To determine if the change of PB2-627 from E to K in VN/1203 influences horizontal transmission and/or pathogenicity, we infected chickens and ducks with HP VN/1203/K and HP VN/1203/E viruses. On the following day, naïve birds were placed in cohabitation with the inoculated animals. Infection was measured by survival to 14 days pi, isolation of virus from swabs and tissues, and seroconversion.

Chickens infected with HP VN/1203/K or HP VN/1203/E viruses died within 18–24 hours pi. There were no significant differences in replication between the two viruses. Virus was isolated from all organs and oral and cloacal swabs at the time of clinical disease or death and most of the swabs collected at 24 hours pi were positive (Table 1). Virus was not isolated from chickens that subsequently survived until 14 days pi nor were they seropositive, indicating they did not receive an infectious dose. Contact transmission from infected chickens was not established at all for HP VN/1203/E and at a low level with HP VN/1203/K with only 3/14 birds positive (Table 1). All inoculated and contact-challenged ducks became infected as determined by disease or seroconversion in survivors.

To determine the virulence for ducks and the extent of horizontal spread to contact ducks and chickens, ducks were infected with 10^5.8 and 10^6.0 EID₅₀ of HP VN/1203/E and HP VN/1203/K, respectively. Inoculation with either virus resulted in similar rates of clinical disease in ducks and in chickens, resulting in euthanasia or death in some ducks and all chickens. Virus was readily isolated from oral swabs from all inoculated ducks (Table 2) and all surviving ducks were seropositive. Both viruses were transmitted to 100% of contact ducks. Ducks infected with HP VN/1203/E transmitted the infection to contact ducks more efficiently than HP VN/1203/K infected animals, as evidenced by virus isolation from oral swabs collected after 5 days in cohabitation (p<0.05, two-tailed Fisher's exact test) (Table 2). The mortality rates and all other data were not significantly different between the two contact-infected duck groups. All contact infected chickens died or were euthanized between 5 and 7 days post contact with both viruses.

The clinical disease and pathology at the time of death were essentially similar in chickens infected with the HP VN/1203/E and VN/1203/K viruses. Establishment of infection caused mortality or severe depression between 21 and 46 hours pi. The morbidity periods were very short, with some chickens requiring euthanasia within 3 hours of the initial manifestation of clinical signs. There was an absence of gross pathology. Microscopically, there were small foci of necrosis, most noticeably in the red pulp of the spleen, but also in the lung interstitium and lamina propria of the intestine. Tissues of infected birds were assessed for viral antigen by immunohistochemistry at time of euthanasia. High levels of antigen expression was detected mainly in the endothelium (Figure 1A, C, E), indicating that viral replication occurred in all vascularised tissues. Antigen was particularly dense in rich vascular networks, such as in the spleen, lung, glomerular tufts and lamina propria of intestinal villi. Dense antigen was also present in inflammatory foci in the submucosa of various tissues, such as intestine, respiratory airways, liver and kidney. The parenchyma of organs contained low to moderate levels of antigen, and this was most frequent in myocardium (Figure 1E), but also in small foci in the brain, lung, liver, pancreas and kidney tubules.

There were also no noticeable differences in the clinical disease and pathology between ducks infected with the HP VN/1203/E and VN/1203/K viruses. Gross lesions in ducks were often absent or subtle and included mild pale streaking on the heart, pale or red spots on the pancreas and mildly increased fluid in the body cavities. Microscopic lesions included acute, diffuse, mild to severe cardiomyopathy, diffuse mild to moderate non-suppurative encephalitis and acute focal necrosis of the pancreas. Influenza virus antigen was detected in the myocardium, in foci of neurons in the brain, in skeletal and smooth muscle, necrotic foci in the pancreas (Table 3) and in air sac epithelium. Antigen was rare in lung interstitium and kidney tubules. The pathology observed in contact-infected ducks was similar to that in inoculated ducks.

To determine whether the PB2-627 amino acid had changed during the course of infection in vivo we sequenced the corresponding region of viruses recovered from two diseased (for HP VN/1203/E and HP VN/1203/K viruses) contact ducks. In both instances the same sequence as the inoculated virus was obtained suggesting that this position was not undergoing rapid host-driven selection in vivo.

