Methods for detecting tuberculosis in body fluid samples

Abstract

Methods and systems for detecting TB in bodily fluid samples are described. By extracting extracellular vesicles (EVs) from bodily fluid samples and using antibody-conjugated nanoparticles to capture Mtb-related biomarkers, the results show that the test can be completed within hours rather than weeks, and the detection limit can be significantly reduced with high accuracy. Furthermore, the methods and systems of the present invention can distinguish between active and latent TB infection.

Description

[0001] Prior related applications

[0002] This application claims priority to U.S. Provisional Application No. 63 / 061,674, filed August 5, 2020, which is incorporated herein by reference in its entirety for all purposes.

[0003] Federally sponsored research statement

[0004] not applicable.

[0005] Field of the present invention

[0006] The present invention generally relates to a method for diagnosing tuberculosis in body fluid samples, and more specifically to a method for detecting tuberculosis-derived antigens on circulating extracellular vesicles (EVs) in body fluid samples using a nanoplasma-enhanced assay with dark-field microscopy.

[0007] Background of the Invention

[0008] Tuberculosis (TB) has become a greater burden globally than previously estimated, with potentially millions of undiagnosed and unreported cases each year. Timely diagnosis of TB remains urgent for positive patient outcomes. However, traditional techniques such as Mycobacterium tuberculosis (MTB) culture and sputum smears are time-consuming and have low sensitivity. The Xpert MTB / RIF test can rapidly diagnose TB and assess rifampicin resistance within 2 hours; however, its sensitivity depends on high bacterial loads in sputum samples. Despite significant efforts to explore TB pathogen-derived antigenic markers, the World Health Organization has only reviewed lipoarabinomannan (LAM), a TB antigen identified as flowing into urine during active TB, for rapid early screening of HIV-positive TB patients. However, the inability to detect LAM in the urine of HIV-negative patients limits the applicability of these tests, as 85% of TB patients are HIV-negative. Furthermore, this test is only applicable to urine samples.

[0009] Ideally, invasive assays should simultaneously measure pathogenic and host targets in bodily fluids to achieve the highest accuracy in tuberculosis detection at different stages, including latent infection. Extracellular vesicles (EVs) closely associated with the pathogenic process can contain numerous host and component elements for marker discovery. The failure to detect LAMs in HIV-negative patients, despite their extensive research and recognition as a tuberculosis-specific marker, may be due to LAMs being masked by immune complex formation. LAMs have been well-established to mediate the release of EVs from macrophages.

[0010] Therefore, there is a need to develop a rapid, quantitative, ultrasensitive, non-sputum-based biomarker detection method for active tuberculosis.

[0011] Overview of the Invention

[0012] In one aspect of the invention, a method for detecting Mycobacterium tuberculosis (MTB)-specific proteins in bodily fluid samples is described. The method comprises the steps of: (a) extracting extracellular vesicles (EVs) from the bodily fluid sample; (b) mixing antibody-conjugated nanoparticles with the EVs from step (a), wherein the antibody-conjugated nanoparticles are conjugated with a first antibody against the MTB-specific protein; and (c) detecting the presence of the MTB-specific protein using dark-field microscopy.

[0013] In another aspect of the invention, a system for detecting MTB-specific proteins in bodily fluid samples is described. The system comprises: (a) a dark-field microscope; (b) a sample matrix; and (c) antibody-conjugated nanoparticles, wherein the antibody-conjugated nanoparticles are conjugated to a first antibody targeting an MTB-specific protein; wherein the sample matrix is ​​coated with a second antibody against an extracellular vesicle-specific (EV-specific) protein.

[0014] In another aspect of the invention, a method for detecting and determining tuberculosis infection status by detecting the presence of first and second Mycobacterium tuberculosis (MTB)-specific proteins in a bodily fluid sample is described. The method includes the following steps: a) extracting extracellular vesicles (EVs) from the bodily fluid sample; b) mixing antibody-conjugated nanoparticles with the EVs extracted in step (a), wherein the antibody-conjugated nanoparticles are conjugated with a first antibody against the first MTB-specific protein and a second antibody against the second MTB-specific protein; c) detecting the presence of the first and / or second MTB-specific proteins using dark-field microscopy; and d) determining the tuberculosis infection status based on the presence of the first and second MTB-specific proteins.

[0015] In another aspect of the invention, a method for screening antibodies against an MTB-specific protein is described. The method includes the steps of: (a) immobilizing an MTB-specific protein on a substrate; (b) introducing a plurality of first antibodies onto the substrate; (c) mixing nanoparticles with the mixture from step (b), wherein the nanoparticles are conjugated with a second antibody and a signal emitting group, wherein the second antibody targets a heavy chain constant region of the first antibody; and (d) detecting the presence of an antibody against the MTB-specific protein by detecting a signal emitted by the signal emitting group on the nanoparticles.

[0016] In another aspect of the invention, a method for detecting the presence of a bacterial-specific protein in a bodily fluid sample is described. The method includes: (a) extracting extracellular vesicles (EVs) from the bodily fluid sample; (b) mixing antibody-conjugated nanoparticles with the EVs, wherein the antibody-conjugated nanoparticles are conjugated with a specific first antibody of the bacterial-specific protein; and (c) detecting the presence of the bacterial-specific protein using dark-field microscopy.

[0017] In one embodiment, the body fluid sample may be obtained from the patient. For example, the body fluid may be blood, serum, sputum, urine, or other available body fluids.

[0018] In one embodiment, the antigen of the first antibody is selected from the Mtb-specific antigen group, and the Mtb-specific antigen group consists of lipoarabinomannan (LAM), antigen 85B, LAM carrier proteins LprG and LpqH, α-crystal protein (HspX), DnaK, GroEL2, KatG, SodA, and GlnA. In one embodiment, the antigen of the second antibody is selected from the same group but different from that of the first antibody.

[0019] In one embodiment, a capture antibody, such as an anti-CD81 or anti-MTB antibody, is used to extract the EVs. Detection antibodies that recognize surface markers on exosomes or MTB-derived EVs can also be used, and non-limiting examples include CD9, CD91, CD63, PDCD6IP, HSPA8, ACTB, ANXA2, PKM, HSP90AA1, ENO1, ANXA5, HSP90AB1, YWHAZ, YWHAE, LprG, LpqH, α-crystallin (HspX), DnaK, GroEL2, KatG, SodA, and GlnA. In one embodiment, an anti-CD81 antibody is used to extract the EVs.

[0020] In one embodiment, the nanoparticles are those that, when excited by a light source, can generate a surface plasmon resonance effect to significantly increase the fluorescence emitted by inorganic fluorescent particles. Gold nanoparticles are most commonly used due to their stability and easily modifiable surface. For antibody conjugation, the surface of gold nanoparticles can be modified, for example, using carboxyl groups. Thus, negatively charged gold nanoparticles can form covalent bonds with positively charged amino groups. However, other nanoparticles can also be used, and non-limiting examples include nanoparticles made of silver (Ag), bismuth oxide (Bi₂O₃), platinum (Pt), gadolinium oxide (Gd₂O₃), iron oxide (Fe₃O₄), etc., as long as these particles can exhibit enhanced light scattering.

[0021] In another aspect of the invention, a system for screening antibodies against an MTB-specific protein is described. The system comprises: a substrate having a coating layer covalently bonded to the MTB-specific protein; the MTB-specific protein; and nanoparticles conjugated to an antibody against the constant region of the anti-IgG heavy chain.

[0022] As used in this article, "Mycobacterium tuberculosis" or "MTB" refers to a pathogenic bacterium in the family Mycobacterium tuberculosis and the cause of tuberculosis. MTB grows very slowly, approximately doubling in size every day.

[0023] As used in this article, "sample" refers to a small amount of biological material collected from a subject.

[0024] As used herein, "sample substrate" refers to a substrate on which a sample can be placed and examined by microscopy. In one embodiment, the sample substrate is a glass slide, such as a patterned glass slide on which multiple samples can be deposited. However, other sample substrates may also be used, such as glass slides or transparent plates made of other materials.

[0025] As used herein, “extracellular vesicles” or “EVs” refer to particles defined by a lipid bilayer that are naturally released from cells and bacteria and cannot self-replicate. EVs range in diameter from approximately 20-30 nm to approximately 10 μm or larger. EVs are capable of transferring nucleic acids, such as RNA, between cells. EVs are typically separated from bodily fluid samples by ultracentrifugation or density gradient ultracentrifugation, size exclusion chromatography, ultrafiltration, and affinity / immunoaffinity capture. Certain EV enrichment markers can be used to better separate EVs. Examples of EV enrichment markers include, but are not limited to, CD81, PDCD6IP, HSPA8, ACTB, ANXA2, CD9, PKM, HSP90AA1, ENO1, ANXA5, HSP90AB1, CD63, YWHAZ, YWHAE, and antibodies against LprG, LpqH, α-crystallin (HspX), DnaK, GroEL2, KatG, SodA, and GlnA.

[0026] As used herein, “dark-field microscopy” refers to a microscopic technique that removes unscattered light beams from an image, thereby darkening the background around the sample. In optical microscopy, dark-field microscopy describes an illumination technique used to enhance contrast in unstained samples, achieved by illuminating the sample with light that is not collected by the objective lens and therefore does not form part of the image. When combined with hyperspectral imaging, dark-field microscopy can be used to characterize nanomaterials embedded in cells, such as gold nanoparticles targeting cells with specific markers.

[0027] As used herein, "antibody-conjugated nanoparticles" refers to nanoparticles conjugated with specific antibodies against the target antigen. The conjugation between the antibody and the nanoparticle can be achieved through electrostatic interactions (physical adsorption) or through antibody-oriented covalent conjugation on a metal surface. Static ion adsorption of nanoparticles to antibodies has been reported to exhibit poor reproducibility, random antibody orientation, and low stability under varying pH conditions. Covalent conjugation with the amino acid side chains of antibodies is promoted by modifying the nanoparticle surface with active groups (such as carboxyl and amino groups).

[0028] As used herein, antigen 85B refers to a subunit of the antigen 85 (Ag85) complex (Ag85A, Ag85B, Ag85C) found to be produced in MTB culture medium. The 85A, 85B, and 85C proteins are encoded by three genes located at different positions in the mycobacterial genome and exhibit broad cross-reactivity and homology at both the amino acid and gene levels.

