Journal of Virological Methods 213 (2015) 5–11
Contents lists available at ScienceDirect
Journal of Virological Methods journal homepage: www.elsevier.com/locate/jviromet
Short communication
Development and evaluation of an N9-specific enzyme-linked immunosorbent assay to detect antibodies in duck and chicken sera Audrey Schmitz ∗ , Marie-Odile Le Bras, Katell Louboutin, Véronique Jestin French Agency for Food, Environmental and Occupational Health & Safety, Ploufragan/Plouzané Laboratory, Avian and Rabbit Virology, Immunology and Parasitology Unit, VIPAC, French National Reference Laboratory for Avian Influenza and Newcastle disease, B.P.53, 22440 Ploufragan, France
a b s t r a c t Article history: Received 30 May 2014 Received in revised form 14 October 2014 Accepted 21 October 2014 Available online 18 November 2014 Keywords: Influenza N9 antibodies ELISA Duck Chicken
A serological test for detecting N9-specific antibodies may be useful as a DIVA strategy to differentiate vaccinated from infected animals or simply for direct serological detection of infection with N9-subtype virus. The method currently recommended for the detection of antibodies against neuraminidase is neuraminidase inhibition (NI), which is a laborious method using toxic chemicals and has low sensitivity. The present study describes the development and validation of an N9-specific ELISA. Data obtained with this N9 ELISA were compared to those obtained with nucleoprotein-based ELISA, haemagglutination inhibition test using homologous antigen and NI assay. 785 sera from ducks and chickens were used, from flocks previously determined to be AI negative or from experimentally infected or immunized flocks. Sensitivity and specificity were evaluated, and a ROC curve and kappa values, which provide a comparison between methods, were calculated. The results obtained in this study indicate that the N9 based-ELISA is effective in detecting N9-specific antibodies with high specificity and with better sensitivity than the recommended NI method; using data from 177 common sera tested with N9 ELISA and NI assay both compared to NP-based ELISA, their specificity were evaluated at 93.6% and 91.5% respectively, and sensitivity at 90.8% and 39.2% respectively. © 2014 Elsevier B.V. All rights reserved.
Avian Influenza A viruses (AIV) belong to the Orthomyxoviridae family and subtypes have been identified in AIV based on the antigenic properties of the 16 haemagglutinin (HA) (H1 to 16) and 9 neuraminidase (NA) (N1 to 9) surface glycoproteins. Each subtype is then defined by a combination of one haemagglutinin (H1 to 16) and one neuraminidase (N1 to 9) (Alexander, 2000; Fouchier et al., 2007). These two proteins are the major surface integral membrane glycoproteins of influenza viruses: HA binds to sialic acid and NA catalyzes the removal of terminal sialic acid residues from oligosaccharide chains (Colman, 1994). NA may be important in virulence and the early stage of infection by promoting virus entry or in facilitating release of progeny virions from the surface of infected cells (Matrosovich et al., 2004; Su et al., 2009). Generic tests based on the detection of antibodies against the conserved AIV nucleoprotein (NP) are routinely used as screening tests (Brown et al., 2009). Further to these preliminary tests, haemagglutination inhibition (HI) tests allowing the detection of specific HA subtype antibodies are used as the gold standard either for diagnostic purpose to detect infection by a given HA subtype
∗ Corresponding author. Tel.: +33 2 96 01 64 11; fax: +33 2 96 01 62 63. E-mail address: [email protected] (A. Schmitz). http://dx.doi.org/10.1016/j.jviromet.2014.10.016 0166-0934/© 2014 Elsevier B.V. All rights reserved.
or for assessment of humoral immunity following vaccination using a given HA subtype; although a recent paper demonstrated that for sera from ducks infected or immunized with LP AIV belonging to the H5 subtype of the Eurasian lineage, an H5-based ELISA was more sensitive (Schmitz et al., 2013). In addition to HA identification, a serological test for detecting N9-specific antibodies may be useful in two ways: (i) it could be simply used for direct serological detection of infection in poultry, other captive birds or wild birds with N9-subtype AIV or (ii) in a country which considers AI vaccination to prevent/control AI outbreaks in poultry, a N9-based ELISA test could be used for DIVA strategy to differentiate vaccinated from infected birds either by using a vaccine strain with a N9 subtype different from the NA subtype displayed in the AIV circulating in poultry or the reverse, i.e. by using a vaccine strain with a NA subtype different from the N9 subtype displayed in the AIV circulating in poultry (Avellaneda et al., 2009; World Organization for Animal Health, 2013). N9 viruses are indeed commonly present in both wild birds and poultry, in association with many HA subtypes. The most recent influenza cases concern human and poultry infections with an H7N9 LPAIV which emerged in eastern continental China during the winter of 2013 (Chen et al., 2013; Gao et al., 2013; Li et al., 2014). According to the most recent OMS data available dated
6
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
from the end of June, the number of cases was 450, by the end of summer 2014 the total number of human cases seem to exceed 450. The cases were reported in (i) 15 provinces including the autonomous province of Xingjiang and the capital of Mainland China – including 11 provinces also reporting outbreaks in poultry – and (ii) Hong Kong SAR, Macao and Taiwan and (iii) Malaysia (the cases in the latter four having been contracted in one of the former provinces) (http://www.chp.gov.hk/en/content/599/34480.html; http://www.cidrap.umn.edu/sites/default/files/public/downloads/ topics/cidrap h7n9 update 080414.pdf; http://www.who.int/ influenza/human animal interface/influenza h7n9/riskassessment h7n9 27june14.pdf?ua=1). By the end of June the overall mortality was around 37% (http://www.who.int/influenza/human animal interface/influenza h7n9/riskassessment h7n9 27june14.pdf?ua =1). Forty three outbreaks of low pathogenic avian influenza (H7N9) have been officially notified involving pigeons, chickens and ducks present essentially in live bird markets and/or environment samples collected in the same area, located in twelve provinces of eastern, southern China, (http://www.oie.int/wahis 2/public/ wahid.php/Reviewreport/Review/viewsummary?reportid=13225). Natural infections with this H7N9 virus in poultry are asymptomatic or induce few signs that are difficult to detect; however they may induce a humoral response that can be detected serologically. Domestic ducks play a major epidemiological role in AI, since they are often associated with higher prevalence of subclinical AI infection even in some case of infection with AIV that are highly pathogenic in gallinaceous poultry. In addition, China produces about 5.5 billion chickens for consummation a year and 830 million ducks for consummation a year (number of heads) (FAO data: http://faostat3.fao.org/faostat-gateway/go/to/browse/Q/QA/E), the latter corresponding to 70% of the world’s population of ducks, often as free-range productions which are therefore in frequent contact with wild birds. Improved detection of H7N9 AIV infection in domestic anseriformes would therefore be beneficial in preventing the spread of those viruses to terrestrial poultry. The neuraminidase inhibition (NI) assay is the miniaturized test used for detection of NA subtype specific antibodies (Van Deusen et al., 1983; WHO Global Influenza, 2011). The NI assay is a laborious technique and requires handling toxic chemicals. Its substitution by a more sensitive and specific ELISA test would be useful. The aim of the present paper is to describe the development of a blocking N9based ELISA test and its validation with a panel of approximately 800 sera essentially from ducks and chicken. The present in-house blocking ELISA test was developed using a N9-specific mouse monoclonal antibody (reference IdVet 6B1 A12) produced by IdVet, Montpellier, France. The concentrations of the reagents used, such as coating antigen and monoclonal antibody, as well as the serum to be tested, were optimized. The first step of this ELISA is to coat ELISA plates with a N9 antigen. When a whole AIV is used as antigen and not a recombinant N9 protein, the choice of the antigen for coating is important because HA subtype-specific antibodies can block the access of specific NA (N9 presently) antibodies. This happens because HA subtype-specific antibody binding hampers the access of NA (N9 presently)-specific antibodies to NA (N9 presently)-epitopes. Using a very rare subtype such H15 – that has not been reported (to date) in poultry – should reduce the chances of these unwanted HA antibody/antigen interactions when using whole virus as antigen for a NA specific ELISA. For this reason, the N9 ELISA reported in this study was developed using an uncommon subtype H15N9 A/shearwater/WestAustralia/2576/1979. This virus was grown in 9-day-old embryonated chicken eggs, inactivated with formol at 1‰ overnight at 37 ◦ C, purified with 20–50% gradient sucrose and concentrated 17.5×. Microplates were then coated with the inactivated and purified H15N9 antigen diluted to 1/600 in 0.05 M carbonate buffer (pH 9.6) and incubated at 4 ◦ C
