Abstract
Vaccination against avian influenza (AI) is now included amongst the prevention and control measures recommended by international animal health organizations to combat the disease in poultry. For optimal control of human influenza infections, the antigenic variability within subtypes requires the annual update of seed strains for inclusion in vaccines. The decisions taken are based on serological cross-reactivity of viral strains measured by haemagglutination inhibition (HI) tests. The reason for this is to ensure that the vaccine contains strains that are related antigenically to the current circulating field strain as field viruses evolve or are substituted by variants of distinct antigenicity. Such an annual approach is not viable economically for the poultry industry. In the current study, we have applied a similar HI-based approach to demonstrate, as proof of principle, that cross-reactive strains can be identified. Applying the same approach used by the World Health Organization to investigate antigenic differences among human influenza viruses, we assessed the serological cross-reactivity of a selection of natural H5 and H7 subtype viruses. Analysing HI data, we have identified strains that are cross-reactive and may have the potential to act as seed viruses for future vaccine development. This study should be considered a starting point for a more informed approach to the selection of seed strains for the development of avian influenza vaccines against field infections caused by viruses of H5 and H7 subtypes.
Introduction
Avian influenza (AI) is considered one of the major threats for the global poultry industry and for rural and hobby flocks in Eurasia and Africa. In addition, some strains such as H5, H7 and H9 viruses may have important human health implications (Capua & Alexander, 2004).
Until recent times AI, in its highly pathogenic form (HPAI), was a sporadic disease in poultry and was controlled principally through eradication using defined humane slaughter policies for infected flocks. Recent changes in nomenclature for trade purposes (OIE, 2005) have resulted in the inclusion of all H5 and H7 subtype viruses under the definition of notifiable AI. A similar definition has been adopted under European Union legislation (CEC, 2006a). The reason for this modified nomenclature is that H5 and H7 AI viruses of low virulence (LPAI) may act as progenitors of the HPAI viruses. In addition, recent legislation includes revised control strategies to include vaccination systems that allow a Differentiation of Infected from Vaccinated Animals (DIVA) strategy (CEC, 2005). For trade purposes, vaccination—enacted and monitored appropriately—does not impact negatively on the continuation of international trade (OIE, Citation2008).
The results of surveillance efforts in wild and domestic birds in Eurasia and Africa indicate that there is an ongoing threat of introduction of notifiable AI viruses into domestic poultry. Recommendations issued by bodies such as the European Food Safety Authority (2007) include the establishment of vaccine banks for poultry containing isolates of H5 and H7 subtypes for emergency or prophylactic use.
Human influenza is controlled by annual vaccination campaigns in defined “at-risk” populations. The updating of the antigenic component of the vaccine is based on the antigenic characteristics of viruses circulating currently. Such characteristics are defined by the analysis of haemagglutination inhibition (HI) data. HI tests and the analysis of HI data are therefore the main tools for the selection of vaccine candidates for the control of human influenza infections.
In contrast, little information is available on the antigenic characteristics of AI viruses and the selection of vaccine candidates for AI does not follow any defined criteria.
Studies on the antigenicity of human, equine and swine influenza viruses have shown that there is significant variability in antigenic properties even if they belong to the same H subtype and genetic lineage (Bryant et al., Citation2009; Kuntz-Simon & Madec, Citation2009).
Studies on AI viruses have shown that the antigenic characteristics of avian and human H5 viruses are also variable and that the efficacy of a vaccine in terms of clinical protection and reduction of virus shedding is influenced by the antigenic relationship between the challenge and vaccine virus (Shortridge et al., Citation1998; Horimoto et al., Citation2004). Phylogenetic studies on H5N1 HPAI viruses have indicated that antigenic diversity, even within the same subtype, is likely to be significant (Shortridge et al., Citation1998; Horimoto et al., Citation2004; Soda et al., Citation2008; Wu et al., Citation2008).
