Severity of Clinical Disease and Pathology in Ferrets Experimentally Infected with Influenza Viruses Is Influenced by Inoculum Volume

ABSTRACT

Ferrets are a valuable model for influenza virus pathogenesis, virus transmission, and antiviral therapy studies. However, the contributions of the volume of inoculum administered and the ferret's respiratory tract anatomy to disease outcome have not been explored. We noted variations in clinical disease outcomes and the volume of inoculum administered and investigated these differences by administering two influenza viruses (A/California/07/2009 [H1N1 pandemic] and A/Minnesota/11/2010 [H3N2 variant]) to ferrets intranasally at a dose of 10⁶ 50% tissue culture infective doses in a range of inoculum volumes (0.2, 0.5, or 1.0 ml) and followed viral replication, clinical disease, and pathology over 6 days. Clinical illness and respiratory tract pathology were the most severe and most consistent when the viruses were administered in a volume of 1.0 ml. Using a modified micro-computed tomography imaging method and examining gross specimens, we found that the right main-stem bronchus was consistently larger in diameter than the left main-stem bronchus, though the latter was longer and straighter. These anatomic features likely influence the distribution of the inoculum in the lower respiratory tract. A 1.0-ml volume of inoculum is optimal for delivery of virus to the lower respiratory tract of ferrets, particularly when evaluation of clinical disease is desired. Furthermore, we highlight important anatomical features of the ferret lung that influence the kinetics of viral replication, clinical disease severity, and lung pathology.

IMPORTANCE Ferrets are a valuable model for influenza virus pathogenesis, virus transmission, and antiviral therapy studies. Clinical disease in ferrets is an important parameter in evaluating the virulence of novel influenza viruses, and findings are extrapolated to virulence in humans. Therefore, it is highly desirable that the data from different laboratories be accurate and reproducible. We have found that, even when the same virus was administered at similar doses, different investigators reported a range of clinical disease outcomes, from asymptomatic infection to severe weight loss, ocular and nasal discharge, sneezing, and lethargy. We found that a wide range of inoculum volumes was used to experimentally infect ferrets, and we sought to determine whether the variations in disease outcome were the result of the volume of inoculum administered. These data highlight some less explored features of the model, methods of experimental infection, and clinical disease outcomes in a research setting.

INTRODUCTION

Influenza viruses cause highly contagious respiratory infection in humans and have been estimated to cause more than 36,000 deaths and over 200,000 hospitalizations annually (1). Since the 1930s, the domestic ferret (Mustela putorius furo) has been used to propagate influenza viruses and as a model for human disease (2). The ferret is a valuable model because it manifests clinical signs of influenza virus infection similar to those manifested by humans (3, 4) and has also been shown to have an anatomical expression of sialic acid (SA) receptors, the receptors by which influenza viruses attach to cells, similar to that found in the human respiratory tract (5). Postinfection ferret antisera are highly strain specific and are used in public health labs to evaluate the antigenic drift of human influenza viruses (6). In addition, ferrets are relatively easy to maintain and handle in a research setting and can be naturally infected with most human influenza viruses (2). Use of the ferret in influenza virus research has expanded to include virus transmission (7–12), vaccine efficacy (13, 14), and antiviral therapeutics studies (15, 16).

Clinical disease in ferrets is an important parameter in evaluating the virulence of novel influenza viruses, and findings are extrapolated to virulence in humans. Therefore, it is highly desirable that the data from different laboratories be accurate and reproducible. In 2009, a number of investigators used the ferret model to evaluate the replication and pathogenesis of the newly emerged H1N1 pandemic (H1N1pdm) virus (10, 17–19). Although the viruses and doses administered were similar, different investigators reported a range of clinical disease outcomes, from asymptomatic infection to severe weight loss, ocular and nasal discharge, sneezing, and lethargy (14, 17, 18, 20, 21). Our review of the studies revealed that a wide range of inoculum volumes (0.2 to 3 ml) was used to experimentally infect the ferrets; we sought to determine whether the variations in disease outcome were the result of the volume of inoculum administered.

We evaluated the importance of inoculum volume on clinical disease and pathology in ferrets infected with two influenza viruses, A/California/07/2009 (H1N1pdm) and A/Minnesota/11/2010 [H3N2 variant (H3N2v)]. We found that the severity of clinical disease worsened and the level of virus replication and respiratory tract pathology increased as the volume of inoculum increased with both viruses.

MATERIALS AND METHODS

Viruses.

Viruses were propagated in the allantoic cavity of 9- to 11-day-old specific-pathogen-free embryonated hen's eggs (Charles River Laboratories, Franklin, CT) at 37°C. Virus titers were determined in Madin-Darby canine kidney (MDCK) cells (ATCC, Manassas, VA) and calculated using the Reed and Muench method (22). The method of virus propagation in eggs or MDCK cells was similar to that described in published studies with the H1N1pdm virus (21, 24–27).

Animals.