In order to determine the influence of amino acid substitution at PB2-627 on the pathogenesis of infection, chickens and ducks were inoculated with virus and euthanased at intervals for sample collection. Two experiments were conducted in chickens, both with similar dose and route of inoculation, but sampled at different time points. In the first experiment chickens were inoculated with 10^6.0 TCID₅₀ of either virus. Neither HP VN/1203/E nor HP VN/1203/K were isolated from spleen, lung, and brain tissue samples as well as cloacal and oral swabs collected at 6 and 12 hours pi, however, all specimens collected at 24 hours pi were positive. In the second experiment, chickens infected with 10^6.0TCID₅₀ of HP VN/1203/E had virus in different organs at 12 hours pi in contrast with birds infected with the same dose of HP VN/1203/K (Figure 2). At 18 hrs pi, titers in the different organs and swabs were not significantly different between the two viruses. All tissue samples and swabs contained infectious virus by the time clinical disease developed (21–24 hours pi), with titers between 10² to 10⁸ TCID₅₀/mL (Figure 2).

Antigen was not detected in any of the chickens at 6 and 12 hours pi, but was present in all birds sampled at or after 18 hours pi (data not shown). There was no difference in antigen levels between the two viruses at these time points and at time of clinical disease.

With both HP VN/1203/E and HP VN/1203/K virus antigen was detected in duck tissues with peak levels found between 3 and 4 days pi, with a decline in antigen levels by day 6 (Table 3). There was no apparent difference in the tissue tropism between the two viruses with heart and brain most frequently positive. There appeared to be a higher overall prevalence of heart and brain infection in ducks infected with HP VN/1203/E, but the overall differences with HP VN/1203/K were not significant.

Virus was isolated from all ducks in both virus groups at 3 and 4 days pi, but not all organs were positive. Titers in the spleen, lung, brain and cloacal and oral swabs at the different time points are presented in Figure 3. Although there were minor differences between the two virus groups all time points, the differences were not statistically significant. Taken together, these findings suggested that in this model there are no significant differences in the kinetics of virus infection and virus distribution in the major organ systems of chickens and ducks as a result of the change of amino acid at position 627 of PB2.

The virulence of wild-type HP VN/1203 virus in chickens was overwhelming and precluded detection of moderate differences between PB2-627K and PB2-627E infections. Therefore, we developed isogenic viruses with a deletion of the multibasic cleavage site of the HA; i.e. LP VN/1203/E and LP VN/1203/K. To confirm that the

LP VN/1203/K and LP VN/1203/E viruses were indeed of low pathogenicity, we inoculated chickens (n = 6) with 10^6.1 EID₅₀ of either virus and evaluated the response for 14 days. Three of the 6 birds infected with LP VN/1203/E developed ataxia and torticollis and were euthanased, one 8 days pi and the other two 13 days pi. These 3 birds had non-suppurative encephalitis and dermatitis, as described below, and virus antigen associated with lesions. While 5/6 chickens seroconverted after infection with LP VN/1203/K, none showed clinical signs (Table 4). The remaining birds of both groups were euthanased and sampled at 14 days pi; although they survived apparently healthy, many had microscopic lesions in brain and skin. Characteristic lesions that were associated with antigen included focal glial nodules with the presence of mononuclear cell perivascular cuffing in the brain and focal dermal and subdermal necrosis, particularly in the comb, and folliculitis. In the absence of viral antigen in such lesions were considered highly likely as indicative of prior influenza virus replication. Such lesions were present in the remaining three chickens that were euthanased at 14 days pi for LP VN/1203/E and in 1 of 5 chickens surviving to 14 days pi for LP VN/1203/K (Table 4). Viral antigen was present in single neurons in glial nodules in the brain (Figure 1B), and at the stratum granulosum of epidermis of skin (Figure 1D) and feather follicles. The incidence of histopathologic lesions in the brain and skin were significantly higher in birds infected with LP VN/1203/E than with LP VN/1203/K (p<0.05, two-tailed Fisher's exact test, Table 4). All birds that survived until termination developed HI antibodies except one bird infected with LP/VN1203/K. However, the antibodies failed to completely clear the virus based on immunohistochemistry and did not prevent the development of clinical disease in the case of the two birds with nervous symptoms at 13 days pi. To confirm that both LP VN/1203/E and LP VN/1203/K viruses replicated in vivo, oral secretions collected at 6 days pi were analyzed by RT-PCR; viral RNA was detected in both groups of chickens.

