|Home | About | Journals | Submit | Contact Us | Français|
Amantadine is one of the antiviral agents used to treat influenza A virus infections, but resistant strains have widely emerged worldwide. In the present study, we developed a novel method to detect amantadine-resistant strains harboring the Ser31Asn mutation in the M2 gene based on the cycling probe method and real-time PCR. We also studied the rate of amantadine resistance in the 2007-2008 influenza season in Japan. Two different primer and cycling probe sets were designed for A/H1N1 and A/H3N2 each to detect a single nucleotide polymorphism corresponding to Ser/Asn at residue 31 of the M2 protein. By using nasopharyngeal swabs from patients with influenza-like and other respiratory illnesses and virus isolates, the specificity and the sensitivity of the cycling probe method were evaluated. High frequencies of amantadine resistance were detected among the A/H1N1 (411/663, 62%) and A/H3N2 (56/56, 100%) virus isolates collected from six prefectures in Japan in the 2007-2008 influenza season. We confirmed that the cycling probe method is suitable for the screening of both nasopharyngeal swabs and influenza virus isolates for amantadine-resistant strains and showed that the incidence of amantadine resistance among both A/H1N1 and A/H3N2 viruses remained high in Japan during the 2007-2008 season.
Adamantanes (amantadine and rimantadine) have been used for the prevention and treatment of influenza A virus infections (25). The molecular basis of resistance has been identified as single nucleotide changes that lead to corresponding amino acid substitutions at one of four critical amino acid residues (residues 26, 27, 30, and 31) in the transmembrane region of the M2 ion channel protein (19, 26, 27). Recent studies suggest that the rates of influenza virus A/H3N2 resistance to amantadine and rimantadine have been high globally since 2005 (2, 6, 7, 9, 29, 30), while the rates of resistance among A/H1N1 viruses varied from country to country but increased sharply from 2006 onwards (1, 9, 32). It should be noted that resistance in both subtypes was almost exclusively associated with one amino acid substitution at residue 31 (Ser to Asn) of the M2 ion channel protein after 2005 (1, 2, 6, 7, 9, 29, 30, 32).
We have previously established methods for the detection of amantadine susceptibility, such as the virus titration method with comparison of the 50% tissue culture infectious doses (TCID50s; TCID50/0.2 ml) in the presence and the absence of amantadine (24) and PCR-restriction fragment length polymorphism (PCR-RFLP) analysis (21, 31). Other methods for the detection of resistant strains have also been reported, such as enzyme-linked immunosorbent assay (ELISA) (4), plaque reduction assay (13), and DNA sequencing (24). In general, however, conventional methods are time-consuming. Recently, a high-throughput method of genetic analysis called pyrosequencing was used as a rapid method for screening for amantadine and neuraminidase inhibitor resistance (6-9, 11); however, the cost of pyrosequencing is not always coverable for every laboratories.
In the study described here, we developed a rapid assay using a chimera probe-adapted real-time PCR, or the cycling probe method, to detect amantadine-resistant viruses with the Ser31Asn substitution in the M2 ion channel protein. Furthermore, we report the frequency of amantadine resistance among influenza A viruses in six prefectures in Japan in the 2007-2008 influenza season.
The cycling probe technology is a unique nucleic acid-based method that detects single nucleic acid polymorphisms (SNPs) in a target DNA sequence by using a probe-adapted real-time PCR (3, 10) (Fig. (Fig.1).1). The cycling probe method involves a reaction between a chimeric fluorescence- and quencher-labeled DNA/RNA oligonucleotide probe (cycling probe) and RNase (RNase H). This cycling probe is a short DNA fragment (normally 10- to 20-mer) accommodating an RNA complementary to the nucleotide of interest that undergoes degeneration by RNase H activity, once a DNA-RNA complex is formed during annealing. This degeneration leads to the emission of strong fluorescence (17), and by measuring the intensity of the fluorescence, the amount of amplified product can be quantified (Fig. (Fig.1a).1a). For SNP typing, two cycling probes labeled with two different fluorescence dyes (6-carboxyfluorescein [FAM] or 6-carboxy-X-rhodamine [ROX]) are used, with each probe harboring RNA corresponding to the wild-type nucleotide or the nucleotide with a mutation at the SNP position (Fig. (Fig.1b1b).
