The H5-subtype HPAIVs, such as the H5N1, H5N2, H5N6, and H5N8 viruses, are of both veterinary and public health concern worldwide. Infection with these viruses leads to multiorgan disease and death in domestic birds [5, 6]. After bird-to-human transmission, the H5N1 viruses cause severe disease with a frequently fatal outcomes, and their pandemic potential is commonly recognized [7]. For these reasons, H5-subtype HPAIVs and LPAIVs with a propensity to mutate into HP viruses are under epidemiologic surveillance as NAIVs, which is accomplished using a variety of methods recommended by OIE [3]. Our initial work in the area of H5Nx AIV diagnostics provided 6 different antibody clones against the HA1 subunit of HA [9]. Each of these mAbs showed broad strain specificity against AIVs of the H5 subtype and did not cross-react with non-H5 subtype virus strains. The study presented here resulted in a prototype H5 EB-ELISA for detecting anti-H5 HA antibodies in chicken sera, which was based on the G-7-27-18 mAb selected from the newly generated antibody clones.
Consistent with the characteristics of the blocking mAb [9], all of the analyzed antisera against HAs of the H1–H4 and H6–H16 subtypes tested negative in the H5 EB-ELISA (Fig. 1, Table 2), which proves that the assay is truly H5-subtype specific. In addition, the anti-H5 HA negative, various-origin sera yielded a mean inhibition of 22.9 ± 7.8%, indicating their negativity in the developed ELISA (Fig. 1). However, 5 out of 209 samples negative against H5 HA caused inhibitions to be higher, though close to the cutoff value of 38.5% (Table 2). These false positive samples were responsible for the reduction of the Dsp value to 97.6% (Table 2). To refine the H5 EB-ELISA, a larger number of anti-H5 HA negative specimens need to be analyzed to correct the cutoff value and/or indicate the doubtful threshold. Alternatively, the cutoff value could be optimized by using the receiver operating characteristic (ROC) analysis on the larger set of data.
The H5 EB-ELISA enabled not only specific but also sensitive detection of H5 subtype-specific antibodies, especially when applied to examine sera from chickens immunized using the H5 HA protein. Experimental antisera raised in commercial chickens against rH5-E. coli yielded mean inhibitions ranging from 76.9 ± 8.7% to 89.8 ± 3.5% (Fig. 1). They were determined with 99.1% sensitivity (Table 2). For comparison, the same set of samples was analyzed in the commercial FluAC H5 test with sensitivity as low as 80.9% (Additional file 3: Table S2). Thus, the developed assay is particularly suitable for the differentiation of infected chickens from chickens vaccinated with anti-H5Nx virus vaccines based on recombinant H5 HA proteins. Most recently, the Volvac® B.E.S. T AI + ND vaccine containing the H5 HA protein of baculovirus-expression system origin has been positively evaluated by efficacy trials in commercial chickens [13]. In general, recombinant DNA technology has been widely explored to obtain influenza virus HA protein, as exemplified by rH5-E. coli [10, 12]. Thus, the contribution of such vaccines in anti-AIV vaccination programs can be expected to increase, similar to the utility of the H5 EB-ELISA with its current characteristics.
Our results confirmed that the H5 EB-ELISA performed well in screening the SPF chickens immunized with H5-subtype LPAIVs. Under the applied assay conditions, different batches of antisera against H5N1, H5N2, H5N3 and H5N2 LPAIVs caused inhibitions varying between 40.4 ± 2.6% and 77.8 ± 2.3% (Fig. 1) and were determined with 98.0% sensitivity (Table 2). Nevertheless, both characteristics specified for the reference antisera were lower than for the experimental antisera (Fig. 1, Table 2). In addition, the sensitivity of detecting the anti-H5 HA antibodies in the reference antisera using the H5 EB-ELISA was decreased in comparison to those achieved with the FluAC H5 test, as indicated by Dse values of 97.8% vs. 100% established for the same set of samples (Additional file 3: Table S1). Therefore, it would be advisable to further optimize the H5 EB-ELISA to increase the sensitivity of detecting H5 subtype-specific serum antibodies induced with H5Nx influenza viruses. Thus, the optimized assay would be more effective as a DIVA test in vaccination programs utilizing predominantly oil-emulsified, inactivated whole AIV vaccines [4]. It would probably work better when used for the diagnosis of infection with H5-subtype influenza viruses.
