ABSTRACT
Among the different hemagglutinin (HA) subtypes of avian influenza (AI) viruses, H5, H7, and H9 are of major interest because of the serious consequences for the poultry industry and the increasing frequency of direct transmission of these viruses to humans. The availability of new tools to rapidly detect and subtype the influenza viruses can enable the immediate application of measures to prevent the widespread transmission of the infection. In this study, a novel one-step real-time reverse transcription-PCR (RRT-PCR) was developed to detect simultaneously the H5, H7, and H9 subtypes of AI viruses from clinical samples of avian origin. The sensitivity of the RRT-PCR assay was determined by using in vitro-transcribed RNA and 10-fold serial dilutions of titrated AI viruses. High sensitivity levels were obtained, with limits of detection ranging from 101 to 103 RNA copies and from 101 50% egg infectious dose (EID50)/100 μl to 102.74 EID50/100 μl with titrated viruses. Excellent results were achieved in the intra- and interassay variability tests. The comparison of the results with those obtained from the analysis of 725 avian samples by means of the reference method (virus isolation [VI]) showed a high level of agreement. To date, this is the first real-time PCR protocol available for the simultaneous detection of AI viruses belonging to subtypes H5, H7, and H9, and the results obtained indicate that this method is suitable as a routine laboratory test for the rapid detection and differentiation of the three most-important AI virus subtypes in samples of avian origin.
Influenza viruses are enveloped negative-strand RNA viruses belonging to the family of Orthomyxoviridae. Viruses of the Influenzavirus A genus cause avian influenza (AI) when infecting birds. AI is a disease of varying severity but may be of great importance for animal health, with serious implications for the poultry industry and, in some cases, for human health. Subtyping of influenza A viruses is based on antigenic differences between the two surface glycoproteins hemagglutinin (HA) and neuraminidase (NA). To date, 16 HA and 9 NA subtypes of influenza A viruses have been identified, and all of these subtypes have been isolated from avian species (10) in most possible combinations. Influenza A viruses infecting poultry can be divided into two distinct groups on the basis of their ability to cause disease in susceptible birds: low-pathogenicity AI (LPAI) and highly pathogenic AI (HPAI). The virulent viruses (HPAI) are restricted to subtypes H5 and H7, although not all viruses of these subtypes cause HPAI.
Only three HA and two NA subtypes (H1, H2, and H3 and N1 and N2, respectively) of influenza A viruses have become established in the human population, although in recent years AI viruses of subtypes H5, H7, and H9 have caused an increasing number of cases of infection in the human host (6, 16).
AI subtypes H5, H7, and H9 are also of major interest to the poultry industry. The highly pathogenic subtype H5 and H7 viruses have caused several outbreaks with devastating economic consequences (3). Viruses belonging to the H9 subtype are LPAI viruses, but in the last decade, several outbreaks caused by the H9N2 virus have occurred across a wide geographical area, causing serious disease problems in commercial poultry in Iran, Pakistan, and the Middle East (1, 2, 3, 17). In several geographical areas of Eurasia and Africa, these three subtypes have been reported to cocirculate, particularly in areas in which live-bird markets are a common practice and when a superinfection with a distinct subtype occurs in an area in which another subtype is already endemic. This occurrence poses problems for the correct diagnostic interpretation of results in the case of the application of virus isolation (VI) techniques or monovalent PCR assays, in which only the predominant virus may be detected. This can hamper the appropriate management of outbreaks and may result in the application of incomplete intervention strategies.
The significant problems caused for the poultry industry by subtype H5, H7, and H9 viruses and the increased risk of direct transmission of these viruses to humans highlight the need for a highly sensitive, accurate, and rapid test to reveal, as early as possible, the circulation of these viral subtypes in the susceptible avian population.
Conventional AI diagnostic tools (i.e., VI and hemagglutinin inhibition) are time consuming and require facilities not easily available in some affected areas. Because of their rapidity and sensitivity, molecular tests, such as reverse transcription-PCR (RT-PCR) and real-time RT-PCR (RRT-PCR), are being used more and more by medical and veterinary diagnosticians for the diagnosis of AI (4, 24).
