All animal trials were conducted in an animal facility approved for animal experiments (n°C-22–745–1). They were approved by Anses Ploufragan ethical committee (committee number C2EA-016/ ComEth ANSES/ENVA/UPEC) and authorized by French Ministry for higher education and research (permit number APAFiS#3925-2016020310538529, application number 16-016, statement number 08/03/16-1). Chickens were raised and humanely euthanized in agreement with EU directive number 2010/63/UE.

Vaccine viruses used in this study were obtained from a local veterinary practice, reconstituted, and administered according to the manufacturer's instructions. To avoid any conflict of interest, currently distributed IBD vaccines used in this study were anonymized. All IBD I+ vaccines (coded as I+1, I+2, I+3, and I+4) marketed in France as of 2016 were included. As representative intermediate vaccines, two intermediate vaccines (I1 and I2) were also included. The NDV Poulvac Hitchner B1 strain (Zoetis) was used for NDV vaccination and the velogenic NDV strain “Herts33” (Anses Ploufragan laboratory) was used to challenge the birds. The classical IBDV virulent strain “F52/70” and the mild IBD vaccine strain “PBG98” (no longer marketed in France; passaged twice on chicken fibroblasts) were obtained from the collection of IBDV strains maintained by the OIE reference Laboratory (Anses, Ploufragan, France).

In total three experiments were performed. The layouts of the experiments are presented in Figure 1A (experiments 1 and 2) and Figure 3A (experiment 3), respectively. For each experiment, six groups of 10 1-day-old SPF chicks (Anses, Ploufragan, France) were housed in negative-pressure filtered-air isolators and were administered (or not, in case of controls) one dose of IBD vaccine by the eye-drop route. In the case of PBG98 or F52/70, chicks received 100 tissue culture median infectious doses (TCID₅₀) or 100 embryo median infectious doses (EID₅₀, corresponding for this virus to 740 TCID₅₀), respectively. Chicks received (or not, in case of the non-vaccinated controls) one dose of ND vaccine by the eye-drop route after 14 days of IBDV vaccination. Controls included a group that received neither IBD nor ND vaccine (“no vaccine control”), and a group that received only the ND vaccine (“NDV vaccine control”). In the case of experiment 3, prior to ND vaccination, chicks were blood-sampled at the occipital sinus with EDTA-coated tubes (S-Monovette EDTA K2 2.7 mL, Sarstedt, reference 04.1915.100) for blood cell quantification. For all experiments, chickens were blood-sampled at the occipital sinus 2 weeks post-NDV vaccination, either in serology tubes (S-Monovette serum-gel, 2.6 mL, Sarstedt, reference 04.1904) for serological analyses only (experiments 1 and 2) or in EDTA-coated tubes for blood cell quantification as well as serological analyses (experiment 3). On the same day, chicks were inoculated intra-muscularly with 10⁵ EID₅₀ of NDV strain Herts33. A clinical follow-up was performed for 14 days following the challenge. Birds were rated as “clinically protected” when no clinical sign was observed or “non-protected” when they displayed neurological, respiratory signs, or prostration. Animals showing clinical signs clearly unrelated to the NDV challenge were excluded from the clinical follow-up. Animals exhibiting neurological signs and/or prostration reached the ethical endpoint of the experiment and were consequently humanely euthanized.

In all experiments, serum (experiments 1 and 2) or plasma (experiment 3) obtained after blood sampling were analyzed by hemagglutination inhibition (HI) to quantify levels of anti-NDV antibodies according to Agence Française de Normalisation Norm U 47-011. Four hemagglutination units of NDV LaSota strain were used as antigen. Results were expressed as the reciprocal of the highest dilution of serum able to completely inhibit the hemagglutination of 4 units of NDV antigen in log₂. In this assay, titers equal to or below 2 log₂ are considered negative. To exclude any impact of EDTA from the blood-collection tube on HI titers, plasmas from experiment 3 were treated with CaCl₂ (to a final concentration of 2.5 mM) to induce fibrin precipitation (7). After 2 h of incubation at room temperature, the fibrin clot was mechanically disrupted by stirring with micropipette tips, the tubes were centrifuged at room temperature for 10 min at 14 000 × g and supernatants were collected.

