Biomarker composition for diagnosing tuberculosis

Abstract

The invention provides a group of biomarker compositions for diagnosing tuberculosis, the marker compositions obtained by screening can be used for diagnosing tuberculosis, the model is used for tuberculosis diagnosis, the area under ROC curve (AUC) is greater than 0.9, and when the cut-off value (diagnosis threshold) is 0.696, the sensitivity and specificity of diagnosis are both greater than 80%.

Description

Technical Field

[0001] The invention belongs to the field of biomedicine, and in particular relates to a group of biomarker compositions for diagnosing tuberculosis. Background Art

[0002] Tuberculosis is an infectious disease caused by Mycobacterium tuberculosis (MTB). MTB is a common pathogen that usually parasitizes in host macrophages and can be effectively transmitted through the respiratory system of the population, causing more serious lung diseases in infected hosts (PREDA M, BC, ZOB DL, et al. The bidirectional relationship between pulmonary tuberculosis and lung cancer [J]. Int J Environ Res Public Health, 2023, 20 (2): 1282.). In addition, MTB is a highly adaptable Gram-positive bacterium that is mainly transmitted between people (HUNTER R L. The pathogenesis of tuberculosis: the ear.ly infiltrate of post-primary (adult pulmonary) tuberculosis: a distinct disease entity [J]. Front Immunol, 2018, 9: 2108.). Tuberculosis has always been a key infectious disease controlled globally and in my country. Early diagnosis and treatment are important means to control the spread of tuberculosis. However, due to the lack of rapid and accurate diagnostic technology, the diagnosis of tuberculosis has always failed to meet clinical needs. Therefore, finding a suitable method to diagnose tuberculosis (TB) as early as possible is of great significance for formulating clinical treatment plans and improving prognosis.

[0003] At present, there are several main methods for diagnosing tuberculosis. The first is traditional bacteriological testing, including sputum smear microscopy and Mycobacterium tuberculosis culture, but these methods are not only time-consuming but also have low sensitivity. The second is molecular biological testing, including Xpert MTB / RIF and linear probes, which mainly target the DNA of Mycobacterium tuberculosis for detection. Although these methods are fast and accurate, molecular biological methods have poor sensitivity in low-bacteria samples, have high requirements for the quality of sputum specimens, and need to be equipped with corresponding equipment for detection, which has limited application in primary medical institutions and comprehensive medical institutions. The third method is based on the detection of human immune response to Mycobacterium tuberculosis, that is, tuberculosis-infected T cell detection technology, but this method cannot distinguish between latently infected people and active tuberculosis patients. Some other conventional biochemical tests or imaging tests have poor specificity in distinguishing tuberculosis from other non-tuberculosis diseases. WHO recommends the identification of tuberculosis biomarkers based on non-sputum specimens and the development of corresponding detection reagents. Therefore, there is an urgent need to find new biomarkers for the diagnosis of tuberculosis and develop an early and rapid diagnosis method.

[0004] Non-coding RNA (non-coding RNA) is produced during the transcription of the human genome into primary transcription products. They are stronger than messenger RNA (mRNA) in terms of tissue specificity and spatiotemporal specificity, and participate in the body's immune response and pathological damage process in various ways. Among them, long non-coding RNA (lncRNA) accounts for 70%-80% of non-coding RNA. lncRNA is a class of RNA molecules with a length of more than 200 nucleotides. Although lncRNA does not encode proteins, they can regulate epigenetic modifications, transcription, translation and post-translational modifications through interactions with DNA, RNA and proteins, thereby playing an important biological function. Abnormal expression of lncRNA has been found in many diseases, including cancer, cardiovascular disease, autoimmune diseases, neuroinflammation, and various viral and bacterial infections.

[0005] More and more studies have shown that lncRNA plays an important regulatory role in the process of tuberculosis infection. lncRNA can not only induce autophagy and apoptosis of macrophages, but also participate in regulating the immune defense response of immune cells to Mycobacterium tuberculosis. Given the key role of lncRNA in the immune response to tuberculosis, it shows great potential as a new biomarker for the diagnosis of tuberculosis. lncRNA can be detected in almost all body fluids, including blood, urine, sputum, cerebrospinal fluid and pleural effusion. Summary of the invention

[0006] The present invention screens out a group of biomarker compositions for diagnosing tuberculosis, which have strong diagnostic efficacy for diagnosing tuberculosis. Based on this, the present invention is completed.