Recently, RRT-PCR assays have been developed for the detection of type A influenza viruses (7, 23) and for the specific diagnosis of H5 and H7 viruses (8, 12, 14, 15, 19, 21, 23, 27). To date, the only published RRT-PCR assay designed for subtype H7 was validated for viruses belonging to the American lineage, and no primer and probe sets are currently available for the identification of subtype H9 viruses.
Here we report the development and validation of a sensitive and specific RRT-PCR assay with hydrolysis-type probes to detect simultaneously subtypes H5, H7, and H9 of the AI virus from clinical samples of avian origin.
MATERIALS AND METHODS
Viruses and bacterial strains.Selected avian viruses and bacteria were used to test the specificity and sensitivity of the RRT-PCR assay (Table 1).
Viral and bacterial strains used in this study
To produce viral working stocks for the standardization of the assay, all avian viruses were propagated in the allantoic cavities of 9- to 11-day-old embryonated fowl eggs, whereas avian pneumovirus type A and B isolates were grown and harvested from tissue cultures. Bacterial strains were cultured and propagated using standard methods (25).
The median egg infectious dose (EID50) of each of the AI viruses used in the sensitivity tests was calculated according to the Reed and Muench formula (18).
RNA extraction.Viral RNA was extracted from clinical samples, supernatant of cell culture, and allantoic fluid by using a Qiagen RNeasy mini kit according to the manufacturer's directions (Qiagen, Hilden, Germany). Two-hundred-microliter samples of allantoic fluid or of phosphate-buffered saline (PBS) suspensions of cloacal and tracheal swabs and samples of feces and organs, as described below, were used in the extraction. RNA was eluted in a final volume of 50 μl and stored at −80°C.
Primer and probe set design.Viral subtype H5-, H7-, and H9-specific primer and probe sets for conserved regions in the HA2 subunit of the H5, H7, and H9 HA gene sequences were designed (Table 2). Because of the significant sequence variability of the H5, H7, and H9 genes belonging to viruses isolated in different parts of the world, Eurasian and African H5, H7, and H9 influenza viruses were chosen as the main targets for primer and probe design. Multiple alignments of historical and recent H5, H7, and H9 subtypes were performed to minimize primer and probe mismatches. The alignment was performed using, respectively, 166, 81, and 131 HA nucleotide sequences for subtypes H5, H7, and H9. Primers and probes were designed and optimized to have compatible melting temperatures, enabling them to be used with identical thermal profiles. The hydrolysis probes for the H9 and H5 genes contained 6-carboxyfluorescein as a fluorescent reporter dye at the 5′ end and 6-carboxytetramethylrhodamine as a quencher dye at the 3′ end. The H7 hydrolysis probe was labeled with VIC at the 5′ end, and the 3′-end label was 6-carboxytetramethylrhodamine.
RRT-PCR primer and probe sequences
RRT-PCR.The reagents contained in a QuantiTect multiplex RT-PCR kit (Qiagen, Hilden, Germany) were used for RRT-PCRs. All but one of the primers targeting the HA gene were applied to the PCR at the optimized concentration of 300 nM each. The exception was the H7-specific reverse primer, which was used at a concentration of 900 nM. Specific fluorescently labeled probes were used at a final concentration of 150 nM. The RRT-PCR took place in a final volume of 25 μl using a RotorGene 6000 (Corbett, Australia) apparatus. Each PCR tube contained a single primer/probe set (i.e., for H5 or H7 or H9). The identical thermal profile was adopted in order to detect the distinct subtypes simultaneously and within the same run. The following protocol was used for all primer/probe sets: 20 min at 50°C and 15 min at 95°C, followed by 40 cycles at 94°C for 45 s and 54°C for 45 s.
