An in vivo experimental model to determine antigenic variations among infectious bursal disease viruses

Abstract

Infectious bursal disease virus (IBDV) is a double-stranded RNA virus causing infectious bursal disease in chickens. IBDV undergoes antigenic drift, so characterizing the antigenicity of IBDV plays an important role for identification and selection of vaccine candidates. In this study, an in vivo experimental model was developed to differentiate a new antigenic variant of IBDV. To this end, a hyper-immune serum to IBDV E/Del-type virus was generated in specific pathogen-free chickens and a standard volume of the hyper-immune serum was serially diluted and injected in specific pathogen-free birds via intravenous, subcutaneous, or intramuscular routes. The chickens were bled at different time points in order to evaluate the dynamics of virus neutralization titres. Based on the results, chickens were injected with different serum dilutions by the subcutaneous route. Twenty-four hours later, chickens were bled and then challenged with 100 median chicken infectious doses of the E/Del virus and a new IBDV variant. Chickens were euthanized at 7 days post infection and the bursa of Fabricius was removed for microscopic evaluation to determine the bursal lesion score. The determined virus neutralization titre along with the bursal lesion score was used to determine the breakthrough titre in the in vivo chicken model. Based on the data obtained, an antigenic subtype of IBDV was identified and determined to be different from E/Del. This model is a sensitive model for determination of IBDV antigenicity of non-tissue culture adapted IBDV.

Introduction

Infectious bursal disease (IBD) is an immunosuppressive disease of young chickens and is a threat to the poultry industry worldwide (Allan et al., 1972; Faragher et al., 1972; Hirai et al., 1974d; Rosenberger & Gelb, 1976; Giambrone et al., 1977; Saif, 1991). The causative agent, infectious bursal disease virus (IBDV), affects the bursa of Fabricius (BF) due to a lytic infection of proliferating lymphocytes of the B-cell lineage (Ivanyi & Morris, 1976; Hirai & Calnek, 1979; Käufer & Weiss, 1980; Müller, 1986). The disease was first described by Cosgrove (1962) as avian nephrosis and was later recognized as IBD due to the pathologic/histologic changes in the BF (Hitchner, 1963; Helmboldt & Garner, 1964). Chickens between 3 and 6 weeks of age are highly susceptible for clinical IBD (Cosgrove, 1962; Hirai et al., 1974d; Ley et al., 1983) when protective levels of neutralizing antibody levels are absent. IBDV is a double-stranded, non-enveloped, icosahedral virus (Delmas et al., 2005) with diameter ranging from 55 to 65 nm (Hirai et al., 1974b; Nick et al., 1976). IBDV belongs to the family Birnaviridae, genus Avibirnavirus. IBDV is a bisegmented virus containing segments A and B (Dobos et al., 1979; Müller et al., 1979; Kibenge et al., 1988). Segment A encodes on two open reading frames (ORFs) for viral proteins (VPs). ORF1 encodes for the viral polyprotein, which is autoproteolitically cleaved by its viral protease to three VPs, the premature preVP2, VP4 and VP3 (Birghan et al., 2000; Lejal et al., 2000). A second ORF (Spies et al., 1989, Bayliss et al., 1990) precedes and partially overlaps the first ORF and encodes for VP5 (Mundt & Müller, 1995), which is non-essential for virus replication in cell culture (Mundt et al., 1997) but attenuated IBDV when the expression was omitted (Yao et al., 1998). Segment B encodes for VP1, which represents the viral RNA-dependent RNA polymerase (Spies et al., 1987; von Einem et al., 2004).

