Marker combination for diagnosing active tuberculosis and distinguishing latent tuberculosis infection and active tuberculosis and application thereof

Abstract

The invention relates to the technical field of diagnostic markers, in particular to a marker combination for diagnosing active tuberculosis and distinguishing latent tuberculosis infection and active tuberculosis and application of the marker combination. The lectin combination which can be used for ATB diagnosis and can distinguish LTBI and ATB is obtained by detecting the IgG glycosylation condition in the onset process of tuberculosis, and the lectin combination can be used as a marker for ATB diagnosis and distinguishing LTBI and ATB, and has high specificity and sensitivity; a new thought and method are provided for distinguishing different stages of tuberculosis and early accurate diagnosis of ATB, and the lectin combination is combined with the tuberculosis antigen to detect the specific antibody, so that the diagnosis effect can be further improved. The marker combination and the detection product thereof provided by the invention have relatively good application potential in ATB diagnosis and LTBI and ATB distinguishing.

Description

Technical Field

[0001] This invention relates to the field of diagnostic biomarker technology, and more particularly to biomarker combinations and their applications for diagnosing active tuberculosis and differentiating between latent tuberculosis infection and active tuberculosis. Background Technology

[0002] Tuberculosis (TB) is a chronic infectious disease primarily transmitted through the respiratory tract, caused by Mycobacterium tuberculosis, and poses a serious threat to human health. Clinically, approximately 90% of people infected with Mycobacterium tuberculosis do not immediately develop the disease but enter a latent infection stage, known as latent infection individuals (LTBI). LTBI is the main source of active tuberculosis (ATB), with about 5%-10% of LTBI cases eventually progressing to ATB. One of the most critical issues in global TB control strategies is the early detection and diagnosis of cases. Therefore, timely diagnosis of TB and the differentiation between LTBI and ATB are crucial for the prevention and control of TB.

[0003] Currently, tuberculosis diagnosis faces many challenges, particularly for long-term tuberculosis (LTBI), where there is still no clear gold standard for diagnosis and an effective biomarker for diagnosing acute tuberculosis (ATB) and differentiating between LTBI and ATB. The most widely used methods for diagnosing LTBI are the tuberculin skin test (TST), the interferon-gamma release assay (IGRA), and some emerging technologies developed based on these two methods. These methods are mainly used for indirect diagnosis of LTBI, but they still have certain limitations, especially since none of them can directly distinguish between LTBI and ATB populations. Patent application WO2019 / 084051A1 discloses a method for differentiating ATB from LTBI, which uses four different tuberculosis antigens (PPD, LAM, Ag85A, CFP10) to capture antibodies and uses HRP-labeled anti-human IgG antibodies and two lectins (SNA and RCA1, which have affinity for sialic acid and galactose, respectively) for detection to determine the antibody glycosylation status. This method uses the specific antibody titers of four different Mycobacterium tuberculosis antigens and their associated lectin binding titers to distinguish between ATB and LTBI. While it can differentiate between ATB and LTBI, it only applies to TB patients and cannot diagnose ATB in healthy individuals. Furthermore, it relies on the pre-capture of specific antibodies by the four tuberculosis antigens, resulting in cumbersome and costly procedures. Similarly, WO2016 / 064955A1 also suffers from the problem of only applying to TB patients and not being able to diagnose ATB in healthy individuals. Moreover, it only discloses that compared to ATB patients, LTBI patients have reduced pro-inflammatory glycans in the galactose structure of tuberculosis-specific IgG and increased anti-inflammatory glycans in the 2-galactose (G2) unit, without disclosing biomarkers that can accurately distinguish between ATB and LTBI.

[0004] The lack of effective diagnostic tools to differentiate between LTBI and ATB significantly hinders the early identification and accurate diagnosis of ATB, and is a potential cause of high morbidity and mortality in tuberculosis. Therefore, developing biomarkers and diagnostic methods to differentiate between LTBI and ATB is of great importance for tuberculosis prevention and control.

[0005] Antibodies are important immune effector molecules in serum, especially IgG molecules, which can exert different immune effector mechanisms through various glycosylation modifications. Studies have shown that the glycosylation level of IgG changes with physiological and immune status, and its glycosylation pattern can reflect disease progression, indicating its great potential as a biomarker in early diagnosis, disease differentiation, and prognosis. Lectins are widely distributed proteins in nature that can specifically bind to glycan molecules at the ends of glycoproteins and are widely used in research related to glycosylation. Summary of the Invention

[0006] This invention provides a combination of biomarkers for diagnosing active tuberculosis and for differentiating between latent tuberculosis infection and active tuberculosis, and their applications.

