|Home | About | Journals | Submit | Contact Us | Français|
The pandemic caused by a new type of influenza virus, pandemic H1N1 (2009) influenza virus A (AH1pdm), has had a major worldwide impact. Since hemagglutinin (HA) genes are among the most specific genes in the influenza virus genome, AH1pdm can be definitively diagnosed by viral gene analysis targeting the HA genes. This type of analysis, however, cannot be easily performed in clinical settings. While commercially available rapid diagnosis kits (RDKs) based on immunochromatography can be used to detect nucleoproteins (NPs) of influenza A and B viruses in clinical samples, there are no such kits that are specific for AH1pdm. We show here that an RDK using a combination of monoclonal antibodies against NP can be used to specifically detect AH1pdm. The RDK recognized AH1pdm virus isolates but did not recognize seasonal H1N1 and H3N2 and influenza B viruses, indicating that the specificity of the RDK is 100%. A parallel comparison of RDK with a commercial influenza A/B virus kit revealed that both types of kits had equal sensitivities in detecting their respective viruses. Preliminary evaluation of clinical samples from 5 individuals with PCR-confirmed human AH1pdm infection showed that the RDK was positive for all samples, with the same detection intensity as that of a commercial influenza A/B virus kit. This RDK, together with a new vaccine and the stockpiling of anti-influenza drugs, will make aggressive measures to contain AH1pdm infections possible.
The pandemic caused by a new type of influenza virus, pandemic H1N1 (2009) influenza virus A (AH1pdm), has had a major worldwide impact. As of 27 September 2009, more than 4,100 deaths from AH1pdm infection have been reported to the World Health Organization (WHO) (http://www.who.int/csr/don/2009_10_02/en/index.html). Current methods used to diagnose AH1pdm virus in clinical specimens are based on viral RNA analysis targeting hemagglutinin (HA) genes, because the HA genes are among the most specific genes in the influenza virus genome. Although these methods are highly sensitive, they usually take more than 2 to 6 h to complete and require well-equipped laboratories with virologists or well-trained medical technicians and specialized tools for virus genome isolation and amplification (6, 8) (http://www.who.int/entity/csr/resources/publications/swineflu/CDCRealtimeRTPCR_SwineH1Assay-2009_20090430.pdf). Rapid diagnostic kits (RDKs) based on immunochromatography utilize antibodies (Abs) against antigens of interest. Although RDKs are usually less sensitive than genetic assays, they do not require the isolation of a viral genome, thus overcoming the intrinsic difficulties of viral gene analyses. RDKs for many infectious diseases (2, 4, 9, 11-14), including influenza viruses A and B (1), are commercially available. However, RDKs capable of distinguishing AH1pdm viruses from seasonal influenza viruses have yet to be implemented in a clinical setting.
Nucleoproteins (NPs) of influenza A, B, and C viruses have important differences in their antigenicities that enable them to be distinguished from one another but are highly conserved within each major serotype. Thus, antibodies to NPs have been utilized in commercially available RDKs to distinguish between influenza A and B viruses (15). In a monoclonal antibody (MAb) preparation procedure targeting NPs derived from highly pathogenic H5N1 avian influenza (HPAI), we obtained 2 MAbs that reacted with NPs of AH1pdm as well as that of HPAI but not those of seasonal influenza A virus. We have therefore utilized these MAbs in the development of novel RDKs for AH1pdm, and we have validated these RDKs in laboratory environments.
Recombinant NP of influenza A virus [A/Viet Nam/VL-020/2005 (H5N1)] (GenBank accession number AAZ72762), a virus isolated from a patient infected with HPAI, was prepared from Escherichia coli BL21(DE3) CodonPlus-RIPL cells (Stratagene), which carry a TAGZyme pQE2 (Qiagen) derivative carrying the NP protein gene (7). The NP was used to immunize 7- to 9-week-old female WKY rats (Oriental Yeast Co., Ltd.), and rat MAbs were prepared as described previously (10).
The reactivity of the MAbs with NPs derived from seasonal influenza and AH1pdm was analyzed by conventional enzyme-linked immunosorbent assay (ELISA) using microplates coated with NPs or by sandwich ELISA using microplates coated with polyclonal antibodies prepared from rabbits immunized with recombinant NPs.