Six-month-old male and female ferrets (Triple F Farms, Sayre, PA) were used. The source from which ferrets were procured was similar to those in published studies with the H1N1pdm virus (21, 24–27). All of the ferrets in each of the cohorts infected with the H1N1pdm and H3N2v viruses were born on the same day. All ferrets were seronegative for hemagglutination inhibition (HAI) antibodies against currently circulating human H1N1 and H3N2 viruses. All ferret experiments were performed at the National Institutes of Health (NIH) with the approval of and in compliance with the guidelines of the National Institute of Allergy and Infectious Diseases (NIAID), NIH, Institutional Animal Care and Use Committee.

Experimental influenza virus infection.

Ferrets (n = 12) were lightly anesthetized via inhalation of isoflurane and then held in an upright position and inoculated intranasally (i.n.) with 0.2, 0.5, or 1.0 ml containing 10⁶ 50% tissue culture infective doses (TCID₅₀) of the H1N1pdm or H3N2v influenza virus. The total volume of each inoculum was administered in roughly equal portions into the two nares. The ferrets were held upright for 10 to 15 s after virus administration to allow them to take a few normal breaths before they were placed prone in their cages. The ferrets were sedated deeply enough to suppress sneezing, and they regained consciousness within 5 min of intranasal administration of the inoculum.

Clinical illness.

Prior to inoculation, a subcutaneous transponder was implanted between the shoulders of all animals for monitoring of body temperature and for animal identification. The ferrets were monitored daily over a 6-day period for temperature changes, weight loss, and clinical signs of influenza virus infection, including nasal signs of infection and changes in activity related to disease, based on methods established by Reuman et al. (23). Briefly, nasal symptoms in ferrets were assessed and scored on the basis of the following criteria: no evidence of nasal symptoms (score, 0), nasal rattling or sneezing (score, 1), nasal discharge on external nares (score, 2), or evidence of mouth breathing (score, 3). Changes in the ferret's level of activity were assessed and scored on the basis of the following criteria: the ferret was fully playful (score, 0), the ferret responded to play overtures but did not initiate play (score, 1), the ferret was alert but not playful (score, 2), or the ferret was not playful or alert (score, 3).

Kinetics of virus replication.

Two separate studies were performed in 37 ferrets; 12 ferrets in each group were inoculated i.n. with 0.2, 0.5, or 1.0 ml containing 10⁶ TCID₅₀ of the H1N1pdm or H3N2v influenza virus. One ferret was mock infected with 1 ml of diluent (Leibovitz-15 [L-15] medium; Invitrogen-Gibco). Four ferrets from each of the volume groups (groups inoculated with 0.2, 0.5, or 1.0 ml) were sacrificed at 1, 3, or 6 days postinfection (p.i.), and the mock-infected control ferret was sacrificed on day 6. Samples from the nasal turbinates (NTs) and lungs were harvested and stored at −80°C. Tissue samples were thawed, weighed, and homogenized in L-15 medium containing a 2× concentration of antibiotic-antimycotic (penicillin, streptomycin, and amphotericin B; Invitrogen-Gibco) to make 5% or 10% (wt/vol) NT or 10% (wt/vol) lung tissue homogenates. Tissue homogenates were clarified by centrifugation at 1,500 rpm for 10 min and titrated in 24-well and 96-well tissue culture plates containing a monolayer of MDCK cells. The virus titers were calculated using the Reed and Muench method (22) and expressed as log₁₀ TCID₅₀ per gram of tissue.

Histology.

Sections of lungs and NTs were collected for histological evaluation as follows. Sections of the right middle, left cranial, and left caudal lung lobes were collected for microscopic evaluation. Prior to collection, the lung lobes (with trachea intact) were insufflated with 10% neutral buffered formalin (NBF) and thereafter submerged in 10% NBF for 2 to 3 days. After adequate fixation, the desired sections of lung were placed in tissue cassettes.

In order to collect the NTs for histological evaluation, the animal's head was sectioned along the midline into equal halves. Nasal turbinate tissues from one half were removed with forceps and frozen for virus titration. The remaining half, with NTs left in situ, was then placed in 10% NBF for 2 to 3 days, followed by placement in decalcifying solution (Richard-Allan Scientific) for 12 to 24 h. To permit examination of the different epithelia within the rostral, middle, and caudal aspects of the nasal passages, the nasal passages were then sectioned transversely at three levels: (i) immediately caudal to the canine tooth, (ii) bisecting the 3rd premolar, and (ii) immediately rostral to the cribriform plate.

The tissues were embedded in paraffin, sectioned (5 μm), placed on glass slides, and stained with hematoxylin and eosin (H&E). The stained tissue sections were randomized and graded for the presence and severity of pathology as follows: a score of 1 for inflammatory changes in <20% of the examined section, a score of 2 for inflammation comprising 20 to 50% of the examined section, and a score of 3 for inflammation comprising 50% or greater of the examined section.

Dye distribution.

To evaluate the role of volume on the spread of the inoculum in the respiratory tract of the ferret, four adult ferrets (two male and two female ferrets) were lightly anesthetized via inhalation of isoflurane and inoculated i.n. with 0.2, 0.5, or 1.0 ml of a tissue-marking dye (Polysciences Inc., Warrington, PA). Following administration of the tissue-marking dye, the ferrets were allowed to recover fully and move freely within their cages. Approximately 30 min later, the animals were anesthetized and humanely euthanized, and the entire length of the respiratory tract was opened (in situ) and assessed grossly for the presence and extent of tissue-staining dye distribution.

