Abstract
Highly pathogenic avian influenza virus (HPAIV) H5N1 can infect mammals via the intestine; this is unusual since influenza viruses typically infect mammals via the respiratory tract. The dissemination of HPAIV H5N1 following intestinal entry and associated pathogenesis are largely unknown. To assess the route of spread of HPAIV H5N1 to other organs and to determine its associated pathogenesis, we inoculated infected chicken liver homogenate directly into the intestine of cats by use of enteric-coated capsules. Intestinal inoculation of HPAIV H5N1 resulted in fatal systemic disease. The spread of HPAIV H5N1 from the lumen of the intestine to other organs took place via the blood and lymphatic vascular systems but not via neuronal transmission. Remarkably, the systemic spread of the virus via the vascular system was associated with massive infection of endothelial and lymphendothelial cells, resulting in widespread hemorrhages. This is unique for influenza in mammals and resembles the pathogenesis of HPAIV infection in terrestrial poultry. It contrasts with the pathogenesis of systemic disease from the same virus following entry via the respiratory tract, where lesions are characterized mainly by necrosis and inflammation and are associated with the presence of influenza virus antigen in parenchymal, not endothelial cells. The marked endotheliotropism of the virus following intestinal inoculation indicates that the pathogenesis of systemic influenza virus infection in mammals may differ according to the portal of entry.
INTRODUCTION
Highly pathogenic avian influenza viruses of the H5N1 subtype (HPAIV H5N1) can infect mammals via the intestine. This is unusual since influenza viruses typically infect mammals via the respiratory tract. Consumption of sick or dead birds infected with HPAIV H5N1 is a likely route of infection for different species of carnivores found naturally infected with the virus (17) and first supported the possibility that HPAIV H5N1 infection in mammals may occur via the digestive tract. An intestinal route of infection by HPAIV H5N1 was also suggested in humans following reports of patients with gastrointestinal symptoms as the only initial symptoms and reports of patients exposed to the virus via consumption of raw duck blood (3). To further investigate HPAIV H5N1 intestinal route of infection, domestic cats, red foxes (Vulpes vulpes), and ferrets were experimentally fed infected chickens (10, 14, 18). They developed lesions associated with the presence of the virus in multiple organs, confirming that consumption of infected birds can result in infection and systemic disease in these species. Direct intragastric inoculation of HPAIV H5N1 in mice, ferrets, hamsters, and cats was shown to result in systemic spread of the virus, further proving the intestine as a portal of entry for HPAIV H5N1 in these species (14, 24, 26).
The route of spread of HPAIV H5N1 from the lumen of the digestive tract to other organs and associated pathogenesis remain largely undetermined. Thus far, intestinal replication of HPAIV H5N1 has been detected in epithelial and mononuclear cells of the intestinal mucosa of fatal human cases (6), in ganglion cells of the submucosal and myenteric plexi of cats experimentally fed infected chickens (21), and in mononuclear cells of the Peyer's patches of a naturally infected cat (8). Based on these findings, it was speculated that HPAIV H5N1 may spread from the digestive tract to other organs via several routes: HPAIV H5N1 may infect mononuclear cells of gut-associated lymphatic tissues and spread via blood, lymph, or a combination of both routes (21, 24, 26) or the virus may infect ganglion cells of the intestinal plexi and spread via neuronal transmission (21).
HPAIV H5N1 cause systemic disease in poultry, and particularly chickens. Systemic spread of the virus occurs in this species via the vascular system. However, the pathogenesis of HPAIV infection in chickens is distinctive. Infection of endothelial cells is a hallmark of HPAIV infection in this species and results in rapid dissemination to and high virus titers in numerous organs (25). In marked contrast to HPAIV pathogenesis in chickens, infection of endothelial cells by influenza virus has rarely been reported in mammals (17). HPAIV H5N1 infection of endothelial cells was detected in rare endocardial cells of the heart and endothelial cells of the pulmonary vein of experimentally infected cats (21). Recently, however, endotheliotropism of HPAIV H5N1 was demonstrated in vitro in cultures of human endothelial cells (15).
Neuronal transmission of HPAIV H5N1 has been described in mammals. In mice, neuronal transmission of HPAIV H5N1 from the nasal cavity to the olfactory bulb, along the olfactory nerves, has been experimentally shown to occur (16). In ferrets, lesion patterns in the olfactory bulb indicate a similar neuronal spread of HPAIV H5N1 from the nasal cavity to the brain (5, 23). Ferrets inoculated either intratracheally or intranasally developed distinct lesion and influenza virus antigen distribution patterns, suggesting different routes of viral spread depending on the virus site of entry (2). While intratracheal inoculation resulted in pneumonia associated with viral infection of respiratory epithelial cells, intranasal inoculation resulted in meningitis and encephalitis, associated with viral infection of neurons. These findings support neuronal transmission from the nasal cavity to the olfactory bulb but also suggest that the pathogenesis of HPAIV H5N1 infection may differ according to the site or route of entry.
In the present study, we inoculated liver homogenate of HPAIV H5N1-infected chickens directly into the small intestine of cats by use of enteric-coated capsules and performed necropsies at different time points after inoculation. This was done in order to (i) assess the route of spread of HPAIV H5N1 from the lumen of the intestinal tract to other organs and (ii) determine its associated pathogenesis.
MATERIALS AND METHODS
Virus preparation.