Chickens infected with 10^6.1 EID₅₀ of LP VN/1203/E and LP VN1203/K were sacrificed at 1, 2, 3, 4, and 5 days pi and organs were examined for the presence of viral antigen by immunohistochemistry and virus isolation by inoculation of 9- to 10-day old embryonated chicken eggs. None of the chickens showed clinical signs, although some sampled at 4 and 5 days pi had lesions, including glial nodules in the brain and mild, focal necrosis in the dermis. Overall, virus isolation rates were higher for LP VN/1203/E than for LP VN/1203/K. The difference was significant at 3 days pi (p<0.05, two-tailed Fisher's exact test, Table 5).

Viral antigen was detected in neurons and glial cells (with or without associated lesions), in single myocardial fibers (Figure 1F), kidney tubular epithelium, single cells in dermis, connective tissues (follicle sheath, periosteum), nasal gland and feather pulp. Definitive antigen in endothelium, as seen with the HP AIV infections, was not found. Interestingly, LP VN/1203/E caused a higher incidence of virus-antigen positive tissues than LP VN/1203/K starting at 3 days pi. The difference between the total number of antigen positive tissues of all birds combined for 3–5 days pi is significantly higher for LP VN/1203/E (14/18) than LP VN/1203/K (3/18) (p<0.05, two-tailed Fisher's exact test).

Sequence analysis confirmed that the virus obtained from the bird with neurological signs at 8 days pi and from the four birds that we viral-antigen positive still retained the HA cleavage site associated with the LP form as well as the original E at PB2-627.

In conclusion, these studies indicated that there are differences associated with infection in chickens between viruses with E versus K at PB2-627, but that these changes can only be demonstrated when birds were infected with the LP viruses.

Previous studies by Sterz and Weiss [33]–[34] suggested that AIV can replicate in thrombocytes, we therefore examined whether thrombocytes obtained from infected chickens and ducks contained infectious virus. Thrombocytes from all chickens contained infectious virus at 18 and 21–24 hours pi with either HP VN/1203/E or HP VN/1203/K, and at 12 hours pi for 5/6 chickens infected with HP VN/1203/E (Figure 2). In ducks, thrombocyte preparations from only some of the animals contained infectious virus but in contrast with the findings in chickens the titers were extremely low (≤30 TCID₅₀, data not shown). In contrast, thrombocytes from chickens infected with LP VN/1203/E and LP VN/1203/K were negative for virus isolation except for one sample from the LP VN/1203/E group taken at 4 days pi.

Both HP VN/1203/E and VN/1203/K viruses could be isolated at high levels from chicken thrombocytes suggesting that AIV might replicate in these cells or serve as a vehicle for the spread virus in the vasculature. To quantify the viral RNA in thrombocytes, we analyzed purified cells by real-time RT-PCR using primers specific for viral genomic RNA. Chicken thrombocytes collected at 18 hours pi with HP VN/1203/K or HP VN/1203/E had Ct values of 18.6 and 17.0 for genomic RNA and 17.8 and 16.2 for cRNA specific cDNA reactions, respectively. In contrast thrombocytes from uninfected birds had Ct values of >40 for both primers.

The profound transcriptome changes observed in thrombocytes collected at 18 hours pi suggested that HP VN/1203/K and HP VN/1203/E may have initiated replication in chicken thrombocytes or progenitor megakaryocytes. Table 6 shows the complete list of significant changes, defined as ≥2-fold increase in transcription relative to uninfected controls. Analysis of the TLR receptor signaling pathway indicated significant upregulation of TLR3, which can be triggered by dsRNA, and involve STAT1, IRF7, and IFNAR2. In addition several cell adhesion molecules (SELE, VCAM1, CD274), SOCS 1 and MX were also highly upregulated in thrombocytes.

These results showed a major difference in the response of thrombocytes between chickens and ducks inoculated with HP AIV and between chickens inoculated with HP versus LP viruses. The microarray results further support the postulated importance of thrombocytes for the pathogenesis and pathology of HP virus in chickens.

This study was carried out in strict accordance with the recommendations in the National Health and Medical Research Council's Australian Code of Practice for the Care and Use of Animals for Scientific Purposes 7^th edition. All experimental protocols were approved by the CSIRO Australian Animal Health Laboratory Animal Ethics Committee (Approval numbers AEC 1152, 1174, 1300 and 1301) at the Australian Animal Health Laboratory (Bureau of Animal Welfare Scientific Procedures Premises Licence SPPL 113). All efforts were made to minimize suffering.