Micro-CT imaging.

Formalin-fixed lungs (with trachea attached) from four ferrets were transferred to 70% alcohol and imaged using an in vivo micro-computed tomography (micro-CT) scanner (SkyScan 1176; Bruker microCT, Kontich, Belgium). The alcohol was removed from the airways, and the trachea was intubated and attached to an oxygen source. The air pressure was increased until all lung lobes were distended. The inflated lungs were then placed in the image chamber and scanned for 3 to 4 h at 18 μm using a 0.2-mm aluminum filter set at 45 kV and 550 mA for an exposure time of 180 ms. After scanning, lung images were reconstructed using NRecon software of Bruker microCT (Kontich, Belgium) and analyzed using manufacturer-supplied software.

Statistical analysis.

To examine the associations between response variables, such as ferret weight, body temperature, nasal turbinate virus titer, and right or left lung virus titer, and the explanatory variables, such as gender, day postinfection, inoculum volume, and virus subtype, we used a linear mixed model. Statistical analysis was performed using the statistical software S-Plus. P values of less than 0.01 were considered significant.

RESULTS

Clinical data.

Weight, temperature, and clinical signs of influenza were monitored daily in experimentally infected ferrets.

(i) H1N1pdm virus infection.

All animals receiving 0.5 or 1.0 ml of inoculum lost weight, with peak percent weight losses being 5.6 and 5.7%, respectively, while 3 of 4 animals receiving 0.2 ml inoculum had a peak weight loss of 3.4% (Fig. 1). Animals in the 1.0-ml inoculum group exhibited elevations in body temperature that ranged from 1 to 2°C above those at the baseline. The elevations were first observed on day 1 postinfection (p.i.) and were maintained through day 6 in all 4 animals (Fig. 2). Fewer animals in the 0.2-ml (1/4 animals) and 0.5-ml (3/4) inoculum groups had elevations in temperature above those at the baseline (Fig. 2). In addition, the body temperatures of nearly all animals in the 0.2- and 0.5-ml inoculum groups with elevated body temperatures displayed a return to the baseline body temperature or lower by day 6 p.i.

FIG 1.

Percent weight loss in ferrets inoculated i.n. with 0.2, 0.5, or 1.0 ml containing 10⁶ TCID₅₀ of H1N1pdm (A) or H3N2v (B) influenza virus. Animals were monitored for weight loss daily over a period of 6 days. Dotted line, results for the single mock-infected ferret; solid lines, results for animals that received virus at one of the three different volumes.

FIG 2.

Change in body temperature in ferrets inoculated i.n. with 0.2, 0.5, or 1.0 ml containing 10⁶ TCID₅₀ of H1N1pdm (A) or H3N2v (B) influenza virus. Animals were monitored for changes in body temperature over a period of 6 days. Dotted line, results for the single mock-infected ferret; solid lines, results for animals that received virus at one of the three different volumes.

Ferrets that received 1.0 ml inoculum had a more substantial decrease in the level of activity (higher scores) than the other inoculum volume groups (0.2 and 0.5 ml) (Fig. 3). Activity scores for the 0.2- and 0.5-ml inoculum volume groups ranged from clinically normal (score, 0) to mild (score, 1) or moderate (score, 2). Nasal symptom scores were the highest (score, 3; most severe) on day 6 p.i. for animals in the 1.0-ml inoculum group, whereas nasal symptom scores for the 0.2- and 0.5-ml inoculum groups ranged from clinically normal (score, 0) to mild (score, 1) or moderate (score, 2) (Fig. 4).

FIG 3.

Activity scores for ferrets infected with H1N1pdm (A) or H3N2v (B) influenza virus. Activity scores are as follows: 0, activity within normal limits; 1, mild decrease in activity; 2, moderate decrease in activity; and 3, severe decrease in activity. Symbols represent activity scores for individual ferrets within each inoculum volume group at a given time point.

FIG 4.

Nasal symptom scores for ferrets infected with H1N1pdm (A) or H3N2v (B) influenza virus. Nasal symptom scores are as follows: 0, within normal limits; 1, mild; 2, moderate; 3, severe. Symbols represent nasal symptom scores for individual ferrets within each inoculum volume group at a given time point.

(ii) H3N2v virus infection.