Based on virulence in susceptible chickens, IBDV serotype I strains were broadly classified into three pathotypes: classical, very virulent (vv), and variant strains. Classical IBDV strains cause a strong inflammatory response in the BF, and chickens infected with virulent classical IBDV show clinical signs (Cosgrove, 1962; Käufer & Weiss, 1980). In contrast, variant IBDV strains result in atrophy of the BF in the absence of any inflammatory changes (Rosenberger & Cloud, 1986; Sharma et al., 1989). The subclinical disease caused by variant IBDVs may barely be noticed and sometimes only an increase in the incidence of respiratory diseases might be observed (Saif 1984; Rosenberger & Cloud, 1986). More dramatic are vvIBDV, which cause a fast depletion of the B lymphocytes in the BF that is also associated with petechial haemorrhage in the muscles and a haemorrhagic BF (van den Berg & Meulemans, 1991). IBDV strains described in the early 1960s from the USA belong to what is now referenced as the classical subtype. IBDV strains isolated in the USA in the mid-1980s were described as variant strains due to change in their pathogenic phenotype and were later determined to be antigenically different from earlier viruses (Saif, 1984; Rosenberger et al., 1985; Snyder et al., 1988a, 1988b; Snyder, 1990). All three pathotypes of IBDV induce immunosuppression, leading to an increase in the opportunity for secondary pathogens to invade the host and cause multisystemic diseases (Cho, 1970; Fadly et al., 1976; Rosenberger & Gelb, 1978; Pejkovski et al., 1979; Yuasa et al., 1980; Moradian et al., 1990). There are several variant strains of IBDV described—such as Md (Saif, 1984), E/Del (Rosenberger et al., 1985), variant A (Rosenberger et al., 1987), IN (Ismail et al., 1990), GLS and DS 326 (Snyder et al., 1992), and AL2 (Toro et al., 2009)—circulating in US commercial poultry operations. Based on the variability of IBDV strains in the field, immunization with vaccines that are antigenically similar to the circulating IBDV strains plays a key role in IBDV control. Breeder hens are commonly vaccinated with a series of live and killed vaccines in an effort to induce antibodies that are transferred to the offspring via the eggs (Wyeth & Cullen, 1978, 1979; Wyeth 1980). Vaccination programmes differ between poultry operations and are highly dependent upon the virus challenge circulating in the field.

Identification of circulating field strains of IBDV plays an important role in prevention and control. Field surveillance studies are necessary in the poultry flocks within commercial operations to select the most appropriate vaccines for use in vaccination programmes as well as to identify possible vaccine candidates (Dormitorio et al., 2007). Various diagnostic tools were developed for identification of IBDV in poultry flocks. Initially, isolation and identification of IBDV by electron microscopy was used to confirm the presence of the virus (Hirai et al., 1974a, 1974c). Diagnosis of IBDV was later based on the agar gel diffusion test (Hirai & Shimakura, 1974b) and also by virus neutralization (VN) assays (Weisman & Hitchner, 1978). Advanced molecular methods, such as reverse transcription-polymerase chain reaction (RT-PCR), supported rapid diagnosis of IBDV (Lee et al., 1992; Clementi et al., 1995; Moody et al., 2000). RT-PCR followed by restriction fragment length polymorphism (RFLP) was used for genotyping IBDV (Jackwood & Jackwood, 1994; Lin et al., 1994; Jackwood & Nielsen, 1997; Jackwood & Sommer, 1998) and was intended to diagnose IBDV based on comparison of RFLP patterns from unknown IBDVs with patterns of known viruses. However, these methods did not provide the most important information for determining the antigenicity of the viruses. Antigenic characterization of the IBDVs was necessary to determine whether changes in field viruses would result in a diminished protection from current vaccines. For this purpose, an antigen-capture enzyme-linked immunosorbent assay was developed that provided a system for differentiation between different antigenic subtypes of IBDV based on certain reaction patterns using a panel of neutralizing monoclonal antibodies (mAbs; Van der Marel et al., 1990). In-depth analysis of amino acid exchanges causing differences in the antigenicity of IBDV lead to the conclusion that IBDV antigenicity is more complex than expected (Letzel et al., 2007; Icard et al., 2008; Durairaj et al., 2011). It also became clear that amino acid exchanges outside the projection domain might result in antigenic differences (Durairaj et al., 2011). Using this approach, several IBDVs lacking reactivity with any neutralizing mAb and indicating a different antigenic make-up (Icard et al., 2008; Durairaj et al., 2011) were detected and isolated (IBDVn-Var). Attempts to characterize the viruses in cross-neutralization assays in embryonated eggs were subjective since the viruses were not lethal in embryos, as observed with some IBDV strains, and embryo lesions were variable. We therefore developed an in vivo model to evaluate the antigenicity of IBDVn-var strains. Since E/Del-like IBDV are the most common variant strains circulating in the USA, a system based on E/Del neutralizing antibodies was developed with a read-out system focused on lesions in the BF. Using this system, we were able to demonstrate that one virus field isolate was antigenically different from E/Del IBDV. The in vivo experimental model described will serve as a defined platform for biological evaluation of IBDV field isolates, lacking an antigenic signature in our in vitro characterization model, with known vaccine strains.