[0029] As used in this article, "LAM carrier protein LprG" refers to lipoarabinomannan carrier protein LprG.

[0030] As used in this article, "LpqH" refers to the lipoprotein LpqH found in MTBs. The 19 kDa Mycobacterium tuberculosis lipoprotein (LpqH) induces macrophage apoptosis through external and internal pathways: the role of mitochondrial apoptosis-inducing factors.

[0031] As used in this article, "α-crystal protein (HspX)" refers to the 16 kDa heat shock protein HspX required for the continued existence of mycobacteria in microphages.

[0032] As used in this article, “DnaK” refers to the bacterial molecular chaperone protein DnaK. Molecular chaperones are proteins that bind to other proteins, thereby stabilizing them in an ATP-dependent manner. DnaK is an enzyme that combines the cycle of ATP binding, hydrolysis, and ADP release caused by the N-terminal ATP-hydrolyzing domain with the cycle of chelation and release of unfolded proteins caused by the C-terminal substrate-binding domain.

[0033] As used in this article, "GroEL2" refers to 60 kDa chaperone protein 2 (also known as Cpn60.2), which is closely associated with the Cpn60.1 molecular chaperone located on the outer layer of the Mycobacterium tuberculosis cell wall. GroEL2 has been found in the cerebrospinal fluid of patients with tuberculous meningitis.

[0034] As used in this article, "KatG" refers to the catalase-peroxidase encoded by the katG gene in Mycobacterium tuberculosis, which activates the prodrug INH. Mutation of the katG gene in Mycobacterium tuberculosis is a major mechanism of INH resistance.

[0035] As used herein, “SodA” refers to superoxide dismutase [Fe]. For MTB assays, unless otherwise specified, SodA specifically refers to MTB SodA.

[0036] As used in this article, "GlnA" refers to glutamine synthase. For MTB detection, GlnA specifically refers to MTB GlnA.

[0037] As used in this article, "PDCD6IP" refers to the programmed cell death 6-interacting protein, which encodes a protein believed to be involved in programmed cell apoptosis.

[0038] As used in this article, "HSPA8" refers to human heat shock 70 kDa protein 8, also known as heat shock homolog 71 kDa protein, Hsc70, or Hsp73. As a member and chaperone protein of the heat shock protein 70 family, it contributes to the correct folding of newly translated and misfolded proteins, as well as to stabilizing or degrading mutant proteins.

[0039] As used in this article, “ACTB” refers to human β-actin, which is one of six different actin subtypes that have been identified in humans.

[0040] As used in this article, "ANXA2" refers to annexin A2, which is involved in a variety of cellular processes, such as cell movement, the connection of membrane-associated protein complexes to the actin cytoskeleton, endocytosis, fibrinolysis, ion channel formation, and cellular matrix interactions.

[0041] As used in this article, "PKM" refers to pyruvate kinase M1 / 2, which catalyzes the transfer of phosphoryl groups from phosphoenolpyruvate to ADP, generating ATP and pyruvate.

[0042] As used in this article, "HSP90AA1" refers to the human heat shock protein HSP90α (cytoplasm), member A1. Hsp90A is expressed when cells experience proteotoxic stress, complemented by the constitutively expressed homologous Hsp90B, which has over 85% amino acid sequence identity. Once expressed, the Hsp90A dimer functions as a molecular chaperone, binding to other proteins and folding them into their functional three-dimensional structures.

[0043] As used in this article, “ENO1” refers to α-enolase, a glycolytic enzyme expressed in most tissues. Each isoenzyme is a homodimer composed of two α, two γ, or two β subunits, functioning as a glycolytic enzyme. In addition, α-enolase also functions as a monomeric structural crystal protein (tau-crystal protein).

[0044] As used in this article, "ANXA5" refers to annexin A5, a cellular protein in the annexin group. ANXA5 can bind to phosphatidylserine, a marker of apoptosis located on the outer lobe of the plasma membrane.

[0045] As used in this article, "HSP90AB1" refers to the heat shock protein HSP90-β, a molecular chaperone.

[0046] As used in this article, "YWHAZ" refers to the 14-3-3 protein ζ / δ, a member of the 14-3-3 protein family and a central protein in many signal transduction pathways. It is a major regulator of the apoptosis pathway, which is essential for cell survival, and plays a key role in many cancers and neurodegenerative diseases.

[0047] As used in this article, “YWHAE” refers to 14-3-3 protein ε, a member of the 14-3-3 family, which mediates signal transduction by binding to proteins containing phosphoserine.

[0048] Unless the context otherwise requires, when used with the term "comprising" in the claims or specification, the use of the word "a" or "an" refers to one or more.

[0049] Unless otherwise specified, the term “about” means the value plus or minus a measurement error margin, or plus or minus 10%.

[0050] Unless explicitly stated otherwise, referring only to alternatives or if the alternatives are mutually exclusive, the use of the term "or" in the claims means "and / or".

[0051] The terms “comprising,” “having,” “including,” and “containing” (and variations thereof) are open-ended connecting verbs that allow for the addition of other elements in the claims.

[0052] The phrase “composed of” is closed and does not include any other elements.

[0053] The phrase “consistent essentially of” does not include additional material elements, but may include non-material elements that do not substantially alter the nature of the invention.

[0054] The following abbreviations are used in this article:

[0055] abbreviation the term Ag85b Antigen 85B AuNRs Gold nanorods EVs Extracellular vesicles or exosomes LAM Lipo-arabinomannan LC Liquid chromatography MS Mass spectrometry MTB Mycobacterium tuberculosis nPES Nanoplasma Enhanced Scattering TB tuberculosis

[0056] Brief description of the attached figures

[0057] Figure 1Purification and characterization of EVs from TB-infected macrophages for marker discovery, including protein identification and lipoprotein glycan confirmation. (a) In vitro TB infection model protocol for marker discovery. (b) Number of different protein types identified in EVs isolated from TB-infected macrophages using LC-MS proteomics. (c) Heatmap showing folding changes of several proteins between culture medium and EVs purified from TB-infected macrophages using LC-MS proteomics. (d) Western blotting used to confirm LAM (37 kDa) in the culture filter protein (CFP) from different TB strains, EVs, and cytoplasm of TB-infected macrophages.

[0058] Figure 2 Nanoplasma-enhanced immunoassay for highly sensitive detection of TB antigen on macrophage-derived EVs infected with TB. (a) Protocol for detecting TB antigen on EVs using AuNRs-based nanoplasma-enhanced immunoassay. (b) UV-vis spectrum of AuNRs. (c) TEM image of AuNRs. (dg) Dark-field images of different concentrations of AuNRs. (h) Quantification of signals from dark-field images of different concentrations of AuNRs. (i) Quantified intensity from dark-field images of samples containing different concentrations of LAM. Mean ± SD; n = 6. (j) Comparison of the intensity of dark-field images of samples containing 10 μg / mL macrophage-derived EVs infected with TB with the intensity in the absence of EVs, using different combinations of capture and detection antibodies. (k) Intensities quantified from dark-field images of samples containing macrophage-derived EVs with different concentrations of TB infection, using anti-CD81 antibody as the capture antibody and anti-LAM antibody (A194-01) as the detection antibody, mean ± SD; n = 6.

[0059] Figure 3Performance application of nanoplasma-enhanced immunoassay in diagnosing tuberculosis in HIV-negative children. (a) Strength of purified EVs from serum, purified EVs, or purified supernatant from 4 TB patient serum samples and 2 control serum samples. (b) Strength and OD450 of nanoplasma-enhanced immunoassay in 15 TB patient serum samples and 5 control serum samples using anti-LAM as the detection antibody, as measured by ELISA. (c) Strength of TB patient samples and control samples when anti-LAM antibody is used as the detection antibody in conventional ELISA. (df) Strength of TB patient samples and control samples when (d) anti-LAM, (e) anti-LprG, and (f) anti-LpqH are used as detection antibodies. t-test, **, p < 0.01, ***, p < 0.001. (g) ROC curves for diagnosing 20 Vietnamese samples using nanoplasma-enhanced immunoassay with different detection antibodies and ELISA using anti-LAM antibody.

[0060] Figure 4A The schematic diagram and analysis diagram are shown. OMVs: Mtb outer membrane vesicles.

[0061] Figure 4B Size distribution of EVs secreted by uninfected and Mtb H37Rv-infected macrophages.

[0062] Figure 4C The values ​​of LAM and LprG (mean ± SD of three technical replicates) on serum EVs of nonhuman primates with pulmonary tuberculosis (PTB), latent TB infection (LTBI) or their healthy controls (Ctrl) are shown.

[0063] Figure 4D The expression of LAM, LprG, and combined LAM and LprG (LAM+LprG) in serum EVs of children with TB (N=10) and without evidence of TB was shown by EV-ELISA (control; N=5; mean ± SE by nonparametric Kruskal-Wallis univariate ANOVA and Dunn's posttest, *p<0.05 and **p<0.01).

[0064] Figure 5A This is a schematic diagram of the NEI image capture workflow and signals.

[0065] Figure 5B The .EV LAM and LprG NEI signals were linearly correlated with the Mtb EV concentration curve generated using EVs from Mtb-infected macrophages.

[0066] Figure 5C The ability to differentiate between children with and without TB by single and combined EV LAM and LprG NEI signals was analyzed by receiver operating characteristic (ROC) analysis, showing the region under the ROC curve.

[0067] Figure 5D -F. NEI signals of (D) LAM, (E) LprG, and (F) combined LAM and LprG expression in serum EVs of children with TB (N=10) and without evidence of TB (N=5). Solid lines represent mean ± SE; dashed lines represent the positive signal threshold (C) determined in the corresponding ROC analysis.

[0068] Figure 6 Diagnostic manifestations of Mtb EV NEI in HIV-infected children at high risk of tuberculosis. NEI signals are defined as follows for children with confirmed TB, pending TB, and unlikely TB: positive respiratory culture / Xpert or stool Xpert results, duration of TB-related symptoms meeting NIH criteria, chest X-ray (CXR) results, close contact with TB or positive TST, positive ATT response, and / or TB-related death.

