Quantitation of Circulating Mycobacterium tuberculosis Antigens by Nanopore Biosensing in Children Evaluated for Pulmonary Tuberculosis in South Africa

Abstract

Nanopore sensing of proteomic biomarkers lacks accuracy due to the ultralow abundance of targets, a wide variety of interferents in clinical samples, and the mismatch between pore and analyte sizes. By converting antigens to DNA probes via click chemistry and quantifying their characteristic signals, we show a nanopore assay with several amplification mechanisms to achieve an attomolar level limit of detection that enables quantitation of the circulating Mycobacterium tuberculosis (Mtb) antigen ESAT-6/CFP-10 complex in human serum. The assay's nonsputum-based feature and low-volume sample requirements make it particularly well-suited for detecting pediatric tuberculosis (TB) disease, where establishing an accurate diagnosis is greatly complicated by the paucibacillary nature of respiratory secretions, nonspecific symptoms, and challenges with sample collection. In the clinical assessment, the assay was applied to analyze ESAT-6/CFP-10 levels in serum samples collected during baseline investigation for TB in 75 children, aged 0-12 years, enrolled in a diagnostic study conducted in Cape Town, South Africa. This nanopore assay showed superior sensitivity in children with confirmed TB (94.4%) compared to clinical "gold standard" diagnostic technologies (Xpert MTB/RIF 44.4% and Mtb culture 72.2%) and filled the diagnostic gap for children with unconfirmed TB, where these traditional technologies fell short. We envision that, in combination with automated sample processing and portable nanopore devices, this methodology will offer a powerful tool to support the diagnosis of pulmonary TB in children.

Keywords: CFP-10; ESAT-6; blood test; nanopore; pediatric tuberculosis.

Figure 1.

Schematic illustration of the nanopore assay for ESAT-6/CFP-10 antigen complex quantification. a: Capture antibody-modified dynabeads immunoprecipitate ESAT-6/CFP-10 antigen complex from serum samples, followed by specific binding of detection antibody-modified CuO nanoparticles to form a sandwich structure. Sandwich structures were then magnetically separated and Cu⁺ was released by hydrochloric acid to catalyze the click reaction between DNA-alkyne (DNA-A) and azido adamantane (AA) to form DNA-AA with amplification. Finally, DNA-AA@CB[6] probes were formed via host-guest interaction with CB[6]. b: Translocation recordings of DNA-A through a single α-HL nanopore and the corresponding nanopore steric hindrance status. c: Translocation recordings of DNA-AA through a single α-HL nanopore and the corresponding nanopore steric hindrance status. d: Translocation recordings of DNA-AA@CB[6] through a single α-HL nanopore. Three blockade levels of the characteristic signal indicate the molecular status inside the nanopore: DNA translocation, CB[6] dissociation, and CB[6] oscillation, respectively. Data was acquired using 3 M KCl, 10 mM Tris buffer at pH 8.0 and under a 160 mV trans potential unless otherwise stated.

Figure 2.

Characterization of DNA complexes and their translocation signals. a: Chemical structure of DNA-A, DNA-AA, and DNA-AA@CB[6]. DNA (5′-CCCCCCCCCCTCCCCCCCCCC) modified with an alkyne moiety on the T base formed DNA-A; a click reaction between the alkyne moiety and azide adamantane formed DNA-AA; non-covalent host-guest interaction between DNA-AA and CB[6] formed a DNA-AA@CB[6] probe. b: Electrospray ionization mass spectrograms of reactant DNA-A and product DNA-AA of the click reaction. c: ¹H NMR (300 MHz) spectra of reactant CB[6] and achieved DNA-AA@CB[6] probes. The host-guest interaction can be confirmed by the chemical shift of protons of the CB[6] molecule from methylene (CH₂) groups (5.66 ppm and 5.44 ppm) and tertiary C-H groups (5.61 ppm). d-f: Blockade vs. dwell time translocation profiles of (d) DNA-A, (e) DNA-AA, and (f) DNA-AA@CB[6]. Top: histograms of relative frequency of I/I₀ with Gaussian fitting. Right: histograms of relative frequency of dwell time with exponential fitting. g: A typical characteristic oscillation signal generated by the DNA-AA@CB[6] probe translocating an α-HL nanopore. The signal is characterized by a pattern with 3 levels. Average currents of Level 1, 2, and 3 in oscillation signals were analyzed by one-way ANOVA. **** indicates P<0.001 Level 1 vs. Level 2, Level 2 vs. Level 3 by multiple comparisons, N=30). Data was acquired using 3 M KCl, 10 mM Tris buffer at pH 8.0 and under a 160 mV trans potential unless otherwise stated.

Figure 3.

Quantification of ESAT-6/CFP-10 in human serum. a: Raw translocation recordings of serially diluted ESAT-6/CFP-10 antigens (0 aM, 10 aM, 10 fM, and 10 pM) through single α-HL nanopores; triangles, squares, and circles indicate oscillation signals in different data sets. b: Capture rate of oscillation signals in different ESAT-6/CFP-10 antigen concentrations. The cumulative numbers of signals were counted every 20s. Solid lines indicate linear regression. c: Correlations between oscillation signal frequency and ESAT-6/CFP-10 antigen concentration in human serum within the range of 0.1–10⁶ fM. Inset shows the correlations within attomolar range (0, 10 and 100 aM). Data represents mean ± SD of three replicates. Solid line indicates linear regression. Shadow indicates limits of 95% confidence interval. Data was acquired using 3 M KCl, 10 mM Tris buffer at pH 8.0 and under a 160 mV trans potential unless otherwise specified.

Figure 4.

Clinical validation in a pilot pediatric cohort. Study participants numbered P1 to P75 were classified into three groups according to NIH definitions: confirmed TB (N=18), unconfirmed TB (N=36), and unlikely TB (N=21). a: Quantification of ESAT-6/CFP-10 antigen complex in different groups by the nanopore assay. Insets show cases with low concentration ESAT-6/CFP-10. The green line indicates LOD (0.01 fM); the yellow line indicates LOQ (0.10 fM); the red line indicates the optimal threshold (1.15 fM) from the receiver operating characteristic (ROC) curve. Data represents mean ± SD of three replicates. Data was acquired using 3 M KCl, 10 mM Tris buffer at pH 8.0 and under a 160 mV trans potential unless otherwise stated. b: ROC curve of the 75 clinical samples. Area under the ROC curve is 0.8757, P<0.0001, threshold is 1.15 fM, sensitivity is 94.44%, specificity is 80.95%. c: Cluster map of clinical test results and nanopore test results for all participants. Clinical tests include TST, smear, Xpert, Mtb culture. Nanopore results are called based on the threshold. Blue stars indicate cases confirmed at follow up; green stars indicate cases redesignated to unlikely TB; and black stars indicate cases redesignated to unconfirmed TB upon further review of treatment and follow up information. d: Sensitivity and specificity comparison among nanopore, Xpert, and Mtb culture tests, including updated sensitivities and specificities after reclassification (refer to Discussion).