Nasopharyngeal swab samples that previously tested positive for influenza virus by virus isolation and for which their genetic substitution of interest was confirmed by sequencing were selected for evaluation of the assay's specificity. These included 20 amantadine-sensitive and 20 amantadine-resistant samples each of the influenza virus A/H1N1 and A/H3N2 subtypes. The viral RNA of influenza viruses A/H1N1 and A/H3N2 and other common respiratory viruses (respiratory syncytial virus, parainfluenza virus, enterovirus, and rhinovirus) and the DNA of adenovirus were extracted from 100 μl of the supernatants of the nasopharyngeal swabs or the virus culture supernatant by using an Extragen II kit (Kainos, Tokyo, Japan), according to the manufacturer's instructions. Reverse transcription was performed in a reaction separate from the real-time PCR in order to obtain 25 μl of the first-strand cDNA of the influenza virus genome by using influenza A virus universal primer Uni12, as reported elsewhere (16). The RNA of the other respiratory viruses used to check for cross-reactions was reverse transcribed by using random primers (Invitrogen Corp., Carlsbad, CA).
PCR primers (sets of forward and reverse primers for H1N1 and H3N2) were designed to specifically amplify the M2 gene transmembrane region of influenza viruses A/H1N1 and A/H3N2; the PCR product sizes were 155 bp and 98 bp, respectively. The chimera probes H1N1-AS (where AS indicates amantadine sensitive), H1N1-AR (where AR indicates amantadine resistant), H3N2-AS, and H3N2-AR were created to detect Ser31 (AGT) and Asn31(AAT) in the M2 gene, respectively (the underscores indicate the nucleotide replaced by RNA) (TaKaRa Bio Inc.) (Table (Table1).1). A CycleavePCRCore kit (TaKaRa Bio Inc.) was used for the PCR and the simultaneous cleavage of RNase H. The amplification was carried out in a total volume of 25 μl. The reaction mixture contained (final concentrations are given) 1× CycleavePCR buffer, 3 mM Mg2+, 0.3 mM each deoxynucleoside triphosphates, 5 pmol of each PCR primer (forward and reverse), 5 pmol of each probe (the FAM-labeled probe and the ROX-labeled probe), 100 U of Tli RNase HII, and 1.25 U of Ex Taq HS (TaKaRa Bio Inc.). One microliter of cDNA was added to the reaction mixture as the DNA template. To detect A/H1N1 viruses, a set of primers and probes consisting of the H1N1 forward primer, the H1N1 reverse primer, and the H1N1-AS and H1N1-AR probes was used; and to detect A/H3N2 viruses, a relevant set of H3N2 primers and probes was employed (Table (Table1).1). PCR amplification and fluorescence detection were performed with a TP800 thermal cycler Dice real-time PCR system (TaKaRa Bio Inc.). The conditions of the PCR cycles were as follows: a hold at 95°C for 10 s, followed by 40 cycles of denaturation at 95°C for 5 s, primer annealing at 57°C for 10 s, and extension and emission of fluorescence at 72°C for 15 s. Duplicate wells were used for each sample, and amantadine-sensitive (Ser31) and -resistant (Asn31) control plasmids were included in each run in a 96-well plate.
Four positive control plasmids, H1N1-AS, H1N1-AR, H3N2-AS, and H3N2-AR, were made from the PCR product of each subtype amplified with the same PCR primers used in this study. Both the H1N1-AS and H3N2-AS controls contained sequences that code for serine at position 31 (amantadine sensitive), while the H1N1-AR and H3N2-AR controls possessed sequences with mutations that code for asparagine at position 31. The purified PCR product was ligated and cloned into a pMD20-T vector (TaKaRa Bio Inc.) and was transformed into JM109 competent cells (TaKaRa Bio Inc.) with a Mighty TA cloning kit (TaKaRa Bio Inc.), according to the manufacturer's instructions. Positive clones were selected by the blue-white colony pickup method and were then cultured in Luria-Bertani broth and incubated overnight at 37°C in a shaking incubator. The bacterial culture was pelleted by centrifugation, and the plasmids were extracted with a Wizard Plus SV Minipreps DNA purification system (Promega, Madison, WI), according to the manufacturer's instructions. Sequencing was performed with an ABI Prism BigDye Terminator (version 3.1) cycle sequencing kit (Applied Biosystems, Foster City, CA), according to the manufacturer's instructions, and the products were sequenced with an ABI 3100 automatic sequencer (Applied Biosystems) to confirm the presence of the insert. The sequences were aligned and compiled by using BioEdit software (version 7.0.7) (14).
