Respiratory syncytial virus, human metapneumovirus, and influenza virus infection in Bangkok, 2016-2017

Study design and specimens

We retrospectively tested 8,842 stored respiratory samples obtained from both in-patient and out-patient individuals of all ages with ILI who sought medical care at Bangpakok 9 International Hospital in Bangkok and collected consecutively between January 2016 and December 2017. ILI was defined as fever (>38 °C) and accompanying respiratory symptoms such as cough, sore throat, or pharyngitis. This study analyzed de-identified convenient samples and extended an earlier investigation of an ongoing influenza virus prevalence in Thailand (Suntronwong et al., 2017). Available patient information included gender and age, but not extensive clinical information nor disease severity. The Institutional Review Board of the Faculty of Medicine of Chulalongkorn University approved this study (IRB number 609/59).

Real-time RT-PCR

RNA was extracted from 200 µL of specimens using the Viral Nucleic Acid Extraction Kit (RBC Bioscience, Taiwan, R.O.C.) according to the manufacturer’s instructions. RSV and hMPV detections were performed using an in-house TaqMan-based multiplex one-step real-time RT-PCR. The primers and probes targeted the M gene of RSV and the F gene of hMPV (Table 1). The RSV probe was labeled with 6-carboxy-fluorescein (FAM) at the 5′ end and Black Hole Quencher-1 (BHQ-1) at the 3′ end. The hMPV probe was labeled with 6-carboxy-fluorescein (HEX) at the 5′ end and Black Hole Quencher-1 (BHQ-1) at the 3′ end. The reaction mixture contained 2 µL RNA, 10 µmol of each of the primers and probes, and SensiFAST Probe No-ROX One-Step reagent (Bioline, London, UK). Cycling parameters included 1 cycle for 20 min at 42 °C, initial denaturation for 3 min at 95 °C, 50 cycles for 10 s at 95 °C and 20 s at 60 °C. This assay has a limit of detection of 100 genome copies per reaction for both viruses. No cross-detections were observed between the two viruses and other respiratory viruses including influenza A and B viruses, adenovirus, enterovirus, rhinovirus, and coronavirus. Real-time RT-PCR assay of influenza A and B viruses was previously described (Suwannakarn et al., 2008). Parallel detection of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene served as an internal control. Fluorescence signals cycle threshold (Ct) was based on optimization and values ≤38 were considered positive.

Conventional RT-PCR

Samples tested positive for RSV and/or hMPV were genotyped. Complementary DNA was synthesized using the ImProm-II Reverse Transcription System (Promega, Madison, WI, USA) according to the manufacturer’s instructions. RNA and random hexamers were incubated at 70 °C for 5 min, followed by extension for 2 h at 42 °C and inactivation at 70  °C for 15 min. Amplification of the partial RSV glycoprotein (G) gene inclusive of the second hypervariable region (HVR2) and the F gene was performed using semi-nested RT-PCR as previously described (Auksornkitti et al., 2014). Cycling parameters were initial denaturation at 94 ° C for 3 min, 40 cycles of denaturation at 94 °C for 20 s, annealing at 55 °C for 20 s, elongation at 72 °C for 90 s, and a final extension at 72  °C for 10 min. Identical amplification parameters were carried out in the second-round PCR for 30 cycles. Partial F-gene of hMPV was subjected to nested-PCR as previously described (Chung et al., 2008). The PCR conditions were initial denaturation at 95  °C for 3 min, 35 cycles of 95 °C for 1 min, 55 °C for 1 min, 72 °C for 1 min, and a final extension at 72 °C for 3 min. The PCR products for RSV-A (840 bp), RSV-B (720 bp), and hMPV (750 bp) were visualized using 2% agarose gel electrophoresis and purified using the GeneAll Expin gel extraction kit (GeneAll Biotechnology, Seoul, South Korea) according to the manufacturer’s instructions. Purified PCR products were subjected to Sanger sequencing.

Sequence and phylogenetic analyses of RSV and hMPV genotypes

Nucleotide sequences of RSV and hMPV strains were aligned using ClustalW implemented in BioEdit (version 7.0.9) by comparison to the sequences previously assigned to specific genotypes (Table S1 and Table S2). Phylogenetic trees were constructed using the maximum likelihood method implemented in the MEGA6 (Tamura et al., 2013). The reliability of the tree based on the Tamura–Nei model was estimated using 1,000 bootstrap pseudo-replications. Sequences were considered the same genotype if they clustered together with bootstrap values of 70–100% (Venter et al., 2001).