Influenza virus A/Vietnam/1194/2004 (H5N1), kindly provided by W. Lim, Queen Mary Hospital, Hong Kong, was propagated once on Madin-Darby canine kidney (MDCK) cells. It was titrated on MDCK cells according to standard methods (19) and reached an infectious virus titer of 106.2 median tissue culture infectious doses (TCID50)/ml.
Twenty-four 1-day-old specific-pathogen-free (SPF) Leghorn chicks were inoculated intratracheally with 2.5 × 104 TCID50 of H5N1 virus in 0.5 ml of phosphate-buffered saline (PBS). Four 1-day-old SPF Leghorn chicks housed in a separate isolator were sham inoculated intratracheally with 0.5 ml of PBS. All chicks were euthanized by cervical dislocation at 2 days postinoculation (dpi) or earlier if exhibiting clinical signs.
Twenty liver samples were collected from each chick, weighed, and frozen at −80°C. They were homogenized with a tissue homogenizer (Kinematica Polytron, Lucerne, Switzerland) in 3 ml of transport medium. One liver homogenate sample per chick was thawed and titrated in triplicate on MDCK cells according to standard methods (19). Virus titers ranged from 106.7 to 107.8 TCID50 per g of liver tissue (106.5 to 107.4 TCID50 per ml). Liver homogenates from sham-infected chicks were negative.
Experimental protocol.
To avoid virus contamination of the respiratory tract in the procedure of intestinal inoculation, we delivered enteric-coated capsules filled with infected chicken liver homogenate directly into the stomachs of anesthetized cats. The production and filling of the capsules are described in the supplemental material.
A gavage tube was introduced in the esophagus up to the stomach of nine 4- to 6-month-old SPF cats, under anesthesia with ketamine, medetomidine, and atropine. Each cat was inoculated with four capsules containing 0.6 ml of chicken liver homogenate each, i.e., a total of 107.8 TCID50 of HPAIV H5N1 per cat. One negative control cat was inoculated in the same way with four capsules each filled with 0.6 ml of liver homogenate from chickens sham-inoculated with PBS. The cats were observed for clinical signs daily. In addition, the cats were observed closely for regurgitation in the first hours following the inoculation procedure. Nasal, pharyngeal, and rectal swabs were taken from all cats under anesthesia with ketamine and medetomidine at 12 h, 24 h, and daily thereafter and placed in 3 ml of transport medium. Swabs were also taken minutes after inoculation to control for possible pharyngeal contamination. The swabs were frozen at −80°C until titration. Blood was sampled from all cats under anesthesia, several minutes and 12 h after inoculation and then daily. Blood was allowed to clot, and serum was frozen at −80°C until titration. The cats were euthanized by exsanguination under anesthesia and subsequently examined by necropsy. It was planned to examine tissues of three cats per time point at time points 1, 3, and 7 dpi. However, of the three cats euthanized at 3 dpi, two were excluded because of virus contamination of the pharynx just after inoculation. Of the three cats to be euthanized at 7 dpi, two died or were moribund by 4 and 5 dpi, and the third remained virus negative in all swabs and tissues and was excluded. Therefore, the results of the remaining three cats that were euthanized or died between 3 and 5 dpi were analyzed together. The negative control cat was euthanized at 3 dpi. All experimental procedures on chickens and cats were approved by an independent Animal Care and Use Committee.
Pathology and immunohistochemistry.
Necropsies and tissue sampling were performed according to a standard protocol. Special attention was paid to the sampling of the potential routes of virus spread. Detailed description of the sampling protocol is given in the supplemental material. Tissues of the digestive, respiratory, cardiovascular, nervous, endocrine, urinary, and lymphoid systems (a list is provided in the supplemental material) were examined by histopathological and immunohistochemical methods.
After fixation in 10% neutral-buffered formalin and embedding in paraffin, tissue sections were stained with hematoxylin and eosin (H&E) for histological evaluation or with an immunohistological method using a monoclonal antibody against the nucleoprotein of influenza A virus as a primary antibody for detection of influenza viral antigen (20). Lung tissue of a cat experimentally infected intratracheally with influenza virus A/Vietnam/1194/2004 (H5N1) was included as a positive control. Isotype-matched and omission controls were included as negative controls (20). To identify the nature of the cells found infected, we additionally stained consecutive slides of selected tissue sections with an immunohistological method using monoclonal antibodies against the nucleoprotein of influenza virus, against pankeratin or against von Willebrand factor as a primary antibody for identification of cells infected with HPAIV H5N1, and for identification of epithelial and endothelial cells, respectively. We used consecutive slides of the same tissues from a cat infected intratracheally with HPAIV H5N1 (21) for comparison of cell tropism and associated lesions.
Virology.
Tissues of the digestive, respiratory, cardiovascular, nervous, endocrine, urinary, and lymphoid systems (a list is provided in the supplemental material) were further sampled, weighed, and frozen at −80°C. The tissue samples were homogenized with a FastPrep-24 (MP Biomedicals, Eindhoven, Netherlands) in 1 ml of transport medium and centrifuged briefly. Tenfold serial dilutions of the tissue homogenates, swabs, and serum samples were inoculated on MDCK cells in triplicate as described previously (19). The minimal detectable titer was 100.8 TCID50/ml. All experiments were performed under biosafety level 3+ conditions.
RNA isolation, reverse transcription-PCR (RT-PCR), sequencing, and genetic analyses.
