1, SLC25A17, and STT3A by reverse transcription real-time quantitative PCR, i.e., RT-qPCR.

4. The detection kit according to claim 3, characterized in that, The primer pairs include SEQ ID NO:1 and SEQ ID NO:2 for the TMEM273 gene, SEQ ID NO:3 and SEQ ID NO:4 for the HSF1 gene, SEQ ID NO:5 and SEQ ID NO:6 for the SLC25A17 gene, and SEQ ID NO:7 and SEQ ID NO:8 for the STT3A gene. SEQ ID NO: 1: TGGCTGGTGCTATGTGGTTAC; SEQ ID NO:2: GCAAGCAGTCAGAAGGGAAC; SEQ ID NO:3:ATGGAAGCGTGCGTGTCTAC; SEQ ID NO:4:TCGGGACGCATAGAGGAA; SEQ ID NO:5: GCCCGGGAGTTTCTACTCATT; SEQ ID NO:6: GGCATACCCGTATTCTTGACC; SEQ ID NO:7: GCAGACAGACGAAACAGGGA; SEQ ID NO:8:GGCAGTAAGAATCCCAGCAG.

5. The detection kit according to claim 4, characterized in that, The detection kit also includes reagents required for RT-qPCR detection, which include reverse transcriptase, DNA polymerase, dNTPs, fluorescent dyes or fluorescently labeled probes, and PCR reaction buffer.

6. The detection kit according to claim 1, characterized in that, The reagents include library construction reagents for detecting the expression levels of the genes TMEM273, HSF1, SLC25A17, and STT3A by RNA sequencing, i.e., RNA-seq.

7. The detection kit according to claim 1, characterized in that, The kit also includes reagents for detecting the expression levels of internal reference genes, which are selected from one or more of GAPDH, ACTB, 18S rRNA, and β-actin, and are used to correct the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A.

8. The use of a reagent in the preparation of a reagent kit, characterized in that, Specifically, the use of expression level detection reagents for genes TMEM273, HSF1, SLC25A17, and STT3A in the preparation of a kit for predicting the efficacy of vedelizumab in patients with ulcerative colitis, the kit being used to predict and evaluate the treatment response of vedelizumab by detecting the expression levels of the four genes in samples from patients with ulcerative colitis.

9. The use according to claim 8, characterized in that, The prediction and assessment are based on a comprehensive score calculated from the expression levels of genes TMEM273, HSF1, SLC25A17, and STT3A. When the expression level of a specific gene combination is higher or lower than a preset cutoff value, the patient is determined to be responsive or non-responsive to vedelizumab.

10. An integrated detection system for predicting the efficacy of vedelizumab in patients with ulcerative colitis, characterized in that, include: The reagent compartment is used to carry the reagents in the detection kit according to any one of claims 1 to 7, the amplification detection module is used to perform RT-qPCR or RNA-seq, and the data processing module is used to collect fluorescence signals and output efficacy prediction results according to a preset algorithm.

Images