Antigens

Equad proteins (ectodomains of the E protein containing four point mutations near the FL domain) for WNV, USUV and TBEV were produced in a Drosophila S2 cell expression system (Thermo Fisher, Waltham, USA) and purified as previously described by Rockstroh et al.³⁴ and Berneck et al.³⁵.

Avian serum samples

Avian serum samples (n = 271) from chickens (Gallus sp.), ducks (Anatidae sp.) and geese (Anserinae sp.) from experimental studies as well as field sera from the National Reference Laboratory (NRL) were included in this study. Flavivirus IgY-positive poultry samples (n = 127) were tested positive for either WNV, TBEV or USUV. A comparable set of flavivirus IgY-negative sera from chickens, ducks and geese (n = 144) was also included. A detailed list of the origin of each serum sample can be found in Supplementary Tables S1 – S7. All serum samples were characterized with regard to their flavivirus antibody titers using a commercially available competitive ELISA (ID Screen^® WN competition ELISA, IDVet, Grabels, France) followed by virus neutralization tests (VNT) as described by Seidowski et al.³⁸ using the mammalian cell lines Vero B4, Vero 76 and PK15, which were obtained from the collection of Cell Lines in Veterinary Medicine (Friedrich-Loeffler-Institut (FLI), Germany). Duck and goose sera were combined into a single group because preliminary tests during protocol establishment indicated that both require the same procedures in the different ELISAs. Detailed information on the flavivirus antibody status and the amount of tested avian sera is provided in Tables 1 and 2 for Equad ELISAs and Equad pre-absorption ELISAs, respectively.

Equine serum samples

Equine field serum samples from the NRL as well as sera from routine diagnostics in veterinary clinics are part of this study. A detailed list of the origin of each serum sample can be found in Supplementary Tables S8 – S11. Information regarding the antibody status of the equine serum samples (n = 136) and the number of samples tested is presented in Tables 1 and 2. Samples were classified as either positive for WNV, USUV or TBEV or flavivirus antibody negative. The samples were examined to determine their Flavivirus IgG antibody titers using a commercially available competition ELISA (ID Screen^® WN competition ELISA, IDVet, Grabels, France) followed by VNTs on sera that yielded positive ELISA results.

Determination of antibody status of the serum samples

Pre-characterization of the sera utilized for the Equad and Equad pre-absorption ELISAs was conducted employing commercially available ELISAs and VNTs. Sera were characterized with regard to their flavivirus antibody titers using the ID Screen^® Flavivirus competition ELISA (formerly named West Nile competition ELISA at the time of testing, IDVet, Grabels, France). Sera that yielded positive results were further tested in WNV, USUV and TBEV virus neutralization tests for which the mammalian cell lines Vero B4, Vero 76 and PK15 were used. The corresponding VNT titers to the positive sera of the panel are provided in the Supplementary Tables S1-S11. Positive sera from naturally infected birds were tested against WNV and USUV in the VNT to ensure reliable differentiation. The experimentally infected geese and chickens were purchased as day-old chicks or SPF-free hatching eggs^39,40,41. The TBEV positive duck sera originated from an experimental infection where animals were purchased at four weeks of age and tested negative for TBEV in the VNT⁴², which theoretically leaves the possibility of a heterologous flavivirus infection. Positive horse sera were tested against WNV, USUV and TBEV in the VNT. However, a few isolated sera constitute exceptions to this pattern. These sera originate from regions and years in which the presence of antibodies against WNV, USUV, or TBEV could be ruled out with certainty, as these viruses were not present at the time and in the area in question. These details are provided for each serum in the Supplementary Tables S8 – S11. The VNT titer for the virus for which the sample was tested positive was always at least four times higher than VNT titers of the other flaviviruses tested, in accordance to standard procedures⁴³. This does not exclude the slight possibility that animals do have antibody titers against a heterologous flavivirus that are less than four times lower than those against the flavivirus under investigation.