[0069] Figure 7A This is a schematic diagram of a portable smartphone-based DFM device used for NEI analysis readings.

[0070] Figure 7B Showing Figure 7A The dark-field focusing mask used. Invention Details

[0072] This invention provides a novel method and system for detecting the presence of MTB in bodily fluid samples and extracted EVs, which differs from traditional TB detection methods that require sputum samples. Bodily fluid samples, such as blood, saliva, or urine, are easier to extract from patients, while sputum samples are difficult to obtain.

[0073] Furthermore, the method and system of this invention can be used to rapidly determine the presence of MTB within hours, even if the body fluid sample contains only a small amount of MTB protein. Traditional TB detection methods require sputum culture, which takes 1-8 weeks. The rapid turnaround time helps physicians make treatment decisions and prevents the spread of tuberculosis.

[0074] Furthermore, the method of this invention enables the high-precision detection of TB by extracting EVs without concentration. In other practices, once EVs are extracted, a further centrifugation step is typically performed to concentrate the EVs. However, in this invention, the highly sensitive nanoparticles and dark-field microscopy make it easier to detect TB even in bodily fluid samples without further increasing the concentration of EVs in the sample.

[0075] Furthermore, the methods and systems of this invention can be used to detect TB in non-HIV patients and latent patients. Traditional TB tests are only sensitive to HIV-positive patients and cannot distinguish between latent TB patients and TB-negative patients. The methods and systems of this invention have been shown to detect the presence of MTB in HIV-negative patients and also to detect the presence of MTB in latent, asymptomatic TB patients. The high sensitivity and specificity of the methods and systems of this invention allow for detection using bodily fluid samples instead of sputum, as bodily fluid samples are more readily available.

[0076] Perhaps more importantly, the methods and systems of the present invention provide a combination of biomarkers that can be used to distinguish between patients with active TB and patients with latent TB.

[0077] To achieve these results, the present invention describes a method for detecting the presence of an MTB-specific protein in a body fluid sample, comprising the steps of: a) extracting extracellular vesicles (EVs) from the body fluid sample; b) mixing nanoparticles with the EVs, wherein the nanoparticles are conjugated with a specific first antibody against the MTB-specific protein; and d) detecting the presence of MTB using dark-field microscopy.

[0078] Alternatively, the present invention describes a method for screening antibodies against an MTB-specific protein, comprising the steps of: a) immobilizing the MTB-specific protein on a substrate; b) introducing a plurality of first antibodies onto the substrate; c) mixing nanoparticles with the mixture from step (b), wherein the nanoparticles are conjugated with a second antibody and a signal emitting group, wherein the second antibody targets the heavy chain constant region of the first antibody; and d) detecting the presence of an antibody against the MTB-specific protein by detecting a signal emitted by the signal emitting particles. This screening method can effectively screen a large number of antibodies to obtain antibodies targeting the MTB-specific protein.

[0079] The present invention also describes a system for detecting MTB protein in body fluid samples, comprising: a) a dark-field microscope; b) a sample substrate; and c) antibody-conjugated nanoparticles, wherein the antibody-conjugated nanoparticles comprise an anti-MTB antibody targeting the MTB protein, and wherein the sample substrate is coated with a second antibody against an extracellular vesicle-specific protein.

[0080] Alternatively, the present invention describes a method for detecting and determining tuberculosis infection status by detecting the presence of first and second Mycobacterium tuberculosis (MTB)-specific proteins in a bodily fluid sample. The method includes the following steps: a) extracting extracellular vesicles (EVs) from a bodily fluid sample; b) mixing antibody-conjugated nanoparticles with the EVs extracted in step (a), wherein the antibody-conjugated nanoparticles are conjugated with a first antibody against a first MTB-specific protein and a second antibody against a second MTB-specific protein; c) detecting the presence of the first and / or second MTB-specific proteins using dark-field microscopy; and d) determining the tuberculosis infection status based on the presence of the first and second MTB-specific proteins.

[0081] The methods and systems of this invention focus on extracellular vesicles (EVs) in subjects that possess at least one MTB protein. EVs have specific surface markers that can be targeted by antibodies, and at least one MTB protein also has an epitope that can be targeted by antibodies. Therefore, pathogenic targets and host targets can be detected simultaneously in bodily fluids containing EVs.

[0082] This invention describes a novel method for detecting the presence of MTB in a sample, which involves first extracting extracellular vesicles (EVs) from the sample and then detecting MTB-specific markers from the EVs. To do this, the first step is to identify the MTB-specific markers present in the EVs, thus enabling the capture of these markers.

[0083] 1. Identification of TB-specific markers

[0084] To explore TB-specific markers for EVs, macrophages were infected with TB to isolate EVs in vitro for proteomics studies via the following steps. Figure 1 As shown in figure a, marker discovery was performed using an in vitro TB infection model. Extracellular vesicles (EVs) secreted by macrophages were purified after TB infection. Furthermore, liquid chromatography-mass spectrometry-based proteomics (LC-MS) studies were conducted for protein marker identification, and immunoassays were performed for glycolipid confirmation.

[0085] Collection of culture medium from macrophages infected with active MTB

[0086] To provide more accurate and realistic results, it is important to conduct experiments based on TB-infected macrophages. The following materials were used:

[0087] Total required sample: 1400-2100 mL of culture medium in T175 flasks (4 x 10 x 35-52.5 mL = 1400-2100 mL), each flask containing 2.5 x 10⁷ activated THP-1 infected with MTB (MOI = 10:1) [4-hour infection].

[0088] Cells, bacteria, and reagents: THP-1 TIB202 TM M.tb H37Rv (ideally some other strains with different virulence), RPMI 1640 without PS medium, RPMI 1640 (without FBS, without PS) medium.

[0089] THP-1 cell culture ( TIB202 TM Thaw the frozen cells in a 37°C water bath. Thawing should be rapid (approximately 2 minutes).

[0090] 1. Transfer the contents of the vial to a centrifuge tube containing 9.0 mL of complete culture medium. Then rotate at approximately 125 x g for 5 minutes.

[0091] 2. Count the number of cells and the viability (not less than 95%). Resuspend the cell particles in intact culture medium with a final concentration of 3 x 10⁵ live cells / ml.

[0092] 3. The culture was incubated at 37°C in a suitable incubator containing 5% CO2.

[0093] 4. When the cell concentration reaches 8 x 10⁵ cells / ml, perform subculture. The cell concentration should not exceed 1 x 10⁶ cells / ml. Centrifuge the culture medium and resuspend the cell particles in the same manner as described above.

[0094] 5. When the total number of cells reaches 2.5 x 10⁸, the cells are evenly distributed into 10 T175 culture flasks (35-52.5 ml) (2.5 x 10⁷ cells / flask), and 500 ng / ml PMA is added for 24 hours to differentiate THP-1 cells into macrophages.

[0095] 6. Examine the cell morphology. If the cells have fully expanded, gently remove the culture medium and wash the cells three times with 37°C PBS.

[0096] Then the cells can accept infection.

[0097] Active MTB infection with THP-1

[0098] i) MOI = 10, the total number of non-integrative nucleolytic mycobacteria required for THP-1 infection (M.tb H37Rv stock solution) = 2.5 x 10⁹.

[0099] ii) Resuspend the Mtb particles in a small amount (5-10 mL) of RPMI 1640 medium containing 10% FBS, remove clumps by brief sonication in a water bath, and pass the Mtb particles through a syringe fitted with a 27-gauge needle at least 10 times.

[0100] iii) Add more RPMI 1640 PS-free medium (PS = penicillin and streptomycin) and 10% FBS, and mix thoroughly. Then divide the mixture into 10 equal portions in 10 macrophage culture flasks (20 mL RPMI PS-free medium in each flask).

[0101] iv) Infect under culture conditions for 4 hours.

[0102] v) After infection, remove the mixture and gently wash the cells with 37°C PBS (10 ml / time, 3 or more times) (examine under a microscope).

[0103] vi) Add RPMI 1640 (FBS-free, PS-free) to the culture flask and incubate for 48 hours.

[0104] vii). Collect the culture medium. Filter the culture medium using a 0.45 μm filter to remove bacteria. Store the culture medium in a -80°C freezer and inoculate a small amount of the medium for 3-4 weeks (Mtb) to ensure that no live bacteria are present.

[0105] The supernatant from the culture medium was collected for EV isolation, and then LC-MS proteomics studies were performed using standard procedures. EV isolation is well-known in the field and will not be described further here.

[0106] like Figure 1 As shown in (b), 78 TB-derived proteins were detected in macrophage-derived EVs infected with TB, of which 36 were membrane proteins that could be present on the surface of the EV membrane for further detection without destroying the EVs. Seventeen of these 36 proteins showed high abundance and showed promise as markers for further detection in TB diagnosis.

[0107] The TB-derived EV-specific markers identified in this study include lipoarabinomannan (LAM), antigen 85B (Ag85B), LAM carrier proteins LprG and LpqH, α-crystal protein (HspX), DnaK, GroEL2, KatG, SodA, and GlnA.

[0108] like Figure 1As shown in (c), several proteins (including Ag85B, DnaK, EsxA, EsxO, EsxN, LprO, and PepA) were significantly overexpressed in TB-infected macrophage-derived EVs compared to the culture medium from TB-infected macrophage-derived EVs. Among these proteins, Ag85b showed the highest fold change in EVs compared to the culture medium from TB-infected macrophage-derived EVs. Since LC-MS is insensitive to LAM (a well-known glycolipid (37 kDa) that plays a crucial role during TB infection), Western blotting was performed using two different anti-LAM antibodies (clone numbers: CS-35 and CS-40) as detection antibodies to confirm the presence of LAM in EVs isolated from different strains of TB-infected macrophages. Figure 1 The results in (d) confirm the existence of LAM.

[0109] 2. Validation of markers

[0110] The markers identified in the screening experiment were then verified using Western blotting (data not shown).