Genetic analyses were performed to determine (i) whether the inoculum viruses (input viruses) used in the present study (virus isolated from chicken liver) and past study (A/Vietnam/1194/2004 passaged twice on MDCK) differed genetically and (ii) whether the output viruses isolated from cats inoculated via the intestine, via the trachea, or by feeding on infected chickens (21) differed genetically from the respective input viruses, and from each other. Included in the analyses were (i) the inoculum virus that was used to infect the cats in the present study (chicken liver homogenate isolate); (ii) the inoculum virus that was used in the previous study (A/Vietnam/1194/2004) (21); (iii) pharyngeal swab, rectal swab, trachea, lung, cerebellum, and mesenteric lymph node isolates from one cat infected via the intestine (present study); (iv) rectal swab, trachea, and lung isolates from one cat infected intratracheally (21); and (v) rectal swab, trachea, lung, and cerebellum isolates from one cat fed infected chickens (21).
Total RNA was extracted by using a MagnaPure LC system with the MagnaPure LC total nucleic acid isolation kit (Roche Diagnostics, Almere, Netherlands), according to the manufacturer's instructions. Viral RNA was reverse transcribed and PCR amplified using Platinum Taq Hot Start One Step RT-PCR kit (Invitrogen, Ltd., Paisley, United Kingdom) and a modified version of the PCR method described by Zhou et al. (27), which amplifies all eight influenza segments in a single PCR. The primers used in the present study are described in the supplemental material. For each sample, 2 to 5 μg was sequenced using 454 and/or Illumina technology. All samples were run through QUASR_v6.07 for quality control. Library preparation and sequencing were performed according to the manufacturer's instructions. Reads were assembled to A/Vietnam/1194/2004 (H5N1) genome as a reference sequence and a consensus sequence generated using the majority high-quality base present at each position using Samtools software (13). Accession numbers are given in the supplemental material.
RESULTS
Clinical signs.
No cat was observed regurgitating any administered capsules following inoculation. The cats that were necropsied at 1 dpi did not show visible clinical signs. The cats necropsied at 3 to 5 dpi were lethargic with reduced appetite from 3 dpi onward (see Table S1 in the supplemental material). The negative control cat showed no clinical signs.
Virology.
HPAIV H5N1 was typically isolated from nasal, pharyngeal, and rectal swabs from 2 dpi onward (Fig. 1). However, in two cats, low titers (100.8 TCID50/ml) were determined in the nasal swabs at 0.5 and 1 dpi. Titers from nasal, pharyngeal, and rectal swabs continued to increase and plateaued at 3 to 5 dpi, reaching 104.2 to 105.2 TCID50/ml. Similarly, HPAIV H5N1 was typically isolated from serum from 2 dpi onward, reaching up to 106.2 TCID50/ml at 4 dpi (Fig. 1). A low titer (100.8 TCID50/ml) was determined in the serum of one cat sampled at 0.5 dpi, suggesting rapid onset of viremia following intestinal inoculation.
Fig 1.
(A to D) Virus titers in nasal swabs (A), pharyngeal swabs (B), rectal swabs (C), and serum samples (D) from cats infected with HPAIV H5N1 via the intestine. The serum samples collected at 5 dpi were not available.
No virus was isolated from any organs of the cats necropsied at 1 dpi. In contrast, HPAIV H5N1 was isolated from nearly all organs at 3 to 5 dpi (Table 1). The highest titers were consistently determined (i) in organs associated with the digestive tract, i.e., the ileum, cecum, colon, and liver, (ii) in the kidneys, (iii) in lymphoid organs, i.e., the tracheobronchial lymph node, tonsils, mesenteric lymph node, and spleen, and (iv) in organs associated with the respiratory tract, i.e., the nasal concha, nasal septum, and lungs. Overall, the virus titers were highest in the liver, reaching more than 108 TCID50/g. Along the intestinal tract, the virus titers typically increased from the duodenum to the colon, with the highest titers in the ileum, cecum, and colon. No virus was isolated from the swabs or organs of the negative control cat.
Table 1.