overnight. After washing with PBS (pH 7.4) and 0.05% Tween 20, plates were blocked with PBS-Tween 20 containing 5% milk and 10% foetal bovine serum for 1 h at 37 ◦ C. After washing, serum samples and controls (both previously diluted to 1/5) were added for 1 h at 37 ◦ C, followed after washing by incubating N9-specific monoclonal antibodies diluted to 1/600 for 1 h at 37 ◦ C. The unbound monoclonal antibodies were removed by washing and then phosphatase labelled goat anti-mouse IgG (H + L) (KPL) diluted to 1/400 was added and incubated for 1 h at 37 ◦ C. After washing, the presence of secondary-conjugated-antibody was revealed by incubating the p-nitrophenyl phosphate substrate (pNPP) (Uptima) for 30 min at room temperature. The intensity of the colour measured as the optical density (OD) at 405 nm with a spectrophotometer (Dynex Technologies–MRX Revelation), was correlated inversely to the amount of N9-specific antibodies present in the sample to be tested. Results were expressed as a competition percentage (CP) corresponding to the ratio of the sample analyzed to the negative control [S/N%] = 100 × (OD Sample)/(OD Negative control)). A total of 785 sera were tested. Details of their origin and numbers are provided in Table 1. Status of 521 sera with a known negative AIV were used for establishing the threshold, then 57 reference sera against different AIV subtypes or against pathogens other than AI produced in SPF chickens and ducks were used for checking preliminary specificity. The remaining 207 sera, collected from SPF chickens and N9 ducks experimentally infected or immunized with AIV belonging to the N9 subtype, were used for validation of this N9 ELISA. Most of them were collected from ducks (85.2%), mainly from muscovy ducks. SPF animals were maintained in air-filtered facilities and monitored on a regular basis for ∼20 bacterial and viral avian pathogens including avian influenza virus, to check for their negative status. All experimental immunisations were performed in biological safety level BSL2+ (intermediate between BSL level 2 and 3 – this level requires complete change of clothes, double door access, sealed windows, negative airflow, shower before exit) or BSL3 facilities, approved by the French veterinary authority (entitled officer from the French Ministry of Agriculture) according to EU standards; animal experimentation was performed by duly authorized personnel following reception of approval from the Ethical Committee for experimentation, Cometh ANSES/ENVA/UPEC (National Ethical Committee for Animal Experimentation registration no. 16). All 785 sera were tested with the NP ELISA using a commercial NP-based ELISA test (IDEXX – Elisa IA Multiscreen, recently re-named Influenza A Ab Test). 34 positive sera by NP-ELISA collected from infection/immunization with AIV not belonging to the N9 subtype were removed from the comparative study. Therefore 751 were used for comparison between NP-ELISA and N9-ELISA (Table 1). Sera from experimental infection or immunization (n = 177) with AIV belonging to the N9 subtype were also tested with the HI test (using the same antigen inoculated for immunization or the homologous virus, although inactivated, as the one used for infection) and NI assay. The HI test was performed according to the World Organization for Animal Health manual (World Organization for Animal Health, 2012), using four haemagglutinating units of homologous antigen; HI titres ≥16 (or log2 titre ≥ 4) were scored positive. NI assay was performed as described by Aymard-Henry et al. (1973), except the method was miniaturized and the results were given as OD (Van Deusen et al., 1983). The antigen used was the same as the one used with N9-based ELISA, i.e. the uncommon subtype H15N9 A/shearwater/WestAustralia/2576/1979: as well as N9-ELISA, again as previously described using a very rare subtype such H15 should prevent unspecific reactions. The threshold was determined beforehand using an Inhibition Percentage (IP) (corresponding to [(N − S)/N%] = 100 × (OD negative control − OD sample)/(OD negative control) equivalent to 30% – positive with
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
7
Table 1 Listing of sera used for the development of N9 ELISA. Categories
Species
Age at the moment of sampling
Number of sera tested with N9-based ELISA test
Historical data
Purpose in the present study
SPF ducks
Muscovy duck
19–79 weeks
101
Determination of threshold
Conventional ducks with AI negative status
Muscovy duck
7–49 weeks
420
Reference sera against other avian pathogens than AIV
Muscovy duck and chicken
9–11 weeks
16
Reference SPF pooled sera
Muscovy duck and chicken
Reference sera AIV positive (not N9)
Chicken
8 weeks 2 (duck) 78 weeks (chicken) 9–11 weeks 26
SPF, notably negative AI status as monitored periodically using NP-based ELISA test Negative AI status as tested with NP-based ELISA Experimental immunization with either avian viral pathogen different from AI as mentioned in Fig. 1 Pooled sera of SPF chicken and ducks
Reference sera N9 positive
Chicken
Experimental N5-N6-N7-N8 positive sera
Muscovy duck
6 weeks
8
Experimental N9 sera
Chicken
9–11 weeks
2 × 12
Muscovy duck
16 weeks
2 × 12
Experimental N9 sera
Muscovy duck
10–16 weeks
80
Experimental N9 sera
Chicken
8–10 weeks
3 × 10
Experimental N9 sera
Chicken
8–9 weeks
14
5
Monospecific AIV reference seraa (N1-2-3-4-5-6-78) (cf list of strains in legend) Monospecific AIV reference seraa (N9) (cf list of strains in legend) Experimental infection with inactivated virusb (N5-6-7-8) (cf list of strains in legend) Reference N9 serum from H9N9 LP A/turkey/France/03295/2003 and H11N9 LP A/duck/Memphis/546/1974 serially diluted in SPF pooled chicken sera Serial dilution of 2 positive sera obtained with experimental infection with H9N9 LP A/turkey/France/03295/2003 Experimental infection with wild virus H9N9 LP A/turkey/France/03295/2003 (kinetic) Experimental immunization with inactivated virus H7N9 LP A/Anhui/01/2013 Experimental immunization with inactivated virus H7N9 LP A/teal/France/110007d/2011
n = 521
Set of sera also tested with NP ELISA
√
√
Preliminary evaluation of specificity
n = 57
√
√
Ability to detect n = 207 the smallest antibody titre and inclusive specificityc
Ability to detect the earliest antibody response and inclusive specificityc
Set of sera also tested with HI test using homologous antigen
Set of sera also tested with NI
–
–
–
–
–
–
√
√
√
–
–
√
√
–
–
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
8
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
Table 1 (Continued) Categories
Species
Age at the moment of sampling
Number of sera tested with N9-based ELISA test
Historical data
Purpose in the present study
Experimental N9 positive sera
Muscovy duck
11–12 weeks
35
Experimental vaccination with H5N9 vaccine (containing H5N9 A/chicken/Italy/22A/1998)
Inclusive specificityc
Total
Set of sera also tested with NP ELISA
√
785
Set of sera also tested with HI test using homologous antigen
√
–
751
Set of sera also tested with NI
177
264
SPF birds were maintained in air-filtered facilities and monitored on a regular basis against ∼20 bacterial and viral avian pathogens including avian influenza virus, to check for their negative status. All experimental sera were obtained starting with SPF animals, submitted to the schedule mentioned in historical data. AI, Avian influenza; AIV, avian influenza virus; LP, low pathogenic; HP, highly pathogenic. 751 sera tested with both N9-based and NP-based ELISA test. 177 sera tested with both N9-based ELISA, HI test and NI test. 230 sera tested with both NI test and NP-based ELISA. √ Done. – Not done. a Each monospecific reference chicken AI serum (N1 to N9) was prepared by prime immunization of 6-week-old SPF chickens followed by boosting with the same inactivated virus about 2 weeks later; one to three week after the boost, all sera with a positive HI titre of 512-2048 against the specific viral haemagglutinin subtype were pooled. Influenza virus used were (in order): H1N1 A/duck/Alberta/35/1976, H5N1 A/duck/France/05066b/2005, H5N1 A/common pochard/France/06167/2006, H5N1 A/muteswan/France/070203tr/2007, H5N1 A/duck/France/080036/2008, H7N1 A/mallard/France/06159/2006, H7N1 A/africa starling/983/1979, H3N2 A/turkey/England/1969, H5N2 A/chicken/France/03426a/2003, H5N2 A/duck/France/05056a/2005, H5N2 A/duck/France/05057b/2005, H2N3 A/duck/Germany/1215/1973, H5N3 A/duck/France/02166/2002, H5N3 A/duck/France/05054a/2005, H5N3 A/mallard/France/061054/2006, H5N3 A/muscovy duck/France/070090/2007, H7N3 A/turkey/England/1963, H8N4 A/turkey/Ontario/6118/1968, H12N5 A/duck/Alberta/60/1976, H14N5 A/mallard/Gurjev/263/1982, H4N6 A/duck/Czechoslovakia/1956, H13N6 A/gull/Maryland/704/1977, H5N7 A/duck/Denmark/64650/2003, H7N7 A/turkey/England/647/1977, H10N7 A/chicken/Germany/N/1949, H6N8 A/turkey/Canada/ 1963, H7N9 A/mallard/France/090034d/2009, H7N9 A/mallard/France/110007/2011, H9N9 A/turkey/France/03295/2003, H11N9 A/duck/Memphis/546/1974 and H15N9 A/shearwater/WA/2576/1979. b Each monospecific duck AI serum (N5 to N8) was prepared by prime infection of 3 week-old SPF ducks followed by boosting with the same wild virus 2 weeks later. Influenzavirus used were (in order): H14N5 A/mallard/Gurjev/263/1982, H4N6 A/sarcelle/France/05104/JD7/2005, H10N7 A/chicken/Germany/N/1949 and H6N8 A/duck/France/05063/2005. c Ability of the method to detect antibodies induced by several different AIV strains belonging to the N9 subtype.
