A combination of sugar chain markers for identifying severity of pancreatitis and application thereof

Abstract

The application discloses a sugar chain marker combination for identifying severity of pancreatitis and application thereof, and belongs to the technical field of biomedical detection. The sugar chain marker combination is composed of 10 N-glycan chains of NGA2F, NG1A2F-1, NG1A2F-2, NA2, NA2F, NA2FB, NA3Fb, NA4, NA4Fb and NA4F2b, and can specifically distinguish mild, moderate and severe pancreatitis. The marker combination has high specificity and strong stability, the detection sample is non-invasive and convenient, the prediction model has high accuracy, the limitation of traditional detection means can be effectively overcome, reliable basis is provided for grading diagnosis and individualized treatment of pancreatitis, and the marker combination has important clinical transformation value and application prospect.

Description

Technical Field

[0001] This invention belongs to the field of biomedical detection technology, specifically relating to a combination of glycan biomarkers for identifying the severity of pancreatitis and their application. Background Technology

[0002] Pancreatitis is a common acute abdominal condition in clinical practice. Based on severity, it can be classified into mild acute pancreatitis (MAP), moderate to severe acute pancreatitis (MSAP), and severe acute pancreatitis (SAP). SAP is particularly dangerous, often accompanied by serious complications such as pancreatic necrosis, infection, systemic inflammatory response syndrome (SIRS), and multiple organ failure (MOF), with a mortality rate as high as 15%–30%. Therefore, early and accurate identification of the severity of pancreatitis is of crucial clinical significance for timely and targeted treatment (such as fluid resuscitation, nutritional support, intensive care intervention, and even surgery), rational allocation of medical resources, and improvement of patient prognosis.

[0003] Currently, the main methods used in clinical practice to assess the severity of pancreatitis include the following categories: 1. Clinical scoring system Currently, widely used scoring systems include the Ranson score, APACHE II score, BISAP score, and CTSI score. These scoring systems assess the condition by integrating multiple clinical indicators (such as age, laboratory test results, and organ function status), but they have significant shortcomings: First, some scoring indicators can only be collected and calculated 48 to 72 hours after admission, resulting in delayed assessment results and making it difficult to meet the needs of early intervention; second, the numerous scoring indicators and complex calculation process limit their operability in practical clinical applications; third, scoring results are affected by subjective judgment factors, and differences may exist between different assessors, affecting the consistency and comparability of the scores.

[0004] 2. Imaging examinations Enhanced CT scans are currently an important imaging tool for assessing the extent of pancreatic necrosis and peripancreatic complications. Especially when combined with the CTSI scoring system, they can provide a relatively intuitive anatomical basis for the severity of the disease. However, imaging examinations have the following limitations: First, in the early stages of the disease (usually within 72 hours of onset), pancreatic necrosis is not yet fully apparent, resulting in lower sensitivity and accuracy of CT scans, which can easily lead to missed diagnoses or underestimations. Second, CT scans are expensive and carry radiation exposure risks, making them unsuitable for repeated dynamic monitoring. Third, imaging assessment is limited for some patients who cannot tolerate contrast agents or are critically ill and unable to be transported.

[0005] 3. Serum biochemical indicators Currently, commonly used serological markers in clinical practice include C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α). These indicators can reflect the level of the body's inflammatory response to some extent, but they have several limitations: while CRP is widely used, its peak value appears relatively late (usually 48–72 hours after onset) and lacks specificity; various factors such as infection, trauma, and surgery can cause its elevation. PCT has some value in differentiating infectious pancreatic necrosis, but its ability to differentiate non-infectious severe pancreatitis is limited. Although cytokines such as IL-6 react earlier, their short half-life and poor detection stability limit their clinical application. Overall, existing serological markers do not achieve ideal levels of sensitivity and specificity in differentiating between mild and severe pancreatitis, making them unsuitable as independent diagnostic criteria.

