Pathology of spontaneous and experimental infections by Goose haemorrhagic polyomavirus

Abstract

Haemorrhagic nephritis enteritis of geese (HNEG) is a fatal disease of geese aged from 3 to 12 weeks. The causative virus, Goose haemorrhagic polyomavirus (GHPV), is a member of the Polyomaviridae family We examined goslings either spontaneously or experimentally infected with GHPV. Tissues were sampled for histology, GHPV DNA detection and electron microscopy. Clinical signs and gross lesions observed in experimentally infected goslings were largely consistent with those noticed in field cases. Histological examination showed that, in the acute phase of HNEG, GHPV replicates in almost all the tissues with a particular tropism for endothelial and lymphoid cells. Haemorrhagic foci were widespread in many tissues, including brain. Ultrastructural features were largely consistent with other polyomavirus infections, with accumulation of virions in the nucleus. Non-typical, double-membraned organelles were observed in the cytoplasm. GHPV DNA distribution was widespread in tissues of infected birds, from day 5 post-infection. GHPV therefore induces a systemic disease in its host, leading to severe vascular dysfunction and immunosuppressive B-cell depletion.

Introduction

In the past 20 years, French geese flocks have been regularly affected by a fatal disease, named hæmorrhagic nephritis enteritis of geese (HNEG). This disease, suspected to be associated with a virus, was subjected to pathological and virological investigations.

HNEG was first described in 1969 in Hungary (Bernath & Szalai, 1970), where occurrence of field cases was always associated with the administration of sera collected from convalescent flocks previously affected by Derzsy's disease, in order to confer passive immunity to young geese. No spontaneous case of HNEG had been observed in Hungary until 2001 (Bernath et al., 2001). HNEG was described a few years later in Germany (Schettler, 1977) and France (Schettler et al., 1980), where it has occurred as a sporadic disease, with the exception of a few epizootics in the late 1980's (Sans, 1992), and again in 1997. Recently, epizootics were described in Hungary and Germany (Bernath et al., 2001; Miksch et al., 2002). In a recent report, we demonstrated that the agent of HNEG is a novel polyomavirus, clearly distinct from other polyomaviruses isolated in birds and we thus suggested that the agent of HNEG should be named GHPV for Goose haemorrhagic polyomavirus (Guerin et al., 2000). Phylogenetic and structural analysis performed on the whole GHPV genome confirmed that this virus is different from Budgerigar fledgling disease virus (BFDV; Johne & Muller, 2003). We report here the gross, microscopic and ultrastructural lesions associated with spontaneous and experimental GHPV infections.

Materials and Methods

Natural cases of HNEG

Goslings (n=5) from 5 infected flocks, with typical clinical signs of HNEG, were collected on farms and sent for necropsy and sampling to our laboratory.

Reproduction of HNEG in day-old goslings

Thirty day-old goslings were obtained from a local hatchery. Birds were housed according to the guidelines of the European Community on Animal Care (European Council directive 86/609/ECC, 24 November 1986), in wire-floored cages with infrared lamps for heating, and were provided with food and water ad libitum. Twenty birds were inoculated subcutaneously with 200 μl of viral inoculum, purified from cell culture as previously described (Guérin et al., 2000). This inoculum corresponded to 10⁸ genome-equivalent viruses, as determined by real-time quantitative PCR analysis (data not shown). Ten birds were mock-infected and housed separately. All the birds were clinically monitored on a daily basis. At post-infection days (dpi) 2, 3, 4, 5 and 8, 4 inoculated birds and 2 controls were sacrificed, necropsied and sampled for virological and morphological examinations. At day 8, all infected birds showed terminal clinical signs when sacrificed.

Histological examination

After examination of experimentally or spontaneously infected birds, samples of the following tissues were taken and fixed in 10% buffered formaldehyde solution: lung, trachea, liver, spleen, pancreas, kidney, thymus, cloacal bursa, sections of the digestive tract, sciatic nerve, brain and feather follicle. These tissues were routinely processed, embedded sectioned and stained with haematoxylin, paraffin and eosin.

