Wild bird surveillance around outbreaks of highly pathogenic avian influenza A ( H 5 N 8 ) virus in the Netherlands , 2014 , within the context of global flyways

J H Verhagen1, H P van der Jeugd2,3, B A Nolet2, R Slaterus4, S P Kharitonov5, P P de Vries2, O Vuong1, F Majoor4, T Kuiken1, R A Fouchier (r.fouchier@erasmusmc.nl)1 1. Erasmus MC, Department of Viroscience, Rotterdam, the Netherlands 2. Netherlands Institute of Ecology [NIOO-KNAW], Department of Animal Ecology, Wageningen, the Netherlands 3. Vogeltrekstation Dutch Centre for Avian Migration and Demography [NIOO-KNAW], Wageningen, the Netherlands 4. Sovon, Dutch Centre for Field Ornithology, Nijmegen, the Netherlands 5. Bird Ringing Centre of Russia, Moscow, Russia


Introduction
Wild aquatic birds are the natural reservoir for low pathogenic avian influenza A (LPAI) viruses, which are classified based on their surface proteins haemagglutinin (HA, H1-H16) and neuraminidase (NA, N1-N9) [1,2].These viruses can be carried over long distances along migratory flyways [3][4][5].LPAI viruses of the H5 and H7 subtype can evolve into highly pathogenic avian influenza (HPAI) viruses upon introduction into poultry.HPAI H5N8 viruses, such as A/duck/Jiangsu/ k1203/2010, were first detected in birds on live bird markets in China in 2010 [6].These H5N8 viruses contain genes derived from HPAI H5N1 viruses of the socalled A/Goose/Guangdong/1/1996 (GsGd) lineage [7] that have caused outbreaks in numerous countries of the eastern hemisphere since 1997.
In January 2014, HPAI H5N8 viruses were detected in South Korea, where they infected birds of 161 poultry farms and resulted in the culling of 14 million poultry by September 2014 [8].In April 2014, HPAI H5N8 virus was detected on a chicken farm in Japan.Over the summer of 2014, no new cases were reported outside South Korea.In September, HPAI H5N8 virus was detected in China in a domestic duck and an environmental sample.During the same month, H5N8 virus was also detected in north-eastern Russia in a Eurasian wigeon (Anas penelope).From November 2014 to February 2015, HPAI H5N8 virus has been found in poultry and/ or free-living wild birds in Asia (Japan and Taiwan), Europe (Germany, Hungary, Italy, the Netherlands and the United Kingdom (UK)), and North America (US) [9,10].HPAI H5N8 virus was also detected in captive wild birds: dead gyrfalcons (Falco rusticolus) in the north west of the United States (US) and white storks (Ciconia ciconia) in a zoo in Germany (Table 1) [11].The HA of HPAI H5N8 viruses detected in domestic and wild birds in Asia, Europe and North America belonged to the GsGd H5 clade 2.3.4.4 [12].Genetic closely related H5N8 viruses belonging to the same GsGd H5 clade 2.3.4.4 were detected in China since 2010.So far, HPAI H5N8 virus has been isolated from free-living wild birds of the orders Accipitriformes, Anseriformes, Charadriiformes, Falconiformes and Gruiformes in several countries including Germany, Japan, Russia, South Korea, Taiwan, the Netherlands,

Table 1
Global detection of highly pathogenic avian influenza A(H5N8) virus and other viruses belonging to the H5 clade 2. Data from [8,9].
and the US (Table 1).In live wild birds, H5N8 virus detections were limited to ducks (order: Anseriformes) of the species common teal (Anas crecca), mallard (Anas platyrhynchos), spot-billed duck (Anas poecilorhyncha), Eurasian wigeon, American wigeon (Anas americana) and gadwall (Anas strepera) [8,9] (Table 1).In addition, H5N8-virus-specific antibodies were detected in 10 to 53% of ducks of the species Baikal teal (Anas formosa), common teal, mallard, Eurasian wigeon and spot-billed duck in South Korea [8], suggesting that this virus had been circulating in these species for some time and that these individual birds had survived infection and thus may have played a role in the dispersal of H5N8 This study presents data on wild bird surveillance activities in the Netherlands that were intensified in the country, in response to the HPAI H5N8 virus outbreaks on poultry farms at the end of 2014.We present our findings in the perspective of the distribution and migratory flyways of H5N8-virus-positive bird species.[13].For cloaca and oropharynx samples, the number of tested birds depended on the bird species, capture method and capture success.For fresh faeces, swab samples were collected from flocks of single species.The number of faeces droppings sampled per flock was on average less than 40% of the total number of birds in the flock with at least one metre in between each dropping (to limit sampling the same individual twice).

Virus detection, isolation and characterisation
Samples for virus detection were analysed for presence of H5N8 virus using a matrix-specific and H5-specific polymerase chain reaction (PCR) followed by H5 sequencing.Samples that tested positive in matrixspecific PCR were used for virus isolation in embryonated chicken eggs as described previously [13].