Weight loss was observed in all animals receiving each inoculum volume. Ferrets receiving 1.0 ml of inoculum reached a peak percent weight loss of 13.9% (4/4 animals), while animals that received 0.2 ml and 0.5 ml had peak weight losses of 7.6% (4/4) and 10.6% (4/4), respectively (Fig. 1). Peak elevations in body temperature occurred on day 2 p.i. for all inoculum groups (Fig. 2). The largest increase in body temperature (∼2.5°C above that at the baseline) was observed in the 0.5-ml inoculum group, and elevations in body temperature were observed in 3/4 animals. All animals in the 1.0-ml inoculum group had elevations in body temperature that ranged from 1.0 to 1.5°C. Two of four animals in the 0.2-ml inoculum group exhibited elevations in body temperature above that at the baseline (Fig. 2). Although the magnitude of the body temperature elevation was greater in the 0.5-ml inoculum group, the consistency (by day postinfection and kinetics) with which the temperature elevation occurred was more appreciable in the animals that were administered 1.0 ml of virus inoculum. Additionally, we observed a relatively blunt response according to the body temperature change compared with the response according to the weight change. The temperature was recorded using a subcutaneous transponder at a similar time each day. However, it is possible that a core body temperature would have been a more sensitive measure.

Clinical activity scores were similar for animals in the 0.5- and 1.0-ml inoculum groups, ranging from clinically normal (score, 0) to mild (score, 1) or moderate (score, 2), and were observed as early as day 1 p.i. (Fig. 4). Animals in the 0.2-ml inoculum group achieved activity scores similar to those for the animals in the 0.5- and 1.0-ml inoculum groups, but the changes were of a shorter duration and were observed only between days 2 and 4 p.i. (Fig. 4). Nasal symptom scores were the highest for animals in the 1.0-ml inoculum group, and nasal symptoms were the most severe on day 6 p.i. (3/4 animals). The duration and degree of nasal symptoms were similar for animals in the 0.2- and 0.5-ml inoculum groups.

Viral replication in the upper respiratory tract (URT) and lower respiratory tract (LRT). (i) H1N1pdm virus infection.

In the NTs, the H1N1pdm virus replicated to high titers at all inoculum volumes and in all animals inoculated. Animals that received 0.2 ml inoculum (4/4 animals) had a mean virus titer of 10^7.5 TCID₅₀/g on day 1 p.i. Similar levels of virus replication were observed on day 3 p.i. in animals that received the virus in a volume of 0.5 ml (4/4; 10^7.4 TCID₅₀/g) and on day 1 p.i. in animals that received the virus in a volume of 1.0 ml (4/4; 10^6.6 TCID₅₀/g) (Fig. 5).

FIG 5.

Replication kinetics of H1N1pdm (A) or H3N2v (B) influenza virus in ferrets following i.n. inoculation of 10⁶ TCID₅₀ of virus. Virus titers in the NTs and lungs (RML, right middle lung; LcrL, left cranial lung) of 4 ferrets per group sacrificed on days 1, 3, and 6 p.i. are expressed as the log₁₀ TCID₅₀ per gram of tissue. Bars, mean titers; symbols, titers from individual ferrets; dashed horizontal line, lower limit of detection (10^1.8 and 10^1.5 TCID₅₀ per gram for the NTs and lungs, respectively).

In the LRT, sections of right middle (RM) and left cranial (LCr) lung lobes were sampled. The kinetics and levels of virus replication for animals in the 1.0-ml inoculum group were similar when the right and left lung lobes were compared. The mean titers on the day that the highest titer was achieved in the 1.0-ml inoculum group were 10^5.1 TCID₅₀/g for the right lung and 10^5.2 TCID₅₀/g for the left lung (Fig. 5). The mean titers on the days that the highest titer was achieved for animals in the 0.5-ml inoculum group were 10^3.3 TCID₅₀/g and 10^3.8 TCID₅₀/g for the right and left lung lobes, respectively. Animals in the 0.2-ml inoculum group showed the most noticeable disparity in viral replication between the two sides of the lung, with mean peak titers of 10^2.5 TCID₅₀/g for the right lung samples versus 10^3.8 TCID₅₀/g for the left lung samples (Fig. 5). At all inoculum volumes, viral titers in the lungs declined sharply or were undetectable by day 6 p.i.

(ii) H3N2v virus infection.

Similar to the replication in H1N1pdm virus-infected animals, the H3N2v virus replicated efficiently and to high titers at all inoculum volumes. Animals in the 0.5-ml and 1.0-ml inoculum groups showed the highest titers of virus in the NTs (10⁸ and 10^7.9 TCID₅₀/g, respectively), followed by animals in the 0.2-ml inoculum group (10^7.2 TCID₅₀/g). In contrast to the consistent decline in viral titers observed with the H1N1pdm virus, virus was detected at moderate to high levels on day 6 p.i. in the 0.5- and 1.0-ml inoculum groups (Fig. 5).

In the LRT, peak mean virus titers were achieved on day 1 p.i. for the 1.0-ml inoculum group. and the titers for the right and left lungs were of similar magnitudes, with 10^7.1 TCID₅₀/g for the right lung and 10^7.3 TCID₅₀/g for the left lung. Similar to observations made in H1N1pdm-infected animals, the two smaller inoculum groups (0.2 and 0.5 ml) showed more noticeable inconsistencies in the amount of virus isolated between right and left lung samples (Fig. 5). The mean titer on the day that the highest titer was achieved for the 0.5-ml inoculum group was 10^3.1 TCID₅₀/g in the right lung and 10^5.3 TCID₅₀/g in the left lung. Similar to the level of inconsistent virus replication observed in the 0.2-ml inoculum group for animals infected with H1N1pdm virus, the peak mean virus titers from the right and left lungs in the 0.2-ml inoculum group were 10^2.7 and 10^4.1 TCID₅₀/g, respectively. In all inoculum volume groups, viral titers decreased to levels near the limit of detection (LOD) by day 6 p.i.