After the DNA concentrations of the plasmids with the wild-type and mutant sequences were measured with a spectrophotometer (GeneQuant 1300; GE Healthcare UK Ltd., Buckinghamshire, England), serial 10-fold dilutions were made to determine the detection sensitivities of this method for the wild-type and mutant sequences of each subtype. In addition, mixtures of plasmids with the wild-type and mutant sequences were tested at ratios of 1:1, 1:10, 1:100, and 1:1,000 to examine the detection sensitivity of the method with a mixed population.
Clinical samples were collected from patients with influenza-like illnesses at 14 outpatient clinics and hospitals from October 2007 to April 2008 in six prefectures (Hokkaido, Niigata, Gunma, Kyoto, Hyogo, and Nagasaki) in Japan. Briefly, after written or oral informed consent was obtained, nasopharyngeal swab specimens were collected from the patients with influenza-like illnesses, and the medication used in the clinic for the treatment of influenza (amantadine, oseltamivir, or zanamivir) was recorded. The swabs were stored in viral transport medium and were kept at 4°C until transportation to the Department of Public Health, Niigata University, for virus isolation within 1 week. Then, 100-μl aliquots of supernatants of the nasopharyngeal swab specimens were inoculated into Madin-Darby canine kidney cells (MDCK), prepared in 48-well plates. The plates were incubated at 34°C under a 5% CO2 atmosphere for up to 10 days to assess the samples for the presence of cytopathic effects (CPE). Fifty-microliter aliquots of the supernatants of CPE-positive samples were then passaged twice to obtain a sufficient virus titer for virus identification. The influenza virus isolates were typed and subtyped by a hemagglutination inhibition (HAI) assay with commercially available antisera to the influenza virus vaccine strain for the 2007-2008 season in Japan (Denka Seiken Co., Ltd., Tokyo, Japan), namely, A/Solomon Islands/3/2006 (A/H1N1), A/Hiroshima/52/2005 (A/H3N2), and B/Malaysia/2506/2004, and with guinea pig red blood cells. RNA extraction, cDNA synthesis, and the cycling probe real-time PCR were employed with the nasopharyngeal swab specimens and virus isolates as described above to examine whether they possessed the S31N substitution.
A FAM-labeled cycling probe was designed to detect the sequence for amantadine sensitivity (AGT) at amino acid position 31 in the M2 protein by replacing guanine DNA with RNA, while the ROX probe replaced the adenine DNA with RNA corresponding to the sequence for amantadine resistance (AAT) (Table (Table1).1). If the sample in question had a sequence conferring sensitivity, the fluorescent emission of FAM was detected after the hybridized RNA and DNA complex at the guanine position was degenerated by RNase H and the dye was subsequently liberated from the quencher. The same reaction occurred with the sequence conferring amantadine resistance, which harbored adenine at the identical position (position 31) and which is reported by the detection of ROX.
The cycling probes were first tested with DNA plasmids with the known wild-type and mutant sequences in the M2 protein for both subtype A/H1N1 and subtype A/H3N2, which served as controls for our two-step cycling probe real-time PCR. Fluorescence intensities with threshold cycle (CT) values of between 16 and 20 for 2.9 × 107 copies for A/H1N1 and 2.95 × 107 copies for A/H3N2 were detected for the plasmids with the wild-type sequence and the plasmids with the mutant sequence (Fig. 2a and b). The detection limits for the controls were 2.9 × 102 copies for the A/H1N1 plasmid and 2.95 × 102 copies for the A/H3N2 plasmid, giving CT values of about 38 (Fig. 2c and d).
The study of mixtures of control plasmids with the wild-type and mutant sequences showed that the cycling probes gave CT values with each dye in a plasmid concentration-dependent manner. A ratio of 1:1 gave similar CT values for both the wild-type and the mutant plasmids, and the CT values gradually increased as the proportion of the wild-type or mutant plasmids became smaller (data not shown). The detection limit was a ratio of 1:100 for the mutant and wild-type plasmids for both subtype H1N1 and subtype H3N2.