[0007] This invention utilizes lectin microarray technology to detect IgG glycosylation during the pathogenesis of tuberculosis. After processing the microarray data, lectins and their combinations suitable for the differential diagnosis of tuberculosis were screened, resulting in lectins and their combinations that can be used for the diagnosis of acute tuberculosis (ATB) and can differentiate between long-term tuberculosis-related intraepithelial neoplasia (LTBI) and ATB. Further model fitting of the screened lectins and their combinations confirmed that the lectin combinations of this invention have good diagnostic efficacy, accurately diagnosing ATB and accurately differentiating between ATB and LTBI.

[0008] Specifically, the present invention provides the following technical solutions.

[0009] In a first aspect, the present invention provides a combination of biomarkers for diagnosing active tuberculosis or distinguishing between latent tuberculosis infection and active tuberculosis, the combination of biomarkers comprising the following lectin combination: GNL, LAL and Black.bean.crude.

[0010] The combination of the aforementioned lectins GNL, LAL, and Black bean crude specifically binds to the glycosyl groups in the IgG glycan chain to form a lectin-IgG complex. This invention reveals that the levels of the corresponding complexes of lectins GNL, LAL, and Black bean crude differ significantly between healthy controls and patients with active tuberculosis, as well as between patients with active tuberculosis and patients with latent tuberculosis infection. These complexes can serve as biomarkers for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, exhibiting high specificity and sensitivity.

[0011] The combination of lectins GNL, LAL, and Black bean crude can be used alone as a marker for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, or it can be used in combination with certain lectins or Mycobacterium tuberculosis-related markers.

[0012] Furthermore, the lectin combination may also include one or two of the following lectins: PNA, HPA.

[0013] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, and PNA, or the lectin combination includes GNL, LAL, Black.bean.crude, and HPA.

[0014] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, PNA, and HPA.

[0015] This invention has found that the combination of four lectins—GNL, LAL, Black bean crude, and PNA—as well as the combination of four lectins—GNL, LAL, Black bean crude, and HPA, and the combination of five lectins—GNL, LAL, Black bean crude, PNA, and HPA—can further improve the accuracy of diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis compared to the combination of three lectins—GNL, LAL, and Black bean crude. Specifically, this is manifested in a significant improvement in AUC value, specificity, and sensitivity.

[0016] Preferably, the lectin specifically binds to the glycosyl groups in the IgG glycan chain to form a complex.

[0017] In some embodiments, the biomarker combination comprises a lectin-IgG complex formed by the specific binding of the lectin to glycosyl groups in the IgG glycan chain. The IgG is IgG from blood, serum, or plasma. Specifically, the biomarker combination comprises a complex formed by the binding of GNL, LAL, and Black bean cake to IgG, or a complex formed by the binding of GNL, LAL, Black bean cake, PNA, and HPA to IgG.

[0018] Furthermore, the combination of markers may also include Mycobacterium tuberculosis antigen. The Mycobacterium tuberculosis antigen preferably includes LAM.

[0019] LAM, or lipoarabinomannan, is an important antigen of Mycobacterium tuberculosis. This invention discovers that combining the above-mentioned lectin combination with LAM can further improve the accuracy of diagnosing active tuberculosis and differentiating between latent tuberculosis infection and active tuberculosis.

[0020] Secondly, the present invention provides the use of a lectin combination or a lectin combination combined with Mycobacterium tuberculosis antigen in the preparation of products for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination includes: GNL, LAL and Black.bean.crude.

[0021] Specifically, the product uses the lectin combination to bind to IgG glycans in blood, plasma, or serum, and detects the content of glycosyl groups in the IgG glycans that specifically bind to the lectin, thereby diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis.

[0022] Thirdly, the present invention provides the application of a lectin combination or a lectin combination combined with Mycobacterium tuberculosis antigen in the diagnosis of active tuberculosis or in differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination includes: GNL, LAL and Black.bean.crude.

[0023] Fourthly, the present invention provides the use of a detection product for lectin combination specifically binding to glycosyl groups or a detection product for lectin combination specifically binding to glycosyl groups combined with Mycobacterium tuberculosis antigen in the preparation of products for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination includes: GNL, LAL and Black.bean.crude.

[0024] Fifthly, the present invention provides a detection product for lectin combinations that specifically bind to glycosyl groups, or a detection product for lectin combinations that specifically bind to glycosyl groups combined with Mycobacterium tuberculosis antigen, for the diagnosis of active tuberculosis or for differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination includes: GNL, LAL, and Black.bean.crude.

[0025] In the applications described in the fourth and fifth aspects above, the lectin combination specifically binds to glycosyl groups that are glycosyl groups in blood, plasma, or serum IgG glycans that specifically bind to the lectin combination.

[0026] The detection products for the specific binding of glycosyl groups of lectins may include reagents and instruments for detecting the specific binding of glycosyl groups of lectins.

[0027] Preferably, the detection product for the specific binding of glycosyl groups by the lectin combination includes the lectin combination or includes a chip loaded with the lectin combination.

[0028] The detection products for the specific binding of glycosyl groups of lectins may also include fluorescent dyes, anti-IgG antibodies, blocking buffers, PBS, mouse serum, etc. The detection products for the specific binding of glycosyl groups of lectins may also include microarray scanners.