[0111] 3. Detect the presence of Mtb in EVs

[0112] We developed and optimized a nanoplasma-enhanced immunoassay to achieve highly sensitive detection of TB antigens on macrophage-derived EVs infected with TB. 1 μL of mouse anti-CD81 antibody (5 μg / mL) was coated onto a patterned glass slide. After washing and blocking, 1 μL of serum was added, and the slide was incubated at 37°C for 1 hour, followed by two washes with PBS. LAM on the EVs was then captured using 1 μL of biotin-functionalized human anti-LAM antibody (1 μg / mL, A194-01). Subsequently, the immobilized antibody could conjugate avidin-functionalized gold nanoparticles to produce a scattering signal observable by dark-field microscopy.

[0113] The human anti-LAM antibody (A194-01) used in this invention is a monoclonal antibody targeting epitopes found in lipoarabinomannan (LAM) and phosphatidyl-myo-inositol mannoside 6 (PIM6), and is used for the diagnosis and treatment of MTB infection. The human monoclonal anti-LAM antibody A194-01 was isolated from cultured memory B cells obtained from TB-194 patients infected with TB.

[0114] Please refer to Figure 2 a. EVs were captured on a glass slide using a capture antibody, and the TB antigen was bound to the EV membrane using a detection antibody. The detection antibody described herein was further conjugated with gold nanorods (AuNRs) for signal reading. The signal readings of the AuNRs are shown below. Figure 2 As shown in b.

[0115] like Figure 2 As shown in c, AuNRs exhibit an absorption peak with a uniform size distribution at 650 nm under TEM, and as... Figure 2 As shown in dg, under dark-field microscopy, the red scattered light gradually increases with increasing AuNRs concentration. Figure 2 As shown in h, the signal is quantified by calculating the number of scattering objects and the average intensity of the pixels forming these images. The average intensity shows a better linear response with AuNR concentration.

[0116] like Figure 2 As shown in (i), LAM was used as an exemplary target for this method, and several anti-LAM antibodies (CS-35, CS-40, and A194-01) were tested. CS-35 showed the strongest signal with a relatively high background, while A194-01 showed the best linear response to LAM. Different combinations of capture and detection antibodies were further tested to detect macrophage-derived EVs infected with 10 μg / mL TB. Figure 2 As shown in Figure j, using anti-CD81 antibody as the capture antibody and anti-LAM antibody (A194-01) as the detection antibody, a strong signal ratio relative to the blank control (in the absence of EVs) was generated under nanoplasma-enhanced immunoassay, and a dynamic range of 0-15 μg / mL was obtained for the detection of macrophage-derived EVs infected with TB under this antibody combination. Figure 2 k). This range provides a practical detection limit for TB-specific EVs. The CD-81 / A194-01 combination is further used to detect EVs in serum samples for TB diagnosis.

[0117] The method of this invention is also used to detect TB antigen on EVs from patient serum samples, with results as follows: Figure 3 As shown. In Figure 3 In this study, anti-CD81 antibody was used as the capture antibody and anti-LAM antibody (A194-01) was used as the detection antibody. Both whole serum and purified EVs from these samples from TB patients showed positive results.

[0118] Further testing was conducted on serum samples from 15 untreated TB patients and 5 control serum samples, with results as follows: Figure 3 As shown in b. Most of these samples produced higher signals compared to the control. Conversely, conventional ELISA could not distinguish between TB and the control group.

[0119] Anti-LprG, anti-LpqH (two LAM carrier proteins), and anti-Ag85B were also used as detection antibodies in nanoplasma-enhanced immunoassays, with results as follows: Figure 3 As shown in ce. When using anti-LAM ( Figure 3 c) Anti-LprG ( Figure 3 d) and resistance to LpqH ( Figure 3 e) When used as a detection antibody, a significant difference was observed between TB and the control group. The clear difference between TB patient and control samples indicates that the method of the present invention can effectively detect the presence of TB in samples.

[0120] The area under the curve (AUC) is a measure of the sensitivity / specificity of a test method; a larger AUC indicates better test results. For example... Figure 3 As shown in f, the AUC for conventional ELISA is only 0.68, while the AUC for conventional Ag85B-based nanoplasma-enhanced immunoassay is 0.627. Conversely, according to the present invention, using anti-LAM, anti-LprG, and anti-LpqH as detection antibodies, and A194-01 as a capture antibody, the nanoplasma-enhanced immunoassay achieves AUCs of 0.91, 0.90, and 0.83, respectively. The results indicate that the assay of the present invention can achieve higher sensitivity and selectivity for effective TB diagnosis.

[0121] Additional samples were collected and tested from 15 TB patients. Dark-field imaging results and standardized test intensity results (not shown) from 15 TB patients (TB1-14 and 16), 5 control serum samples (C1-5), and two exosome control samples (THP1 and THP1+CFP) were analyzed.

[0122] These results clearly demonstrate that the methods and systems of the present invention can detect the presence of MTB in bodily fluid samples (such as blood samples) and differentiate between TB-negative and TB-positive samples. The methods and systems of the present invention can also detect the presence of MTB in HIV-negative patients, which has not been demonstrated in the art. Furthermore, the rapid detection process ensures rapid and accurate diagnosis, reducing waiting time compared to conventional culture tests.

[0123] Furthermore, image recognition software can optimize signal readings in immunoassays to lower the detection limit. It has also been reported that different types of EVs may improve detection accuracy; therefore, further focus is placed on quantifying LAM and other biomarkers in different types of EVs from patient urine or serum samples. Finally, correlating LAM and host biomarker concentrations with clinical information in TB patients can also provide insights for selecting appropriate treatment regimens.

[0124] 4. Automatic detection of Mtb in EVs

[0125] The Automated Nanoparticle-Enhanced EV Immunoassay (NEI) uses machine learning to detect EVs secreted by Mtb-infected cells (Mtb-EVs) based on the surface expression of factors abundantly expressed on the Mtb outer membrane. Lipoarabinomannan (LAM) is one of the targets in this embodiment because it accounts for 15% of Mtb biomass and modulates Mtb virulence. Another target is LprG, which is required for the distribution of LAM to the outer cell membrane. It is hypothesized that EVs secreted by Mtb-infected macrophages will exhibit surface expression of LAM and LprG, which can serve as biomarkers for TB disease, and that these Mtb EVs will accumulate in peripheral blood circulation and other bodily fluids (such as urine, saliva, etc.), thus allowing for the detection of TB in the lungs and extrapulmonary regions. Figure 4A ).

[0126] Using the NEI method, it was found that when combined Mtb-EVs biomarkers integrating LAM and LprG surface expression can distinguish between active TB and latent TB infection (LTBI) in a non-human primate disease model, they can differentiate children with extrapulmonary and pulmonary TB from TB-negative controls, and can diagnose TB in a diagnostically challenging group—hospitalized, severely immunosuppressed HIV-infected children who are at high risk for TB.

[0127] EVs secreted by Mtb-infected macrophages expressed LAM and LprG. To assess the potential of EV LAM and LprG expression as biomarkers for Mtb infection or TB disease, we first examined the expression of these factors in EVs secreted by Mtb-infected macrophages. EVs isolated from macrophage cultures infected or uninfected with the Mtb H37Rv reference strain showed similar morphology and size distribution (not shown), although significantly more EVs were secreted by Mtb-infected macrophages (…). Figure 4B 2.3 × 10⁹ vs. 1.1 × 10⁹ EVs / mL (p < 0.01, Kruskal-Wallis test). Based on the analysis of equal volumes of cytoplasmic, cell membrane, and EV protein extracts from macrophages infected with Mtb strains exhibiting different growth rates and immunogenicity, LAM and LprG showed significant EV enrichment, suggesting their potential utility as strain-independent biomarkers for Mtb infection (not shown).

[0128] EV-ELISAs are the gold standard method for detecting EV surface biomarker 25. Serial dilutions of EVs generated from macrophages incubated with Mtb culture filtrate protein (CFP) extract also detected dose-dependent increases in LAM and LprG signals (not shown), validating the surface expression of both biomarkers.

[0129] Next, the ability of serum EVs isolated from non-human primates that have never been infected with Mtb or have developed LTBI or pulmonary tuberculosis (PTB) after exposure to Mtb was evaluated to assess the ability of serum EV LAM and LprG to distinguish active TB, latent TB infection (LTBI), and healthy controls. An NHP model was used in this analysis to ensure reliable differentiation between active TB and LTBI, as it is difficult to distinguish LTBI from newly diagnosed or subclinical TB in human patient populations. When analyzed with LAM and LprG EVELISAs, serum EVs isolated from NHPs that have never been infected with Mtb showed low nonspecific background signal (…). Figure 4C In the LTBI group, serum EVs showed elevated expression of either LAM or LprG, but not both; most NHPs with LTBI showed elevated LprG signaling. However, serum EVs from the PTB group showed elevated expression of both LAM and LprG, indicating a linear relationship, suggesting that these two factors could serve as composite biomarkers for specific TB diagnosis. Figure 4D The expression of LAM, LprG, and integrated LAM and LprG (LAM+LprG) in serum EVs was detected by EV-ELISA in children with TB (N=10) and without evidence of TB. The results showed that the expression of both LAM and LprG was significantly increased in TB patients compared to the TB-negative control group. This further indicates that LAM and LprG are positive biomarkers for TB.

[0130] As part of a large population-based tuberculosis surveillance study, a similar analysis (not shown) was performed on EVs isolated from archived serum of a small, well-characterized case-control pediatric population (registered at the National Lung Hospital, NLH, Vietnam). The analysis revealed that the mean EV LAM in the TB and non-TB controls differed from the combined EVALAM and LprG signals, but the EV LprG signal was not different. Other Mtb-derived membrane-associated factors showed no difference in serum EVs between the two groups (not shown), except for LpqH, which exhibited greater signal overlap between the TB and non-TB groups than LAM or LprG.

[0131] 5. NEI was used to detect the expression of LAM and LprG on Mtb EVS with high sensitivity.

