Background
Avian influenza (AI) virus causes one of the most economically important diseases of poultry worldwide. AI is classified by the world organization for animal health (OIE) into two forms, low pathogenicity (LP) and high pathogenicity (HP), based on virulence in chickens [
1]. H7 is one of the two most economically important AI virus subtypes because historically all HP AI viruses have been either the H7 or H5 subtype and it is among the most common subtypes in commercial poultry in the world [
1,
2]. In numerous cases the HP form mutated from a LPAI H7 (or H5) virus that was circulating in chickens or turkeys [
3‐
6]. However, not all H7 LPAI viruses become HP.
In the U.S., H7 AI viruses are sporadically recovered from wild birds (WB) and commercial poultry. Many of the outbreaks in commercial poultry [
7‐
9] can be traced to the live bird market (LBM) system of New York and New Jersey where a single LP H7 genetic lineage persisted from 1994 to 2006 [
10,
11]. Few studies have directly compared the pathogenesis of AI virus in the three primary poultry species: chickens, ducks and turkeys. The aim of this work was to characterize the pathogenesis of selected North American H7 LPAI virus isolates from WB, commercial poultry, and the LBMs in the three primary domestic poultry species; chickens, ducks and turkeys.
Discussion
Twelve H7 LPAI viruses were evaluated for their pathogenesis in chickens, ducks and turkeys. The general patterns of virus shed observed here; OP shed peak at 2-4 days PI and overall higher tittered OP than CL shed which peaked later, at 4-7 days PI, is typical of LPAI virus infections in chickens and turkeys following respiratory inoculation [
13]. Similarly, clinical signs, microscopic lesions, and patterns of IHC staining were consistent with what has been reported previously [
13]. The comparison among the species did show consistent differences among the species in disease severity and virus shed. Although there were no statistically significant trends in disease severity by virus isolate (although most of the genes in 9 of the 12 isolates were relatively closely related to each other) there was a clear and significant trend for clinical disease to be more severe in turkeys than in either chickens or ducks. In turkeys the mean maximum clinical scores, which included mortality, were significantly higher than those of either chickens or ducks with 8 of 12 isolates. Mortality in turkeys possibly had a secondary bacterial component, which is typical in the field with respiratory viruses. Importantly, mortality was not observed in the controls, varied among the AIV isolates, and the birds which died shed the highest virus titers, therefore it appears that AIV was a critical factor for turkey mortality. Additionally, the turkeys generally shed significantly more virus from both the OP and CL routes at numerous time points. This suggests that turkeys may be more susceptible to disease from LPAI virus than chickens or ducks. This is consistent with a report by Tumpey et al. where turkeys were reported to be more susceptible than chickens to LP H7N2 AIV from the 2002 Shenandoah Valley outbreak [
14]. In contrast, Ladman et al. [
15] reported that chickens were more susceptible to disease, when inoculated with LP H7N2 AI viruses isolated from chickens when exposed by the conjunctival sac route.
Although all of the individual birds did not seroconvert, serology indicated that each of the species did become infected at the dose administered and not all birds that seroconverted developed disease. Ducks had the lowest seroconversion rates, interestingly three of the five isolates with 100% seroconversion rates were the WB isolates. In contrast turkeys had the highest rates of seroconversion, which correlated with the trend for disease in turkeys to be more severe than in either chickens or ducks. This suggests that virus dose may have been a factor in disease development.
Also, consistent with the differences in disease severity among the species, susceptibility studies have reported the 50% bird infectious dose (BID
50) for A/Turkey/VA/158512/2002 H7N2, which was isolated from the same outbreak as TK/VA/67, to be 10
2.8EID
50 in chickens, 10
0.8EID
50 in turkeys and 10
3.5EID
50 in Pekin ducks [
16], indicating that it is best adapted to turkeys. In contrast the BID
50 for ML/OH/421 has been determined to be 10
6.6EID
50 in chickens, and10
1.0EID
50 in mallard ducks (data not published). Interestingly, ML/OH/421 had some of the lowest shed titers of all the viruses in all three species and was shed at the highest titers by turkeys. Turkeys also had the highest clinical scores of all three species with ML/OH/421. Importantly, it is unknown how the chicken egg passage that was used to propagate these viruses may affect adaptation to chickens and turkeys.
