During the last decade, an acute paretic syndrome of unclear origin was increasingly reported from flocks of White Leghorn chickens worldwide . In an effort to elucidate the underlying cause and pathobiology of this syndrome we performed an extensive pathohistological and immunological investigation.
All birds of our collective that presented with limb paresis showed severe demyelination of peripheral nerves, with a predilection for craniospinal nerve roots and associated ganglia. This demyelination was associated with multifocal endoneurial infiltration by lymphocytes, plasma cells and macrophages. TEM provided evidence of macrophages invading myelin spirals at the outer mesaxon, between the paranodal loops and Schmidt-Lanterman incisures, leading to stripping off of myelin lamellae from morphologically intact axons. This picture closely resembles human AIDP, which leads us to denote this disease as avian inflammatory demyelinating polyradiculoneuritis (AvIDP) [46, 47].
Several pathways have been proposed by which the attention of macrophages is directed towards the myelin sheath. Firstly, autoreactive T-helper cells may secret chemokines attracting macrophages to the endoneurium and, subsequently, activate them via macrophage-activating factors. Activated macrophages also are capable of recognising the Fc-region of auto-antibodies that opsonise myelin epitopes and/or activate the complement system . In AvIDP there is a considerable body of evidence for employment of both pathways. Immunohistological, flow cytometrical and gene expression studies confirm the presence of significant numbers of αβ T-cells of both T-helper and cytotoxic T-cell (CTL) phenotypes. Gene expression profiling further indicated a dominance of a Th1-like immune response as demonstrated by a significant upregulation of IFNγ mRNA and no changes in IL-13 mRNA levels in affected nerves in comparison to healthy controls [49, 50]. Th1-like immune responses have been described as taking part in acute EAN and in the initial stages of AIDP; findings which further underline the similarity between AIDP and AvIDP [51, 52].
On the other hand, one may ask if the high density of endoneurial plasma cells and deposition of myelin-bound IgG lend credence to a simultaneous recruitment of humoral effector mechanisms . IL-10 expression suggests a Th-cell-dependency rather than a primary humoral response . Morphological examination suggests, and quantitative PCR confirms, that humoral pathways seem to predominate at the stage of AvIDP disease investigated in this study. B-cell and plasma cell activation is reflected by high mRNA expression levels of chB6 (Bu-1) and B lymphocyte-induced maturation protein-1 (Blimp-1) in the affected spinal ganglia [38, 55]. Interestingly, IL-13 was - despite the indications for a predominantly humoral pathway - not significantly elevated [56, 57], which may be explained by the stage of disease (see below).
The increase in mRNA expression of IL-10 is suggestive of down-regulation of the immune response [54, 58, 59] and it is in line with the observation of a remission of the clinical deficits soon after the paralytic phase in an additional group of animals subjected to long-term trials (unpublished data; see below).
Furthermore, there was no elevation of IL-1 and IL-6, which contrasts with common observations of an acute inflammatory reaction [60, 61]. Even though there is an increase in IFNγ-level, it seems not to have led to the transformation of macrophages into inflammatory macrophages that would have produced higher levels of iNOS in addition to IL-1 and IL-6 [62, 63].
In accord with recent data on AIDP and on rodent GBS models , it may be concluded that the paraparetic birds in this study were just about to leave a Th1-driven initial stage for a Th2-dominated plateau or even recovery phase. It is very likely that the occurrence of neurological deficits coincides with late a Th1-stage or a Th1-to-Th2-transition, which implies the existence of a considerable preclinical period in disease progression. Because of the spontaneous appearance of disease, and the low incidence (1 per 100 animals) at earlier stages of AvIDP would be very difficult to assess through random sampling in preparetic chickens.
Concerning the natural history of the disease, however, we have launched preliminary longitudinal trials in order to clarify whether AvIDP resembles an acute monophasic disease or a chronic progressive, stagnant or even remitting-relapsing disease with sudden onset. Our preliminary observations indicate that remission and relapse of clinical signs is possible.
In GBS and CIDP, several auto-antibodies have been demonstrated to react with myelin proteins and peripheral nerve gangliosides. Auto-antibodies against myelin proteins P0, P2 and PMP-22  are associated with AIDP and CIDP in humans. These proteins are uniformly distributed throughout the PNS and do not explain either the proximodistal gradient of inflammatory demyelination or the variations in degree of involvement amongst different nerves. Ganglioside composition, however, does vary among different cranial nerves . It therefore has been shown in AIDP and CIDP in humans, and in corresponding animal models in laboratory animals, that the profile of anti-ganglioside antibodies may predict the clinical phenotype [66, 67]. Hence, oculomotor nerve involvement is a characteristic feature of the Miller-Fisher syndrome and of GBS with ophthalmoplegia associated with antibodies directed against the ganglioside GQ1b that is particularly abundant in CNIII [65, 68]. The binding partners for IgG in AvIDP remain to be identified but the remarkable involvement of CNIII and CNV render similarly distributed avian gangliosides likely target molecules for immunoaggression, even though Miller-Fisher syndrome and GBS with ophthalmoplegia are axonal diseases [48, 69].
