1. Introduction
Since 1997, H5 subtype high pathogenicity avian influenza (HPAI) viruses (HPAIVs) have repeatedly caused epidemics of HPAI and have caused massive damage to the poultry industry around the world. Wild aquatic birds belonging to the orders
Anseriformes and
Charadriiformes, which are the reservoir hosts of avian influenza viruses (AIVs), should be considered as potential original sources of HPAIV to poultry. However, direct contact with wild birds and poultry birds in poultry houses is unlikely to occur, especially in modern poultry systems that apply strict biosecurity measures. There could be intermediate factors in HPAIV transmission from wild birds to poultry. Potential vectors of HPAIV introduction may include feed, water, aviary equipment, the clothing and shoes of farm workers, flying insects, small birds, and mammals. Anything that could contact the feces or other excrement of HPAIV-carrying waterfowl should be considered a source or route of transmission of HPAIVs. To design more practical biosecurity measures in poultry operations, the route by which HPAIV is introduced to farms must be elucidated. In this regard, synanthropic animals that are frequently found around poultry farms, such as terrestrial wild birds and wild and domestic mammals, are some of the most potent risk factors [
1,
2,
3].
Wild rodents, such as house mice (
Mus musculus), brown rats (
Rattus norvegicus), and black rats (
Rattus rattus), may be the most prevalent mammals in the world’s poultry farms. Animals that are frequently observed on farms are more likely to be at risk than those that are rarely observed on farms [
4]. Bacterial infections such as
Salmonella,
Campylobacter, and
Escherichia coli are known as rodent-borne diseases on poultry farms. Wild rodents are also believed to play a role in the transmission of viruses such as AIVs, avian paramyxovirus 1, avian gammacoronavirus, and infectious bursal disease virus. Therefore, rodent control programs have been empirically and aggressively recommended as biosecurity measures for poultry operations to prevent the introduction and spread of rodent-mediated pathogens in farms. In the case of avian influenza, wild rodents are suspected to be HPAI vectors in poultry farms [
4,
5]. During the initial outbreak of H5N1 subtype HPAIV in Hong Kong in 1997, HI antibodies were detected in the sera of some brown rats captured in live poultry markets, although a virus was not isolated [
6]. In 2015, an epidemiological study of house mice on farms experiencing HPAI outbreaks in Ohio, USA, did not find any virus-positive or antibody-positive cases [
3]. However, in May 2024, H5N1 subtype HPAIV infections were reported in 66 house mice in New Mexico, USA, where H5N1 subtype HPAIV infections had occurred in dairy cattle (
https://wahis.woah.org/#/in-review/4451 (accessed on 10 July 2024)). Although few in number, these cases of natural infection in brown rats and house mice imply that at least some synanthropic rodents are susceptible to HPAIV.
Wild house mice and brown rats that were experimentally inoculated with low pathogenicity AIVs were shown to be susceptible to avian-origin viruses and productively replicate viruses in their respiratory tract [
7,
8]. On the other hand, a few studies have shown a direct link between wild rodents and HPAIVs. In bank voles (
Myodes glareolus), which are wild rodents sharing habitat with the wild reservoirs of AIVs, inoculation with avian-origin H5N1 and H7N1 subtype HPAIVs caused asymptomatic infection and resulted in the shedding of high amounts of virus [
6,
9]. In a black rat experimentally infected with H5N1 subtype HPAIV, a small amount of virus was isolated in the kidney and colon 3 days after infection [
2]. In that study, the amount and duration of viral shedding from HPAIV-inoculated black rats after the challenge was not evaluated. Likewise, the susceptibility of house mice and brown rats to HPAIVs has not yet been explored by experimental inoculation.
In the past few years, many cases of HPAIV infection in mammals have been reported, raising concerns about the risk of transmission to humans [
10]. The HPAIV susceptibility of rodents that live close to humans needs to be elucidated not only to protect poultry farms but also as a public health concern. In the present study, wild-caught house mice, brown rats, and black rats were experimentally inoculated with two H5N1 subtype HPAIVs, which had caused outbreaks in poultry and wild birds in the winters of 2007 and 2010, respectively, in Japan. They were observed to determine their susceptibility to HPAIVs as shown by clinical manifestations, viral shedding, and tissue distribution, as well as antibody response after challenge.
2. Materials and Methods
2.1. Animals (Wild and Laboratory Rodents)
Twenty-nine house mice (13 males and 16 females), 25 brown rats (13 males and 12 females), and 24 black rats (11 males and 13 females) captured and reared by Technical Research Laboratory, IKARI SHODOKU CO., LTD., in pest-control efforts were used in the present study. After transport to Tottori University, wild rodents were separately reared in custom-made animal cages (Natsume Seisakusho Co., Ltd., Tokyo, Japan). Because the body weights of most house mice ranged from 10 to 20 g, five house mice with less than 10 g body weight were separated as a small house mice group. Prior to virus challenge, blood was collected from all wild rodents using animal lancets (Medipoint, Inc., Mineola, NY, USA) under anesthesia. The absence of specific antibodies in their sera was confirmed by hemagglutination-inhibition (HI) tests against the challenge viruses, as described below.
