Heterogeneous pathological outcomes after experimental pH1N1 influenza infection in ferrets correlate with viral replication and host immune responses in the lung

Viruses

Two strains of human pH1N1 2009 IAV which had been isolated from geographically similar locations were used in the present study. The viruses were isolated at the National Influenza Centre, Centro Nacional de Microbiología, Instituto de Salud Carlos III (CNM, ISCIII) from respiratory samples sent by the Spanish Influenza Surveillance System for virological characterisation.

The pH1N1 2009 subtype isolates A/CastillaLaMancha/RR5661/2009 and A/CastillaLaMancha/RR5911/2009 were isolated from respiratory samples of two patients who were both Caucasian and free of previous co-morbid conditions at the time of infection. Hereafter, the following designations will be used: virus R61 for the A/CastillaLaMancha/RR5661/2009 virus, which was isolated from a 23-year old man exhibiting mild respiratory disease, and virus R11 for the A/CastillaLaMancha/RR5911/2009 virus, which was isolated from a 35-year old woman who developed a severe disease that resulted in a fatal outcome. Both viruses were isolated via bronchoalveolar lavage as described by Rodríguez et al. [11] and were passaged three times in Madin-Darby Canine kidney (MDCK) cells. The R61 virus had a titre of 10^8.3 50% Tissue Culture Infective Dose (TCID₅₀) per mL, and the R11 virus had a titre of 10^8.2 TCID₅₀/mL as determined with the Reed and Muench method [25].

Ferrets

Ten ferrets between the ages of twelve and twenty-four months that were seronegative for influenza A virus (influenza A antibody competition multi-species ELISA, ID Screen®, France) and Aleutian disease virus were randomly selected from a stable, purposely bred colony (Isoquimen, Spain). Upon arrival at the Centre de Recerca en Sanitat Animal (CReSA), the animals were placed in biosafety level 3 (BSL-3) facilities. The ferrets were randomly assigned to different experimental groups, separated into experimental isolation rooms and then maintained for one week for acclimation. The animals were maintained in standard housing cages and provided with commercial food pellets and tap water ad libitum throughout the experiment. All experiments were performed under a protocol (n° 1976) that was reviewed and approved by the Animal and Human Experimentation Ethics Commission of the Universitat Autònoma de Barcelona. All ferrets were monitored daily for clinical signs, and any animal that was determined to be in a moribund state was ethically euthanized. Any animal losing 25% of its original body weight or presenting severe respiratory signs such as severe tachypnea, nasal and conjunctival exudates and/or severe apathy (irresponsiveness to external stimulus) were considered to be in a moribund state and was humanely euthanized. Anorexia was measured daily by observing the amount of faeces and the amount of food that needed to be replaced and the attitude of the animals towards food. Depression was evaluated with reference to the activity of the animals and their response to external stimulus (such as sound or human presence). Tachypnea was considered when respiratory rate was higher than 70 breaths per minute, accompanied by abdominal breathing.

The animals were divided into three groups. The control group included two ferrets that were inoculated intratracheally with 0.1 M phosphate-buffered saline (pH = 7.5) (PBS). The R61 group included 4 ferrets that were infected with the R61 virus, and the R11 group included 4 ferrets that were infected with the R11 virus. The ferrets were intratracheally inoculated with 10⁶ TCID₅₀/mL of the appropriate virus, and necropsies were performed on days 4 and 7 postinfection. The animals were euthanized by intravenous injection of sodium pentobarbital (100 mg/kg) under anaesthesia with ketamine (5–10 mg/kg) (Imalgene 1000® Merial, S.A., Spain) and medetomidine (0.05 mg/kg) (Domtor® Pfiser, S.A., Spain) and were then necropsied according to a standard protocol.

Histopathology

Right lung lobe sections (cranial, medial and caudal), nasal turbinate and trachea were taken for histological examination according to standard protocols. The tissues were fixed for 48 h in neutral-buffered 10% formalin, then embedded in paraffin wax, sectioned at 3 μm, and stained with haematoxylin and eosin (HE) for examination by light microscopy.

