Virus shedding and transmission
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 (). 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 (). All inoculated and contact-challenged ducks became infected as determined by disease or seroconversion in survivors.
Experimental infection of chickens by inoculation or contact challenge with HP VN/1203/K and HP VN/1203/E.
To determine the virulence for ducks and the extent of horizontal spread to contact ducks and chickens, ducks were infected with 105.8 and 106.0 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 () 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) (). 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.
Infection of ducks and chickens by inoculation or contact with HP VN/1203/K (K) and HP VN/1203/E (E).
Clinical disease and pathology in chickens
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 (), 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 (), but also in small foci in the brain, lung, liver, pancreas and kidney tubules.
Virus antigen expression in tissues of infected chickens.
Clinical disease and pathology in ducks
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 () 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.
Presence of viral antigen in ducks infected with HP VN/1203/K and HP VN/1203/E virus.
Absence of genetic changes following infection and transmission
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.
Pathogenesis of infection and disease
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 106.0 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 106.0TCID50 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 (). 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 ().
Virus isolation from tissues of infected chickens.
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 (). 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 . 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.
Virus isolation from tissues of infected ducks.
Infection of chickens with low pathogenic viruses lacking polybasic cleavage site in the HA
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 106.1
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 (). 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 (). Viral antigen was present in single neurons in glial nodules in the brain (), and at the stratum granulosum of epidermis of skin () 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, ). 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.
Clinical, pathological and virological responses in chickens inoculated with LP VN/1203/K or LP VN/1203/E and monitored for 14 days piA.
Chickens infected with 106.1 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, ).
Pathogenesis of infection with LP VN/1203/K and LP VN/1203/E in chickens.
Viral antigen was detected in neurons and glial cells (with or without associated lesions), in single myocardial fibers (), 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.
Thrombocyte infection with HP and LP viruses
Previous studies by Sterz and Weiss 
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 (). 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. 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.
Two-fold or more upregulated selected genes in thrombocytes 18 hours pi of chickens with HP VN/1203/K and HP VN/1203/E relative to uninfected controls.
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.