[0006] 4. Current Status of Research on Emerging Biomarkers In recent years, with the development of glycomics technology, the role of glycosylation modification in inflammatory responses, immune regulation, and disease progression has received increasing attention. As an important component of protein post-translational modifications, the structural changes of glycans can directly reflect dynamic changes in cellular state, signaling pathway activation, and the tissue microenvironment. Studies have shown that under acute inflammation and tissue injury conditions, the expression profiles of glycans in serum and tissues undergo significant alterations, such as increased fucosylation levels and changes in sialylation patterns. Compared with traditional protein biomarkers, glycan biomarkers exhibit higher stability and disease specificity, and can be accurately detected using high-throughput technologies such as mass spectrometry and lectin chips, demonstrating their potential as novel biomarkers.

[0007] Although existing studies have explored the diagnostic value of glycans in tumors, autoimmune diseases, and inflammatory diseases, research on specific glycan biomarkers for differentiating the severity of pancreatitis is still in its early stages, with few related reports and no mature products approved for clinical application. Therefore, developing a combination of glycan biomarkers and a detection method that can identify the severity of pancreatitis early, rapidly, and accurately has significant clinical translational value and application prospects. Summary of the Invention

[0008] Technical problem solved: To address the above-mentioned technical problems, this invention provides a combination of glycan biomarkers for identifying the severity of pancreatitis and its application, which can effectively solve the problems of long detection time, low sensitivity, poor specificity and complicated operation of biomarkers in the prior art, and realize early, rapid and accurate identification of the severity of pancreatitis.

[0009] Technical Solution: A combination of glycan biomarkers for identifying the severity of pancreatitis, wherein the combination of glycan biomarkers comprises NGA2F (non-galactosyl α-1,6 core fucosylated biantennary N-glycan), NG1A2F-1 (mono-branched galactosyl α-1,6 core fucosylated biantennary N-glycan), NG1A2F-2 (mono-branched galactosyl α-1,6 core fucosylated biantennary N-glycan), NA2 (galactosylated biantennary N-glycan), and NA2F (galactosylated α-1,6 core fucosylated biantennary N-glycan). The formula consists of NA2FB (galactosyl α-1,6 core fucosylated bi-antenna N-glycan), NA3Fb (P8, galactosylated α-1,3 branched fucosylated triantenna N-glycan), NA4 (galactosylated tetraantenna N-glycan), NA4Fb (galactosylated α-1,3 core fucosylated tetraantenna N-glycan), and NA4F2b (galactosylated di-α-1,3 branched fucosylated tetraantenna N-glycan); among which, NG1A2F-1 and NG1A2F-2 are isomers.

[0010] Secondly, the present invention provides a method for constructing a predictive model for identifying the severity of pancreatitis, comprising the following steps: Step 1) Obtain a training sample set, which includes biological samples with known pancreatitis grades and their corresponding grade labels; the pancreatitis grades include mild acute pancreatitis (MAP), moderate to severe acute pancreatitis (MSAP), and severe acute pancreatitis (SAP). Step 2) Detect the abundance of each glycan in the above glycan marker combination for each sample in the training sample set to obtain glycan abundance data; Step 3) Using the glycan abundance data as the independent variable and the graded labels as the dependent variable, a machine learning algorithm is used to train the model to obtain the prediction model.

[0011] Preferably, the biological sample in step 1) is blood, serum or plasma derived from venous blood or peripheral blood of a pancreatitis patient; the sample is easy to obtain and non-invasive, and can meet the needs of repeated clinical sampling and monitoring.

[0012] Preferably, the method for detecting oligosaccharide chain abundance in step 2) is capillary electrophoresis.

[0013] Preferably, the machine learning algorithm in step 3) is at least one of logistic regression, support vector machine (SVM), random forest, gradient boosting decision tree, extreme gradient boosting, artificial neural network or convolutional neural network.

[0014] Thirdly, the present invention provides a pancreatitis grading prediction model, which is obtained by the construction method described in the second aspect. Its input is the abundance data of 10 glycan markers in the sample to be tested, and its output is the grading label of mild, moderate or severe pancreatitis and the corresponding prediction probability, which can quickly output accurate grading results.