In the case of experiment 3, collected blood was analyzed by flow cytometry to quantify leukocytes following the approach developed by Seliger et al. (8) with the few modifications already described (9). Accordingly, blood thrombocytes (defined as K1⁺ cells), blood monocytes (defined as K1⁺, Kul01⁺ and CD45⁺ cells), blood T cells (defined as CD45⁺ cells that are additionally either CD4⁺ or CD8⁺ or TCRγδ⁺), and blood B cells (defined as CD45⁺ and Bu1⁺ cells) were quantified.

Bursal cells were isolated from 9 to 15 week-old SPF chickens as previously described (10). Cells were maintained at 40°C, 5% CO₂ in a lymphocyte culture medium consisting of Iscove's Modified Dulbecco's Medium (IMDM) with L-glutamine and HEPES (reference 21980–032, Gibco, Thermo Fisher), supplemented with 8% FBS, 2% SPF chicken serum (ANSES, Ploufragan, France), 1X Insulin Transferrin Selenium (reference 41400–045, Gibco, Thermo Fisher), 50 μM beta-mercaptoethanol, 1 μg/mL Phorbol 12-myristate 13-acetate (reference tlrl-pma, Invivogen), penicillin (200 IU/mL), streptomycin (0.2 mg/mL), and fungizone (2 μg/mL).

For infection, one million bursal cells per well seeded in 96-well plates were inoculated at a target multiplicity of infection of 5 × 10⁻⁴. Inocula were titrated and titers are presented as the 0 h post-infection (h.p.i.) time-point in Figure 2A. Supernatants were collected at 16, 23, and 48 h.p.i. and were frozen at −70°C until titration. Additionally, at 16 and 23 h.p.i., cells were collected for immunolabeling and flow cytometry analyses. A total of 5 biological replicates (consisting of cells from distinct birds) were analyzed during two distinct experiments.

Bursal cells collected at 16 and 23 h.p.i. were labeled as previously described (10). Briefly, viability was evaluated using the Fixable Viability Dye eFluor 750 (Thermofisher, reference 65-0865-14) and cells were immunostained using an anti-VP3 monoclonal antibody [mAb 18, (11)] followed by a goat anti-mouse IgG2a Alexa Fluor 546 antibody (Thermofisher, reference A21133). Cells were analyzed on a FC500 MPL flow cytometer (Beckman Coulter).

Virus titration of inocula and supernatants at 16, 23, and 48 h.p.i. was carried out as previously described (10). Briefly, serial dilutions of viral suspensions were incubated into 96-well plates with chicken bursal cells in lymphocyte culture medium at 40°C for 48 h in a humidified 5% CO₂ incubator. Cells were then subjected to immunocytochemistry using the IBDV VP3 specific mAb 18, a horse-radish peroxidase coupled anti-mouse IgG serum and KPL Trueblue peroxidase substrate (Seracare, reference 5510-0030). Viral titers, expressed as TCID₅₀/mL, were determined using Reed and Muench formula (12).

Statistical analyses were performed using R version 4.0.3 (13). Differences in percentages of protection were analyzed using the Fisher's exact test followed by pair-wise comparisons using the “fisher.multcomp” function from RVAideMemoire package version 0.9-79 (14). All other quantitative parameters were analyzed using Kruskal-Wallis test followed by Fisher's least significant difference test with Holm adjustment method for multiple comparisons using the “kruskal” function from the package “Agricolae” version 1.3-2 (15).

In both experiments 1 and 2, following challenge with the virulent NDV strain Herts 33, all the NDV non-vaccinated control birds rapidly exhibited intense prostration and died or were euthanized within 72 h post-inoculation. All the NDV vaccine control birds showed complete clinical protection, without any sign of disease. Chicks that received both IBD and ND vaccines showed varying degrees of protection from the NDV challenge. Chicks from the PBG98-treated group showed 100% clinical protection. Those that received I vaccines had 80–90% protection, which did not differ statistically from those who received NDV vaccine controls. In contrast, protection in chicks that received I+ vaccines was significantly lower than that received NDV vaccine controls, with only 10–40% protection (Figure 1B). Finally, no animals from the F52/70 group were protected. HI testing for anti-NDV antibodies revealed expected background levels (1–2 log₂) for non-vaccinated controls and high titers for NDV-vaccinated controls, with median values of 6 and 4.5 log₂ in experiments 1 and 2, respectively (Figure 1C). Chicks vaccinated with PBG98 or I vaccines did not show any statistically significant difference in HI titers compared with NDV vaccine controls. In contrast, birds that received I+2, I+3, I+4, and F52/70 showed severe reduction in HI titers, almost reaching the background levels of non-vaccinated controls. Interestingly, chicks that received I+1 vaccine showed a numerically mild and statistically non-significant reduction in HI titers in spite of a statistically reduced clinical protection (only 40% birds protected).