For laboratory species of rodents, 22 four-week-old female BALB/c mice (Mus musculus) and 22 four-week-old female Sprague-Dawley (SD) rats (Rattus norvegicus) were purchased from Clea Japan, Inc., Tokyo, Japan.
2.2. Viruses
An HA clade 2.3.2.1 H5N1 subtype HPAIV, A/chicken/Shimane/1/2010 (Shima10), which was an index strain from the 2010–2011 poultry outbreaks in Japan [
11], was kindly provided by the National Institute of Animal Health, Japan. An HA clade 2.2 H5N1 subtype HPAIV, A/mountain hawk-eagle/Kumamoto/1/2007 (Kuma07), was isolated from a mountain hawk-eagle (
Nisaetus nipalensis) found dead on 4 January 2007 [
12]. This isolate is genetically close to poultry isolates such as A/chicken/Miyazaki/K11/2007 (H5N1), which caused HPAI outbreaks in poultry on January and February 2007 in Japan [
13]. The deduced amino acid sequence of the 2007 epidemic strains including Kuma07 have a lysine at amino acid 627 of the PB2 protein (PB2-627K), a mutation known to be associated with mammalian adaptation [
14,
15].
2.3. Inoculation of Rodents with HPAIVs
Eight each of house mice, black rats, brown rats, BALB/c mice, and SD rats were lightly anesthetized with 3 to 5% isoflurane gas inhalation and intranasally inoculated with 100 μL of 107 fifty-percent egg infectious dose (EID50) of Shima10 or Kuma07 diluted in phosphate-buffered saline (PBS). The inoculum volume for the house mice experiment was reduced to 50 μL, containing 107 EID50 of viruses. Three of the eight individuals of each species were euthanized at 3 days post-inoculation (dpi) (at 2 dpi for the house mice experiment) and the brain, nasal turbinate, trachea, lung, heart, spleen, liver, kidney, duodenum, rectum, and heparinized whole blood were collected for virus isolation. The remaining five animals were clinically observed for 12 days after virus challenge. Because of a scheduling problem, the observation period was 10 days for the SD rats inoculated with Shima10. Only for the BALB/c mice and SD rat experiments, another three animals as a control group were mock inoculated with 100 μL of PBS instead of viruses. Five small house mice with less than 10 g body weight were also challenged intranasally with 50 μL of 107 EID50 of Kuma07 and clinically observed for 12 days.
To monitor virus shedding, oral and rectal swabs were collected from five individuals under anesthetized conditions at 1, 2, 3, 5, 7, and 12 dpi and suspended in 1 mL of PBS containing 10,000 unit/mL penicillin (Meiji Seika Pharma, Tokyo, Japan) and 10,000 μg/mL streptomycin (Meiji Seika Pharma, Tokyo, Japan). If available, fresh fecal samples of house mice and laboratory mice were collected instead of rectal swabs.
Tissue samples and fecal samples were homogenized in PBS containing antibiotics to obtain 20% suspensions. The swab, whole blood, fecal, and tissue samples were inoculated into 10-day-old embryonated chicken eggs (ECEs) (Aoki Breeder Farm, Tochigi, Japan) for calculation of EID50. At the end of the observation period, post-challenge sera were collected from all surviving animals and tested for specific antibodies against challenge viruses using an HI test to confirm infection. All experiments with infectious HPAIVs were performed in the biosafety level-3 containment laboratory at the Avian Zoonosis Research Center, Faculty of Agriculture, Tottori University.
2.4. Viral Shedding of Wild Rodents after Exposure to Shima10
To further evaluate the risk of wild rodents as potent vectors of HPAIVs, the amount and length of oral shedding in wild rodents were examined after dose-dependent exposure to Shima10 HPAIV. In the case of house mice, 8 animals were divided into two groups of four mice and intranasally inoculated with 105 and 103 EID50 of Shima10. For brown rats, 9 animals were divided into three groups of 3 rats and intranasally inoculated with 105, 103, and 101 EID50 of Shima10. Eight black rats were intranasally inoculated with 105 (n = 3), 103 (n = 2), and 101 (n = 3) EID50 of Shima10. These animals were monitored for viral shedding in oral swabs for 7 days after challenge.
