Abstract
Background
Previously, hepatitis E in humans was believed to be caused exclusively by species A variants of the Orthohepevirus genus (HEV-A) of the family Hepeviridae. However, we have previously demonstrated that Orthohepevirus species C (HEV-C), also known as rat hepatitis E virus, also causes hepatitis in humans. Due to high sequence divergence between HEV-A and HEV-C, serological tests based on HEV-A are often insensitive for HEV-C diagnosis. Therefore, we developed an enzymatic immunoassay (EIA) system for differentiating HEV-A and HEV-C antibody signatures in patient sera.
Methods
HEV-A and HEV-C peptide homologs spanning the entire immunogenic E2s region of HEV ORF2 capsid protein were expressed in E. coli. These peptides, HEV-A4 p239 and HEV-C p241, form virus-like particles (figure 1). Both peptides were coated in separate 96-well plates. Sera obtained from patients with RT-PCR confirmed acute HEV-A infection (n = 54), HEV-C infection (n = 15), and uninfected HEV seronegative controls (n = 126) were tested in parallel in HEV-A4 p239 and HEV-C p241 EIAs for respective IgG antibodies. Sample optical densities (ODs) were divided by mean OD + 3SD of the seronegative controls to generate signal/noise (S/N) ratios. S/Ns marking positivity in either assay were determined by ROC analysis. An algorithm for assay interpretation was developed (table 1) and the performance of this algorithm was measured against the RT-PCR gold standard.
HEV-A4 p239 and HEV-C1 p241 virus like particles
Transmission electron microscopy images of the two peptides used in this study
Diagnostic testing algorithm interpretation
Interpretation of results of testing using parallel enzymatic immunoassays
Results
Using cutoffs determined by ROC analysis (figure 2), HEV-A4 p239 and HEV-C p241 EIAs detected species-specific antibody responses well (sensitivity: 92.6% and 80% respectively) and were specific (92.9% and 98.3% respectively). The DC was 100% congruent with HEV-C RT-PCR and 88.9% congruent with HEV-A RT-PCR in RT-PCR positive samples. Incorporating all three cutoffs into the algorithm, we derived a 3×3 confusion matrix of RT-PCR sample assignation vs EIA algorithm classification (table 2). The Cohen’s κ value was 0.883 indicating excellent inter-rater reliability.
ROC analysis for determining S/N cutoffs
Curve for A/A represents analysis for HEV-A4 p239 EIA. Curve for C/¬C represents analysis for HEV-C p241 EIA. Curve for r(C/A) represents analysis for the differentiating ratio.
3×3 confusion matrix comparing sample assignations by RT-PCR vs. EIA algorithm
Conclusion
A parallel EIA system accurately differentiated HEV-A and HEV-C serological signatures in acute patient sera. This method can now be applied to seroprevalence studies to determine seroprevalence of rat hepatitis E in human populations.
Disclosures
Siddharth Sridhar, FRCPath, Abbott (Other Financial or Material Support, Speaker’s honoraria)
Antigenic Characterization of ORF2 and ORF3 Proteins of Hepatitis E Virus (HEV)