Extraction and detection of GHPV DNA

At days 0, 2, 3, 4, 5 and 8 post-infection, plasma, tissues (i.e. liver, spleen, kidney, cerebral cortex, pancreas, lung, cloacal bursa, thymus) and cloacal swabs were sampled and stored at −20°C until DNA extraction, using the High pure PCR template Preparation kit (Roche). PCR amplification was performed as previously described, using the set of primers VP1F (5′-GAGGTTGTTGGAGTGACCACAATG-3′) and VP1R (5′-ACAACCCTGCAATTCCAAGGGTTC-3′) (Guérin et al., 2000). PCR products were analyzed by electrophoresis on a 1.5% agarose gel.

Electron microscopy

Cloacal bursa, kidney, feather follicle and lung from infected goslings were processed as routinely described: briefly, samples were fixed with 2% glutaraldehyde in 0.2M Sorensen phosphate buffer, embedded in Embed 812 resin, and finally sliced into 60 nm-thick sections positively stained with uranyl acetate and lead citrate. Examinations were carried out on a transmission Hitachi HU-12A electron microscope.

Results

Clinical signs, gross and microscopic lesions in spontaneously infected geese

In field cases, morbidity was 30 to 80% and the lethality was nearly 100% in geese 4 to 10 weeks old. In these field conditions, death generally followed coma and, in few acute cases, nervous symptoms. Necropsy findings were a subcutaneous oedema, ascites and swelling of the kidneys (Figure 1A). Less frequently, haemorrhagic enteritis was observed in all areas of the intestine and caeca. Older geese, which died after several days of clinical disease, had visceral gout with urate deposits in the joints (Figure 1B). These findings are similar to those in previous reports of the disease (Schettler, 1980; Kisary, 1993).

Figure 1. Gross and microscopic lesions in a 8-week-old gosling, after fatal evolution of spontaneous HNEG. (A) Macroscopic appearance of the kidney, showing severe swelling and hemorrhages. (B) Macroscopic view of urate deposits within the tibio-tarsometatarsal joint. (C) Photomicrograph of a normal cloacal bursa. Haematoxylin & Eosin.×100 (Bar 100 μm). (D) Photomicrograph of the cloacal bursa: multifocal, marked centrofollicular lympholysis without inflammatory cells. Haematoxylin & Eosin.×400. (Bar 25 μm). (E) Photomicrograph of kidney tissue section, showing severe, multifocal necrosis of tubules (arrows). Haematoxylin & Eosin.×400. (Bar 25 μm).

Histopathologically, the most obvious features observed in spontaneous cases were (i) interstitial nephritis and necrosis of the kidney tubular epithelium (Figure 1C), and (ii) a moderate to severe lymphocytolysis in both cortex and medulla of follicles of the cloacal bursa (Figure 1D), suggestive of B-lymphocyte depletion (Guerin et al., 2000). In grossly affected intestine, there was necrosis of the intestinal epithelium. Haemorrhagic foci were also observed in most tissues, particularly after the acute course. No viral inclusions were detected in tissues of birds diagnosed with HNEG by PCR.

Figure 2. Microscopic lesions observed day-8 post-inoculation. Experimental reproduction of HNEG in goslings. (A) At the clinical stage of the disease (dpi 8), moderate lymphocyte depletion was observed in cloacal bursa (A, Bar 100 μm) while there was extensive tubular necrosis (B, Bar 250 μm). (C) A discrete nuclear enlargement of endothelial cell involved various organs as soon as dpi3; (kidney, (C), Bar 50 μm). (D) Later in the course of the disease, multifocal marked arteriolitis, with vessel wall infiltration and major endothelial cell activation, (D) was observed in most tissues (C, Bar 15 μm). Haematoxylin & Eosin.

Histological findings of experimentally infected goslings

Inoculation of day-old goslings resulted in a 100% mortality within 8 days. Clinical signs were limited to a comatose state and final nervous signs such as opisthotonus and paresis. At necropsy, all inoculated birds had ascites, subcutaneous oedema and a severe, acute nephritis, which in some cases was haemorrhagic. Necrotic-haemorrhagic foci were noticed on the duodenum and/or jejunum in 4/20 birds. These lesions concerned either the villi or the intercrypt regions, were limited in extent and were likely subsequent to underlying arteriolitis.

