Sensitive and Specific Detection of Low-Level Antibody Responses in Mild Middle East Respiratory Syndrome Coronavirus Infections

Abstract

Middle East respiratory syndrome coronavirus (MERS-CoV) infections in humans can cause asymptomatic to fatal lower respiratory lung disease. Despite posing a probable risk for virus transmission, asymptomatic to mild infections can go unnoticed; a lack of seroconversion among some PCR-confirmed cases has been reported. We found that a MERS-CoV spike S1 protein-based ELISA, routinely used in surveillance studies, showed low sensitivity in detecting infections among PCR-confirmed patients with mild clinical symptoms and cross-reactivity of human coronavirus OC43-positive serum samples. Using in-house S1 ELISA and protein microarray, we demonstrate that most PCR-confirmed MERS-CoV case-patients with mild infections seroconverted; nonetheless, some of these samples did not have detectable levels of virus-neutralizing antibodies. The use of a sensitive and specific serologic S1-based assay can be instrumental in the accurate estimation of MERS-CoV prevalence.

Keywords: ELISA; MERS; MERS-CoV; Middle East respiratory syndrome coronavirus; Qatar; S1; South Korea; antibodies; camels; coronavirus; diagnostics; human; neutralizing; serology; spike; the Netherlands; viruses.

Figure 1

Detection of MERS-CoV–specific antibody responses 6–12 months following PCR-diagnosed mild and severe infections using different assays. Spike S1–specific antibody responses were tested with a routinely used S1 ELISA (rELISA) (A), in-house S1 ELISA (iELISA) (B), and S1 microarray (C). Nucleocapsid-specific antibody responses were tested using a luciferase immunoprecipitation assay (D). Severe infections (red, n = 5; cohort H) resulted in antibody responses detected for up to 1 year by all assays, while detection of mild infections (green, n = 6; cohort G) varied among assays. Horizontal dotted line indicates cutoff for each assay; yellow shaded area indicates serum undetected by each assay. CoV, coronavirus; LU, luminescence units; MERS, Middle East respiratory syndrome; OD, optical density.

Figure 2

MERS-CoV–specific antibody responses detected by different assay platforms. A) In-house IgG of S1 ELISA (iELISA); B) MERS-CoV S1 protein microarray; C) HCoV S1 microarray reactivity of non-MERS-CoV–infected serum samples to the S1 proteins of 6 different HCoVs; D) nucleocapsid-luciferase immunoprecipitation assay; E) IgG S2 ELISA; F) routinely used IgG S1 ELISA expressed as the ratio of optical density of sample to kit calibrator; G) plaque reduction neutralization test (PRNT), expressed as endpoint titer for 90% plaque reduction. Serum samples tested were obtained from healthy blood donors (n = 50, cohort A); patients with PCR-diagnosed respiratory infections including human coronaviruses (n = 145, cohorts B and C); S1-microarray positive (n = 18, cohort D1) and negative (n = 19, cohort D2) camel contacts; and longitudinal serum samples from 2 PCR-confirmed MERS-CoV–infected patients taken 15–228 days after diagnosis (n = 7, cohort F). Cohort E is not included because patients in this cohort were in the acute phase of infection (<14 days postdiagnosis), in which seroconversion may not have occurred. Cohorts A, B, C, and F are from the Netherlands, cohort D from Qatar. Serum samples were tested at dilutions 1:101 for ELISA and N-LIPS, 1:20 for S1 microarray, and 1:20 to 1:2,560 for PRNT. Dotted lines indicate cutoff for each assay. CoV, coronavirus; LU, luminescence units; MERS, Middle East respiratory syndrome; OD, optical density; RFU, relative fluorescence units.

Figure 3

Low sensitivity of commercial S1 ELISA shown as the effect of lowering coating antigen concentration (A) or antigen denaturation (B) on the sensitivity of antibody detection among Middle East respiratory syndrome coronavirus–infected persons with camel contact. All samples were seropositive by in-house S1 ELISA and microarray. Dark blue indicates those that tested seropositive by commercial S1 ELISA.

Figure 4

Reactivity to Middle East respiratory syndrome coronavirus of serum samples from 2 patients with human coronavirus OC43 in different assays. Longitudinal serum samples, collected before and after OC43 infection, from the 2 patients (red, patient 1; black, patient 2) were analyzed by commercial IgG S1 ELISA (A); in-house IgG S1 ELISA (B); S1 protein microarray(C); and PRNT₉₀ (D). Dotted line indicates the cutoff for each assay. Error bars indicate 95% CIs. OD, optical density; PRNT₉₀, 90% reduction in plaque reduction neutralization test; RFU, relative fluorescence units.

Figure 5

Correlation between neutralizing and S1 antibody responses and comparison of different S1 platforms. A) PRNT₉₀ neutralization titers and IgG titers obtained by in-house S1 ELISA among PCR-confirmed MERS-CoV patients and persons with camel contact. Spearman correlation r value and 2-tailed p-value are shown. Yellow shading indicates S1-reactive nonneutralizing antibodies. B) Receiver operator characteristic (ROC) curves comparing the specificity and sensitivity of different MERS-CoV S1–based platforms for the diagnosis of MERS-CoV infections among PCR-confirmed cases. AUC for iELISA (blue) is 1; for S1 microarray (red) is 0.9893; for rELISA (green) is 0.9481. Dotted lines show the cutoff for each assay. AUC, area under the curve; dpd, days postdiagnosis; PRNT₉₀, 90% reduction in plaque reduction neutralization test.