[0029] In the applications described in the second to fifth aspects above, the lectin combination may further include one or two of the following lectins: PNA, HPA.

[0030] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, and PNA, or the lectin combination includes GNL, LAL, Black.bean.crude, and HPA.

[0031] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, PNA, and HPA.

[0032] In the applications described in the second to fifth aspects above, the Mycobacterium tuberculosis antigen includes LAM.

[0033] In this invention, the lectin combination itself has high accuracy in diagnosing active tuberculosis and differentiating between latent tuberculosis infection and active tuberculosis, and can be applied in diagnostic or differential diagnostic practice. To further improve diagnostic accuracy, it can be combined with Mycobacterium tuberculosis antigen detection. This invention does not have specific limitations on the combined Mycobacterium tuberculosis antigens; the LAM mentioned above is merely an example, and other known or new Mycobacterium tuberculosis antigens that can be used to diagnose active tuberculosis or differentiate between latent tuberculosis infection and active tuberculosis can also be selected. These antigens include: secretory protein antigens (e.g., ESAT-6, CFP-10, Ag85 complex), cell wall-associated antigens (e.g., LAM, lipoprotein antigens such as LpqH, 38kDa / MPT64), and other important antigens (e.g., TB7.7 (Rv2654c), Rv2031c (HspX / α-crystallin)).

[0034] In this invention, the products include detection reagents, detection kits, detection chips, etc.

[0035] Sixthly, the present invention provides a method for constructing a diagnostic model for active tuberculosis or a model for distinguishing between latent tuberculosis infection and active tuberculosis, the method comprising: All samples are randomly divided into training and test sets; Obtain sample data to characterize the content of lectin combinations-specifically bound glycosyl groups in the IgG glycans contained in the training set samples; A training dataset is constructed using the tuberculosis diagnosis status corresponding to the sample as the label for the sample data; Construct a diagnostic model for active tuberculosis or a model to distinguish between latent tuberculosis infection and active tuberculosis based on the training dataset; The effectiveness of the model is verified using the test set. The lectin combination includes: GNL, LAL, and Black.bean.crude.

[0036] The samples mentioned above are preferably blood, plasma, or serum samples.

[0037] Preferably, the lectin combination further includes one or two of the following lectins: PNA, HPA.

[0038] Preferably, the model is constructed using an Artificial Neural Network (ANN) algorithm.

[0039] In the above construction method, the classification model is trained based on the training dataset, and the trained classification model is determined as a diagnostic model for active tuberculosis or a model for distinguishing between latent tuberculosis infection and active tuberculosis.

[0040] In the above construction method, the construction of the training dataset includes: grouping samples according to the tuberculosis diagnosis status of each sample as the label, using the detection signal of the agglutinin combination in each group of plasma samples, or the detection signal of the agglutinin combination in each group of plasma samples and the antibody data of Mycobacterium tuberculosis antigen detection as the sample data, and using the tuberculosis diagnosis status corresponding to the grouping of the plasma samples as the label of the sample data to construct the training dataset.

[0041] Specifically, the method for constructing the model includes the following steps: (1) Constructing the training dataset The training dataset may include sample data and the labels corresponding to the sample data; Here, sample data refers to the detection results of the lectin combination in each group of plasma samples (i.e., the signal intensity obtained by combining the lectin combination with plasma IgG specific glycans); for example, each group of plasma samples specifically includes plasma samples from the healthy control group HC, the latent tuberculosis infection group LTBI, and the active tuberculosis group ATB, and the training dataset contains sample data from the above three groups of plasma samples; the label corresponding to the sample data is the tuberculosis diagnosis status corresponding to the plasma sample. For example, based on the above grouping, the label corresponding to the sample data from HC and LTBI can be negative, that is, tuberculosis has not been diagnosed, and the label corresponding to the sample data from ATB can be positive, that is, tuberculosis has been diagnosed.

[0042] (2) Training the classification model based on the training dataset The classification model is a neural network model used to perform classification functions, such as an artificial neural network (ANN).

[0043] In the process of training a classification model based on a training dataset, sample data from the training dataset can be input into the classification model to obtain the classification results output by the model. The classification results are then compared with the labels in the training dataset to calculate the loss function. Based on the loss function, the parameters of the classification model are iterated, thereby training the classification model. During training, the classification model can learn the mapping relationship between the lectin combination detection results of the sample data and the tuberculosis diagnosis status as the label. Thus, the trained classification model has the ability to diagnose whether it is active tuberculosis or to distinguish between latent tuberculosis infection and active tuberculosis based on the input lectin combination detection results. In other words, the trained classification model can be used as a model for diagnosing active tuberculosis or distinguishing between latent tuberculosis infection and active tuberculosis.

