Spread of Influenza Virus A (H5N1) Clade 2.3.2.1 to Bulgaria in Common Buzzards

Detection of this highly pathogenic clade in Europe poses a health threat to poultry and humans

W ild aquatic birds are considered natural reservoirs of all known infl uenza virus subtypes (1). Highly pathogenic avian infl uenza viruses (HPAIVs) usually cause asymptomatic infections in waterfowl. Compared with that for poultry, the number of reported outbreaks of HPAIVs in wild birds (aquatic or terrestrial) before 2002 was low; 1 infl uenza A (H5N3) outbreak occurred in wild common terns (Sterna hirundo) in South Africa in 1961 (2), and H7 subtype HPAIV was isolated from a Saker falcon (Falco cherrug) in Italy in 2000 (3). In December 2002, a die-off of aquatic waterfowl caused by an HPAIV (H5N1) occurred in Penfold Park in Hong Kong. That event was followed by a second outbreak a week later in Kowloon Park, also in Hong Kong (4). Since then, cases of HPAIVs (H5 and H7 subtypes) in wild birds have been reported often.
In May 2005, a massive HPAIV (H5N1) outbreak occurred in wild aquatic birds in Qinghai Lake, western People's Republic of China; 6,184 gulls, geese, great cormorants, and ruddy shelducks died (5). The lake is a staging area for migratory waterfowl, and the scientifi c community's fear that the virus would spread during migration (6) was realized when the so-called Qinghailike infl uenza virus A (H5N1) clade 2.2 spread to western Siberia in Russia and then to many countries in Asia, the Middle East, Europe, and Africa at the end of 2005 and during 2006, killing poultry fl ocks and wild birds.

Virus Isolation and Initial Characterization
Pooled lung, trachea, liver, cecal tonsil, and gizzard tissue from a common buzzard were injected into embryonated chicken eggs, and an avian infl uenza virus (A/ common buzzard/Bulgaria/38WB/2010) was isolated. The isolate was subjected to hemagglutination inhibition (HI) assays, as specifi ed in the World Organisation for Animal Health (OIE) Manual of Diagnostic Tests and Vaccines for Terrestrial Animals 2011 (11). The 50% egg infectious dose (EID 50 ) was assayed in 10-day-old embryonated chicken eggs after a 40-hour incubation at 35°C and calculated according to Reed and Muench (12).

Sequence Analysis
Viral RNA was extracted from virus-containing allantoic fl uid by using the MagMax-96 AI/ND RNA extraction kit (Applied Biosystems/Ambion, Austin, TX, USA). Reverse transcription PCR was performed by using a universal set of primers (13,14); whole-genome sequencing was performed by using an Illumina Genome Analyzer (Illumina, San Diego, CA, USA), as described by Ducatez et al. (15).

Phylogenetic Analysis
We performed a phylogenetic analysis of all gene segments of A/common buzzard/Bulgaria/38WB/2010. Sequences were retrieved from the National Center for Biotechnology Information infl uenza virus database (16) and aligned by using the ClustalW tool in MEGA4 (17). Phylogenetic relationships were estimated by using the MrBayes program (18) to apply a Bayesian method. The model of nucleotide substitution that best fi ts our data was selected by using the Modeltest 3.7 program (19).

Antigenic Characterization and Pathogenicity Studies
We assessed the antigenic relationship of A/common buzzard/Bulgaria/38WB/2010 with other infl uenza (H5N1) viruses by performing HI assays with a panel of postinfection ferret antiserum against viruses from clades 2  Table). All animal studies were conducted in United States Department of Agriculture-approved biosafety level 3 enhanced facilities at St. Jude Children's Research Hospital.

Pathogenicity Tests in Chickens
The intravenous pathogenicity index test was conducted according to the OIE Manual of Diagnostic Tests and Vaccines for Terrestrial Animals 2011 (11) to determine whether A/common buzzard/Bulgaria/38WB/2010 is pathogenic in chickens. A natural route of infection study was performed in fi ve 8-week-old specifi c pathogen-free chickens. A/common buzzard/Bulgaria/38WB/2010 (10 6 EID 50 in 0.5 mL) was administered to each bird as follows: 0.1 mL in the nares, 0.1 mL in the trachea, 0.2 mL in the throat, and 1 drop in each eye. The chickens were examined daily for clinical signs of disease.

