Acute infectious bursal disease in poultry: A review

Transcription

1 Avian Pathology ISSN: (Print) (Online) Journal homepage: Acute infectious bursal disease in poultry: A review Thierry P. Van Den Berg To cite this article: Thierry P. Van Den Berg (2000) Acute infectious bursal disease in poultry: A review, Avian Pathology, 29:3, , DOI: / To link to this article: Published online: 17 Jun Submit your article to this journal Article views: 4005 View related articles Citing articles: 202 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 19 December 2017, At: 04:05

2 Avian Pathology (2000) 29, REVIEW ARTICLE Acute infectious bursal disease in poultry: a review Thierry P. van den Berg* Veterinary and Agrochemical Research Centre, Section of Avian Virology and Biotechnology, Groeselenberg 99, 1180 Brussels, Belgium This review is focused on the acute form of infectious bursal disease (IBD) caused by very virulent IBD virus (vvibdv). First described in Europe about 10 years ago, this new form of the disease has rapidly spread all over the world, causing dramatic losses; after a decade, it still represents a considerable threat to the poultry industry. Emergence of the acute forms of the disease has drastically changed the epidemiology of IBD. Although their origin is still under investigation, vvibdvs have spread all over the world in a very explosive but conserved manner. This raises the question of the origin of vvibdvs, of the possible existence of reservoirs and of the possible emergence of new, distinct lineages in the future. While it has become clear that amino acids within the variable region of virus protein VP2 account for the molecular basis of antigenic variation, no definite hot spot that determines pathogenicity has been identified. Fingerprints of VP2 on vvibdvs have to be considered more as common evolutionary markers than as virulence markers. The search for such markers is in progress. Pathogenesis of the disease is still poorly understood, and cytokines might play a crucial role in the onset of the disease and in the development of immunosuppression. Mechanisms such as apoptosis and necrosis have been described in lymphoid organs and are involved in the severity of the disease. Macrophages, especially, could play a specific role in the acute phase. Classical serotype 1 vaccines still induce good protection, but the actual problem for control of the disease has became the interference of maternally derived antibody in the establishment of the vaccination schedule. The development of safe vaccines that could either transmit a high passive immunity which could protect broilers during the whole growing period or prime an immune response before or at hatching in the presence of passive immunity might be established in the near future. In this context, recombinant vaccines and virus-neutralizing factor technology might have an advantage over other approaches. Introduction Infectious bursal disease (IBD) has been a great concern for the poultry industry for a long time, but particularly for the past decade. Indeed, its reemergence in variant or highly virulent forms has been the cause of significant economic losses. Until 1987, the strains of virus were of low virulence, causing less than 2% specific mortality, and satisfactorily controlled by vaccination. But in 1986 and 1987, vaccination failures were described in different parts of the world. In the US, it was demonstrated that the new isolates had been affected by antigenic drift against which classical IBD virus (IBDV) vaccines were not satisfactorily protective (Jackwood & Sa õ f, 1987; Snyder et al., 1992), whereas in Europe, the first cases of acute IBDV were described (Chettle et al., 1989; van den Berg et al., 1991). Surprisingly, some of these first acute outbreaks occurred in broilers, at the end of the fattening period, at farms where all the hygienic and prophylactic measures had been taken. These findings indicated a dramatic change in the field situation. Although no antigenic drift was detected, strains of increased virulence were identified. This sudden onset of hypervirulent IBD created the need for a better characterization of the circulating strains so that, in the future, the vaccination * Tel: ; Fax: ; thvan@var.fgov.be Accepted 20 January ISSN (print)/issn (online)/00/ Houghton Trust Ltd

3 176 T.P. van den Berg Figure 1. Structure of IBDV. Electron microscopy of (a) IBDV particles; (b) tubule-like structures obtained when VP2a is expressed via the baculovirus system: (c) virus-like particle (VLP) obtained when the polyprotein is expressed via the baculovirus system; (d) VLPs obtained when the polyprotein is expressed via the vaccinia virus system (kind gift of Paco Rodriguez, CSIC, Madrid, Spain). schedule could be adapted faster to a new epidemiological situation. Further improvement in the control of the disease will only be realized through a better understanding of the viral structure and the mechanisms of pathogenesis. During the 63rd General Session of the Office International des Epizooties (OIE, Paris, 15 to 19 May 1995), it was estimated that IBD has considerable socio-economic importance at the international level, as the disease is present in more than 95% of the Member Countries (Eterradossi, 1995). In this survey, 80% of the countries reported the occurrence of acute clinical cases. Although isolates from different countries have been examined, the current typing confuses antigenic and pathogenic criteria. The situation is sufficiently unclear as to require more extensive comparative studies. Particularly, the absence of known markers to easily characterize very pathogenic viral strains is a serious hindrance, preventing early detection and application of specific prophylactic measures as soon as they appear. Moreover, there is a wide variation in disease control procedures that seldom conform to a specific or standard plan. These features justified the elaboration of a specific resolution (Resolution XVIII in 1995). The European picture has been dominated for a decade by the emergence of very virulent (vv) IBDV strains of infectious bursal disease. These strains have now spread all over the world. Therefore, this review will focus on the acute form of the disease, referring to outbreaks due to vvibdv, as proposed by Stuart (1989) in his letter. Properties of the Virus Several structures and sequences are essential for the viability of IBDV, while others are specific for strains and types, including serotypes and pathotypes. Each modification in the genetic make-up of these structural and regulatory proteins and/or sequences could influence the viral cycle, the host specificity and the virulence of the strain. In order to understand the molecular basis of virulence, a better knowledge of the natural life cycle of IBDV is necessary. Three years ago, Nagarajan & Kibenge (1997) reviewed the viral genome structure, organization and replication. Interesting recent progress has been made in the structure of the virus and, therefore, this section will focus on viral morphogenesis. Structure of the virus IBDV is a small, non-enveloped virus, belonging to the family Birnaviridae, which is characterized by a bisegmented dsrna genome (Kibenge et al., 1988). The prototype of the family is infectious pancreatic necrosis virus (IPNV) of fish. Other members of the family can be found in insects (Drosophila X virus, DXV) and bivalve molluscs (tellina and oyster virus). The virion has a single capsid shell of icosahedral symmetry composed of 32 capsomers and a diameter of 60 to 70 nm (Figure 1a). The viral genome structure and organization are presented in Figure 2. The larger segment A (approximately 3400 base pairs (bp)) contains two open reading frames (ORF). The larger ORF of segment A is monocistronic and encodes a polyprotein that is auto-processed after several steps into mature VP2, VP3 and VP4 (Müller & Becht, 1982; Azad et al., 1985, 1987; Hudson et al., 1986; Kibenge et al., 1997). Segment A can also encode VP5, a short 17 kda protein, from a short, partially overlapping ORF (Mundt et al., 1995). The smaller genome segment B (approximately 2800 bp) encodes VP1, the viral RNA polymerase of 90 kda (Müller & Nitschke, 1987; Spies et al., 1987).