[0007] In the first aspect, the present invention provides a group of biomarker compositions for diagnosing tuberculosis, the biomarker composition comprising 5 lncRNAs, the 5 lncRNAs are NR_104128, Inc-MXRA7, ENST00000565797, Inc-NT5C3A and ENST00000583184, wherein the sequence of NR_104128 is shown in SEQ ID NO.1, the sequence of Inc-MXRA7 is shown in SEQ ID NO.2, the sequence of ENST00000565797 is shown in SEQ ID NO.3, the sequence of Inc-NT5C3A is shown in SEQ ID NO.4, and the sequence of ENST00000583184 is shown in SEQ ID NO.5.

[0008] Furthermore, the biomarkers are derived from whole blood, plasma or serum of the patient.

[0009] Furthermore, the marker composition calculates the predicted probability value of the sample according to the relative expression levels of the five lncRNAs, and when the probability value is greater than or equal to 0.696, the patient is diagnosed with tuberculosis.

[0010] Furthermore, the predicted probability value of the sample is calculated by the AdaBoost model.

[0011] Furthermore, the tuberculosis is a disease caused by infection with Mycobacterium tuberculosis.

[0012] Furthermore, the tuberculosis includes primary tuberculosis, secondary tuberculosis, blood type disseminated tuberculosis, tracheobronchial tuberculosis, tuberculous pleurisy, septic tuberculosis, etc.

[0013] Furthermore, the Mycobacterium tuberculosis infection includes: primary infection, secondary infection, and extrapulmonary infection.

[0014] In a second aspect, the present invention provides a group of biomarker compositions for use in preparing a reagent for diagnosing tuberculosis, wherein the biomarker composition comprises the five lncRNAs described in the first aspect of the present invention, and the reagent is a reagent that can detect the relative expression levels of the five lncRNAs in a patient's biological sample.

[0015] Furthermore, the biomarkers are derived from whole blood, plasma or serum of the patient.

[0016] Furthermore, the predicted probability value of the sample was calculated based on the relative expression levels of the five lncRNAs. When the probability value was greater than or equal to 0.696, the patient was diagnosed with tuberculosis.

[0017] Furthermore, the predicted probability value of the sample is calculated by the AdaBoost model.

[0018] Furthermore, the tuberculosis is a disease caused by infection with Mycobacterium tuberculosis.

[0019] Furthermore, the tuberculosis includes primary tuberculosis, secondary tuberculosis, blood type disseminated tuberculosis, tracheobronchial tuberculosis, tuberculous pleurisy, septic tuberculosis, etc.

[0020] Furthermore, the Mycobacterium tuberculosis infection includes: primary infection, secondary infection, and extrapulmonary infection.

[0021] In a third aspect, the present invention provides a kit for diagnosing tuberculosis, wherein the kit comprises the reagent as described in the second invention of the present invention.

[0022] Furthermore, the kit may be one or more of a nucleic acid detection kit, an immunofluorescence kit, a gene chip detection kit, a molecular hybridization kit and / or an in situ hybridization staining kit.

[0023] Furthermore, the diagnostic method of the kit includes one or more of PCR method / qPCR method, linear probe method, high-resolution melting curve method and / or gene chip method.

[0024] Beneficial Effects

[0025] The marker composition screened by the present invention can be used to diagnose tuberculosis. The area under the ROC curve (AUC) of the model for diagnosing tuberculosis is greater than 0.9. When the cut-off value (diagnostic threshold) is 0.696, the sensitivity and specificity of the diagnosis are both greater than 80%. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is the ROC of the AdaBoost model for diagnosing tuberculosis in the training set.

[0027] Figure 2 This is the ROC of the AdaBoost model for diagnosing tuberculosis in the validation set.

[0028] Figure 3 ROC of AdaBoost model for diagnosing tuberculosis in the test set. DETAILED DESCRIPTION

[0029] The specific embodiments of the present invention are further described below. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention. In addition, the technical features involved in the embodiments described below can be combined with each other as long as they do not conflict with each other.

[0030] The experimental methods in the following examples are conventional methods unless otherwise specified, and the experimental materials used in the following examples are commercially available unless otherwise specified.

[0031] Example 1 Screening of 5 lncRNAs

[0032] 1. Inclusion criteria

[0033] 1) Patients diagnosed with tuberculosis: referring to the industry standard "WS-288-2017 Diagnosis of Pulmonary Tuberculosis", patients with clinical symptoms of tuberculosis, imaging showing active tuberculosis lesions, and one of the following conditions: positive sputum / body fluid smear examination, positive Mycobacterium tuberculosis culture, positive Mycobacterium tuberculosis molecular biological detection, or positive histopathology are included.