Distributionof influenza virus in organs of cats infected with HPAIV H5N1 via the intestine
| Type of organ | No. of cats with positive organs (virus titer range [log TCID50/g of tissue]) |
||
|---|---|---|---|
| Negative control (n = 1) | 1 dpi (n = 3) | 3–5 dpi (n = 3) | |
| Digestive system and associated organs | |||
| Esophagus | 0 (<0.8) | 0 (<0.8) | 3 (3.0–5.3) |
| Stomach | 0 (<0.8) | 0 (<0.8) | 2 (3.6–4.0) |
| Duodenum | 0 (<0.8) | 0 (<0.8) | 2 (2.1–5.0) |
| Jejunum | 0 (<0.8) | 0 (<0.8) | 3 (1.9–4.9) |
| Ileum | 0 (<0.8) | 0 (<0.8) | 3 (2.3–5.0) |
| Cecum | 0 (<0.8) | 0 (<0.8) | 3 (4.3–6.9) |
| Colon | 0 (<0.8) | 0 (<0.8) | 3 (3.4–5.4) |
| Pancreas | 0 (<0.8) | 0 (<0.8) | 1 (2.9) |
| Liver | 0 (<0.8) | 0 (<0.8) | 3 (6.7–>8.8) |
| Lymphoid organs | |||
| Tonsils | 0 (<0.8) | 0 (<0.8) | 3 (6.0–6.9) |
| Tracheobronchial lymph node | 0 (<0.8) | 0 (<0.8) | 3 (5.4–6.3) |
| Mesenteric lymph node | 0 (<0.8) | 0 (<0.8) | 3 (3.1–7.2) |
| Spleen | 0 (<0.8) | 0 (<0.8) | 3 (5.0–7.1) |
| Nervous system | |||
| Vagus nerve | 0 (<0.8) | 0 (<0.8) | 3 (3.7–5.7) |
| Spinal cord | 0 (<0.8) | 0 (<0.8) | 3 (3.5–6.8) |
| Olfactory bulb | 0 (<0.8) | 0 (<0.8) | 3 (3.9–6.0) |
| Brain stem | 0 (<0.8) | 0 (<0.8) | 3 (2.2–5.0) |
| Cerebellum | 0 (<0.8) | 0 (<0.8) | 3 (3.0–5.4) |
| Cerebrum | 0 (<0.8) | 0 (<0.8) | 3 (3.5–5.5) |
| Respiratory system and conjunctiva | |||
| Eyelid | 0 (<0.8) | 0 (<0.8) | 3 (3.0–5.7) |
| Nasal concha and septum | 0 (<0.8) | 0 (<0.8) | 3 (6.5–6.9) |
| Trachea | 0 (<0.8) | 0 (<0.8) | 3 (2.9–5.6) |
| Lung | 0 (<0.8) | 0 (<0.8) | 3 (4.9–7.5) |
| Cardiovascular system | |||
| Heart | 0 (<0.8) | 0 (<0.8) | 3 (3.5–6.6) |
| Portal vein | 0 (<0.8) | 0 (<0.8) | 2 (3.0–6.5) |
| Urinary system | |||
| Kidney | 0 (<0.8) | 0 (<0.8) | 3 (5.8–6.9) |
| Urinary bladder | 0 (<0.8) | 0 (<0.8) | 3 (2.5–5.5) |
Gross pathology and histopathology.
At necropsy, no gross lesion could be seen in any organs of the negative control cat and of the cats necropsied at 1 dpi (see Table S1 in the supplemental material). In contrast, the livers of the cats necropsied at 3 to 5 dpi showed a mild to moderate acute necrosis, characterized by the presence of multiple white spots of 0.5 to 1 mm in diameter on the surface of each liver. In these cats, moderate to marked enlargement of the lymphoid organs associated with multifocal hemorrhages was frequently observed. Digested or fresh blood was occasionally present in the stomach, duodenum, and proximal parts of the jejunum. In one cat, the duodenum and jejunum had diffuse mucosal hemorrhages. Other significant findings included multifocal adrenocortical hemorrhages in two cats and mild multifocal acute pancreatic necrosis, characterized by multiple white areas 1 cm in diameter in one cat.
Histological lesions were observed in all cats inoculated with HPAIV H5N1 (see Table S2 in the supplemental material). Leakage of erythrocytes outside blood vessels into neighboring tissues (hemorrhage) was the most common lesion, often associated with enlargement (hypertrophy) of endothelial cells, and was widespread at 3 to 5 dpi. Hemorrhages were often associated with mild to moderate inflammation, characterized by the presence of few to moderate numbers of neutrophils in the affected tissue and mild to moderate lesions of necrosis (Fig. 2). The most severely affected organs were the lymphoid organs and the liver. In lymphoid organs, hemorrhages, inflammation, and necrosis were most severe in the lymph nodes associated with the digestive tract (Fig. 2D) and in the Peyer's patches, all the more as they were located more distally along the digestive tract. In the liver, acute hepatic necrosis and inflammation was multifocal to coalescing and moderate to severe at 3 to 5 dpi (Fig. 2G).
Fig 2.
Viral replication sites and associated histological lesions in cats inoculated via the intestine with HPAIV H5N1. (A to C) There is hemorrhage in the lumen of the jejunum (asterisk) and inside the villi (arrowheads) (A), with influenza viral antigen expression in the lamina propria of the villi (B), including in the hypertrophic endothelial cells of small blood vessels (C). (D to F) There is focal necrosis and hemorrhage in the cortex of the mesenteric lymph node (D), with expression of influenza viral antigen expression in the affected tissue (E), including lymphendothelial cells of the subcapsular sinus (F). (G to I) There are discrete foci of necrosis and inflammation (arrowheads) in the liver (G), with influenza viral antigen expression in hepatocytes at the transition of lesional and normal liver parenchyma (H and I). (J) There is a rare focus of necrosis and inflammation (arrowheads) at the transition of gray and white matter in the spinal cord and mild focal hemorrhage (asterisk) around a nearby blood vessel. (K and L) Influenza viral antigen expression is present both in the focus of necrosis and in the blood vessel (K), where hypertrophic endothelial cells stain positive (L). (M) The alveolar walls of the lung are diffusely thickened and hypercellular. Note that the alveolar lumina are virtually clear. (N and O) Influenza viral antigen expression is present throughout the alveolar walls (N), mainly in endothelial cells (O). (P to R) The adrenal gland has multiple hemorrhages in the cortex (P), with influenza viral antigen expression in both adrenal cortical cells (arrowhead) and hypertrophic endothelial cells (arrow) (Q and R). Tissues were collected between 3 and 5 dpi. Serial sections of tissues were stained either with H&E or by immunohistochemistry with a monoclonal antibody against the nucleoprotein of influenza A virus as a primary antibody.
