Guinea Pig Model for Evaluating the Potential Public Health Risk of Swine and Avian Influenza Viruses

Background The influenza viruses circulating in animals sporadically transmit to humans and pose pandemic threats. Animal models to evaluate the potential public health risk potential of these viruses are needed. Methodology/Principal Findings We investigated the guinea pig as a mammalian model for the study of the replication and transmission characteristics of selected swine H1N1, H1N2, H3N2 and avian H9N2 influenza viruses, compared to those of pandemic (H1N1) 2009 and seasonal human H1N1, H3N2 influenza viruses. The swine and avian influenza viruses investigated were restricted to the respiratory system of guinea pigs and shed at high titers in nasal tracts without prior adaptation, similar to human strains. None of the swine and avian influenza viruses showed transmissibility among guinea pigs; in contrast, pandemic (H1N1) 2009 virus transmitted from infected guinea pigs to all animals and seasonal human influenza viruses could also horizontally transmit in guinea pigs. The analysis of the receptor distribution in the guinea pig respiratory tissues by lectin histochemistry indicated that both SAα2,3-Gal and SAα2,6-Gal receptors widely presented in the nasal tract and the trachea, while SAα2,3-Gal receptor was the main receptor in the lung. Conclusions/Significance We propose that the guinea pig could serve as a useful mammalian model to evaluate the potential public health threat of swine and avian influenza viruses.


Introduction
Five influenza pandemics have resulted in significant morbidity and mortality around the world. Genetic analyses revealed that these pandemic strains were entirely or partially derived from the viruses of animal origin, including swine and avian strains [1]. So far, animal influenza viruses have repeatedly transmitted to humans with increasing genetic diversity, such as avian H5N1 and H9N2 influenza viruses, and swine H1N1, H1N2, H3N2 influenza viruses [2,3], which served as an important reminder that another influenza pandemic is highly likely. Among these swine and avian influenza viruses, in addition to highly pathogenic H5N1 avian influenza viruses, many low pathogenic viruses are also high on the list of candidates that could potentially cause the next human pandemic. The efficiency of transmission is a key factor in determining the severity of influenza epidemics. Despite evidence for limited human-to-human transmission of some animal influenza viruses, they have yet to exhibit sustained transmission among humans [4,5]. However, with typical human flu-like illness in humans [6,7] that may be unrecognized, these less pathogenic animal viruses can develop the ability of human-tohuman transmission. Thus, a mammalian disease model is needed to evaluate the epidemic potential of these viruses.
Mice are a common mammalian model used in influenza research but are not ideal for transmission studies and even the 1918 pandemic strain could not transmit between mice [8]. Ferrets, however, are susceptible to influenza infection and develop similar symptoms to humans and effectively transmit between individuals. Nevertheless, the ferret model presents several practical disadvantages [9]. Recently, guinea pigs have been shown to be an acceptable alternative mammalian model for the study of human influenza A virus transmission [10][11][12]. Guinea pigs are an attractive model host due to their high susceptibility to and ability to transmit low-passage human isolates (in contrast to mice), and based on several practical considerations such as size and cost, make them considerably more convenient for research purposes than ferrets [8]. Additionally, guinea pigs were also proposed as a useful model to understand the virus-host interactions in influenza A virus infection [13] because they possess similar innervations in the airway to humans [14] and also have human-like bronchusassociated lymphoid tissue in lungs [15].
Host receptor expression is one of the important factors influence the infection of the respiratory tracts by influenza viruses. The HA protein of influenza viruses initiates infection by binding sialic acid (SA) which are bound to glycans through an a2,3 or a2,6 linkage [16]. In general, human influenza viruses have a binding preference for SAa2,6-Gal and avian influenza viruses preferentially binding SAa2,3-Gal [17,18]. The distribution of these two forms of sialic acid compounds is known to be cell and species specific. Thus, host receptor distribution is an important factor in consideration when evaluating an animal model.
Thus far, influenza studies using the guinea pig model have been performed with 1918 pandemic [19], 2009 pandemic H1N1 [11], seasonal human H1N1 and H3N2 [10,13], and avian H5N1 influenza viruses [15,20,21], but little is known about other swine and avian influenza virus infections, particularly those isolated in recent years. To evaluate the applicability of the guinea pig model for swine and avian influenza viruses, we describe the infectivity and transmissibility in guinea pigs of the subtypes of viruses that repeatedly transmit to humans (swine H1N1, H1N2, H3N2 and avian H9N2 influenza viruses) compared with 2009 pandemic and human seasonal influenza viruses. Additionally, the expression of SAa2,3-Gal and SAa2,6-Gal linked receptors in guinea pig respiratory tissues were determined.