Pathogenicity Tests in Mice and Ferrets
The 50% mouse lethal dose was determined to assess the pathogenicity of the buzzard infl uenza (H5N1) virus in mammals. The experiment was conducted as described by Boon et al. (20), and the titer was calculated according to Reed and Muench (12).
We also performed a pathogenicity and transmissibility study in a ferret model (21). We collected nasal washes from all ferrets (i.e., inoculated ferrets, naive direct-contact ferrets, and naive respiratory droplet-contact ferrets) every other day and nasal swabs from only the donors (i.e., inoculated ferrets) on day 4 postinfection. EID 50 was calculated according to Reed and Muench (12).

Gross Pathologic Findings
The fresh common buzzard carcass that was initially found was in good condition, without any discharge or other clinical signs of disease. Necropsy revealed a suffi cient quantity of yellow subcutaneous and abdominal fat tissue. Gross pathologic changes were not observed in the lungs, trachea, heart, liver, spleen, gizzard, stomach, pancreas, or intestines.

Antigenic Characterization and Pathogenicity in Chickens
In the HI test, A/common buzzard/Bulgaria/38WB/2010 had a titer of 80 to antiserum from infl uenza (H5N1) clades 2.3.2 and 2.3.2.1 (Table). The intravenous pathogenicity index of A/common buzzard/Bulgaria/38WB/2010 was scored 3.0 because all 10 birds were found dead within 24 hours after inoculation. The natural route of infection study resulted in 2 chickens being found dead on day 2 postinfection, and the other 3 were sick. On day 3 postinfection, the remaining 3 chickens were found dead. The clinical signs observed before death were cloudy eyes, cyanosis of the exposed skin and wattles, edema of the face, and diarrhea.

Pathogenicity in Mice and Ferrets
The 50% mouse lethal dose of A/common buzzard/ Bulgaria/38WB/2010 was 30 EID 50 (EID 50 of the virus was 10 9 /mL). All 3 donor ferrets were shedding virus by day 5 postinfection; however, only 1 was still shedding virus by day 7 postinfection (Figure). Virus titers were not detected in the nasal washes of contact ferrets, indicating that A/ common buzzard/Bulgaria/38WB/2010 is not transmissible by direct contact or respiratory droplets. All ferrets used in this study were healthy during the experiment and showed no clinical signs of disease; they were alert, playful, and eating and drinking normally.

Hemagglutinin
The virus's hemagglutinin (HA) cleavage site had the polybasic amino acid sequence PQRERRRKRGLF, which is characteristic of HPAIVs (22). The cleavage site also had a K329 deletion (H5 numbering used throughout). The receptor-binding pocket of HA1 retained residues E186, Q222, and G224, which preferentially bind to avian α-2,3-NeuAcGal receptors (23). All conserved residues within the HA receptor-binding domains of A/common buzzard/Bulgaria/38WB/2010 were identical to those of other clade 2.3.2 HA sequences available in GenBank (online Technical Appendix Figure).

Neuraminidase
Compared with that of A/goose/Guangdong/1/96, the amino acid sequence of A/common buzzard/ Bulgaria/38WB/2010s neuraminidase had a 20-residue deletion in the stalk region (residues 49-68), which was thought to be required for infl uenza viruses to adapt from wild aquatic birds to domestic chickens (24). This deletion causes a loss of the N terminal NQS glycosylation site (positions 50-52). Residues E119, H275, or N295 (N1 numbering) were not mutated, which suggests sensitivity to oseltamivir and zanamivir (25,26).  consisting of >78 aa are intact and functional. The PB1-F2 N66S mutation, which is characteristic of increased viral pathogenicity and contributed to the high lethality of the 1918 pandemic infl uenza virus (28), was not observed in A/ common buzzard/Bulgaria/38WB/2010. However, several PB2 mutations were present (online Technical Appendix Table).