Immunohistochemistry.
No cells positive for influenza viral antigen were seen in the three cats necropsied at 1 dpi and in the negative control cat. The possible exception was a suspect positive cell of unidentified type in the liver of one cat necropsied at 1 dpi. At 3 to 5 dpi, cells positive for influenza viral antigen were present in most organs examined (see Table S3 in the supplemental material). The cell types most frequently seen positive were the endothelial cells lining veins, venules, capillaries, and lymphatic vessels, and to a lesser extent, arteries and arterioles. Endothelial cells were found to be positive in virtually all tissues examined.
Along the digestive tract, endothelial cells of blood and lymphatic vessels stained positive in the intestinal mucosa within the lamina propria and along the villi (Fig. 2B and C), as well as in the mesentery. In addition, mononuclear cells were frequently found to be positive in the lamina propria and in the Peyer's patches of the digestive tract, notably of the ileum. Ganglion cells of the submucosal and myenteric plexi were rarely seen positive.
In all of the lymphoid organs examined, particularly those associated with the digestive tract, endothelial cells of blood vessels and lymphatic vessels were positive, as well as lymphendothelial cells lining the sinuses of the lymph nodes, and reticular cells in these sinuses (Fig. 2E and F).
In the liver, few to moderate numbers of hepatocytes stained positive for influenza viral antigen at the edges of lesions of acute hepatic necrosis and inflammation in one cat (Fig. 2H and I), and virtually all hepatocytes were positive in the two other cats. Endothelial cells were found to be positive along the portal vein and other vessels of the portal system.
In organs of the peripheral and central nervous system (i.e., peripheral ganglia, spinal cord, brain stem, cerebellum and cerebrum), most cells found positive for influenza viral antigen were identified as endothelial cells (Fig. 2K and L and Fig. 3B). Neurons were rarely seen infected, in contrast to what has been described in cats infected via the trachea (21) (Fig. 3E).
Fig 3.
Comparison of cellular tropism of highly pathogenic avian influenza virus H5N1 in cats inoculated via different routes. After inoculation via intestine (upper row), endothelial cells are the main cell type expressing influenza viral antigen in lung (alveoli) (A), kidney (glomerulus) (B), and brain (cerebrum at the level of the hippocampus) (C). In contrast, after inoculation via trachea (21) (lower row), influenza viral antigen expression is seen in organ parenchymal cells as follows: type 2 pneumocytes in the lungs (D), visceral epithelial and/or mesangial cells in the kidney (E), and neurons in the brain (F). Tissues were collected between 4 and 6 dpi. Tissues were stained by immunohistochemistry with a monoclonal antibody against the nucleoprotein of influenza A virus as a primary antibody.
In the respiratory tract, cells were positive in the nasal concha, nasal septum, trachea, and alveolar walls and were mostly identified as endothelial cells. In addition, many mesothelial cells of the visceral pleura were found to be positive for influenza viral antigen in one cat. In the nasal concha and septum, a few respiratory epithelial cells as well as chondrocytes of the cartilaginous tissue were also found positive, associated with moderate to severe lesions of necrosis. In the lung, the cells staining positive for influenza viral antigen were diffusely distributed in the alveolar walls (Fig. 2N and O). They were identified as capillary endothelial cells based on their encircling of erythrocytes and on elongated and thin nuclei that were not protruding into the alveoli (Fig. 3A) and by positive von Willebrand factor staining (Fig. 4). There was no evidence of influenza virus antigen expression in type I pneumocytes, type II pneumocytes, bronchiolar epithelial cells or bronchial epithelial cells, in contrast to what has been described in cats infected via the trachea (21) (Fig. 3D).
Fig 4.
Confirmation of the endothelial nature of infected cells following infection of with HPAIV H5N1 cats via the intestine. The cells expressing influenza virus antigen in pulmonary alveoli are endothelial cells (A), as confirmed by expression of the von Willebrand factor in the cytoplasm of these cells (B). Tissues were collected at 4 dpi. Tissues stained by immunohistochemistry, with a monoclonal antibody against the nucleoprotein of influenza A virus as the primary antibody, were destained in eluting buffer and restained with a monoclonal antibody against von Willebrand factor as a primary antibody.
In other organs, endothelial cells also were the cell type most frequently found positive for influenza virus antigen. In the kidney, infection of endothelial cells in the glomeruli of cats infected via the intestine (Fig. 3C) contrasted with the infection of visceral epithelial or mesangial cells in the glomeruli of cats infected via the trachea (Fig. 3F). In the heart, there also were few to moderate numbers of infected epicardial cells, cardiomyocytes and Purkinje cells. In the adrenal glands, there were also multiple foci of infected cortical cells (Fig. 2Q and R).
Genetic analyses.
The original inoculum virus (A/Vietnam/1194/2004 passaged twice on MDCK cells) used in the past study (21) and used in the present study for the inoculation of chickens had three nucleotide substitutions compared to the A/Vietnam/1194/2004 reference sequences obtained from GenBank. These were in the PB2 gene (C178G), the PA gene (A9C), and the HA gene (T665A). Only the change in the PB2 gene was nonsynonymous, resulting in the substitution of histidine by aspartic acid at position 60 (H60D). The sequences of the original inoculum virus are used as references for all subsequent analyses.