Replication of human, swine and avian influenza viruses in guinea pigs
To evaluate the replication of human, swine and avian influenza viruses in guinea pigs, groups of 15 animals were anesthetized and intranasally inoculated with 10 6 50% egg infective dose (EID 50 ) of each virus listed in Table 1. On days 2, 4, 6, and 8 post-inoculation (p.i.), three animals from each group were anesthetized and the tracheas, lungs, hearts, livers, spleens, kidneys, brains, and colons were collected for virus titration. Nasal washes from the remaining three animals were collected on days 2, 4, 6, and 8 p.i., and titrated by EID 50 assay. The study indicated that, similar to human influenza viruses, all swine and avian influenza viruses used in this study replicated efficiently in the nasal passages of guinea pigs, with peak virus titers in nasal washings ranging from 10 4.4 to 10 6.6 EID 50 /mL. The virus titer of BJ/317/09 (H1N1) was lower than those of other pandemic 2009 (H1N1) virus reported previously [22], possibly due to strain specificity. Most viruses grew to peak titers on day 2 p.i. in the nasal passages of inoculated guinea pigs and dropped to undetectable levels by day 8 p.i.. However, Sw/ GD/211/06 (H3N2), Sw/GD/968/06 (H3N2) and Qa/HK/G1/ 97 (H9N2) strains possessed different shedding kinetics and were cleared more slowly from the guinea pig nasal passages than other viruses ( Fig. 1 A-D). The nasal virus titers of these three virusinfected animals remained high (.10 5.0 EID 50 /mL) on day 6 p.i.. In particular, the Qa/HK/G1/97 (H9N2) virus still replicated well in nasal passages on day 8 p.i. with mean titers of 10 5.3 EID 50 /mL.
Compared to nasal passages, tracheas and lungs had much lower virus titers and viruses were cleared more quickly from these organs ( Table 2). Two human seasonal influenza viruses were not detected in the lower respiratory tract, while the 2009 pandemic H1N1 virus could replicate in the lung. For animal influenza viruses, only the Sw/FJ/204/07 (H1N1), Sw/GD/1222/06 (H1N2), Qa/HK/G1/ 97 (H9N2) and Ck/SD/1/06 (H9N2) viruses could be detected in the lung. All of these viruses could not be detected in tracheas and lungs on day 8 p.i.. It is interesting that the European avian-like swine H1N1 (Sw/FJ/204/07) and North American triple H1N2 (Sw/GD/1222/06) viruses, that were proposed to be the progenitors of 2009 pandemic H1N1 influenza viruses, both replicated more efficiently and distributed in more tissues than other swine influenza viruses, similar to the 2009 pandemic H1N1 virus (BJ/317/09). No virus was detected in other organs, including heart, liver, spleen, kidney, brain and colon during the observation period. Taken together, the replication of tested influenza viruses was restricted to the respiratory system of guinea pigs and virus replication in the upper respiratory tract was higher than in the lower respiratory tracts. Similar to human strains, swine and avian influenza viruses replicated efficiently in nasal tracts without prior adaptation.

Histopathology of human, swine, and avian influenza viruses in guinea pigs
To compare the histopathological changes in human, swine, and avian influenza viruses in guinea pigs, the nasal, trachea, and lung specimens from infected animals on day 4 p.i. were fixed in 10% neutral buffered formalin and processed for routine histology. The representative histopathology changes are shown in Fig. S1. A-L. Nasal lesions in infected guinea pig were characterized by phlebectasia and congestion in the submucosal capillaries and veins ( Fig S1 A-C), dropout of the mucous epithelium and erythrocytes adhering to the surface of the mucosa (Fig S1 D). The tracheal mucous membrane showed edema and thickening (Fig S1 E) and dropout of tracheal epithelial cells (Fig S1 G, H). The infected lungs showed bronchiolitis with inflammatory cellular infiltrate and dropout of mucous epithelial cells in the bronchioles (Fig S1 I), peribronchiolitis was also observed with edema and inflammatory cells infiltrate around bronchioles and blood vessels (Fig S1 I, K).
The results suggested that the tested viruses showed varying levels of pathological lesions. In particular, the histopathological changes in guinea pigs infected with Qa/HK/G1/97 were more severe than those caused by other viruses. In this group, the epithelial cells of the tracheal mucosa were severely desquamated resulting in exposure of the lamina propria (Fig S1 H). Severe bronchiolitis, peribronchiolitis, and bronchopneumonia were also observed with alveolar lumens and bronchioles flooded with edema fluid mixed with fibrin, erythrocytes, and inflammatory cells (Fig S1 L). For the tissues in virus was not detected, lesions were mild or absent (Fig S1 F, J).