[0044] Understandably, the model obtained thereby takes the detection results of the agglutinin combination in the plasma sample of the person being tested, or the detection results of the agglutinin combination and the antibody detection results of the Mycobacterium tuberculosis antigen, as input, and can output the diagnosis result of active tuberculosis for the person being tested, i.e., negative or positive, and can also output the diagnosis result of whether the person being tested has latent tuberculosis infection or active tuberculosis.

[0045] Specifically, the classification model is trained based on the training dataset, and the trained classification model is determined as the tuberculosis diagnostic model, including: The training dataset is split into multiple data subsets; The classification model is trained based on the first part of the multiple data subsets, and the trained classification model is tested based on the second part of the multiple data subsets to obtain the test results of the classification model. Based on the test results of the classification model, a tuberculosis diagnostic model is determined from the classification model.

[0046] Specifically, during the training of a tuberculosis diagnostic model, the training dataset can be split to obtain multiple data subsets. This splitting can be random. For example, the training dataset can be split into 10 data subsets.

[0047] After obtaining multiple data subsets, cross-validation can be performed on multiple classification models based on these subsets. Here, the multiple classification models can be classification models with the same structure and parameters, or they can be classification models with different structures and / or different parameters. This embodiment of the invention does not impose specific limitations on this.

[0048] The cross-validation process involves dividing the multiple data subsets obtained from the above split into two parts: a first part and a second part. The first part is used for model training, and the second part is used for model testing. For example, in the case of 10 data subsets, 9 subsets can be used as the first part for model training, and the remaining subset as the second part for model testing. Thus, for each classification model, the first part can be used for training, and the second part can be used for testing to obtain the test result. This test result can be AUC (Area Under the Curve), which specifically refers to the area under the ROC (Receiver Operating Characteristic) curve and is commonly used to evaluate the performance of binary classification models.

[0049] Since the splitting of training data into subsets is random and the resulting subsets are different each time, and the way the first and second parts are split can be different each time, the classification model can be trained and tested again after the first and second parts are split each time. This will generate more trained classification models and obtain the corresponding test results.

[0050] Based on this, the model with the best test results from all trained classification models can be selected as the final diagnostic model or discrimination model.

[0051] (3) Validate the diagnostic model or differentiation model. After model training is complete, the trained diagnostic model or differentiation model can be validated to verify the effectiveness of the tuberculosis diagnostic model. The validation dataset used for model validation can be constructed in the same way as the training dataset.

[0052] In a seventh aspect, the present invention provides a diagnostic model for active tuberculosis or a model for distinguishing between latent tuberculosis infection and active tuberculosis, constructed using the construction method described above.

[0053] Eighthly, the present invention provides products for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, said products comprising a lectin combination, or a detection product comprising a lectin combination specifically binding to glycosyl groups; wherein the lectin combination comprises: GNL, LAL, and Black.bean.crude.

[0054] Preferably, the lectin combination further includes one or two of the following lectins: PNA, HPA.

[0055] Preferably, the product includes a chip loaded with the lectin combination.

[0056] The chip loaded with the aforementioned lectin combination can be a lectin chip, specifically a chip obtained by fixing each lectin in the lectin combination onto a support.

[0057] Preferably, the product further includes an anti-IgG antibody. The anti-IgG antibody is labeled with a fluorescent dye.

[0058] The product may include other reagents and instruments required for detecting the specific binding of the lectin combination to glycosyl groups, wherein the reagents may be selected from one or more of blocking buffer, PBS, and mouse serum. The instruments include microarray scanners, etc.

[0059] Furthermore, the product may also include Mycobacterium tuberculosis antigen. The Mycobacterium tuberculosis antigen preferably includes LAM.

[0060] Preferably, the Mycobacterium tuberculosis antigen is used to detect antibodies against the Mycobacterium tuberculosis antigen in the sample. The product may also include reagents required for detecting antibodies against the Mycobacterium tuberculosis antigen in the sample using the Mycobacterium tuberculosis antigen, such as ELISA reagents.

[0061] Ninthly, the present invention provides an apparatus for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, the apparatus comprising: The detection module includes a module for detecting the content of lectin combinations that specifically bind glycosyl groups in the IgG glycan chains contained in the test sample; wherein the lectin combination includes: GNL, LAL, and Black.bean.crude; The input module is used to obtain the detection results from the detection module; The analysis module is used to diagnose active tuberculosis or differentiate between latent tuberculosis infection and active tuberculosis based on the test results obtained from the input module. The output module is used to output the analysis results.

[0062] Preferably, the lectin combination further includes one or two of the following lectins: PNA, HPA.

[0063] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, and PNA, or the lectin combination includes GNL, LAL, Black.bean.crude, and HPA.

[0064] In some embodiments, the lectin combination includes GNL, LAL, Black.bean.crude, PNA, and HPA.

[0065] In the aforementioned detection module, the content of the lectin combination specifically binding glycosyl groups in the IgG glycan chains contained in the test sample is the content of complexes formed by the specific binding of each lectin in the lectin combination with the corresponding glycosyl groups in the IgG glycan chains. The complex is a lectin-IgG complex.