Materials and Methods

Cells

The VN assay was performed in the chicken embryo fibroblast cell line DF1 (Himly et al., 1998) grown in Dulbecco's modified Eagle's medium with 4.5 g/l glucose (DMEM-4.5; Thermo Scientific, Waltham, Massachusetts, USA) supplemented with 10% foetal bovine serum (Mediatech, Manassas, Virginia, USA), penicillin (100 IU/ml) and streptomycin (100 µg/ml). Cells were incubated at 37°C with 5% CO₂.

Virus propagation in specific pathogen-free chickens

The viruses used in this experiment were E/Del (kindly provided by Ruud Hein, Intervet, Millsboro, Delaware, USA) and one IBDV field isolate from Alabama (IBDVn-var, Genbank accession number JF748992) previously described (Durairaj et al., 2011). For virus propagation, 3-week-old specific pathogen-free (SPF) chickens (Merial, Gainesville, Georgia, USA) were orally infected with E/Del and IBDVn-var in separate Horsfall–Bauer-type isolation units in a forced-air positive-pressure system (Poultry Diagnostic and Research Center, The University of Georgia, Athens, Georgia, USA). The SPF chickens were maintained in isolators and given feed and water ad libitum for the duration of the study. All chicken experiments were approved by the University of Georgia animal care and use committee (AUP number A2010 04-064-Y2-A1). The chickens were humanely euthanized and necropsied 96 h after infection. The BFs were harvested and homogenized in viral transport medium as previously described (Durairaj et al., 2011). The virus stocks were aliquoted and stored at −80^oC until further use.

Virus titrations

Both viruses were titrated in 9-day-old embryonated SPF eggs (Merial) and 3-week-old SPF chickens to determine the median embryo infectious dose (EID₅₀) and the median chicken infectious dose (CID₅₀), respectively. The determination of the EID₅₀ was performed by inoculating 100 µl per dilution into five embryonated eggs via the chorio-allantoic membrane route (Hitchner, 1970). The inoculated embryonated eggs were maintained in an incubator at 37°C with 55% relative humidity and candled daily to check for viability. Eggs with dead embryos were removed and stored at 4°C until 7 days after inoculation when the remaining embryonated eggs were evaluated for the presence of lesions characteristic of IBDV (retarded growth, green-spotted liver, enlarged spleen). For the determination of the CID_50, the SPF chickens were divided into groups of five chickens each. Chickens were bled prior to inoculation and the sera were tested by VN assay to confirm the absence of pre-existing IBDV antibodies. Seven groups per virus were infected by the ocular route with serially diluted viral inoculum of E/Del or IBDVn-var. One group was not inoculated and served as the negative control. The chickens were maintained in Horsfall–Bauer isolation units and given feed and water ad libitum. Seven days after infection, the chickens were humanely euthanized, and BFs were collected and preserved in 10% neutral buffered formalin for microscopic evaluation. The lesions in the BF were scored based on the degree of B-lymphocyte depletion using a bursal lesion score (BLS) from 1 to 4: score of 1, up to 10% of the bursal follicles show depletion; score of 2, 10 to 30% of the bursal follicles show depletion; score of 3, 31 to 70% of the bursal follicles show depletion; and score of 4, >70% of the bursal follicles show depletion. Any BLS >1 was regarded as positive. Calculations of both the EID₅₀ and CID₅₀ were performed according to the method of Reed & Muench (1938).

Generation of hyperimmune serum in chickens

Three-week-old SPF chickens were orally infected with 100 µl E/Del-like virus 8903 (kindly provided by Ruud Hein, Intervet), with a titre of 10^1.5 median tissue culture infective dose (TCID₅₀). Three weeks later, 100 µl E/Del-like 8903 strain (10^2.0 TCID₅₀) was administered by the intramuscular route. Chickens were intramuscularly injected 3 weeks later with an oil-emulsion vaccine containing one volume part β-propiolactone inactivated 8903 virus containing 10⁷ TCID₅₀/ml prior to inactivation and one part incomplete Freund's adjuvant. The inactivated virus was passaged twice in DF1 cells to confirm successful inactivation due to absence of any cytopathic effect. Five weeks later, chickens were exsanguinated under aseptic conditions. The clotted blood samples were centrifuged by 700×g at 4°C, and serum individually harvested and heat inactivated at 56°C for 30 min. The virus neutralizing antibody titre was determined (see below) and serum samples with a virus neutralizing antibody titre > 2¹³ were pooled, aliquoted and the final virus neutralizing antibody titre was determined for the pool. The serum samples were stored at −80^oC.