4 Review of infectious bursal disease 177 Figure 2. (a) Genomic organization of IBDV; (b) post-translational modifications in the open reading frame of segment A; j, Regulatory region; NC, non-coding sequence. Morphogenesis The external surface of the virion is composed of trimeric sub-units formed by VP2 and the inner capsid is built of trimeric subunits formed by VP3 (Böttcher et al., 1997; Lombardo et al., 1999). The positively charged C terminus of VP3 might interact with the dsrna genome (Hudson et al., 1986; Böttcher et al., 1997). Infection of mammalian or insect cells with recombinant viral vectors like vaccinia virus (Fernandez-Arias et al., 1998; Lombardo et al., 1999) and baculovirus (Hu et al., 1999; Kibenge et al., 1999; Martinez-Torrecuadrada et al., 1999), which express different parts of the IBDV genome, have given more insight into the structure and components of the virus. Single expression of VP3 results in massive expression of the protein but no structure could be detected. The expression of VP2a alone, not VP2b, leads to the formation of tubule-like structures (Figure 1b). Only expression of the polyprotein gives the formation of virus-like particles (VLPs) with size and shape very similar to those of authentic IBDV particles (Figure 1c,d). Moreover, the final processing of VP2a into VP2b does not occur in VLPs that remain in the cell cytoplasm, but only in IBDV particles. This reinforces the assumption that the final processing of VP2 might be associated with the last steps of the viral cycle (maturation or release process; Müller & Becht, 1982; Kibenge et al., 1999). Viral proteins VP1, the RNA-dependent RNA polymerase of the virus, is present in small amounts in the virion, both as a free polypeptide and as a genome-linked

5 178 T.P. van den Berg Figure 3. Worldwide geographical distribution of the acute forms of IBDV (updated from Eterradossi, 1995). In gray, countries where acute forms have been reported. In black, countries where no acute forms have been reported. In white, countries with no report. protein (Müller & Nitschke, 1987; Kibenge & Dhama, 1997). It plays a key-role in the encapsidation of the viral particles (Lombardo et al., 1999). VP2 has long been identified as the hostprotective antigen as it contains the antigenic region responsible for the induction of neutralizing antibodies and for serotype specificity (Fahey et al., 1989). This protein is highly hydrophobic and conformation dependent, as demonstrated by the observation that all neutralizing monoclonal antibodies (Mabs) react in immunoprecipitation but not in Western blot (Öppling et al., 1991; Schnitzler et al., 1993; van den Berg et al., 1996). VP3 is a group-specific antigen that is recognized by non-neutralizing antibodies, some of which cross-react with both serotypes 1 and 2 (Becht et al., 1988; Öppling et al., 1991). Inoculation of baculovirus-derived VP3 alone failed to induce neutralizing antibodies (Pitcovski et al., 1999). As suggested by Böttcher et al. (1997), VP3 would act as an intermediary, interacting with both VP2 and VP1, and the formation of VP1 VP3 complexes is likely to be an important step in the morphogenesis of IBDV particles (Lombardo et al., 1999; Tacken et al., 2000). VP4 is a fourth, minor and non-structural polypeptide. It is involved in the auto-processing of the polyprotein as a virus-encoded protease producing VP2a, VP3 and VP4 itself (Azad et al., 1987). Although VP4 has little homology with any other known protease, a specific proteolytic activity could be demonstrated (Hudson et al., 1986; Kibenge et al., 1997). The amino acids responsible for this proteolytic activity have been recently characterized as a serine-lysine catalytic dyade (Birghan et al., 2000). In addition, Sanchez & Rodriguez (1999) have carried out a systematic analysis, using a series of plasmids encoding polyproteins containing either deletions or single amino acid substitutions, to identify the processing sites. Electron microscopy of density gradient purified viral preparations using a specific anti-vp4 Mab has shown that VP4 is associated with the formation of specific microtubules present in infected cells and that it is not a constituent of the mature virion (Granzow et al., 1997). This was in contrast with previous studies describing this protein as a minor structural component present in mature virion purified by various methods (reviewed in Kibenge et al., 1988). VP5 was first described in IPNV particles (Havarstein et al., 1990) and has been identified recently in IBDV infected cells (Mundt et al., 1995). Its participation in the viral structure could not be demonstrated. This viral protein more likely has a regulatory function and could play a key role in virus release and dissemination (Mundt et al., 1997). Short History and Epidemiology of Acute IBD First described in Europe at the end of the 1980s (Chettle et al., 1989; van den Berg et al., 1991; Eterradossi et al., 1992), the acute forms of the disease were then described in Japan in the early 1990s (Nunoya et al., 1992; Lin et al., 1993), and they have rapidly spread all over Asia and to other major parts of the world (reviewed in Eterradossi, 1995). Since then, they have been isolated in many countries (Figure 3) including Central Europe (Savic et al., 1997) and Russia (Scherbakova et al., 1998), the Middle East, South America (Di Fabio et al., 1999) and Asia (Cao et al., 1998; Chen et al., 1998; To et al., 1999). On the other hand, Australia, New Zealand, Canada and the US are so far unaffected (Snyder, 1990; Proffitt et al., 1999; Sapats & Ignjatovic, 2000). Moreover, only a sporadic severe outbreak has been described in Finland (Nevalainen et al., 1999), whereas the other northern European countries are still free (Czifra & Janson, 1999).