Primary bursal cells obtained from SPF chickens were infected with various IBDV strains to determine whether replication parameters (e.g., replicative fitness or cytopathogenicity) could mirror immunosuppressive properties. Strain F52/70 as well as the six vaccines tested in vivo were included in this approach. Since the PBG98 vaccine does not replicate in bursal B cells (data not shown), it was not included in the panel. For each of the tested strains, a dramatic increase in viral titers measured from bursal cells supernatants was visible as early as 16 h.p.i. and viral titers increased until 48 h.p.i. (Figure 2B). Interestingly, two trends in replication curves could be observed. On the one hand, I2, I+2, I+3, I+4, and strain F52/70 reached a viral titer higher than 10⁷ TCID₅₀/mL at 23 h.p.i. and then reached almost 10⁹ TCID₅₀/mL at 48 h.p.i. On the other hand, I1 and I+1 presented viral titers of about 10⁶ TCID₅₀/mL at 23 h.p.i. and then reached a plateau phase of about 10⁷ TCID₅₀/mL at 48 h.p.i., which was statistically lower compared with the first trend for this time point. In parallel of these infectious titrations, both the percentages of viable and IBDV-infected cells were analyzed by flow cytometry (example of dot-plot in Figure 2C). Significant differences in the percentage of live infected cells were detected at 16 h.p.i. with higher values for I2, I+2, I+4, and F52/70 (Figure 2D). At 23 h.p.i., a high variability in percentages between cells originating from distinct birds was observed, and differences between viruses were no longer significant. It is noteworthy that a bursa-specific trend was observed: for instance, at least 3 fold-higher percentages of infected cells were observed for bursa number 3 compared with bursa 4 for almost all viruses at 23 h.p.i.

Based on the hypothesis that I+ vaccines, inducing higher bursal damages, may reduce blood B cell concentrations, a third in vivo experiment was implemented in which four vaccines were tested (Figure 3A): I1 and I+3, for which ex vivo replication levels were correlated to the intensity of in vivo immunosuppression, and I2 and I+1, for which this correlation was not observed. Clinical protection from NDV challenge was consistent with results from experiments 1 and 2, except for the I+1 treated group, which no longer differed statistically from the NDV vaccine control (Figure 3B). Anti-NDV antibody levels were significantly reduced for all vaccines except I1 compared with NDV vaccine controls (Figure 3C), unlike in experiment 1 (Figure 1C). To check the lack of interference of the anti-coagulant, HI testing was performed again after adding CaCl₂ to the plasmas to induce clotting. No difference in HI titers was observed after this treatment (data not shown). At both time-points, blood leukocyte counts revealed significant differences in B cell concentrations between virus strains. At 15 days of age, prior to NDV vaccination, both control groups exhibited concentrations close to 1,000 B cells/μL (Figure 3D, left panel). For I1 vaccinated birds, the median concentration significantly dropped to 143 B cells/μL. For I2 vaccinated birds, the drop was even more pronounced (median of 42 cells/μL, p < 0.05). For I+1 and I+3 vaccinated birds, B cell counts were statistically lower than in I vaccine groups, with medians below the detection limit of 8 cells/μL. In addition, no bird in these two groups presented more than 60 B cells/μL. Blood cell counts performed at 29 days of age (prior to NDV challenge) revealed a global increase of B cell counts in all groups compared with 15 days of age (Figure 3D, right panel). Both controls groups showed median concentrations above 2,000 B cells/μL. Median B cell concentrations significantly dropped to 956 for I1 vaccinated group. Those median concentrations were lower in I2, I+1, and I+3 vaccinated groups, with 130, 91, and 10 B cells/μL, respectively. Some birds from I+1 and I+3 treated groups still did not show any blood B cell at that time-point. Some milder variations were observed in other blood leukocyte concentrations: for instance, at 15 days of age, a statistically significant reduction in both monocytes and T cell concentrations were seen for I2 and I+3 groups compared with control groups (Supplementary Figure 1).