Cross sections of the cranial, middle and caudal pulmonary lobes of each animal were separately evaluated, and semiquantitative assessments of IAV-associated microscopic lesions in the lungs were performed. The lesional scoring was graded on the basis of lesion severity as follows: grade 0 (no histopathological lesions observed), grade 1 (mild to moderate necrotising bronchiolitis), grade 2 (bronchointerstitial pneumonia characterised by necrotising bronchiolitis and diffuse alveolar damage in adjacent alveoli), and grade 3 (necrotising bronchiolitis and diffuse alveolar damage in the majority of the pulmonary parenchyma). Microscopic lesional scores were assigned for each lobe, and the means of the three lobes were used for the final histopathological score for each animal. Animals with lung lesion scores over 2 that presented with diffuse alveolar damage in at least 2 of the examined lung lobes were considered to have presented with severe lung lesions, and animals with scores below 2 were considered to have developed mild lung lesions.

Viral detection and quantification

Immunophenotyping and quantification of lung inflammatory cells

Phenotyping of the different inflammatory cell lineages following IHC were performed on the lungs of the infected and control animals to determine which inflammatory cell types were associated with the different lung lesional patterns. Neutrophils and macrophages were detected by IHC using anti-lysozyme polyclonal antibodies (Dako, Polyclonal Rabbit Anti-Human Lysozyme EC 3.2.1.17, ref A0099), T and B cells were detected using anti-CD3 (Dako, Polyclonal Rabbit Anti-Human CD3, n° IS503) and anti-CD20 (Thermo Scientific, CD20 Rabbit Polyclonal Antibody, ref RB-9013-P) polyclonal antibodies as previously described [29],[30].

Cytotoxic lymphocytes and natural killer (NK) cells were detected using the anti-CD8a (Sino Biological, Polyclonal Rabbit anti-ferret CD8a Antibody, ref 60001-RPO2, China) polyclonal antibody. Briefly, tissue sections were deparaffinised with xylene and rehydrated through graded concentrations of alcohol. Endogenous peroxidase activity was blocked by incubation with 3% H₂O₂ in methanol for 30 min. Tissue sections were rinsed in PBS and immersed in Retrieval Solution (Dako, Target Retrieval solution 10x concentrate, n°S1699) for antigen retrieval. Later, the slides were blocked with 2% bovine serum albumin (85040C, Sigma-Aldrich Química, S.A., Spain) for one hour at RT and incubated with the primary antibody overnight at 4 °C at a dilution of 0.5 μg/mL. Next, a polymer-based non-avidin-biotin peroxidase system (Dako EnVision° + System, Peroxidase-HRP, Dako, Denmark) was applied directly to the slides and incubated for 30 min at RT. The reaction was developed with DAB (Sigma-Aldrich, Madrid, Spain) at RT, followed by counterstaining with Mayer's haematoxylin. Ferret lymph node sections were used as positive controls. The same sections in which the specific primary antibodies were substituted with PBS were used as negative controls.

Semiquantitative assessments of the inflammatory cells detected in the lungs were also performed. The cells that were positive in the IHC for the anti-lysozyme (macrophages and neutrophils), anti-CD3 (T cells), anti-CD8a (cytotoxic T cells and NK cells) and anti-CD20 (B cells) antibodies were quantified for every animal and antibody. Positive cells in 6 arbitrarily chosen 40× objective fields in alveolar areas and 5 arbitrarily chosen 40× objective fields in bronchi or bronchiolar areas were counted separately for each lobe (cranial, medial and caudal) of every animal. The means of the total cell counts per field across the three lung lobes were then calculated for each animal and antibody.