6 Review of infectious bursal disease 179 Figure 4. Generalized phylogenetic tree of IBDVs (non-exhaustive) based on the comparison of vvp2 nucleotide sequences and simplified from Yamaguchi et al. (1997), Cao et al. (1998), Chen et al. (1998), Eterradossi et al. (1999), Sapats & Ignjatovic (2000) and Zierenberg et al. (2000). Branch lengths have no particular meaning.

7 180 T.P. van den Berg Figure 5. Molecular basis of the antigenicity of IBDV. Neutralizing antibodies have been shown to bind to VP2, within a minimal region, called the variable domain or vvp2, which is highly hydrophobic with a small hydrophilic region at each terminus. These two hydrophilic peaks are located at the surface of the virus and constitute the neutralizing epitopes. Important changes in the sequences of these peaks might determine an antigenic shift as described for serotype 2 viruses. Point mutations inside or outside these peaks might produce an antigenic drift, giving raise to sub-types such as the serotype 1 variant strains described in the US. Molecular epidemiology Due to the high mutation rate in the VP2 variable domain (vvp2) sequence, comparison of this region among strains offers the best evolutionary clue for IBDVs (Figure 4). These studies, together with epidemiological observations and mortality studies, clearly suggest that vvibdv strains belong to the same genetic lineage (Brown et al., 1994, van den Berg et al., 1996; Yamaguchi et al., 1997; Eterradossi et al., 1997b). The first published sequence, strain UK661, is now considered as the reference strain for European vvibdvs (Brown & Skinner, 1996). The Asiatic very virulent strains were probably derived from Europe and then spread throughout Asia in an extremely explosive and conserved manner (Lin et al., 1993; Yamaguchi et al., 1997; Cao et al., 1998; Chen et al., 1998; To et al., 1999). Moreover, some recent phylogenetic analyses performed on the vvp2 sequences of vvibdv strains isolated in Africa in the late 1980s (Eterradossi et al., 1999; Zierenberg et al., 2000) demonstrated that they belong to the common very virulent lineage. There are, however, significant distances between these strains and the European and Asiatic ones, indicating independent evolution. Taken together, all these data might indicate the possible emergence of all vvibdv from an unique event and, hence, a common ancestor. However, comparison of total viral genome sequences should be performed for a more detailed analysis of the spatio-temporal relationships among strains. Changes in vvp2 have to be considered as a common evolution, not as a virulence marker, and the occurrence of new and diverging lineages of vvibdvs should not be excluded in the future. Origin and phylogeny The question of the origin of vvibdv is still open. Phylogenetic analyses performed on segment A of vvibdvs (Brown & Skinner, 1996; Yamaguchi et al., 1997; Pitcovski et al., 1998) confirm that they constitute a specific cluster and that they are more closely related to classical virulent strains, e.g. 52/70, than to other lineages. On the other hand, the topology tree performed on segment B is quite different, indicating that a genetic re-assortment from an unidentified reservoir (wild birds, fish or insects) might have played an important role in the emergence of hypervirulent strains (Howie & Thorsen, 1981; Lasher & Shane, 1994; Yamaguchi et al., 1997). Moreover, although no data on viral shedding have been reported, serological surveys in wild birds (Wilcox et al., 1983; Gardner et al., 1997; Ogawa et al., 1998b) suggest their possible role as a reservoir. Finally, the possible existence of asymptomatic carriers or latently infected birds should also be considered. Antigenic and Pathotypic Variation The high mutation rate of the RNA polymerase of RNA viruses generates a genetic diversification that could lead to emergence in the field of viruses, with new properties allowing them to persist in immune populations. In the case of IBDV, these mutations lead to antigenic variation and modification in virulence in vivo and attenuation in vitro. Antigenic variation Two serotypes of IBDV are described and distinguished by cross-neutralization and cross-protec-