[0132] Since target EVs cannot be directly captured and analyzed from serum by EV ELISA, we investigated the feasibility of an NEI method based on dark-field microscopy (DFM) to improve the detection of target EVs. Such an assay can capture EVs from complex biological samples without time-consuming separation steps and identify specific EV populations by detecting light scattered from gold nanorod (AuNR) probes, which bind to target biomarkers on the outer membrane of the EV population. NEI signals are read by analyzing high- or low-magnification DFM images of the sample wells. High-magnification analysis allows for ultrasensitive detection but requires manual focusing to detect plasmonic signals from interacting nanoparticles, which can introduce selection bias due to the limited area of ​​the sampling analysis well. Low-magnification DFM analysis can be automated and is therefore more suitable for clinical applications, but is susceptible to artifacts that can increase background and reduce sensitivity.

[0133] Then the NEI workflow was established. Figure 5A This workflow allows for the automated capture of low-magnification DFM images, which are processed using a custom denoising algorithm to reduce artifacts introduced during the assay from serum aggregates, microparticles, and surface scratches. The algorithm detects and crops regions of interest (ROIs); converts each image to an HSB (hue, saturation, brightness) color space; and normalizes a threshold and applies it to each HSB channel to identify AuNR-derived signals (not shown). This method significantly reduces signal artifacts, manifested as an enhancement of the AuNR signal in the processed image and its pixel intensity map (not shown). For signal detection, analysis of AuNR dilution curves by quantizing positive pixels and average pixel intensity revealed that the latter method exhibits greater linearity and less variability (not shown).

[0134] To determine the probe concentration required to maximize the signal-to-noise ratio of target EVs at low concentrations (close to the detection limit of EV-ELISA), EVs isolated from Mtb-infected macrophages were captured by nonspecific binding and hybridized with dilutions of AuNRs conjugated with anti-LAM antibody (not shown). Evaluation of capture and detection antibody pairs determined that the anti-CD81 capture antibody and two anti-LAM (A194-01, Creative Biolabs) and anti-LprG (clone B) detection antibodies produced the optimal signal-to-noise ratios (10.21 and 7.25, respectively) over a wide range of EV concentrations (0–150 ng EV protein / mL). Figure 5B They all exhibit good linearity and variability when used.

[0135] NEI analysis of serum from NHL-detected individuals revealed that serum EV LprG and LAM signals showed similar ability to distinguish between TB and non-TB groups, but EV LAM and LprG signals integrated using a logistic model (see Methods) exhibited superior differential expression. Figure 5C ), as shown when using the thresholds of each of these analyses to distinguish between TB and non-TB groups ( Figure 5D -F). This analysis showed significant signal overlap between the TB and non-TB groups near the threshold of EV LAM or LprG signals, resulting in three false negatives and one false positive classification, while integrating LAM and LprG signals produced only one false negative identification. NEI signals for LAM and LprG were detected in EV-enriched rather than EV-depleted serum fractions, confirming NEI signaling as EV-specific (not shown). Unlike the results for EV-enriched samples, when EV-ELISA was used to directly analyze the serum of these groups, EV LAM and LprG signals failed to distinguish between TB and non-TB cases, indicating a lack of sensitivity required for EV-ELISA application to serum. No difference in NEI signal was observed when pediatric TB cases were stratified by age, sex, or TB presentation (pulmonary or extrapulmonary), suggesting that NEI signaling is a general biomarker for TB disease.

[0136] 6. Validation of Mtb EV biomarkers in HIV-infected children with TB symptoms

[0137] Next, we evaluated the performance of the NEI test in a diagnostically challenging population of HIV-infected children who are at high risk for TB-related morbidity and mortality, risks that are often missed by respiratory sampling. This analysis used serum collected from HIV-infected children in the multi-site Pediatric Emergency and Stable Post-ART Initiation Trial 145 (PUSH) conducted in Kenya. These children were hospitalized but had not yet started ART at admission. Most children in this study presented with severe immunosuppression and some symptoms consistent with TB. Children were retrospectively categorized according to the 2015 NIH definition of TB in pediatric clinical practice: diagnosed TB if they had microbiological evidence of TB (positive Mtb culture or positive Xpert result from respiratory or fecal samples); unconfirmed TB (based on two or more of the following criteria: TB symptoms, abnormal CXR, close contact with TB or evidence of Mtb infection (i.e., positive TST), or a positive response to TB treatment); or if they lacked both criteria required for a diagnosis of unconfirmed TB, they were defined as unlikely to have TB. Children diagnosed with TB and those with pending TB had a high probability of having symptoms (72.7% vs. 85.5%) and chest imaging findings (90.9% vs. 80.6%) consistent with TB, and most children showed positive results in both TB criteria (63.6% vs. 81.1%). Figure 6 Despite the fact that, based on the anti-TB treatment (ATT) response or TB-related death determined by the expert review panel, one subgroup of children was reclassified from unlikely to have TB to unconfirmed TB. Figure 6 Subgroup A). ​​Children who were unlikely to have TB frequently presented with symptoms or chest imaging findings consistent with TB (65.7% vs. 31.8%), but the subgroup of children with positive results for both criteria were classified as unlikely to have TB because their TB-related symptoms improved after starting ART without ATT (Automatic Therapy). Figure 6 Subgroup B).

[0138] The NEI analysis, performed by operators unaware of clinical information and using a positive signal threshold previously established in the NLH discovery population, detected confirmed TB and pending TB with diagnostic sensitivities of 90.9% and 72.5%, respectively (Table 2). The NEI results showed similar sensitivity for pending TB cases with or without a clinical TB diagnosis. Figure 6The study detected all but one confirmed TB case (not shown). NEI results also detected the majority (52.7%) of children unlikely to have TB who met at least one criterion required for a confirmed TB diagnosis (Table 2). Due to the inapplicability of age-matched low-risk groups, the specificity of the NEI diagnosis was assessed in a subgroup of children unlikely to have TB who did not have any clinical outcomes meeting the NIH case definition thresholds. The median Mtb EV signal in this group was significantly lower than in all other groups. Successful ATT responses were higher in children with confirmed TB than in children with confirmed TB (87.2% vs. 66.7%; Table 1). Mortality rates were higher in children with confirmed TB than in children with confirmed TB after ATT initiation (33.3% vs. 8.5%; Table 2), but higher (45.5%) mortality rates were observed in children with undiagnosed and untreated confirmed TB during the study period (31.9% 22 / 69) compared to children unlikely to have TB (18.2%), although deaths in children unlikely to have TB often lacked microbiological or CXR results. Figure 6 Subgroup B). Urine LAM results showed poor diagnostic sensitivity for confirmed TB (42.8%; 3 of 7 cases), valid results for pending TB (5.6%; 3 of 53 cases), and moderate specificity for unlikely TB cases (88.7%; 47 of 53 cases), including children who did not meet the criteria for TB (88.3%; 10 of 12 cases).

[0139] 7. Correlation between serum Mtb EV signal and TB reclassification and treatment

[0140] Most children with suspected TB were diagnosed using symptom or chest imaging data collected at admission, but based on subsequent outcomes (including deaths deemed TB-related by an expert panel or a positive response to ATT) (not shown), one subgroup of children was reclassified from unlikely to have TB to suspected TB (17.9%; 12 / 67). The reclassification decision was made without knowledge of EV outcomes, and most children reclassified from unlikely to have TB to suspected TB (83.3%; 10 / 12) had detectable Mtb EV NEI signal at baseline, and these values ​​remained positive in children whose serum was available after TB reclassification (not shown). Conversely, most cases of suspected TB reclassified as unlikely to have TB (without ATT initiation due to symptom improvement) had a positive Mtb EV NEI signal at baseline (75%; 9 / 12), which significantly decreased at reclassification (88.9%; 8 / 9; data not shown). In summary, these results suggest that a longitudinal decrease in serum NEI signaling, reflecting reduced MtbEV, is associated with resolution of TB disease with an ATT response and may also be associated with clearance or suppression of immune reconstitution.

[0141] 8. Design and validation of a portable DFM device for measuring Mtb EV signal readings

[0142] To adapt this measurement platform to the resource-constrained environments prevalent in TB measurement, we used a 3D printer to manufacture an inexpensive and portable smartphone-based DFM device that can scan a 144-well measurement slide in 5 minutes. Figure 7A The device employs an aluminum sliding support to increase stability and maintain focus during automated scanning, and an illumination mask to block stray light and improve DFM signal strength. Figure 7B A smartphone application developed for the device allows manual centering and focusing of the first slide well, after which the application automatically centers, focuses, and captures images of the remaining wells, saving all images to separate files for download and analysis. When used to read NEI data from the discovery population, the device produced results similar to the baseline DFM results, accurately identifying 73% (11 / 15; vs. 80%; 12 / 15) of TB cases and 87% (13 / 15; vs. 93%; 14 / 15) of non-TB cases. Normalized NEI signal intensities for both composite and individual biomarker signals were also similar on both devices (data not shown).

[0143] A novel NEI method can be used to detect Mtb EV biomarkers in body fluids, enabling rapid diagnosis of TB by improving upon standard NEI methods, thereby achieving automated image capture and ultrasensitive detection of the target signal. The method and system of this invention can be further applied to other bacterial pathogens that can be found in EVs. This method employs a workflow suitable for clinical settings to detect TB in children at high risk of TB, where detection using respiratory samples often misses TB. This assay captures EVs directly from serum or other body fluids, eliminating the common requirement for EV immunoassays on purified EV samples, which are typically isolated using methods that trade off time, labor, cost, and EV yield, purity, and integrity, limiting their clinical feasibility. Since there are currently no widely accepted EV biomarkers for TB, this study evaluated the potential diagnostic value of measuring serum levels of EVs expressing two Mtb-derived factors, LAM and LprG, which are associated with TB virulence. The study found that EV expression of these biomarkers is enhanced in TB cases, and multiplex detection of these two factors can distinguish between LTBI and TB in an NHP model. Both LAM and LprG are highly expressed in Mycobacterium tuberculosis, which may contribute to their diagnostic performance. However, other abundant Mtb-derived factors identified as potential targets through literature search did not demonstrate similar ability to differentiate between TB and non-TB cases, including the membrane protein Ag85b, a mycoyltransferase required for efficient biosynthesis of the Mtb cell wall. In this analysis, only the Mtb membrane protein LpqH showed a significant difference between the two groups, although it remains unclear whether the secreted Mtb proteins (CFP10, ESAT6, MPT51, and MPT64) analyzed in this study form stable membrane interactions, thus allowing for their detection on serum EVs.

