Development of lateral flow assays to detect host proteins in cattle for improved diagnosis of bovine tuberculosis

Abstract

Bovine tuberculosis (bTB), caused by Mycobacterium bovis (M. bovis) infection in cattle, is an economically devastating chronic disease for livestock worldwide. Efficient disease control measures rely on early and accurate diagnosis using the tuberculin skin test (TST) and interferon-gamma release assays (IGRAs), followed by culling of positive animals. Compromised performance of TST and IGRA, due to BCG vaccination or co-infections with non-tuberculous mycobacteria (NTM), urges improved diagnostics. Lateral flow assays (LFAs) utilizing luminescent upconverting reporter particles (UCP) for quantitative measurement of host biomarkers present an accurate but less equipment- and labor-demanding diagnostic test platform. UCP-LFAs have proven applications for human infectious diseases. Here, we report the development of UCP-LFAs for the detection of six bovine proteins (IFN-γ, IL-2, IL-6, CCL4, CXCL9, and CXCL10), which have been described by ELISA as potential biomarkers to discriminate M. bovis infected from naïve and BCG-vaccinated cattle. We show that, in line with the ELISA data, the combined PPDb-induced levels of IFN-γ, IL-2, IL-6, CCL4, and CXCL9 determined by UCP-LFAs can discriminate M. bovis challenged animals from naïve (AUC range: 0.87-1.00) and BCG-vaccinated animals (AUC range: 0.97-1.00) in this cohort. These initial findings can be used to develop a robust and user-friendly multi-biomarker test (MBT) for bTB diagnosis.

Keywords: DIVA; UCP-LFA; biomarkers; bovine tuberculosis; chemokines; cytokines; diagnostics; upconverting reporter particles.

Figure 1

Schematic diagram of lateral flow (LF) strip design (A) and workflow of UCP-LFAs performed in the study (B). UCP particles are first conjugated to target-specific antibodies to form the UCP conjugate. LF strips are made by stripping a test (T) line, composed of a second antibody specific for the target analyte and an isotype flow control (FC) antibody on a nitrocellulose membrane. The sample pad (the end added to the sample) and the absorbent pads are assembled to the nitrocellulose membrane on a plastic backing, and individual strips are cut and stored at room temperature. UCP-LFAs are performed by mixing a specific UCP conjugate with a sample diluted in LF buffer, followed by adding target-specific LF strips. Once immunochromatography is complete and the strips are dry, a dedicated infrared reader is used for scanning, and the signal intensity (expressed in Relative Fluorescence Units, RFUs) is calculated for both T and FC lines. Ratio (R) corresponds to the quantity of target analyte and is obtained by dividing signal intensities at T by the FC lines.

Figure 2

Serial dilutions in triplicate for IFN-γ, IL-2, IL-6, CCL4, CXCL9, and CXCL10. Respective recombinant bovine cytokines were spiked in assay buffer and analyzed using UCP-LFAs. Shown are the ratios (R) obtained by dividing the signal at the respective test (T) lines by the signal at the flow control (FC) lines for each recombinant bovine protein. Mean values with error bars (±1 SD) are shown.

Figure 3

IFN-γ, IL-2, IL-6, CCL4, CXCL9, and CXCL10 levels in PPDb-stimulated whole blood supernatants were measured by UCP-LFAs. The ratios among groups were compared using the Kruskal–Wallis test with Dunn's multiple comparison post-test. N: Naive Animals (n = 16, empty circles); V: BCG-vaccinated animals (n = 10, green circles); V/C: BCG-vaccinated and Mycobacterium bovis challenged (n = 12, blue circles); C: M. bovis challenged only (n = 9, red circles). The bars of scatter dot plots show mean values, and error bars show ±1 SD. ^*p < 0.05; ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001.

Figure 4

ROC curve analysis of IFN-γ (red), IL-2 (green), IL-6 (pink), CCL4 (yellow), CXCL9 (black), and CXCL10 (blue) detected in PPDb-stimulated whole blood supernatants to discriminate naïve (N) from Mycobacterium bovis-challenged group (C) (A), BCG-vaccinated (V) from M. bovis-challenged group (C) (B), and BCG-vaccinated group (V) from those that were M. bovis-challenged post-BCG (V/C) (C).

Figure 5

A NUM score was calculated based on IFN-γ, IL-2, IL-6, CCL4, CXCL9, and CXCL10 ratios in PPDb-stimulated whole blood supernatants for naïve (N; n = 16), BCG-vaccinated (V; n = 10), and Mycobacterium bovis-challenged animals with (V/C; n = 12) or without prior BCG vaccination (C; n = 9). The NUM score (y-axis) combines the results of six proteins, indicating the number of proteins with levels above a threshold based on the maximal Youden's index for each marker [Table 1(a)]. Group differences were determined using the Kruskal-Wallis test; the statistical significance level used was ^***p < 0.001; ^****p < 0.0001 (A). Heatmap showing Pearson correlation among the ratios of the evaluated host proteins. The color corresponds to the Pearson R value as indicated in each square (B).