8 Review of infectious bursal disease 181 Figure 6. Classification of IBDV strains as pathotypes. IBDV strains can be defined as apathogenic (serotype 2); mild, intermediate or hot (serotype 1 vaccines); classical virulent (IBDV), variant, or very virulent (serotype 1). Serotype 2 strains cause neither mortality nor bursal lesions in specified pathogen free birds. Serotype 1 vaccines cause no mortality but possess residual pathogenicity with bursal lesions varying from mild to moderate or even severe. Virulent serotype 1 strains induce both mortality and bursal lesions. tion tests. Antigenic variation among serotype 1 isolates of IBDV has been shown in the US since These antigenic variants were of different subtypes compared with classical strains, as determined by serum neutralization tests, and could be antigenically differentiated by the use of a selected panel of neutralizing monoclonal antibodies (Snyder et al., 1992). Even though only one of these subtypes could be considered as truly variant based on cross-protection experiments (Jackwood & Saif, 1987), important economic losses have been sustained due to the emergence of these antigenic mutants. Neutralizing Mabs have been shown to bind to VP2, within a minimal region called the variable domain between amino acids 206 and 350, which is highly hydrophobic with a small hydrophilic region present at each terminus (Bayliss et al., 1990). Sequencing of the VP2 gene of numerous different IBDV strains and selection of escape mutants have proven that this variable domain represents the molecular basis of antigenic variation (Figure 5) (Öppling et al., 1991; Schnitzler et al., 1993; van den Berg et al., 1994a; Vakharia et al., 1994b). Vaccination failures due to vvibdvs have caused great concern for possible antigenic variation among the recent isolates. There is no evidence of antigenic variation in the very virulent strains as described in the US: they belong to classical serotype 1 (van der Marel et al., 1990; van den Berg et al, 1991; Eterradossi et al., 1992). Nevertheless, a modified epitope could be identified on all the vvibdvs tested by Eterradossi et al. (1997b) by the use of a panel of neutralizing Mabs. This corresponded to a mutation of amino acid at position 222 (numbering following Bayliss et al., 1990) that is located in the first hydrophilic peak, as demonstrated by the selection of an escape mutant. Anyway, no drift could be demonstrated by crossneutralization tests (Eterradossi et al., 1998). Other amino acid changes have been shown in the hydrophilic peaks of the variable domain in vvibdvs but their antigenic relevance and epidemiological significance is questionable. For instance, in China, where poultry is one of the fundamental industries of animal production, there have been recent molecular indications for the emergence of variant very virulent strains (Cao et al., 1998) but their biological and epidemiologica l relevance still needs to be established. In France, during their monitoring of the field, Eterradossi et al. (1998) have also shown atypical antigenicity in some vvibdvs due to critical amino acid changes in the second hydrophilic peak, but these strains were not shown to replace the more typical prevalent vvibdvs. All these observations indicate that vvibdvs are evolving but, in contrast to the US, where a change in the field situation was demonstrated, the biological significance of several antigenic differences has to be demonstrated by cross-neutralization tests. Moreover, molecular investigations must be related to the field situation, with a good characterization of the circulating strains in terms of prevalence and virulence. Pathotypic variation In addition to antigenic differences in serotypes and subtypes, the viral strains can also be classified according to their virulence (Figure 6). But there has been a great deal of confusion in these definitions. In particular, the term very virulent has been used to describe both European hypervirulent strains and variant American strains that cause less than 5% mortality but are able to multiply to a higher degree in the bursa of Fabricius of vaccinated animals. In the absence of the identification of specific virulence determinants, the only valuable criteria for the classification of

9 182 T.P. van den Berg Figure 7. The reverse genetics system is based on the full-length cdna cloning of the IBDV segments in a vector producing fulllength RNAs and their subsequent transfection into eukaryotic cells, allowing the generation of completely new synthetic viruses (Mundt & Vakharia, 1996). The synthesis of a re-assortant between segment A of a serotype 1 strain and segment B of a serotype 2 strain is illustrated. This allows the directed mutagenesis of cdnas prior to reverse genetics in order to study the effect of any mutation. IBDV strains as pathotypes should refer to their virulence (mortality or lesions) in 3- to 6-week-old specific pathogen free birds and not to any antigenic specificity. Virulence markers The search for virulence markers, now considered to be the holy grail by most IBDV researchers, is still in progress, and no biological or molecular structure has been identified as responsible for the virulence of IBDV strains. The finding of a molecular marker like, for instance, the presence of basic residues in the cleavage site of the fusion protein of Newcastle disease virus (Collins et al., 1993; Kant et al., 1997; Oberdörfer & Werner, 1998) can only progress through a better knowledge of the viral structure and infectious cycle. While it has become obvious that amino acids within the variable region of VP2 represent the molecular basis for antigenic variation, no definite hot spot that determines pathogenicity has been identified. As demonstrated by the selection of re-assortant IBDV strains that possessed segment A of the pathogenic serotype 1 strain Cu-1 and segment B from serotype 2 strain 23/82, pathogenicity is not driven by one of the two genomic segments. Both segments contribute to the replication in the bursa of Fabricius and virulence (Müller et al., 1992; Mundt, 1999). Sequence comparisons between pathogenic and non-pathogenic serotype 1 strains showed nucleotide changes throughout the genome on segments A and B (Brown & Skinner, 1996; Yamaguchi et al., 1997; Pitcovski et al., 1998; Yehuda et al., 1999), indicating that nucleotide changes in different areas of the genome probably contribute to a multigenic nature of virulence. Since the two genomic segments of IBDV are relatively short, one may expect that additional sequences will be available in the near future that will allow the mapping of the principal pathogenicity determinants. Moreover, the development of the reverse genetics system (Figure 7), based on the full-length cdna cloning of the IBDV segments in vector producing full-length RNAs, and subsequent transfection in eukaryotic cells, allows the generation of completely new synthetic viruses (Mundt & Vakharia, 1996; Mundt et al., 1997; Yao et al., 1998; Boot et al., 1999). More recently, simplifications of the method have been proposed for the