2.5. Hemagglutination-Inhibition (HI) Test Against the Challenge Viruses
The pre- and post-serum samples were tested for HI antibodies according to the standard method (WHO Manual on Animal Influenza Diagnosis and Surveillance,
http://www.who.int/iris/handle/10665/68026 (accessed on 9 May 2020)), with slight modifications. For this assay, a solution of receptor-destroying enzyme II (Denka Seiken Co., Niigata, Japan) was freshly prepared by dissolving 2 mg lyophilized enzyme in 20 mL sterile saline. For each assay, 75 μL of enzyme solution was mixed with 25 μL of serum. The mixture was incubated at 37 °C for 18 h and heat-inactivated at 56 °C for 30 min. Viruses such as hemagglutinin (HA) antigen were adjusted to 4 HA units/25 μL, and then the HI titer was measured with 0.5% packed chicken red blood cells. The lower limit of detection was 1:4 for this HI test (i.e., values less than 4 HI were undetectable), except for the SD rats after Kuma07 exposure, for whose serum the lower limit was 1:10.
4. Discussion
Wild house mice, brown rats, and black rats, which are abundant in most poultry farms, are thought to have been a contributing factor to HPAI outbreaks across the world [
4,
5]. However, little is known about their HPAIV susceptibility, clinical signs, route of viral shedding, or duration of shedding. To the best of our knowledge, the only rodents that have been experimentally infected with HPAIV are black rats and bank voles. One study [
2] showed black rats have low susceptibility to various H5N1 subtype HPAIVs, since little or no viruses were recovered at 3 dpi. Another study [
9] showed that bank voles, which are common wild mice throughout Europe and Asia, were infected asymptomatically with H5N1 and H7N1 subtype HPAIVs and shed viruses in their nasal washes. In the present study, we infected a larger variety of rodents with two H5N1 subtype HPAIVs. Shima10 belongs to HA clade 2.3.2.1 HPAIVs, which caused a nationwide epidemic among poultry and wild birds in Japan in the 2010–2011 season [
11]. Kuma07 belongs to HA clade 2.2 HPAIVs (formerly called Qinghai Lake lineage), which were isolated from a raptor and poultry in Japan during the 2006–2007 season [
12]. The present results strongly suggest that synanthropic wild rodents are susceptible to infection of avian-origin H5N1 subtype HPAIVs. Detection of infectious virus in the oral cavity indicates that wild rodents exposed to HPAIVs shed virus through saliva and could contaminate food, water, and equipment in the poultry house.
Among the three rodent species examined in this study, black rats appeared to shed the most virus and for the longest duration (
Figure 2) and thus were assumed to have the most potential to transmit HPAIVs to poultry. Hiono et al. (2016) infected four black rats with HA clade 2.3.2.1 virus but detected the virus in only one of them, and in that rat, the virus loads were extremely low and confined to the kidney and intestine [
2]. In the present study, our finding of HA clade 2.3.2.1 or clade 2.2 viruses in the nasal turbinates or lungs, but not the kidney or intestine, of black rats (
Table 1) suggests that the viruses replicates in respiratory tissues, although it was unclear which cells were infected. The analysis of oral swabs collected periodically revealed that black rats shed viruses for several days after inoculation of both HPAIVs. Stable viral shedding was observed in black rats even after exposure to smaller doses (10
5 EID
50) of Shima10 (
Figure 3C). Although the virus strains and the amount of virus inoculum differed from the strains and amount used in the previous study [
2], the present results indicate that black rats are more susceptible to HPAIVs than previously reported [
2]. The differences of viral shedding among species or individuals are likely partly due to differences in innate immune responses to H5N1 subtype HPAIVs.
As for the Shima10 virus, a fifty percent chicken infectious dose (CID
50) for 4-week-old specific pathogen-free (SPF) chickens was calculated as 10
3 EID
50 [
11], which was detected in oral swabs from one black rat (
Figure 2F). Although the CID
50 of Kuma07 was not defined, that of an HA clade 2.2 H5N1 virus (A/chicken/Miyazaki/K11/2007), which is genetically close to Kuma07, was shown to be 10
2.5 EID
50 using 4-week-old SPF chickens [
13]. In the Kuma07 experiments, two black rats, two brown rats, and one house mouse discharged virus exceeding the deemed CID
50 (10
2.5 EID
50) at 1 dpi (
Figure 2B,E,G). A high titer of virus shedding (10
3 EID
50) was also detected in one of the black rats at 2 dpi. These results suggest that some wild rodents, notably black rats, exposed to HPAIV shed enough virus from the oral cavity to infect chickens.
