Infectivity, Transmission, and Pathology of Human-Isolated H7N9 Influenza Virus in Ferrets and Pigs

See allHide authors and affiliations

Science  12 Jul 2013:
Vol. 341, Issue 6142, pp. 183-186
DOI: 10.1126/science.1239844

This article has a correction. Please see:

Avian Flu in Ferrets

A recent outbreak of avian H7N9 influenza in humans in eastern China has been closely monitored for any evidence of human-to-human transmission and its potential for sparking a pandemic. Zhu et al. (p. 183, published online 23 May) examined the behavior of the avian virus in the ferret, a mammalian model for human influenza. The virus was excreted by the ferrets and could be transmitted readily by contact but displayed limited capacity for airborne infectivity. The pathology of H7N9 is similar to H1N1, and it seems that factors other than the intrinsic pathogenicity of the virus contribute to the reported high fatality rate.


The emergence of the H7N9 influenza virus in humans in Eastern China has raised concerns that a new influenza pandemic could occur. Here, we used a ferret model to evaluate the infectivity and transmissibility of A/Shanghai/2/2013 (SH2), a human H7N9 virus isolate. This virus replicated in the upper and lower respiratory tracts of the ferrets and was shed at high titers for 6 to 7 days, with ferrets showing relatively mild clinical signs. SH2 was efficiently transmitted between ferrets via direct contact, but less efficiently by airborne exposure. Pigs were productively infected by SH2 and shed virus for 6 days but were unable to transmit the virus to naïve pigs or ferrets. Under appropriate conditions, human-to-human transmission of the H7N9 virus may be possible.

On 31 March 2013, the Chinese National Health and Family Planning Commission announced the occurrence of three human infections with H7N9 subtype influenza viruses (1). Analyses of the sequences of the human H7N9 isolates indicate that the virus was derived by reassortment events between H7 and N9 subtype viruses, possibly from aquatic birds, and enzootic H9N2 viruses from chickens (1, 2). As of 1 May 2013, more than 125 human cases have been confirmed, with the majority of the patients hospitalized and many suffering acute respiratory distress syndrome (35). More than 75% of human cases had a history of contact with, or exposure to, poultry before disease onset (4), suggesting a zoonotic origin of the infections. Identification of some family clusters raised concerns of human-to-human transmission by the H7N9 virus (4). Sequence analyses showed that the H7N9 viruses might have undergone mutations that are favorable for efficient replication in mammalian hosts (2, 68).

To characterize this H7N9 virus, we assessed its infectivity, transmissibility, and pathogenicity in ferrets, the primary mammalian model for human influenza. Influenza-free ferrets (n = 6) were inoculated intranasally with 106 times the median tissue culture infectious dose (TCID50) of A/Shanghai/02/2013 (SH2), a human isolate from a fatal index case (see supplementary materials and methods) (1). These ferrets displayed a brief fever at 1 to 2 days postinoculation (dpi) and robust sneezing and nasal discharge throughout the experiment (Table 1 and fig. S1, A, C, and D). Coughing and mild lethargy occurred periodically during the course of the disease, but weight change was essentially negligible (fig. S1, B, E, and F). Ferrets (n = 6) inoculated with the pandemic A (H1N1) 2009 virus, A/California/07/2009 (CA07), showed similar signs as the SH2-inoculated animals, except for more prominent nasal discharge and slightly greater weight loss, but without statistically significant differences (fig. S1). All animals in both virus groups displayed near-normal activity levels by 14 dpi but had some residual sneezing and nasal discharge. SH2-infected animals were normal for body temperature, body weight, sneezing, coughing, and activity by 16 dpi (table S1).

Table 1 Transmissibility of SH2 and CA07 viruses in ferrets.

nd, not detectable; na, not applicable.

View this table:

The virus load in the nasal washes of each inoculated ferret was determined daily by TCID50 assays in Madin-Darby Canine Kidney (MDCK) cells. Virus shedding was detected at 1 dpi and was sustained at high titers (3.1 to 5.4 log TCID50/ml) for 7 days (Fig. 1A and Table 1). Thus, virus shedding occurred before the development of most clinical signs. For the CA07-infected ferrets, virus shedding began at 1 dpi and continued at high titers (3.1 to 5.9 log TCID50/ml) for 6 days (Fig. 1A). Overall, SH2 and CA07 infection showed comparable clinical profiles and virus shedding kinetics in ferrets with no statistically significant differences.

Fig. 1 Shedding of virus by inoculated and exposed ferrets.