In addition, Yassine et al. (2008) demonstrated the existence of antigenic variants amongst the H3N2 viruses circulating in swine and turkey populations in the US by the Archetti and Horsfall method (Archetti & Horsfall, Citation1950) and proposed that the antigenic diversity was the cause of a H3N2 subtype vaccine failure in turkeys.
These reports imply that vaccine viruses developed in the past may not be related sufficiently to contemporary strains to elicit an adequate immune response against contemporary viruses.
To date no data are available on the degree of intra-subtypic cross-reactivity amongst Eurasian H5 and H7 viruses. It would seem reasonable for the intended development of a vaccine bank for Europe, to be able to select as vaccine seed viruses, strains that offer the greatest breadth of antibody response to cover the greatest number of circulating influenza strains possible for that subtype.
The aim of the present study was to investigate the antigenic relatedness and intra-subtype serological cross-reactivity patterns amongst selected avian H5 and H7 viruses by the analysis of HI data. In order to mimic the field situation, we used an adjuvant vaccine formulation based on whole virus antigens.
The present study can be considered a starting point for a more informed approach to the study of cross-reactivity patterns of AI viruses.
Material and Methods
To evaluate the potential use as vaccine candidates, viruses were tested against a panel of homologous and heterologous antisera. Sera were produced using adjuvant to evaluate how a serum produced following a vaccine immunization cross-reacts in vitro, thus mimicking the method of vaccine preparation and administration used most commonly in the field situation.
Selection of viruses
Avian H5 and H7 viruses available at the IZSVe repository prior to 2007 were selected on the basis of their epidemiological relevance, country and year of isolation. The majority of strains were isolated in Italy as part of national surveillance studies and from Africa and the Middle East as a result of collaborative studies and as part of the OIE/FAO mandate requiring confirmatory diagnosis. The amino acid sequences of the HA gene of the H7 and H5 viruses were also analysed phylogenetically in order to include strains representative of different genetic clades.
H5 viruses
Twenty-seven isolates of H5 subtype obtained between 1980 and 2006 were used in the present study, as detailed in . Twenty-one H5N1 HPAI viruses were selected according to the genetic clade. Viruses of clades 0, 1, 2.1, 2.2, 2.3 and 2.5 (WHO, 2008; Cattoli et al., 2009), collected in Asia, Europe, Africa and the Middle East between 2004 and 2006, were investigated. The remaining six strains were selected as viruses unrelated genetically to the H5N1 HPAI viruses isolated between 2004 and 2006 ().
Table 1. Viruses selected for cross-HI tests.
H7 viruses
Twenty-six H7 isolates were used in this study. Representative H7N3 LPAI strains isolated between 2001 and 2004 in Italy in domestic birds and H7 LPAI strains obtained from the wild-bird surveillance programme between 2004 and 2007 in Italy were selected. Other H7 isolates obtained from European and non European countries were also included: A/turkey/England/647/77; A/chicken/England/4054/2006; A/macaw/England/626/80; A/African starling/England-Q/983/79; A/chicken/Pakistan/447/95; A/turkey/Italy/1067/99; and A/chicken/Netherlands/2586/2003.
Sequencing of the HA gene and phylogenetic analysis
Viral RNA was extracted from infectious allantoic fluid using the High Pure™ RNA Isolation Kit (Roche) and was reverse-transcribed with the SuperScript III Reverse Transcriptase kit (Invitrogen, Carlsbad, California, USA). Polymerase chain reaction amplification was performed using specific primers (primer sequences available on request). The complete coding sequences were generated using the Big Dye Terminator v3.1 cycle sequencing kit (Applied Biosistems, Foster City, California, USA). The products of the sequencing reactions were purified using PERFORMA DTR Ultra 96-Well kit (Edge BioSystems, Gaithersburg, Maryland, USA) and uploaded onto a 16-capillary ABI PRISM 3130xl Genetic Analyzer (Applied Biosistems). Phylogenetic analysis of the HA amino acid sequences was carried out using the neighbour-joining method in the MEGA 3 program (Kumar et al., Citation2004). The accession numbers of viruses used are as follows: CY024738, AF202226, EPI167285, DQ83851, EPI167295, DQ838513, EPI154974, EPI167288,CY020589, CY034750, CY022613, EPI154958, CY020613, EPI185457, AY586411, AF202247, EF467826, AF149295,AJ584647, DQ838515, GQ247853, CY021493, CY020584, AF202250, DQ412997, DQ862001, DQ862002, DQ862000, CY020661, CY016795,CY016867, CY016875, CY021517, CY016923, CY016931, EF541401, AF084267, EU146622, EF619980, AY684894, CY016803, CY016883, DQ661910, DQ44053J, CY016284, GQ247850, CY021525, AB166862, and DQ838508.