Histopathology of upper and lower respiratory tracts. (i) H1N1pdm virus infection.

Sections of NTs collected from animals on days 1, 3, and 6 p.i. were evaluated histologically. In the sections of NTs from day 6 p.i., the inflammatory changes were most intense and were often associated with abundant cellular debris that, in some cases, completely filled the NT lumina (Fig. 6). The overlying respiratory epithelium consistently displayed both regions of cellular hyperplasia and occasional areas of squamous metaplasia; intact ciliated cells were infrequently observed. Inflammation was first observed on day 3 p.i. and was characterized by the marked expansion and infiltration of the submucosal stroma by edema fluid and inflammatory cells (Fig. 6A). Ciliated cells along the mucosal surface were variably sloughed or intact and necrotic. In many cases, the lumina of the NTs contained variable amounts of viable and degenerate neutrophils and admixed necrotic cellular debris. The findings for sections of NTs collected on day 1 p.i. from animals in all three groups were within normal limits, without evidence of inflammation (Fig. 6A).

FIG 6.

Pathology in NTs of ferrets following i.n. inoculation of H1N1pdm (A) and H3N2v (B) influenza viruses at 10⁶ TCID₅₀ of virus administered in a 0.2-, 0.5-, or 1.0-ml volume. Photomicrographs are representative of the histological changes present in tissues from the 4 ferrets per group that were sacrificed on days 1, 3, and 6 p.i. Arrows, areas of NT luminal inflammatory cell debris. Magnifications, ×40. Bars, 100 μm.

Sections of the right middle, left cranial, and left caudal lung lobes from four animals in each group at each of the three time points (1, 3, and 6 days p.i.) were sampled for histopathology. The presence and general degree of inflammation within the sections of lung examined correlated with the volume of inoculum administered. Lung pathology was widespread and the most severe for the 1.0-ml inoculum group, while the 0.2- and 0.5-ml inoculum groups showed less severe pathology, and fewer lung lobes were affected in those groups (Table 1). While the frequency of lesions differed by volume, the cellular composition of the inflamed regions and the localization of the inflammatory lesions were consistent at all inoculum volumes. Similar to the kinetics of the pathology in the URT, inflammatory changes were most intense in the lung by day 6 p.i. (Fig. 7A). The regions immediately surrounding the bronchi were often characterized by expansion and infiltration of the peribronchial stroma by edema and an abundant mixed cellular infiltrate composed of neutrophils, macrophages, lymphocytes, and plasma cells. In most cases, within these inflammatory foci, the submucosal glands (SMGs) exhibited variable degrees of neutrophilic inflammation and necrosis (Fig. 7A). Sections of lung from day 3 p.i. exhibited a more extensive inflammatory process characterized by intense peribronchiolar inflammation that also extended into and expanded the alveolar interstitium adjacent to the affected airways (Fig. 7A). Inflammatory changes in the lung were present as early as day 1 p.i. (Fig. 7A) and began with a mild and predominantly neutrophilic and lesser histiocytic (alveolar macrophage) infiltrate in both peri- and intrabronchiolar regions.

TABLE 1.

Distribution and severity of pathology in lungs of ferrets infected with H1N1pdm or H3N2v influenza virus administered at different volumes

Virus	Animal no.	No. of lung lobes affected/lung pathology score(s) in each lobe for the following inoculum vol (ml) and day p.i.a:
		0.2			0.5			1.0
		Day 1	Day 3	Day 6	Day 1	Day 3	Day 6	Day 1	Day 3	Day 6
H1N1pdm	1	0/0	1/1	1/1	1/3	0/0	2/2, 2	1/1	2/2, 1	2/3, 3
	2	1/1	1/1	1/1	0/0	1/1	2/3, 2	2/1, 1	1/1	2/3, 3
	3	1/1	1/2	1/3	1/1	1/1	2/2, 2	2/3, 1	2/3, 2	2/3, 3
	4	0/0	1/1	0/0	0/0	1/2	2/2, 3	1/2	2/2, 1	2/3, 2
H3N2v	1	1/1	0/0	0/0	0/0	0/0	2/2, 1	0/0	1/1	2/3, 1
	2	0/0	0/0	0/0	0/0	1/1	1/2	0/0	1/1	2/1, 3
	3	0/0	0/0	0/0	1/1	1/1	1/1	1/2	1/2	2/2, 2
	4	0/0	0/0	0/0	0/0	1/2	1/1	1/2	1/3	2/3, 2

^aTwo sections of lung (the right middle and left cranial [caudal half] sections) were sampled for animals infected with the H1N1pdm or H3N2v influenza virus. Lung pathology scores ranged from 0 (minimal) to 3 (severe) and were based on the extent of pathology present in each tissue section examined: 0, within normal histological limits; 1, 5 to 25%; 2, 25 to 50%; 3, 50% or greater.