Transmission of human, swine, and avian influenza viruses in guinea pigs
We next tested the propensity of the viruses to be transmitted by direct contact. Briefly, groups of three animals were intranasally inoculated with 10 6 EID 50 of tested virus and housed in a cage. At 24 h p.i., three naïve guinea pigs were placed in the same cage for each virus. On days 2, 4, 6, and 8 p.i., the nasal washings of the three inoculated animals and three contact animals were collected and titered by EID 50 assay. Evidence of transmission was based on the detection of virus in the nasal washing and seroconversion at the end of the three-week observation period. All of the viruses replicated efficiently in inoculated animals, as for the replication study; however, the transmissibility of these viruses in guinea pigs was variable. As shown in Fig. 2 and Table 1, transmission of BJ/317/09 (H1N1) virus occurred with 100% efficiency: all three exposed animals were detected to be positive on day 3 post-exposure (corresponding to day 4 p.i.). The two seasonal human influenza viruses transmitted less efficiently with only one of three and two of three exposed animals becoming infected, respectively. However, no virus shedding or seroconversion was detected in any direct contact animal of swine and avian influenza viruses. These results demonstrate that although the swine and avian influenza viruses replicated effectively in the inoculated guinea pigs, they lacked successful direct contact transmissibility. As no contact transmission occurred for swine and avian influenza viruses, we did not test the propensity of influenza viruses to be transmitted via the aerosol route. Of note, we observed that guinea pigs in the inoculated group and contact group of BJ/317/09 (H1N1) infections began to sneeze on days 3 and 4 p.i., respectively. Some guinea pigs in the GD/41/06 (H1N1) infected group also sneezed during the observation period. Except for sneezing, no other clinical abnormalities were observed in any of the animals during the observation period.

Receptor distribution in the respiratory system of guinea pigs
We examined the receptor binding specificity in the respiratory tract of guinea pigs by lectin histochemistry assay. Biotinylated Maackia amurensis lectin II (MAA II), and Sambucus nigra agglutinin (SNA) were used to detect SAa2,3-Gal and SAa2,6-Gal linked receptors, respectively. In the nasal tract, both the SAa2,3-Gal and SAa2,6-Gal receptors were widely expressed by stratified squamous epithelial cells, vascular endothelial cells and the epithelial cells of the serous gland (Fig. 3 A, B). In the trachea, both of SAa2,3-Gal and SAa2,6-Gal receptors were widely expressed by epithelial cells of the tracheal mucosa (Fig. 3 D, E). In the lungs, SAa2,3-Gal receptor was mainly expressed by the alveolar cells and the vascular endothelial cells and was also expressed by the bronchial mucosa epithelial cells, however, SAa2,6-Gal receptor was only rarely expressed in the alveolar cells (Fig. 3 G, H). Taken together, the data suggested that SAa2,3-Gal and SAa2,6-Gal receptors were present in the nasal and tracheal areas of the guinea pig. In contrast, SAa2,3-Gal was the dominant receptor type in the lung.