Determination of virus neutralizing antibody titre

The VN assay was performed to determine levels of E/Del-specific neutralizing antibodies in the chicken sera obtained. First, the TCID₅₀ for the E/Del like strain 8903 was determined following standard methods. For the VN test, 50 µl foetal bovine serum-free DMEM-4.5 was added to all wells of a 96-well tissue culture plate. Next 50 µl of the serum was added to the first column (2^–1) and serially two-fold diluted through the 12th column of the plate (2^–12). In case the VN titre was greater than 2¹², a second plate was used to continue the dilutions. Next, 50 µl of virus containing 100 TCID₅₀ was added to each serum dilution and incubated for 1 h at 37°C. During that time, DF1 cells were trypsinized, resuspended in 10% foetal bovine serum-containing DMEM-4.5 and adjusted to a cell density of 5×10⁵ cells/ml. One hundred microlitres of DF1 cells were added to all wells of the virus–serum suspension and incubated for 5 days at 37°C and 5% CO₂. After incubation, cells were scored for the presence of a CPE. The end-point of the VN test for a serum sample was determined to be the reciprocal of the highest dilution, expressed in log₂, in which there was no visible CPE. During each test, the diluted virus was back-titrated to determine the true TCID₅₀. If the TCID₅₀ was greater than a 0.25 log₁₀ difference from 100 TCID₅₀, the VN test was regarded as invalid and was repeated.

Administration of hyperimmune serum

Two 3-week-old SPF chickens per group were used for this study. Three hundred microlitres of the hyperimmune serum were administered to each chicken via either the intramuscular (i.m.), intravenous (i.v.) or subcutaneous (s.c.) route and two chickens were used as negative control. Each group was kept in a Horsefall–Bauer isolation unit and given water and feed ad libitum. Chickens were bled at 24, 48, 72, and 168 h post injection via the brachial vein on the side opposite of the inoculation. The clotted blood samples were centrifuged at 700×g for 5 min and the serum was harvested. The serum samples were heat inactivated at 56°C for 30 min and the VN test was performed as described above to study the dynamics of VN antibody titre.

Differentiation of IBDV strains in vivo

Three-week-old SPF chickens were divided into 21 groups of five chickens per group. Each chicken was wing-banded for identification. Serum was diluted up to a dilution of 1:128. Ten groups of five chickens per group were used for each virus during the experiments: Group 1 (serum undiluted), Group 2 (1:2 diluted), Group 3 (1:4 diluted), Group 4 (1:6 diluted), Group 5 (1:8 diluted), Group 6 (1:16 diluted), Group 7 (1:32 diluted), Group 8 (1:64 diluted), Group 9 (1:128 diluted), and Group 10 (phosphate-buffered saline). Twenty-four hours after injection of the appropriate serum dilution, the chickens were bled via the brachial vein on the opposite wing and blood samples were processed for the VN test as described above. One hour later, one group from each serum dilution level was challenged with either 100 CID₅₀ of E/Del or 100 CID₅₀ IBDVn-var. One group of SPF chickens did not receive serum and was left unchallenged. Seven days after challenge, chickens were humanely euthanized and BF collected for microscopic evaluation to determine the BLS.

Histopathology

Bursal tissues were collected at the time of necropsy on an individual basis to allow the matching of each serum sample to the appropriate chicken. The bursal sample was placed in 10% neutral buffered formalin, and was paraffin embedded. Sections of the paraffin-embedded BF were stained with haematoxylin and eosin following standard histologic procedures. The stained sections were microscopically examined for the presence of bursal lesions. Based on the depletion of the B lymphocytes, the BLS was determined for each BF based on the scoring system described above.