While wild mice and rats stayed active in their cages after HPAIV challenge, laboratory BALB/c mice showed marked weight loss and changes in appearance, such as ruffled fur and hunched posture, and finally died. This noticeable difference between biologically identical species (
Mus musculus) could be explained in part by the finding that the immune systems of wild mice are more active than those of laboratory bred mice [
16]. Although a high titer of Kuma07 virus exceeding the assumed CID
50 (10
2.5 EID
50) was detected in oral swabs of BALB/c mice at 3 dpi (
Figure 2I), these mice did not really contaminate the ambient surroundings because they had laid still from 2 dpi to death at 4-5 dpi. If virus carriers were too sick to enter and run around poultry houses or died outside the poultry house, the risk of transmission and spread of viruses would be low. The lack of clinical manifestations in SD rats appears to resemble those of black rats and other wild rodents, although the length and extent of viral shedding were slightly less compared to black rats (
Figure 2J,K). Bank voles were also shown to be infected asymptomatically with HPAIV and shed viruses in their nasal washes [
9]. From these findings, based on the challenge experiments, it was shown that wild rodents are subclinically infected with HPAIVs. This means that wild synanthropic rodents after HPAIV exposure can pollute the environment of poultry premises by roaming around while shedding infectious viruses and thus play a role transmitting or spreading viruses.
Even though wild rodents can be infected subclinically with HPAIVs, there are exceptions, like the small house mice that died suddenly after Kuma07 HPAIV infection. The wild house mice had greater interindividual variation in their immune ability [
16]. Immune status may be compromised by other pathogens or stress from the environment. Thus, there are cases in which wild animals, especially immature or compromised animals, die after HPAIV infection. Immunity in wildlife is still largely unexplored. But a better understanding of the viral susceptibility of wild animals requires analysis of innate and acquired immunity. Postmortem observations revealed that Kuma07 HPAIVs highly replicated in the lungs and hearts of the dead house mice (
Table 1), indicating that the viruses might be transmitted to poultry through pecking at the rodent carcasses, in which viable viruses remained.
More recently, natural infection of house mice with HA clade 2.3.4.4b H5N1 subtype HPAIV was reported to the World Organization for Animal Health (WOAH) in June 2024 (
https://wahis.woah.org/#/in-review/4451 (accessed on 10 July 2024)). The route of infection has not been elucidated, but it is assumed that the infection was caused by ingestion of unpasteurized milk from HPAIV-infected dairy cattle [
17]. It was previously reported that influenza virus nucleic acids were detected in Norway rats (brown rat) inhabiting a city area [
18]. Both cases imply that the source of infection is not related to waterfowl reservoirs. However, these findings provided a direct connection between wild rodents and influenza viruses. If wild rodents are infected with HPAIVs by association with waterfowl in ponds or puddles around poultry farms, they will remain active and will enter and contaminate poultry houses with HPAIVs through their saliva or respiratory excretions. That is one way in which HPAIVs can be transmitted from wild rodents to poultry. Alternatively, wild rodents may also spread HPAIVs between poultry houses. In this case, wild rodents might be infected with HPAIVs, or they might pick up the virus on their fur or palms in poultry houses containing infected poultry. Rodents carrying HPAIV may also be a source of infection for predators such as weasels and martens, which target rodents and invade poultry houses [
19].
In the present study, no HI antibodies were detected in the sera of any of the Shima10-inoculated group (
Table 2). In some studies in which wild rodents were experimentally infected with avian-origin viruses, antibodies to the viruses were not detected. These included a black rat infected with HA clade 2.3.4.4 H5N1 subtype HPAIV [
2] and Norway rats infected with LPAIVs [
7]. These viruses were used in inoculation trials prior to adaptation to mammals. Kuma07 has a lysine residue at position 627 in PB2 (627K), which reportedly contributes to the adaptation to mammals [
14,
15,
20], while Shima10 has a glutamate at this position (627E). The reason for the lack of antibody response in Shima10-inoculated groups remains unclear. However, Shima10 was less adapted to mammals than Kuma07. Shima10 and Kuma07 do not harbor the HA mutations Q226L and G228S, which have been implicated in increasing the affinity to human-type α2,6 receptor [
21]. If Shima10 acquires these adaptive mutations after infecting rodents, it may change enough to induce antibody production. Future experimental infections of rodents with HPAIVs should monitor the well-known PB2 mutation (E627K) and HA mutations (Q226L and G228S) to estimate the potential adaptive capacity of the viruses.
Demonstrating viral infection at the tissue level requires genetic detection, virus isolation, or immunohistochemical staining [
22]. A weakness of this study was that Shima10 infection was not confirmed by histopathological examinations or detection of an antibody response. However, our results strongly suggest that both Shima10 and Kuma07 can replicate in the respiratory tract of the synanthropic wild rodents. Future studies of susceptibility to HPAIV in wild rodents and other mammals should use a variety of epidemic strains, especially HA clade 2.3.4.4b lineage viruses that have been circulating in recent years [
10].