Nucleic Acid Dipstick Test for Molecular Diagnosis of Pandemic H1N1

Control virus cultures.Cultured A/England/195/2009 pandemic H1N1 virus was provided by the Centre for Infections (CfI), Health Protection Agency (HPA), London, United Kingdom. The cultured virus was heat inactivated (56°C for 90 min) and diluted in Copan universal transport medium (UTM). Quality Control for Molecular Diagnostics (QCMD) influenza virus hemagglutinin typing panels 2007 and 2008 and cultures of 10 common respiratory pathogens were obtained from Qnostics Ltd. (Glasgow, United Kingdom). RNA extracts from an external quality assessment (EQA) panel were provided by CfI, HPA.

Clinical samples.A total of 265 mixed nasal/throat swab specimens submitted to the Clinical Microbiology and Public Health Laboratory at Addenbrooke's Hospital, Cambridge, United Kingdom, for pandemic (H1N1) 2009 virus testing between 12 and 27 November 2009 were analyzed in a double-blinded manner by both the United Kingdom National Standard Method and the SAMBA for pandemic H1N1. All of these specimens were randomly selected from individuals who presented with influenza-like symptoms. Three postmortem samples were excluded, leaving 262 valid samples. Patient ages ranged from 19 days to 91 years, with a mean of 22.4 years, and the male-to-female ratio was 1.2 to 1. The specimens were decoded, and the results were compared at the end of each working day. Samples that gave discordant results with qRT-PCR and the SAMBA for pandemic H1N1 were tested again by the SAMBA as well as sent to CfI for confirmation by qRT-PCR and sequencing.

Nucleic acid extraction.RNA from cultured virus was extracted using a QIAamp viral RNA minikit (Qiagen, Crawley, United Kingdom). Clinical samples (500 μl) were processed with the NucliSENS easyMAG system (bioMérieux, Lyon, France), taking 40 min. As an internal control, bacteriophage MS2 (ATCC 15597-B1) was included (at ∼4,600 PFU) in each sample prior to extraction (24). Viral nucleic acid (all from RNA viruses except two adenovirus isolates, which are DNA viruses) for the specificity study was extracted with the use of a QIAamp DNA minikit (Qiagen). Extracted nucleic acid was used as the template for the SAMBA and the qRT-PCR assay.

Detection of pandemic (H1N1) 2009 virus by the United Kingdom National Standard Method.Clinical specimens were screened for influenza virus with qRT-PCR assays by biomedical scientists at the regional clinical microbiology laboratory in the Cambridge Health District, United Kingdom. A generic quadriplex assay capable of detecting all influenza A virus subtypes, influenza B virus, the hemagglutinin H5 subtype, and bacteriophage MS2 was performed as described previously (15). A minor modification was made to the specific minor grove binding (MGB) TaqMan probe for the matrix protein gene (at the third position) to improve its similarity to the pandemic H1N1 sequence (VIC-TCYTGTCACCTCTGAC-MGBNFQ [where MGBNFQ is the minor-groove binder nonfluorescence quencher]; Applied Biosystems, Warrington, United Kingdom). The second assay was specific for pandemic (H1N1) 2009 virus and was originally designed and developed at CfI but was modified into a one-step quantitative assay for the Rotor-Gene 3000 or 6000 real-time thermal cycler (Corbett Research, Qiagen). The primers (Metabion, Martinsried, Germany) H1 Sw Forward (5′-TTACCAGATTTTGGCGATCTAYT-3′) and H1 Sw Reverse (5′-CCAGGGAGACTASCARTACCA-3′) target the 5′ terminus of RNA segment 4 (hemagglutinin gene) and yield a 62-bp amplicon. The MGB probe (5′-ACWGTCGCCAGTTC-3′) contains a 6-carboxyfluorescein (FAM) label at its 5′ end. Degeneracies were introduced into the primers and probe to facilitate the additional amplification of positive-control material (A/Aragon/RR3218/2008 virus) supplied by CfI to laboratories in England at the start of the influenza pandemic. Both assays were performed with the use of a SuperScript III Platinum one-step qRT-PCR kit (Invitrogen, Paisley, United Kingdom) in a reaction volume of 25 μl (containing 3 mM MgSO₄) using the Rotor-Gene 3000 or 6000 instrument. The same amplification conditions were applied for each assay: incubation at 50°C for 30 min and at 95°C for 2 min, followed by 45 cycles of denaturation at 95°C for 15 s and annealing and extension at 60°C for 1 min. The fluorescence was measured at the end of each cycle. For the pandemic (H1N1) 2009 virus-specific assay, the primer (forward or reverse) and probe concentrations were 600 and 250 nM, respectively. Excluding sample extraction, the assay took 125 min.