Sources of NPs for the sandwich ELISA included cultured human A/New York/55/2004 (H3N2) and A/New Caledonia/20/1999 (H1N1) viruses in tissue culture and recombinant NPs from HEK293 cells transfected with cytomegalovirus (CMV) promoter-driven plasmids (7) carrying an NP gene with the sequence of A/California/04/2009 (H1N1) (GenBank accession number ACP44151), a virus isolated from a patient infected with AH1pdm; that of H5N1 HPAI virus [A/Viet Nam/VL-020/2005 (H5N1)] (accession number AAZ72762), a virus isolated from a patient infected with HPAI; and chimeric NPs derived from those of H5N1 HPAI and seasonal H3N2 viruses (as described above) (see Fig. Fig.3b).3b). To construct chimeric NPs, three regions of each NP were amplified separately by PCR using the primers containing restriction sites and ExTaq polymerase (Takara). Primer sequences are available upon request. The PCR products were subsequently purified, digested with the corresponding restriction enzymes, and ligated to insert the DNA fragments into the expression plasmid as described above.
The concentration of each NP was normalized by conventional Western blotting with rabbit anti-NP polyclonal antibody. To perform sandwich ELISA, 250 ng of rabbit anti-NP polyclonal Ab dissolved in 50 mM sodium-carbonate buffer (pH 9.0) was fixed to each well of a 96-well microtiter plate (Corning) at room temperature for 1 h. After washing with phosphate-buffered saline containing 0.02% Tween 20 (PBS-T) and blocking with SuperBlock (Pierce), 10 ng of the NPs dissolved in PBS-T was added to each well. Following incubation and washing, the wells were incubated with culture supernatants of hybridomas producing the indicated MAbs. In a conventional ELISA, 50 ng/well of antigens was fixed onto the plates directly. The binding of MAbs was detected with horseradish peroxidase (HRP)-goat anti-rat IgG (GE Healthcare) and tetramethylbenzidine (TMB) (Bio-Rad).
Eight NP fragments (see Fig. Fig.2a)2a) derived from NP [A/California/04/2009 (H1N1)] (GenBank accession number ACP44151) were prepared in E. coli cells as described above and used for the epitope mapping of MAbs based on ELISA results. Synthetic peptides prepared by a commercial service (>70% purity; Invitrogen) (500 ng each) were fixed onto the plates by incubation in 50 mM carbonate buffer (pH 9.0) containing the chemical cross-linker disuccinimidyl suberate (DSS) (1 mM; Pierce) at room temperature for 1 h, followed by epitope mapping.
The RDK was assembled based on the Quickchaser Flu A,B test (Mizuho Medy, Japan), a commercially available rapid diagnosis kit used to detect influenza A and B viruses in clinical specimens. Sample migration was assayed by using rabbit immunoglobulin (rIg) (Rockland) and anti-rabbit IgG, with rIg conjugated to gold particles as the mobile phase (3). Ab1 recognizes NPs from AH1pdm, seasonal H1N1, and H3N2 viruses but not NP from HPAI virus, while Ab2 and Ab3 recognize NPs from AH1pdm and HPAI viruses but not NPs from seasonal H1N1 or H3N2 virus (see Results). To detect AH1pdm specifically and to exclude the reactivity against NP from HPAI virus in the RDK, Ab1 was used as the gold particle-labeled mobile antibody, while Ab2 or Ab3 was used as the capture antibody. In such a composition of the RDK, NP from H5N1 HPAI virus will migrate to and bind to Ab2 or Ab3, although it will not generate a line because it is not attached to colloidal gold-labeled Ab1. To prepare test lines, Ab2 or Ab3 (0.76 μg/test) was coated onto nitrocellulose membranes (Millipore) at a position 30 mm from the sample dropping point and allowed to dry at room temperature. To prepare control lines, anti-rabbit IgG (0.2 μg/test) (Rockland) was coated onto the membrane at a position 39 mm from the sample dropping point and allowed to dry. Pads were prepared by dropping anti-influenza A virus NP MAb (clone M322211; Fitzgerald) (named Ab1 in this study) and rIg, each conjugated with colloidal gold, onto glass filters, followed by freeze-drying. The nitrocellulose membrane and pad were assembled with filter papers as sample application pads and absorption pads on a plate within a plastic housing (see Fig. Fig.4).4). These assembled RDKs were stored in a waterproof bag with desiccant at room temperature (1°C to 30°C).