Nasal washes from ferrets inoculated with or exposed to SH2 or CA07 were tested for infectious virus titer over a 14-day period after intranasal inoculation (A) or direct-contact (B and D) or airborne (C and E) exposure. Results are expressed as log10 TCID50/ml. (A) Virus titers for six ferrets [mean ± SEM (error bars)]. Red, SH2; blue, CA07. (B to E) Virus titers for individual ferrets; different symbols and lines are used for each individual ferret. Red, SH2; blue, CA07.

Efficient transmission of influenza viruses in ferrets is considered as a predictor of human-to-human transmissibility (9). Six ferrets inoculated with 106 TCID50 of virus were placed, two each, in three transmission cages. At 1 dpi, a naïve ferret was introduced into each cage with the inoculated ferrets to measure direct-contact transmission (fig. S2A). An additional naïve ferret was placed in an adjacent cage, separated by a distance of 10 cm for airborne exposure (fig. S2A), with airflow toward this cage at a rate of <0.2 m/s. An identical experiment was conducted with the CA07 virus.

All three direct-contact ferrets of the SH2 inoculated group shed virus within 3 days postexposure (dpe) and showed sneezing, nasal discharge, coughing, and inactivity by 6 dpe, but only one developed fever for 1 day at 4 dpe (Fig. 1B and Table 1). One of the airborne-exposed ferrets began shedding virus at 3 dpe and continued shedding virus at high titers for 6 days (Fig. 1C and Table 1). The two remaining airborne-exposed ferrets did not shed detectable virus and had few clinical signs (Fig. 1C and Table 1). All inoculated and direct-contact ferrets seroconverted by 14 dpi or 14 dpe. The airborne-exposed ferret that shed virus seroconverted with an hemagglutination inhibition (HAI) titer of 1:320, whereas one airborne-exposed ferret that did not shed virus had an HAI titer of 1:40 at 14 dpe (Table 1). Thus, ferrets infected with SH2 can transmit the virus via direct contact and airborne exposure, albeit the latter less efficiently. All naïve ferrets placed in CA07 direct-contact and airborne-exposure cages began shedding virus between 1 and 4 dpe (Fig. 1, D and E, and Table 1), consistent with earlier studies (10, 11).

Eighteen ferrets inoculated with SH2 as above were killed at 1, 3, 5, 7, 10, and 14 dpi to examine the gross pathology and infected tissue types to study disease progression. Respiratory and other major organs were collected for histopathologic examination and immunostaining for the presence of viral nucleoprotein (NP). Virus load was also determined by detection of a viral matrix gene by real-time polymerase chain reaction (RT-PCR) and TCID50 assays.

NP-positive cells were detected in the respiratory epithelial cells of the nasal turbinate and trachea at 3 dpi (Fig. 2, A and B). Bronchiolar epithelial cells were also positive for NP at 3 dpi (Fig. 2C). Gross pathological examination of the lungs revealed multifocal lesions on days 3, 5, and 7. At 3 dpi, histopathological examinations showed focal bronchopneumonia with acute neutrophil-predominant inflammatory infiltrates in the bronchioles and alveoli (Fig. 2, D and G). By 5 and 7 dpi, some acute but predominantly chronic lymphoplasmacytic inflammatory infiltration was seen in the bronchioles and alveoli (Fig. 2, E, F, and H).

Fig. 2 Histopathology and immunohistochemical analyses of ferret respiratory tissues after SH2 infection.

Influenza NP antigen staining is visible (brown) in the nasal turbinate (A), trachea (B), and lung (C) at 3 dpi. Pulmonary tissues were harvested from ferrets infected with SH2 and hematoxylin-and-eosin stained at 3 (D), 5 (E), and 7 dpi (F). Pathological changes characteristic of bronchopneumonia with mixed inflammatory infiltrates of bronchioles and alveoli were observed. At 3 dpi, neutrophils were the predominant infiltrating cell type (G). By 5 dpi, chronic lymphoplasmacytic infiltrates are more prominent (H). Scale bars indicate 100 μm (A and B), 200 μm (C to F), or 20 μm (G and H).

Viral RNA was detected in ferret nasal turbinate, trachea, lungs, hilar lymph nodes, and brain (Fig. 3). The presence of infectious virus was confirmed by TCID50 assays (fig. S3) and NP staining in the brain, although this was predominately cytoplasmic, and in hilar lymph nodes (fig. S4). Thus, inoculation of SH2 can result in infection of the upper and lower respiratory tracts, lymph nodes, and, potentially, the brain. A different site of inoculation, such as the trachea (12), may result in a different pathology. Clinically, the cellular tropism of H7N9 viruses may determine its spectrum of clinical disease, and it may be advisable to examine human cases for signs of central nervous system affects.