We next evaluated the sensitivity and the specificity of the method for the detection of strains previously determined to be amantadine sensitive and resistant using clinical influenza isolates (high template concentration) and nasopharyngeal swab specimens (low template concentration). The H1N1-AS probe successfully detected the amantadine-sensitive virus (Fig. (Fig.3a)3a) and the H1N1-AR probe successfully detected the amantadine-resistant virus (Fig. (Fig.3b)3b) from both clinical isolates and nasopharyngeal swab specimens. Similar results for the differentiation (specific detection) of sensitive and resistant strains were obtained with A/H3N2 probes and clinical samples of A/H3N2 (Fig. 3c and d). The genetic sequencing results matched the cycling probe results. The average CT value observed for the isolates was low (15 to 20 cycles) compared with that for the nasopharyngeal swab specimens (25 to 35 cycles). A reaction was considered positive only when the CT value did not exceed 38 cycles after a 40-cycle PCR run.
We evaluated the specificities of the probes for human influenza virus and other common respiratory viruses (Table (Table2).2). The H1N1 probes reacted only with amantadine-sensitive and -resistant A/H1N1 strains and did not react with human influenza virus A/H3N2, whereas the H3N2 probes reacted only with amantadine-sensitive and -resistant A/H3N2 strains and did not react with human influenza virus A/H1N1. Neither probe reacted with samples containing influenza B virus, other respiratory virus-positive samples, or negative (no-template) samples. Furthermore, 99 swab specimens, including 85 isolation-positive samples and 14 isolation-negative samples, were directly tested with the cycling probe. Forty isolates each of H1N1 and H3N2 were identified, and 20 isolates each of amantadine-sensitive and isolation-positive H1N1 and H3N2 strains were also identified by these methods. Neither probe reacted with influenza B viruses, respiratory syncytial virus-positive samples, or isolation-negative samples.
A total of 1,027 primary clinical samples were collected from clinicians, and eventually, 756 (73.4%) influenza viruses were isolated by the use of MDCK cells. Of these, 672 (88.9%) isolates were influenza virus A/H1N1, 62 (8.2%) were influenza virus A/H3N2, and 22 (2.9%) were influenza B virus by HAI testing. The cycling probe assay successfully identified 663 (98.7%) of the 672 A/H1N1 viruses and 56 (90.0%) of the 62 A/H3N2 viruses. In this assay, 62.0% (411 of 663) of the A/H1N1 isolates and 100% (56 of 56) of the A/H3N2 isolates were amantadine resistant (Table (Table3).3). None of the patients received amantadine before sampling or after diagnosis.
We developed a rapid and high-throughput real-time PCR assay for the detection of the Ser31Asn mutation in the M2 gene transmembrane region in both the influenza virus A/H1N1 and the influenza virus A/H3N2 subtypes using specific florescent-labeled chimeric probes. The assay is called the cycling probe technology (3, 10). We demonstrated in the study described here that the method is highly specific for the detection of the SNP for amantadine resistance in the M2 gene and could successfully differentiate the two influenza A virus subtypes. Eventually, we showed a high frequency of occurrence of amantadine-resistant strains of both subtypes during the 2007-2008 season in Japan.
The results of the cycling probe assay described in this paper demonstrated agreement with the results of gene sequencing. Virus detection was successful by the use of both nasopharyngeal swabs and virus isolates from clinical samples, despite the differences in the virus concentration between the two types of samples. In addition, the cycling probe sets used to detect both subtypes did not show false-positive reactions with the other influenza A virus subtypes, influenza B virus, or other respiratory viruses. The sensitivity of the method for the detection of influenza virus A/H3N2, based on virus isolation, was lower than that for the detection of A/H1N1, which was due to sporadic nucleotide mismatches in the primers and/or probes for A/H3N2. At present, the proportions of viruses with a mismatch is not sufficiently significant to require a change to the sequence of the primers or the probes. Thus, our method is highly specific and is suitable for the subtyping of human influenza A viruses, along with the identification of amantadine-resistant viruses. In addition, our method successfully detected mixed populations of plasmids with mutant and wild-type sequences at a ratio as low as 1:100. This method can be used to quantify the virus loads in both clinical samples and samples tested in vitro containing a mixed mutant and wild-type virus population.
Most protocols currently employed are based on one-step reverse transcription real-time PCR (22, 33), and thus, it is common to use as the control a known concentration of RNA. However, in the cycling probe method, it is not possible to adopt a one-step real-time PCR, because it includes RNase H and the template could not achieve a sufficient length of cDNA during the reverse transcription. Other common controls are cDNA (two-step method), but plasmid preparations have higher degrees of stability and reproducibility (20, 23, 28).