Production of hyperimmune sera
Nine H5 isolates and 11 H7 isolates were selected to produce hyperimmune sera (). All antigens used to produce sera had an HA titre ranging between 1:128 and 1:256. The purity of the antigens was confirmed by HI tests and sequencing methods as conducted by an OIE reference laboratory for the production of certified reagents.
The abbreviations used for H5 and H7 antisera are reported in . Viruses were inactivated with 0.05% (v/v) β-propiolactone and mixed in 30/70 water-in-oil emulsion adjuvant (Montanide ISA 70 VG [Seppic®]) as recommended by the manufacturer. This adjuvant was selected as it is used widely in the field for avian vaccines. The emulsion was used to immunize 6-week-old specific pathogen free chickens by administering 0.5 ml doses of the preparation subcutaneously twice, 2 weeks apart. Hyperimmune sera were collected by bleeding the birds from the wing vein 2 weeks after the second administration. Blood was allowed to clot in plastic vessels at room temperature overnight. Vessels were centrifuged at 1000 x g for 10 min. Sera were collected and pooled, and 1 ml aliquots were placed in vials stored at –20°C until use. Prior to testing, sera were treated with receptor-destroying enzyme to remove any non-specific reactions (WHO, 2002; OIE, 2004).
Cross-HI tests
The HI test was carried out according to the European Union diagnostic manual (CEC, Citation2006a). Briefly, 4 HA units of virus and a 1% chicken red blood cell suspension were used in the test.Viruses of the H5 and H7 subtypes used to produce hyperimmune chicken antisera are presented in .
Antigenic relatedness study
The antigenic relatedness between isolates was evaluated using the Archetti and Horsfall formula based on cross-HI data (Archetti & Horsfall, Citation1950). The degree of antigenic relatedness, indicated by the R value, is estimated for each pair of viruses by calculation of the homologous and heterologous titre ratios. As a result, this analysis can only be carried out between viruses for which homologous antisera are produced. For example, the antigenic relatedness is measured between virus x and virus y. If virus x has an HI titre of 1:256 with the homologous antiserum and an HI titre of 1:32 with the heterologous antiserum (antiserum y), the ratio of these two values (1:256 and 1:32) is used to calculate the R value (Archetti & Horsfall, Citation1950). The general consensus is that viruses related serologically have R values greater than 50% (Archetti & Horsfall, Citation1950; Lee et al., Citation2004).
One-way serological tests
The HI test was used to determine the intra-subtypic cross-reactivity of the nine anti-H5 and 11 anti-H7 polyclonal antisera against all H5 and H7 viruses investigated (). The degree of cross-reactivity was assessed as the degree of reactivity of a given serum with a selection of viruses measured by HI titre.
Statistical analysis
HI data generated from the one-way serological tests were subjected to statistical analysis in order to determine significant differences in the range of cross-reactivity.
A box plot was used to represent data distribution and the Grubbs test was performed to identify the outliers in a univariate dataset. The Shapiro–Francia test was applied to evaluate the normality of the data. When data were not normal and no mathematical functions were applicable to normalize them, the non-parametric Wilcoxon signed-rank test was used. When data could be normalized (i.e. log) and the homoschedasticity hypothesis was verified by the F-Snedecor test, the parametric Student t test for paired data was applied (Siegel & Castellan, Citation1992; Fisher & van Belle, Citation1993).