FIG 7.

Pathology in the lungs of ferrets following i.n. inoculation of H1N1pdm (A) or H3N2v (B) influenza virus at 10⁶ TCID₅₀ of virus administered in a 0.2-, 0.5-, or 1.0-ml volume. Animals were sacrificed on days 1, 3, and 6 p.i. Photomicrographs are representative of the histological changes present in tissues from the 4 ferrets per group that were sacrificed on 1, 3, and 6 p.i. *, areas of SMG inflammation and necrosis. Magnifications, ×40. Bars, 100 μm.

(ii) H3N2v virus infection.

Similar to the findings for the URT in ferrets infected with the H1N1pdm virus, infection with the H3N2v virus resulted in NT inflammation that was first observed on day 3 p.i. but was most severe on day 6 p.i. Sections of NTs from day 6 p.i. were characterized by an intense neutrophilic inflammatory infiltrate that often completely filled the nasal passages. Frequently, the mucosa within these intensely inflamed regions was devoid of ciliated cells and displayed both epithelial hyperplasia and squamous metaplasia (Fig. 6B). On day 3 p.i., sections were characterized by expansion and infiltration of the submucosal stroma by edema and inflammatory cells. The overlying respiratory epithelium was often segmentally populated by ciliated cells, and the changes in those exhibiting changes were histologically consistent with degeneration and necrosis. Sections of NTs from day 1 p.i. showed no evidence of inflammation (Fig. 6B).

We found that, similar to animals infected with the H1N1pdm virus, lung pathology was more severe and more widespread, affecting a larger proportion of the lung sections examined, as the inoculum volumes were increased (Table 1). In the case of H3N2v, inflammatory lesions were the most severe on day 6 p.i. for animals that received virus at volumes of 0.5 ml and 1.0 ml (Fig. 7B). These inflammatory foci were intensely neutrophilic and were closely associated with the bronchi and bronchioles. Day 3 p.i. (in the 0.5-ml inoculum group) was the earliest time point at which inflammation was observed. The inflammation consisted of variably dense neutrophilic and, to a lesser extent, histiocytic infiltrates that were situated intimately around small bronchi and bronchioles (Fig. 7B). In contrast to the findings for H1N1pdm-infected animals, inflammation was minimal to absent at all time points in animals that received virus in a 0.2-ml inoculum volume.

Distribution of inoculum in ferret respiratory tract.

In order to determine the basis for differences in viral replication and the distribution of the pathology in the respiratory tract, we sought to investigate the ability of different inoculum volumes to coat the URT and LRT. We inoculated ferrets intranasally with tissue-staining dye in volumes of 0.2, 0.5, or 1.0 ml. Introduction of this dye into the nasal passages at different volumes allowed us to characterize the spread and localization of the inoculum immediately following administration.

At all three volumes, dye was detected in the URT and in the esophagus. However, we cannot estimate the proportion of the inoculum that was swallowed. Animals in the 0.2-ml inoculum group showed small accumulations of dye in the rostral-most regions of the nasal passages (Fig. 8A). Inoculum was also present in the proximal esophagus, but not in the trachea (Fig. 8B). Animals in the 0.5-ml inoculum group showed dye over a greater area of the nasal passages and abundant staining of the esophageal mucosa but no evidence of dye in the trachea. In the 1.0-ml inoculum group, dye coverage of the URT was extensive, extending caudally toward the margin of the cribriform plate. The dye was present throughout the length of the esophagus and the full length of the trachea and was also observed grossly within the parenchyma of the left caudal lung lobe (Fig. 8C). On cut section, dye was present within the lung parenchyma (Fig. 8C, inset).

FIG 8.

Distribution of tissue-staining dye within the upper and lower respiratory tracts of ferrets. Tissue-staining dye was administered i.n. to 4 ferrets per group in a 0.2-, 0.5-, or 1.0-ml volume. Thirty minutes later, the animals were euthanized and gross examination of the respiratory tract was conducted. The examined tissues included the nasal passage (np) (A), trachea (tr) and esophagus (e) (B), and lung (C). Arrowheads, deposition of tissue-staining dye within the trachea and/or esophagus; arrows, localization of dye in the lung. Dye can be observed on the surface of the lung (arrow) and on cut sections within the lung parenchyma (inset).

MicroCT imaging and analysis of the ferret respiratory tract.

Using micro-computed tomography (micro-CT), we found that, consistent with published data, ferrets and humans share a number of anatomic similarities of the lung (28), characterized by three right (cranial, middle, caudal) lung lobes and two left (cranial and caudal) lung lobes. Interestingly, we found that the diameter of the right main-stem bronchus was considerably and consistently larger than that of the left main-stem bronchus. In addition, the right bronchus arose from the tracheal bifurcation at an angle with a more pronounced lateral deviation and one that is more acute than that for the left bronchus, whereas the narrower left bronchus consistently arose from the bifurcation in a straighter fashion, projecting toward the left caudal lung lobe (Fig. 9).

FIG 9.