Various methods has been used to examine virus isolates for their amantadine susceptibilities, such as ELISA (4), plaque reduction assay (13), the TCID50/0.2-ml method (24), PCR-RFLP analysis (21, 31), and DNA sequencing (24). However, the time to the retrieval of results by these methods may be from several hours to a few days. The advantage of our approach is that we can detect amantadine-resistant strains directly from patients' nasopharyngeal swab specimens in only 3 h: 1.5 h for RNA extraction and cDNA synthesis and 1.5 h for the real-time PCR run. Our method can be also used for dual subtyping and resistance detection if the two subtype-specific reactions are performed in parallel in a 96-well plate for sample sizes of ≥45. Pyrosequencing (6-9, 11) and DNA microarray analysis (34) were recently developed for the detection of amantadine-resistant strains. These methods have high throughputs and are rapid, just as our cycling probe technique is, but unlike our method, they have the disadvantage of the costs for the machine and the reagents, which are too high for their routine use in laboratories. Thus, our method is perhaps one of the quickest and most affordable with respect to the running cost (which is equivalent to that of the TaqMan probe method) among the methods currently available for the identification of amantadine resistance, and it is ideal for large-scale screening for resistant mutants and subtyping even with specimens with low template concentrations, such as nasopharyngeal swab specimens. Of note, the TaqMan probe real-time PCR was applied for the detection of oseltamivir-resistant influenza virus A/H1N1 possessing His274Tyr (N2 numbering) in the neuraminidase gene and used for monitoring for resistant viruses (5). We are currently developing new cycling probe sets to detect other amantadine resistance mutations in the M2 gene and the oseltamivir resistance mutation (His274Tyr) in the neuraminidase gene.
None of the patients in our study were known to have received amantadine. However, we detected amantadine-resistant viruses among both A/H1N1 (62.0%) and A/H3N2 (100%) viruses at a high frequency during the 2007-2008 influenza season. The high prevalence of amantadine-resistant A/H1N1 strains detected in Japan was one season before the 2006-2007 season, and that of the A/H3N2 strains was the previous two seasons (2005-2006 and 2006-2007) (29, 30, 32). The source of the international spread of amantadine-resistant strains is speculated to be Southeast and East Asia (6). In particular, in China, information from the Nonprescription Medicines Association shows that amantadine is available in over-the-counter formulations and is included in various cold remedies for which humans do not need prescriptions, and chicken farmers frequently add amantadine to chicken food or water for the treatment and prophylaxis of avian influenza virus and other viral diseases with low levels of pathogenicity (15). In addition, amantadine-resistant strains that could be efficiently transmitted without drug pressure were eventually generated by genetic reassortment (18).
Recently, a high rate of resistance to oseltamivir was detected among A/H1N1 viruses in several countries in different regions of the world in 2007 and 2008 (35), and combined treatment with adamantane and neuraminidase inhibitors should be considered for high-risk patients when the subtype is unknown (12). However, monitoring of influenza viruses for resistance to both drugs is crucial when combination treatment is used.
In conclusion, this study shows that the cycling probe method can specifically and rapidly detect amantadine-resistant influenza A viruses with the Ser31Asn mutation in the M2 gene directly from nasopharyngeal swab specimens and is quite useful for monitoring for drug-resistant strains to elicit the judicial use of amantadine for chemoprophylaxis and the treatment of influenza.
This study was supported by a research grant from the Kurozumi Medical Foundation.
We thank Junko Yamamoto, Kazuhide Okazawa, and Kentaro Moro of Takara Bio Inc. for technical assistance in developing the cycling probe assay. We thank clinical doctors Rika Sugai in Hokkaido Prefecture; Takashi Kawashima in Gunma Prefecture;, Isamu Sato in Niigata Prefecture; Shigeyoshi Hibi, Satoshi Ikushima, Fumitomo Fujiwara, and Kentaro Tsunamoto in Kyoto Prefecture; Tetsuo Hashida in Hyogo Prefecture; and Hironori Masaki, Yutaka Shirahige, Hidehumi Ishikawa, Satoshi Degawa, Noritika Asou, and Hironobu Kageura in Nagasaki Prefecture. We are grateful to Akinori Miyashita and Ryozo Kuwano in the Department of Molecular Genetics, Bioresource Science Branch, Center for Bioresources, Brain Research Institute, Niigata University, for utilization of their DNA sequencer. We thank Akemi Watanabe for technical assistance with virus isolation and Yoshiko Kato for intensive secretarial work.
Published ahead of print on 4 November 2009.