[0015] Fourthly, the present invention provides a pancreatitis grading prediction system, comprising: 1) Data acquisition module, used to acquire the abundance information of each glycan in the above glycan marker combination in the sample to be tested; 2) Prediction module, which stores the pancreatitis grading prediction model mentioned above, and calls the model to analyze and calculate the abundance information; 3) Output module, configured to output the grading prediction results calculated by the pancreatitis grading prediction model based on abundance information, including grading labels and corresponding prediction probabilities.

[0016] Fifthly, the present invention provides the application of the combination of glycan markers described in the first aspect in the preparation of a kit for assisting in the differentiation of pancreatitis severity; the kit can rapidly detect the abundance of 10 glycans in a sample, providing a reliable basis for the differentiation of pancreatitis severity.

[0017] In a sixth aspect, the present invention provides a kit for assisting in the identification of pancreatitis grading, comprising reagents for detecting the abundance of each glycan in the above-mentioned glycan marker combination, wherein the reagents can specifically bind to each target glycan to achieve accurate detection of glycan abundance.

[0018] Beneficial effects: (1) This invention first discovered and verified a combination of markers composed of 10 specific N-glycans, whose abundance information is highly specific to different grades of pancreatitis (mild, moderate and severe, severe), providing a new biomarker system for the field of pancreatitis grading, breaking through the limitations of traditional biochemical indicator detection, and filling the technical gap of specific glycan markers for pancreatitis grading. (2) The present invention combines specific glycan markers with machine learning algorithms (such as support vector machines) to construct a glycan pancreatitis grade prediction model (GPGI), which shows excellent predictive performance. Experimental results show that the model has a total grade accuracy rate of over 90% in both the training and validation sets. Among them, the grade accuracy rate can reach 100% for severe pancreatitis that urgently needs timely intervention in clinical practice. It effectively overcomes the problems of insufficient sensitivity and high false negative rate of traditional markers, and provides a reliable means for the early identification of critically ill patients. (3) The test sample used in this invention is blood (blood, serum, or plasma from venous blood or peripheral blood). The sample acquisition method is simple and non-invasive, and the patient compliance is good. Compared with imaging examinations, the detection method of this invention does not require special equipment for transportation, and is easier to repeat sampling and dynamic monitoring, which creates favorable conditions for the promotion and application of primary hospitals and the long-term management of patients' conditions. (4) The glycans detected by this invention can undergo specific structural and abundance changes before the occurrence of cytokine storm, which can reflect the severity of the disease earlier than traditional biochemical indicators (such as CRP), and can accurately reflect specific pathophysiological changes (such as immune cell activation and tissue remodeling) during the progression of pancreatitis. Its diagnostic timeliness and specificity are superior to single protein markers. (5) Through the grading prediction method provided by the present invention, clinicians can quickly and accurately obtain the grading information of patients' pancreatitis, and then formulate individualized treatment plans (such as whether to conduct close follow-up, intensive care or surgical intervention), effectively optimize the allocation of medical resources, and ultimately significantly improve the treatment effect and quality of life of patients. Attached Figure Description

[0019] Figure 1 The images show the serum glycan profiles of patients with mild, moderate, and severe pancreatitis in Example 1; where the horizontal axis represents time and the vertical axis represents altitude. Detailed Implementation

[0020] The present invention will be described in detail below with reference to specific embodiments. Unless otherwise specified, the technical means used in the following embodiments are all conventional means well known to those skilled in the art, and the experimental methods without specific conditions are all conventional methods in this field. Unless otherwise specified, the materials and reagents used can be obtained from commercial channels.

[0021] Example 1

[0022] 1. Test Sample This study collected serum samples from 447 subjects with different grades of pancreatitis as a dataset, including 294 cases of mild acute pancreatitis (MAP), 117 cases of moderate to severe acute pancreatitis (MSAP), and 36 cases of severe acute pancreatitis (SAP). All samples were obtained from Zhuzhou Central Hospital, and the experimental protocol has been filed with and approved by the hospital's ethics committee.