The specificity of the RDK for AH1pdm was assessed by using reverse transcription (RT)-PCR-confirmed AH1pdm and seasonal influenza A and B viral strains isolated during 2009 from infected individuals in Japan by culturing on Madin-Darby canine kidney (MDCK) cells. The viral isolation procedures were performed at the Osaka Prefectural Institute of Public Health. Experiments using clinical samples were reviewed and approved by the Institutional Review Board of Mizuho Medy. All patients provided written informed consent, and all clinical samples were assayed anonymously.
Each isolate (ca. 108 viral copies/100 μl) was extracted into sample extraction tubes containing sample extraction buffer (0.4 M Tris buffer containing 1% nonionic detergent and 0.09% sodium azide) and assayed by RDK by visual assessment of line intensity on a scale from − to +++. In parallel, these samples were analyzed with a commercial influenza A/B virus kit to validate reactivity. In addition, these samples were diluted (see Table Table2)2) to assess the sensitivity of the RDK for AH1pdm. The specificity of the RDKs was also tested by using clinical specimens obtained by nasal swabs.
Analyses of viral copy numbers and detection of the influenza viruses by real-time RT-PCR were performed according to WHO criteria (http://www.who.int/csr/resources/publications/swineflu/WHO_Diagnostic_RecommendationsH1N1_20090521.pdf) for AH1pdm and influenza A viruses and according to methods described previously by van Elden et al. (16) for influenza B virus.
During the preparation of MAbs against NP of HPAI virus, we obtained 2 MAbs, Ab2 and Ab3, that were highly reactive against NP from AH1pdm as well as HPAI virus but not against NPs from seasonal influenza A H1N1 [A/New Caledonia/20/1999 (H1N1)] and H3N2 [A/New York/55/2004 (H3N2)] viruses (Fig. (Fig.1a).1a). Both MAbs showed 6-fold-greater reactivity with NP from AH1pdm than with NPs from the seasonal H1N1 and H3N2 viruses.
To identify the determinant(s) recognized by Ab2 and Ab3 in the AH1pdm virus NP, we performed epitope mapping with 8 protein fragments derived from the NP of AH1pdm virus (Fig. (Fig.2).2). Both Ab2 and Ab3 reacted with the most N-terminal fragment, fr 1-1-1, containing amino acids (aa) 1 to 56 of NP (Fig. (Fig.2b)2b) only. Since amino acids 37 to 56 are overlapped with those of fr 1-1-2, we compared the amino acid sequences of amino acids 1 to 36 of NPs from seasonal H3N2, seasonal H1N1 HPAI, and AH1pdm viruses. As a result, we found mutations specific for HPAI and AH1pdm viruses at amino acid positions 16 to 18. NPs from AH1pdm and H5N1 HPAI viruses have the sequence GGE, whereas NPs from the seasonal H1N1 and H3N2 viruses have the sequences DGE and DGD, respectively. To confirm that Ab2 and Ab3 could distinguish the GGE sequence from DGE and DGD, we synthesized three 15-mer peptides, QGTKRSYEQMETDGE (peptide H1 from seasonal H1N1 virus), QGTKRSYEQMETDGD (peptide H3 from seasonal H3N2 virus), and QGTKRSYEQMETGGE (peptide AH1pdm from AH1pdm and HPAI viruses) and analyzed their reactivities with Ab2 and Ab3 (Fig. (Fig.2c).2c). Both Ab2 and Ab3 reacted with peptide AH1pdm but not with seasonal peptides H1 and H3.