10 generation of synthetic particles by direct transfection of cdna vector into chick embryo fibroblast cells (Lim et al., 1999) and by improved methods of reverse transcription, polymerase chain reaction (PCR) and cloning of full-length segments of both strands of IBDV (Akin et al., 1999, Boot et al., 2000). The genetic engineering of re-assortants, recombinants or mutants will be of considerable help to elucidate the role of segments, genes, regions or even single amino acids in the disease. Attenuation and adaptation to cell culture In the same context, a characteristic of serotype 1 field strains, especially vvibdvs, is their inability to grow in cell culture. Adaptation requires several blind passages in cell culture or embryonated eggs and leads to attenuation for chickens (Spies et al., 1987; Yamaguchi et al., 1996a). This property is used for the production of live vaccine strains. In order to identify the sequences involved in attenuation and tissue culture adaptation, comparison of the nucleotide sequences of wild type and its attenuated counterpart has been performed by different groups with classical (Müller et al., 1992; Hassan et al., 1996) or hypervirulent (Yamaguchi et al., 1996b) IBDV strains. Sequence comparisons confirmed a multigenic nature of attenuation, as mutations are located in different parts of the genome. Taking into account that the complete 5 and 3 termini of the adapted strains have not been fully determined, the significance of each change for attenuation remains to be established. Study of infection at the level of virus binding is also important for understanding the virus host cell interactions and subsequent pathogenesis of the disease (Nieper & Müller, 1996; Ogawa et al., 1998a). As previously mentioned, two types of serotype 1 viruses can be classified on the basis of their ability to infect and replicate in cultured cells and/or in the B lymphocytes of the bursa of Fabricius. Several amino-acid exchanges in vvp2 have been identified, using the reverse genetics system, as being responsible for cell culture adaptation (Lim et al., 1999; Mundt, 1999). Although these findings might be an indication of a possible role of VP2 in virulence, this is probably limited to cell tropism, as each modification in the fitness of VP2 for its target cell might increase infectivity. Diagnosis of Acute IBD and Characteristics of vvibdvs Symptomatology and lesions Hypervirulent IBDV infections are characterized by severe clinical signs and high mortality. Indeed, the vvibdvs produce disease signs similar to conventional type 1 infection, with the same incubation period (4 days), but the acute phase is exacerbated and more generalized in the affected Review of infectious bursal disease 183 flock. Severe outbreaks are characterized by sudden onset of depression in susceptible flocks. Animals in the acute phase of the disease are prostrate and reluctant to move, with ruffled feathers and frequently watery or white diarrhoea. The age susceptibility is extended, covering the entire growing period in broilers, and the peaks of mortality show a sharp death curve followed by rapid recovery (Chettle et al., 1989; van den Berg et al., 1991; Nunoya et al., 1992; Tsukamoto et al., 1992). On post mortem examination of birds that died during the acute phase of vvibd, the bursa of Fabricius is the principal diagnostic organ: it is turgid, oedematous, sometimes haemorrhagic and turns atrophic within 7 to 10 days. This atrophy might be more rapid, even 3 to 4 days after inoculation (Tsukamoto et al., 1992). In addition, dehydration and nephrosis with swollen kidneys are common, and ecchymotic haemorrhages in the muscle and the mucosa of the proventriculus are observed in the majority of the affected birds. Severe depletion of lymphoid cells is observed not only in the bursa of Fabricius, but also in the non-bursal lymphoid tissues. Pathogenicity of IBDV has been associated with virus distribution in non-bursal lymphopoietic and haematopoietic organs. Indeed, using various immunostainin g methods, a higher frequency of antigen-positiv e cells could be demonstrated after infection of birds with vvibdv compared with other strains, in the thymus (Nunoya et al., 1992; Sharma et al., 1993; Inoue et al., 1994), the spleen and the bone marrow (Tanimura et al., 1995; Tsukamoto et al., 1995; Inoue et al, 1999). In particular, atrophy of the thymus has been associated with the acute phase of the disease and might be indicative of the virulence of the isolate, although it is not associated with extensive viral replication in thymic cells (Sharma et al., 1993). An increased number of macrophages are found in various organs (Tanimura et al., 1995). Thrombocytes also represent a target for IBDV, and acute disease is characterized by disseminated haemorrhages probably related to an impairment of the clotting mechanism (Skeeles et al., 1980). Molecular tools for diagnosis Antigenic and molecular similarity among the new vvibdv isolates from different parts of the world is an indication of a common origin and a similar antigenic evolution. Nevertheless, although the marked increase in acute IBD in different parts of the world dominates the field picture, strains of different virulence still co-exist, warranting the need for a rapid discrimination between circulating strains. So far, no Mab specific for the very virulent strains have been obtained but, as previously mentioned, a modified neutralizing epitope has been identified by Eterradossi et al. (1997a). This modification is present on all tested vvibdvs, and

11 184 T.P. van den Berg the usefulness of such a marker for epidemiologica l investigations is considerable. On the other hand, vvp2 sequence can also be used as a molecular marker, and generalization of the molecular testing like reverse transcription (RT)-PCR followed by restriction enzyme digestion or restriction fragment length polymorphism (RFLP) analysis of the amplified fragment might also prove very helpful in the near future (Jackwood & Sommer, 1999). These approaches, however, are less related to the biological properties of the virus, as mutations at nucleotide level are not subjected to the same selection pressure as amino acids. Only time and the accumulation of sequences will prove the usefulness of such approaches. So far, RT-PCR techniques on selected fragments of the genome (essentially the variable domain of VP2) have to be followed by sequencing and phylogenetic comparison. This represents actually the only valuable molecular alternative for the classification of IBDV strains. Such alignments have shown that vvibdvs share unique amino acid residues at positions 242A, 256I and 294I (numbering following Bayliss et al., 1990) that might represent the genetic fingerprints of vvibdv (Yamaguchi et al., 1997; Cao et al., 1998; Eterradossi et al., 1999). Recently, Cardoso et al. (2000) have shown that the Lukert strain of IBDV can be grown in the chicken embryo rough (CER) cell line (Smith et al., 1997). This may be an advantage in a diagnostic laboratory if it were to be shown that field isolates could grow readily in this cell line, as CER cells may also be used for the propagation of avian pneumovirus (Arns & Hafez, 1995; Dani et al., 1999). Vero and other mammalian cell lines, e.g. baby grivet monkey kidney cells, have also been used to grow IBDV (Jackwood et al., 1987; Cardoso et al., 1998). Pathogenesis and Immunosuppression Pathogenesis can be defined as the method used by the virus to cause injury to the host with mortality, disease or immunosuppression as a consequence. These injuries can be evaluated at different levels: the host, the organ and the cell, and are exacerbated in the acute forms of the disease. The selected host of the virus is young chickens where a clinical disease occurs, while in older birds the infection is essentially subclinical. Susceptibility of different breeds has been described with higher mortality rates in light than in heavier breeds (Bumstead et al., 1993; Nielsen et al., 1998). Inoculation of IBDV in other avian species fails to induce disease (McFerran, 1993). The target organ of IBDV is the bursa of Fabricius at its maximum development, which is a specific source for B lymphocytes in avian species. Bursectomy can prevent illness in chicks infected with virulent virus (Hiraga et al., 1994). The severity of the disease is directly related to the number of susceptible cells present in the bursa of Fabricius; therefore, the highest age susceptibility is between 3 and 6 weeks, when the bursa of Fabricius is at its maximum development. This age susceptibility is broader in the case of vvibdv strains (van den Berg et al., 1991; Nunoya et al., 1992). After oral infection or inhalation, the virus replicates primarily in the lymphocytes and macrophages of the gut-associated tissues. Then virus travels to the bursa via the blood stream, where replication will occur. By 13 h post-inoculatio n (p.i.), most follicles are positive for virus and by 16 h p.i., a second and pronounced viraemia occurs with secondary replication in other organs leading to disease and death (Müller et al., 1979). Similar kinetics is observed for vvibdvs but replication at each step is amplified. Actively dividing, surface immunoglobuli n M-bearing B cells are lysed by infection (Hirai & Calnek, 1979; Hirai et al., 1981; Rodenberg et al., 1994), but cells of the monocyte macrophage lineage can be infected in a persistent and productive manner, and play a crucial role in dissemination of the virus (Burkhardt & Müller, 1987; Inoue et al., 1992; van den Berg et al., 1994b) and in the onset of the disease (Sharma & Lee, 1983; Kim et al., 1998; Lam, 1998). Indeed, the exact cause of clinical disease and death is still unclear but does not seem to be related only to the severity of the lesions and the bursal damage. Indeed, after infection, some birds with few bursal lesions can be found dead, while others can survive despite extensive bursal damage. Moreover, mortality rates are often variable and the establishment of median lethal dose for standardization has always been hazardous. In addition, the narrow age range for susceptibility to clinical disease has not yet been clearly explained. Prostration (with ruffled feathers, diarrhoea and inappetence) preceding death is very similar to what is observed in acute coccidiosis, and is reminiscent of a septic shock syndrome (Figure 8). The macrophage could play a specific role in this pathology by an exacerbated release of cytokines such as tumor necrosis factor or interleukin 6 (Kim et al., 1998). However, an intermediate role of TH cells in this pathophysiological mechanism should also be considered (Tanimura & Sharma, 1997; Vervelde & Davison, 1997). As chicken macrophages are known to be activated by interferon (Dijkmans et al., 1990), this role could occur through an increased secretion of interferon as has been described in vitro after infection of chicken embryo cultures or in vivo in chicken (Gelb et al., 1979a,b). Depletion of lymphoid cells in the bursa of Fabricius after IBDV infection is due to both necrosis and apoptosis. Apoptosis, or programmed cell death, is a process where, in response to specific stimuli, cells die in a controlled, programmed manner. Many different cell species can undergo apoptosis but immature B and T cells are particularly susceptible to apoptotic cell death.