These results are similar to recent comparable studies with 20 H5 subtype LPAIV viruses [
17], where turkeys were more susceptible to infection and disease, however mortality was not observed with the H5 isolates. Another comparable study with 16 H4, H6 and H9 subtype viruses [
18] did not show differences in disease between chickens and turkeys, but based on serology, turkeys could be infected with more viruses that chickens. In contrast, taking into account data from Halvorson et al. [
19] that showed that not all duck origin AI viruses will readily transmit to turkeys, it is unclear whether turkeys are more susceptible to infection with influenza or are just more susceptible to disease when they do become infected.
The details of species susceptibility to LPAIV infection and disease needs to be explored with reverse genetics which can more definitively identify markers of host restriction and virulence. Additionally, minimum infectious dose studies would be necessary to establish host adaptation of the isolates. In the case of the wild bird isolates one may conclude that the isolates are more duck or shorebird adapted which seems to be supported by the virus shed levels by different species in this study. However, since numerous avian species are housed in close proximity in the live-bird markets, neither the exact host passage history nor the host adaptation can be inferred from the species of origin.
One practical implication of this is that infection would likely be detected more easily in turkeys, whereas sub-clinical infection may spread unnoticed until the birds are tested prior to slaughter (currently 100% of turkey and chicken flocks are tested for AI virus prior to slaughter in the US).
Since both LPAI viruses and HPAI viruses are reported to primarily cause sub-clinical disease in wild mallard ducks, the induction of mild disease by LPAI virus in Pekin ducks which are most closely related to mallard ducks but which are bred for rapid growth, is important for the commercial duck industry. It has been reported that some Asian lineage H5N1 HPAI virus isolates can cause disease in ducks, but the severity of disease depends on duck age and species [
20,
21]. Two-week old Pekin ducks, like what was used here, were among the most susceptible to disease with the Asian H5N1 HPAI virus [
20].
To complement the clinical data and to provide a more complete characterization of the isolates, full genome sequence was produced for all 12 isolates and a basic analysis was conducted. In depth analysis of the HA and NA genes of the H7N2 isolates has already been reported [
10,
11]. There was no clear correlation between gene constellation, or a particular gene and any biological or pathogenic characteristic evaluated here. Further work with reverse genetics would need to be conducted to identify markers for species adaptation and virulence.
Methods
Viruses
Twelve North American origin LPAI viruses were selected to represent different H7 HA genetic groups, different species of origin and different dates of origin within the available H7 subtype domestic LPAI viruses (Table
1). Viruses were propagated and titrated in 9 to 11 day-old embryonated chicken eggs by standard procedures [
22].
Pathogenesis studies
Pathogenesis studies were conducted with specific pathogen free (SPF) white leghorn chickens (Gallus gallus domesticus), broad breasted white turkeys (Meleagris galopova) and Pekin ducks (Anas platyrhynchos domesticus). Chickens were obtained at 2 weeks of age from a commercial supplier of SPF animals (Charles-River SPAFAS, Franklin, CT) and were housed in isolators until they were exposed to the virus at 4 weeks of age. There are no sources of turkeys and ducks which are maintained as SPF or free from viral or bacterial respiratory diseases therefore turkeys and ducks were obtained from commercial hatcheries at hatch and were housed in isolators until they were exposed to virus at 2 weeks of age. All birds were obtained from flocks with no antibody or prior exposure to AI virus. The birds were housed in glove-port isolators (Allentown Caging, Allentown, NJ) with ad libitum access to feed and water before and after exposure to the viruses. Ducks and turkeys were exposed to the virus at 2 weeks instead of 4 weeks of age to accommodate the larger size and faster growth rates of these species in the animal facilities. Since the immune systems of all three species at these ages are considered to be relatively immature, this difference is not expected to impact species associated differences in susceptibility to LPAIV infection and disease. Animals were cared for in accordance with established humane procedures and biosecurity guidelines.