Four-week-old, straight-run broilers were obtained from a commercial producer. Five-week-old White Pekin ducks were obtained from Luv-a-Duck (Victoria, Australia). All birds were identified individually with leg bands and were assigned to the different treatment groups for each experiment using a randomization tool (www.graphpad.com). Birds were housed in Microbiological Containment level 3 Zoonotic rooms with the temperature set at 22°C and the airflow set at approximately 15 air changes/hour [23].

HP VN/1203/K was derived from the parental wild-type virus by standard reverse genetics methods [70]. Mutations at positions 627 of the PB2 were generated by standard oligonucleotide mutagenesis methods [71] to generate HP VN/1203/E. An additional mutation was generated in both VN/1203/K and VN/1203/E by deleting 4 basic amino acids in the HA cleavage site. These LP viruses are referred to as LP VN/1203/K and LP VN/1203/E, respectively. All viruses were propagated by two passages in 10-day-old embryonated specific-pathogen-free chicken eggs and stored at −80°C until use. Virus stocks were titrated by endpoint dilution to determine the EID₅₀. Following propagation, the full genomes of rescued viruses were sequenced to confirm the presence of the appropriate wild type or mutant sequence.

Birds were inoculated by placing one or two drops of virus in each eye and nostril, and instilling the remainder into the oral cavity. Virus titers were confirmed by back titration in 9- to 10-day-old embryonated chicken eggs.

Pharyngeal and cloacal swabs were collected into 2 mL of phosphate-buffered saline containing penicillin, streptomycin and gentamycin. After euthanasia, fragments of spleen, brain, lung and heart tissues were frozen at −80°C until processing for virus isolation. For some experiments, small amounts of 1.0 mm silicon carbide beads (Daintree Scientific, St Helens, Tasmania, Australia) were added to the collection tubes with phosphate-buffered saline.

Thrombocytes were prepared as described by Scott and Owens [39]. Cells were resuspended in 0.2 ml of HBSS and stored at −80°C until use. The dilution factor of plasma in the final cell preparation was estimated to be between 5 and 6 log₁₀.

Tissue samples were ground by mortar and pestle using sterile sand or pulverized using a MP Fast-Prep24 instrument (MP Biomedicals, Inc, Solon, OH) when collected with silicon carbide beads and diluted in PBS to 10% w/v homogenates. Virus isolations for the transmission experiment in chickens were performed by inoculating 10-day-old embryonated chicken eggs. Allantoic fluids from all eggs were examined for the presence of HA using chicken red blood cells [72]. Tissue extracts and fluids from swabs obtained from birds inoculated with HP AIV virus were titrated directly in Vero cells as described [23] except that low dilutions of samples causing non-specific cytopathic effects (CPE) in cell culture were tested for HA if higher dilutions did not show specific CPE. The limit of detection of HPAI virus in Vero cells was 1.25log₁₀ TCID₅₀/mL.

Blood samples were obtained from all experimental birds prior to challenge and from survivors at termination. Sera were stored at −20°C until used for the hemagglutination inhibition test [72].

Tissues were collected and processed for histology and immunohistochemistry as described by Bingham et al. [23].

Total RNA was extracted from animal tissues using an RNA isolation kit (RNEasy, Qiagen, Doncaster, Australia) according to the manufacturer's instructions. A viral RNA extraction kit, QIAamp Viral RNA Mini Kit, (Qiagen) was used to extract RNA from oral and cloacal swabs according to the manufacturer's instructions. RT-PCR of the HA cleavage site was conducted using the oligonucleotide primers Fwd 5′-TGTCAAGAAAGGGGACTC-3′ and Rev 5′-CTTTGTCTGCAGCGTACC-3′. RT-PCR of the PB2 gene was conducted using the primers Fwd 5′-CCAAAGATTAAACCCCATGCATCA-3′ and Rev 5′-ACATATGGTCTCGTATTAGTAGAAACAAGGTCGTTT-3′ All samples were subjected to electrophoresis and amplicons were visualised and photographed (Kodak Digital Science, Rochester, NY).

PCR products were purified using a QIAquick® PCR Purification Kit (Qiagen, Germany) according to the manufacturer's instructions. DNA sequencing was performed by the sequencing facility at CSIRO-AAHL using an ABI Prism 377 automated DNA sequencer (Applied Biosystems) and the oligonucleotide primers shown above. Nucleotide sequence data and the translated amino acid sequence were subjected to sequence alignments using clone manager 9 (Version 9.0, Scientific and Educational Software).