12 Review of infectious bursal disease 185 Figure 8. The septic shock syndrome or cytokine storm. (a) Sepsis is a systemic clinical situation caused by toxic substances released by microorganisms during severe infection and coincides with a rapid increase in circulating levels of inflammatory cytokines such as tumour necrosis factor alpha (TNFa ), gamma interferon (IFNg ), and interleukins IL8 and IL6. (b) Macrophages can be activated either directly (persistent infection by IBDV) or indirectly (stimulation of T cells to secrete high levels of IFN (that activates macrophages) and produce apoptotic mediators, e.g. NO or TNFa. (c) Inflammatory mediators identified in the chicken with tests and references.

13 186 T.P. van den Berg Figure 9. Measurement of immunosuppression. Although the immunosuppression caused by IBDV is principally directed towards B lymphocytes, an effect on cell-mediated immunity has also been demonstrated. Therefore, immunosuppression can be measured in vitro by using proliferation tests or by measuring cytokine (ChIFNg ) release after mitogen activation of T cells using either the HD11 biological assay or a specific capture enzyme-linked immunosorbent assay (ELISA) for chicken IFNg. Apoptosis is usually initiated by a variety of physiological stimuli, although pathological stimuli, such as viral infections, can also trigger the phenomenon. Recent studies have shown that immunosuppression induced by IBDV is caused, at least in part, by apoptosis (Vasconcelos & Lam, 1994; Ojeda et al., 1997; Tanimura & Sharma, 1998, Nieper et al., 1999). A direct effect of viral proteins like VP2 and VP5 has been implicated in the induction of the mechanism (Fernandez-Arias et al., 1997; Yao et al., 1998) but further investigations are needed to establish their exact role in pathogenesis and immunosuppression, notably by comparing them in strains with different virulence. On the other hand, apoptotic cells have also been observed in viral antigen-negative bursal cells (Tanimura & Sharma, 1998; Nieper et al., 1999), reinforcing the possible role of immunological mediators in the process. Recovery from disease or subclinical infection will be followed by immunosuppression with more serious consequences if infection occurs early in life. Although the immunosuppression caused by IBDV is principally directed towards B lymphocytes, an effect on cell-mediated immunity (CMI) has also been demonstrated (Sharma & Fredricksen, 1987; Sharma et al., 1989; Cloud et al., 1992a,b). Mechanisms such as the development of suppressor cells and the impairment of helper T cells have been suggested (Sharma & Fredricksen, 1987; Vervelde & Davison, 1997). This can be demonstrated in vitro by using proliferation tests (Confer et al., 1981, Confer & MacWilliams, 1982; Sharma & Lee, 1983; Karaca et al., 1996; McNeilly et al., 1999) or by measuring cytokine release after mitogen activation of T cells (Lambrecht et al., 2000) (Figure 9). Prevention and Control Due to the high resistance of IBDV to environmental exposure and its wide distribution, hygienic measures alone, while essential, are often insufficient. Vaccination is thus essential (reviewed in Lütticken, 1997). The economic impact of both clinical and sub-clinical diseases warrants the search for and the use of efficient vaccines. While the protective role of cellular immunity cannot be ruled out, good protection is achieved by the induction of neutralizing antibodies, as proven by the excellent passive protection of young chicks against infection. This satisfactory protection can be achieved by immunization with live or inactivated vaccines. Classical live vaccines achieve lifelong and broad protection, but possess residual pathogenicity and a proportional risk of reversion to virulence. Inactivated vaccines, although costly, were used successfully until the emergence of the hypervirulent strains. Indeed, it was a normal practice in broiler production to vaccinate hens with an oil-emulsion vaccine just before laying in order to induce a high level of passive immunity in the offspring, which could protect them until an age where infection is less detrimental with regards to immunosuppression (Box, 1989). This procedure was satisfactory until the emergence of vvibdvs when all classical prophylactic measures were called into question. The first cases occurred in flocks where all hygienic measures and vaccinations had been properly applied. It was no longer possible to protect broilers passively during the whole growing period and a live vaccination became necessary. But the interference of maternally derived antibody (MDA) (see Figure 10) became the crucial problem in establishment of the vaccination schedule. Serological monitoring is usually necessary to determine the optimal timing for vaccination (van den Berg & Meulemans, 1991; Kouwenhoven & van den Bos, 1994). In this context, the development of tests allowing the differentiation between passive (antibody-positive, CMI-negative) and active immunity (antibodypositive, CMI-positive) could be of considerable help (Lambrecht et al., 2000). Inactivated vaccines might prove helpful if they can induce higher antibody levels in breeders, which will then be passively transmitted to the offspring and protect them during their entire growing period. Subunit IBDV proteins expressed in yeast (Macreadie et al., 1990) or via the baculovirus system (Vakharia et al., 1994a; van den Berg et al., 1994a; Pitcovski et al., 1996; Dybing & Jackwood, 1998; Yehuda et al., 2000) might help to reach this goal. Another advantage of these technologies is that a vaccine based on VP2 alone should allow monitoring of the field situation by the