Thirteen to 15 of each species were inoculated with 106 EID50 per bird in 0.1 ml by the intrachoanal route. The birds were monitored daily for clinical disease signs which were scored as follows: 0 = no signs, 1 = mild to moderate respiratory signs (mild depression in ducks), 2 = moderate to severe (i.e. depressed, not eating, neurological signs), 3 = Dead. Oro-pharyngeal and CL swabs were each collected at days 2, 4, 7, 10 and 14 post inoculation (PI) to evaluate virus shed by quantitative real-time RT-PCR (qr-RT-PCR). Three days PI, 3-5 birds from each group were euthanized and necropsied to evaluate gross lesions. Tissues (heart, lung, pancreas/duodemun, kidney, liver, ileum, jejunum, ceca, bursa, thymus, spleen, breast muscle, thigh muscle, brain, nasal cavity, adrenal glands, cecal tonsils, trachea, and reproductive organs) were collected for microscopic evaluation. Serum was collected from ducks at 18 days PI and from chickens and turkeys at 21 days PI to confirm infection status.
qrRT-PCR
RNA was extracted with the MagMAX-96 Viral Isolation Kit (Ambion Inc. Austin, TX) with the KingFisher (Thermo-Fisher Scientific, Waltham, MA) magnetic particle processor in accordance with the manufacturer's instructions. Quantitative real-time RT-PCR was conducted with a primer-probe set that targeted the matrix gene as described previously [
23] using the AB 7500 FAST (Applied Biosystems, Foster City, CA) instrument and the AgPathID (Ambion) one-step RT-PCR kit in accordance with kit instructions. Standard curves for virus quantification were established with RNA extracted from dilutions of the same titrated stock of the virus being evaluated.
Immunohistochemistry
Because of the sensitivity limitations of immunohisotochemical (IHC) staining for AI virus antigen, tissues were only processed for IHC from birds with OP or CL shed titers greater than 10
4EID
50 2 days PI. Tissue sections were cut (4 μm thick) from paraffin-embedded tissue samples and mounted on charged glass slides (Superfrost/Plus; Fisher Scientific). Deparaffinization, antigen retrieval and blocking procedures have been previously described [
24]. A 1:2,000 dilution of a mouse-derived monoclonal antibody (P13C11) specific for a type A influenza virus nucleoprotein (developed at Southeast Poultry Research Laboratory, USDA) was applied and allowed to incubate for 2 hours at 37°C. The primary antibody was then detected by the application of biotinylated goat anti-mouse IgG secondary antibody using a biotin-streptavidin detection system (Supersensitive Multilink Immunodetection System, Biogenex). Fast Red TR (Biogenex) served as the substrate chromagen, and hematoxylin was used as a counterstain.
Serology
Antibody collected at 18 (ducks) or 21 (chickens and turkeys) days PI from surviving birds was used to confirm infection status. Sera were tested by commercial ELISA (FlockCheck, IDEXX Inc., Westbrook ME). Sera were tested at the manufacturer's recommended dilution of 1:500 and also at 1:100 and 1:50. The dilution of 1:100 was selected for final analysis because at this dilution there were no false-positives among the sera from negative control birds and there appeared to be better sensitivity.
Sequencing and Phylogenetic analysis
Full genome sequencing of all isolates was performed as previously described [
11]. Genbank accession numbers for new sequence generated for this study are and previously reported sequences are provided in Additional File
3. Phylogenetic analysis was performed using either ClustalV or ClustalW (Lasergene 7.1, DNASTAR, Madison, WI). Trees were constructed with BEAST v. 1.4.8 [
25] using HKY substitution, empirical base frequency, Gamma heterogeneity, codon 2 partitions, relaxed lognormal clock, Yule Process tree prior with default operators with UPGMA starting tree and MCMC length of 10
6.
Statistical analysis
Mean maximum clinical scores were compared by poultry species for each isolate and virus shed was evaluated by species for each day PI and swab type (OP or CL). All comparisons were conducted with the Student's T-test and the Mann-Whitney rank-sum test if the normality test failed (Sigmaplot 11.0, Systat Inc. San Jose, CA). All statistics were evaluated with a significance threshold of p ≤ 0.05
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ES was involved in virus selection and experimental design and conducted sequence and phylogenetic analysis. JG was involved in virus selection and experimental design. BL and LP conducted animal experiments, collected specimens, ran qrRT-PCR, and serological assays. CP evaluated tissue sections for microscopic lesions. MJP conducted the immuno-histochemical staining and evaluation. ETM contributed to data analysis and conducted minimum infectious dose studies. All authors contributed to data analysis and manuscript preparation.