Nonstructural Proteins
The nonstructural (NS) 1 protein of A/common buzzard/Bulgaria/38WB/2010 is encoded by 230 aa. This sequence differs from other closely related NS sequences in the National Center for Biotechnology Information database in that it lacks the 5-aa deletion (residues 80-84) that became common in HPAIV (H5N1) NS sequences after they were fi rst observed during 2001 in poultry in Hong Kong.

Phylogenetic Analysis
To determine the place of A/common buzzard/ Bulgaria/38WB/2010 in the modern classifi cation of infl uenza (H5N1) viruses, we phylogenetically analyzed HA nucleotide sequences of viruses representing all clades that had been reported as of the end of 2010. The HA of the Bulgarian isolate clustered with subtypes from clade 2.3.  Figure).
To assess possible reassortment in the A/common buzzard/Bulgaria/38WB/2010 genome, we performed phylogenetic analysis of the remaining genes using the same group of viruses that had been used to make the HA tree.  Figure). The evolutionary distance between the 2 groups of isolates in the M tree is long, indicating that the M gene of A/common buzzard/Bulgaria/38WB/2010 originates from a non-clade 2.3.2 ancestor.

Discussion
The day after we received the HPAIV (H5N1)containing common buzzard carcass, offi cials in Romania notifi ed the OIE of an outbreak of an HPAIV (H5N1) in Letea Village (Danube Delta); 47 backyard chickens were found dead. These reports are considered to be the introduction of HPAIV (H5N1) clade 2.3.2.1 into Europe. Furthermore, the European Reference Laboratory on Avian Infl uenza and Newcastle Disease (Weybridge, UK) found that the HA gene sequence of the Bulgarian isolate is 99.9% similar to that of the Romanian isolates from Letea Village (34), confi rming that both viruses are derived from a common source, which is most likely wild birds.
St. Konstantin and Helena Black Sea Resort, where the common buzzard carcass was found, is located on the border of Batova (43°21′11″N, 27°57′33″E), which is a complex of habitats typical for woodland bird species, waterfowl, and poultry. This area is defi ned as a bottleneck migration site of global importance because 3 fl ows of migratory birds meet over the Batova River Valley (35). Each spring (March-April), migratory waterfowl from Africa, Bosphorus, and the Dardanelle Straits stop at the lakes along the coast of the Black Sea in Bulgaria on their way north (36). Their next resting spot is the Danube Delta, where the Romanian outbreak occurred and <250 km from where the Bulgarian buzzard carcass was found. Thus, if we are correct in our hypothesis that birds from Via Pontica are the hosts that carried the HPAIV (H5N1) to Bulgaria and Romania, the virus most likely came from Tyva and Mongolia, where its ancestral viruses had been isolated, to Via Pontica by using >2 other overlapping fl yways. Although this is the most likely scenario, it is still unconfi rmed. To confi rm this hypothesis, the pathogenicity of A/common buzzard/Bulgaria/38WB/2010 in ducks and its possible transmission among them must be defi ned, which will require additional biologic studies.
To trace the possible routes of introduction of HPAIVs (H5N1) into Bulgaria, the veterinary authorities and ornithologists from the Bulgarian Society for the Protection of Birds organized continuous monitoring of the bird areas along the Black Sea coast, near the Danube River and around the Ogosta Dam and Lom River. From January 1, 2010, through April 30, 2010, a total of 812 cloacal, fecal, and tissue samples from wild birds collected from these areas were tested for avian infl uenza virus; 269 samples were collected after March 15, 2010. All samples tested were negative for HPAIV (H5N1). Five carcasses of common buzzards found in different areas were submitted to the Regional Diagnostic Lab on Avian Infl uenza after March 15, 2010; only 1 was carrying HPAIV (H5N1).