Results

Determination of viral titres

Initially, cross-neutralization assays in SPF embryos inoculated via the chorio-allantoic membrane were performed to determine the antigenic relatedness of IBDVn-var with E/Del. However, IBDVn-var was not embryo lethal and identification of embryo lesions was minimal and somewhat subjective. Although results from the initial cross-neutralization studies in embryos indicated IBDVn-var was antigenically different from E/Del, additional testing was needed to provide more definitive and reliable data. Determination of the CID₅₀ was thus undertaken to determine the titre of the viruses. The read-out system was the presence of lesions in the BF, thus providing the most sensitive system for determining infection in each dilution. The infectious titre was determined using material generated from the BF. This would provide a standard and comparable system for evaluating infection at later time points, during titration and challenge. Although the EID₅₀ was not used during the study, the calculated titres for the E/Del bursa material was 10^4.1/100 µl for EID₅₀ and 10^4.5/100 µl for CID₅₀. Interestingly, titration of IBDVn-var resulted in a titre of 10^3.5/100 µl (EID₅₀) and 10^4.5/100 µl (CID₅₀). Thus if the calculation for the infectious dose would have been based on 100 EID_50, the infectivity for the BF would have been underestimated by a factor of 10 for IBDVn-var and a higher dose would have been used for infection in comparison with the E/Del virus.

Generation and administration of hyperimmune serum

The final VN titre of the pooled hyperimmune serum from chickens vaccinated with the E/Del-like strain 8903 as described above was 2¹⁴. This serum was used for all subsequent experiments. Three routes of application (i.m., s.c., i.v.) were tested for their practicality and reliability. From the user's point of view, the easiest route was the i.m. application. The s.c. route of application had the same practicality and reliability. The i.v. route was not easy to perform and failed several times during experiments (data not shown). Based on the VN titres measured at several time points after serum administration, it was observed following i.m. application that the VN titre decreased relatively fast (Table 1). As expected, the VN titre was highest after i.v. application in comparison with both i.m. and s.c. application. The s.c. administration resulted in a higher titre than the i.m., but lower than the i.v. route of administration; however, the decline of the s.c. VN titre was comparable with the i.v. application. After taking into consideration all aspects of the serum administration, the s.c. route was chosen since it was easy to perform, reliable and resulted in VN titres that lasted for a relatively long period of time.

Table 1.  Virus neutralization titre after different applications methods and time points.

		Hours after serum administration
Chicken wing-band ID	Route of application	24 h	48 h	72 h	168 h
494	i.m.	256^a	32	64	128
475	i.m.	128	32	32	32
487	i.v.	512	128	128	64
489	i.v.	512	256	64	64
490	s.c.	256	64	64	32
493	s.c.	256	256	64	128
486	Control	<8	<8	<8	<8
491	Control	<8	<8	<8	<8
^aThe VN titre shown as the reciprocal of the final two-fold dilution where no cytopathic effect was observed.

Comparison of E/Del and IBDVn-var in the in vivo model

The experiments were performed using 3-week-old SPF chickens. The chickens were s.c. injected with E/Del-specific antiserum at different dilutions. The VN titre prior to dilution was 2¹⁴ when 100 TCID₅₀ of the E/Del-like strain 8903 were used for the VN test. In a separate experiment, the identity of the tissue-culture-adapted 8903 was confirmed by indirect immunofluorescence using the panel of mAbs (10, 57, R63, 67, B69), which resulted in a positive reaction with only R63 and 67 as previously described (Icard et al., 2008). This confirmed the E/Del-like antigenic subtype. Twenty-four hours later the chickens were bled and the VN titres were determined using the E/Del-like strain 8903 (Tables 2 and 3). As expected the VN titres declined in parallel with the dilution of the serum. Furthermore, the VN titres ranged from one titre (Group 9) to five different titres (Group 5) and indicated variable uptake of the antibodies into the bloodstream when administered by the s.c. route. The highest VN titre observed was 1024 in three chickens injected with either the undiluted or the 1:2 diluted serum. The control chickens were free of any IBDV strain 8903 neutralizing antibodies. The result of the challenge infection with IBDV strain E/Del and the IBDVn-Var was evaluated using the BLS at day 7 following challenge. In the group of the chickens infected with only virus and no serum, a BLS of 4 was detected in every chicken indicating a sufficient infectious dose of virus. Chickens that did not receive serum or virus had a BLS of 1. Chickens that received serum preparations and were challenged were regarded as protected if the BLS was between 1 and 2. The results showed that E/Del-challenged chickens were fully protected up to a VN titre of 128 (see Table 3). A few chickens were protected when the VN titre was ≤64. Even at a VN titre of 8, two out of six chickens were fully protected—as indicated by a BLS of 1.