Maximum-intensity projection (MIP) of normal ferret lung anatomy. Lungs from uninfected adult ferrets were imaged (ex vivo) using a modified air bronchogram method and micro-CT to assess the normal anatomy and conformation of the ferret LRT. Dashed lines, orientation and diameter of the ostia of the right (R) and left (L) main-stem bronchi as they arise from their origin at the tracheal bifurcation; solid vertical line, a bisector for the angles created by the ostia of the right and left main-stem bronchi at the level of the tracheal bifurcation.

DISCUSSION

In this study, we demonstrate that inoculum volume affects the severity of clinical disease in ferrets experimentally infected with two different influenza viruses. Understanding the contribution of inoculum volume to disease outcomes in this model will reduce the level of variability in clinical disease and viral replication and lead to standardization of the model. Miller et al. (29) demonstrated variable clinical outcomes in mice infected with an H3N2 influenza A virus (X31) that was administered to mice intranasally at a dose of 10³ TCID₅₀ in a volume of 25, 35, or 50 μl. They found that mice infected with virus in a volume of 25 μl readily recovered from disease, while the same dose of virus administered in 35-μl and 50-μl volumes was associated with a fatal outcome. Cook et al. (30) also reported considerable variations in clinical disease outcomes in mice that were infected with 120 PFU of pneumovirus administered in 10, 25, and 50 μl. With a lowering of the inoculum volume from 50 to 25 μl, the severity and duration of clinical signs were diminished. With a further reduction of the inoculum to 10 μl, signs of disease were completely absent.

Importantly, in addition to clinical observations, we found that the inoculum volume had a significant effect on the level and consistency of viral replication in the LRT of ferrets. Furthermore, we found that smaller inoculum volumes (0.2 and 0.5 ml) were associated with viral replication in the lungs less consistent than that in animals that received the 1.0-ml inoculum. In some cases, animals that received a 0.2- or 0.5-ml inoculum had 10,000- to 100,000-fold differences in virus titer between the right and left lung lobes. Differences of this magnitude were rarely observed in the 1.0-ml inoculum group, demonstrating that the largest inoculum volume resulted in more uniform viral replication in the LRT, likely a result of a more uniform distribution of the inoculum. van den Brand et al. (31) reported inconsistencies in the ability to elicit clinical signs of disease in ferrets using a seasonal human influenza virus administered intranasally at a dose of 10⁶ TCID₅₀. Thereafter, they modified their experimental design to intratracheal administration of 10⁹ TCID₅₀ of virus in an inoculum volume of 3 ml. However, we found that a dose of 10⁶ TCID₅₀ of virus administered intranasally in a volume of 1.0 ml was sufficient to induce clinical disease and pulmonary pathology in ferrets.

The H1N1pdm and H3N2v viruses replicated to high titers in the NTs, though the H3N2v virus replicated to levels that were significantly (P = 0.0016) higher than those to which the H1N1pdm virus replicated. Animals that received H1N1pdm virus in 0.2 ml of inoculum had viral titers that were slightly higher than those in animals that received this virus in a 0.5- or 1.0-ml volume. Furthermore, the level of virus replication in the animals in the 0.2-ml inoculum group (n = 4) was more consistent than that observed in animals in the 0.5- and 1.0-ml inoculum groups. In contrast, NT virus titers for animals infected with the H3N2v influenza virus were the highest and most consistent for the 0.5- and 1.0-ml inoculum groups and slightly lower (10- to 100-fold) and less consistent than those in animals that received virus at the lowest inoculum volume (Fig. 5). The H3N2v virus replicated to levels that were significantly higher than those of the H1N1pdm virus in the NTs, but in the LRT, both viruses replicated poorly, except when the inoculum volume was 1.0 ml. Smith et al. (32) also reported that when ferrets were infected with a dose of 10² TCID₅₀/ml in 0.2 ml, the H1N1pdm virus replicated well (∼6 log units) in the NTs, but virus was not isolated from the lungs of any of the animals. However, in contrast to our findings, Pearce et al. (11) reported that the influenza A/MN/10 (H3N2v) virus replicated to high titers in the NTs but was undetectable in the lungs of ferrets that were infected intranasally with 10⁶ PFU administered in a 1.0-ml volume. It is possible that their inability to detect viral replication in the LRT was influenced by sampling of lung tissue for virus titration, because we noted significant differences in the levels of virus replication between the right and left lung samples, irrespective of virus subtype. We found that analysis of the right middle portion and a caudal portion of the left cranial lung lobe for virus titration provides a more comprehensive assessment of viral replication in the lungs. We also found that when we compared the pathology associated with the two viruses in the LRT of ferrets, pathological changes were minimal in the 0.2-ml inoculum group but were more severe and evenly distributed (between the right and left lungs) when animals received either virus in a 1.0-ml volume (Table 1). In a multivariate analysis, when we examined gender, day p.i., inoculum volume, and virus subtype, we found that both day p.i. and inoculum volume were significant predictors of the level of virus replication in the lungs of ferrets experimentally infected with either virus subtype. Specifically, we found that increasing the inoculum volume led to significant increases in the virus titer in both the right (P < 0.0001) and left (P = 0.0001) lung lobes. In addition, we found that over the 6-day period, there was a significant decrease in virus in both the right (P = 0.0001) and left (P < 0.0001) lung lobes.