Discussion
In the present study, we investigated the guinea pig as a mammalian model for the study of the replication and transmission characteristics of selected swine H1N1, H1N2, H3N2 and avian H9N2 influenza viruses. We compared these characteristics to those of pandemic (H1N1) 2009, and seasonal human H1N1 and H3N2 influenza viruses. Our results showed that all the tested swine and avian influenza viruses replicated effectively in guinea pig nasal tracts without prior adaptation and the replication of these viruses was restricted to the respiratory system. The replication characteristics of swine and avian influenza viruses were similar to those of pandemic (H1N1) 2009 and seasonal human H1N1, H3N2 influenza viruses in addition to those of other influenza viruses reported by previous studies [13,19,21]. Moreover, some swine and avian viruses (namely, Sw/GD/211/ 06 (H3N2), Sw/GD/968/06 (H3N2) and Qa/HK/G1/97 (H9N2) viruses) were cleared more slowly from the guinea pig nasal passages than pandemic (H1N1) 2009 and seasonal human viruses. Histopathological analyses indicated that no significantly different pathological changes between human, swine and avian influenza viruses upon infection of the guinea pig. Of note, however, the histopathology after Qa/HK/G1/97 (H9N2) infection was the most severe of the viruses tested.
Pandemic (H1N1) 2009 influenza viruses possess effective human-to-human transmission ability and rapidly spread around the world. Similar to the transmissibility in humans, the transmission results in the present study in guinea pigs indicated that the 2009 pandemic virus can be passed from infected guinea pigs to all the uninfected contacts. Steel et al. has also demonstrate that transmission of a pandemic (H1N1) 2009 strain occurred with 100% efficiency in guinea pigs [11]. Two human seasonal viruses also showed transmission between guinea pigs. Additionally, Lamb et al. reported that during the 1918 epidemic, a parallel outbreak of pneumonial disease among their laboratory guinea pigs emerged [23]. Although sporadic instances of human infection with swine  and avian influenza viruses occur, human-to-human transmission is rare. The transmission phenotype of swine and avian influenza viruses in guinea pigs was similar to those found in the human population. Tested swine and avian influenza viruses could replicate effectively in the respiratory system of guinea pigs, however, none could be transmitted from infected guinea pigs to contact animals. Taken together, the experimental transmission of human, swine and avian influenza viruses in guinea pigs correlated well with the transmissibility of these viruses in humans. These studies show the utility of the guinea pig model for not only the evaluation of the transmissibility of pandemic influenza strains, human seasonal viruses and H5N1 influenza viruses, but could also be used for low pathogenic swine and avian influenza viruses to help recognize the strains that possess potential human-tohuman transmission ability. The genetic composition of the 2009 pandemic H1N1 viruses probably resulted from the reassortment of recent North American H3N2 and H1N2 swine viruses (i.e., avian/human/swine ''triple'' reassortant viruses) with Eurasian avian-like H1N1 swine viruses [24]. In the present study, the shedding kinetics and tissue tropism of Sw/GD/1222/06 (H1N2, North American triple) and Sw/FJ/ 204/07 (H1N1, European avian-like swine) were similar to BJ/ 317/09 (H1N1, pandemic) and support the proposed origin of pandemic H1N1 influenza virus. However, the transmission phenotypes of these swine influenza viruses were significantly different from those of the 2009 pandemic strain. The 2009 pandemic influenza virus was able to transmit efficiently from inoculated guinea pigs to all exposed animals, while no transmission was observed for the swine influenza viruses. Thus, further studies are needed to determine how swine viruses that lack human-to-human transmissibility acquired efficient transmissibility resulting in current pandemic strains.
Many studies have demonstrated that the guinea pig was a good animal model for influenza transmission research, however, the reason for the susceptibility of guinea pigs to influenza infection is not yet clear. Tang et al. proposed that the nasal epithelial cells of guinea pigs support influenza virus growth along with the excessive nasal mucus secretions and these factors may contribute to the susceptibility of guinea pigs to droplet spread [13]. In the present study, sneezing was often observed in all the guinea pigs infected by the 2009 pandemic H1N1 influenza virus and this strain had the most efficient transmissibility in this study. Sneezing is a clinical sign of humans infected with influenza virus and is usually observed in ferrets infected with influenza virus, and is considered an important ability of a virus for the efficient respiratory droplet transmission to elicit symptoms that promote expulsion of virus from the host [25]. Although previous studies have not reported such clinical signs in influenza virus infected guinea pigs, sneezing in guinea pigs is usually measured in studies for antigen-induced rhinitis [26,27]. Here, we propose that sneezing is perhaps another reason that guinea pigs should be considered an ideal influenza transmission model host.
Host receptor expression is one of the factors important for infection of host cells by influenza viruses. We examined the receptor specificity of the guinea pig respiratory system. The results obtained in this study by histochemistry were similar to those of Gao et al. found by fluorescence [21]. We found a widespread presence of both SAa2,6-Gal (SNA) and SAa2,3-Gal (MAA II) receptors in the nasal tract of guinea pigs, which suggests that these organs are potential targets for avian, swine and human influenza viruses. The results of animal experiments performed in this study, in addition to previous studies, demonstrated that the human, swine and avian influenza viruses replicated efficiently in this area of guinea pigs. Furthermore, virus replication at this site is likely to result in viruses that can be readily shed to infect other animals leading to the efficient transmission of influenza viruses in guinea pigs. The SAa2,6-Gal receptor is rarely detected in the upper respiratory tract of BALB/c mice, while SAa2,6-Gal is widespread on epithelial cells of the upper airways of humans [18,28]. Therefore, compared with mice, guinea pigs that express both SAa2,3-Gal and SAa2,6-Gal in the upper respiratory tract reflect a similar anatomy to that of humans.
In summary, we found the high susceptibility of guinea pigs to swine and avian influenza viruses, the correlation of transmissibility of these animal influenza viruses in guinea pigs with those in humans and the similar receptor distribution in the respiratory system of guinea pigs to those of humans. Additionally, guinea pigs possess many other human-like characteristics, including airway innervation [14], bronchus-associated lymphoid tissue in the lung [15] and response to various mediators [29]. Thus, we suggested that guinea pig could be used as an appropriate animal model for evaluating the potential threat of swine and avian viruses for humans. Animal influenza viruses are repeatedly introduced into humans and frequent reassortment and evolution often occurs. To this end, the guinea pig model offers a useful animal model to assess circulating strains that could challenge human public health and an important tool for pandemic preparedness worldwide.