Design of primers and probes for SAMBA for pandemic H1N1.Primer and probe sets were designed to target conserved sequences specific for the pandemic (H1N1) influenza virus. Virus sequences were obtained from the Influenza Virus Resource (National Center for Biotechnology Information; http://www.ncbi.nlm.nih.gov/genomes/flu/flu.html). The sequences of the hemagglutinin genes of A/California/07/2009, A/Brisbane/59/2007 (H1N1), and A/Brisbane/10/2007 (H3N2) viruses were first aligned with the use of the multiple alignment using fast Fourier transform (MAFFT) program, available from the European Molecular Biology Laboratory-European Bioinformatics Institute (EMBL-EBI; http://www.ebi.ac.uk/tools/mafft), and were then analyzed with the Jalview program (version 2.3; University of Dundee, Dundee, United Kingdom). Regions that were conserved among pandemic H1N1 strains but that contained mismatched sequences relative to the sequences of seasonal influenza A viruses were selected as targets for the primers. Detector and capture probes were also designed to target similarly mismatched nucleotides. The primers and probes were compared with the sequences in the Nucleotide Collection database at NCBI with the use of the Basic Local Alignment Search Tool (BLAST) in order to determine their specificities. The primer set was also aligned with all H1, H3, and H5 sequences to confirm subtype specificity. The oligonucleotides were analyzed for predicted secondary structure, GC content, and potential hetero- and homodimer formation with the use of an OligoAnalyzer (version 3.1; IDT Integrated DNA Technologies).

Detection of pandemic (H1N1) 2009 virus by SAMBA for pandemic H1N1.Amplification of RNA extracted from pandemic (H1N1) 2009 virus was performed by isothermal amplification (18). The amplified products were then detected with a visual detection system on a dipstick, as described previously (5, 17). In brief, after the amplification step, the amplification product was incubated in a 2-ml microcentrifuge tube at 41°C on a heating block. Twenty microliters of amplification product was added to our proprietary detection mixture, and the dipstick was inserted in the reaction mixture. The test results were examined after 25 min of incubation, and the signal on the dipstick was scored by at least two operators, according to the in-house scoring chart (Fig. 1). The total time for amplification and detection was 85 min. The procedure was performed by a research student and a research assistant in the University of Cambridge.

• Open in new tab
• Download powerpoint

FIG. 1.

Detection of pandemic (H1N1) 2009 virus on the dipstick. After amplification of the influenza viral genome, the amplification product was visually detected on the dipstick by naked eye and the signal strength can be scored according to the scoring chart. (a) Reference scoring chart showing signal strength from 0.5 to 5 in 0.5-point increment. (b) Samples of test results on the dipsticks showing strong positive, intermediate positive, and negative (from left to right) samples. The upper line is the internal control line, which should be present in all tests to validate the test run; the lower line is the test line indicating the presence of the pandemic (H1N1) 2009 viral genome in the samples.

Statistical analysis.Pearson's correlation coefficient was used to compare the threshold cycle (C_T) value of the qRT-PCR and the signal strength of the SAMBA for pandemic H1N1 test. A P value of <0.05 was considered statistically significant.