Cytokine and TLR gene expression profiles

The gene expressions of interferon (IFN) α, IFNγ, TNFα, interleukin (IL) 6, IL-1α, CXCL8, CCL5, CXCL10, CCL3, CCL2 and TLR3 were detected by RT-qPCR. Briefly, RNA extraction was performed on the ferret lung tissue samples with an RNeasy Mini Kit using the RNA stabilisation and on-column DNase digestion protocols (Qiagen). Reverse transcription was performed using an ImProm-II reverse transcription system (Promega) at 0.5 μg RNA. PCR was performed using a Power SYBR green kit (Applied Biosystems) and Fast 7500 equipment (Applied Biosystems). PCR reactions were performed in 10 μL reaction volumes using the Power SYBR green kit (Applied Biosystems); 40 amplification cycles were used, and the annealing temperature was 60°. The expression levels were normalised using the house-keeping gene β-actin, and the results are expressed as arbitrary units. The primer sequences and their publication sources are presented in Table 1. Primer sequences for the β-actin, IL-1α and CCL3 genes were designed as described previously [31]. The ferret-specific genes are available in the NCBI nucleotide database [32]. The designed primer sequences and the GenBank accession numbers are presented in Table 1. The amplification products were detected by electrophoresis to validate the sizes of the product in accordance with the primer design, and the products were purified using a QIAquick PCR Purification Kit (Qiagen). Sequencing reactions were performed with ABI Prism BigDye Terminator Cycle Sequencing v.3.1 Ready Reaction (Applied Biosystems) and analysed using an ABI PRISM model 3730 automated sequencer (Applied Biosystems). The amplified sequences correlated with the ferret specific target sequences.

Apoptotic cell detection and quantification

The expression of the proapoptotic genes caspase 8 (CASP8) and BAX were quantified by RT-qPCR. The primer sequences are presented in Table 1. The RT-qPCR techniques were performed as previously described [31]. Briefly, RNA extraction was performed on the ferret lung tissue samples with an RNeasy Mini Kit using the RNA stabilisation and on-column DNase digestion protocols (Qiagen). Reverse transcription was performed using an ImProm-II reverse transcription system (Promega) at 0.5 μg RNA. PCR was performed using a Power SYBR green kit (Applied Biosystems) and Fast 7500 equipment (Applied Biosystems). PCR reactions were performed in 10 μL reaction volumes using the Power SYBR green kit (Applied Biosystems); 40 amplification cycles were used, and the annealing temperature was 60°. The expression levels were normalised using the house-keeping gene β-actin, and the results are expressed as arbitrary units. Primer sequences of CASP8 and BAX were designed as described previously [31]. The designed primer sequences and the GenBank accession numbers are presented in Table 1. Primer validation was performed as described above in the cytokine gene expression profile section.

Apoptotic cells in the lungs were also detected by IHC using the anti-caspase 3 polyclonal antibody (Cell Signalling, Cleaved Caspase-3 (Asp175) Antibody, ref 9661) diluted 1:100 in PBS using Ethylenediaminetetraacetic acid (EDTA) as the antigen retrieval method. The IHC technique was performed as described above for inflammatory cell immunophenotyping. Swine and ferret lymph node sections were used as positive controls, and the sections in which the specific primary antibodies were substituted with PBS were used as negative controls.

Statistical analyses

Histopathological scores and positive cell quantification data were expressed as the means ± the standard errors of the mean (SEMs). The correlations between the IAV antigen quantifications and the histopathological scores of each animal were calculated by linear regression. Despite the small number of animals used in each infectious group, statistical analyses were performed. Comparisons of the immunophenotypes and the IAV antigen quantifications between the animals that were infected with the R11 and the R61 viruses and comparisons between animals with severe and mild lung lesions were performed using a non-parametric one-way analysis of variance (Kruskal-Wallis) followed when necessary by a Mann–Whitney U test. A robust analysis (one iteration) was used to obtain the means ± the SEMs for the RT-qPCR data. RT-qPCR comparisons between animals infected with the R11 and the R61 viruses and comparison between animals with severe and mild lung lesions were performed using Student's t-tests. In all cases, the results were considered statistically significant when the P value was < 0.05. All statistical analyses and data visualisation were performed with GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA).