Table 2.  Virus neutralization titre at day of challenge infection.

		Virus neutralization titre^b
Group	Serum dilution^a	1024	512	256	128	64	32	16	8	<8
1	Undiluted	2^c	7	1	0	0	0	0	0	0
2	1:2	1	5	4	0	0	0	0	0	0
3	1:4	0	0	5	4	0	1	0	0	0
4	1:6	0	0	1	2	4	1	2	0	0
5	1:8	0	0	0	1	2	2	4	1	0
6	1:16	0	0	0	1	0	3	1	5	0
7	1:32	0	0	0	0	0	0	1	8	1
8	1:64	0	0	0	0	0	0	1	2	7
9	1:128	0	0	0	0	0	0	0	1	9
10	No serum	0	0	0	0	0	0	0	0	10
^aSerum dilution factor prior to s.c. administration. ^bThe VN titre is the reciprocal of the final two-fold dilution where no cytopathic effect was observed. ^cNumber of chickens in the corresponding serum dilution group that showed the appropriate VN titre.

Table 3.  Protection of chickens after challenge infection with two different IBDV strains.

	Challenge virus
	E/Del		IBDVn-var
VN titre^a	Protected/total	BLS^b	Protected/total	BLS
1024	1/1^c	1	2/2	1,1
512	6/6	1,1,1,2,2,1	6/6	1,1,1,2,2,2
256	5/5	1,1,1,1,1	4/6	1,1,4,4,1,1
128	3/3	1,1,1	3/5	4,4,2,2,1
64	2/5	4,4,4,2,1	1/1	1
32	3/5	1,4,4,1,1	0/2	4,4
16	1/4	4,1,4,4	1/5	1,4,4,4,4
8	2/6	4,1,1,4,4,4	0/11	4 (11×)
<8	0/15	4 (15×)	0/12	4 (12×)
^aVN titre at the day of challenge/infection. ^bBLS of each chicken in the corresponding group. ^cNumber of chickens protected/total number of chickens in corresponding VN titre group. Chickens were regarded as protected from challenge when BLS ≤2.

Results from the challenge infection with IBDVn-var were different. Only chickens with a VN titre ≥512 were fully protected. Fifty to 60% (VN titres of 256 and 128, respectively) of birds were protected from challenge infection, which indicated that IBDVn-var was able to break through VN titres at levels where chickens were still fully protected when challenged with the homologue E/Del strain. The data clearly showed that only one chicken out of 30 was protected from IBDVn-var challenge at VN titre levels ≤32, while six out of 30 chickens were still protected at these titres after challenge with E/Del. It needs to be mentioned that no clinical signs of disease were observed in any chickens, also indicating that the IBDVn-var clinically belongs to the large group of variant IBDV strains.

Discussion

IBDV is ubiquitous in poultry operations worldwide. Mutations in the viral progeny can arise during replication as the viral replicase probably lacks a proofreading mechanism, responsible for repair of misincorporated nucleotides. In addition, changes in the antigenic make-up can occur by selection of escape mutants due to presence of neutralizing antibodies and may result in IBDVs that differ from vaccines in either their antigenicity or virulence or both. The results of this selection process are not predictable. Consequently, this requires the continuous characterization of field isolates for their pathogenic potential and antigenic make-up. The development of neutralizing mAbs that were used to classify different IBDV antigenic subtypes (Snyder et al., 1988a, 1988b, 1992; Snyder, 1990) and their subsequent use in an antigen-capture enzyme-linked immunosorbent assay was an important step for antigenic characterization of field isolates (van der Marel et al., 1990) and served as an effective surveillance tool. More rapid techniques were developed and the use of RT-PCR followed by restriction enzyme digestion (RFLP) of the amplified cDNA fragment (Jackwood & Jackwood, 1994) became mainstream for typing IBDV. Using this approach it was possible to distinguish between different IBDV subtypes, but did not provide essential information regarding the virulence or antigenicity of viruses (Jackwood & Nielsen, 1997; Jackwood & Sommer, 1997). In some cases, the data generated by RT-PCR/RFLP led to the incorrect designation of IBDV subtype. One example was the detection of two IBDV isolates that were grouped into the vvIBDV group based on RFLP. Surprisingly, one isolate caused 70 to 80% mortality, typical for vvIBDV, while the other isolate caused only 10% mortality (Hoque et al., 2001).