Detailed histological observations are reported in Table 2. Most salient lesions observed at dpi 8 were tubular epithelial necrosis in kidney, moderate lymphocytic depletion of the cloacal bursa and spleen, similar to that observed in natural infection (Figure 2). Lymphocyte depletion in germinal centres was observed as early as dpi 3 and dpi 4 in spleen and bursa respectively, while kidney lesions were only observed from dpi 5. No lymphocyte depletion was observed in thymus.

Table 2. Histological findings during the course of experimental infection of goslings^a

Tissues	Day 2		Day 3		Day 4		Day 5		Day 8
	Birds	Lesion intensity	Birds	Lesion intensity	Birds	Lesion intensity	Birds	Lesion intensity	Birds	Lesion intensity
Spleen
Follicular depletion	–		2/4 (a-b)	0/+	4/4	++/+++	4/4	+++	4/4	+++
Arteriolitis	–		–				4/4	+	3/4(a-c-d)	++/+++
Kidney
Epithelial nephritis	–		–		–		1/4 (d)	+	4/4	+++
Arteriolitis	–		–		–		1/4 (a)	+	4/4^b	++
Eye
Arteriolitis	–		–		–		2/4 (a-b)	+/++	4/4	+++
Pancreas
Nuclear swelling, endothelial cells	–		–		–		–		4	++
Arteriolitis	–		–		–		–		4/4	+/++
Lung
Nuclear swelling, endothelial cells	–		2/4 (a-d)	+	4/4	+	4/4	+/++	4/4	+++
Arterilolitis	–		1/4 (d)	0/+	1/4 (d)	+	3/4 (a-b-d)	+	4/4	++
Heart
Nuclear swelling, endothelial cells	–		1/4 (d)	+	2/4 (b-d)	+	4/4	+/++	4/4	++/+++
Arteriolitis	–		1/4 (d)	0/+	–		3/4 (a-b-d)	+	3/4 (a-c-d)	++
Cloacal bursa
Follicular lympholysis	–		–		3/4 (b-c-d)	+	4/4	++	4/4	+++
Heterophilic infiltrate	–		–		1/4 (d)	0/+	2/4 (b-c)	+/++	4/4	+/++
Small intestine
Nuclear swelling, endothelial cells	–		3/4 (a-b-d)	+	4/4	+	4/4	+/++	4/4	+++
Arteriolitis	–		1/4 (d)	0/+	1/4 (d)	+	3/4 (a-b-d)	+	4/4	+/++
Caeca
Nuclear swelling, endothelial cells			3/4 (a-b-d)	+	4/4	+	4/4	+/++	4/4	+++
Arteriolitis			3/4 (a-b-d)	0/+	1/4 (d)	+	3/4 (a-b-d)	+	4/4	+/++
^a Day 8 corresponds to the death or sacrifice of diseased birds. For each day post-inoculation, 4 birds (noted a, b, c, d) and 2 mock-infected birds were analyzed. Lesion severity was graded from 0 to+++:+ minimal;++ mild;+++ severe.
^b Microthrombosis (kidney).

Despite systemic arteriolitis, only few haemorrhages and thrombotic lesions were observed at clinical stage of the experimental disease. The vascular lesions were first observed at dpi 3; they consisted of a multifocal moderate endothelial cell nuclear enlargement, sometimes associated with a light lymphocytic infiltration of the vessel wall. Such images were observed in various organs. At dpi 4 to 5, arteriolar lesions involved almost all examined tissues. Central nervous system was affected, with mild to moderate vascular lesions, but no major degenerative lesions were observed. Control birds analyzed at each time point showed none of these lesions.

GHPV genome distribution during experimental infection

Results of GHPV distribution in tissues are summarized in Table 1. During the course of experimental infection, PCR allowed detection of GHPV DNA as early as 3 dpi. In most cases, DNA detection occurred in tissues showing histological changes (Table 1). In thymus, DNA was detected early, whereas no lesion could be noticed in lymphoid cells, endothelial cells only showing a mild activation. At 8 dpi, all the tissues examined were PCR positive. Plasma and cloacal swabs remained negative until 8 dpi. None of the control goslings was found PCR positive for GHPV DNA.