IP > 30% – obtained by adding three standard deviations to the negative mean value. Using 751 sera, relative sensitivity and specificity of the N9specific ELISA were evaluated using NP ELISA as the theoretical “gold standard”. Similarly, relative sensitivity and specificity of NI assay versus NP ELISA were determined using 230 sera as mentioned in Tables 1 and 4. In addition, the N9 ELISA and NI assay were compared to NP ELISA and HI test using these 177 common sera. The methods were also compared based on their Receiver Operating Characteristic (ROC) curve, corresponding to plots of the sensitivity versus ‘1 − specificity’ as the discrimination threshold varied from 0 to 1 (Bamber, 1975). The predictions are better when the ROC curve is furthest from the diagonal (or delineates the highest area with the diagonal as defined by the area under the ROC curve (AUC)). The AUC was calculated and was expressed as an overall statistic of diagnostic accuracy. The ROC curve was also used to identify suitable cutoff scores and improved the sensitivity and/or the specificity of the model, which was determined using the Younden index. This was the point on the ROC curve where the sum of sensitivity and specificity was the highest. ROC analyses were performed using the XLStat software add-in for Microsoft Excel (http://www.xlstat.com/en/products-solutions.html). Cohen’s kappa value was also calculated as described by Cohen (1960). It expressed the agreement between two compared methods: values 0–0.20 correspond to no agreement, 0.21–0.40 to slight agreement, 0.41–0.60 to moderate agreement, 0.61–0.80 to substantial agreement and ≥0.81 to almost perfect agreement. The fidelity of this evaluated ELISA was checked using positive control sera. Fidelity, including repeatability and intermediate fidelity, was calculated as mentioned by Feinberg (2010). Repeatability corresponds to the agreement between results obtained following analysis carried out the same day on the same sample
by the same person with the same material. Intermediate fidelity corresponds to the agreement between results from one sample tested several times with at least two different operators, or in two different places, or at two different times (AFNOR, 1994). So, the repeatability of the N9 ELISA was obtained using control serum present at least twice per plate, and intermediate fidelity was calculated using control serum results performed on different days, with different operators. The result was expressed as a percentage. Applying the threshold calculated by subtracting three standard deviations from the mean value (positive with CP ≤ 70% and negative with CP > 70%) of the 521 SPF and AI-negative conventional sera, exclusive and inclusive specificity was satisfactory: all 97 Table 2 Ability to detect the smallest antibody titre of N9-based ELISA in comparison with other serological tests. Last positive dilution with
Serum 1 (chicken) Serum 2 (chicken) Serum 3 (duck) Serum 4 (duck)
N9 ELISA
HI test
NP ELISA
64 128 16 16
128 128 8 2
128 64 16 16
NI 16 16 4 2
Positive sera from immunization n or infection have been gradually diluted into SPF serum from the same species and dilutions used were: 1/1, 1/2, 1/4, 1/8, 1/16, 1/32, 1/64, 1/128, 1/256, 1/512, 1/1024 and 1/2048. These dilutions were tested with N9 ELISA, NP ELISA, HI test – using homologous antigen – and NI. 1: monospecific reference chicken serum against H9N9 Serum A/turkey/France/03295/2003. 2: monospecific reference chicken serum against H11N9 Serum A/duck/Memphis/546/1974. Serum 3 and 4: positive duck sera experimentally infected with H9N9 A/turkey/ France/03295/2003.
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
9
Fig. 1. Exclusive and inclusive specificity of N9 ELISA using sera with a known AI N9 negative or positive status. The details of preparation and their respective number are mentioned in Table 1. If several sera corresponding to a same pathogen or same subtype were tested, only the mean and corresponding standard deviation appear. N9 specificity was checked using chicken hyperimmune sera against five different AIV strains (H7N9 A/mallard/France/090034d/2009, H7N9 A/mallard/France/110007/2011, H9N9 A/turkey/France/03295/2003, H11N9 A/duck/Memphis/546/1974 and H15N9 A/shearwater/WA/2576/1979) (n = 5), sera collected during experimental vaccination containing H5N9 A/chicken/Italy/22A/1998 (n = 35) and sera collected at the peak of humoral response using H9N9 LP A/turkey/France/03295/2003 (n = 30), H7N9 LP A/teal/France/110007d/2011 (n = 7) and H7N9 LP A/Anhui/01/2013 (n = 20). Exclusive specificity corresponds to the ability of the method to conclude negative all avian viruses other than AIV and all avian AIV not belonging to the N9 subtype. Inclusive specificity corresponds to the ability of the method to detect antibodies induced by several different AIV strains belonging to the N9 subtype.
N9-specific sera were positive and all 57 non-N9 sera were negative (Fig. 1). Preliminary assessment of analytical sensitivity, based on the detection of the smallest antibody quantity in four sera diluted serially (Table 2) showed that the N9 ELISA presented a sensitiv-
ity equivalent to that of the NP ELISA. HI presented slightly better sensitivity than N9-based ELISA when considering chicken sera, and conversely N9 ELISA presented better results when considering only duck sera. The analytical sensitivity obtained with the NI assay was lower regardless of the sera studied. The ability of the
Table 3 Comparative ability of several methods to detect an early antibody response following infection (N9 ELISA versus other serological tests). Method
Time post-infection W + 0.5
W+1
W+2
W+3
W+4
W+5
W+6
(a) Infection with French H9N9 0/10 N9 ELISA 0/10 HI test 0/10 NP ELISA 1/10 NI
D0
0/10 0/10 0/10 1/10
9/10 9/10 10/10 4/10
8/10 10/10 9/10 0/10
10/10 10/10 10/10 2/10
10/10 10/10 10/10 0/10
10/10 10/10 10/10 3/10
10/10 10/10 10/10 9/10
(b) Immunization with French LP H7N9 nd N9 ELISA nd HI test nd NP ELISA nd NI
nd nd nd nd
nd nd nd nd
5/7 6/7 7/7 0/7
6/7 7/7 7/7 1/7
nd nd nd nd
nd nd nd nd
nd nd nd nd
(c) Immunization with Chinese LP H7N9 N9 ELISA nd HI test nd nd NP ELISA nd NI
nd nd nd nd
nd nd nd nd
3/10 7/10 7/10 0/10
10/10 10/10 10/10 5/10
10/10 10/10 10/10 7/10
nd nd nd nd
nd nd nd nd
Number of positive sera detected with N9 ELISA, HI test using homologous antigen, NP ELISA or NI/total number of tested sera following: (a) infection with French H9N9 A/turkey/France/03295/2003 virus: infection of ten 10-weeks-old ducks or (b) immunization with inactivated French LP H7N9 A/teal/France/110007d/2011virus: immunization of seven 6-weeks-old chicken or (c) immunization with inactivated Chinese LP H7N9 A/Anhui/01/2013 virus: immunization of seven 6-weeks-old chicken.
10
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
Table 4 Characteristics of the N9 ELISA and NI methods compared with NP ELISA and HI test. Methods compared
N9 ELISA versus NP ELISA
N9 ELISA versus NP ELISA
N9 ELISA versus HI test
NI versus NP ELISA
NI versus NP ELISA
NI versus HI test
No. of sera used for calculation Relative sensitivity 95% CI Relative specificity 95% CI AUC (ROC curve) Younden index–optimal threshold (ROC curve) Kappa
751 88.9% 83.9–93.9 98.7% 97.7–99.6 0.995 81.0%
177 90.8% 85.8–95.7 93.6% 82.5–98.7 0.982 67.3%
177 91.3% 86.3–96.2 88.2% 76.1–95.6 0.963 68.9%
230 33.3% 25.9–40.8 94.8% 87.2–98.6 0.709 −2.2%
177 39.2% 30.8–47.6 91.5% 79.6–97.6 0.803 −2.2%
177 40.5% 31.9–49.0 92.2% 81.1–97.8 0.800 2.3%
0.895
0.795
0.773
0.214
0.204
0.231
N9 ELISA to detect an early antibody response was evaluated using sera kinetics following experimental infections of SPF ducks with a French H9N9 strain or experimental immunization of chickens with inactivated recent French or Chinese H7N9 strains (Table 3). As for the NP ELISA and HI tests, the first antibodies were detected with the N9 ELISA 1 week after infection (W + 1) with the H9N9 virus and 2 weeks after immunization with the French and Chinese inactivated H7N9 antigen (Table 3). Results obtained with the NI assay indicated a lower sensitivity and a lack of consistency from 1 week to the next in comparison with the three other tests and a problem of specificity with one positive result before infection (Table 3). Finally relative specificity and sensitivity of both the N9 ELISA and NI assay were also evaluated by comparing both methods with either the NP ELISA or HI test (Table 4). The N9 ELISA’s relative sensitivity and specificity estimates respectively 88.9–91.3% and 88.2–98.7% depending on the number of sera and “reference” method used, were satisfactory and in agreement with the high AUC values – estimate between 0.963 and 0.995 – determined with the ROC curve (Fig. 2). The N9 ELISA threshold was estimated using the Younden index between 67.3% and 81.0% and was in agreement with a threshold previously calculated at 70%. Kappa
calculation also found substantial or almost perfect agreement (values between 0.773 and 0.895). Conversely, the NI assay presented low relative sensitivity estimate between 33.3% and 40.5% depending on the number of sera and “reference” method used, an inconsistent threshold lower than 0% and kappa values defined as slight agreement (values between 0.204 and 0.231). This lower sensitivity generally hampered an early and/or consistent detection of N9 antibodies. Positive control sera were used to assess the fidelity of the N9 ELISA method, defined by repeatability and intermediate fidelity. Thirty-one plates were tested, containing 64 deposits of control serum, and were obtained over 11 days of analysis, carried out by four different users. The fidelity calculated here was less than 2%, equivalent to a low value and to a very good result. Sandbulte et al. (2009) describe a miniaturized NI assay and emphasize the interest of this method to evaluate human vaccination by measuring NA-specific antibodies. The method used is a miniaturized and optimized format of the conventional NI assay described in the WHO (2002) Manual on Animal Influenza Diagnosis. The NI method thus developed is close to the one used in this study, except the end revelation (using extraction of chromophore to the organic layer before absorbance measuring) and
Fig. 2. ROC curve – different couples evaluated according to the compared methods and the number of data used. In a Receiver Operating Characteristic (ROC) curve the true positive rate (sensitivity) is plotted in function of the false positive rate (100 − specificity) for different cut-off points. Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a particular decision threshold. A test with perfect discrimination (no overlap in the two distributions) has a ROC curve that passes through the upper left corner (100% sensitivity, 100% specificity). Therefore the closer the ROC curve is to the upper left corner, the higher the overall accuracy of the test.