Various antigenic strains of IBDV have been described in the USA based on their reactivity with mAbs: E/Del-like IBDV reacted with mAbs R63 and 67, GLS-like IBDV reacted with mAbs 10 and 57, and classical IBDV reacted with mAbs 10, R63, and B69 (Snyder et al., 1988a, 1988b, 1992; Snyder, 1990). In addition, an IBDV isolated in Belgium reacted with mAbs 10, R63, 67, and B69, a combination previously not described (Letzel et al., 2007). Furthermore, a new group of viruses was described that did not react with any of the mAbs using a diagnostic reverse genetics system for IBDV (Icard et al., 2008; Durairaj et al., 2011). It is important to keep in mind that virus isolates represent only a snapshot of IBDVs evolving in the field. Based on an evolutionary advantage they might become established in an environment, or in the case of an evolutionary disadvantage they may go unnoticed. Surveillance is thus a necessary tool for monitoring IBDV evolution in the field. The recently described IBDVs that did not react with any mAbs have also been identified in South America and Europe (unpublished data). Consequently these IBDV isolates need to be characterized either in vitro or in vivo. One virus isolate from Alabama, belonging to the group of IBDVs that do not react with any of the mAbs, was evaluated in vitro and in vivo (Durairaj et al., 2011). Virus titrations in SPF chicken embryos repeatedly resulted in no embryo mortality and minimal embryo lesions that were inconsistent between virus dilutions, making it difficult to quantify the virus after infection of chicken embryos (data not shown). This particular virus appeared to be less embryo adapted than observed for other IBDV field isolates. Ultimately the virus was titrated in chickens and provided the read-out system for subsequent in vivo VN assays. The use of susceptible IBDV antibody-free chickens represents the most sensitive and reliable system for detection and quantification of IBDV. In the in vivo model described, the IBDV antibody titres detected in the serum after administration of IBDV hyperimmune serum using the s.c. route were comparable with those obtained following i.v. administration, although only two chickens were used for the experiments. The results showed that, independent of the serum dilution, IBDV antibody titres were present in a gradient in the chicken serum. The VN titre levels were not always in agreement with the dilution of the serum used in a particular group. This may be a result in differences in the uptake of IBDV antibodies into the bloodstream. Owing to this finding, the data were arranged by chicken and corresponding VN titre followed by the addition of the BLS score. The results obtained showed that the E/Del was neutralized, as expected, at E/Del-specific VN titre levels, whereas IBDVn-var was able to break through even higher VN titres and induced bursal lesions. These results confirmed the assumption that a virus lacking reactivity with any of the neutralizing mAbs in the in vitro assay was also antigenically different from E/Del since the same IBDV titres for infection were used. The neutralizing epitopes that have been recognized by the immune system of the mice are thus also important for VN in chickens. Besides this, other, so far unidentified, neutralizing epitopes must also be present on the viral capsid that may be more cross-reactive. This assumption is based on the observation that at least partial VN of IBDVn-var was observed. These epitopes are shared between both viruses investigated here and requires the analysis of such cross-neutralizing epitopes in more detail for a better understanding of the biology of the virus. This might help to improve virus vaccines by molecular techniques that are available. The findings described here do not show that IBDVn-var was also different from classical as well as GLS-like IBDVs, but allows the assumption. Hyperimmune sera specific for classical IBDV (e.g. D78) or GLS-like IBDV (GLS-05) are being prepared as described in this paper and will be used similarly in this in vivo system for a more in-depth analysis of newly emerging IBDV field isolates.

In summary, an in vivo experimental model was developed that can be used for the biological evaluation of IBDV field isolates in unvaccinated susceptible chickens. The use of IBDV subtype-specific antisera for passive immunization will allow the determination of the break-through titres between IBDV isolates in a titre-dependent manner. The in vivo experimental model is a highly sensitive method for evaluating IBDV infections since the results are based on lesions in the BF that represents the target organ for IBDV. It is clear that this study provides data to support the use of an in vivo neutralization model to identify relevant antigenic differences between IBDVs and further refinement of the model is needed to decrease the numbers of chickens necessary for the analysis and to finally establish an in vitro method.