Virus sequencing and phylogeny
Of the HPAI H5N8 viruses isolated within this study, the sequences of the complete genome were obtained and deposited in a public database (http://www.gisaid.com).Sequencing was performed using specific primers as described previously [14].

Wild bird surveillance activities to detect H5N8 virus in the Netherlands: newly acquired and historical data
Surveillance of avian influenza virus in wild birds in the Netherlands has been in place in the country since 1998.After the first HPAI H5N8 detection in poultry on 14 November 2014, activities to detect the virus were increased and a total of 4,018 wild birds of 25 different species belonging to five orders were sampled (Table 3).Of those, 623 birds (16%) were sampled within 10 km of farms previously affected by HPAI H5N8-virus.In the six months before the first detection of HPAI H5N8 in poultry, a total of 2,745 wild birds of nine different

Table 2A
Information on influenza A virus sequences obtained from the Global Initiative on Sharing Avian Influenza Data used for the study  The birds were located ca 15 to 28 km away from three of five H5N8-virus-infected poultry farms; the remaining two H5N8-virus-infected farms were located ca 80 km away.In the Netherlands, the affected poultry farms were located in wild-bird-rich areas where water is abundant and with low to medium poultry densities.
The distribution in time of sampled birds is shown per age, location, sample type and species in Figure 1.

Genetic analyses of H5N8 viruses
Genetic analyses of the HA and NA gene showed that H5N8 viruses from Europe and Russia were genetically most closely related to H5N8 viruses detected in Japan in November and December of 2014 followed by viruses detected in South Korea in 2014 (Figure 2).Also, genetic analyses of the HA gene showed that H5N8 viruses from North America were genetically most closely related to HPAI H5N2 and H5N1 viruses detected in North America followed by H5N8 virus detected in South Korea and Japan.The NA of North American H5N8 viruses was genetically most closely related to H5N8 viruses from South Korea and Japan (i.e.A/crane/Kagoshima/KU1/2014, Figure 2).
Genetic analyses of all gene segments showed that the gene constellation of H5N8 viruses from domestic and wild birds in Europe and from birds in North America was very similar to H5N8 viruses from domestic and wild birds in South Korea and Japan (data not shown).Of these viruses, four of eight gene segments (i.e.basic polymerase 2 (PB2), HA, nucleoprotein (NP) and NA) were derived from viruses similar to A/Duck/Jiangsu/ k1203/2010 (H5N8).Of those, PB2 and HA genes were derived from viruses of the HPAI H5 GsGd lineage.The remaining four gene segments (i.e.basic polymerase 1 (PB1), acidic polymerase (PA), matrix protein (MP) and non-structural protein (NS)) were derived from common LPAI viruses [6,7]   and wild birds [7,8,16].H5N8 viruses detected in Europe (Germany, Italy, the Netherlands, and the UK), Russia and in North America belonged to lineage A based on analyses of the HA gene [8].The close genetic relationship between European, Asian and North American isolates suggested that these H5N8 viruses have a common origin.

Distribution and migratory flyways of H5N8virus-positive bird species
Migrating birds from which H5N8 viruses have been isolated (Table 1) and that have circumpolar breeding grounds (e.g.northern pintail, Anas acuta) or that cover multiple major migratory flyways (e.g.Eurasian wigeon) are of specific interest with respect to global H5N8 virus epidemiology (Figure 3).Most of those species can be divided into distinct populations based on their geographically separate wintering areas.However, less is known about the degree of mixing among these populations in their breeding areas in Russia, and to which degree birds are loyal to their wintering areas.
Ring recoveries suggest that some waterfowl species (including ducks and geese) with populations wintering in East Asia and populations wintering in western Europe may have overlapping breeding grounds.For instance, ring recoveries of Eurasian wigeon and northern pintail ringed in Japan indicate that they migrate mostly north to north-east to the Russian Far East during spring migration, but a minority strays more north-west, some as far as the Western Siberian Lowlands [17] (Figure 3A and 3B).Here, ring recoveries indicate that some conspecifics originating from western Europe also may be found [18] (Figure 3A and 3B).Hence, although the probability of an actual meeting between east and west seems low, ring recoveries suggest it is not impossible.Furthermore, ring recoveries of Eurasian wigeon and northern pintail indicated a direct migratory connection between north Russia and north India (Figure 3A and 3B).Baikal teal and spotbilled duck, from which H5N8 viruses have also been isolated, have more restricted ranges, but could be involved in transport of virus from wintering grounds to breeding grounds in north-eastern Russia (Figure 3C  and 3D).Mallards and teals have extensive ranges, and potentially can also be involved in transport of virus, but ring-recovery data from Russia were not available (Figure 3E and 3F).
Ring recoveries and satellite tracking have shown various waterfowl species from East Asia to be in indirect and sometimes even direct migratory connection with North America.Satellite tracking and colour banding of various waterfowl species, including emperor goose (Chen canagica) [19], black brant (Branta bernicla nigricans) [20], lesser snow goose (Chen caerulescens caerulescens) [21] and northern pintail have shown them to cross the Bering Strait [22].Ring recoveries of northern pintail in particular show that the connection between East Asia and North America is quite strong, albeit most likely still indirect with contact zones in the Table 3B Wild bird species sampled for highly pathogenic avian influenza (HPAI) H5N8 virus before and after the first detection of HPAI H5N8 virus in poultry on 14 November 2014, the Netherlands, May Russian Far East and Wrangel Island [17,23].The same is true for some other species than waterfowl, which have not been identified as H5N8 virus hosts, but may play a role in the epidemiology of influenza, such as waders [24,25].