Results
Phylogenetic analysis
H7 viruses. A phylogenetic tree was constructed for the H7 subtype viruses used in the present study ().
Figure 1. Box plot showing the distribution of HI titre values obtained testing all H5 viruses against all H5 antisera. Each box refers to HI data generated testing each serum and indicates the degree of dispersion in the data; the line inside the box is the median, the points above and below the box are the outliers.
Figure 2. Box plot showing the distribution of HI titre values obtained testing all H7 viruses against all H7 antisera. Each box refers to HI data generated testing each serum and indicates the degree of dispersion in the data; the line inside the box is the median, the points above and below the box are the outliers.
Figure 3. Phylogenetic tree of the amino acid sequences of the HA gene of H7 viruses constructed by the neighbour-joining method. The viruses included in the present antigenic study are labelled with a circle. The numbers at each branch point represent bootstrap values and were determined by bootstrap analysis using 1000 replications. Scale bar = 0.005 amino acid substitutions/site.
The phylogenetic analysis revealed the existence of distinct clusters of viruses. The Italian H7N3 viruses isolated between 2001 and 2004 grouped together in a distinct branch of the phylogenetic tree, with the distribution of the respective clusters reflecting the collection dates of the samples.
Most of the H7N3 viruses isolated in 2002 formed a separate group from the strains isolated between 2003 and 2004. The H7N3 viruses collected in 2007 were also distributed in a separate branch of the phylogenetic tree. An amino acid sequence similarity less than 96% was found between the H7N3 viruses isolated in 2007 and the remaining Italian H7N3 viruses.
The majority of the Italian H7N7 viruses were also distributed in a distinct cluster of the phylogenetic tree containing two main sublineages. One lineage was formed by the H7N7 viruses of wild bird origin, most of which were isolated in 2004, and the other was composed of the H7N7 viruses obtained subsequently from poultry in 2006. Clustering was not absolute: A/mallard/Italy/4810-79/2004 (H7N4) clustered with the H7N7 viruses, and a mallard isolate of the H7N7 subtype isolated in 2004 was placed in the H7N3 cluster (). These events are most probably the result of genetic reassortment.
Antigenic relatedness study
H5 viruses. Results of the antigenic relatedness study, based on the Archetti and Horsfall formula, are presented in . A high level of antigenic relatedness (R ≥ 50%) to each virus tested was displayed by A/duck/Italy/775/2004 (H5N3, IT/775) and by A/mallard/Italy/3401/2005 (H5N1, IT/3401). Viruses A/chicken/Italy/22A/1998 (H5N9, IT/22A), A/chicken/Yamaguchi/7/2005 (Yamaguchi) and A/Nigeria/957/2006 (Nigeria/957) showed high R values with the majority of viruses tested. The exceptions to this observation were the H5N9 IT/22A virus exhibiting R values lower than 50% with the H5N1 Yamaguchi and Iranian strains, and the Nigeria/957 strain for which R was 44% with the A/turkey/Italy/1980 (H5N2/80). Strains H5N2/80, IT/1258 (H5N2), Yamaguchi and IT/3401 (H5N1) were all related antigenically to each other with R values of 80% or greater. The Middle Eastern and African H5N1 HPAI viruses (isolated in Iran, Afghanistan and Nigeria) tested were related antigenically to each other with R values ≥69% and showed lower antigenic relatedness to each of the other isolates examined (≤50%). The exception to this observation was the LPAI H5N1 virus IT/3401, for which the R values obtained were either 62% or 63% ().
Table 2. Antigenic relatedness of H5 viruses determined from cross-HI test results.
The antigenic relatedness study showed that all of the Middle Eastern and African H5N1 HPAI viruses shared lower R values (<52) with the H5N2/80 strain and strain IT/1258 (H5N2) ().
One-way serological tests
Data from the one way serological tests were analysed statistically, and the results are presented in .
Table 3. Results of statistical analysis on one-way serological data: comparison of all possible pairs of data.
