Establishment and clinical validation of an in-cell-ELISA-based assay for the rapid quantification of Rabies lyssavirus neutralizing antibodies

Abstract

Neutralizing antibodies (nAbs) prevent the entry of viruses into permissive cells. Since nAbs represent correlates of protection against the Rabies lyssavirus, the presence of sufficient nAbs indicates effective vaccination. Accordingly, Rabies lyssavirus-specific nAb titers need to be determined in routine diagnostics to identify individuals being at risk of Rabies lyssavirus infections due to insufficient immunity. The current gold standard for the quantification of Rabies lyssavirus-specific nAbs is the rapid fluorescent focus inhibition test (RFFIT). However, RFFITs are expensive and labor-intensive since multiple microplate wells must be evaluated one-by-one by trained personnel through microscopic inspection, which limits the number of samples that can be processed. To overcome this disadvantage, we established a novel assay for Rabies lyssavirus-specific nAbs relying on an in-cell-ELISA (icELISA)-based neutralization test (icNT). The icNT differs from the RFFIT in the readout phase, and can be automatically quantified in minutes using broadly available microplate readers. During the establishment, icNT parameters such as antibody concentrations, permeabilization procedures, blocking reagents, infectious doses, and the duration of infection were optimized. Afterwards, a dose-dependent detection of Rabies lyssavirus neutralization was demonstrated using the WHO Standard Rabies Immunoglobulin reference. A panel of 200 sera with known RFFIT titers revealed very good sensitivity and specificity of the icNT. Furthermore, the icNT showed very good intra- and inter-assay precision. By recognizing Rabies lyssavirus-specific antigens, the assay can be applied immediately to automatically quantify the concentration of Rabies lyssavirus nAbs in routine diagnostics or for various basic research questions such as screening for antiviral compounds.

Fig 1. The icNT results correlate with the RFFIT.

(A) Schematic overview of the newly established assay, icNT, and the currently used assay, RFFIT. Serum samples were incubated with Rabies lyssavirus. After 70 minutes, a suspension of BHK-21 cells was added. For the analysis by icNT at 48 hours post-infection (h p. i.), cells were permeabilized and infected cells were detected by a specific primary antibody. A peroxidase-labelled secondary goat anti-mouse antibody was added and the ELISA reaction was started by use of TMB substrate and stopped with 0.5 M HCl. The absorbance was quantified by use of a microplate multireader. For the analysis by RFFIT at 22 h p. i., infected cells were detected by a specific FITC-labelled primary antibody and evaluated microscopically. (B) Rabies lyssavirus was incubated with a negative control serum, a positive control serum or left untreated. A twofold dilution was done with the positive control. Neutralization was evaluated by icNT. 4-fold replicates of samples were determined. The boxes range from the 25^th to the 75^th percentile. The line within the box depicts the median value. The whiskers indicate minimum and maximum values. Significance was calculated by an unpaired two-tailed t test, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Linear regression for the linearity was analyzed by coefficient of determination, R² = 0.94. (C) icNT was performed with different doses of infection of Rabies lyssavirus (MOI 5, 1 and 0.2). (D)(E) RFFIT and icNT were performed in parallel. Same sera and same dilutions were used. Bars depict the mean values ± SD. Dots show the values of the individual measurements. Samples were analyzed in sixfold determinations.

Fig 2. The increased virus inoculum does not affect icNT results.

Two icNT protocols were performed in parallel either using the usual “high” dose of Rabies lyssavirus (see standard protocol) or a “RFFIT-like” low-dose infection (~100 TCID50 per well). The high-dose version was fixed, processed, and evaluated by icELISA at 48 h p. i., while the low dose icNT was analyzed at 72 h p. i. to allow replication of the non-neutralized residual virus. Samples were determined in duplicate.

Fig 3. The icNT shows very good intra- and interassay precision.

(A) HRIG, WHO-2 SRIG, negative control, twofold dilution of positive control, six negative, three intermediate, three positive, and 3 highly positive serum samples with known RFFIT titers were analyzed by icNT. 6-fold replicates of samples within one microplate were determined (intraassay precision). Bars depict the mean values ± SD. Dots show the values of the individual measurements. (B)-(G) Same samples as in (A) were determined on two different days (interassay precision). The abscissa and the ordinate each represent one determination. The dashed line represents a Coefficient of Variability (CV) of 0%. Dots show the values of the individual measurements. The determinations were grouped and visualized in separate graphs. Linear regression for the interassay precision was analyzed by the Pearson correlation coefficient, r = 0.98, p (two-tailed) = 0.0001. Samples were determined in duplicate.

Fig 4. The icNT shows very good sensitivity and specificity.

(A) 25 seronegative, 25 positive, and 25 highly positive exemplarily chosen serum samples with known RFFIT titers were analyzed by icNT at indicated serum dilutions. (B) Concordance of icNT versus RFFIT according to nAb titer, both calculated by the Reed-Muench method for n = 200 sera. Since our diagnostic RFFIT does not further stratify negative (<0.3 IU/ml) and highly positive (>4 IU/ml) samples, negative and highly positive sera were arbitrarily set to 0.1 and 5 IU/ml, respectively, in both assays. Please note: 80 negative and 29 highly positive sera are superimposed (indicated by bigger dots). The dashed line depicts the ideal linear regression line. The dotted line depicts the cut-off titer of 0.5 IU/ml and the intermediate range (0.3–0.7 IU/ml) is shaded in grey. Linear regression was analyzed by the Pearson correlation coefficient, r = 0.9505, p (two-tailed) < 0.0001. (C) icNT results of the sera shown in (B). Serum samples were stratified based on the RFFIT results to the indicated groups.