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
serial dilutions of tested sera (for Sandbulte’s method) which allow determination of NA titre. Sandbulte’s paper also describes the potential interference of HA-specific antibodies with substrate cleavage by NA and importance to use a subtype HA in the general absence of human immunity. A sensitivity similar to that achieved with the standard test is described, but results are not confronted with another method. A study presented by Cattoli et al. (2003) demonstrated the simultaneous appearance of anti-HA and NA antibodies respectively detected with the HI test and indirect immunofluorescence assay (sera from vaccinated or naturally infected birds). The results obtained in this study with the kinetics of infection were consistent with these findings as in most cases both HA and NA antibodies were detected at the same date post-infection/immunization. Importantly in the current study, almost all sera identified as positive by NP ELISA (coming from infection or immunization with N9 strain) were detected as positive by N9 ELISA in which a rare and old inactivated whole virus antigen was used. One problem of using such older rarer strains could be that antibodies produced in response to infections with a more current strain will no longer recognize the epitopes of the older virus. However, in the current study the epitope of N9 strains was shown to have remained highly conserved. The positive results obtained with sera induced following infection or immunization with both old and recent American and Eurasian strains also indicated that the monoclonal antibody used in the present study indeed recognized a conserved N9 epitope; however, it would be interesting to complete those data with sera specific to American lineages. Finally and in support of this N9 epitope conservation, antibodies specific to H7N9 virus circulating in China since 2013 have since been detected with this N9-ELISA, but the study was limited to chicken sera. Presently only analytical sensitivity data are shown: indeed, apart from the 512 sera collected from AI negative conventional ducks used to define the threshold of the test (Table 1), all the other sera were collected from experimental chicken and ducks and were so considered of “laboratory grade”. Unfortunately, no field sera collected from avian influenza N9 naturally-infected or immunized poultry were available to assess diagnostic sensitivity. The results obtained show satisfactory ability to detect N9-specific antibodies with satisfactory specificity and with a much higher sensitivity than the recommended NI assay. These results are similar to those obtained in the study of Moreno et al. (2009) with presentation of development, validation and comparison of N1-, N2- and N3-ELISA with NI assays: those data also provide evidence of a higher sensitivity of NA-ELISA tested compared to NI assay. This ELISA method also presents many advantages such as the use of non-toxic reagents, and the possibility of automatisation, and can be performed easily and quickly. This method can be used for serological surveys aimed at detecting infection of poultry with AIV belonging to the N9 subtype including the recent H7N9 virus circulating in China since spring 2013 or as a DIVA test to differentiate poultry vaccinated with AIV vaccine. References AFNOR, 1994. Application de la statistique- Exactitude (justesse et fidélité) des résultats et méthodes de mesure – Partie 1: principes généraux et définitions. AFNOR, Paris. Alexander, D.J., 2000. A review of avian influenza in different bird species. Vet. Microbiol. 74, 3–13.
11
Avellaneda, G., Sylte, M.J., Lee, C.W., Suarez, D.L., 2009. A heterologous neuraminidase subtype strategy for the differentiation of infected and vaccinated animals (DIVA) for avian influenza virus using an alternative neuraminidase inhibition test. Avian Dis. 54, 272–277. Aymard-Henry, M., Coleman, M.T., Dowdle, W.R., Laver, W.G., Schild, G.C., Webster, R.G., 1973. Influenzavirus neuraminidase and neuraminidase-inhibition test procedures. Bull. World Health Organ. 48, 199–202. Bamber, D., 1975. The area above the ordinal dominance graph and the area below the receiver operating characteristic graph. J. Math. Psychol. 12, 387–415. Brown, J.D., Stallknecht, D.E., Berghaus, R.D., Luttrell, M.P., Velek, K., Kistler, W., Costa, T., Yabsley, M.J., Swayne, D., 2009. Evaluation of a commercial blocking enzyme-linked immunosorbent assay to detect avian influenza virus antibodies in multiple experimentally infected avian species. Clin. Vaccine Immunol. 16, 824–829. Cattoli, G., Terregino, C., Brasola, V., Rodriguez, J.F., Capua, I., 2003. Development and preliminary validation of an ad hoc N1–N3 discriminatory test for the control of avian influenza in Italy. Avian Dis. 47, 1060–1062. Chen, Y., Liang, W., Yang, S., Wu, N., Gao, H., Sheng, J., Yao, H., Wo, J., Fang, Q., Cui, D., Li, Y., Yao, X., Zhang, Y., Wu, H., Zheng, S., Diao, H., Xia, S., Zhang, Y., Chan, K.H., Tsoi, H.W., Teng, J.L., Song, W., Wang, P., Lau, S.Y., Zheng, M., Chan, J.F., To, K.K., Chen, H., Li, L., Yuen, K.Y., 2013. Human infections with the emerging avian influenza A H7N9 virus from wet market poultry: clinical analysis and characterisation of viral genome. Lancet 381, 1916–1925. Cohen, J., 1960. A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 20, 37–46. Colman, P.M., 1994. Influenza virus neuraminidase: structure, antibodies, and inhibitors. Protein Sci. 3, 1687–1696. Feinberg, M., 2010. Interprétation du profil d’exactitude in Validation des méthodes d’analyse quantitative par le profil d’exactitude. Cahier des Techniques de l’INRA, numéro special., pp. 45–60. Fouchier, R., Munster, V.J., Keawcharoen, J., Osterhaus, A., Kuiken, T., 2007. Virology of avian influenza in relation to wild birds. J. Wildl. Dis. 43, S7–S14. Gao, R., Cao, B., Hu, Y., Feng, Z., Wang, D., Hu, W., Chen, J., Jie, Z., Qiu, H., Xu, K., Xu, X., Lu, H., Zhu, W., Gao, Z., Xiang, N., Shen, Y., He, Z., Gu, Y., Zhang, Z., Yang, Y., Zhao, X., Zhou, L., Li, X., Zou, S., Zhang, Y., Li, X., Yang, L., Guo, J., Dong, J., Li, Q., Dong, L., Zhu, Y., Bai, T., Wang, S., Hao, P., Yang, W., Zhang, Y., Han, J., Yu, H., Li, D., Gao, G.F., Wu, G., Wang, Y., Yuan, Z., Shu, Y., 2013. Human infection with a novel avian-origin influenza A (H7N9) virus. N. Engl. J. 368, 1888–1897. Li, Q., Zhou, L., Zhou, M., Chen, Z., Li, F., Wu, H., Xiang, N., Chen, E., Tang, F., Wang, D., Meng, L., Hong, Z., Tu, W., Cao, Y., Li, L., Ding, F., Liu, B., Wang, M., Xie, R., Gao, R., Li, X., Bai, T., Zou, S., He, J., Hu, J., Xu, Y., Chai, C., Wang, S., Gao, Y., Jin, L., Zhang, Y., Luo, H., Yu, H., Gao, L., Pang, X., Liu, G., Shu, Y., Yang, W., Uyeki, T.M., Wang, Y., Wu, F., Feng, Z., 2014. Epidemiology of human infections with avian influenza A (H7N9) virus in China. N. Engl. J. Med. 370, 520–532. Matrosovich, M.N., Matrosovich, T.Y., Gray, T., Roberts, N.A., Klenk, H.D., 2004. Neuraminidase is important for the initiation of influenza virus infection in human airway epithelium. J. Virol. 78, 12665–12667. Moreno, A., Brocchi, E., Lelli, D., Gamba, D., Tranquillo, M., Cordiolo, P., 2009. Monoclonal antibody based ELISA tests to detect antibodies against neuraminidase subtypes 1, 2 and 3 of avian influenza viruses in avian sera. Vaccine 27, 4967–4974. Sandbulte, M.R., Gao, J., Straight, T.M., Eichelberger, M.C., 2009. A miniaturized assay for influenza neuraminidase inhibiting antibodies utilizing reverse geneticsderived antigens. Influenza Other Respir. Viruses 3, 233–240. Schmitz, A., Le Bras, M.O., Guillemoto, C., Pierre, I., Rose, N., Bougeard, S., Jestin, V., 2013. Evaluation of a commercial ELISA for H5 low pathogenic avian influenza virus antibody detection in duck sera using Bayesian methods. J. Virol. Methods 193, 197–204. Su, B., Wurtzer, S., Rameix-Welti, M.A., Dwyer, D., Van der Werf, S., Naffakh, N., Clavel, F., Labrosse, B., 2009. Enhancement of the influenza A hemagglutinin (HA)-mediated cell-cell fusion and virus entry by the viral neuraminidase (NA). PLoS One 4, e8495. Van Deusen, R.A., Hinshaw, V.S., Senne, D.A., Pellacani, D., 1983. Micro neuraminidase-inhibition assay for classification of influenza A virus neuraminidases. Avian Dis. 27, 745–750. WHO Manual on Animal Influenza, 2002. Diagnosis and Surveillance. http:// www.who.int/csr/resources/publications/influenza/whocdscsrncs20025.pdf WHO Global Influenza – Surveillance Network, 2011. Manual for the laboratory diagnosis and virological surveillance of influenza. http://whqlibdoc. who.int/publications/2011/9789241548090 eng.pdf World Organization for Animal Health (OIE), 2012. Avian influenza. In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals. World Organization for Animal Health, Paris, France, pp. 1–19 (Chapter 2.3.4), http://www. oie.int/fileadmin/Home/eng/Health standards/tahm/2.03.04 AI.pdf World Organization for Animal Health (OIE), 2013. Avian influenza. In: Terrestrial Animal Health Code. http://www.oie.int/fileadmin/Home/fr/Health standards/tahc/current/chapitre 1.10.4.pdf
Contents lists available at ScienceDirect
Journal of Virological Methods journal homepage: www.elsevier.com/locate/jviromet
Short communication
Development and evaluation of an N9-specific enzyme-linked immunosorbent assay to detect antibodies in duck and chicken sera Audrey Schmitz ∗ , Marie-Odile Le Bras, Katell Louboutin, Véronique Jestin French Agency for Food, Environmental and Occupational Health & Safety, Ploufragan/Plouzané Laboratory, Avian and Rabbit Virology, Immunology and Parasitology Unit, VIPAC, French National Reference Laboratory for Avian Influenza and Newcastle disease, B.P.53, 22440 Ploufragan, France
a b s t r a c t Article history: Received 30 May 2014 Received in revised form 14 October 2014 Accepted 21 October 2014 Available online 18 November 2014 Keywords: Influenza N9 antibodies ELISA Duck Chicken
A serological test for detecting N9-specific antibodies may be useful as a DIVA strategy to differentiate vaccinated from infected animals or simply for direct serological detection of infection with N9-subtype virus. The method currently recommended for the detection of antibodies against neuraminidase is neuraminidase inhibition (NI), which is a laborious method using toxic chemicals and has low sensitivity. The present study describes the development and validation of an N9-specific ELISA. Data obtained with this N9 ELISA were compared to those obtained with nucleoprotein-based ELISA, haemagglutination inhibition test using homologous antigen and NI assay. 785 sera from ducks and chickens were used, from flocks previously determined to be AI negative or from experimentally infected or immunized flocks. Sensitivity and specificity were evaluated, and a ROC curve and kappa values, which provide a comparison between methods, were calculated. The results obtained in this study indicate that the N9 based-ELISA is effective in detecting N9-specific antibodies with high specificity and with better sensitivity than the recommended NI method; using data from 177 common sera tested with N9 ELISA and NI assay both compared to NP-based ELISA, their specificity were evaluated at 93.6% and 91.5% respectively, and sensitivity at 90.8% and 39.2% respectively. © 2014 Elsevier B.V. All rights reserved.