Fig. 3 Detection of influenza virus RNA in ferret tissues after intranasal inoculation with SH2.

The nasal turbinate (A), trachea (B), lungs (C), hilar lymph node (D), and brain (E) and other major organs were harvested (n = 3 animals) at various time points after inoculation. Tissue homogenates were processed for RNA extraction and quantitative RT-PCR to detect the influenza virus matrix gene. Results are mean ± SEM (error bars). Dotted lines represent the limit of detection. Samples collected at 1, 3, 5, 7, 10, and 14 dpi from the serum, spleen, intestine, mesenteric lymph node, liver, heart, and kidney were all under the limit of detection.

The domestic pig is a major mammalian host of influenza A viruses and may have played a key role in facilitating the emergence of human pandemic influenza viruses (13). To evaluate the infectivity and transmissibility of SH2 in pigs, four animals were inoculated with 106 TCID50 of SH2. Virus shedding was detected as early as 1 dpi and lasted for 5 to 6 days, with peak virus titers ranging from 3.49 to 5.16 log TCID50/ml (Table 2 and fig. S5A). Sneezing, nasal discharge, and diminished activity were observed from 1 to 1.5 days after virus shedding.

Table 2 Transmission of SH2 and CA07 in pigs and to ferrets.

nd, not detectable; na, not applicable.

View this table:

To determine transmissibility, at 1 dpi, two naïve pigs were housed with each of the two groups of inoculated animals, and a further naïve pig and ferret were placed in separate cages to assess airborne transmission (fig. S2B). None of the direct-contact pigs or airborne-exposed pigs and ferrets shed virus via the nasal or rectal routes (Table 2 and fig. S5, C and D). One of the four direct-contact pigs and both airborne-exposed ferrets seroconverted by 14 dpe. Neither airborne-exposed pig seroconverted (Table 2). Three additional inoculated pigs were killed at 4 dpi to detect the presence of virus in major organs. Viral RNA was detected in the the nasal turbinate, trachea, lungs, and lymph nodes of these pigs and NP in nasal turbinates (fig. S6). Additionally, viral RNA was detected in the kidney, heart, liver, spleen, and intestine, albeit at levels closer to the detection limit. Thus, SH2 could productively infect domestic pigs after intranasal inoculation. An identical transmission experiment using the CA07 virus showed that all direct-contact pigs and airborne-exposed pigs and ferrets could shed virus (Table 2 and fig. S5, B to D).

Thus, ferrets can be infected by SH2 and can shed virus that may transmit to direct-contact and airborne-exposed animals, resulting in productive infections. Shedding of this virus occurred before the development of the majority of clinical signs. This trait has been observed previously for pandemic and seasonal influenza (10, 11, 14, 15). If this virus acquires the ability to efficiently transmit from human to human, extensive spread of this virus may be inevitable, as quarantine measures will lag behind its spread. Assuming that poultry is the source of the H7N9 virus, continued prevalence of this virus could lead to it becoming enzootic in poultry, as has occurred with the Asian highly pathogenic H5N1 and H9N2 virus lineages (16, 17). If so, the opportunities for the H7N9 virus to evolve to acquire human-to-human transmissibility, or to be introduced to pigs, would greatly increase. To prevent this from happening, it may be advisable to reconsider the management of live poultry markets, especially in the urban areas.

Supplementary Materials

Materials and Methods

Figs. S1 to S6

Table S1

References (1821)

References and Notes

  1. Acknowledgments: We gratefully acknowledge our colleagues from the Joint Influenza Research Center (SUMC/HKU) and the State Key Laboratory of Emerging Infectious Diseases for their excellent technical assistance and D. K. Smith for editorial assistance. This study was supported by the NIH (National Institute of Allergy and Infectious Diseases contract HSN266200700005C), Li Ka Shing Foundation, the Area of Excellence Scheme of the University Grants Committee of the Hong Kong SAR (grant AoE/M-12/06), Shenzhen Peacock Plan High-End Talents Program (KQTD201203), and Emergency Research Project on human infection with avian influenza H7N9 virus from the National Ministry of Science and Technology (no. KJYJ-2013-01-01-01 to Y.S.). The sequences generated by this study were deposited in GenBank under accession nos. KF021594 to KF021601; other data can be found in the supplementary materials.
View Abstract

Navigate This Article