Thus far, only a few experimental bELISA and cELISA tests designated to detect anti-H5 HA antibodies have been developed and evaluated [14–20]. These assays have employed mAbs predominantly to the highly variable HA1 subunit of HA [14–16, 20], and as an exception against the relatively well-conserved HA2 subunit of the antigen [17]. The tests reported by Chen et al. [14] and Yang et al. [15] did not show cross-reactions with non-H5 subtype AIVs; however, their confirmed H5-subtype specificities were restricted to a single H5N2 virus strain [14] or three strains representing the H5N1 and H5N2 AIVs [15]. Prabakaran et al. [16] developed bELISA based on the mAb to the linear epitope in the HA1 subunit with 100% and 96.9% conservation rates among the H5N1 virus isolates from humans and avian sources, respectively, while 54.3% conservation rates were observed in the H5-subtype viruses with an NA subtype other than N1. Consequently, the assay provided the highly sensitive detection of antibodies against HAs of various-origin H5N1 viruses in chicken and human sera, but predictably, it would not detect the anti-HA antibodies in approximately 50% of antisera raised against H5N2-N9 viruses. In contrast, Postel et al. [17] set up the broadly reacting cELISA for anti-H5 HA antibody detection using mAb, which recognized a linear epitope conserved within the H5-subtype influenza viruses and located in the HA2 subunit. The assay demonstrated good diagnostic specificity and sensitivity in testing sera from different avian species but also substantial cross-reactivity with H2 subtype-specific sera and, to a lesser extent, with sera specific for H1 and H6 subtypes. Thus, this H5 cELISA did not allow for a clear distinction between antisera against HAs of the H5 and H2, H1 or H6 subtypes. In the H5-subtype cELISA described by Dlugolenski et al. [18], no intersubtype cross-reactions were found, but the accuracy of the test was low and varied between the avian species tested. The diagnostic specificity and sensitivity of this assay calculated across all chicken, duck and turkey sera were 32% and 85%, respectively, and those for chicken sera were 63% and 66%, respectively. The high specific and sensitive detection of antibodies to H5 HA in chicken sera was achieved in the bELISA designed by Jensen et al. [19]. However, the high accuracy of the test was achieved only after two subsequent ELISAs were performed using the H5N7 and H5N2 inactivated viruses as the coating antigens to circumvent interference with the NA protein. In addition, the H5-subtype specificity of this bELISA was not adequately validated, as only five subtypes other than H5 were considered in the cross-reactivity testing.
The most promising studies presenting mAbs against H5 HA and the application of one of them, the 5D8 mAb, in the development of H5 cELISA were described by Moreno et al. [20]. The assay clearly differentiated homologous and heterologous positive sera and performed very well in terms of diagnostic specificity and sensitivity for different avian species. Such performance characteristics of the assay were due to the wide intra-H5-subtype reactivity of the 5D8 mAb. It was shown that the competitor antibody recognizes a conformational epitope in the receptor-binding domain of the HA1 subunit and inhibits hemagglutination by H5-subtype AIVs. This implies that the cELISA will not detect antibodies to influenza viruses mutated under selection pressure of HI antibodies targeting the binding epitope of the 5D8 mAb. Owing to usage of the G-7-27-18 mAb with no HI activity (Additional file 2: Tables S3 and S4), detectability of the H5-subtype-specific antibodies in the EB-ELISA presented here is not affected by mutations induced with HI antibodies. In this respect, the assay developed by us is advantageous over the cELISA reported by Moreno et al. [20].
The blocking antibody in the present H5 EB-ELISA combines the inability to inhibit hemagglutination with high and broad-range specificity against H5 HAs [9], indicating conservation of its target epitope within HAs of the H5-subtype influenza viruses. Thus, the assay can be expected to diagnose infections with the currently circulating and novel H5Nx viral strains. As such, the H5 EB-ELISA meets the demand for diagnostic tools to accurately identify the HA-subtype specificity of antibodies to AIV, which is reported despite the availability of commercial kits [4]. The examples are the FluAC H5 test (IDVet, France) and H5-HA antibody ELISA kit (Dialab, Germany), which are designed to detect the anti-H5 HA antibodies in bird and human sera, respectively. The diagnostic performance of the FluAC H5 test has been evaluated in the domestic poultry population of Vietnam, partially vaccinated with reassortant H5N1 LP virus vaccine [21], ducks experimentally infected with LP and HP H5-subtype AIVs or immunized with H5 HA-encoding DNA vaccine [22], waterfowl experimentally infected with LP and HP H5N1 AIVs [23], mute swans [24], and zoo birds vaccinated with inactivated H5N9 AI vaccine [25]. In some of these studies, the assay showed low degree of cross-reactivity with antisera to the non-H5 subtype AIVs [21], a specificity value of only 89.4% [22] and variable sensitivity depending on the tested viral strain [17, 23]. In particular, low rates of detecting anti-H5 HA antibodies, 14% and 22%, were noted for antisera against the A/Chicken/West Java/SMI-PAT/2006 H5N1 HPAIV [23] and Egyptian HPAIV H5N1 antigenic drift variant [17], respectively. To improve the performance of the FluAC H5 test, modifications to the manufacturer's protocols [23, 24] or revalidation of the cutoff value [21] are suggested. Evaluation of the H5-HA antibody ELISA kit showed that the assay detects only high levels of anti-H5 HA antibodies and may produce false positive results for antisera towards seasonal H3N2 and H1N1 influenza viruses [26].
The disadvantages of the referenced experimental and commercial ELISAs justify our efforts to provide a novel screening test for subtype-specific serodiagnosis and surveillance. The accuracy of the current H5 EB-ELISA is sufficiently high to consider its further optimization and validation in accordance with the OIE guidelines [27]. Future analyses should include sera from chickens infected with H5Nx viruses. Depending on the scope of validation work, the test could be used for serological analyses in chickens and other poultry species, such as geese, ducks or turkeys, as well as in wild birds. The present H5 EB-ELISA also provides a basis for developing an assay designed to diagnose infection with H5Nx influenza viruses in humans.