Avian Influenza A viruses (AIV) belong to the Orthomyxoviridae family and subtypes have been identified in AIV based on the antigenic properties of the 16 haemagglutinin (HA) (H1 to 16) and 9 neuraminidase (NA) (N1 to 9) surface glycoproteins. Each subtype is then defined by a combination of one haemagglutinin (H1 to 16) and one neuraminidase (N1 to 9) (Alexander, 2000; Fouchier et al., 2007). These two proteins are the major surface integral membrane glycoproteins of influenza viruses: HA binds to sialic acid and NA catalyzes the removal of terminal sialic acid residues from oligosaccharide chains (Colman, 1994). NA may be important in virulence and the early stage of infection by promoting virus entry or in facilitating release of progeny virions from the surface of infected cells (Matrosovich et al., 2004; Su et al., 2009). Generic tests based on the detection of antibodies against the conserved AIV nucleoprotein (NP) are routinely used as screening tests (Brown et al., 2009). Further to these preliminary tests, haemagglutination inhibition (HI) tests allowing the detection of specific HA subtype antibodies are used as the gold standard either for diagnostic purpose to detect infection by a given HA subtype
∗ Corresponding author. Tel.: +33 2 96 01 64 11; fax: +33 2 96 01 62 63. E-mail address: [email protected] (A. Schmitz). http://dx.doi.org/10.1016/j.jviromet.2014.10.016 0166-0934/© 2014 Elsevier B.V. All rights reserved.
or for assessment of humoral immunity following vaccination using a given HA subtype; although a recent paper demonstrated that for sera from ducks infected or immunized with LP AIV belonging to the H5 subtype of the Eurasian lineage, an H5-based ELISA was more sensitive (Schmitz et al., 2013). In addition to HA identification, a serological test for detecting N9-specific antibodies may be useful in two ways: (i) it could be simply used for direct serological detection of infection in poultry, other captive birds or wild birds with N9-subtype AIV or (ii) in a country which considers AI vaccination to prevent/control AI outbreaks in poultry, a N9-based ELISA test could be used for DIVA strategy to differentiate vaccinated from infected birds either by using a vaccine strain with a N9 subtype different from the NA subtype displayed in the AIV circulating in poultry or the reverse, i.e. by using a vaccine strain with a NA subtype different from the N9 subtype displayed in the AIV circulating in poultry (Avellaneda et al., 2009; World Organization for Animal Health, 2013). N9 viruses are indeed commonly present in both wild birds and poultry, in association with many HA subtypes. The most recent influenza cases concern human and poultry infections with an H7N9 LPAIV which emerged in eastern continental China during the winter of 2013 (Chen et al., 2013; Gao et al., 2013; Li et al., 2014). According to the most recent OMS data available dated
6
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
from the end of June, the number of cases was 450, by the end of summer 2014 the total number of human cases seem to exceed 450. The cases were reported in (i) 15 provinces including the autonomous province of Xingjiang and the capital of Mainland China – including 11 provinces also reporting outbreaks in poultry – and (ii) Hong Kong SAR, Macao and Taiwan and (iii) Malaysia (the cases in the latter four having been contracted in one of the former provinces) (http://www.chp.gov.hk/en/content/599/34480.html; http://www.cidrap.umn.edu/sites/default/files/public/downloads/ topics/cidrap h7n9 update 080414.pdf; http://www.who.int/ influenza/human animal interface/influenza h7n9/riskassessment h7n9 27june14.pdf?ua=1). By the end of June the overall mortality was around 37% (http://www.who.int/influenza/human animal interface/influenza h7n9/riskassessment h7n9 27june14.pdf?ua =1). Forty three outbreaks of low pathogenic avian influenza (H7N9) have been officially notified involving pigeons, chickens and ducks present essentially in live bird markets and/or environment samples collected in the same area, located in twelve provinces of eastern, southern China, (http://www.oie.int/wahis 2/public/ wahid.php/Reviewreport/Review/viewsummary?reportid=13225). Natural infections with this H7N9 virus in poultry are asymptomatic or induce few signs that are difficult to detect; however they may induce a humoral response that can be detected serologically. Domestic ducks play a major epidemiological role in AI, since they are often associated with higher prevalence of subclinical AI infection even in some case of infection with AIV that are highly pathogenic in gallinaceous poultry. In addition, China produces about 5.5 billion chickens for consummation a year and 830 million ducks for consummation a year (number of heads) (FAO data: http://faostat3.fao.org/faostat-gateway/go/to/browse/Q/QA/E), the latter corresponding to 70% of the world’s population of ducks, often as free-range productions which are therefore in frequent contact with wild birds. Improved detection of H7N9 AIV infection in domestic anseriformes would therefore be beneficial in preventing the spread of those viruses to terrestrial poultry. The neuraminidase inhibition (NI) assay is the miniaturized test used for detection of NA subtype specific antibodies (Van Deusen et al., 1983; WHO Global Influenza, 2011). The NI assay is a laborious technique and requires handling toxic chemicals. Its substitution by a more sensitive and specific ELISA test would be useful. The aim of the present paper is to describe the development of a blocking N9based ELISA test and its validation with a panel of approximately 800 sera essentially from ducks and chicken. The present in-house blocking ELISA test was developed using a N9-specific mouse monoclonal antibody (reference IdVet 6B1 A12) produced by IdVet, Montpellier, France. The concentrations of the reagents used, such as coating antigen and monoclonal antibody, as well as the serum to be tested, were optimized. The first step of this ELISA is to coat ELISA plates with a N9 antigen. When a whole AIV is used as antigen and not a recombinant N9 protein, the choice of the antigen for coating is important because HA subtype-specific antibodies can block the access of specific NA (N9 presently) antibodies. This happens because HA subtype-specific antibody binding hampers the access of NA (N9 presently)-specific antibodies to NA (N9 presently)-epitopes. Using a very rare subtype such H15 – that has not been reported (to date) in poultry – should reduce the chances of these unwanted HA antibody/antigen interactions when using whole virus as antigen for a NA specific ELISA. For this reason, the N9 ELISA reported in this study was developed using an uncommon subtype H15N9 A/shearwater/WestAustralia/2576/1979. This virus was grown in 9-day-old embryonated chicken eggs, inactivated with formol at 1‰ overnight at 37 ◦ C, purified with 20–50% gradient sucrose and concentrated 17.5×. Microplates were then coated with the inactivated and purified H15N9 antigen diluted to 1/600 in 0.05 M carbonate buffer (pH 9.6) and incubated at 4 ◦ C