Analytical specificity and sensitivity of the method.The specificity of the primer/probe sets was tested on nucleic acids extracted from a diverse array of microorganisms that may be naturally present in samples of avian origin (Table 1). Each strain used was tested in triplicate.
In the present study, the term “sensitivity of the method” reflects the efficacy of the entire method applied to recover the target organism in the field specimens, including the RNA extraction procedure and the RRT-PCR protocol (26). For this reason, allantoic fluid containing 10-fold serial dilutions of titrated AI viruses belonging to the H5, H7, and H9 subtypes was prepared, and the RNA was extracted and then used for the sensitivity test (Table 3). To establish whether the different types of sample matrices could influence the analytical sensitivity, lungs obtained from specific-pathogen-free chickens were weighed (0.1 g) and homogenized with sterile quartz sand in 1 ml (1:10 wt/vol) of PBS, pH 7.4. Lung homogenates were then blended with 10-fold dilutions of titrated subtype H5, H7, and H9 viruses and processed for RNA extraction. Similarly, feces obtained from specific-pathogen-free chickens were used for sensitivity tests. One-gram samples of feces were weighed and homogenized with sterile quartz sand to obtain a 1:5 wt/vol suspension in PBS. Blending and dilution were performed as described for lung samples.
Sensitivity of the method
Evaluation of the analytical sensitivity of the method was done by testing each dilution in five replicates. The sensitivity of the method was determined as the last dilution at which at least 4 of 5 replicates of each dilution was positive.
LoD of the RRT-PCR assay.In the present study, the PCR detection limit reflects the sensitivity of the RRT-PCR procedure, which includes the sensitivity of the primers and probes as well as the preparation of the master mix and the optimization of thermocycling conditions (26). To determine the limit of detection (LoD) of the assay in terms of RNA copy numbers, in vitro-transcribed RNAs of the H5, H7, and H9 genes were analyzed. Briefly, the HA genes of A/chicken/Yamaguchi/7/04 (H5N1), A/turkey/Italy/4580/99 (H7N1), and A/turkey/Wisconsin/66 (H9N2) strains were amplified by RT-PCR and the amplification products were cloned into the PCR-II vector using a dual-promoter TOPO TA cloning kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Plasmids with the HA insert were isolated from positive Escherichia coli colonies by using a GenElute plasmid miniprep kit (Sigma-Aldrich, St. Louis, MO). The H5, H7, and H9 insert control plasmids were linearized by using the restriction enzyme HindIII (MBI Fermentas, Lithuania) for the H9 and H7 genes and NotI (New England Biolabs, MA) for the H5 gene. The in vitro-transcribed RNA was generated from the T7 promoter by using a RiboMax kit (Promega, Madison, WI) according to the manufacturer's recommendations and then quantified by using a UV BioPhotometer (Eppendorf, Hamburg, Germany).
The number of RNA copies was calculated by following the formula reported in a previous study (11). Tenfold dilutions of the RNA transcripts, ranging from 1 to 1010 copies/μl, were prepared. The LoD of the assay was determined from three independent replicates.
Intra- and interassay variability.The repeatability of the H5, H7, and H9 RRT-PCR assay was determined using three different concentrations (high, medium, and low) of each viral subtype tested. The selected concentrations were 105.83, 104.83, and 103.83 (H5); 106.37, 105.37, and 103.37 (H7); and 106, 104, and 102 (H9) EID50/100 μl. For intra-assay variability, each dilution was analyzed in triplicate. For interassay variability, each dilution was analyzed in six different runs performed by two distinct operators on different days. The coefficient of variation was determined in accordance with previously published guidelines (26).
Detection of virus RNA in samples collected from field-exposed and experimentally infected birds.To evaluate whether or not the RRT-PCR assay could be used as a diagnostic tool in surveillance programs for AI, we analyzed retrospectively by VI and RRT-PCR 725 samples collected from different avian species of poultry (n = 234) and wild birds (n = 491) during field and laboratory investigations in Eurasia and Africa in 2006 to 2007. The samples consisted of tracheal swabs (n = 114), cloacal swabs (n = 504), and organs (n = 107; consisting of trachea, lungs, intestines, and brain) collected during necropsies. The number of samples analyzed for clinical validation is in accordance with the guidelines proposed in a previous study (5).