Table 1. Detection of GHPV by PCR during the course of an experimental infection of goslings^a

Tissues	Day 2	Day 3	Day 4	Day 5	Day 8
Liver	0/4	3/4 (a,b,c)	4/4	4/4	4/4
Spleen	0/4	3/4 (a,b,d)	4/4	4/4	4/4
Kidney	0/4	3/4 (a,b,d)	4/4	4/4	4/4
Cerebral cortex	0/4	1/4 (a)	2/4 (a,b)	3/4 (b,c,d)	4/4
Pancreas	0/4	2/4 (b,d)	4/4	3/4 (a,c,d)	4/4
Lung	0/4	2/4 (b,d)	4/4	4/4	4/4
Cloacal bursa	0/4	3/4 (a,b,d)	4/4	3/4 (a,c,d)	4/4
Thymus	1/4 (b)	3/4 (a,b,d)	4/4	4/4	4/4
Cloacal swab	0/4	0/4	0/4	0/4	4/4
Plasma	0/4	0/4	0/4	0/4	4/4
^a Detection of GHPV DNA from tissues of goslings inoculated with 10^8 genome-equivalent GHPV. Day 8 corresponded to the death or sacrifice of diseased birds. For each day post-inoculation, 4 birds (a, b, c, d) and 2 mock-infected birds were analyzed. Control birds remained PCR negative throughout the experiment.

Ultrastructural analysis of GHPV-infected cells

Virus-like particles were largely detected in the nuclei of different cell types in kidneys, lungs, feather follicles, where endothelial cells appeared to be the most regularly infected. In cloacal bursa, virions were mostly observed in lymphoid cells. The ultrastructural appearance was consistent with polyomaviruses (Figure 3): an intranuclear formation of crystalline networks in a few cells (Figure 3A), disruption of nuclear membranes, and cytoplasmic virus particles in some infected cells (Figure 3B). Chromatin margination was noticed in most nuclei. No linear aggregation of virions was detected in GHPV-infected cells, as seen in BFDV-infected budgerigar kidney cells (Dykstra & Bozeman, 1982). In most infected cells, vesicles of about 2 μm in diameter were found aggregated around the nucleus. No mitochondria were recognisable in the cytoplasm of these cells.

Figure 3. Electron microscopy on tissues of goslings day 8 post-inoculation. (In all pictures, bar=500 nm). (A) Crystalline organization of virions in the nucleus of an endothelial cell adjacent to a feather follicle. (B) Endothelial cell in the pulp of a feather follicle. Note the virions, margination of chromatin and organization of organelles around the nucleus. The cytoplasm of GHPV-infected cells contain no recognizable mitochondria. (C) Capillary epithelial cell of the lung. Virions are released from the nucleus to the cytoplasm of the cell (top of the picture). (D) Proximal convoluted tubular epithelial cell in the kidney. Margination of chromatin. Note the mitochondrial vacuolation.

Discussion

Clinical signs and gross lesions observed in this study were very similar to those published in previous clinical reports in Hungary, Germany or France, confirming that the clinical manifestations of HNEG are identical over time and space, as is the virus (Sans, 1992; Kisary, 1993; Bernath et al., 2001; Miksch et al., 2002; Guerin, 2003). In young goslings, the main lesions are ascites and oedema, associated with vascular lesions. When older geese are affected, clinical disease developed more slowly and was secondary to kidney lesions, mainly through urates deposition. Nervous signs are often observed in experimental infections but are uncommonly reported in field cases (Guerin, 2003).

Recently, PCR using VP1-specific primers was applied to tissues from the first cases described in Hungary in 1969. There was complete sequence homology with French GHPV isolates (Bernath et al., 2001). These findings are consistent with the well-known genetic stability of polyomaviruses (Shah, 1996).

During the course of infection, GHPV seems to replicate first in endothelial cells, nuclear enlargement of endothelial cells and arteriolitis being the first lesions noticed. These histological findings suggest a selective tropism for endothelial cells, which might be, quantitatively, the main target of GHPV replication and qualitatively, of great relevance in pathogenesis of HNEG. Endothelial cells are indeed known to play a critical role in many biological pathways, to which many results of vascular dysfunctions as ascites or oedema, can be directly related (Michiels, 2003). The dissemination through all the tissues occurred at dpi 5, as shown by the PCR analysis, while GHPV was still not detectable in serum. The virus was likely associated with circulating cells, but at a very low rate, since our technique allows routine detection of GHPV in serum of acutely or chronically infected geese (unpublished data). Further studies including in situ hybridization experiments on infected tissues will allow a more accurate analysis of target cells.

