Multiplex sample-sparing assay for detecting type-specific antibodies to Zika and dengue viruses: an assay development and validation study

Abstract

Background: Serology for dengue viruses (DENV) and Zika virus (ZIKV) has been hindered by antibody cross-reactivity, which limits the utility of these tests for surveillance and assessment of sero-status. Our aim was to develop a multiplexed IgG-based assay with increased accuracy to assess the history of previous DENV and ZIKV infections.

Methods: We developed and assessed the analytical performance of a sample-sparing, multiplexed, microsphere-based serological assay using domain III of the envelope protein (EDIII) of DENV serotypes 1-4 and ZIKV, the most variable region between each virus. We used a reference panel of well-characterised serum samples from US-based travellers or residents of southeast Asia, central America, or Puerto Rico, who were naive or immune to either or both DENV and ZIKV, to develop an algorithm for detecting previous exposure to DENV and ZIKV and identify optimal positivity cutoffs to maximise assay performance. To independently confirm the performance of the assay and algorithm, we used a second test set of previously collected samples from healthy children (aged 9-16 years) living in Puerto Rico, whose DENV and ZIKV serostatus had been defined using the gold-standard virus neutralisation assay. We evaluated the performance of the multiplex assay compared with the gold-standard assay by estimating sensitivity and specificity for identification of past exposure to ZIKV and DENV.

Findings: The multiplexed EDIII assay showed reproducible results over different days and a linearity range from μg to pg levels for various EDIII antigens. Using a reference panel of serum samples from individuals who were DENV naive (n=136), DENV immune (n=38), ZIKV naive (n=67), and ZIKV immune (n=28), we optimised the assay and developed a testing algorithm that was 94·9% (95% CI 83·1-99·1) sensitive and 97·1% (92·7-98·9) specific for identifying previous exposure to DENV, and 100% (95% CI 88·0-100) sensitive and 97·0% (89·8-99·5) specific for identifying previous exposure to ZIKV. In an analysis with an independent test set of 389 samples, the assay and algorithm had 94·2% (89·9-97·1) sensitivity and 92·9% (87·3-96·5) specificity for DENV, and 94·1% (88·7-97·4) sensitivity and 95·0% (90·0-98·0) specificity for ZIKV.

Interpretation: The multiplexed EDIII serology assay can accurately identify the history of previous infection with either DENV or ZIKV. This high-throughput and sample-sparing assay is a promising new tool for supporting flavivirus surveillance, epidemiological and clinical studies, and serological testing for dengue vaccine eligibility. Further studies are needed to reduce the cost of the assay, eliminate high background in some samples, and to assess performance in DENV-endemic and ZIKV-endemic countries.

Funding: US National Institutes of Health.