Primers and probes for SAMBA for pandemic H1N1.Primer and probe design for the SAMBA for pandemic H1N1 took advantage of the limited number of mismatched nucleotides in the hemagglutinin gene between the pandemic (H1N1) 2009 and seasonal H1N1 strains. A 199-bp target region was amplified by the forward primer (5′-ACTTTGTTGGTCAGCACTAGTAGAT-3′) and reverse primer (5′-CAATCATGACTCGAACAAAGGT-3′), which was attached to a T7 promoter sequence and a linker sequence (5′-AATTCTAATACGACTCACTATAGGGAGAAGG-3′, at the 5′ end). The detector probe (5′-TTCCTTTTTTAACTAGCCAT-3′), which was labeled with multiple hapten moieties, and the capture probe (5′-CTTTCCCTTTATCATTAATG-3′) were designed to anneal to sequences within the single-stranded amplification product. The underlined nucleotides are specific to pandemic (H1N1) 2009 virus. Comparison of the primer sequences with the sequences in the Nucleotide Collection database at NCBI predicted no cross-reaction of the primer set with common H1, H3, H5, H7, or H9 influenza A viruses, influenza B viruses, or other common human pathogens. A panel of common H1, H3, H5, H7, and H9 influenza A viruses, influenza B viruses, and four swine influenza viruses as well as other common respiratory pathogens was used to test the specificity of the SAMBA for pandemic H1N1 (Table 1). All virus strains in the panel, including the two H1N1 and two H1N2 swine viruses, tested negative.

Analytical sensitivity.The limit of detection of the SAMBA for pandemic H1N1 was determined by testing diluted RNA extracts from a titrated stock of reference virus (A/England/195/2009). The limit of detection was defined as the lowest concentration of extract at which 95% of replicates tested positive (signal, ≥0.5). Testing of 20 replicates of 5, 1, 0.5, 0.25, 0.125, 0.05, and 0.01 PFU of cultured virus per test by three operators independently yielded an analytical sensitivity of 0.25 PFU per test.

Evaluation of SAMBA for pandemic H1N1 with clinical samples.Among 262 nose/throat swab samples tested, 82 initially yielded positive results with both the SAMBA pandemic H1N1 test and the comparator method. Seven samples were reactive only by qRT-PCR, and one sample was reactive only by the SAMBA. These eight discordant samples were tested again by the SAMBA for pandemic H1N1, and two of the seven that initially tested negative by this test were found to be positive. The remaining six discordant samples were sent to CfI for further testing. Four samples were not confirmed to harbor pandemic (H1N1) virus, and the remaining two samples were reactive by qRT-PCR, although sequencing failed as a result of a low viral load. In summary, among 86 true-positive and 176 true-negative samples, the SAMBA for pandemic H1N1 detected 82 positive and 175 negative samples but gave one false-positive and four false-negative results. Among the four samples with false-negative results, two yielded a positive result on retesting with a lower volume of extract (5 μl instead of 15 μl), possibly as a result of dilution of residual inhibitor in the extract. The proportion of positive samples detected was 31.3% (82/262), which was similar to the prevalence (34.5%) of pandemic H1N1 between week 46 (week ending 15 November 2009) and week 48 (ending 29 November 2009) in England (13).

To address the reason for the apparent deficit in sensitivity of the SAMBA for pandemic H1N1, we examined the correlation between the SAMBA signal strength and the C_T value for qRT-PCR testing (Fig. 2). The visual signal of the SAMBA pandemic H1N1 test showed a significant inverse correlation with the C_T value of qRT-PCR (r = −0.68, P < 0.0001). The sensitivity deficit of the SAMBA appeared to be related to the low H1N1 RNA load in several samples. Compared with that of the gold standard method, the performance of the SAMBA test was thus characterized by a sensitivity of 95.3%, specificity of 99.4%, positive predictive value (PPV) of 98.8%, and negative predictive value (NPV) of 97.8% (Table 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Correlation of the dipstick signal for the SAMBA for pandemic H1N1 and the C_T value for qRT-PCR with clinical samples. Samples that tested positive by both the SAMBA for pandemic H1N1 and qRT-PCR (○), those that tested initially negative but subsequently positive by the SAMBA (•), and those that tested negative by the SAMBA but positive by qRT-PCR (▵) are indicated. The C_T value is inversely proportional to the viral load. A significant inverse correlation was apparent between the two sets of results (r = −0.68, P < 0.0001).