[0023] The dataset was randomly divided into a training set and a validation set in an 8:2 ratio: the training set consisted of 357 cases (MAP 235 cases, MSAP 93 cases, SAP 29 cases) for model building; the validation set consisted of 90 cases (MAP 59 cases, MSAP 24 cases, SAP 7 cases) for model performance validation.

[0024] 2. Instruments and equipment Capillary electrophoresis analyzer, PCR, centrifuge.

[0025] 3. Test reagents Reagent A: 5 mM NH4HCO3 added to 1% SDS solution; Reagent B: Add 2 U / μL of exoglycoside exonuclease solution to 1% NP-40; Reagent C: Add 2 U / μL of sialidase solution to 100 mM, pH 5 NH4AC; Reagent D: ddH2O; Reagent E: A solution prepared by mixing 5 mM fluorescent labeling solution (trisodium 8-aminopyrene-1,3,6-trisulfonic acid) with DMSO solution (organic reducing agent NaBH3CN concentration of 1 M).

[0026] 4. Glycan mapping detection and abundance collection 1) Release of sugar chains Take 5 μL of serum sample and place it in a centrifuge tube. Add 3 μL of reagent A, vortex to mix, and heat at 95℃ for 5 min to denature the protein in the sample. After cooling to room temperature, add 3 μL of reagent B and 4 μL of reagent C, vortex to mix, and react at 37℃ for 4 h to complete the release of glycans from the protein. After the reaction is complete, add 80 μL of reagent D, vortex to mix, and terminate the enzymatic reaction.

[0027] 2) Fluorescent labeling of sugar chains Take 10 μL of the above enzymatically digested sample solution and dry it at 70℃ for 30 min. After drying, add 3 μL of reagent E, vortex mix, and react at 90℃ for 2 h to achieve fluorescent labeling of the glycans. After the reaction is completed, add 80 μL of reagent D, vortex mix, and terminate the labeling reaction.

[0028] 3) Glycan mapping detection and abundance collection Take 10 μL of fluorescently labeled glycan sample, place it in an ABI 96-well plate, and detect it using an ABI 3500 sequencer to obtain N-glycan maps.

[0029] The glycan maps were analyzed using spectral analysis software to obtain the relative abundance of 10 target glycans (NGA2F, NG1A2F-1, NG1A2F-2, NA2, NA2F, NA2FB, NA3Fb, NA4, NA4Fb, NA4F2b). Among them, NG1A2F-1 and NG1A2F-2 are isomers.

[0030] 5. Data Analysis and Processing from Figure 1It can be seen that there are significant differences in serum glycan profiles among patients with different grades of pancreatitis, indicating that the target glycan combination has the potential to differentiate the severity of pancreatitis. The specific analysis process is as follows: (1) Screening of characteristic sugar chains The abundance data of 10 glycans in patients with different grades of pancreatitis in the training set (357 cases) were compared and analyzed. Glycans with statistical p-values ​​less than 0.05 were selected as biomarkers for the prediction model, as shown in Table 1 below: Table 1. Comparative analysis of glycan abundance in patients with different grades of pancreatitis in the training set (mean ± standard deviation) , As shown in Table 1, the abundance of the 10 sugar chains differed significantly among patients with different grades of pancreatitis (P < 0.05), and can be used as characteristic markers for pancreatitis grading.

[0031] (2) Construction and validation of the pancreatitis grading prediction model (GPGI) Using the abundance data of 10 characteristic glycans in the training set as independent variables and the clinically diagnosed pancreatitis grading labels (mild, moderate, and severe) as dependent variables, a grading prediction model GPGI was constructed using the Support Vector Machine (SVM) algorithm. During model construction, the model parameters were optimized using a grid search method, and the optimal parameters were finally determined as: kernel='rbf', C=1, gamma=1, decision_function_shape=ovr. The model outputs the predicted probability for each grade, and the grade with the highest probability value is taken as the final prediction result.