overnight. After washing with PBS (pH 7.4) and 0.05% Tween 20, plates were blocked with PBS-Tween 20 containing 5% milk and 10% foetal bovine serum for 1 h at 37 ◦ C. After washing, serum samples and controls (both previously diluted to 1/5) were added for 1 h at 37 ◦ C, followed after washing by incubating N9-specific monoclonal antibodies diluted to 1/600 for 1 h at 37 ◦ C. The unbound monoclonal antibodies were removed by washing and then phosphatase labelled goat anti-mouse IgG (H + L) (KPL) diluted to 1/400 was added and incubated for 1 h at 37 ◦ C. After washing, the presence of secondary-conjugated-antibody was revealed by incubating the p-nitrophenyl phosphate substrate (pNPP) (Uptima) for 30 min at room temperature. The intensity of the colour measured as the optical density (OD) at 405 nm with a spectrophotometer (Dynex Technologies–MRX Revelation), was correlated inversely to the amount of N9-specific antibodies present in the sample to be tested. Results were expressed as a competition percentage (CP) corresponding to the ratio of the sample analyzed to the negative control [S/N%] = 100 × (OD Sample)/(OD Negative control)). A total of 785 sera were tested. Details of their origin and numbers are provided in Table 1. Status of 521 sera with a known negative AIV were used for establishing the threshold, then 57 reference sera against different AIV subtypes or against pathogens other than AI produced in SPF chickens and ducks were used for checking preliminary specificity. The remaining 207 sera, collected from SPF chickens and N9 ducks experimentally infected or immunized with AIV belonging to the N9 subtype, were used for validation of this N9 ELISA. Most of them were collected from ducks (85.2%), mainly from muscovy ducks. SPF animals were maintained in air-filtered facilities and monitored on a regular basis for ∼20 bacterial and viral avian pathogens including avian influenza virus, to check for their negative status. All experimental immunisations were performed in biological safety level BSL2+ (intermediate between BSL level 2 and 3 – this level requires complete change of clothes, double door access, sealed windows, negative airflow, shower before exit) or BSL3 facilities, approved by the French veterinary authority (entitled officer from the French Ministry of Agriculture) according to EU standards; animal experimentation was performed by duly authorized personnel following reception of approval from the Ethical Committee for experimentation, Cometh ANSES/ENVA/UPEC (National Ethical Committee for Animal Experimentation registration no. 16). All 785 sera were tested with the NP ELISA using a commercial NP-based ELISA test (IDEXX – Elisa IA Multiscreen, recently re-named Influenza A Ab Test). 34 positive sera by NP-ELISA collected from infection/immunization with AIV not belonging to the N9 subtype were removed from the comparative study. Therefore 751 were used for comparison between NP-ELISA and N9-ELISA (Table 1). Sera from experimental infection or immunization (n = 177) with AIV belonging to the N9 subtype were also tested with the HI test (using the same antigen inoculated for immunization or the homologous virus, although inactivated, as the one used for infection) and NI assay. The HI test was performed according to the World Organization for Animal Health manual (World Organization for Animal Health, 2012), using four haemagglutinating units of homologous antigen; HI titres ≥16 (or log2 titre ≥ 4) were scored positive. NI assay was performed as described by Aymard-Henry et al. (1973), except the method was miniaturized and the results were given as OD (Van Deusen et al., 1983). The antigen used was the same as the one used with N9-based ELISA, i.e. the uncommon subtype H15N9 A/shearwater/WestAustralia/2576/1979: as well as N9-ELISA, again as previously described using a very rare subtype such H15 should prevent unspecific reactions. The threshold was determined beforehand using an Inhibition Percentage (IP) (corresponding to [(N − S)/N%] = 100 × (OD negative control − OD sample)/(OD negative control) equivalent to 30% – positive with
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
7
Table 1 Listing of sera used for the development of N9 ELISA. Categories
Species
Age at the moment of sampling
Number of sera tested with N9-based ELISA test
Historical data
Purpose in the present study
SPF ducks
Muscovy duck
19–79 weeks
101
Determination of threshold
Conventional ducks with AI negative status
Muscovy duck
7–49 weeks
420
Reference sera against other avian pathogens than AIV
Muscovy duck and chicken
9–11 weeks
16
Reference SPF pooled sera
Muscovy duck and chicken
Reference sera AIV positive (not N9)
Chicken
8 weeks 2 (duck) 78 weeks (chicken) 9–11 weeks 26
SPF, notably negative AI status as monitored periodically using NP-based ELISA test Negative AI status as tested with NP-based ELISA Experimental immunization with either avian viral pathogen different from AI as mentioned in Fig. 1 Pooled sera of SPF chicken and ducks
Reference sera N9 positive
Chicken
Experimental N5-N6-N7-N8 positive sera
Muscovy duck
6 weeks
8
Experimental N9 sera
Chicken
9–11 weeks
2 × 12
Muscovy duck
16 weeks
2 × 12
Experimental N9 sera
Muscovy duck
10–16 weeks
80
Experimental N9 sera
Chicken
8–10 weeks
3 × 10
Experimental N9 sera
Chicken
8–9 weeks
14
5
Monospecific AIV reference seraa (N1-2-3-4-5-6-78) (cf list of strains in legend) Monospecific AIV reference seraa (N9) (cf list of strains in legend) Experimental infection with inactivated virusb (N5-6-7-8) (cf list of strains in legend) Reference N9 serum from H9N9 LP A/turkey/France/03295/2003 and H11N9 LP A/duck/Memphis/546/1974 serially diluted in SPF pooled chicken sera Serial dilution of 2 positive sera obtained with experimental infection with H9N9 LP A/turkey/France/03295/2003 Experimental infection with wild virus H9N9 LP A/turkey/France/03295/2003 (kinetic) Experimental immunization with inactivated virus H7N9 LP A/Anhui/01/2013 Experimental immunization with inactivated virus H7N9 LP A/teal/France/110007d/2011
n = 521
Set of sera also tested with NP ELISA
√
√
Preliminary evaluation of specificity
n = 57
√
√
Ability to detect n = 207 the smallest antibody titre and inclusive specificityc
Ability to detect the earliest antibody response and inclusive specificityc
Set of sera also tested with HI test using homologous antigen
Set of sera also tested with NI
–
–
–
–
–
–
√
√
√
–
–
√
√
–
–
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
√
8
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
Table 1 (Continued) Categories
Species
Age at the moment of sampling
Number of sera tested with N9-based ELISA test
Historical data
Purpose in the present study
Experimental N9 positive sera
Muscovy duck
11–12 weeks
35
Experimental vaccination with H5N9 vaccine (containing H5N9 A/chicken/Italy/22A/1998)
Inclusive specificityc
Total
Set of sera also tested with NP ELISA
√
785
Set of sera also tested with HI test using homologous antigen
√
–
751
Set of sera also tested with NI
177
264
SPF birds were maintained in air-filtered facilities and monitored on a regular basis against ∼20 bacterial and viral avian pathogens including avian influenza virus, to check for their negative status. All experimental sera were obtained starting with SPF animals, submitted to the schedule mentioned in historical data. AI, Avian influenza; AIV, avian influenza virus; LP, low pathogenic; HP, highly pathogenic. 751 sera tested with both N9-based and NP-based ELISA test. 177 sera tested with both N9-based ELISA, HI test and NI test. 230 sera tested with both NI test and NP-based ELISA. √ Done. – Not done. a Each monospecific reference chicken AI serum (N1 to N9) was prepared by prime immunization of 6-week-old SPF chickens followed by boosting with the same inactivated virus about 2 weeks later; one to three week after the boost, all sera with a positive HI titre of 512-2048 against the specific viral haemagglutinin subtype were pooled. Influenza virus used were (in order): H1N1 A/duck/Alberta/35/1976, H5N1 A/duck/France/05066b/2005, H5N1 A/common pochard/France/06167/2006, H5N1 A/muteswan/France/070203tr/2007, H5N1 A/duck/France/080036/2008, H7N1 A/mallard/France/06159/2006, H7N1 A/africa starling/983/1979, H3N2 A/turkey/England/1969, H5N2 A/chicken/France/03426a/2003, H5N2 A/duck/France/05056a/2005, H5N2 A/duck/France/05057b/2005, H2N3 A/duck/Germany/1215/1973, H5N3 A/duck/France/02166/2002, H5N3 A/duck/France/05054a/2005, H5N3 A/mallard/France/061054/2006, H5N3 A/muscovy duck/France/070090/2007, H7N3 A/turkey/England/1963, H8N4 A/turkey/Ontario/6118/1968, H12N5 A/duck/Alberta/60/1976, H14N5 A/mallard/Gurjev/263/1982, H4N6 A/duck/Czechoslovakia/1956, H13N6 A/gull/Maryland/704/1977, H5N7 A/duck/Denmark/64650/2003, H7N7 A/turkey/England/647/1977, H10N7 A/chicken/Germany/N/1949, H6N8 A/turkey/Canada/ 1963, H7N9 A/mallard/France/090034d/2009, H7N9 A/mallard/France/110007/2011, H9N9 A/turkey/France/03295/2003, H11N9 A/duck/Memphis/546/1974 and H15N9 A/shearwater/WA/2576/1979. b Each monospecific duck AI serum (N5 to N8) was prepared by prime infection of 3 week-old SPF ducks followed by boosting with the same wild virus 2 weeks later. Influenzavirus used were (in order): H14N5 A/mallard/Gurjev/263/1982, H4N6 A/sarcelle/France/05104/JD7/2005, H10N7 A/chicken/Germany/N/1949 and H6N8 A/duck/France/05063/2005. c Ability of the method to detect antibodies induced by several different AIV strains belonging to the N9 subtype.