Clinical signs

Two animals infected with the R61 virus and two animals infected with the R11 virus began to present with anorexia, depression and tachypnea at 2 days post infection (dpi).

Of these four animals, one animal infected with the R11 virus died at 4 dpi, and one animal infected with the R61 virus had to be sacrificed at 4 dpi for ethical reasons. Animals showing more severe clinical signs also presented higher body weight loss (data not shown) which also correlated with more severe pathological findings in the lungs. Temperature was not affected by viral group or by disease severity (data not shown).

Gross lesions and histopathology

Gross pathological changes were observed mainly in the lungs of the infected animals that were necropsied at 4 dpi. One and two ferrets infected with the R11 and R61 viruses, respectively, presented with macroscopic lung lesions. The macroscopic lesions were characterised by heavy, firm, red and oedematous lungs. The animals in the control group did not exhibit any gross or histopathological changes.

The pulmonary histopathological scores are presented in Table 2. In general, the three lung lobes of any given animal were equally affected. No major differences in terms of histopathological scores were observed between the viral groups. Two animals infected with the R11 viral strain and two animals infected with R61 viral strain presented with severe lung lesions that were characterised by severe and diffuse alveolar damage. In general, the three lung lobes were equally affected, alveolar epithelial necrosis was observed, and the alveoli were distended and contained dense proteinaceous debris, desquamated cells, and high numbers of macrophages and neutrophils; in some cases, alveolar haemorrhage and hyaline membranes were also observed (Figure 1).

Animal	Sex	Viral inoculums	Days post infection	Histopathological scorea	Pathological classification
A	Male	R11	4†	2.67 (0.47)	Severe
B	Male	R11	4	1 (0.00)	Mild
C	Female	R61	4	3 (0.00)	Severe
D	Female	R61	4	1 (0.00)	Mild
E	Male	R61	4	3 (0.00)	Severe
F	Male	R11	7	3 (0.00)	Severe
G	Male	R11	7	1.67 (0,94)	Mild
H	Male	R61	7	1 (0.00)	Mild
I	Female	PBS	Control	0 (0.00)	Control
J	Male	PBS	Control	0 (0.00)	Control

†Animal found dead.

^aMean and SEM of the histopathological scores of the three right lung lobes.

The remaining infected animals (two infected with the R11 virus and two infected with the R61 virus) presented with mild lung lesions that consisted of moderate to severe bronchial, bronchiolar and glandular necrosis with lymphoplasmacytic inflammation (Figure 1). The histopathological lesions observed at 7 dpi were similar to those described at 4 dpi but also included bronchiolar epithelial regeneration (data not shown).

The trachea and nasal turbinates were also microscopically examined. In general, the ferrets that presented with severe lung lesions also presented with necrotising rhinitis with mucous and tracheitis with a lymphoplasmacytic infiltration. The ferrets that presented with mild lesions also presented with lesions in the trachea and nasal turbinates that were characterised by a catarrhal rhinitis with mucous secretion and tracheal glandular necrosis (data not shown).

Viral detection and quantification

IAV antigen expression was only detected in the lungs of the infected animals at 4 dpi, and no differences in IAV expression, as quantified by IHC (Additional file 1A) or RT-qPCR, were observed between animals in the R11 and R61 viral groups (Additional file 1B).

Interestingly, IAV antigen quantifications in the lungs were significantly correlated with the histopathological scores of the infected animals (Figure 2A). Higher viral antigen quantifications and higher viral loads (Figure 2B and C) were observed in the lungs of the animals that presented with severe lung lesions than in those that presented with mild lung lesions at 4 dpi.