These observations prompted us to investigate the anatomy of the ferret's respiratory tract in order to determine whether anatomical features contribute to the variable distribution of the inoculum in the LRT of the ferret model. Micro-CT analysis of the ferret lung demonstrated that the right main-stem bronchus of ferrets (n = 4; irrespective of gender) consistently had a larger diameter than the left main-stem bronchus and often arose from the tracheal bifurcation earlier and at a more acute angle, whereas the left bronchus often exhibited a straighter projection into the caudal lung lobe. In addition to the observations made on micro-CT, we examined gross specimens (n = 10) and found that at the level of the tracheal bifurcation, the mean diameter of the left main-stem bronchus was 3.08 ± 0.12 mm and the right main-stem bronchus was approximately 1.5 times larger than the left main-stem bronchus, measuring 4.85 ± 0.20 mm in diameter. Irrespective of virus subtype, there was a significantly higher titer of virus in both the right and left lung lobes, based on the volume of inoculum administered. In addition, we noted that at the two smaller inoculum volumes (0.2 and 0.5 ml) there was a trend toward higher virus titers in the left lung lobe, irrespective of the virus subtype administered. However, the variability in virus titer was reduced when a volume of 1.0 ml of inoculum was administered. We speculate that the conformation of the conducting airways influences the distribution of the inoculum in the LRT.

Ferrets are commonly held in an upright position while the inoculum is instilled into the nasal passages, and they are maintained in this position long enough to ensure full inhalation of the inoculum. It is possible that anatomical features of the ferret airways, in conjunction with the vertical positioning and the administration of a particular volume of inoculum, could influence the direction and immediate localization of the inoculum in the LRT. Swallowing of the inoculum likely contributes to the high variability of virus titers in the lung and the volume-dependent differences in clinical disease severity when smaller inoculum volumes are administered. Three of four animals that received 1.0 ml of dye showed evidence of staining in the LRT, in addition to a considerable amount of dye in the esophagus. It is likely that the larger volume of inoculum easily reaches the lower respiratory tract, though a significant proportion of the inoculum is swallowed. It is important to note that the dye inoculation study shows only the immediate distribution of an inert inoculum; clearly, influenza virus administered in 0.2 or 0.5 ml is able to infect the LRT, despite our inability to detect dye in the LRT of ferrets immediately following administration. The association of clinical signs with the larger volume of inoculum suggests that the virus replication and inflammatory changes that occur in the LRT are important determinants of clinical illness.

In the lung, both influenza viruses were associated with a predominantly neutrophilic inflammatory response at the middle and late time points of the study. However, the severity and distribution of inflammation in the lung were greater in the animals infected with the H1N1pdm virus. The H1N1pdm virus was largely associated with changes in the pulmonary interstitium at early time points and became more airway centered by day 6 p.i. A remarkable feature of H1N1pdm virus infection at day 6 p.i. was the presence of submucosal gland (SMG) inflammation and viral antigen. Bissel et al. (33) also identified H1N1pdm virus within the SMG of experimentally infected ferrets between days 2 and 3 p.i. by in situ hybridization (ISH). The significance of inflammation at this location is not well understood; however, the localization of viral antigen in the submucosal glands has been demonstrated in humans who succumbed to H1N1pdm infection (34–36), ferrets infected with seasonal human H3N2 virus (37), and formalin-fixed paraffin-embedded (FFPE) tissues from humans and mice infected in vitro with H5N1 virus using immunohistochemistry (36).

Ferrets express influenza virus SA receptors in a pattern that is similar to that in humans. In general, human influenza viruses preferentially bind α2,6 SAs, which are found in the URT, while avian influenza viruses preferentially bind α2,3 SAs, which are much more prevalent in the LRT. In this study, we did not examine the effect of inoculum volume on infection with seasonal human influenza viruses or avian influenza viruses that differ significantly in their sialic acid receptor preferences. However, in a separate study, we observed that a highly pathogenic H5N1 influenza virus administered in 0.2 ml caused only mild clinical disease in ferrets (unpublished data), while the same amount of virus administered in an inoculum volume of 0.5 ml caused neurological disease, including hind limb paresis (38). Taken together with the findings from the current study, it is likely that the animals that received 0.2 ml of the H5N1 virus had milder clinical disease than the animals that received the 0.5-ml inoculum because the smaller volume did not efficiently infect the LRT, where the α2,3 SA receptors predominate.

In summary, we have demonstrated the importance and influence of inoculum volume on clinical illness and pathology and viral replication in ferrets experimentally infected with two different influenza A viruses. We have shown the importance of using a 1.0-ml volume of inoculum to deliver virus to the LRT. In addition, we have identified anatomical features of the ferret LRT that influence the disease process. Therefore, we recommend administering 1.0 ml of inoculum intranasally for experimental infection in ferrets, particularly when clinical disease outcomes are a desired study parameter.