[0066] In the above detection module, the detection of the content of lectin combination-specifically bound glycosyl groups in the IgG glycan chains contained in the test sample includes: The lectin assembly is contacted with the sample to be tested so that the lectin assembly forms a complex with IgG containing its specifically binding glycosyl group. After removing unbound IgG, the complex is contacted with a labeled anti-IgG antibody; Signal detection was performed after removing unbound anti-IgG antibodies.

[0067] In the above analysis module, diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis based on the test results obtained from the input module includes: comparing the content of the lectin combination specifically binding glycosyl group in the IgG glycan chain contained in the test sample with the content of the glycosyl group in the IgG glycan chain contained in the control sample, and diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis based on the comparison results.

[0068] Preferably, the detection module further includes a module for detecting the content of antibodies against Mycobacterium tuberculosis antigen. The Mycobacterium tuberculosis antigen preferably includes LAM.

[0069] Preferably, the sample to be tested is blood, serum, or plasma.

[0070] In this invention, compared with healthy controls, ATB patients showed decreased levels of the GNL-related complex, decreased levels of the LAL-related complex, and increased levels of the Black-bean-crude-related complex. Compared with healthy controls, ATB patients also showed decreased levels of the PNA-related complex and the HPA-related complex.

[0071] In this invention, compared with LTBI patients, ATB patients showed decreased levels of the GNL-corresponding complex, increased levels of the LAL-corresponding complex, and increased levels of the Black-bean-crude-corresponding complex. Compared with LTBI patients, ATB patients showed decreased levels of the PNA-corresponding complex and the HPA-corresponding complex.

[0072] In a tenth aspect, the present invention provides a method for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, the method comprising: The content of lectin combination-specifically bound glycosyl groups in the IgG glycan chains contained in the test sample was detected. Based on the test results, diagnose active tuberculosis or differentiate between latent tuberculosis infection and active tuberculosis; The lectin combination includes: GNL, LAL, and Black.bean.crude.

[0073] Preferably, the lectin combination further includes one or two of the following lectins: PNA, HPA.

[0074] Preferably, the content of the lectin combination specifically binding glycosyl groups in the IgG glycan chains contained in the test sample is the content of the complex formed by the specific binding of each lectin in the lectin combination with the corresponding glycosyl group in the IgG glycan chain. The complex is a lectin-IgG complex.

[0075] Preferably, the detection of the content of lectin combination-specifically bound glycosyl groups in the IgG glycan chains contained in the test sample includes: The lectin assembly is contacted with the sample to be tested so that the lectin assembly forms a complex with IgG containing its specifically binding glycosyl group. After removing unbound IgG, the complex is contacted with a labeled anti-IgG antibody; Signal detection was performed after removing unbound anti-IgG antibodies.

[0076] The beneficial effects of this invention include at least the following: This invention obtains a combination of lectins that can be used for ATB diagnosis and can differentiate between LTBI and ATB by detecting IgG glycosylation during the pathogenesis of tuberculosis. These lectin combinations can serve as biomarkers for ATB diagnosis and differentiation between LTBI and ATB, exhibiting high specificity and sensitivity. This provides new ideas and methods for differentiating different stages of tuberculosis and for early, accurate diagnosis of ATB. The combination of these lectin combinations with specific antibodies detected by tuberculosis antigen can further improve diagnostic efficacy. The biomarkers and their detection products provided by this invention have good application potential in ATB diagnosis and differentiation between LTBI and ATB. Attached Figure Description

[0077] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0078] Figure 1 This describes the lectin chip dot matrix distribution and fluorescence detection in Embodiment 2 of the present invention.

[0079] Figure 2 This is a list of the contribution rankings of lectin markers in Embodiment 3 of the present invention.

[0080] Figure 3 This refers to the differences between the HC, LTBI, and ATB groups of GNL, PNA, HPA, LAL, and Black.bean.crude in Embodiment 3 of the present invention.

[0081] Figure 4 This is a schematic diagram of the ANN model workflow (A) and the diagnostic model (B) in Embodiment 4 of the present invention.

[0082] Figure 5 The ROC curve analysis of the ANN diagnostic model in Embodiment 4 of the present invention is shown in (A) and (B) respectively. (A) is the ROC curve analysis based on the training set and (B) is the ROC curve analysis based on the test set.

[0083] Figure 6 The diagnostic efficacy analysis of different lectin combinations in Embodiment 5 of the present invention includes (A) ROC curve analysis based on the training set and (B) ROC curve analysis based on the test set.

[0084] Figure 7 The ROC curve analysis of the LAM combined with lectin diagnostic model in Embodiment 6 of the present invention is shown in (A) and (B) respectively.

