Introduction

Influenza A viruses are likely to have originated from wild waterfowl, which act as a natural reservoir [5, 9]. Given the long period of coexistence, these viruses are thought to be evolutionarily stable and non-pathogenic to their waterfowl hosts [11, 13]. There are several subtypes of the viruses according to the antigenicity of surface glycoproteins, but only certain subtypes, generally subtypes H5 and H7, are associated with acute highly pathogenic avian influenza (HPAI) disease in birds. The acute HPAI disease outbreaks usually become a matter of public health concern only among populations of domestic land-based poultry, because the natural hosts, waterfowl, are believed to be resistant to HPAI disease. Since late 2003, however, several endemics of highly pathogenic H5N1 avian influenza viruses have also demonstrated pathogenicity among waterfowl and wild birds, indicating that the viruses have distinct biological properties compared to previously endemic HPAI viruses [1, 2].

Pigeons, often sold in live-bird markets in East Asian countries, are believed to exhibit resistance against HPAI viruses [3, 4, 7, 8]. Therefore, the general consensus had been that pigeons play a minimal role in spreading H5 viruses. Following reports starting in 2002 that described pigeons potentially involved in and perishing during H5N1 outbreaks of domestic and wild birds, the general consensus has begun to change on the possible susceptibility of pigeons to HPAI [2, 12]. The susceptibility of pigeons to H5N1 virus and the role of pigeons in the transmission of avian influenza viruses to domestic birds and humans remain controversial topics. Further study is therefore required to elucidate the susceptibility of pigeons to recent H5N1 HPAI A viruses.

This study was thus undertaken to compare the susceptibility of pigeons to a panel of ten strains of H5N1 HPAI viruses isolated from several different species of infected birds during the outbreaks of 2004 H5N1 HPAI in China. We examined the extent to which these viruses are able to replicate and be shed in pigeons.

Materials and methods

Viruses

A total of ten H5N1 HPAI viruses (four chicken-derived, three duck-derived, and three pigeon-derived) were selected: A/Chicken/Anhui/85/2004 (Ck/Ah/04), A/Chicken/Guangxi/12/2004 (Ck/Gx/04), A/Chicken/Hubei/14/2004 (Ck/Hb/04), A/Chicken/Tianjin/65/2004 (Ck/Tj/04), A/Duck/Guangdong/23/2004 (Dk/Gd/04), A/Duck/Guangxi/13/2004 (Dk/Gx/04), A/Duck/Hunan/15/2004 (Dk/Hn/04), A/Pigeon/Hunan/39/2002 (Pg/Hn/02), A/Pigeon/Jilin/30/2004 (Pg/Jl/04), and A/Rock Pigeon/Shanxi/47/2004 (Pg/Sx/04). Nine viruses were isolated from tissue samples collected from birds with significant disease signs during the H5N1 HPAI outbreaks in China in 2004, and one virus was isolated from a dead pigeon in 2002 (Table 1). All H5N1 viruses had been grown in the allantoic cavities of 10-day-old embryonated chicken eggs, and harvested virus fluid titers ranged from 7.5 to 8.9 log 10 EID50/ml.

Table 1 H5N1 highly pathogenic avian influenza viruses used in this study
Full size table

Pigeons

Seven-week-old pigeons, from a local farm in Harbin, China, with no antibody against H5, H7 and H9 subtypes of influenza viruses were used in these experiments. All pigeons were housed separately in each isolation unit, ventilated under negative pressure with HEPA-filtered air. Feed and water were provided ad libitum. All experiments using H5N1 viruses were performed in a biological safety level 3 facility at the Harbin Veterinary Research Institute of CAAS and were authorized by the Chinese Ministry of Agriculture.

Experimental design

One hundred sixty pigeons were divided into ten groups (16 birds/group) and intranasally inoculated with 0.1 ml of virus fluid containing 106 EID50 of an H5N1 virus. All pigeons were monitored daily for clinical signs. Oropharyngeal and cloacal swabs were taken from live pigeons at 3, 5, 7, 10, and 14 days postinfection (dpi) and were collected in PBS with antibiotics for virus isolation and titration. Blood samples were collected from live pigeons at 21 dpi for analysis of specific serum antibody. Two pigeons in each infected group were necropsied at 3, 5, 7, 10 and 14 dpi for virus titration. Tissues of the brain, heart, liver, spleen, lungs, kidneys, pancreas, trachea, small intestine and bursa of fabricius were collected from each bird.

