Raccoons are considered an invasive alien species in Japan. Recently, the growing population of feral raccoons has resulted in significant agricultural damage and prompted the initiation of eradication programs in several areas. We used a total of 1,088 serum samples collected from animals captured under this official eradication program over 3 periods in the western region of Japan and 1 period in eastern Japan during 2005–2009 for a serologic survey of avian influenza virus (H5N1) infection (Table 1). To detect antibodies specific to the H5 hemagglutinin (HA) in the serum samples, we performed a virus neutralization (VN) test (6) with 2 subtype H5N1 viruses, A/Indonesia/3006/2005 (clade 2.1.3) and A/whooper swan/Mongolia/4/2005 (clade 2.2). As an initial screening step, we used the serum specimens (1:5 dilution) after receptor-destroying enzyme treatment of the serum to remove nonspecific inhibitors. The VN antibody-positive serum samples were then further tested for their reactivity by using a panel of influenza viruses of multiple subtypes (Table 2) as well as Western blot analysis (Figure 1). In these assays, we found a total of 10 serum specimens that were positive for VN antibody to subtype H5N1 viruses, representing 0.9% positivity. The A-1 to A-6 serum specimens, which were collected from animals captured within a 10 km² area, strongly reacted to A/whooper swan/Mongolia/4/2005 (clade 2.2) and more weakly to other clades of subtype H5N1, H5N2, and H5N3 viruses. These serum specimens did not react to viruses of other HA subtypes, including H1, H3, H7, and H9. Of note, the A-2, A-3, and A-4 animals were from the same litter captured at a lair, which suggests that the detected VN antibodies in these samples might be maternal antibodies from their uncaptured mother, who may have been infected with a subtype H5N1 virus. It is possible that 2 viruses of clade 2.2, which had slightly different antigenicities, may have infected raccoons in this area, as indicated by the different patterns of cross-reactive VN titers to subtype H5N1 clade 1 and H5N3 viruses. One group consisted of A-1 to A-4 and the other of A-5 and A-6. The B-1 and B-2 samples from animals captured at a 25-km distance strongly reacted to both subtype H5N1 clades 2.2 and 2.5 viruses. Given that the subtype H5N1 clade 2.5 virus has not circulated since 2004 and that the clade 2.2 virus was more highly reactive than the clade 2.5 virus, these raccoons were likely infected with clade 2.2 viruses, as supported by timing with poultry outbreaks. By contrast, the C-1 and C-2 samples from raccoons captured in eastern Japan reacted strongly to A/whooper swan/Akita/1/2008 (clade 2.3.2), unlike the samples from western Japan, indicating that the C-1 and C-2 animals were infected with a virus of this clade. Together, these data suggest that feral raccoons in Japan have been infected with subtype H5N1 viruses of different clades.

To assess the presence of anti-neuraminidase (NA) antibodies in the serum samples, we used an NA-inhibition (NI) assay for the VN-positive samples and found marked inhibition of the NA activity of the N1 subtype (Figure 2). We also performed the standard NI assay using another N1 virus, A/swine/Iowa/15/30 (H1N1), to avoid nonspecific NA inhibition by H5 antibodies, for 2 VN-positive serum specimens (A-6 and C-2) and found that A-6 and C-2 had positive NI titers of 20 and 80, respectively. These data demonstrate that VN-positive raccoon serum specimens contain anti-H5N1 antibodies, indicating that raccoons have been infected with subtype H5N1 viruses.

Japan has experienced 3 outbreaks of highly pathogenic subtype H5N1 viruses. In the first in early 2004, clade 2.5 subtype H5N1 viruses were detected in poultry farms in western Japan. The second, in early 2007, involved the isolation of clade 2.2 subtype H5N1 viruses from poultry in western Japan. The third occurred in mid-2008, when clade 2.3.2 viruses were isolated from diseased swans in the lakes in the northern area of eastern Japan. All of these outbreaks were contained by prompt culling of birds. Since 2008, subtype H5N1 viruses have not been reported in any poultry or wild migratory birds under the government surveillance program. Our data indicate that raccoons in western Japan were likely infected with the clade 2.2 viruses, whereas those in eastern Japan were infected with the clade 2.3.2 virus. Notably, some antibody-positive raccoons in western Japan were captured 6 months before the poultry outbreak with clade 2.2 virus, suggesting that a clade 2.2 subtype H5N1 virus had invaded Japan by 2006.

We cannot determine by seropositive text results the exact date when the viruses infected the raccoons, because the duration of naturally acquired antibody to subtype H5N1 virus in this species is unknown. Recent data indicate that this animal maintains a detectable serum antibody response for at least 9 months after natural exposure to influenza viruses of other HA subtypes such as H1, H3, and H4 (7). In humans, a detectable antibody response to seasonal viruses can last >5 years (8) and in swine antibodies to the virus have been detected 28 months postinfection (9).

Because wild raccoons are omnivores and highly opportunistic at exploiting foods they prefer, whenever available they could eat diseased or dead migratory birds from areas where subtype H5N1 viruses are enzootic. They also sometimes attack poultry farms for food, creating the potential to transmit virus to domestic poultry. In addition, the increasing likelihood for contact between wild raccoons and humans elevates the possibility of human infection with these viruses, posing risks to public health and increasing the possibility of the emergence of mammalian-adapted mutant viruses with pandemic potential. Further investigation and surveillance of influenza virus infections in peridomestic animal species are needed to better understand influenza ecology.