[0144] NEI analysis of NHP models of LTBI and TB showed that serum EV expression of LAM and LprG was necessary to differentiate TB from LTBI. LAM is a virulence factor expressed on the outer cell wall of Mtb, where it can bind to macrophage mannose receptors, promoting Mycobacterium tuberculosis entry into host phagocytes and inhibiting phagosome-lysosome fusion, thus regulating the immune response to promote the sustained intracellular survival of Mtb required for TB development. LprG plays an important role in the localization of LAM to the outer cell membrane of Mycobacterium tuberculosis. LprG-deficient Mtb mutants exhibited reduced LAM surface expression and virulence, decreased Mtb entry into host macrophages, reduced Mtb cell membrane biogenesis and / or integrity, inability to inhibit phagosome-lysosome fusion, and decreased intracellular replication rate. Therefore, LprG is crucial for LAM activity, as LprG deficiency attenuates Mtb virulence without altering LAM expression. Therefore, LprG expression might be downregulated in NHPs with LTBI; however, our results suggest the opposite: LAM was downregulated and LprG expression was upregulated in serum EVs from NHPs diagnosed with LTBI, while both markers were elevated in serum EVs from NHPs diagnosed with TB. Several mechanisms could explain this finding, including LprG downregulation of LAM expression or inhibition of LPRG activity to limit LAM transport, both of which are expected to limit LAM expression on the outer membrane of Mtb and potentially limit LAM expression in EVs secreted by Mtb-infected cells during LTBI. However, the mechanisms leading to this differential expression and its functional significance remain unclear and warrant further investigation, including replication in human studies with well-defined LTBI and TB populations.

[0145] NEI analysis of Mtb EVs showed good diagnostic sensitivity in a diagnostically challenging population of HIV-infected children (including children who, despite extensive TB testing, were not clinically diagnosed during the maternal study period), for both confirmed and unconfirmed TB. The diagnostic sensitivity of NEI was particularly important in this group because the mortality rate of untreated children not diagnosed with TB during clinical assessment was more than 5 times higher than that of children diagnosed and treated at assessment, consistent with the reported high mortality rate of children who did not receive ATT due to missed diagnoses. Notably, several children showed positive Mtb EV signals in their serology prior to their clinically-based TB diagnosis, suggesting the potential of serum Mtb EV signaling as an early diagnostic tool for TB. This is especially important in this population, as one-third of children identified in this way died at or shortly after their diagnosis using conventional methods.

[0146] In patients with confirmed and suspected TB, NEI Mtb EV signaling was significantly reduced after ATT initiation, consistent with the improvement in TB symptoms, suggesting that Mtb EV levels can serve as a surrogate for ATT response. A similar reduction was observed in a pediatric subgroup of children who met the criteria for suspected TB at baseline data assessment but were reclassified as unlikely to have TB due to improvement in their TB-suggestive symptoms after ART initiation without ATT. The reduced NEI Mtb EV signaling observed in all these cases suggests that children may have neonatal TB, and the improved immune function after ART initiation at least partially suppresses this possibility.

[0147] As described above, by first extracting EVs from bodily fluid samples and then using a NEI with anti-LAM and LprG antibodies, the sensitivity and specificity of the test can be significantly improved. This method can accurately detect the presence of Mtb in bodily fluid samples, and based on the detected expression of LAM and LprG, it can also determine whether the infection is latent or active. The combination of NEI and dark-field microscopy allows for highly sensitive Mtb detection without further enrichment of EV concentration in the sample. Automated NEI detection further improves detection efficiency while maintaining high accuracy.

[0148] The present invention uses the following methods and materials.

[0149] method

[0150] Mtb culture. Mtb strains H37Rv, CDC1551, and HN878 were obtained from Houston Methodist Hospital and cultured in rollers at 37°C in a protein-free basal medium (containing 0.1% (v / v) glycerol, 1 g / L KH2PO4, 2.5 g / L Na2HPO4, 0.5 g / L asparagine, 50 mg / L ferric ammonium citrate, 0.5 g / L MgSO4×7H2O, 0.5 mg / L CaCl2, and 0.1 mg / L ZnSO4, with or without 0.05% tyloxapine (v / v)) to mid-log (OD600 = 0.5–0.6).

[0151] THP-1 cell culture and differentiation. THP-1 mononuclear cells were purchased from the American Type Culture Collection (ATCC; Manassas, Virginia) and cultured at 37°C in RPMI 1640 supplemented with 10% FBS in a 5% CO2 incubator. Ten T175 culture flasks (each containing approximately 2.5 x 10⁸ cells) were incubated with nM PMA (Sigma Aldrich, 389, USA). 7 5-7 live THP-1 monocytes (10⁶ each) 5THP-1 macrophages were cultured for 24 hours (cells / mL; ≥95% viability) to differentiate into macrophages. After washing three times with PBS at 37°C to remove PMA and non-adhesive cells, the adhesive and differentiated THP-1 macrophages were cultured in RPMI 1640 supplemented with 10% FBS.

[0152] For experiments using Mtb-infected macrophages, Mtb H37Rv, CDC1551, and HN878 cultures in mid-log phase were clumped by centrifugation at 3000g for 10 minutes at 4°C. The resulting bacterial clumps were resuspended in 10 mL of penicillin- and streptomycin-free RPMI 1640 / 10% FBS, and the clumps were removed using a brief sonication step. The mixture was then passed through a syringe fitted with a 27-gauge needle (VWR, Norm-Ject, USA) 10 times. The Mtb suspension was then mixed with another 10 mL of antibiotic-free RPMI 1640 / 10% FBS, and 0.1 mL aliquots of the suspension were added to T175 culture flasks containing approximately 2.5 x 10⁻⁶ cells cultured in 20 mL of antibiotic-free RPMI 1640 / 10% FBS. 7 Differentiated THP-1 macrophages were used to obtain a multiplicity of infection (MOI) of 10. After 4 hours of culture, the cell culture was washed three times with PBS at 37°C to remove extracellular Mycobacterium tuberculosis and cultured in FBS-free RPMI 1640 for 48 hours. The culture supernatant was then passed through a 0.22 μm filter to remove Mycobacterium tuberculosis and generate samples for EV and soluble protein analysis. The culture filtrate was stored at -80°C, and aliquots were inoculated into mycobacterium growth indicator tubes. Mtb growth was assessed after 3–4 weeks of culture to confirm the absence of viable Mycobacterium tuberculosis in these samples. Cultured macrophages were recovered by trypsin digestion and aliquoted for viability analysis and cell lysate was used to generate Western blot analysis of target proteins.

[0153] For experiments using culture filtrate protein (CFP), an aliquot containing 100 μg CFP was added to a T175 culture flask containing approximately 2.5 x 10⁻⁶ CFP cultured in 20 mL RPMI 1640 / 10% FBS. 7 Differentiated THP-1 macrophages. After culturing for 4 hours, the cell culture was washed three times with PBS at 37°C to remove extracellular CFP, and then cultured for 48 hours in FBS-free RPMI 1640. The culture supernatant was then collected to generate samples for EV and soluble protein analysis. Cultured macrophages were recovered by trypsin digestion and aliquoted for viability analysis and to generate cell lysates for Western blot analysis of target proteins.

[0154] EV isolation and characterization. Frozen serum aliquots (1 mL) used for EV isolation were rapidly thawed in a 37°C water bath, vortexed for 3 s, and then 300 μL aliquots were transferred to 2 mL centrifuge tubes, mixed with 1.2 mL PBS, and centrifuged at 1200 rpm for 15 min at 4°C to precipitate large particles and debris. The supernatant was centrifuged at 2,000 g for 30 min at 4°C, passed through a 0.45 μm filter, and centrifuged twice at 10,000 rpm and 4°C for 45 min, followed by centrifugation at 110,000 g and 4°C for 90 min. The obtained EV aggregates were suspended in 1 mL PBS and centrifuged at 110,000 g and 4 °C for 3 h. They were then suspended in 50 μL PBS and characterized by dioctanedinic acid (BCA) assay and NanoSight nanoparticle tracking analysis (Malvern Panalytics, USA; 5 μg / mL, 5 replicates) to determine protein content, EV concentration, and size distribution. Cell culture supernatant used for EV separation was concentrated from 200 mL of starting material to 100 μL by centrifugation at 4 °C and 3000 rpm in 50 mL 10 kDa copolymer styrene ultrafiltration tubes (Millipore Sigma, USA) for 15 min. These concentrated samples were then used for EV separation and characterized as described above. For transmission electron microscopy analysis, a carbon-coated copper grid (400 mesh) was incubated for 30 seconds at room temperature with 5 μL of EV aliquots containing 1 mg / mL protein. Excess liquid was blotted with filter paper. The membrane was then stained with 2 μL of 1% osmium tetroxide at room temperature for 2 minutes. The solution was blotted to remove the stain. The membrane was then imaged using a Tecani F30 microscope (FEI Corporation, USA) at an accelerating voltage of 200 kV and the indicated magnification.

[0155] EV ELISA Analysis. 100 μL of human CD81-specific mouse antibody aliquots (BioLegend, USA) (5 μg antibody / mL dissolved in PBS) were added to each well of a 96-well microtiter plate, and the plates were incubated at room temperature for 16 hours to produce a standard EV capture plate for EV-ELISA. The wells were then washed three times with 260 μL PBST (30 s each time), followed by blocking with 250 μL blocking buffer (1% wgt / vol BSA dissolved in PBST) at 37°C for 2 hours, and then washed three more times with PBST. To analyze the optimal EV capture conditions for NEI, the wells were alternately incubated with 100 μL of specific mouse antibody aliquots against EV surface markers (including different clones of CD9, CD63, CD81, and anti-LAM / LprG antibodies).