Virus isolation

Oropharyngeal and cloacal swabs and all tissues collected from infected and control pigeons were stored at −70°C until virus titrations were performed. Standard procedures were applied for the titration using 10-day-old embryonated chicken eggs.

Serological analysis

Sera collected from pigeons in the infected and control groups were pretreated by receptor destroying enzyme (RDE; Sigma, USA) at 37°C overnight and inactivated for 30 min at 56°C. They were then processed for detection of H5-specific antibodies by hemmaglutinin-inhibition test (HI test) using chicken red blood cells and an inactivated homologous H5N1 influenza viruses as antigen.

Pathological examination

Pigeons in the Dk/Gd/04- or Pg/Sx/04-infected groups were selected for pathological examination. Excised tissues of the liver, kidneys, heart, lungs, trachea, pancreas, intestine, brain, spleen and bursa of fabricius from pigeons were preserved in 10% neutralized buffered formalin. Tissues were then processed for paraffin embedding and cut into 5-μm-thick sections. One of the sections was stained with routine hematoxylin-and-eosin, and the other was processed for immunohistological staining using rabbit anti-H5N1 influenza virus polyclonal antibody (anti-A/Vietnam/1203/04[H5N1]). Specific antigen-antibody reactions were visualized by 3,3′diaminobenzidine tetrahydrochloride treatment using the Dako EnVision system (Dako Co. Ltd., Tokyo, Japan).

Results

Clinical findings of pigeons

No deaths were observed in any of the infected birds during the observation period. Among the ten different viruses examined, only two duck-derived strains (Dk/Gd/04 and Dk/Gx/04) and two pigeon-derived strains (Pg/Hn/02 and Pg/Sx/04) caused observable clinical symptoms such as decreased activity and neurological signs (Table 2). All of these clinical signs resolved within 3 dpi.

Table 2 Clinical manifestations observed in pigeons after intranasal inoculation with 106EID50 of H5N1 avian influenza viruses
Full size table

Viral titers in oral/cloacal swabs of infected pigeons

Viruses were inconsistently re-isolated from oral and cloacal swabs, with titers ranging from 100.98 EID50/ml to 102.25 EID50/ml from pigeons infected with viruses derived from chickens, ducks and pigeons. Although there was a definitive association between viral shedding and viral origin, it is noteworthy that many viral strains (Ck/Ah/04, Ck/Gx/04, Ck/Hb/04, Dk/Gd/04, Dk/Hn/04, Pg/Jl/04, and Pg/Sx/04) infected pigeons and caused a certain amount of viral shedding from pigeon cloaca even at 5 and 7 dpi (Fig. 1).

Fig. 1
figure 1

Viral titers in oral (blue) and cloacal (dark red) swabs from infected pigeons. Several of the viral strains tested (Ck/Ah/04, Ck/Gx/04, Ck/Hb/04, Dk/Gd/04, Dk/Hn/04, Pg/Jl/04, and Pg/Sx/04) led to some viral shedding

Full size image

Isolation of H5N1 viruses from the organs of infected pigeons

Viruses were inconsistently isolated from various organs of pigeons in infected groups (Table 1). Isolation patterns were: (1) systemic, mostly including the brain; (2) respiratory tropic; (3) brain and respiratory tropic; and (4) sporadic isolation from the other organs. Systemic infection through 7 dpi was confirmed in some of the pigeons infected with Dk/Gx/04 or Pg/Sx/04. The virus titers ranged from 100.98 EID50/ml to 104.5 EID50/ml. However, all infections terminated in a self-limited manner, as indicated by negative virus isolation results at 10 and 14 dpi (data not shown) (Table 3).

Table 3 Replication of H5N1 viruses in pigeons following intranasal inoculation
Full size table