[0032] The model performance was independently validated using a validation set (90 cases). The concordance rates between the model prediction results and the clinical grading results are shown in Tables 2 and 3 below.

[0033] Table 2. Concordance rate between the GPGI model in the training set and the clinical pancreatitis grading results. , As shown in Table 2, the overall concordance rate between the GPGI model in the training set and the clinical pancreatitis grading results reached 99.44%, with a prediction concordance rate of 100.00% for mild and severe cases and 97.85% for moderate and severe cases. The model demonstrated excellent predictive performance in the training set.

[0034] Table 3. Concordance rate between the GPGI model in the validation set and the clinical pancreatitis grading results. , As shown in Table 3, the overall concordance rate between the GPGI model in the validation set and the clinical pancreatitis grading results was 96.67%. Among them, the prediction concordance rate for mild and severe cases remained at 100.00%, and the prediction concordance rate for moderate and severe cases was 87.50%, indicating that the model has good generalization ability and stability and can be effectively applied to the clinical differentiation of pancreatitis severity.

[0035] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A combination of glycan biomarkers for identifying the severity of pancreatitis, characterized in that: The glycan marker combination consists of NGA2F, NG1A2F-1, NG1A2F-2, NA2, NA2F, NA2FB, NA3Fb, NA4, NA4Fb and NA4F2b; wherein NG1A2F-1 and NG1A2F-2 are isomers.

2. A method for constructing a predictive model for identifying the severity of pancreatitis, characterized in that, Includes the following steps: Step 1) Obtain the training sample set, which includes biological samples with known pancreatitis grades and their corresponding grade labels; Step 2) Detect the abundance of each glycan in the glycan marker combination of each sample in the training sample set to obtain glycan abundance data; Step 3) Using the glycan abundance data as the independent variable and the graded labels as the dependent variable, a machine learning algorithm is used to train the model to obtain the prediction model; The glycan marker combination consists of NGA2F, NG1A2F-1, NG1A2F-2, NA2, NA2F, NA2FB, NA3Fb, NA4, NA4Fb and NA4F2b; wherein NG1A2F-1 and NG1A2F-2 are isomers.

3. The construction method according to claim 2, characterized in that: Step 1) The biological sample is blood, serum or plasma derived from venous or peripheral blood of a pancreatitis patient.

4. The construction method according to claim 2, characterized in that: The method for detecting oligosaccharide chain abundance in step 2) is capillary electrophoresis.

5. The construction method according to claim 2, characterized in that: Step 3) The machine learning algorithm is at least one of logistic regression, support vector machine, random forest, gradient boosting decision tree, extreme gradient boosting, artificial neural network or convolutional neural network.

6. A pancreatitis grading prediction model, characterized in that: The prediction model is obtained by the construction method described in any one of claims 2-5. Its input is the abundance data of 10 glycan markers in the sample to be tested, and its output is the grade label of mild, moderate or severe pancreatitis and the corresponding predicted probability.

7. A pancreatitis grading prediction system, characterized in that, include: 1) Data acquisition module, used to acquire the abundance information of each glycan in the combination of glycan markers in the sample to be tested; 2) A prediction module, which stores the pancreatitis grading prediction model as described in claim 6, and calls the model to analyze and calculate the abundance information; 3) Output module, configured to output the grading prediction results calculated by the pancreatitis grading prediction model based on abundance information, including grading labels and corresponding prediction probabilities.

8. The use of the combination of glycan markers according to claim 1 in the preparation of a kit for assisting in the differentiation of pancreatitis grading.

9. A kit for assisting in the differentiation of pancreatitis grading, characterized in that, The apparatus contains reagents for detecting the abundance of each glycan in a glycan biomarker combination, the glycan biomarker combination consisting of NGA2F, NG1A2F-1, NG1A2F-2, NA2, NA2F, NA2FB, NA3Fb, NA4, NA4Fb and NA4F2b; wherein NG1A2F-1 and NG1A2F-2 are isomers.

Images