The agreement between VI and RRT-PCR results was investigated using Cohen's K statistics (a statistical measure of interrater agreement) and associated evaluation of statistical significance (20). The results were compared with the Cohen's K standard values proposed by Landis and Koch (13).
RESULTS
Specificity, analytical sensitivity, and intra-/interassay variability.The H5, H7, and H9 primer and probe sets were able to detect RNA of virus strains of their respective subtype only. No positive results were obtained with any of the other organisms listed in Table 1.
The sensitivity of the RRT-PCR assay was determined using in vitro-transcribed RNA and titrated reference viruses.In terms of HA gene copy number, the LoDs for the H5, H7, and H9 subtypes were 103, 101, and 103 gene copies/μl of in vitro-transcribed RNA, respectively. To determine the linearity of the reaction and PCR efficiency, the threshold cycle values of individual dilutions were plotted against the initial gene copy number, leading to typical standard curves. The linear ranges of the RRT-PCR assay span within 1010 and 102 copies/μl for the H7 gene and within 1010 and 104 copies/μl for the H5 and H9 genes. The reaction efficiencies for H5, H7, and H9 genes were 0.97, 0.98, and 0.97, respectively. The correlation coefficient (R2) was higher than 0.99 in all measurements.
The sensitivity of the method relative to the detectable infectious virus titer ranged between 102.74 EID50/100 μl and 101 EID50/100 μl. The results obtained for the sensitivity of the method in different samples are summarized in Table 3.
To assess the intra- and interassay reproducibility, three different concentrations (high, medium, and low) of each reference virus were tested in triplicate in six different runs performed by two distinct operators on different days. The coefficients of variation within runs (intra-assay variability) ranged from 0.12% to 2.64%. The interassay variability was in the range of 2.26% to 4.11% (Table 4).
Intra- and interassay coefficients of variation
Detection of viral RNA in samples collected from field-exposed and experimentally infected birds.A total of 725 samples was analyzed for subtypes H5, H7, and H9 by VI and RRT-PCR (Table 5). Of these, 141 samples tested positive for subtype H5 by means of RRT-PCR (4/114 tracheal swabs, 54/504 cloacal swabs, and 83/107 organ samples). For subtype H7, the RRT-PCR assay identified 58 positive samples (44/114 tracheal swabs and 14/504 cloacal swabs), and 30 specimens resulted in positive results for subtype H9 (21/114 tracheal swabs and 9/504 cloacal swabs). The comparison of the results of the two methods, summarized in Table 5, showed agreements of 94.06%, 99.17%, and 98.89% for the H5, H7, and H9 subtypes, respectively. The Cohen's K coefficients were 0.80, 0.94, and 0.85, respectively. The difference between the two methods was not statistically significant (P < 0.01). The percentage of agreement between the results of RRT-PCR and VI was influenced by the higher number of samples that tested positive by RRT-PCR but negative by VI. However, 43/55 RRT-PCR-positive/VI-negative samples were sequenced and their identities confirmed (data not shown). Only 2 of 26 samples that were positive by VI for subtype H9 tested negative in the molecular assay.