Ethics Statement
All animal research was approved by the Beijing Association for Science and Technology, the approve ID is SYXK (Beijing) 2007-0023, and complied with the guidelines of Beijing laboratory animal welfare and ethical of Beijing Administration Committee of Laboratory Animals.

Replication of viruses in guinea pigs
Hartley strain female SPF/VAF guinea pigs weighing 300-350 g that were seronegative for influenza virus were used in these studies and were obtained from Vital River Laboratory Animal Technology. Zoletil 100 (tiletamine-zolazepam; Virbac S.A., Garros, France) (10-15 mg/kg) were used to anesthetize animals by intramuscular injection.
To investigate the replication of human, swine and avian influenza viruses in the guinea pig model, groups of 15 animals were anesthetized and inoculated intranasally with 10 6 EID 50 of tested virus in a 300 mL volume (150 mL on each side). On days 2, 4, 6, and 8 post-inoculation (p.i.), three animals from each group were euthanized and tracheas, lungs, hearts, livers, spleens, kidneys, brains, and colons were collected for virus titration in eggs. Nasal washes from the three remaining animals were collected on days 2, 4, 6, and 8 p.i., and titrated by EID 50 assay. Nasal washing was performed by instilling a total of 1 mL of PBS containing antibiotics (Gibco) into the nostrils and collecting liquid runoff into a sterile Petri dish. These three animals were also observed for two weeks for body weight and temperature, signs of disease, and tested for seroconversion on day 21 p.i..

Contact transmission in guinea pigs
Groups of three animals were intranasally inoculated with 10 6 EID 50 of tested virus and housed in a cage. At 24 h p.i., three naïve guinea pigs were placed in the same cage and co-housed in this way for a total of 20 days. On days 2, 4, 6, and 8 p.i., the nasal washes of the three inoculated animals and three contact animals were collected and titered for EID 50 assay. All of the animals were humanely euthanized on day 21 p.i. and tested for seroconversion.

Histopathology
On day 4 p.i., the nasal, trachea and lung specimens from inoculated guinea pigs were fixed in 10% neutral buffered formalin, routinely processed, and embedded in paraffin. Four-micrometer sections were stained with hematoxylin and eosin (HE).
Two female Hartley SPF/VAF guinea pigs were euthanized and the nasal, trachea and lung tissues were fixed in 10% (w/v) neutral-buffered formalin for 24-48 h, and 4-mm thick paraffinembedded tissue sections were made. Sections were stained with biotinylated sialic acid-specific lectins (MAA II or SNA) and the method of lectin histochemistry was performed as described by Yao et al. [32]. Briefly, paraffin-embedded serial sections were deparaffinized and immersed in 3% hydrogen peroxide to eliminate endogenous peroxidase activity. A solution of 5% bovine serum albumin was used as the blocking agent to avoid non-specific staining. Sections were next incubated with SNA (0.625 g/mL) and MAA II (1.25 g/mL) at 4uC overnight. Serial sections of the same tissue were also incubated with PBS instead of lectin as negative controls. Optimal contrast between the specific labeling and the background was obtained with the use of a SABC kit (Dako, Carpinteria, CA, USA). Biotinylated lectin binding was visualized using a DAB (3,39-diaminobenzidine-tetrahydrochloride) substrate chromogen kit (Zymed Labs, San Francisco, CA, USA) and slides were counterstained with hematoxylin.

Neuraminidase pre-treatment
To confirm the specificity of the lectin stains, neuraminidase pre-treatment was performed. Because neuraminidase digests both a-2,3-linked and a-2,6-linked sialic acid residues, negative lectin staining after neuramidase pre-treatment would indicate that the lectin used was specific for detection of a-2,3-linked and a-2,6linked sialic acids. Neuraminidase pre-treatment was performed as described by Yao et al. [32]. Briefly, the paraffin-embedded tissue sections were de-paraffinized and immersed in 3% hydrogen peroxide to eliminate endogenous peroxidase activity. Slides were covered with 12.5 U/mL neuraminidase solution (NEB, Ipswich, MA, USA) and incubated for 24 h at 37uC. Slides were washed three times with PBS and 5% BSA was used to block non-specific straining. Lectin staining was subsequently performed as described above. Additional negative controls were performed by using slides incubated with PBS instead of neuraminidase.