Real-time RT-PCR to verify the presence of H5N1 was performed to determine if virus was replicating in the thrombocytes, according to the method of Heine et al. [73], using the modifications described below. RNA was isolated from the same samples analyzed in the microarray assay, using the Meridian total RNA isolation kit (Cartagen, Seattle, WA). RNA was converted to cDNA using the Quantitect reverse transcription kit (Qiagen), in two separate reverse transcription reactions with either the forward primer AGCAAAAGCAGGTAGATGTTGAA which anneals to viral HA genomic RNA, or the reverse primer AGTAGAAACAAGGTAGTTTTTTACTCC which anneals to HA cRNA. Following reverse transcription each cDNA sample was subjected to real-time PCR to verify the presence of H5N1 viral RNA and cRNA as described [73]. The thermal cycling parameters consisted of an initial Taq polymerase hot start activation at 95°C for 10 min, followed by 45 cycles of 15 sec at 94°C and 1 min at 60°C. Assays were performed on an AB7900HT real time PCR system (Applied Biosystems) and analyzed using the associated sequence detection software version 2.3 (Applied Biosystems).

Total RNA for all samples was isolated using the Meridian total RNA isolation kit (Cartagen, Seattle, WA). One microgram of total RNA was reverse transcribed into cDNA and indirectly labeled with Cy3 using the NimbleGen One-Color DNA labeling kit (Roche, Basel, Switzerland). Each Cy3 labeled cell culture sample was individually hybridized to the NimbleGen whole genome chicken array (designed in-house) and hybridized according to manufactures instructions. The custom array design has been deposited into ArrayExpress (accession number A-MEXP-2133) and all data is MIAME compliant and has been deposited in ArrayExpress (accession number E-MEXP-3432). Arrays were scanned at 2.0 microns on a NimbleGen MS 200 scanner. Raw data were extracted using NimbleScan software v2.1 (Roche). All arrays were subjected to quality control measures before being included in the analysis. The signal intensities from all arrays were normalized using Robust Multiple-chip Analysis (RMA) [74]. Subsequent statistical tests were carried out using Genespring version 7.2 (Silicon Genetics, Santa Clara, CA) to determine all genes differentially regulated at a false discovery rate (FDR) of p = 0.05. Pathway analysis was conducted using Genowiz Version 4.0.5.3 (Ocimum Biosciences, Gaithersburg, MD).

To determine pathology, viral shedding and transmission, four-week-old broilers or 5-week-old ducks were infected with HP VN/1203/K or HP VN/1203/E. One day later age-matched uninfected chickens or ducks were added to each infected group. Oral and cloacal swabs were collected daily from all birds. Birds were euthanized prior to the development of severe clinical illness and tissue and blood samples were collected for virus isolation, immunohistochemistry and histology. Sera were collected from all surviving birds at the trial termination, 14 days pi.

To study the early development of infection and disease, groups of chickens and ducks were inoculated with HP VN/1203/K and HP VN/1203/E and individual birds were euthanased and sampled at defined time intervals. Chickens were euthanized at 6, 12, 18 and 24 hours pi. Infected ducks were euthanized at 1, 2, 3, 4 and 6 days pi. At each time point, oral and cloacal swabs, spleen, lung and brain tissues and thrombocytes were collected for virus isolation. In addition, tissues were collected for immunohistochemistry and histopathology. Uninfected control birds from the same batch were euthanized at the start of the experiments (chickens) or towards the end (ducks).

Because both of these HP viruses caused disease and or mortality in chickens within 24 hrs, we used LP VN/1203/K and LP VN/1203/E to determine if the presence of E versus K at PB2-627 influences the pathogenicity of infection in chickens. In a pilot trial, infected chickens were observed for 14 days pi to confirm that these viruses were low pathogenicity. Oral swabs were collected from these chickens at 6 days pi and analyzed by real-time RT-PCR. Surviving birds were bled at 14 days pi for serum collection and euthanized. Samples from different organs were collected for immunohistochemistry and histopathology. To further study the pathogenesis of infection with both low pathogenic viruses, chickens were infected and euthanized at 1, 2, 3, 4 and 5 days pi and samples were collected for virus isolation, immunohistochemistry and histology. Uninfected control birds from the same batch were euthanized simultaneously with infected chickens at 1 and 5 days pi.

Where appropriate, the two-tailed Fisher's exact test was used to determine whether numerical differences between virus isolation rates, pathology and immunohistochemistry are significant. Significant differences for all tests were declared at P<0.05.