Results of RRT-PCR and VI from clinical samples
DISCUSSION
Several diagnostic methodologies are currently available for the detection of AI infection, with VI in eggs universally recognized as the gold standard. However, this method is time consuming and requires facilities (e.g., BSL3 laboratories) that are not available in many affected areas. Recently, molecular diagnostic tests have proven themselves to be invaluable as a first step in the identification and control of disease outbreaks. Conventional RT-PCR and RRT-PCR have been applied successfully to the diagnosis of AI (7, 8, 12, 14, 15, 19, 21, 23, 27). In this study, we present data on the development and validation of a real-time hydrolysis probe-based RT-PCR assay for the simultaneous detection of AI viruses belonging to subtypes H5, H7, and H9. Our results prove that the assay is highly specific and sensitive. In a previous study (23), an RRT-PCR assay was developed for the sequences of North American AI virus H5 and H7 subtypes. In that assay, the sensitivity data obtained were comparable to the results described here. However, the protocol described previously was a one-step RT-PCR with different thermal profiles for H5 and H7 detection. In addition to its sensitivity and specificity, the method described in the present paper offers several advantages over conventional diagnostic methods, including rapidity, flexibility, and ease of use. This assay makes results for the three major AI viruses currently prevalent in poultry in large areas of the world available in approximately 3 h, and the use of a single-step RRT-PCR procedure provides some protection against contamination events. Based on its technical characteristics, this assay could be used for large-scale screening and subtyping of viral RNA during influenza A virus outbreaks and for surveillance programs.
This RRT-PCR assay was developed and validated using the same annealing temperature in order to identify the H5, H7, and H9 subtypes in a single analytical session. In the literature, the use of multiplexed PCRs has been reported as resulting in a decreased sensitivity of the method (22, 28), and the optimization of the concentration of the multiplex PCR components to achieve optimal amplification can pose several difficulties (9). For these reasons, the development of a multiplex assay was not attempted, as it was so important to identify the three subtypes and maximize the assay's performance. The application of this RRT-PCR format was also due to the necessity of having a cost-effective and flexible diagnostic tool that could be easily switched to a single-subtype identification method that would be applicable during the investigation of outbreaks caused by only one of these subtypes. It should also be considered that not all of the existing real-time PCR platforms are capable of detecting more than two fluorophores simultaneously, making a triplex PCR protocol inapplicable.
The suitability of the RRT-PCR test described in the present study as a diagnostic tool to rapidly recognize the three most-important HA subtypes of the AI virus is confirmed by the results obtained using samples from birds infected naturally. These clinical samples were obtained from a wide range of avian species and geographical areas during field and laboratory investigations. The assay has been used for monitoring the AI virus in poultry and wild birds, and it has proved capable of identifying the presence of several distinct genetic lineages of subtype H5, H7, and H9 viruses, including the H5N1 sublineages circulating in Eurasia and Africa (data not shown). The comparison of the results obtained from applying the conventional diagnostic method (VI) and the RRT-PCR assay to these clinical samples showed good agreement. The lowest level of agreement was observed in the RRT-PCR/VI results for subtype H5. Negative results by VI for H5 could be explained by considering the condition of the clinical samples at the time of their arrival. Many of the samples that proved positive for subtype H5 by RRT-PCR but negative by VI were submitted to the OIE/FAO Reference Laboratory from Africa and the Middle East, and in some cases, the cold chain was not maintained during shipping, compromising the viability of the viruses. Based upon the results obtained in the present study, the RRT-PCR assay for simultaneous detection of subtypes H5, H7, and H9 could be a useful instrument for rapid screening and surveillance in wild and domestic birds. Although this method cannot replace the standard VI technique, this RRT-PCR assay offers several advantages over standard methods, and it could be used as a reliable tool for the rapid detection of the three AI virus subtypes, including identification of cocirculating strains. Routine application in critical environments, such as live-bird markets, or for samples obtained from wild birds in their breeding or resting sites could give an indication of the degree of coinfection with these subtypes, providing insight into the complex ecoepidemiology of AI infections in such birds.
ACKNOWLEDGMENTS
This study was supported by the European Union FLUTRAIN project (training and technology transfer of AI diagnostics and disease management skills) and by the Italian Ministry of Health (RF 2007).
We thank Laura Gagliazzo for statistical assistance.
FOOTNOTES
- Received 14 November 2007.
- Returned for modification 9 January 2008.
- Accepted 17 March 2008.
- Copyright © 2008 American Society for Microbiology