Rapid antigen tests for dengue virus serotypes and Zika virus in patient serum

Abstract

The recent Zika virus (ZIKV) outbreak demonstrates that cost-effective clinical diagnostics are urgently needed to detect and distinguish viral infections to improve patient care. Unlike dengue virus (DENV), ZIKV infections during pregnancy correlate with severe birth defects, including microcephaly and neurological disorders. Because ZIKV and DENV are related flaviviruses, their homologous proteins and nucleic acids can cause cross-reactions and false-positive results in molecular, antigenic, and serologic diagnostics. We report the characterization of monoclonal antibody pairs that have been translated into rapid immunochromatography tests to specifically detect the viral nonstructural 1 (NS1) protein antigen and distinguish the four DENV serotypes (DENV1-4) and ZIKV without cross-reaction. To complement visual test analysis and remove user subjectivity in reading test results, we used image processing and data analysis for data capture and test result quantification. Using a 30-μl serum sample, the sensitivity and specificity values of the DENV1-4 tests and the pan-DENV test, which detects all four dengue serotypes, ranged from 0.76 to 1.00. Sensitivity/specificity for the ZIKV rapid test was 0.81/0.86, respectively, using a 150-μl serum input. Serum ZIKV NS1 protein concentrations were about 10-fold lower than corresponding DENV NS1 concentrations in infected patients; moreover, ZIKV NS1 protein was not detected in polymerase chain reaction-positive patient urine samples. Our rapid immunochromatography approach and reagents have immediate application in differential clinical diagnosis of acute ZIKV and DENV cases, and the platform can be applied toward developing rapid antigen diagnostics for emerging viruses.

Conflict of interest statement

Competing interests: R. Warke, HiMedia, has applied to license monoclonal antibody technology from the Massachusetts Institute of Technology. Relevant patent applications: I.B., K. H-S, J.G-M. and L.G. are inventors on patent US9488613 B2 filed by M.I.T. that covers devices and methods for multiplexed diagnostics; I.B., K.H-S, J.G-M, H.dP, and L.G. are inventors on MIT case number 18262 (pending), which covers anti-dengue virus NS1 protein monoclonal antibodies, and I.B., K.H-S, J.G-M, H.dP, and L.G. are inventors on MIT case number 18671 (pending), which covers anti-Zika virus NS1 protein monoclonal antibodies, and pair-wise detection of Zika virus NS1 protein.

Copyright © 2017 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works.

Figures

Fig. 1.. NS1 protein alignment and linear epitope mapping of the ten antibodies used to run the dengue virus serotype-specific NS1 rapid tests, the pan-dengue NS1 test, and the Zika virus NS1 test.

(A) Table listing mAb names, mAb immunochromatography applications, mAb linear epitope sequences and starting amino acid positions, and NS1 domain positions. (B) Comparison of amino acid similarity based on analysis of NS1 protein sequences from the following viruses: DENV1- Strain Singapore/S275/1990, accession number P33478; DENV2 -Strain NGC, accession number AAA42941; DENV3- Philippines/H87/1956, accession number AAA99437; DENV4- Singapore/8976/1995, accession number AAV31422; Zika virus, accession number KU497555.1. Amino acid sequences were compared using Color Align Conservation http://www.bioinformatics.org/sms2/color_align_cons.html to enhance the output of sequence alignment program. Residues that are identical among the sequences are boxed. Linear peptide epitopes (B) are italicized and indicated in color on the figure, with the key to the right of the figure.

Fig. 2. Rapid immunochromatography for specific detection of DENV NS1 proteins (serotypes 1–4) and ZIKV NS1 protein.

(A-B) Images of rapid test strips. Strip numbers refer to the DENV serotype NS1 (–4), pan-dengue (P; all four DENV serotype NS1 proteins), or ZIKV virus NS1 (Z) detected. Recombinant NS1 proteins, indicated with a lower case “r’ preceding the virus name, were prepared at 500 ng/ml, and the strips were run using 50 μl of solution. Strip #1 (detects DENV serotype 1): mAb pair 912/271; Strip #2 (detects DENV serotype 2): mAb pair 243/1; Strip #3 (detects DENV serotype 3): mAb pair 411/55; Strip #4 (detects DENV serotype 4): mAb pair 626/55; Strip P (“pan dengue”; detects all four DENV serotypes): mAbs 271–243-411–626/323; Strip Z (detects ZIKV): mAbs 130/110. The test proteins run on the strips are recombinant DENV NS1, serotypes 1–4 (rDENV1-rDENV4), as well as recombinant NS1 proteins from ZIKV (rZIKV), West Nile Virus (rWNV), Yellow Fever Virus (rYFV), Japanese encephalitis virus (rJEV), and Tick Borne encephalitis virus (rTBEV). C = control, NS1 = detection site for specific NS1 protein. (C-E) Limits of detection for viral NS1 proteins using the serotype-specific (SSp) DENV strips 1–4 (C), the pan-dengue strip (D), and the ZIKV virus strip (E). The limits of detection, representing three independent determinations, are recorded on the figures. Each point (C-E) is presented as the mean and standard deviation. (F) NS1-containing supernatants from Vero cells infected with DENV serotypes 1–4 (Vs DENV1–4) or ZIKV virus (Vs ZIKV-U (Uganda) or ZIKV-B (Brazil)) were chromatographed on strips 1–4, pan-dengue (P), or ZIKV (Z) NS1 strips. The arrows indicate the strips with positive NS1 signals. Horizontal test lines (panel F) result from applying antibodies to the nitrocellulose using a mechanical striper device; the circular dot signals result from applying antibodies to the nitrocellulose using a standard pipettor.