[0034] 2) Patients with clinically diagnosed tuberculosis: referring to the industry standard "WS-288-2017 Diagnosis of Pulmonary Tuberculosis", patients with clinical symptoms of tuberculosis, imaging manifestations of tuberculosis, positive IGRA test, and effective anti-tuberculosis treatment were included.

[0035] 3) Non-tuberculosis disease controls: patients with suspected tuberculosis based on clinical symptoms and imaging findings, but ultimately diagnosed with other diseases, including tumors that need to be differentially diagnosed from tuberculosis (confirmed by pathology or cytology), pneumonia (other pathogens found and effectively treated), etc.

[0036] 4) Latent tuberculosis infection: no clinical symptoms of tuberculosis, no abnormalities in imaging examinations, positive tuberculosis infection T cell test, and no previous history of tuberculosis.

[0037] 5) Healthy controls: no clinical symptoms of tuberculosis, no abnormal imaging examinations, negative IGRA test results, and no previous history of tuberculosis or contact with tuberculosis.

[0038] 2. Chip detection

[0039] 1) Isolation of PBMCs from peripheral blood

[0040] Collect whole blood, dilute the blood sample with culture medium, mix well and add it to sample density separation solution to extract PBMCs.

[0041] 2) RNA extraction

[0042] Total RNA was extracted and the extracted total RNA was kept for later use after passing the electrophoresis quality inspection.

[0043] 3) cRNA labeling

[0044] Total RNA is amplified and labeled, and the labeled cRNA is purified.

[0045] 4) Chip hybridization

[0046] Rolling hybridization was performed in a rolling hybridization oven, with the cRNA loading amount of 1.65 μg, and the slides were washed.

[0047] 5) Chip scanning

[0048] The hybridized chip was scanned, the data was read, and finally normalized using the limma package in the R software, using the Quantile algorithm.

[0049] 6) Chip detection data

[0050] Ten tuberculosis patients, 10 latently infected persons, and 10 healthy subjects were included to complete the chip detection. Based on the original chip data, WGCNA analysis was applied, combined with intergroup difference analysis, to screen 45 lncRNAs that were different between tuberculosis and the other two groups. In the subsequent analysis, 5 lncRNAs were identified through qPCR verification and LASSO regression analysis to form a diagnostic model. The expression of the 5 lncRNAs in the chip data is shown in Table 1.

[0051] Table 1. lncRNA chip results analysis

[0052]

[0053] Note: TB patients, LTBI latently infected people, HC healthy people. Fc, foldchange.

[0054] Example 2 Diagnostic performance of lncRNA combination

[0055] 1. Determination of relative expression

[0056] The relative expression was calculated as follows: ΔCt = Ct 标志物 -Ct 内参基因 ; Relative expression level = 2 -△Ct .

[0058] Reagents, instruments and raw data for qPCR detection:

[0059] ① Reagents

[0060] lncRNA chip: Agilent Human (4*180K) ceRNA chip

[0061] Reference gene for qPCR: GAPDH

[0062] The target primer sequences are shown in Table 2

[0063] Table 2. Target primer sequences

[0064]

[0065] The reagents for RNA extraction (QIAGEN), reverse transcription (TOYOBO), and qPCR detection (Thermo Fisher Scientific) were all commercial reagents.

[0066] ②Instrument: ABI 7500

[0067] ③Original test data

[0068] Table 3. Original expression levels (relative expression levels) detected by qPCR in the training set

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] Note: TB refers to patients with tuberculosis, LTBI refers to persons with latent infection, HC refers to healthy persons, and NTB refers to patients with non-tuberculous lung diseases.

[0080] Table 4. Original expression level (relative expression level) detected by qPCR in the validation set

[0081]

[0082]

[0083]

[0084]

[0085] Note: TB refers to patients with tuberculosis, LTBI refers to persons with latent infection, HC refers to healthy persons, and NTB refers to patients with non-tuberculous lung diseases.

[0086] Table 5. Original expression levels (relative expression levels) detected by qPCR in independent test sets

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095] Note: TB1 refers to patients with tuberculosis; TB2 refers to patients with clinically diagnosed tuberculosis; LTBI refers to persons with latent infection; HC refers to healthy persons; NTB refers to patients with non-tuberculous lung diseases.