To analyze the prevalence of the each sequence in NPs of influenza A virus from human cases, we analyzed the amino acid sequences of NPs from more than 1,000 patients infected from 2007 to 2009, including patients infected with seasonal H3N2, seasonal H1N1, HPAI, and AH1pdm viruses (Table (Table1).1). Of the 543 AH1pdm virus isolates, 537 (97.3%) had the sequence GGE, whereas only 6 AH1pdm virus isolates had sequences other than GGE. All 14 HPAI H5N1 virus isolates from human cases also had the sequence GGE. In comparison, only one seasonal influenza isolate (a seasonal H3N2 virus) had the sequence GGE. These results indicate that Ab2 and Ab3 can be used to distinguish AH1pdm in addition to HPAI viruses from seasonal H1N1 and seasonal H3N2 viruses by recognizing amino acids 16 to 18, located at the N termini of NPs.
We also unexpectedly found a commercial MAb, named Ab1, that failed to recognize NP of H5N1 HPAI virus but that was capable of recognizing NPs of AH1pdm and seasonal H1N1 and H3N2 viruses (Fig. (Fig.3a).3a). Ab1 showed very little reactivity with NP from H5N1 virus, the same as blank. Although we attempted to identify the epitope as we did for Ab2 and Ab3 by using a fragmented NP as shown in Fig. Fig.2b,2b, Ab1 did not react with the fragments at all (data not shown). Thus, we prepared chimeric NPs of H5N1 and H3N2 viruses (Fig. (Fig.3b)3b) as the recombinant proteins in an expression system using HEK293 cells, and their reactivity with Ab1 was analyzed. Ab1 reacted with chimera 2 and chimera 4 as well as NP from H3N2 virus but failed to react with chimera 1 and chimera 3. These results indicated that Ab1 is a MAb recognizing a conformational epitope of NPs that appears to be located between amino acids 1 and 188. This conclusion is based on the fact that Ab1 recognized chimeras containing aa 1 to 188 from viruses that reacted with Ab1 (e.g., H3N2 virus), while Ab1 failed to recognize chimeras that contained aa 1 to 188 from viruses that failed to react with Ab1 (e.g., H5N1 virus).
As shown above, Ab2 and Ab3 reacted with NP from AH1pdm and HPAI virus, while Ab1 reacted with NP from AH1pdm and seasonal H1N1 and H3N2 viruses but failed to react with NP from HPAI virus. These observations suggested that a combination of Ab1, Ab2, and Ab3 could be used to distinguish NPs derived from AH1pdm viruses from those of seasonal influenza and H5N1 HPAI viruses. We tested the feasibility of utilizing prototype RDKs.
Prototype RDKs for the immunochromatographic detection of AH1pdm virus using MAbs were assembled as described in Materials and Methods, based on a commercial influenza A/B virus kit (Fig. (Fig.4a).4a). These RDKs were used to detect viruses by dropping 100 μl of each diluted sample (in extraction buffer) onto the test plate and waiting for 10 min (Fig. (Fig.4b).4b). A positive result for the AH1pdm virus was indicated by a purple line in the test line area as well as by the control line, which was included to ensure a normal flow of reaction mixture. The intensity of the test lines was scored from − to +++, whereas the control lines were scored ++ (Fig. (Fig.4b),4b), making the assay semiquantitative. The RDKs showed positive results with the AH1pdm virus sample, even after a 1,000-fold dilution, but not with seasonal influenza virus samples even at 10-fold dilutions. We also tested the reactivity of our RDKs with A/New Jersey/8/76, obtained from the ATCC (ATCC VR-897), which is known as the prototype of swine influenza virus transmitted to humans. The viral stain harbors the GGE sequence in its NP and reacted with the prototype kits, as expected (data not shown).
To assess the sensitivity of the prototype kits prepared using Ab2 or Ab3, cultured influenza A virus samples were diluted and dropped onto the kits; in parallel, these samples were tested by using a commercial influenza A/B virus kit, which cannot distinguish AH1pdm from other seasonal influenza A viruses (Table (Table2).2). The prototype kits reacted positively with AH1pdm samples diluted 1:16,000, similar to the sensitivity of the commercial kits for seasonal influenza A viruses. In contrast, the prototype kits did not show positive reactions with any of the seasonal influenza A virus samples, even at a 1:10 dilution, while the commercial kits did. The RDKs had an average lower detection threshold for AH1pdm of 2 × 105 viral copies/kit based on analyses using five different AH1pdm viral cultures (data not shown).