IP > 30% – obtained by adding three standard deviations to the negative mean value. Using 751 sera, relative sensitivity and specificity of the N9specific ELISA were evaluated using NP ELISA as the theoretical “gold standard”. Similarly, relative sensitivity and specificity of NI assay versus NP ELISA were determined using 230 sera as mentioned in Tables 1 and 4. In addition, the N9 ELISA and NI assay were compared to NP ELISA and HI test using these 177 common sera. The methods were also compared based on their Receiver Operating Characteristic (ROC) curve, corresponding to plots of the sensitivity versus ‘1 − specificity’ as the discrimination threshold varied from 0 to 1 (Bamber, 1975). The predictions are better when the ROC curve is furthest from the diagonal (or delineates the highest area with the diagonal as defined by the area under the ROC curve (AUC)). The AUC was calculated and was expressed as an overall statistic of diagnostic accuracy. The ROC curve was also used to identify suitable cutoff scores and improved the sensitivity and/or the specificity of the model, which was determined using the Younden index. This was the point on the ROC curve where the sum of sensitivity and specificity was the highest. ROC analyses were performed using the XLStat software add-in for Microsoft Excel (http://www.xlstat.com/en/products-solutions.html). Cohen’s kappa value was also calculated as described by Cohen (1960). It expressed the agreement between two compared methods: values 0–0.20 correspond to no agreement, 0.21–0.40 to slight agreement, 0.41–0.60 to moderate agreement, 0.61–0.80 to substantial agreement and ≥0.81 to almost perfect agreement. The fidelity of this evaluated ELISA was checked using positive control sera. Fidelity, including repeatability and intermediate fidelity, was calculated as mentioned by Feinberg (2010). Repeatability corresponds to the agreement between results obtained following analysis carried out the same day on the same sample
by the same person with the same material. Intermediate fidelity corresponds to the agreement between results from one sample tested several times with at least two different operators, or in two different places, or at two different times (AFNOR, 1994). So, the repeatability of the N9 ELISA was obtained using control serum present at least twice per plate, and intermediate fidelity was calculated using control serum results performed on different days, with different operators. The result was expressed as a percentage. Applying the threshold calculated by subtracting three standard deviations from the mean value (positive with CP ≤ 70% and negative with CP > 70%) of the 521 SPF and AI-negative conventional sera, exclusive and inclusive specificity was satisfactory: all 97 Table 2 Ability to detect the smallest antibody titre of N9-based ELISA in comparison with other serological tests. Last positive dilution with
Serum 1 (chicken) Serum 2 (chicken) Serum 3 (duck) Serum 4 (duck)
N9 ELISA
HI test
NP ELISA
64 128 16 16
128 128 8 2
128 64 16 16
NI 16 16 4 2
Positive sera from immunization n or infection have been gradually diluted into SPF serum from the same species and dilutions used were: 1/1, 1/2, 1/4, 1/8, 1/16, 1/32, 1/64, 1/128, 1/256, 1/512, 1/1024 and 1/2048. These dilutions were tested with N9 ELISA, NP ELISA, HI test – using homologous antigen – and NI. 1: monospecific reference chicken serum against H9N9 Serum A/turkey/France/03295/2003. 2: monospecific reference chicken serum against H11N9 Serum A/duck/Memphis/546/1974. Serum 3 and 4: positive duck sera experimentally infected with H9N9 A/turkey/ France/03295/2003.
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
9
Fig. 1. Exclusive and inclusive specificity of N9 ELISA using sera with a known AI N9 negative or positive status. The details of preparation and their respective number are mentioned in Table 1. If several sera corresponding to a same pathogen or same subtype were tested, only the mean and corresponding standard deviation appear. N9 specificity was checked using chicken hyperimmune sera against five different AIV strains (H7N9 A/mallard/France/090034d/2009, H7N9 A/mallard/France/110007/2011, H9N9 A/turkey/France/03295/2003, H11N9 A/duck/Memphis/546/1974 and H15N9 A/shearwater/WA/2576/1979) (n = 5), sera collected during experimental vaccination containing H5N9 A/chicken/Italy/22A/1998 (n = 35) and sera collected at the peak of humoral response using H9N9 LP A/turkey/France/03295/2003 (n = 30), H7N9 LP A/teal/France/110007d/2011 (n = 7) and H7N9 LP A/Anhui/01/2013 (n = 20). Exclusive specificity corresponds to the ability of the method to conclude negative all avian viruses other than AIV and all avian AIV not belonging to the N9 subtype. Inclusive specificity corresponds to the ability of the method to detect antibodies induced by several different AIV strains belonging to the N9 subtype.
N9-specific sera were positive and all 57 non-N9 sera were negative (Fig. 1). Preliminary assessment of analytical sensitivity, based on the detection of the smallest antibody quantity in four sera diluted serially (Table 2) showed that the N9 ELISA presented a sensitiv-
ity equivalent to that of the NP ELISA. HI presented slightly better sensitivity than N9-based ELISA when considering chicken sera, and conversely N9 ELISA presented better results when considering only duck sera. The analytical sensitivity obtained with the NI assay was lower regardless of the sera studied. The ability of the
Table 3 Comparative ability of several methods to detect an early antibody response following infection (N9 ELISA versus other serological tests). Method
Time post-infection W + 0.5
W+1
W+2
W+3
W+4
W+5
W+6
(a) Infection with French H9N9 0/10 N9 ELISA 0/10 HI test 0/10 NP ELISA 1/10 NI
D0
0/10 0/10 0/10 1/10
9/10 9/10 10/10 4/10
8/10 10/10 9/10 0/10
10/10 10/10 10/10 2/10
10/10 10/10 10/10 0/10
10/10 10/10 10/10 3/10
10/10 10/10 10/10 9/10
(b) Immunization with French LP H7N9 nd N9 ELISA nd HI test nd NP ELISA nd NI
nd nd nd nd
nd nd nd nd
5/7 6/7 7/7 0/7
6/7 7/7 7/7 1/7
nd nd nd nd
nd nd nd nd
nd nd nd nd
(c) Immunization with Chinese LP H7N9 N9 ELISA nd HI test nd nd NP ELISA nd NI
nd nd nd nd
nd nd nd nd
3/10 7/10 7/10 0/10
10/10 10/10 10/10 5/10
10/10 10/10 10/10 7/10
nd nd nd nd
nd nd nd nd
Number of positive sera detected with N9 ELISA, HI test using homologous antigen, NP ELISA or NI/total number of tested sera following: (a) infection with French H9N9 A/turkey/France/03295/2003 virus: infection of ten 10-weeks-old ducks or (b) immunization with inactivated French LP H7N9 A/teal/France/110007d/2011virus: immunization of seven 6-weeks-old chicken or (c) immunization with inactivated Chinese LP H7N9 A/Anhui/01/2013 virus: immunization of seven 6-weeks-old chicken.
10
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
Table 4 Characteristics of the N9 ELISA and NI methods compared with NP ELISA and HI test. Methods compared
N9 ELISA versus NP ELISA
N9 ELISA versus NP ELISA
N9 ELISA versus HI test
NI versus NP ELISA
NI versus NP ELISA
NI versus HI test
No. of sera used for calculation Relative sensitivity 95% CI Relative specificity 95% CI AUC (ROC curve) Younden index–optimal threshold (ROC curve) Kappa
751 88.9% 83.9–93.9 98.7% 97.7–99.6 0.995 81.0%
177 90.8% 85.8–95.7 93.6% 82.5–98.7 0.982 67.3%
177 91.3% 86.3–96.2 88.2% 76.1–95.6 0.963 68.9%
230 33.3% 25.9–40.8 94.8% 87.2–98.6 0.709 −2.2%
177 39.2% 30.8–47.6 91.5% 79.6–97.6 0.803 −2.2%
177 40.5% 31.9–49.0 92.2% 81.1–97.8 0.800 2.3%
0.895
0.795
0.773
0.214
0.204
0.231
N9 ELISA to detect an early antibody response was evaluated using sera kinetics following experimental infections of SPF ducks with a French H9N9 strain or experimental immunization of chickens with inactivated recent French or Chinese H7N9 strains (Table 3). As for the NP ELISA and HI tests, the first antibodies were detected with the N9 ELISA 1 week after infection (W + 1) with the H9N9 virus and 2 weeks after immunization with the French and Chinese inactivated H7N9 antigen (Table 3). Results obtained with the NI assay indicated a lower sensitivity and a lack of consistency from 1 week to the next in comparison with the three other tests and a problem of specificity with one positive result before infection (Table 3). Finally relative specificity and sensitivity of both the N9 ELISA and NI assay were also evaluated by comparing both methods with either the NP ELISA or HI test (Table 4). The N9 ELISA’s relative sensitivity and specificity estimates respectively 88.9–91.3% and 88.2–98.7% depending on the number of sera and “reference” method used, were satisfactory and in agreement with the high AUC values – estimate between 0.963 and 0.995 – determined with the ROC curve (Fig. 2). The N9 ELISA threshold was estimated using the Younden index between 67.3% and 81.0% and was in agreement with a threshold previously calculated at 70%. Kappa
calculation also found substantial or almost perfect agreement (values between 0.773 and 0.895). Conversely, the NI assay presented low relative sensitivity estimate between 33.3% and 40.5% depending on the number of sera and “reference” method used, an inconsistent threshold lower than 0% and kappa values defined as slight agreement (values between 0.204 and 0.231). This lower sensitivity generally hampered an early and/or consistent detection of N9 antibodies. Positive control sera were used to assess the fidelity of the N9 ELISA method, defined by repeatability and intermediate fidelity. Thirty-one plates were tested, containing 64 deposits of control serum, and were obtained over 11 days of analysis, carried out by four different users. The fidelity calculated here was less than 2%, equivalent to a low value and to a very good result. Sandbulte et al. (2009) describe a miniaturized NI assay and emphasize the interest of this method to evaluate human vaccination by measuring NA-specific antibodies. The method used is a miniaturized and optimized format of the conventional NI assay described in the WHO (2002) Manual on Animal Influenza Diagnosis. The NI method thus developed is close to the one used in this study, except the end revelation (using extraction of chromophore to the organic layer before absorbance measuring) and
Fig. 2. ROC curve – different couples evaluated according to the compared methods and the number of data used. In a Receiver Operating Characteristic (ROC) curve the true positive rate (sensitivity) is plotted in function of the false positive rate (100 − specificity) for different cut-off points. Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a particular decision threshold. A test with perfect discrimination (no overlap in the two distributions) has a ROC curve that passes through the upper left corner (100% sensitivity, 100% specificity). Therefore the closer the ROC curve is to the upper left corner, the higher the overall accuracy of the test.