[0096] 2. ROC analysis of the diagnostic performance of the AdaBoost model combined with five lncRNAs

[0097] The training set included 134 tuberculosis patients, 38 latently infected persons, 46 healthy persons, and 54 non-tuberculous lung diseases; the validation set included 58 tuberculosis patients, 17 latently infected persons, 20 healthy persons, and 24 non-tuberculous lung diseases. The receiver operating characteristic (ROC) curve was used to evaluate the diagnostic performance of the AdaBoost model in distinguishing tuberculosis patients from other control groups.

[0098] AdaBoost mainly trains multiple decision trees iteratively and adjusts the weights of each decision tree according to the error rate of each decision tree. Each classifier assigns a higher weight to the misjudged samples, so that the next round of training pays more attention to these samples, thereby gradually improving the accuracy of the overall model. In this study, the AdaBoost model constructed multiple decision trees based on the relative expression levels of 5 lncRNAs in different orders. By weighted summing up the probability values ​​of samples suffering from tuberculosis output by each decision tree, a comprehensive predicted probability value is finally output for each sample, that is, the probability of the sample suffering from tuberculosis, with a probability value between 0 and 1. The probability values ​​of these samples are subjected to receiver operating characteristic curve (ROC) analysis (such as Figure 1The results show that the probability value of 0.696 is used as the cut-off value (diagnostic threshold). In the figure, controls are LTBI+HC+NTB.

[0099] The results showed that the area under the ROC curve (AUC) of the model for tuberculosis diagnosis in both the training set and the validation set was greater than 0.9. When the cut-off value (diagnostic threshold) was 0.696, the sensitivity and specificity in the training set were 89% (83%-93%) and 91% (87%-95%), respectively; the sensitivity and specificity in the validation set were 81% (69%-91%) and 84% (71%-93%), respectively.

[0100] In a separate test set, including 99 tuberculosis patients, 25 clinically diagnosed tuberculosis patients, 10 latently infected people, 36 healthy people and 36 patients with non-tuberculous lung diseases, the model (still using the probability value of 0.696 as the diagnostic threshold) had an AUC of 0.92 (0.88-0.95) for diagnosing tuberculosis, with a sensitivity and specificity of 75% (67%-83%) and 94% (89%-98%).

[0101] Table 5. Performance of AdaBoost model in diagnosing tuberculosis

[0102]

Claims

1. A group of biomarker compositions for diagnosing tuberculosis, the biomarker composition comprising 5 lncRNAs, the 5 lncRNAs are NR_104128, Inc-MXRA7, ENST00000565797, Inc-NT5C3A and ENST00000583184, wherein the sequence of NR_104128 is shown in SEQ ID NO.1, the sequence of Inc-MXRA7 is shown in SEQ ID NO.2, the sequence of ENST00000565797 is shown in SEQ ID NO.3, the sequence of Inc-NT5C3A is shown in SEQ ID NO.4, and the sequence of ENST00000583184 is shown in SEQ ID NO.

5.

2. The composition according to claim 1, wherein the marker composition calculates the predicted probability value of the sample according to the relative expression levels of the five lncRNAs, and when the probability value is greater than or equal to 0.696, the patient is diagnosed with tuberculosis.

3. The composition as described in claim 2, wherein the predicted probability value of the sample is calculated by an AdaBoost model.

4. The composition according to claim 1, wherein the tuberculosis is a disease caused by infection with Mycobacterium tuberculosis.

5. Use of a group of biomarker compositions in the preparation of a reagent for diagnosing tuberculosis, wherein the biomarker composition comprises the five lncRNAs described in the first aspect of the present invention, and the reagent is a reagent that can detect the relative expression levels of the five lncRNAs in a patient's biological sample.

6. The use as claimed in claim 5, wherein the predicted probability value of the sample is calculated according to the relative expression levels of the five lncRNAs, and when the probability value is greater than or equal to 0.696, the patient is diagnosed with tuberculosis.

7. The application as claimed in claim 6, wherein the predicted probability value of the sample is calculated by an AdaBoost model.

8. A kit for diagnosing tuberculosis, comprising the reagent according to claim 5.

9. The kit according to claim 8, wherein the kit can be one or more of a nucleic acid detection kit, an immunofluorescence kit, a gene chip detection kit, a molecular hybridization kit and / or an in situ hybridization staining kit.

10. The kit according to claim 9, wherein the diagnostic method of the kit comprises one or more of PCR method / qPCR method, linear probe method, high-resolution melting curve method and / or gene chip method.

Images