The animals that developed severe lung lesions exhibited viral antigen expression in bronchi/bronchiole epithelial cells and scattered expression in the alveolar parenchyma. Positive labelling was observed in the nuclei of type II pneumocytes and, to a lesser extent, in the type I pneumocyte nuclei and the nuclei and cytoplasm of macrophages and interstitial lymphocytic cells (Figure 1). In contrast, IAV antigen expression in the animals that developed mild lung lesions was restricted to the bronchi and bronchiolar areas (Figure 1), in which IAV antigens were mainly detected in the nuclei of the bronchi or bronchiolar epithelial and glandular cells.

Immunophenotyping and quantification of lung inflammatory cells

Quantifications of the different inflammatory cell populations for the animals grouped by histopathological score are illustrated in Figure 3.

Inflammatory cell averages were higher in the lungs of infected animals than in those of the control group. In general, lysozyme (Lys) positive cells were the most abundantly observed cells in the lungs of the infected animals, and the maximum cell counts were observed at 4 dpi. The lungs of the ferrets that presented with severe lung lesions exhibited elevations in lysozyme + cell averages in both the alveolar and the bronchi-bronchiolar compartments compared to the animals with mild lung lesions (Figure 3), particularly in the alveolar areas (Figure 4).

The CD3+ and CD8+ cell averages of the infected animals were higher at 7 dpi than at 4 dpi. More importantly, the CD3+ and CD8+ cell counts were higher in the ferrets that developed severe lung lesions (Figure 3), and this finding was particularly striking for the alveolar CD3+ cell averages (Figure 4). The CD8+ bronchial cell averages were similar in the animal that exhibited severe and mild lung lesions at 7 dpi.

In contrast to the observations regarding CD3+ and CD8+ cell quantifications, the CD20+ cell averages decreased in both histopathological groups at 7 dpi (Figure 3). In general, the CD20+ cell averages were the lowest among all of the inflammatory cells (Figure 4). The animals that presented with severe lung lesions exhibited slightly higher bronchial CD20+ cell counts at 4 dpi but not at 7 dpi (Figure 3).

Cytokine and TLR gene expression profiles

To determine the relationships between the observed immunophenotypes, the different pathological outcomes and the cytokine, chemokine and TLR gene expressions, RT-qPCR was performed. Both infected groups exhibited elevated gene expression of pro-inflammatory markers and TLR3 compared to the control group at both 4 and 7 dpi (data not shown). The gene expression profiles of the lungs of the animals that presented with severe lung lesions exhibited increased expressions of CXCL8, CCL3, IFNγ, CXCL10, IL-1α, CCL5, IL-6 and TLR3 compared to those of the animals that presented with mild lung lesions at both 4 dpi (Figure 5) and 7 dpi (data not shown). Interestingly, at 4 dpi, a statistically significant elevation in the induction of IFNα was observed in the lungs of the animals that presented with mild lung lesions (Figure 5B). No differences in CCL2 expression were observed between the pathological groups (Figure 5C). The expressions of all cytokines decreased in both pathological groups at 7 dpi with the exceptions of CCL2 and TLR3 in the animals with severe lung lesions (data not shown).

Apoptotic cell detection and quantification

The apoptotic genes BAX and CASP8 in the lungs of the infected animals were analysed with RT-qPCR. The gene expressions of BAX and CASP8 were elevated in both infected groups compared to the control group (data not shown). However, comparison of the apoptotic induction in the lungs of the animals that presented with different grades of lung pathology revealed that the animals with severe lung lesions exhibited elevations in the inductions of both apoptotic markers compared to the animals with mild lung lesions at both 4 and 7 dpi (Figure 6A).

To determine which cells were involved in the increases in pro-apoptotic gene expression in the animals with severe lung lesions, IHC using the anti-caspase 3 antibody was performed. As expected, the numbers of apoptotic cells were elevated in the ferrets with severe lung lesions compared to the animals with mild lesions, but the positive cells were mainly consistent with macrophages and neutrophils (Figure 6B and C), which accords with the phagocytic cell averages observed in each pathological group.