The statistical analysis revealed a higher cross-reactivity of the antisera against strain Iran/754 (H5N1) and Nigeria/957 (H5N1) than all the other antisera (P < 0.05) ( and ). The comparison of the degree of cross-reactivity revealed no significant difference between these two sera ().
The Yamaguchi (H5N1) antiserum exhibited a higher degree of cross-reactivity when compared with the IT/1258 (H5N2), H5N9 and Afg/1207 (H5N1) antisera (P < 0.05). None of the differences of cross-reactivity observed between the Yamaguchi (H5N1) and the anti-H5N2/80, IT/3401 (H5N1) and IT/775 (H5N3) sera reached a level of statistical significance.
H7 viruses
Antigenic relatedness study. The results of the antigenic relatedness study carried out on H7 viruses are presented in . Virus IT/2987 (H7N3) showed high antigenic relatedness with all viruses, with R values of 100% for all comparisons—except IT/X1 (H7N5), Eng/626 (H7N7) and IT/4479 (H7N3), for which R values were 70%, 70% and 50%, respectively. The H7N3 (IT/4479) and H7N7 (NL/621557) viruses were distant antigenically (R < 35%) from all other viruses tested except for IT/2987, for which the R values were 50% and 100%, respectively ().
Table 4. Antigenic relatedness of H7 viruses determined from cross-HI test results.
One-way serological tests
Results of the statistical analysis of one-way serological tests are presented in . The one-way serological analysis revealed the highest degrees of cross-reactivity by antisera against strains IT/2987 (H7N3), IT/9289 (H7N3) and IT/4810 (H7N4).
Table 5. Results of statistical analysis carried out on one-way serological data for H7 viruses: comparison of all possible couples of data.
A higher level of cross-reactivity was observed with antiserum against virus IT/2987 than the other antisera used in this study (P < 0.05) ( and ). The IT/9289 (H7N3) antiserum showed a higher degree of cross-reactivity (P < 0.05) in comparison with other antisera, with the exception of antisera IT/2987 and IT/4810. These antisera exhibited a lesser and an equal degree of cross-reactivity, respectively. The degree of cross-reactivity of the IT/4810 (H7N4) antiserum was lower (P < 0.05) only when compared with the antisera against IT/2987 (H7N3) and IT/9289 (H7N3) ().
Discussion
Studies on the antigenicity of human, equine and swine influenza viruses have shown that there is a significant variability in the antigenic properties of influenza viruses even if they belong to the same subtype and lineage. In humans, the antigenic variability within subtypes requires the annual update of seed strains to be included in vaccines.
Protection against AI is dependent mainly on the generation of antibodies against the HA, which also displays neutralizing activity in vitro. The HA is a surface glycoprotein involved in the immunological response as a major immunogen and is essential for viral attachment to the host cell. The HA protein elicits the production of specific antibodies, which can be measured by means of the HI test, thus yielding quantitative and qualitative information on the immune response against this major antigenic determinant.
Since vaccination against AI infections is now included amongst prevention and control measures recommended by international animal health organizations to combat the disease in poultry, we suggest that HI testing can be used to investigate the serological cross-reactivity properties of AI viruses with a potential application for the selection of candidate vaccine master seeds.
The strategy used to select the influenza A HA antigen(s) to be included in vaccines for seasonal human influenza evaluates the antigenic properties of the circulating viruses in the human population, in order to establish the degree of viral antigenic variability and the virus that shows the broadest cross-reactivity based on the analysis of cross-HI data.
This method is thus used to investigate antigenic differences amongst influenza viruses that may have repercussions on the ability of potential vaccine candidates to elicit a protective immune response. The rationale is that vaccination can only be successful if the antigenicity of the vaccine strain matches that of the circulating viruses (WHO, 2009).
Recent studies on avian human H5N1 viruses have shown that the antigenic characteristics of these viruses are variable and that vaccine efficacy may or may not be influenced by the antigenic relationship, measured by HI tests, between the challenge and vaccine virus (Swayne et al., 2006; Terregino et al., 2009). We believe that this concept may be important in the selection of master seed strains for poultry vaccines as well.