Nucleotide sequences were submitted to the GenBank database under the accession numbers MH447703–MH447725 (RSV-A), MH447726–MH447818 (RSV-B), and MH447819–MH447950 (hMPV).

Statistical analysis

The association between virus prevalence and the patient age at infection was assessed using univariate analyses (SPSS software version 22.0). P-values were calculated using the Chi-squared test or Fisher’s exact test, where cell counts below 5 were used. A p-value of <0.05 was considered statistically significant.

The overall prevalence of RSV, hMPV, and influenza virus

We retrospectively tested 8,842 consecutive respiratory samples (48.5% males, age range 0–106 years). Of these, 30.5% (2,699/8,842) were positive for one or more viruses. Influenza virus was most commonly identified (17.3%, 1,528/8,842), followed by RSV (11.4%, 1,011/8,842) and hMPV (3.6%, 318/8,842) (Table 2). Influenza virus and RSV were more prevalent in 2016 than in 2017. To facilitate analysis, samples were categorized into seven groups in order to examine the distribution of viral infection relative to age (Table 3 and Table S3). Regardless of gender, the burden of RSV was greatest among children 5 years of age and younger (21.2% and 15.4% among those ≤2 and 3-5 years of age, respectively). Frequency of RSV infection appeared to decrease with increasing age and was <9% in those older than 5 years of age (p < 0.0001). In contrast, influenza virus infection was more frequently found among older individuals. Meanwhile, hMPV infection was distributed among all ages (1.7–5.7%).

Seasonal and genotype distribution of RSV, hMPV, and influenza virus

The prevalence of viral etiology of influenza-like infection differed slightly among the viruses examined. Among 1,011 RSV-positive samples, subgroup identification was possible for 488 specimens. Of these, 36.3% (177/488) were RSV-A and 66.4% (324/488) were RSV-B. RSV infection appeared most frequently in the rainy months (between July and November) with the highest annual prevalence of 37% (206/555) and 17.3% (136/784) in August 2016 and September 2017, respectively (Fig. 1A). Although RSV-A and RSV-B were equally detected in 2016, RSV-B was more frequently identified in 2017. From 318 hMPV-positive samples, subgroup identification was possible for 132 specimens. Of these, 80.3% (106/132) were hMPV-B, which was the predominant subgroup in both years (Fig. 1B). From 1,528 samples tested positive for influenza virus, there were more influenza A virus (76.2%, 1,164/1,528) than influenza B virus (23.8%, 364/1,528). In 2016, high prevalence of influenza virus occurred twice, 20.5% (59/288) in March and 34.5% (234/678) in September (Fig. 1C). The following year, peak influenza virus activity occurred in August (25.3%, 185/732). Overall, A/H3N2 accounted for 70% (815/1,164) of all influenza A virus.

Genotyping and phylogenetic analysis of RSV and hMPV

Partial G gene sequences that were randomly selected to identify the RSV genotypes showed all of the RSV-A strains (23/23) were genotype ON1 and all of the RSV-B strains (93/93) were genotype BA9 (Figs. 2A and 2B, respectively). Inter-subgroup diversity between A_ON1 and B_BA9 was relatively high (p-distance value of 2.17–2.44). In contrast, genetic variations among intra-genotype strains were relatively small (p-distance value of 0–0.073 and 0–0.071 within the ON1 and BA9 genotypes, respectively).

Partial F gene sequences were obtained from 132 of the 318 hMPV-positive specimens. Phylogenetic analysis of 132 hMPV strains identified in this study showed two main genetic lineages, A and B. Strains clustered into subgroup A2, B1, and B2, but not subgroup A1 (Fig. 3). The majority of the strains belonged to subgroup B1 (74%, 98/132), while only 6% (8/132) belonged to subgroup B2. The remaining 20% of the strains (26/132) belonged to subgroup A2. The intra-genotype strains were genetically closely related (p-distance values of 0.001–0.019), while the inter-subgroup comparisons were more diverse (p-distance values of 0.087–0.116).

Coinfections among RSV, hMPV, and influenza virus

The frequency of single versus multiple infections and the number of co-occurrences of viruses for each possible virus combination were examined (Table 4). The most common combination observed was RSV(non-typed) and influenza A H3N2 subtype (n = 68). As a percentage, the virus most often found in coinfections was RSV, which was found in 17.4% (176/1,011) of the samples, followed by hMPV (10.4%, 33/318), and influenza virus (8.5%, 130/1,528).