Fig. 3.. Applying the rapid test to analyze human patient sera.

(A) Map showing the endemic virus regions where the rapid tests were deployed to analyze patient serum samples. The areas of the circles correlate with the numbers of samples analyzed. The blue colors, faint to dark, represent DENV serotypes 1–4. ZIKV is indicated in orange color. (B) ELISA results showing the amounts of DENV (left) and ZIKV (right) NS1 found in patient serum and supernatants from infected cell cultures. Lanes 1 and 6 are supernatants from Vero cells infected with DENV; lanes 2 and 7 are supernatants from Vero cells infected with ZIKV. Lanes 3 and 8 are PCR-negative sera; lanes 4 and 9 are sera from PCR-positive DENV patients. Lanes 5 and 10 are sera from PCR-positive ZIKV patients. (C) Images of rapid test analysis of DENV NS1 serotypes 1–4 and ZIKV NS1 on serotype specific strips 1–4, as well as pan-dengue (P) and ZIKV (Z); the upward arrows mark positive tests, and θ is serum from an uninfected patient. (D-G) Quantification of rapid test results. Dipstick tests were run with PCR-confirmed DENV sera or ELISA-validated ZIKV serum (panel C), and the resulting signals were quantified and expressed as box plots. Statistical significance, based on one one-way ANOVA, is indicated as * P < 0.05, ** P < 0.01, *** P < 0.001. (H) Statistical significance, based on an unpaired T-test, is presented as * P < 0.05. Box and whiskers plots: the black ✕ represent the maximum and minimum measured normalized intensity values, while the small square box ☐ represents the mean value, and the larger box represents the 25–75% range of the data. Individual colored points represented individual patient samples measured. (I-J) Images of rapid tests showing that DENV and ZIKV NS1 tests do not cross-react. (I) Supernatants from Vero cells infected with DENV serotype 4 were chromatographed on DENV serotype strips 1–4, on the pan-dengue strip (P), and on the ZIKV NS1 strip (Z). (J) Supernatants from Vero cells infected with ZIKV virus were chromatographed on DENV serotype strips 1–4, on the pan-dengue strip (P), and on the ZIKV NS1 strip (Z). (K) Images of rapid tests showing ZIKV NS1 is detected in serum samples concentrated 5X, but ZIKV virus NS1 is not detected in concentrated urine. Three sets of paired serum and urine samples were concentrated 5X by filter centrifugation and chromatographed on the ZIKV dipsticks. S: serum; U: urine. (I-K) The red boxes and vertical black lines serve as fiducial markers for image recognition and processing. Upward arrows indicate positive tests, using the serum samples. (L-N) Quantification of NS1 protein in supernatants of Vero cells infected separately with three DENV4 patient isolates (L) or three ZIKV patient isolates (M), or five paired serum/urine patient samples (N). Fig. 3L-N: One-way ANOVA was used to calculate statistical significance of the dengue and Zika tests: p<0.05, p<0.01, and p<.0.001 are indicated as *, **, and ***, respectively. Box and whiskers plots: the black ✕ represent the maximum and minimum measured normalized intensity values, while the black ☐ represents the mean value, and the larger box represents the 25–75% range of the data. Individual colored points represented individual patient samples measured.

Fig. 4.. ROC analysis, sensitivity/specificity analysis, and 95% confidence intervals (95% conf. int.) of the dengue and Zika tests.

(A) ROC curve analysis of the patient sample data collected for dengue serotypes 1–4, Dengue Pan, and Zika virus. TPR: true positive rate (sensitivity); FPR: false positive rate (1-specificity). (B) Table listing numerical values of the sensitivity and specificity results.