[0085] Figure 8 This is a schematic diagram of the device for diagnosing active tuberculosis or distinguishing between latent tuberculosis infection and active tuberculosis in Embodiment 7 of the present invention. Detailed Implementation

[0086] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0087] Example 1: Clinical Sample Collection This invention collected samples from three groups: HC, LTBI, and ATB. HC consisted of healthy individuals with negative IGRA test results and no clinical symptoms; LTBI consisted of infected individuals with positive IGRA test results but no clinical symptoms; and ATB patients all met the diagnostic criteria of the "Guidelines for the Diagnosis and Treatment of Pulmonary Tuberculosis." Patients with cancer, diabetes, allergic diseases, any form of infectious disease, or those in an immunocompromised state were excluded. A total of 389 clinical plasma samples were collected, including 136 HC samples, 102 LTBI samples, and 151 ATB samples. Detailed basic information and diagnostic characteristics of the subjects are shown in Table 1.

[0088] Table 1. Basic information and diagnostic characteristics of the subjects

[0089] Example 2: Analysis of plasma IgG glycosylation using lectin microarray This invention uses the Lectin56 microarray chip containing 56 lectins. TM (Purchased from TBHealthcare) The plasma IgG glycosylation status of 389 clinical plasma samples described in Example 1 was tested, including 136 healthy controls (HC), 102 LTBI, and 151 ATB.

[0090] First, the lectin microarray was blocked with blocking buffer (PBS buffer containing 3% BSA). After washing and drying, each plasma sample was diluted 1:200 and added to the microarray, and incubated overnight at 4°C. After washing with PBS, it was blocked with mouse serum diluted 1:50 for 1 hour. Then, it was incubated with Cy3-labeled anti-human IgG antibody at room temperature in the dark for 1 hour. After washing and drying, it was then analyzed using LuxScan. TM The 10K microarray chip scanner performs scans at a wavelength of 532 nm.

[0091] The median values ​​of the foreground and background at 532 nm were read using GenePix Pro v6.0 software. A blank control was used to eliminate non-specific signal errors. Background correction was performed by calculating the difference between the foreground and background values ​​(FB). Using the preprocessed net signal value as the calculation object, differentially expressed lectins were screened based on statistical methods. A Fold change ≥ 1.2 and P < 0.05 were used to determine differential expression between the two groups.

[0092] Converting the chip image into a digital format for analysis allows for the acquisition of signal values ​​indicating the binding strength of each lectin to IgG-specific binding glycans. Figure 1 ).

[0093] Example 3: Screening for lectins that can serve as biomarkers To screen for biomarkers suitable for tuberculosis diagnosis, lectins for diagnostic model fitting were first selected based on their contribution value from the differential lectins obtained in Example 2 (Foldchange ≥ 1.2, P < 0.05). Using 10-fold cross-validation with logistic regression, the dataset was divided into 10 parts, with 9 parts used for training and 1 part for testing, repeated 10 times to obtain 100 combinations. The frequency of each lectin in the 100 combinations was calculated, and contributions were ranked. A frequency exceeding 80% was defined as a significant contribution. Five significant lectins were selected: GNL, PNA, HPA, LAL, and Black.bean.crude. Figure 2 From Table 2 and Figure 3 It can be seen that all five lectins show significant differences between ATB and HC, and except for LAL, GNL, PNA, HPA, and Black.bean.crude also show significant differences between ATB and LTBI.

[0094] Table 2 Key lectin information and analysis results

[0095] Example 4: Construction of an ATB diagnostic model based on a combination of 5 lectins The five lectins (GNL, PNA, HPA, LAL, and Black.bean.crude) obtained in Example 2 above were used as a combination of biomarkers to construct a tuberculosis diagnostic model using an artificial neural network (ANN) algorithm. First, all samples were randomly divided into a training set and a test set, each accounting for 50% of the total, for model training and model validation, respectively. During model building, the area under the curve (AUC) of each model was calculated, and finally, the model with the best diagnostic efficacy was selected from 500 models. Then, the model efficacy was independently validated using blind testing data from the test set. A schematic diagram and workflow of the diagnostic model constructed using the combination of GNL, PNA, HPA, LAL, and Black.bean.crude are shown below. Figure 4 As shown.

[0096] In the model, HC and LTBI samples are defined as Controls and judged as negative results, while ATB is judged as a positive result. By inputting the feature signal values ​​of GNL, PNA, HPA, LAL, and Black.bean.crude of the test sample, ATB can be distinguished from Controls. On the training set, the model achieved AUCs of 0.913, 0.922, and 0.916 for ATB versus HC, ATB versus LTBI, and ATB versus HC+LTBI, respectively. The specificity and sensitivity for distinguishing ATB from HC were 84.9% and 82.50%, respectively; for ATB from LTBI, 85.40% and 81.30%, respectively; and for ATB from HC+LTBI, 84.20% and 82.50%, respectively. Figure 5 (See Table 3). Furthermore, independent sample validation using the test set showed that the AUC for all distinctions exceeded 0.9, demonstrating the model's excellent diagnostic efficacy. This model can not only identify ATB from HC, LTBI, and ATB samples, but also distinguish between ATB and HC, as well as between ATB and LTBI.