A total of 10 nucleotide substitutions were identified in some consensus sequences of the PB2, PB1, PA, HA, MA, and NS genes of the chicken liver isolate and of output viruses from cats inoculated via the intestine, via the trachea, or fed infected chickens, compared to that of the original inoculum. These resulted in six amino acid substitutions in one to three consensus sequences of the respective PB2, PA, NP, and MA proteins (Table 2). No amino acid change was detected in the HA protein. In general, the coding changes found in the chicken liver isolate and output viruses were present as detectable minority variants in the original inoculum (see Fig. S1 in the supplemental material). Overall, no consistent change in the consensus genome of HPAIV H5N1 differentiated the viruses isolated in cats infected via the intestine from those isolated in cats infected intratracheally or fed infected chickens. However, a trend toward higher genetic diversity was observed for samples taken from cats inoculated via the trachea or fed infected chickens than for those taken from cats inoculated via the intestine (see Fig. S1 in the supplemental material). In general, mixed populations of viruses with either the original amino acids or the substitutions were present in the samples taken from these cats. Surprisingly, one of the noncoding changes in the NS gene (A474G) was present almost exclusively as the majority or large minority variant in samples from cats inoculated via the trachea or fed infected chickens and was not detected or detected at a very low frequency in samples from cats inoculated via the intestine. This noncoding change was present at a similar low frequency in the original inoculum and in the chicken liver homogenate.
Table 2.
Amino acid substitutions identified in consensus protein sequences of HPAIV H5N1 used andisolated in the present and previous studiesa
| Route of inoculation | Type of sample | Amino acid substitution per virus gene segmentb |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| PB2 | PB1 | PA | HA | NP | NA | MA | NS | ||
| None | Chicken liver homogenate | M248V* | |||||||
| Intestine | Pharyngeal swab | ||||||||
| Rectal swab | |||||||||
| Trachea | |||||||||
| Lung | |||||||||
| Cerebellum | |||||||||
| Mesenteric lymph node | S262P | ||||||||
| Trachea | Rectal swab | ||||||||
| Trachea | M248V* | ||||||||
| Lung | |||||||||
| Feeding | Rectal swab | I47M | M84I | M248V* | |||||
| Trachea | S556L | ||||||||
| Lung | M84I | ||||||||
| Cerebellum | T65K† | ||||||||
Rimmelzwaan et al. (21).
Compared to the consensus sequence of the original inoculum (A/Vietnam/1194/2004 H5N1 passaged twice on MDCK cells).
, In the M1 reading frame;
, in the M2 reading frame after splicing.
DISCUSSION
In this study, we show that the systemic spread of HPAIV H5N1 from the intestine to other organs took place via the blood and lymphatic vascular systems, while there was no evidence for neuronal transmission in these intestinally inoculated cats. Unexpectedly, the systemic spread of the virus via the vascular system was associated with massive infection of endothelial and lymphendothelial cells, resulting in hemorrhages in numerous organs. The pathogenesis of HPAIV H5N1 infection in cats following the intestinal portal of entry thus differed markedly from that following the respiratory portal of entry (21) and resembled the pathogenesis of HPAIV infection in terrestrial poultry (25).
We show here that the route of spread of HPAIV H5N1 from the intestine to other organs took place via the vascular system. Viremia was observed as early as 12 h after inoculation, supporting rapid systemic spread of the virus, occurring between 1 and 3 dpi. High virus titers were consistently detected in the liver, in distal parts of the intestinal tract, and in highly vascularized organs, such as the kidney and respiratory tract. High virus titers in the latter correlated with high levels of viral excretion detected in nasal and pharyngeal swabs. The source of excreted virus is unclear and unlikely to be respiratory epithelial cells, since these were rarely found infected in these cats. Lower levels of excretion via the rectum despite high virus titers in the distal parts of the intestinal tract may be due to the absence of infection of intestinal epithelial cells, which would be a major source of fecal excretion, and inactivation of intestinally excreted virus by digestive enzymes. The liver, which directly receives venous blood from the intestine via the portal vein, was among the most severely affected organs, with multifocal to coalescing infection by 3 dpi. In one cat, the foci of infected cells were regularly distributed, typically at equidistance from portal triads, which is consistent with seeding of the virus from blood vessels into the hepatic parenchyma. In carnivores naturally infected with HPAIV H5N1, presumably after consumption of infected bird carcasses, the liver often presented with severe lesions of necrosis, which is associated with the infection of hepatocytes (17), and may indicate infection following intestinal entry of the virus. In contrast, there was no evidence of viral spread via the nervous system, as determined by the paucity of virus antigen expression in neurons of the peripheral and central nervous systems despite thorough examination (see the list of nervous tissues examined in the supplemental material).
The systemic spread of HPAIV H5N1 following intestinal inoculation was associated with massive infection of endothelial and lymphendothelial cells. This is unique for influenza in mammals and resembles the pathogenesis of HPAIV infection in terrestrial poultry. In poultry, endotheliotropism is a hallmark of HPAIV infection, and the peracute form of the disease is predominated by vascular damage resulting in the leakage of fluids and erythrocytes from damaged blood vessels into neighboring tissues, resulting in edema and hemorrhage, and eventually coagulation failure (25). Such pathogenesis largely determines the high pathogenicity of these viruses in poultry, characterized by short times to death. Poultry that survive the peracute phase of disease die later of multiorgan failure associated with virus replication in parenchymal cells. In contrast, neither endotheliotropism nor hemorrhage has been observed as a major component of infection with HPAIV H5N1 or other influenza viruses in humans (9, 11) or other mammals (12, 17). However, HPAIV H5N1 recently has been demonstrated in vitro to be capable of infecting human endothelial cell cultures and was proposed to contribute to the severe outcome of HPAIV H5N1 infection in humans, including coagulopathy described in some patients (15). The endotheliotropic pathogenesis of HPAIV H5N1 infection in cats inoculated via the intestine may have contributed to the more severe and rapidly fatal disease seen in these cats compared to the disease seen in cats inoculated via the trachea (21).