[0156] For the EV ELISA, each well was incubated overnight at 4°C with 100 μL of separated EVs or serum, washed three times with PBST, and then incubated at 37°C for 1 hour with 100 μL of PBST containing 1 μg / mL of the indicated detection antibody, washed three times with PBST, followed by incubation at 37°C for 30 minutes with 100 μL of PBST containing 0.5 μg / mL HRP-labeled goat anti-mouse / human IgG (Jackson Immune Lab, USA), washed three times with PBST. After the final washing step, the wells were incubated for 10 minutes at room temperature with 50 μL of TMB (Sigma, USA), then the reaction was terminated by mixing with 50 μL of 2M H2SO4. The target EV in each well was then read using a microplate reader (Tecan, Ultra, Switzerland) after subtracting the OD 650 signal to measure the OD 450 value.

[0157] Mtb injection and sample collection in non-human primates (NHPs). The cryopreserved NHP plasma analyzed in this study was archived material obtained from NHPs infected with Mtb in 35 previously reported studies. Briefly, adult rhesus monkeys that were specific pathogen-free, retrovirus-free, and never infected with mycobacteria were divided into three experimental groups and received different Mtb exposures (none, low, and high). Samples from the negative control (never infected with Mtb) group were taken from four uninfected rhesus monkeys that were not exposed to Mtb during the study period. Samples from the latent TB infection (LTBI) group were taken from four rhesus monkeys that experienced a low-dose Mtb aerosol exposure event (approximately 10 CFU of Mtb CDC1551), were positive for tuberculin skin test (TST) within one month of exposure, but did not show any signs of TB, remaining asymptomatic LTBI-like infection throughout the study period (approximately 22 weeks). Samples from the TB population were obtained from five rhesus monkeys that had undergone a high-dose Mtb aerosol event (approximately 200 CFU of Mtb CDC1551), developing active TB disease characterized by weight loss, fever, elevated serum C-reactive protein levels, elevated chest X-ray scores consistent with TB, detectable Mtb CFU levels in bronchoalveolar lavage fluid, higher lung bacterial load, and associated lung pathology at the study endpoint. As previously described, lung tissue collected at the study endpoint was randomly sampled using a grid by a pathologist unaware of the animals' treatment.

[0158] NEI Analysis. 1 μL of an aliquot (5 μg / mL) of mouse antibody or indicator antibody specific for human CD81 EV capture was added to each well of a 144-well mask fixed to a microscope slide, and the slides were incubated at 4°C for 16 hours to produce EV capture slides for NEI analysis. All incubation steps in this analysis were performed in a humidified chamber to minimize evaporation. After capture antibody binding, the slides were washed three times with PBST and treated with 1 μL / well SuperBlock. TM Blocking buffer (PBS) (Cytiva, USA) was used to block the EVs at 4°C for 1 hour, followed by washing three times with PBST. The EVs were then incubated with 1 μL of serum or isolated EV samples. Serum samples were centrifuged at 10000g for 20 minutes to remove large fragments. Serum or EV samples were incubated on EV capture slides at 4°C for 16 hours, followed by washing three times with PBS. The slides were then hybridized at 37°C with 1 μL of a specific biotinylated detection antibody (1 μg / mL) for 1 hour, washed three times with PBS, and incubated at 37°C with neutral avidin-functionalized AuNR (Nanopartz, USA) at the specified concentration for 1 hour. After the AuNR incubation step, the slides were washed once with PBST and distilled water to remove unbound particles, and then the slides were photographed and analyzed using DFM.

[0159] DFM noise reduction algorithm. A custom-designed algorithm analyzes the NEI signal, identifying the first region to be analyzed for each region, then subtracting DFM artifacts and background signals. The algorithm detects the high-intensity boundary of each aperture, calculates the center position of the circular region formed by its high-intensity boundary, and then selects the center region of that region to avoid the potential "coffee ring effect" caused by the accumulation of residual AuNRs at the edges of these regions after the final washing step. To remove DFM artifacts and background, the image is converted to the HSB (hue, saturation, brightness) color space, and the values ​​of each channel in the 16-bit image are normalized to the range of 0 to 1, since hue is measured in degrees (0° to 360°). The hue and saturation channels are used to identify the AuNR signal, and a set of training images from slides coated with pure AuNRs and AuNRs mixed with human serum are used to remove artifacts. Hue is used to identify pixels matching the AuNR scattering range, and saturation is used to evaluate the color intensity (purity) matching the AuNR range. Pixels with hue channel values ​​outside the red scattering range of the AuNR signal (≥0.8 and ≤0.05 on the hue wheel) were excluded, as were pixels with extremely low saturation values ​​(≤0.05). Using these parameters, 476 AuNR signals in the processed image were measured using MATLAB (software version 2020a).

[0160] Clinical population assessment

[0161] The NLH cohort. The sample and relevant clinical data were collected from 20 children at the National Lung Hospital (NLH) in Hanoi, Vietnam, who visited the NLH's pediatric outpatient clinic consecutively for clinical evaluation and medical assessment. The NLH cohort included children ≤17 years of age who were seen by clinicians at the NLH pediatric outpatient clinic with written informed consent from their parents or legal guardians. Children ≤17 years of age with laboratory-documented anemia (hemoglobin <9 mg / dL) or who did not obtain informed consent for all study procedures were excluded. Children currently living with HIV or receiving antiretroviral therapy (ART) were not excluded. The control group consisted of children with bronchopulmonary disease (BPD) but not TB, whose QuantiFERON TB Gold Plus (QFT; Hilden, Germany), Xpert GeneXpert MTB / RIF (Xpert; Cepheid, USA) and Mtb culture results were negative, and who received clinical assessments for TB exclusion from experienced pediatric TB specialists. TB cases are identified by positive Mtb cultures and / or GeneXpert MTB / RIF (Xpert; Cepheid, USA) results from lung or extrapulmonary samples. The protocol was approved by the NLH Institutional Review Board.

[0162] The PUSH study (Pediatric Emergency and Post-Stability Highly Active Antiretroviral Therapy (HAART) Initiation Trial) was a randomized controlled trial (NCT02063880) conducted in Kenya to evaluate whether emergency care (<48 hours) versus post-stability antiretroviral therapy (7–14 days) improved survival in hospitalized HIV-infected children under 12 years of age. At admission, children underwent systematic screening for TB symptoms and exposure, and intensive TB assessments were performed, including chest X-rays, Xpert and culture of sputum or gastric aspirates (GA), lipofuscin (LAM) antigen detection in urine, and Xpert in stool, regardless of TB symptoms. Serum was collected at admission and at 2, 4, 12, and 24 weeks post-admission and cryopreserved. CXRs were read by radiologists from the South African Tuberculosis Vaccine Initiative using standardized reporting formats to determine results suggestive of TB. A positive tuberculin skin test (TST) result was defined as an induration >5 mm. The diagnosis results are available for clinicians to study and to decide whether to begin TB treatment based on Kenyan guidelines.

[0163] Based on the NIH international consensus definition of pediatric TB clinical cases, children were post-hocly classified as confirmed TB (positive Mtb culture or Xpert result from respiratory or fecal samples), pending TB (meeting >2 or more of the following criteria: TB symptoms, abnormal CXR consistent with TB, positive TST or known TB exposure, and / or TB treatment response), or unlikely to have TB (not meeting other criteria) (not shown). TB symptoms were defined as cough lasting more than 2 weeks, weight loss / stunted growth, fever lasting more than 1 week, and / or lethargy lasting more than 1 week. For children who died during the study, an expert panel reviewed the cases and reached a consensus on whether the death was considered likely, probable, or unlikely to be related to TB. Deaths considered likely or probable to be related to TB were used as additional criteria for the pending TB classification (Table S6). For assay assessment, children with pending TB were further stratified based on whether they received a clinical TB diagnosis and ATT, or whether they did not have a clinical TB diagnosis and did not start ATT. Children classified as unlikely to have TB were further graded based on the presence or absence of the features used for the pending TB diagnosis.

[0164] Children with available cryopreserved serum within 2 weeks of TB diagnosis (including 2 weeks of ATT initiation) were evaluated to assess the diagnostic performance of Mtb EV NEI results for TB. Children who had started TB treatment and had samples from within 2 weeks of TB treatment initiation and at least one additional follow-up sample were eligible for treatment response analysis.

[0165] Participant characteristics were summarized by frequency and proportion of categorical variables and median and interquartile range (IQR) of continuous variables (Table 1). Assuming that NIH pediatric TB clinical cases are defined as binomial distributions, we estimated the performance of NEI EV in detecting TB using 95% confidence intervals (CI) for sensitivity and specificity (Table 2). Notably, many children in the unlikely TB category had CXRs suggestive of TB or TB symptoms (but did not meet the criteria for suspected TB). Therefore, we further stratified the unlikely TB category based on the presence or absence of abnormal CXRs and TB symptoms to identify potentially more appropriate negative controls. Median Mtb NEI EV levels were assessed using the Wilcoxon rank-sum test and compared to reference values ​​for unlikely TB patients without TB symptoms and with negative CXRs. For treatment response, mean Mtb NEI EV levels at the start of TB treatment and in the most recently available samples were assessed using a paired Wilcoxon signed-rank test (median time between the start of TB treatment and the most recently available sample was 5.5 months (IQR 3.1–5.7)).

[0166] Twenty-four children did not undergo CXR studies (4 confirmed, 2 pending, 18 unlikely). Additional CXR results were extracted from hospital records to inform TB classification. This method identified hospital CXR data for 14 of the 24 children lacking CXR study data. If the hospital CXR results showed characteristics matching TB (SATVI criteria), we included this information to determine TB classification (3 children used this information to change from unlikely to have TB to pending TB). Additionally, XX children lacking CXR and with clinical pneumonia were presumed to have suggestive CXR results consistent with TB (i.e., if there were CXR results, there was likely CXR results suggesting TB). If the hospital CXR only showed "abnormal" without defining specific TB CXR characteristics, we considered this insufficient for TB classification.