Lymphoid cells appear to be another main target of GHPV; virions were observed in many bursal lymphoid cells and cloacal bursa systematically shows a significant lympholysis, whereas thymic lymphoid cells are not affected. This feature is fairly relevant with the well-documented tropism of polyomaviruses to B-lymphocytes but not to T-lymphocytes (Shah, 1996; Wei et al., 2000). Histopathology on lymphoid tissues is still the only way to assess this tropism in goose, since there are no immunological tools to discriminate B vs T lymphoid clusters in waterfowl species (Higgins, 1996). Further investigations are needed to assess the possible immunosuppressive nature of this virus, alone or together with other lymphotropic viruses, such as the recently identified goose circovirus (Soike et al., 1999). Most polyomavirus infections result in intranuclear, large basophilic inclusion bodies, mainly associated with karyomegaly; in particular, they are never totally absent in BFDV infections (Ritchie, 1991). Various examinations performed previously (Kisary, 1993; Guerin, 2003) and intensive histopathological analysis during this work failed to identify such inclusions in goslings either spontaneously infected or inoculated with GHPV. This could explain the initial difficulty in elucidating the aetiology of HNEG. It also suggests that infections by particular genotypes or species of polyomaviruses might not have been recognized as such, because those viruses can only be detected using a molecular or alternatively, electron microscopy examination of tissues. Structure and assembly of virions were fairly suggestive of polyomaviruses. Crystalline networks of virions were regularly observed, whereas no filamentous arrangement, as described in kidneys of BFDV-infected budgerigars (Dykstra & Bozeman, 1982) could be seen. A consistent observation in infected cells was both the absence of mitochondria and the presence of perinuclear non-typical double-membraned bodies. Both origin and function of these structures are still unclear, but the size of these organelles, their double-membrane organization and the absence of regularly shaped mitochondria in such cells are suggestive of a degenerative vacuolation of mitochondria. This might be significant, since mitochondrial permeability dysfunctions are critical in both apoptotic and necrotic pathways (Grandville & Gottlieb, 2002).

Avian polyomaviruses were first identified in various psittacine species (Bozeman et al., 1981; Ritchie, 1991) and then in other wild birds, such as finches and falconiformes (Phalen et al., 1999). All the avian polyomaviruses isolated so far are members of one species, commonly referred to as BFDV. Clinical pictures described for BFDV infections in psittacines show a tremendous diversity but are mainly reminiscent of HNEG. Indeed, GHPV shares common features with BFDV: (i) they are pathogenic, whereas mammalian polyomaviruses are quite harmless in the immunocompetent host and (ii) they show an apparent long-term persistence in infected birds (Shettler, 1980; Phalen et al., 1993). BFDV pathogenesis is still largely unknown, since there have been no controlled studies of experimental infections in psittacine species. An alternative model of BFDV infection was developed in immunocompromized chickens, but seems far from the natural infection (Fitzgerald et al., 1996). GHPV may allow in-depth studies of the interactions between a polyomavirus and its natural avian host.

Much is still needed to clarify the biology of GHPV infection. First of all, infection of long-term carrier breeders might have major consequences through a putative vertical transmission. Similarly, the host range of GHPV in birds, particularly in waterfowl species may be worth investigation. Polyomaviruses are poorly investigated and may be widely underdiagnosed. Yet, they may represent emerging and significant pathogens of birds and poultry.

Acknowledgements

We gratefully thank Pr. John H. Barnes (College of Veterinary Medicine, NCSU, Raleigh, US), as our colleagues Drs Christelle Camus-Bouclainville, Frédérique Messud and Stéphane Bertagnoli for their helpful suggestions and careful reading of the manuscript. We recognize the contribution of Isabelle Fourquaux and Jean-Luc Duteyrat to electron microscopy examinations. This work was supported by Régions Aquitaine and Midi-Pyrénées, CIFOG (Comité Interprofessionnel des Palmipèdes à Foie Gras) and OFIVAL (Office Interprofessionnel des Viandes, de l'Elevage et de l'Aviculture).