Although HI testing has been proved to be a valuable tool for the study of the antigenic characteristics of influenza viruses for vaccine selection purposes or to investigate antigenic drift, it has not been used in the same way to evaluate AI viruses in poultry vaccines. The analysis of HI data has often been conducted without mathematical and/or statistical support, until recently when an antigenic cartography approach was proposed for influenza (Smith et al., 2004). This method allows the quantitative analysis of an extensive HI dataset and was developed primarily to investigate the antigenic differences amongst isolates. Several attempts to apply such methods to AI viruses have been made and efforts continue. This variation in approach to the problem suggests that a standardization of a protocol for the selection would be beneficial to optimize levels of protection obtained from future AI vaccines.
In the present study we have investigated the serological cross-reactivity properties of a selection of natural H5 and H7 subtype viruses applying the same approach used by the WHO to investigate antigenic differences amongst human influenza viruses.
Our results indicated that, for H5 viruses, two LPAI isolates obtained from waterfowl—IT/3401 (A/mallard/Italy/3401/2005(H5N1)) and IT/775 (A/duck/Italy/775/2004(H5N3))—showed the broadest cross-reactivity. The one-way serological tests also demonstrated that the H5N1 HPAI strains Iran/754 and Nigeria/957 showed a high cross-reactivity with all strains tested, suggesting that these viruses also have HA antigenic characteristics suitable as seed viruses for the generation of vaccines. However, since they are HPAI strains, they could only be suitable if transformed in to low pathogenic viruses by modifying their cleavage site motif.
With reference to the avian H7 isolates, no data on their antigenic characteristics are available in the published literature and the present study provides, for the first time, such information.
Our study with viruses of the H7 subtype indicates that the H7N3 virus IT/2987 shared the highest R values with all viruses tested. The statistical analysis of the one-way HI titres confirmed this finding with a high level of cross-reactivity of the antiserum against virus IT/2987.
The broad cross-reactivity of anti-H5 and anti-H7 sera used in this study may be explained by the existence of common epitopes amongst intrasubtypic haemagglutinin proteins as described previously for H3 viruses (Sui et al., 2009; Tsai et al., 2010). In vivo studies evaluating cross-protection would be additive to our results allowing the correlation between HI titres and clinical protection.
For the selection of a vaccine candidate to be used for conventional, whole, inactivated virus vaccine formulations to be practicable, the growth characteristics and the yield of viral antigens are important issues to be considered along with the antigenic properties. The use of adjuvant has also been shown to improve the immunogenicity of vaccine preparations, and this may overcome the antigenic variation displayed within a certain subtype. Continued use of this or other strategies to augment the cross-protective properties of the master seeds selected should be encouraged.
Health authorities may find information provided in the study useful to guide decision-making when establishing vaccine banks for use in avian species as proposed for the European Union (European Food Safety Authority, 2007). However, national authorities should base the selection of vaccine candidates on several factors such as the geographical spread, epidemiology, antigenic and genetic properties of the field viruses that threaten their poultry populations. Given the requirement to update constantly datasets to remain relevant to the evolving field situation, it would seem advisable that institutions involved in international efforts on surveillance in wild and domestic birds join forces to generate information to support both the animal and public health aspects of influenza virus infections.
The current study should be regarded as a first step towards the establishment of a constant monitoring system for the antigenic characteristics of circulating AI viruses by testing novel isolates obtained through ongoing surveillance programmes. Additional studies are necessary to determine whether the H5 and H7 viruses identified will continue to provide the broadest cross-reactivity to other isolates as new viruses emerge or as contemporary viruses evolve and to understand the correlation between the cross-reactivity of viruses measured by HI and the amino acid sequence of HA.
The HI testing approach recommended here highlights the continuing importance of intensified surveillance of animal influenza to assess antigenic and genetic properties of circulating viruses. Its adoption will allow any antigenic variation to be detected and considered when deciding on vaccine strain selection.