A. Schmitz et al. / Journal of Virological Methods 213 (2015) 5–11
serial dilutions of tested sera (for Sandbulte’s method) which allow determination of NA titre. Sandbulte’s paper also describes the potential interference of HA-specific antibodies with substrate cleavage by NA and importance to use a subtype HA in the general absence of human immunity. A sensitivity similar to that achieved with the standard test is described, but results are not confronted with another method. A study presented by Cattoli et al. (2003) demonstrated the simultaneous appearance of anti-HA and NA antibodies respectively detected with the HI test and indirect immunofluorescence assay (sera from vaccinated or naturally infected birds). The results obtained in this study with the kinetics of infection were consistent with these findings as in most cases both HA and NA antibodies were detected at the same date post-infection/immunization. Importantly in the current study, almost all sera identified as positive by NP ELISA (coming from infection or immunization with N9 strain) were detected as positive by N9 ELISA in which a rare and old inactivated whole virus antigen was used. One problem of using such older rarer strains could be that antibodies produced in response to infections with a more current strain will no longer recognize the epitopes of the older virus. However, in the current study the epitope of N9 strains was shown to have remained highly conserved. The positive results obtained with sera induced following infection or immunization with both old and recent American and Eurasian strains also indicated that the monoclonal antibody used in the present study indeed recognized a conserved N9 epitope; however, it would be interesting to complete those data with sera specific to American lineages. Finally and in support of this N9 epitope conservation, antibodies specific to H7N9 virus circulating in China since 2013 have since been detected with this N9-ELISA, but the study was limited to chicken sera. Presently only analytical sensitivity data are shown: indeed, apart from the 512 sera collected from AI negative conventional ducks used to define the threshold of the test (Table 1), all the other sera were collected from experimental chicken and ducks and were so considered of “laboratory grade”. Unfortunately, no field sera collected from avian influenza N9 naturally-infected or immunized poultry were available to assess diagnostic sensitivity. The results obtained show satisfactory ability to detect N9-specific antibodies with satisfactory specificity and with a much higher sensitivity than the recommended NI assay. These results are similar to those obtained in the study of Moreno et al. (2009) with presentation of development, validation and comparison of N1-, N2- and N3-ELISA with NI assays: those data also provide evidence of a higher sensitivity of NA-ELISA tested compared to NI assay. This ELISA method also presents many advantages such as the use of non-toxic reagents, and the possibility of automatisation, and can be performed easily and quickly. This method can be used for serological surveys aimed at detecting infection of poultry with AIV belonging to the N9 subtype including the recent H7N9 virus circulating in China since spring 2013 or as a DIVA test to differentiate poultry vaccinated with AIV vaccine. References AFNOR, 1994. Application de la statistique- Exactitude (justesse et fidélité) des résultats et méthodes de mesure – Partie 1: principes généraux et définitions. AFNOR, Paris. Alexander, D.J., 2000. A review of avian influenza in different bird species. Vet. Microbiol. 74, 3–13.
11
Avellaneda, G., Sylte, M.J., Lee, C.W., Suarez, D.L., 2009. A heterologous neuraminidase subtype strategy for the differentiation of infected and vaccinated animals (DIVA) for avian influenza virus using an alternative neuraminidase inhibition test. Avian Dis. 54, 272–277. Aymard-Henry, M., Coleman, M.T., Dowdle, W.R., Laver, W.G., Schild, G.C., Webster, R.G., 1973. Influenzavirus neuraminidase and neuraminidase-inhibition test procedures. Bull. World Health Organ. 48, 199–202. Bamber, D., 1975. The area above the ordinal dominance graph and the area below the receiver operating characteristic graph. J. Math. Psychol. 12, 387–415. Brown, J.D., Stallknecht, D.E., Berghaus, R.D., Luttrell, M.P., Velek, K., Kistler, W., Costa, T., Yabsley, M.J., Swayne, D., 2009. Evaluation of a commercial blocking enzyme-linked immunosorbent assay to detect avian influenza virus antibodies in multiple experimentally infected avian species. Clin. Vaccine Immunol. 16, 824–829. Cattoli, G., Terregino, C., Brasola, V., Rodriguez, J.F., Capua, I., 2003. Development and preliminary validation of an ad hoc N1–N3 discriminatory test for the control of avian influenza in Italy. Avian Dis. 47, 1060–1062. Chen, Y., Liang, W., Yang, S., Wu, N., Gao, H., Sheng, J., Yao, H., Wo, J., Fang, Q., Cui, D., Li, Y., Yao, X., Zhang, Y., Wu, H., Zheng, S., Diao, H., Xia, S., Zhang, Y., Chan, K.H., Tsoi, H.W., Teng, J.L., Song, W., Wang, P., Lau, S.Y., Zheng, M., Chan, J.F., To, K.K., Chen, H., Li, L., Yuen, K.Y., 2013. Human infections with the emerging avian influenza A H7N9 virus from wet market poultry: clinical analysis and characterisation of viral genome. Lancet 381, 1916–1925. Cohen, J., 1960. A coefficient of agreement for nominal scales. Educ. Psychol. Meas. 20, 37–46. Colman, P.M., 1994. Influenza virus neuraminidase: structure, antibodies, and inhibitors. Protein Sci. 3, 1687–1696. Feinberg, M., 2010. Interprétation du profil d’exactitude in Validation des méthodes d’analyse quantitative par le profil d’exactitude. Cahier des Techniques de l’INRA, numéro special., pp. 45–60. Fouchier, R., Munster, V.J., Keawcharoen, J., Osterhaus, A., Kuiken, T., 2007. Virology of avian influenza in relation to wild birds. J. Wildl. Dis. 43, S7–S14. Gao, R., Cao, B., Hu, Y., Feng, Z., Wang, D., Hu, W., Chen, J., Jie, Z., Qiu, H., Xu, K., Xu, X., Lu, H., Zhu, W., Gao, Z., Xiang, N., Shen, Y., He, Z., Gu, Y., Zhang, Z., Yang, Y., Zhao, X., Zhou, L., Li, X., Zou, S., Zhang, Y., Li, X., Yang, L., Guo, J., Dong, J., Li, Q., Dong, L., Zhu, Y., Bai, T., Wang, S., Hao, P., Yang, W., Zhang, Y., Han, J., Yu, H., Li, D., Gao, G.F., Wu, G., Wang, Y., Yuan, Z., Shu, Y., 2013. Human infection with a novel avian-origin influenza A (H7N9) virus. N. Engl. J. 368, 1888–1897. Li, Q., Zhou, L., Zhou, M., Chen, Z., Li, F., Wu, H., Xiang, N., Chen, E., Tang, F., Wang, D., Meng, L., Hong, Z., Tu, W., Cao, Y., Li, L., Ding, F., Liu, B., Wang, M., Xie, R., Gao, R., Li, X., Bai, T., Zou, S., He, J., Hu, J., Xu, Y., Chai, C., Wang, S., Gao, Y., Jin, L., Zhang, Y., Luo, H., Yu, H., Gao, L., Pang, X., Liu, G., Shu, Y., Yang, W., Uyeki, T.M., Wang, Y., Wu, F., Feng, Z., 2014. Epidemiology of human infections with avian influenza A (H7N9) virus in China. N. Engl. J. Med. 370, 520–532. Matrosovich, M.N., Matrosovich, T.Y., Gray, T., Roberts, N.A., Klenk, H.D., 2004. Neuraminidase is important for the initiation of influenza virus infection in human airway epithelium. J. Virol. 78, 12665–12667. Moreno, A., Brocchi, E., Lelli, D., Gamba, D., Tranquillo, M., Cordiolo, P., 2009. Monoclonal antibody based ELISA tests to detect antibodies against neuraminidase subtypes 1, 2 and 3 of avian influenza viruses in avian sera. Vaccine 27, 4967–4974. Sandbulte, M.R., Gao, J., Straight, T.M., Eichelberger, M.C., 2009. A miniaturized assay for influenza neuraminidase inhibiting antibodies utilizing reverse geneticsderived antigens. Influenza Other Respir. Viruses 3, 233–240. Schmitz, A., Le Bras, M.O., Guillemoto, C., Pierre, I., Rose, N., Bougeard, S., Jestin, V., 2013. Evaluation of a commercial ELISA for H5 low pathogenic avian influenza virus antibody detection in duck sera using Bayesian methods. J. Virol. Methods 193, 197–204. Su, B., Wurtzer, S., Rameix-Welti, M.A., Dwyer, D., Van der Werf, S., Naffakh, N., Clavel, F., Labrosse, B., 2009. Enhancement of the influenza A hemagglutinin (HA)-mediated cell-cell fusion and virus entry by the viral neuraminidase (NA). PLoS One 4, e8495. Van Deusen, R.A., Hinshaw, V.S., Senne, D.A., Pellacani, D., 1983. Micro neuraminidase-inhibition assay for classification of influenza A virus neuraminidases. Avian Dis. 27, 745–750. WHO Manual on Animal Influenza, 2002. Diagnosis and Surveillance. http:// www.who.int/csr/resources/publications/influenza/whocdscsrncs20025.pdf WHO Global Influenza – Surveillance Network, 2011. Manual for the laboratory diagnosis and virological surveillance of influenza. http://whqlibdoc. who.int/publications/2011/9789241548090 eng.pdf World Organization for Animal Health (OIE), 2012. Avian influenza. In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals. World Organization for Animal Health, Paris, France, pp. 1–19 (Chapter 2.3.4), http://www. oie.int/fileadmin/Home/eng/Health standards/tahm/2.03.04 AI.pdf World Organization for Animal Health (OIE), 2013. Avian influenza. In: Terrestrial Animal Health Code. http://www.oie.int/fileadmin/Home/fr/Health standards/tahc/current/chapitre 1.10.4.pdf