Different routes of entry through the intestinal barrier and of spread via the vascular system may be used by HPAIV H5N1 following intestinal inoculation. Entry mechanisms may include the uptake by enterocytes and subsequent infection of underlying endothelial cells of blood or lacteal vessels. However, no infected intestinal epithelial cell was detected in the present study, even though more than 70 cross-sections of intestinal epithelium per cat were collected at regular intervals along the intestine and examined microscopically for influenza virus antigen expression. Another entry mechanism may be the uptake of HPAIV H5N1 by microfold (M) cells. These cells are membranous epithelial cells overlying the lymphoid follicles of Peyer's patches. They sample antigens directly in the lumen of the intestine and present them to lymphoid cells. They may infect other antigen-presenting cells and initiate infection of endothelial or lymphendothelial cells in the gut-associated lymphoid tissues. The presence of infected mononuclear cells in the intestinal mucosa at 3 dpi, as well as the higher virus titers and more widespread and severe foci of infection along the distal parts of the intestinal tract, rich in Peyer's patches, may indicate initial infection of M cells. Subsequent spread of HPAIV H5N1 to endothelia of blood vessels and lymphatics likely occurred from virus in blood and lymph, respectively, given the high virus titers in the cats' serum and the rapid dissemination of the virus throughout the body. Basolateral spread from endothelial cell to endothelial cell cannot be ruled out as an additional mechanism of spread.
We investigated a potential genetic basis for the endotheliotropism of HPAIV H5N1 following intestinal inoculation. Variations in the tissue tropism of influenza viruses are known to be associated with genetic differences in their genome, resulting in amino acid changes in the virus proteins. HPAIV H5N1 with amino acid substitution E627K in the polymerase PB2 protein replicated in a variety of organs, in contrast to the original strain, which replicated only in the respiratory tract in mice (7). Differences in viral replication efficiency between the two isolates were at the origin of the differences in tissue tropism (22). Amino acid substitutions in the receptor binding site of the hemagglutinin (HA) protein and insertion of a poly-basic cleavage site in the HA protein are known to directly modulate influenza virus tissue tropism in mammals, allowing or abrogating infection of different cell types (1). The HA protein of HPAIV H5N1 was recently found to be an important determinant of endotheliotropism in cultures of human endothelial cells (15). However, no consistent differences in the consensus sequences of the HA or any other genes of HPAIV H5N1 inoculated into or isolated from cats inoculated via the intestine, via the trachea, or via feeding on infected chicks were detected. The impact of the observed higher diversity of viruses in cats inoculated via the trachea or fed infected chickens or the presence of the noncoding mutation A474G in the NS gene of viruses isolated from these cats and not in viruses isolated from cats inoculated via the intestine are unclear. Host factors are additional determinants of influenza virus tissue tropism. In mice, interferons were shown to contribute to tissue tropism patterns, their absence in knockout mice allowing influenza virus A/WSN/33 to replicate beyond the respiratory tract (4). The observed difference in HPAIV H5N1 phenotype in cats inoculated via the intestine and via the trachea may not be linked to within-host evolution of the inoculated virus but to host factors associated with the route of entry of the virus.
In conclusion, we demonstrate that the spread of HPAIV H5N1 from the lumen of the intestine to other organs in these cats took place via the blood and lymphatic vascular systems but not via neuronal transmission. The systemic spread of the virus via the vascular system was associated with massive infection of endothelial and lymphendothelial cells, which resembles the pathogenesis of HPAIV infection in chickens. The marked endotheliotropism of the virus following entry from the intestine suggests a high plasticity of the phenotype of HPAIV H5N1 and that the pathogenesis of systemic influenza virus infection in mammals may differ according to the portal of entry.
Supplementary Material
ACKNOWLEDGMENTS
This study was supported by the VIRGO Consortium, an innovative cluster approved by the Netherlands Genomics Initiative and partially funded by the Dutch government (grant BSIK03012).
This study was performed at the Department of Virology, Erasmus Medical Centre, P.O. Box 2040, 3000 CA Rotterdam, Netherlands. Preparation of the enteric-coated capsules was done at the Department of Pharmaceutical Technology and Biopharmacy of the University of Groningen, Groningen, Netherlands, and the genetic analyses were performed at the Wellcome Trust Sanger Institute in Cambridge, United Kingdom.
We thank Debby van Riel and Peter van Run for excellent technical assistance and Frank van der Panne for preparing the figures.
Footnotes
Published ahead of print 16 November 2011
Supplemental material for this article may be found at http://jvi.asm.org/.