Acknowledgements
The present work was part of the European-funded project FLUAID, SSPE-CT-2005-022417. All of the virus sequences used in the study are available in Genbank or the GISAID database. The authors acknowledge Dr G. Koch, Dr R. Fouchier, Dr T. Joannis, Dr Mehrez Aly, Dr Azizullah Osmani, Dr Emmanuel Couacy-Hymann, Dr Do Huu Dung, Dr Magdi D. Saad, and Dr Mitsugu Shimizu for the provision of isolates, Dr A. Toffan and Dr R. De Nardi for their assistance, and thank Drs R. Fouchier and D. Smith for the stimulating and valuable discussions.
References
- Archetti , I. and Horsfall , F.L. 1950 . Persistent antigenic variation of influenza A viruses after incomplete neutralization in ovo with heterologous immune serum . Journal of Experimental Medicine , 92 : 441 – 462 .
- Bryant , N.A. , Rash , A.S. , Russell , C.A. , Ross , J. , Cooke , A. Bowman , S. 2009 . Antigenic and genetic variations in European and North American equine influenza virus strains (H3N8) isolated from 2006 to 2007 . Veterinary Microbiology , 138 : 41 – 52 .
- Capua , I. and Alexander , D.J. 2004 . Avian influenza: recent developments . Avian Pathology , 33 : 393 – 404 .
- Cattoli , G. , Monne , I. , Fusoro , A. , Joannis , T.M. , Lombin , L.H. , Aly , M.M. , et al. 2009 . Highly pathogenic avian influenza virus subtype H5NI in Africa: a comprehensive phylogenetic analysis and molecular characterisation of isolates . PLoS ONE , 4 3 : e4842 . doi: 10.1371/journal.pone.0004842 .
- CEC 2005 . Council directive of 20 December 2005 on community measures for the control of avian influenza and repealing 92/40/EEC (2005/94/EC) . Official Journal of the European Union , L10 , 16 .
- CEC 2006a . Commission Decision of 4 August 2006 approving a Diagnostic Manual for avian influenza as provided for in Council Directive 2005/94/EC (2006/437/EC) . Official Journal the of European Union, L237 , 1 .
- European and Food Safety Authority . 2007 . Scientific opinion on vaccination against avian influenza of H5 and H7 subtypes in domestic poultry and captive birds . EFSA Journal , 489 , 1 57
- Fisher , L.D. and van Belle , G. 1993 . Biostatistics. A Methodology for the Health Sciences , New York : J. Wiley & Sons Inc .
- Horimoto , T. , Fukuda , N. , Iwatsuki-Horimoto , K. , Guan , Y. , Lim , W. Peiris , M. 2004 . Antigenic differences between H5N1 human influenza viruses isolated in 1997 and 2003 . Journal of Veterinary Medical Science , 66 : 303 – 305 .
- Kumar , S. , Tamura , K. and Nei , M. 2004 . MEGA3: integrated software for molecular evolutionary genetics analysis and sequence alignment . Briefings in Bioinformatics , 5 : 150 – 163 .
- Kuntz-Simon , G. and Madec , F. 2009 . Genetic and antigenic evolution of swine influenza viruses in Europe and evaluation of their zoonotic potential . Zoonoses and Public Health , 56 : 310 – 325 .
- Lee , C.W. , Senne , D.A. and Suarez , D.L. 2004 . Effect of vaccine use in the evolution of Mexican lineage H5N2 avian influenza virus . Journal of Virology , 78 : 8372 – 8381 .
- OIE 2004 . Manual of Diagnostic Tests and Vaccines for Terrestrial Animals: Swine Influenza Chapter 2.10.11. Available online at http://www.oie.int/eng/normes/mmanual/A_00137.htm (accessed 26 May 2010)
- OIE 2005 . Manual of Diagnostic Tests and Vaccines for Terrestrial Animals: Avian Influenza (Chapter 2.7.12). Available online at http://www.oie.int/esp/normes/mmanual/a_00037.htm (accessed 26 May 2010) .