[0097] Table 3. Efficacy of the lectin combination diagnostic model

[0098] Example 5: Construction of ATB diagnostic models using combinations of 3 and 4 lectins Following the method described in Example 4, a model was constructed using a combination of three lectin markers—GNL, LAL, and Black bean crude—to evaluate its diagnostic efficacy in differentiating between ATB and control samples (HC and LTBI). The model's effectiveness was assessed using ROC curves, with AUC values ​​of 0.784 and 0.748 for the training and test sets, respectively. Figure 6 ).

[0099] Based on 3-lectins, models were constructed by combining one of PNA or HPA, respectively, to evaluate their diagnostic efficacy in differentiating ATB and control samples (HC and LTBI). ROC curve analysis revealed that adding either PNA or HPA to the 3-lectins significantly improved diagnostic efficacy, indicating that combinations of the four lectins—GNL, LAL, Black bean crude, and PNA—and combinations of the four lectins—GNL, LAL, Black bean crude, and HPA—all possess high diagnostic value. Figure 6 (Table 4).

[0100] Table 4. Diagnostic efficacy of different combinations of lectins for ATB.

[0101] Example 6: Construction of a diagnostic model for LAM using lectin combination Lipoarabinomannan (LAM) is an important component of the cell wall of Mycobacterium tuberculosis and possesses strong immunogenicity. Detection of its specific antibody, LAM-IgG, can provide auxiliary diagnostic information for tuberculosis. To further improve the sensitivity and specificity of the diagnostic model in Example 4, LAM-IgG levels in all clinical samples were detected by ELISA. The detection data were then combined with the lectin ANN model using the logistic regression formula Logit(ATB) = 12.4617*ANN + 1.6036*LAM. After combining the LAM antigen with the ANN model, the AUCs in the training and test sets reached 0.959 and 0.940, respectively. The specificity and sensitivity in the training set were 92.50% and 85.10%, respectively, while the sensitivity in the test set was 91.50% and 87.10%, respectively. Figure 7 (See Table 5). These results indicate that LAM further improves the diagnostic efficacy of the ANN model, and also significantly enhances its specificity and sensitivity.

[0102] Table 5. Diagnostic model efficacy of lectin combination with LAM in differentiating ATB from HC and LTBI

[0103] In summary, the following combinations of lectins—GNL, LAL, and Black bean crude; GNL, LAL, Black bean crude, and PNA; GNL, LAL, Black bean crude, and HPA; and GNL, PNA, HPA, LAL, and Black bean crude—can all serve as novel biomarkers for tuberculosis diagnosis. The ANN model constructed using these lectin combinations can accurately diagnose ATB and differentiate between ATB and LTBI. Furthermore, the Mycobacterium tuberculosis cell wall component LAM can be used in combination with lectin combinations to further enhance its diagnostic value.

[0104] Example 7: A device for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis. This embodiment provides a device for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis. Figure 8 The device includes the following modules: The detection module 100 includes a module for detecting the content of glycosyl groups specifically bound by a lectin combination in the IgG glycan chains contained in the plasma sample to be tested; wherein the lectin combination is: GNL, LAL and Black.bean.crude, or: GNL, LAL, Black.bean.crude and PNA, or GNL, LAL, Black.bean.crude and HPA, or GNL, LAL, Black.bean.crude, PNA and HPA; Input module 110 is used to obtain the detection results of the detection module; Analysis module 120 is used to diagnose active tuberculosis or differentiate between latent tuberculosis infection and active tuberculosis based on the test results obtained from the input module; Output module 130 is used to output analysis results; The detection module executes the following program: The lectin assembly is contacted with the sample to be tested so that the lectin assembly forms a complex with IgG containing its specifically binding glycosyl group. After removing unbound IgG, the complex is contacted with a labeled anti-IgG antibody; After removing unbound anti-IgG antibodies, signal detection was performed to obtain the content of lectin-specifically bound glycosyl groups in the IgG glycan chains contained in the test sample.

[0105] The input module executes the following program: The content of the glycosyl group specifically binding to the lectin combination in the IgG glycans of the test sample is compared with the content of the glycosyl group in the IgG glycans of the control sample, and the active tuberculosis is diagnosed or the latent tuberculosis infection and active tuberculosis are distinguished based on the comparison results. The method for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis based on the comparison results is as follows: the results of the agglutinin combination detection of the sample to be tested can be imported into the diagnostic model to output the diagnostic results.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A combination of biomarkers for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, characterized in that, The biomarker combination includes the following lectin combination: GNL, LAL, and Black.bean.crude.

2. The marker combination according to claim 1, characterized in that, The lectin combination also includes one or two of the following lectins: PNA, HPA.

3. The combination of markers according to claim 1 or 2, characterized in that, The lectin specifically binds to the glycosyl groups in the IgG glycan chain to form a complex.