REFERENCES
- 1. Baigent SJ, McCauley JW. 2003. Influenza type A in humans, mammals and birds: determinants of virus virulence, host range and interspecies transmission. Bioessays 25:657–671 [DOI] [PubMed] [Google Scholar]
- 2. Bodewes R, et al. 2011. Pathogenesis of influenza A/H5N1 virus infection in ferrets differs between intranasal and intratracheal routes of inoculation. Am. J. Pathol. 179:30–36 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gambotto A, Barratt-Boyes SM, de Jong MD, Neumann G, Kawaoka Y. 2008. Human infection with highly pathogenic H5N1 influenza virus. Lancet 371:1464–1475 [DOI] [PubMed] [Google Scholar]
- 4. Garcia-Sastre A, et al. 1998. The role of interferon in influenza virus tissue tropism. J. Virol. 72:8550–8558 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Govorkova EA, et al. 2005. Lethality to ferrets of H5N1 influenza viruses isolated from humans and poultry in 2004. J. Virol. 79:2191–2198 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Gu J, et al. 2007. H5N1 infection of the respiratory tract and beyond: a molecular pathology study. Lancet 370:1137–1145 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hatta M, Gao P, Halfmann P, Kawaoka Y. 2001. Molecular basis for high virulence of Hong Kong H5N1 influenza A viruses. Science 293:1840–1842 [DOI] [PubMed] [Google Scholar]
- 8. Klopfleisch R, et al. 2007. Distribution of lesions and antigen of highly pathogenic avian influenza virus A/Swan/Germany/R65/06 (H5N1) in domestic cats after presumptive infection by wild birds. Vet. Pathol. 44:261–268 [DOI] [PubMed] [Google Scholar]
- 9. Korteweg C, Gu J. 2008. Pathology, molecular biology, and pathogenesis of avian influenza A (H5N1) infection in humans. Am. J. Pathol. 172:1155–1170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Kuiken T, et al. 2004. Avian H5N1 influenza in cats. Science 306:241–241 [DOI] [PubMed] [Google Scholar]
- 11. Kuiken T, Taubenberger JK. 2008. Pathology of human influenza revisited. Vaccine 26S:D59–D66 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kuiken T, van den Brand J, van Riel D, Pantin-Jackwood M, Swayne DE. 2010. Comparative pathology of select agent influenza a virus infections. Vet. Pathol. 47:893–914 [DOI] [PubMed] [Google Scholar]
- 13. Li H, et al. 2009. The sequence alignment/map format and SAMtools. Bioinformatics 25:2078–2079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lipatov AS, Kwon YK, Pantin-Jackwood MJ, Swayne DE. 2009. Pathogenesis of H5N1 influenza virus infections in mice and ferret models differs according to respiratory tract or digestive system exposure. J. Infect. Dis. 199:717–725 [DOI] [PubMed] [Google Scholar]
- 15. Ocana-Macchi M, et al. 2009. Hemagglutinin-dependent tropism of H5N1 avian influenza virus for human endothelial cells. J. Virol. 83:12947–12955 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Park CH, et al. 2002. The invasion routes of neurovirulent A/Hong Kong/483/97 (H5N1) influenza virus into the central nervous system after respiratory infection in mice. Arch. Virol. 147:1425–1436 [DOI] [PubMed] [Google Scholar]
- 17. Reperant LA, Rimmelzwaan GF, Kuiken T. 2009. Avian influenza viruses in mammals. Rev. Sci. Tech. 28:137–159 [DOI] [PubMed] [Google Scholar]
- 18. Reperant LA, et al. 2008. Highly pathogenic avian influenza virus (H5N1) infection in red foxes fed infected bird carcasses. Emerg. Infect. Dis. 14:1835–1841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Rimmelzwaan GF, Baars M, Claas EC, Osterhaus AD. 1998. Comparison of RNA hybridization, hemagglutination assay, titration of infectious virus and immunofluorescence as methods for monitoring influenza virus replication in vitro. J. Virol. Methods 74:57–66 [DOI] [PubMed] [Google Scholar]
- 20. Rimmelzwaan GF, et al. 2001. Pathogenesis of influenza A (H5N1) virus infection in a primate model. J. Virol. 75:6687–6691 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Rimmelzwaan GF, et al. 2006. Influenza A virus (H5N1) infection in cats causes systemic disease with potential novel routes of virus spread within and between hosts. Am. J. Pathol. 168:176–183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Shinya K, et al. 2004. PB2 amino acid at position 627 affects replicative efficiency, but not cell tropism, of Hong Kong H5N1 influenza A viruses in mice. Virology 320:258–266 [DOI] [PubMed] [Google Scholar]
- 23. Shinya K, et al. 2011. Subclinical brain injury caused by H5N1 influenza virus infection. J. Virol. 85:5202–5207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Shinya K, et al. 2011. Systemic dissemination of H5N1 influenza A viruses in ferrets and hamsters after direct intragastric inoculation. J. Virol. 85:4673–4678 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Swayne DE. 2007. Understanding the complex pathobiology of high pathogenicity avian influenza viruses in birds. Avian Dis. 51:242–249 [DOI] [PubMed] [Google Scholar]
- 26. Vahlenkamp TW, Teifke JP, Harder TC, Beer M, Mettenleiter TC. 2010. Systemic influenza virus H5N1 infection in cats after gastrointestinal exposure. Influenza Other Respir. Viruses 4:379–386 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Zhou B, et al. 2009. Single-reaction genomic amplification accelerates sequencing and vaccine production for classical and swine origin human influenza A viruses. J. Virol. 83:10309–10313 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.