Common buzzards are considered territorial birds that usually do not migrate long distances. A 3-year study conducted in southern England showed that local radiotagged common buzzards forage within 1 km of their nests during their fi rst winter; most of the birds that do not disperse make only brief excursions before they opt for a stay-at-home strategy, and most of those that disperse return to their natal area during the following breeding season (37). Long-term ornithologic studies conducted during 1979-2005 in Bulgaria, however, showed unambiguously that extensive autumn and spring migrations of common buzzards (as many as 42,100 birds) occur on the western Black Sea Via Pontica fl yway (35,38). Kostadinova et al. reported ≈6 pairs of common buzzards breeding in the Batova habitat, but during autumn and spring migrations, ornithologists counted as many as 19,712 individuals of the species in the area (35). In most cases, common buzzards migrate singly or in loose fl ocks with other raptors, lesser spotted eagles, white pelicans, or black storks (38).
Migration of common buzzards in different parts of Europe appears to depend on the local climate; buzzards from northern Europe fl y to the western Black Sea area during the winter season, whereas buzzards from Bulgaria fl y south. Tracking bands belonging to common buzzards from Finland, Romania, and Israel have been found in Bulgaria (39). The buzzard infected with HPAIV (H5N1) in Bulgaria was not banded; however, even if it had been a migrant, the habitat nearest to St. Konstantin and Helena Black Sea Resort that provides different food sources is Batova, with an area of 38,132.8 ha (35). The migration of the common buzzards suggests that these birds are capable of spreading pathogens over long distances.
Our results show that chickens are highly susceptible to infl uenza virus A/common buzzard/Bulgaria/38WB/2010 (H5N1) and that the virus is highly pathogenic in them. Mammals appear not to be susceptible. Although buzzards can serve as intermediate hosts of HPAIV (H5N1) between migratory birds and poultry, the lack of gross pathologic fi ndings in the buzzard carcass we examined indicates that the bird died shortly after infection. Thus, in this case, the buzzard could not have served as a reservoir of infection to spread the virus over a long distance. Additionally, the lack of poultry farms within 10 km of the area where the buzzard carcass was found may partially explain why no outbreak occurred. As part of a regular avian infl uenza surveillance plan, we tested 1,709 cloacal and fecal samples from mule ducks that were collected monthly during January 1, 2010-April 30, 2010, from 64 farms in 5 regions of Bulgaria (Plovdiv, Pazardjik, Stara Zagora, Haskovo, and Dobrich). No notifi able avian infl uenza viruses were isolated from any sample.
Since it diffi cult to predict the consequences of the clade's arrival on the continent. For example, this clade is already widely distributed in Asia and is being perpetuated in many wild bird species, which is a prerequisite for long-distance distribution through migration. Wild bird species infected with HPAIV (H5N1) from clade 2.3.2.1 include gray herons, peregrine falcons, and great egrets in Hong Kong; whooper swans, ruby shelducks, and bar-headed geese in Mongolia; and grebes and black-headed gulls in Tyva (8).
The potential of clade 2.3.2.1 HPAIV (H5N1) to cause an outbreak is heightened because vaccines currently in use do not effi ciently protect poultry fl ocks from a strain of this clade that was recently identifi ed in Vietnam (40). Now that clade 2.3.2.1 has spread to Europe, implementing active surveillance plans in all high-risk areas and monitoring the wild birds in the region will play key roles in early detection of incidences of HPAIV (H5N1) infection and in prevention of outbreaks. The expansion of the geographic distribution of HPAIV (H5N1) in wild birds and poultry and the virus's repeated interspecies transmission to humans make this virus a substantial pandemic threat. This work was supported by National Institutes of Health contract no. HHSN266200700005C.
Dr Marinova-Petkova is a research associate at the Regional Diagnostic Laboratory on Avian Infl uenza and Newcastle Disease in Birds, Varna, Bulgaria, and a doctoral student at the National Diagnostic and Research Veterinary Medical Institute, Sofi a, Bulgaria. Her research interests include ecology and evolution of infl uenza viruses at the animal-human interface.