- OIE 2008 . Terrestrial Animal Health Code: Avian Influenza Chapter 10.4. Available online at http://www.oie.int/Eng/normes/mcode/en_chapitre_1.10.4.pdf (accessed 26 May 2010)
- Shortridge , K.F. , Zhou , N.N. , Guan , Y. , Gao , P. , Ito , T. Kawaoka , Y. 1998 . Characterization of avian H5N1 influenza viruses from poultry in Hong Kong . Virology , 252 : 331 – 342 .
- Siegel , S. and Castellan , N.J.J. 1992 . Nonparametric Statistics for the Behavioral Sciences , 2nd edn , New York : McGraw-Hill .
- Smith , D.J. , Lapedes , A.S. , de Jong , J.C. , Bestebroer , T.M. , Rimmelzwaan , G.F. , Osterhaus , A.D.M.E. , et al. 2004 . Mapping the antigenic and genetic evolution of influenza virus . Science , 305 , 371 376 .
- Soda , K. , Ozaki , H. , Sakoda , Y. , Isoda , N. , Haraguchi , Y. Sakabe , S. 2008 . Antigenic and genetic analysis of H5 influenza viruses isolated from water birds for the purpose of vaccine use . Archives of Virology , 53 : 2041 – 2048 .
- Sui , J. , Hwang , W.C. , Perez , S. , Wei , G. , Aird , D. , Chen , L.M. , et al. 2009 . Structural and functional bases for broad–spectrum neutralization of avian and human influenza A viruses . Nature Structural & Molecular Biology , 16 , 265 273 .
- Swayne , D.E. , Lee , C.W. and Spackman , E. 2006 . Inactivated North American and European H5N2 avian influenza virus vaccines protect chickens from Asian H5N1 high pathogenicity avian influenza virus . Avian Pathology , 35 : 141 – 146 .
- Terregino , C. , Toffan , A. , Cilloni , F. , Monne , I. , Bertoli , E. , Castellano , S.L. , et al. 2009 . Evaluation of the protection induced by avian influenza vaccines containing Mexican H5N2 LPAI seed strain against contemporary Egyptian H5N1 HPAI virus by means of serological and in vivo test. In Proceedings of the VIIth International Symposium on Avian Influenza 4 7 . Athens. GA, USA
- Tsai , M.T. , Chern , T.C. , Chuang , J.H. , Hsueh , C.W. , Kuo , H.S. , Liau , C.J. , et al. 2010 Efficient simulation of the spatial transmission dynamics of influenza . PLoS Current Influenza, PLoS January 2010, RRN1141 .
- World Health Organization . 2005 . WHO Manual on Animal Influenza Diagnosis and Surveillance . Available online at WHO/CDS/CSR/ NCS/2002.5.Rev.1.2002:77 .
- World Health Organization . 2009 . Recommended composition of influenza virus vaccines for use in the 2009–2010 influenza season (northern hemisphere winter) . Weekly Epidemiological Record , 9 , 65 76 .
- World Health Organization, World Organisation for Animal Health & Food and Agriculture Organization H5N1 Evolution Working Group . 2008 . Toward a unified nomenclature system for highly pathogenic avian influenza virus (H5N1) . Emerging Infectious Diseases 14 . Available online at http://www.cdc.gov/EID/content/14/7/e1.htm (accessed 26 May 2010)
- Wu , W.L. , Chen , Y. , Wang , P. , Song , W. , Lau , S.Y. Rayner , J.M. 2008 . Antigenic profile of avian H5N1 viruses in Asia from 2002 to 2007 . Journal of Virology , 82 : 1798 – 1807 .
- Yassine , H.M. , Lee , C.W. , Suarez , D.L. and Saif , Y.M. 2008 . Genetic and antigenic relatedness of H3 subtype influenza A viruses isolated from avian and mammalian species . Vaccine , 26 : 966 – 977 .