In analogy to EAN models, the main fraction of recruited T-cells carried αβ T-cell receptors . Even though a contribution of γδ T-cells to human IDP has been documented, indicating a role of non-peptide antigens as triggers for autoaggression [71, 72], the prevalence of at least 10% γδ T-cells in AvIDP has to be interpreted with caution since this T-cell population represents about 20 to 50% of the circulating T-cell pool in chickens under physiological conditions . In contrast, humans γδ T-cells represent approximately 1 to 15% of peripheral blood lymphocytes .
So far, the aetiology of AvIDP is still undetermined. Preceding events associated with the onset of GBS range from viral, mycoplasmal and bacterial infections to surgery, vaccination, fever treatment and other stressful conditions [75, 76]. All chickens, affected and unaffected, originated from one single flock and were raised in an identical environment regarding diet, housing and exposure to environmental pathogens. In contrast to SPF-animals, these chickens are exposed to permanent infection pressures that remain to be controlled by tight polyvaccination management.
Thus, an association with preceding infection, as well as with vaccination, may be involved in disease development via molecular mimicry as has been documented in GBS [76, 77].
Outbreaks of viral, mycoplasmal and bacterial infections in chicken flocks are generally limited by strict vaccination programs and by routine health monitoring. However, worldwide distribution, paired with neurotropism and some overlapping clinical and histopathological features, render Marek's disease herpesvirus an important candidate amongst avian infectious agents . Previous studies already have emphasised the similarity of demyelination in Marek's disease (MD) and in EAN [79–81]. Moreover, GBS has been associated with a panel of human herpesviruses, namely Varicella-zoster virus [82–84], cytomegalovirus [84–86], Herpes simplex virus  and more frequently Epstein Barr virus [84–86]. We therefore employed a PCR protocol that allows for the discrimination of MDV vaccine and field strains. The results of this PCR analysis strongly suggest that infections with virulent MDV strains should not be considered as the cause of AIDP.
In addition to preceding infections, vaccinations have been reported risk factors for GBS and CIDP in humans. Anteceding immunizations with vaccines against influenza, hepatitis, measles, mumps, and rubella and others infrequently have been associated with GBS [77, 87–89]. All chickens in this study were vaccinated in the first few weeks post hatching with inactivated or live vaccines against MDV, Newcastle disease virus, infectious bursal disease virus, avian infectious bronchitis virus, Salmonella spp., and coccidiosis. Hence, multivaccination appears a possible immunological trigger. The potential contribution of infections and/or immunizations to disease development is supported by the observation that SPF animals of the same genetic background do not develop AvIDP. Further trials on a population of unvaccinated chickens are mandatory to evaluate the possible role of vaccination programs for AvIDP development.
While external triggers still remain uncertain, we were able to identify a genetic susceptibility factor confined to the avian major histocompatibility complex (MHC), the so called B-complex. The four genotypes studied displayed marked differences in risk of being affected by AvIDP. Most obvious was the increased risk of genotype [261/539] compared to others. In addition, genotype [357/539] also showed higher risk of AvIDP compared to genotype [261/357], and, to a lesser extend, compared to genotype [261/261] as well.
Moreover, the percentage of AvIDP-affected animals was significantly higher for genotypes [261/539] and [357/539] than for the other two haplotype combinations. Results suggest an association of marker LEI0258, located in the MHC region of the chicken, with the occurrence of AvIDP as earlier indicated by Bacon et al. . Thereby, the allele with a fragment size of 539 bp seems to be linked to an elevated risk of developing this disease.
Likewise, in the highly susceptible Lewis rats that are commonly used in EAN trials, a certain allele of a MHC-linked gene - amongst further, non-MHC regions - is necessary in the MHC or RT1 region to confer EAE susceptibility in the F2 progeny . Furthermore, HLA-DR2 is associated with a higher susceptibility to CIDP in humans . In GBS, a genetic background is also suspected but has not yet been confirmed [92, 93].
To date, EAN is the most frequently used animal model for investigation of immunopathological mechanisms in acute inflammatory demyelinating diseases of the PNS . Even though it strongly resembles AIDP histopathologically, there are several disadvantages and dissimilarities to the human disease in terms of CNS involvement - which is very rare in GBS [20, 94] - and a monocausal neuritogenic trigger, that poorly reflects the natural disease development .
Like AIDP and EAN, AvIDP is characterised by infiltration of nerve roots and peripheral nerves with macrophages and lymphocytes and, most importantly, a cell-mediated demyelination [13, 95]. In AvIDP the CNS is not involved at any time, as is described for the vast majority of GBS cases [20, 94].
Compared to experimental immunization with mimicked epitopes, the spontaneous disease development of AvIDP is much closer to the field situation and provides an opportunity to investigate aetiological factors through purposed-based exposure and manipulations of the environment.
A drawback of AvIDP as a disease model lays in the difficulty to identify pre- or subclinical animals and thereby the very early stages of immunopathology, before switching from Th1- to Th2-mediated cascades. However, scientific approaches to AvIDP are facilitated by its availability, reproducibility and economic considerations. In a flock of up to 5000 animals, a mean number of 50 to 200 chickens is affected per 18 weeks of the breeding cycle. In addition, the availability of the chicken genome sequence now greatly facilitates genetic and immunological studies and has lead to the availability of numerous molecular tools for detailed studies . This and the possibility to select susceptible and resistant birds make the AvIDP-chicken a valuable model system for further studies of inflammatory demyelinating polyradiculoneuropathies.