14 Review of infectious bursal disease 187 Figure 10. Interference of MDA in the establishment of the vaccination schedule. (a) A good vaccine will be in balance regarding safety and potency. Attenuated live vaccines achieve lifelong and broad protection, but possess a residual pathogenicity and a risk of reversion to virulence. Inactivated vaccines are safe but more expensive as more antigen will be necessary to induce a satisfactory response. This situation is complicated by the passive transmission of MDA from dams to the offspring via the egg. This passive immunity, although protective, will interfere with vaccination. (b) There is a strong competition between field and vaccine strains to break through MDA, and the optimal timing has became the crucial problem in establishment of the vaccination schedule. At the flock level, this situation is particularly hazardous as immune and susceptible birds coexist. In this context, recombinant vaccine virus like the herpes virus of turkey (HVT) and virus neutralizing factor (VNF) technology might have an advantage over other approaches. discrimination between vaccinial (anti-vp2 only) and infectious antibody (anti-vp2 and VP3). There is no evidence of antigenic variation in the very virulent strains as described in the US: they belong to classical serotype 1 (van der Marel et al., 1990; van den Berg et al., 1991; Eterradossi et al., 1992) and, therefore, they can be controlled adequately under experimental conditions by vaccination with commercial vaccines prepared from classical attenuated strains (Öppling et al., 1991; van den Berg & Meulemans, 1991; Eterradossi et al., 1992). Live vaccines have another advantage in that they are excreted in the environment, where they can compete with field viruses. Unfortunately, most intermediate vaccines are inadequate for interfering with vvibdvs that could break through

15 188 T.P. van den Berg Table 1. Two European Union concerted actions in the field of IBDV (a) COST a Action 839 on Immunosuppressive Viral Diseases of Poultry Definition Objective Participating countries Tasks (five working groups) Framework for scientific and technical co-operation allowing the coordination of national research at an European level To contribute to reduce the economic impact of both clinical and subclinical forms of the Immunosuppressive Viral Diseases of Poultry (IBDV and CAV) Austria, Belgium, Croatia, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Norway, Poland, Spain, Sweden, UK WG1: Epidemiology WG2: Diagnosis and Economic Impact WG3: Vaccination WG4: Pathogenesis WG5: Molecular Virology Duration 5 years (1998 to 2003) Useful addresses < < < (b) INCO-DC b Action 97 on Acute Infectious Bursal Disease of Poultry Definition Objective Participating countries Tasks Duration 3 years (1998 to 2001) Useful addresses Concerted action in the frame of the Cooperation with Third Countries and International Organizations research and technological development programme To increase knowledge of the epidemiology of IBD by establishing common systems of diagnosis and epidemiosurveillance in order to reveal the incidence and prevalence of the different forms of IBD Asia: China (four partners); Europe: Belgium, France, Germany; and three associated partners: Bangladesh, India, Indonesia 1. Selection and characterization of reference material 2. Genomic data bank 3. Studies on pathogenesis 4. Construction of clones by the reverse genetics system 5. Vaccination trials < < a COST, Cooperation in the field of Scientific and Technological research. b INCO-DC, International Cooperation with Developing Countries. higher MDA levels. However, the use of less attenuated ( hot ) vaccines, even with an acceptable reduction of mortality, is dangerous as these vaccines induce immunosuppression and carry the risk of reversion to virulence (Guittet et al., 1992). Although the reverse genetics system (Mundt & Vakharia, 1996) will represent a basis for the genetic attenuation of strains and for the generation of new vaccines, interference of passive immunity will still exist. Therefore, as they are less sensitive to neutralization by anti-ibdv MDA, recombinant viral vaccines expressing the VP2 protein of IBDV, such as fowl pox virus (Bayliss et al., 1991; Heine & Boyle, 1993), herpes virus of turkey (HVT) (Darteil et al., 1995; Tsukamoto et al., 1999) or fowl adenovirus (Sheppard et al., 1998) might prove to be powerful in the near future to prime an active immune response. In this context, DNA vaccines could also be considered but some limitations, especially cost, individual variability in the response and general low humoral response might limit these vaccines to laboratory investigation s (Fodor et al., 1999; van den Berg et al., 1999). The efficacy and safety of HVT vaccines in ovo have now been demonstrated (Johnston et al., 1997) and might give this recombinant technology an advantage over the other approaches. Recently, a new concept, which consists of the in ovo inoculation of a virus antibody complex vaccine, has emerged (Haddad et al., 1997). This novel technology utilizes specific hyperimmune neutralizing antiserum (or virus neutralizing factor (VNF)) with a vaccine virus under conditions that are not sufficient to neutralize the vaccine virus but which are sufficient for delaying the pathological effects of the vaccine alone. This allows young chicks to be