ELISA setup

Different indirect ELISA assays using Equad proteins as antigens were established for the testing of either chicken, duck and goose or horse sera for WNV, USUV or TBEV IgY/IgG antibodies. To define optimal test conditions, different serum sample dilutions (1:10, 1:50, 1:100) and several secondary antibodies at varying dilutions (1:10,000, 1:20,000, 1:50,000) were compared for avian sera. Furthermore, incubation at room temperature or at 37 °C was compared.

After optimization for chicken serum samples either 160 ng of WNV Equad protein or 200 ng of TBEV Equad protein per well in 100 µl ELISA coating buffer (15 mM Na₂CO₃, 35 mM NaHCO₃, pH 9.6) were applied onto Nunc 96 well polysorb plates (Thermo Scientific) at 4°C overnight. The following day, the plates were washed three times with 350 µl PBS-0.05% Tween20 per well and then blocked with 200 µl of 5% skimmed milk (blocking grade) in PBS for two hours at room temperature. After another washing step, 100 µl of 1:100 diluted chicken sera in skimmed milk in PBS were added to each well and incubated for one hour at room temperature. Following another washing step, 100 µl of the HRP-conjugated secondary antibody (goat anti-bird IgG antibody, Bethyl Laboratories) was added to each well at a dilution of 1:50,000 and the plates were incubated for one hour at room temperature. A final washing step was performed, followed by the addition of 100 µl TMB substrate (BioLegend) per well for 30 min at room temperature in the dark. The reaction was then stopped with 50 µl 1M H₂SO₄ per well. The signals were read out at 450 nm with background reduction at 520 nm using a micro plate reader (TECAN infinite M200 or TECAN infinite F200 pro).

For duck and goose serum samples, the chicken serum protocol was adapted to a serum dilution of 1:50 and a secondary antibody dilution of 1:10,000. In addition, duck and goose sera were tested using either WNV Equad protein (160 ng per well), USUV Equad protein (200 ng per well) or TBEV Equad protein (200 ng per well).

Equine serum samples were tested in WNV Equad, TBEV Equad and USUV Equad IgG ELISAs. The protocol for equine samples was identical to the protocol described for chicken sera, except for the additional testing on USUV Equad protein (200 ng per well) and using a HRP-conjugated rabbit anti-horse IgG H&L (abcam) as secondary antibody at a dilution of 1:50,000.

WNV and USUV Equad pre-absorption ELISAs were established to further reduce cross-reactive binding of IgG and IgY antibodies to WNV and USUV Equad proteins in horse, duck and goose sera. The diluted sera were pre-incubated with the heterologous antigen for one hour at room temperature before being added to the wells. For this pre-absorption step, 20 µg per mL of USUV Equad protein in the WNV Equad pre-absorption ELISA or 16 µg per mL of WNV Equad protein in the USUV Equad pre-absorption ELISA were used. The other steps of the ELISA protocol were identical to those described above for duck, goose and horse sera. The amount of heterologous antigen in the pre-absorption step was determined according to Berneck et al. 2022³⁵.

Statistical analysis

The mean value of all measurements for each serum was included in the statistical analysis. Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, Inc, La Jolla, CA, USA). The results from the different ELISA setups were compared among sample groups based on their flavivirus IgG/ IgY antibody status using boxplots. For each ELISA, receiver operating characteristics (ROC) analyses were performed and the area under the curve (AUC) with a 95% confidence interval was calculated. The cut-off values for each ELISA setup were calculated using Youden’s index, which equally weights false positive and false negative values following ROC analysis. Samples were considered positive if their OD values exceeded the calculated cut-off value and sensitivity and specificity were determined.

Ethics declarations

All methods are reported in accordance with ARRIVE guidelines. All methods were carried out in accordance with relevant guidelines and regulations. Duck, goose and chicken sera were used according to the ethical approvals by the administration of the federal state of Mecklenburg-Western Pomerania, Germany (LALLF reference numbers 7221.3-2-042/17, 7221.3-2-003/23, 7221.3-1-075/16, 7221.3-1.1-018/18 and 7221.3-1-031/20). All experimental protocols were approved by the administration of the federal state of Mecklenburg-Western Pomerania. Blood samples from horses were collected during routine diagnostic testing by veterinarians for which, according to German law, ethical approval was not required.