Histopathology

Pigeons in the Dk/Gd/04- or Pg/Sx/04-inoculated groups had respiratory inflammation,such as tracheitis, bronchitis and airsacitis at 3 dpi. Pigeons necropsied at 5 dpi showed an apparent immunologic reaction such as hyperplastic lymph follicles in the spleen and Bursa of Fabricius. In pigeons necropsied at 7 dpi, the most severe lesions were observed in the brain in both the Dk/Gd/04- and Pg/Sx/04-infected groups. Brain lesions included malacia with perivascular cuffing of mononuclear cells, degeneration and necrosis of nerve cells, activation of glial cells and meningitis (Fig. 2a). Many nerve cells and glial cells were positive for immunohistological staining for viral antigens (Fig. 2b). Viral antigens were detectable only within the brain. Developed lymph follicles with atrophic appearance and a vascular fibrin thrombus were detected in the lungs. In pigeons in the Dk/Gd/04-infected group, severe atrophy of lymph follicles in the lymphatic organs, such as the spleen and bursa of fabricius (Fig. 2c), and perivascular infiltration of mononuclear cells in intestinal serosa were prominent. Small infiltrative foci of pseudoeosinophils in kidneys and lymphocyte depletion in Peyer’s patch follicles (Fig. 2d) were also observed in pigeons infected with Pg/Sx/04.

Fig. 2
figure 2

Pathological changes observed in infected pigeons. a Malacia with perivascular cuffing of mononuclear cells, degeneration and necrosis of nerve cells and activation of glial cells were observed in the brain. Pigeons were infected with Pg/Sx/04, 7 dpi. b Nerve cells and glial cells were positive for viral antigens (brown pigment). Pigeons were infected with Pg/Sx/04, 7 dpi. c Severe atrophy of lymph follicles in the bursa of fabricius. Pigeons were infected with Dk/Gd/04, 7 dpi. d Lymphocyte depletion in Peyer’s patch follicles. Pigeons were infected with Pg/Sx/04, 7 dpi

Full size image

Serology

Among the six pigeons in each group that were inoculated with one of the chicken-derived viruses, one had H5-specific HI antibody. Among the pigeons in the group inoculated with duck-derived viruses, two or three of the six pigeons in the group had HI antibody, while two or four of the six pigeons in the pigeon-derived virus group showed a positive reaction (Table 2). These H5-specific titers were detected in the sera collected at 21 dpi and ranged from 1:32 to 1:64.

Discussion

Previous research had indicated that pigeons were completely resistant or minimally susceptible to infection with HPAI viruses and even with low-pathogenic avian influenza viruses [3, 4, 10]. This notion was supported by findings that groups of pigeons inoculated with either HPAI virus (A/chicken/Pennsylvania/1370/83 or A/chicken/Australia32972/85) or non-pathogenic virus (A/chicken/Pennsylvania/13609/93 or A/emu/Texas/42499/93) remained healthy during the 21-day observation period, did not shed virus, and did not even develop antibodies to these pathogens [7]. Further research had suggested that pigeons were mostly resistant and probably played a minimal epidemiological role in the dissemination of the H5N1 Hong Kong-origin influenza viruses compared with chickens, quail and ducks [8].

Contrary to these findings, we found that the H5N1 viruses used in this study can successfully infect and cause disease in pigeons. Although infection efficiency varies among birds, the results of seroconversion (antibody production) suggest that infection is established in pigeons even in temporal or non-efficient infection cycles. These findings support several reports that described some susceptibility of pigeons to recent H5N1 viruses. Namely, during the outbreak of H5N1 avian influenza infection in Hong Kong in 2002, one pigeon was found dead [2]. Furthermore, a recent H5N1 virus from Indonesia (A/chicken/Indonesia/2003) showed significant pathogenicity in experimentally infected pigeons [6]. Therefore, it is likely that recent prevalent viruses with worldwide distribution including China, possess different properties of susceptibility in pigeons than previous H5N1 strains before the early part of 2002.

Pathological examination revealed that some H5N1 influenza viruses caused pathological changes in respiratory organs in early stages and invaded the brain in later stages, accompanying apparent lymphatic atrophy. Although no pigeons died and indeed only showed mild clinical signs in the 48 h after infection, the virus replicated in the brains of pigeons, reaching titers of 104.25 EID50 at 7 dpi. The clinical signs and mortality were not the same as those of the pigeons infected with A/chicken/Indonesia/2003 [6]; however, neurotropism is a noteworthy characteristic for the recent H5N1 viruses. Moreover, our data showed that pigeons with neuronal infections also discharged infectious viruses in oral or cloacal secretions. We did not conduct transmission experiments, but several reports have indicated low transmission efficiency from infected pigeons to sentinel animals. However, during the second wave of avian influenza in Indonesia in 2006, at least one fatal human case from infection by H5N1 HPAI virus was reported in which the victim had a history of contact with pigeons in his house. Our findings demonstrate that pigeons can indeed be susceptible to the more recent Asian H5N1 HPAI viruses and suggest that pigeons could be a source of infection to other animals including humans.