[0167] Portable DFM device. The mobile DFM sliding scan system (270 x 190 x 106 mm) has a sliding scan range of 82 mm x 38 mm and includes a mechanical scanning system with two stepper motors, a dark-field light source, an interchangeable smartphone with a small objective lens, an IOIO-OTG board (DEV-13613, SparkFun Electronics, Colorado), and two motor drive boards (ROB-12779, SparkFun Electronics, Colorado) for controlling the stepper motors and communicating with the smartphone components via Bluetooth. The system uses two lithium batteries: a 7.6V 3500mAh lithium battery (3.35 x 1.97 x 0.55 inches, GAONENG, China) to power the electronics, and an 11.1V 3200mAh lithium battery (5.16 x 1.73 x 0.67 inches, HOOVO, China) to power the motors. Most of the additional parts were printed using a commercial 3D printer (Objet 30prime, Stratasys, Israel), but the slide holder used to ensure slide stability during scanning was made from aluminum sheet using a CNC milling machine. The dark-field light source consists of an integrated dark-field condenser lens containing an illumination numerical aperture (NA) ranging from 0.7 to 0.9 and an array of 3mm white LEDs with a 30-degree viewing angle. A mask placed in front of the condenser lens blocks stray light and improves the contrast of the dark-field image. The objective lens consists of three identical double-lens systems; it has a focal length of 3.4mm, an NA of 0.25, a working distance of 1mm, and a field of view of 1.6 x 1.6mm. The camera in the device's phone assembly is a Moto G6 with a focal length of 3.95mm, a working f / 1.8, and a 12-megapixel sensor resolution (3072 x 4096 pixels, 1.4μm pixel size).

[0168] Statistical analysis. Images and heatmaps were generated and statistical analyses were performed using GraphPad Prism (version 7.0). Latent differences between groups were analyzed using parametric or nonparametric (Kruskal-Wallis) one-way ANOVA, as well as multiple comparison tests, Wilcoxon signed-rank tests, Student's t-tests, or Mann-Whitney U tests (as applicable). Logistic regression was performed to assess the combination of LAM and LprG signals, which best distinguishes positive and negative controls to establish a diagnostic equation, and hypothesis testing was performed with a two-sided α0.05 significance level. Both features were input into this logistic regression model without other covariates. All analyses used... OnDemand version 9.4 Academic Edition was used (https: / / welcome.oda.sas.com / login). Unless otherwise stated, data are expressed as mean ± SD.

[0169] For all purposes, the following references are incorporated herein by reference in their entirety.

[0170] Zumla, Alimuddin et al., “World Health Organization Global Tuberculosis Report 2013: Successes, Threats and Opportunities”, The Lancet 382.9907(2013):1765-1767.

[0171] Zumla, Alimuddin et al., “World Health Organization Global Tuberculosis Report 2014 – Further Developments”, The Lancet Global Health 3.1 (2015):e10-e12.

[0172] Bass Jr, JB et al., “Diagnostic criteria and classification of tuberculosis”, American Journal of Respiratory and Critical Care Medicine 142.3(1990):725-735.

[0173] Boehme, Catharina C. et al., “Rapid molecular detection of tuberculosis and rifampicin resistance”, New England Journal of Medicine 363.11(2010):1005-1015.

[0174] Yerlikaya, Seda et al., “Tuberculosis Biomarker Database: Key to New Tuberculosis Diagnosis”, International Journal of Infectious Diseases 56(2017):253-257.

[0175] Paris, Luisa et al., “Correlation between urinary lipid arabinomannan and disease severity in HIV-negative pulmonary tuberculosis patients”, Science Translational Medicine 9.420(2017):eaal2807.

[0176] Raviglione, Mario, and Giorgia Sulis, “Tuberculosis 2015: The burden, challenges, and strategies for control and elimination,” Infectious Disease Reports 8.2 (2016).

[0177] CMDenkinger, SVKik, DMCirillo, M.Casenghi, T.Shinnick, K.Weyer, C.Gilpin, CCBoehme, M.Schito, M.Kimerling, M.Pai, “Identifying the need for next-generation tuberculosis testing”, J. Infect. Dis. 211, pp. 29-38 (2015).

[0178] Raposo, And Willem Stoorvogel, “Extracellular vesicles: exosomes, microvesicles and friends”, J Cell Biol 200.4(2013):373-383.

[0179] Liang, Kai et al., “Quantitative analysis of tumor-derived extracellular vesicles in plasma microsamples for diagnostic and therapeutic monitoring using nanoplasma,” Nature Biomedical Engineering 1.4 (2017):0021.

[0180] Sun, Dali, and Tony Y. Hu. “Low-cost mobile phone dark-field microscopy for nanoparticle-based quantitative research”, Biosensors & Bioelectronics 99(2018):513-518.

[0181] Russell, David G, “Who put tuberculosis in pulmonary tuberculosis?”, Nature Reviews Microbiology 5.1 (2007):39.

[0182] Athman, Jaffre J. et al., “Bacterial membrane vesicles mediate the release of Mycobacterium tuberculosis lipoproteins and lipoproteins from infected macrophages”, Journal of Immunology 195.3(2015):1044-1053.

[0183]

[0184] N: Number of participants with results; n: 588 participants with positive results.

[0185] aAccording to the World Health Organization, the age-specific CD4% cutoff for severe immunosuppression, or the CD4 count in the absence of CD4% data (<12 months: <25% / <1500 cells / μl; 12–35 months: <20% / <750 cells / μl; >36 months: <15% / <350 cells / μl).

[0186] b Children aged 5 and under: WHZ < -2 or MUAC < 12.5cm

[0187] c Persistent cough (≥14 days), fever (≥7 days), stunted growth, or lethargy (≥7 days). Stunted growth = emaciation (WHZ < -2 or MUAC < 12.5) or underweight (WHZ < -2) on admission (pre-admission growth trajectory is not applicable).

[0188] d sputum or gastric juice

[0189] e Including two children with pending TB whose fecal Xpert results were inconclusive.

[0190] f This included one child with unconfirmed TB and two children unlikely to have TB, whose urine LAM results were invalid. At the time of the study, a color change corresponding to a manufacturer's reference card grade ≥1 was considered positive.

[0191] g Two children diagnosed with TB died before starting TB treatment.

[0192] h The study period was 6 months.

[0193]

[0194] N=147; 137 children underwent serological analysis at baseline; 7 cases were pending confirmation of TB diagnosis (6 at 2 weeks post-admission and 1 at 4 weeks post-admission); and 3 cases with baseline serological absence who were unlikely to have TB were identified by serological analysis at 2 weeks post-admission.

[0195] ATT: Anti-TB Treatment

[0196] aBased on the international consensus definition of pediatric TB clinical cases (Graham et al., 2015): Children with a TB diagnosis and ATT during the study period—those prospectively diagnosed by clinical staff during the study period and receiving ATT during the study period; Children without a TB diagnosis or ATT during the study period—those who did not receive a clinical diagnosis and ATT during the study period; NIH TB criteria—children who meet at least one of the two diagnostic criteria required for a TB diagnosis according to the NIH algorithm.

[0197] b The NIH standards described in Supplementary Table 4

[0198] c Comparison of Wilcoxon rank-sum test with the likelihood of having TB in asymptomatic individuals and CXR610

[0199] d At the time of TB diagnosis or <14 days after TB treatment

[0200] e Newly available samples. The median time between the start of TB treatment and the start of TB treatment was 5.5 months (IQR 3.1–5.7).

[0201] f Paired Wilcoxon Sign-Rank Test

[0202] The claims are as follows:

Claims

1. A system for detecting MTB-specific antigens in body fluid samples, characterized in that, The system includes: a) Dark-field microscope; b) Sample substrate; and c) Antibody-conjugated nanoparticles, wherein the antibody-conjugated nanoparticles are formed by conjugating nanoparticles with a first antibody that targets the MTB-specific antigen, the MTB-specific antigen being selected from the group consisting of: lipoarabinomannan (LAM) and LAM carrier protein LprG. The sample substrate is coated with a second antibody against the extracellular vesicle-specific (EV-specific) antigen CD81; when the MTB-specific antigen is lipoarabinomannan (LAM), the first antibody is A194-01.

2. The system as described in claim 1, characterized in that, The antibody-conjugated nanoparticles are gold, silver (Ag), bismuth oxide (Bi2O3), platinum (Pt), gadolinium oxide (Gd2O3), or iron oxide (Fe3O4) nanoparticles.

3. The system as described in claim 1, characterized in that, The MTB-specific antigen is lipoarabinomannan (LAM) and LAM carrier protein LprG.

4. The system as described in claim 1, characterized in that, The conjugation between the nanoparticles and the first antibody is a biotin-avidin or biotin-streptavidin conjugation.

5. The system as described in claim 1, characterized in that, The nanoparticles are gold nanoparticles.

6. The system as described in claim 1, characterized in that, The bodily fluid samples were obtained from the patient.

7. The system as described in claim 1, characterized in that, The bodily fluids are blood, serum, sputum, urine, or other available bodily fluids.

8. The use of a primary antibody against Mycobacterium tuberculosis primary (MTB) specific antigen and a secondary antibody against Mycobacterium tuberculosis secondary (MTB) specific antigen in the preparation of a detection reagent, characterized in that, The detection reagent is used to detect and determine the tuberculosis infection status by detecting the presence of first and second Mycobacterium tuberculosis (MTB) specific antigens in extracellular vesicles (EVs) in a body fluid sample, wherein the first MTB specific antigen is lipoarabinomannan (LAM) and the second MTB specific antigen is lipoarabinomannan carrier protein (LprG); wherein extracellular vesicles (EVs) in the body fluid sample are extracted using a capture antibody, anti-CD81 antibody.

9. The use as described in claim 8, characterized in that, The method includes the following steps: a) Use the capture antibody anti-CD81 antibody to extract extracellular vesicles (EVs) from body fluid samples; b) Mix the antibody-conjugated nanoparticles with the EVs extracted in step (a), wherein the antibody-conjugated nanoparticles are formed by conjugating nanoparticles with a first antibody against a first MTB-specific antigen and a second antibody against a second MTB-specific antigen. c) Detect the presence of the first and / or second MTB-specific antigens using dark-field microscopy; and d) Determine the tuberculosis infection status based on the presence of the first and second MTB-specific antigens.

10. The use as described in claim 8, characterized in that, If both the first and second MTB-specific antigens are present, active TB infection is determined.

11. The use as described in claim 8, characterized in that, The first antibody is A194-01.

12. The use as described in claim 8, characterized in that, The bodily fluid samples were obtained from the patient.

13. The use as described in claim 8, characterized in that, The bodily fluids are blood, serum, sputum, urine, or other available bodily fluids.

Images