4. The combination of markers according to any one of claims 1 to 3, characterized in that, The biomarker combination also includes Mycobacterium tuberculosis antigen; the Mycobacterium tuberculosis antigen preferably includes LAM.

5. The use of a lectin combination or a lectin combination combined with Mycobacterium tuberculosis antigen in the preparation of products for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination comprises: GNL, LAL, and Black.bean.crude.

6. The application according to claim 5, characterized in that, The product detects the content of glycosyl groups in IgG sugar chains that specifically bind to the lectin by binding the lectin combination to blood, plasma or serum, thereby diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis.

7. The use of a lectin combination-specific glycosyl-binding detection product or a lectin combination-specific glycosyl-binding detection product combined with Mycobacterium tuberculosis antigen in the preparation of products for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, wherein the lectin combination comprises: GNL, LAL, and Black.bean.crude.

8. The application according to claim 7, characterized in that, The lectin combination specifically binds to glycosyl groups, which are glycosyl groups in blood, plasma, or serum IgG glycans that specifically bind to the lectin combination.

9. The application according to any one of claims 5 to 8, characterized in that, The lectin combination also includes one or two of the following lectins: PNA, HPA.

10. The application according to any one of claims 5 to 9, characterized in that, The Mycobacterium tuberculosis antigen includes LAM.

11. A method for constructing a diagnostic model for active tuberculosis or a model for distinguishing between latent tuberculosis infection and active tuberculosis, characterized in that, The method includes: All samples are randomly divided into training and test sets; Obtain sample data to characterize the content of lectin combinations-specifically bound glycosyl groups in the IgG glycans contained in the training set samples; A training dataset is constructed using the tuberculosis diagnosis status corresponding to the sample as the label for the sample data; Construct a diagnostic model for active tuberculosis or a model to distinguish between latent tuberculosis infection and active tuberculosis based on the training dataset; The effectiveness of the model is verified using the test set. The lectin combination includes: GNL, LAL, and Black.bean.crude; The sample is preferably a blood, plasma, or serum sample.

12. The construction method according to claim 11, characterized in that, The lectin combination also includes one or two of the following lectins: PNA, HPA.

13. A product for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, characterized in that, The product includes a lectin combination, or a detection product that includes a lectin combination that specifically binds to glycosyl groups; wherein the lectin combination includes: GNL, LAL, and Black.bean.crude.

14. The product according to claim 13, characterized in that, The lectin combination also includes one or two of the following lectins: PNA, HPA.

15. The product according to claim 13 or 14, characterized in that, The product includes a chip loaded with the lectin combination.

16. The product according to claim 15, characterized in that, The product also includes anti-IgG antibodies.

17. The product according to any one of claims 13 to 16, characterized in that, The product also includes Mycobacterium tuberculosis antigen; the Mycobacterium tuberculosis antigen preferably includes LAM.

18. A device for diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis, characterized in that, The device includes: The detection module includes a module for detecting the content of lectin combinations that specifically bind glycosyl groups in the IgG glycan chains contained in the test sample; wherein the lectin combination includes: GNL, LAL, and Black.bean.crude; The input module is used to obtain the detection results from the detection module; The analysis module is used to diagnose active tuberculosis or differentiate between latent tuberculosis infection and active tuberculosis based on the test results obtained from the input module. The output module is used to output the analysis results.

19. The apparatus according to claim 18, characterized in that, The lectin combination also includes one or two of the following lectins: PNA, HPA.

20. The apparatus according to claim 18 or 19, characterized in that, The content of the lectin combination specifically binding glycosyl groups in the IgG glycan chains contained in the test sample is the content of the complex formed by the specific binding of each lectin in the lectin combination with the corresponding glycosyl group in the IgG glycan chain.

21. The apparatus according to any one of claims 18 to 20, characterized in that, The detection of the specific binding glycosyl groups of the lectin combination in the IgG glycan chains contained in the test sample includes: The lectin assembly is contacted with the sample to be tested so that the lectin assembly forms a complex with IgG containing its specifically binding glycosyl group. After removing unbound IgG, the complex is contacted with a labeled anti-IgG antibody; Signal detection was performed after removing unbound anti-IgG antibodies.

22. The apparatus according to claim 18 or 19, characterized in that, The method of diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis based on the test results obtained from the input module includes: comparing the content of the lectin combination specifically binding glycosyl group in the IgG glycan chain contained in the test sample with the content of the glycosyl group in the IgG glycan chain contained in the control sample, and diagnosing active tuberculosis or differentiating between latent tuberculosis infection and active tuberculosis based on the comparison results.

23. The apparatus according to any one of claims 18 to 22, characterized in that, The detection module also includes a module for detecting the content of anti-tuberculosis antigen antibodies; the tuberculosis antigen preferably includes LAM.

24. The apparatus according to any one of claims 18 to 23, characterized in that, The sample to be tested is blood, serum, or plasma.

Images