Discussion
The detection of the newly emerging HPAI H5N8 virus in at least 17 migratory bird species in Asia, Europe and North America, emphasises the need to study the role of migratory birds in the epidemiology of these H5N8 viruses.After the first detection of H5N8 virus in poultry in the Netherlands, wild bird sampling activities were intensified and HPAI H5N8 virus was detected in samples from two of 4,018 birds sampled within a three months period.The virus was isolated from Eurasian wigeons exclusively, whereas other bird species like mallards, white-fronted geese (Anser albifrons), black-headed gulls (Chroicocephalus ridibundus) and common coots (Fulica atra) also had been sampled intensively.The Eurasian wigeon is a long-distance migrant in which species H5N8-virus-specific antibodies had been detected in South Korea in 2014 [8].As HPAI H5N8 virus, like other avian influenza viruses, causes an infection of short duration in birds [26], the chance of detection is low and large sample sizes are needed to determine its presence in the population.The chance of detection of H5N8-virus-specific antibodies in wild bird sera is much higher, and serology can be used as a tool to target surveillance and determine past exposure to H5N8 virus, as H5 viruses of the HPAI GsGd lineage differ antigenically from common LPAI H5 viruses [27].
The H5N8 viruses isolated from wild birds in the Netherlands were genetically closely related to and had   Breeding and wintering ranges are reproduced from [30].Breeding ranges of Baikal teal and spot-billed duck have been updated from [31].

A B C D E F
the same gene constellation as H5N8 viruses detected elsewhere in Europe, in Asia and in North America, suggesting a common origin.In wild and domestic birds in North America, HPAI reassortant viruses of the subtypes H5N2 and H5N1 have been detected.These viruses contain genes originating from both HPAI H5N8 and LPAI viruses.Reassortant viruses of the subtypes H5N2 and H5N3 have been detected in domestic birds in Taiwan.In Europe, no reassortant viruses with HPAI H5N8 genes have been detected so far.Monitoring wild birds to detect H5N8 virus and derived reassortants is warranted given their potential to cause severe disease and mortality in poultry and some species of wild birds (e.g.eagles and hawks).
Ring recoveries of migratory duck species from which H5N8 viruses have been isolated provide evidence for indirect migratory connections between East Asia and western Europe and between East Asia and North America.In addition, ring recoveries of northern pintails and Eurasian wigeons demonstrated a direct migratory connection between north India and north Russia and between north India and Europe.If these species are involved in the global spread of H5N8 virus, we hypothesise that H5N8 viruses may also spread to north India as occurred previously with HPAI H5N1 virus of clade 2.2 [28].During large-scale surveillance activities in north India from 2009 to 2011, no avian influenza viruses had been detected in 3,522 wild bird samples [29].To which extent migrating bird populations of different flyways come in direct or indirect contact (e.g. using the same water source during stop over) with each other needs further study.To understand the role of wild birds in the epidemiology of H5N8 virus, sampling activities need to aim at detection of both the virus and specific antibodies with an emphasis on migrating birds in north-east Europe, Russia, and north China.

Figure 1
Figure 1Monthly sampling of wild birds for H5N8 virus detection, by species, location, age, and sample type, the Netherlands, 14 May 2014-20 February 2015 (n=6,763)

Figure 3
Figure 3Breeding and wintering range and ring recoveries from 1940-2010 a of wild duck species from which highly pathogenic avian influenza (HPAI) H5N8 viruses have been isolated Information on influenza A virus sequences obtained from the Global Initiative on Sharing Avian Influenza Data used for the study . Nucleotide sequence identity per segment between European, North American and the genetically closest Asian relatives was high (i.e.99 to 100% identical).Two genetic lineages (A and B) of H5N8 virus were identified in both domestic and wild birds from South Korea in January 2014, of which lineage A was more frequently detected in both domestic Table2BCA: Canada; CN: China; DE: Germany; IT: Italy; JP: Japan; KR: South Korea; NL: Netherlands; RU: Russia; UK: United Kingdom.