16 vaccinated more effectively in the presence of passive immunity even with a strain that would be too virulent for use in ovo or at hatching. Although some questions still remain concerning the batch variations and risks of contamination of VNF, the residual pathogenicity of the vaccine strain and the mechanisms involved in the delay of the immune response (Jeurissen et al., 1998), this technology is very promising for the future control of IBDV. Conclusions Due to their complexity and to the multifaceted nature of the infections, immunosuppressive viral diseases require a multidisciplinary approach that can probably only be achieved by combining scientific expertise and resources (strains, reagents) from different countries. As stated by the OIE Resolution XVIII in 1995, the first requirement for the progress in the diagnosis and the control of IBD is a coordinated effort among States. For instance, in Europe, several initiatives following the principle of concerted actions, have been built under the auspices of the European Union (COST Action 839 and INCO-DC Action 97; see Table 1). As continuation of the OIE recommendations, expected outcomes from these working groups are a harmonization between states, a co-ordination of the efforts, the standardization of tools and nomenclature, and the elaboration of guidelines and recommendations for future research and current practices. In recent years, as illustrated by the numerous references presented in this review, considerable effort has been made all over the world to understand the disease. Molecular virology and avian immunology have made considerable progress and should generate new tools in the near future. However, additional research will be needed to overcome some of the current obstacles. First of all, the definition of virulence markers should allow the development of specific and more sensitive diagnostic methods warranting a better definition of the epidemiological situation. In this regards, the reverse genetic system, providing the tool to construct chimeric viruses, might be decisive in the identification of virulence markers and the genetic attenuation of strains. A better knowledge of the immunological mechanisms involved in the disease might give tools for the measurement of immunosuppression in the field situation. Together with a better epidemiological definition of the field situation (circulating strains), this should allow a more accurate estimate of the economic impact of immunosuppressive viral disease and a cost/benefits analysis. Moreover, better identification of the protective criteria and differentiation between active and passive immunity might be of considerable help in the establishment of vaccination schedules. Finally, the development of safe vaccines that could prime an immune response before or at hatching in the presence of passive immunity might be established in the near future. Acknowledgements The author is very grateful to Peter Flanagan and William (Bill) Ragland for proof-reading the manuscript, and to Nicolas Eterradossi, Guy Meulemans and Bénédicte Lambrecht for helpful discussions. Gérard Charlier is also acknowledged for the electron microscopy work. References Review of infectious bursal disease 189 Akin A., Wu, C.C. & Lin, T.L. (1999). Amplification and cloning of infectious bursal disease virus genomic RNA segments by long and accurate PCR. Journal of Virological Methods, 82, Arns, C.W. & Hafez, H.M. (1995). Isolation and identification of avian pneumovirus from broiler breeder flocks in Brazil. In Proceedings of the 44th Western Poultry Disease Conference (pp ). USA. Azad, A.A., Barrett, S.A. & Fahey, K.J. (1985). The characterization and molecular cloning of the double-stranded dsrna genome of an Australian strain of infectious bursal disease virus. Virology, 143, Azad, A.A, Jagadish, M.N., Brown, M.A. & Hudson, P.J. (1987). Deletion mapping and expression in E. Coli of the large genomic segment of a birnavirus. Virology, 161, Bayliss, C.D., Spies, U., Shaw, K., Peters, R.W., Papageorgiou, A., Muller, H. & Boursnell, M.E.G. (1990). A comparison of the sequences of segment A of four infectious bursal disease virus strains and identification of a variable region in VP2. Journal of General Virology, 71, Bayliss, C.D., Peters, R.W., Cook, J.K.A., Reece, R.L., Howes, K., Binns, M.M. & Boursnell, M.E.G. (1991). A recombinant fowlpox virus that expresses the VP2 antigen of infectious bursal disease virus induces protection against mortality caused by the virus. Archives of Virology, 120, Becht H., Müller, H. & Müller, H.K. (1988). Comparative studies on structural and antigenic properties of two serotypes of infectious bursal disease virus. Journal of General Virology, 69, Birghan, C., Mundt, E., & Gorbalenya, A.E. (2000). A non-canonica l Lon proteinase lacking the ATPase domain employs the Ser-Lys catalytic dyad to exercise broad control over the life cycle of a double-stranded RNA virus. EMBO Journal, 4, Boot H.J., ter Huurne, A.H.M., Peeters, B.P.H. & Gielkens, A.L.J. (1999). Efficient rescue of Infectious Bursal Disease Virus from cloned cdna: evidence for involvement of the 3-terminal sequence in genome replication. Virology, 265, Boot H.J., ter Huurne, A.H.M. & Peeters, B.P.H. (2000). Generation of full-length cdna of the two genomic dsrna segments of Infectious Bursal Disease virus. Journal of Virological Methods, 84, Böttcher B., Kiselev, N.A., Stel Mashchuk, V.Y., Perevozchikova, N.A., Borisov, A.V. & Crowther, R.A. (1997). Three-dimensional structure of infectious bursal disease virus determined by electron cryomicroscopy. Journal of Virology, 71, Box, P. (1989). High maternal antibodies help chickens beat virulent virus. World Poultry, 53, Brown, M.D. & Skinner, M.A. (1996). Coding sequences of both genome segments of a European very virulent infectious bursal disease virus. Virus Research, 40, Brown, M.D., Green, P. & Skinner, M.A. (1994). VP2 sequences of recent European very virulent isolates of infectious bursal disease virus are closely related to each other but are distinct from those of classical strains. Journal of General Virology, 75, Bumstead, N., Reece, R.L. & Cook, J.K.A. (1993). Genetic differences in susceptibility of chicken lines to infection with infectious bursal disease virus. Poultry Science, 72,