To evaluate the specificity of our prototype RDK in detecting AH1pdm viruses, influenza viruses isolated in tissue culture during 2009 were screened. Specifically, we assessed 30 AH1pdm, 20 seasonal H1N1, 20 seasonal H3N2, and 5 influenza B viruses independently isolated from infected patients. The RDK was positive for all the AH1pdm virus isolates (Table (Table3),3), with the intensity of the lines essentially identical to those observed with commercial influenza A/B virus kits (data not shown). In contrast, the RDK did not react with any of the H1N1, H3N2, and influenza B virus isolates. These results indicate that our prototype RDK could distinguish AH1pdm from seasonal influenza viruses with a specificity of 100%.
We also tested the reactivities of the RDKs using clinical specimens obtained during 2009 in Japan from 5 AH1pdm, 20 seasonal influenza A, and 9 seasonal influenza B virus-infected individuals as well as with 20 clinical specimens that were negative for influenza virus based on PCR analyses (Table (Table4).4). The RDKs reacted with all 5 AH1pdm virus-positive samples but not with any of the seasonal influenza A virus-, seasonal influenza B virus-, and influenza virus-negative samples. In comparison, the commercial influenza A/B virus kits showed positive reactions with all of the AH1pdm, seasonal influenza A, and seasonal influenza B virus clinical samples. These results indicate that our RDKs could specifically detect AH1pdm in clinical samples.
Here we have described the development of an RDK that can be used to distinguish AH1pdm viruses from seasonal influenza viruses using MAbs Ab2 and Ab3. NPs of influenza A, B, and C viruses have important differences in their antigenicities that enable them to be distinguished from one another but are highly conserved within each major serotype. A detailed analysis of the NPs from influenza A virus, however, showed considerable sequence variation among them (data not shown). Epitope mapping of Ab2 and Ab3 showed that both MAbs recognize a peptide containing residues 16 to 18 of NP from AH1pdm as well as H5N1 HPAI viruses. The corresponding region of NPs from AH1pdm and H5N1 HPAI viruses was GGE, while those of seasonal H1N1 and H3N2 viruses were DGE and DGD, respectively. Prevalence analyses indicated that this amino acid difference could be used to distinguish AH1pdm from seasonal influenza A viruses (Table (Table1).1). Usually, the detection of new-type influenza viruses, such as AH1pdm and H5N1 viruses, in clinical specimens is performed with assays targeting HA of influenza virus (http://www.who.int/entity/csr/resources/publications/swineflu/CDCRealtimeRTPCR_SwineH1Assay-2009_20090430.pdf) because of the subtype specificity of HA genes. Since HA is highly mutagenic (5), assays of this gene would have to take into account mutations in HA that occur every season. In contrast, the amino acid sequences of the NPs, including that of the AH1pdm virus, which emerged in 2009, are, in general, well conserved. Although reassorted viruses may emerge, from which NPs of specific viruses of swine origin may be lost, our results indicate that methods targeting influenza virus NPs to discriminate among viral subtypes are plausible alternatives to HA-directed methods.
The RDK developed in this study represents a means for the rapid, noninvasive, and cost-effective diagnosis of AH1pdm virus in infected individuals by health workers in remote sites, at the bedside, and in quarantined areas such as airports. The use of the RDK not only could reduce the risk of mortality and morbidity but also would reduce the impact of an influenza pandemic by facilitating the more rapid diagnosis, treatment, and quarantine of infected individuals. Furthermore, the ability of our RDK to definitively diagnose AH1pdm virus infection, which cannot be done using conventional influenza A/B virus kits and is not be feasible by gene analyses such as real-time RT-PCR, would reduce the demand for vaccination against AH1pdm virus. Although we could test only one type of NP derived from H5N1 HPAI virus, which showed a negative result (data not shown), because of the limited availability of the viral strains, our RDKs showed sufficient specificity and sensitivity to detect AH1pdm virus. A large-scale multicenter evaluation of an RDK for AH1pdm virus to obtain approval from regulatory authorities is under way.
We thank N. Saito, M. Komiya, and K. Muraoka for preparation and characterization of the antibodies.
This study was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan and the Japan Science and Technology Agency (JST).
Published ahead of print on 13 January 2010.