Distinct genetic clades of enterovirus D68 detected in 2010, 2013, and 2015 in Osaka City, Japan

The first upsurge of enterovirus D68 (EV-D68), a causative agent of acute respiratory infections (ARIs), in Japan was reported in Osaka City in 2010. In this study, which began in 2010, we surveyed EV-D68 in children with ARIs and analyzed sequences of EV-D68 strains detected. Real-time PCR of 19 respiratory viruses or subtypes of viruses, including enterovirus, was performed on 2,215 specimens from ARI patients (<10 years of age) collected between November 2010 and December 2015 in Osaka City, Japan. EV-D68 was identified in 18 enterovirus-positive specimens (n = 4 in 2013, n = 1 in 2014, and n = 13 in 2015) by analysis of viral protein 1 (VP1) or VP4 sequences, followed by a BLAST search for similar sequences. All EV-D68 strains were detected between June and October (summer to autumn), except for one strain detected in 2014. A phylogenetic analysis of available VP1 sequences revealed that the Osaka strains detected in 2010, 2013, and 2015 belonged to distinct clusters (Clades C, A, and B [Subclade B3], respectively). Comparison of the 5′ untranslated regions of these viruses showed that Osaka strains in Clades A, B (Subclade B3), and C commonly had deletions at nucleotide positions 681–703 corresponding to the prototype Fermon strain. Clades B and C had deletions from nucleotide positions 713–724. Since the EV-D68 epidemic in 2010, EV-D68 re-emerged in Osaka City, Japan, in 2013 and 2015. Results of this study indicate that distinct clades of EV-D68 contributed to re-emergences of this virus in 2010, 2013, and 2015 in this limited region.

Recently detected EV-D68 strains belong to four major genetic clades (A, B, C, and D), as determined by molecular analyses of the VP1 gene [11][12][13][14]. Clade B has three subclades (B1, B2, and B3) [11,15,16] The first reported EV-D68 upsurge in Japan occurred in Osaka City, the second largest city in Japan with 2.7 million residents, in 2010; a single genetic clade was responsible (Clade C) [17]. In this study, we surveyed EV-D68 in children with ARIs since that first epidemic in 2010 and molecularly analyzed the strains detected. This is the first report of longitudinal surveillance of EV-D68 in children with ARIs in a limited geographic area in Japan since the 2010 epidemic. Here, re-emergence of EV-D68 was observed in patients with ARIs mainly in 2013 and 2015, with distinct genetic clades responsible for epidemics in each of these years.

Clinical specimens
From November 2010 to December 2015, 1,905 respiratory specimens were obtained by pediatricians from children (<10 years of age) suspected of having viral ARIs in a passive surveillance program (with influenza-diagnosed specimens were excluded) conducted in Osaka City, Japan [18]. In addition, 310 respiratory specimens were collected from hospitals and clinics from April to December 2015 for cooperative research. Surveillance in Osaka City was part of a national surveillance program for viral infectious diseases in Japan based on the Infectious Diseases Control Law. Specimens, collected as part of a passive surveillance, were residues after clinical use in hospitals and clinics. Specimens were anonymized before testing in our public health laboratory. Sample collection, detection of pathogens, and analysis of detected pathogens were approved by domestic law without additional permissions. Specifically, informed consent is considered unnecessary by domestic law and according to ethical guidelines for epidemiological research. However, sample collection for the cooperative research, which was not considered virus surveillance, required both informed consent and ethics committee review. Therefore, written consent was obtained from parents and guardians of the children whose specimens were included in the study. This part of the study was approved by the Ethics Committee of the National Institute of Infectious Diseases, Tokyo, Japan (no. 532). , and human rhinovirus, as described previously [18]. For enterovirus-positive specimens, the VP1 or VP4 region was amplified by conventional reverse transcription-PCR. VP1 was amplified by PCR using primers 224 (5 0 -GCIATGYTIGGIACICAY RT-3 0 ) and 222 (5 0 -CICCIGGIGGIAYRWACAT-3 0 ), followed by a nested PCR using primers AN89 (5 0 -CCAGCACTGACAGCAGYNGARAYNGG-3 0 ) and AN88 (5 0 -TACTGGACCACCTGGNGGNAYRWACAT-3 0 ) [19]. VP4 was amplified using primers OL68-1 (5 0 -GGTAAYTTCCACCACCANCC-3 0 ) and EVP4 (5 0 -CTACTTTGGGTGTCCGTG TT-3 0 ) [20]. PCR products were subjected to direct sequencing, and EV-D68 was identified by phylogenetic analysis.

EV-D68 genome sequence determination
The complete or nearly complete genome sequences of EV-D68 strains from Osaka in 2013 and 2015 were determined as follows: cDNA was synthesized using primers (random hexamer or sequence-specific) and SuperScript III; then, EV-D68 genomes were amplified in nine fragments using the primers shown in S1 Table. PCR was performed using KOD-plus Neo DNA Polymerase (TOYOBO, Tokyo, Japan) with the following conditions: 94˚C for 2 min, followed by 35-45 cycles at 98˚C for 10 s, 55˚C for 30 s, and 68˚C between 30 s and 4 min according to the manufacturer's instructions. PCR products were purified using a QIAquick PCR Purification Kit (Qiagen Inc.). End-specific nucleotide sequences were determined using the 5 0 RACE system (Thermo Fisher Scientific) and 3 0 RACE with primer TX30SXN [21]. Nucleotide sequences were determined using an ABI 3130 Genetic Analyzer (Thermo Fisher Scientific).

Molecular analysis
Available nucleotide sequences were obtained from GenBank. The nucleotide sequences of the 14 EV-D68 strains identified in this study were deposited into the DDBJ/EMBL/GenBank database (accession nos. LC068709-LC068711, LC107890-LC107897, LC107899-LC107901). Complete or near-complete genome sequences of 58 strains of EV-D68 obtained worldwide were used for the phylogenetic analyses. In addition, sequences of VP1 (927 nt, corresponding to nucleotide positions 2,389-3,315 of the Fermon strain) from 110 strains were used to construct the phylogenetic tree. Sequences were aligned in MUSCLE, and the phylogenetic tree was constructed using the maximum likelihood method with 1,000 bootstrap replicates in MEGA 7.0 [22]. Genetic clades were applied based on a previous report [11].

Phylogenetic analysis of the VP1 gene
The phylogenetic analysis was performed using available EV-D68 sequences obtained worldwide and Osaka strains obtained in 2010, 2013, and 2015. One EV-D68 strain detected in 2014 could not be amplified. Osaka strains detected in 2010, 2013, and 2015 belonged to Clade C, Clade A, and Clade B (Subclade B3), respectively (Fig 2). Clade names were assigned according to an earlier report [11].   (Fig 3). This genogrouping was similar to that based on the VP1 analysis (Fig 2).

Amino acid alignment of the partial VP1 region
To analyze VP1 in detail, amino acid sequences were aligned and examined for substitutions. BC and DE loops located on the surface of the VP1 protein play important roles in the antigenicity of HEV, and two specific amino acid substitutions (N642D, S647A) were observed in the BC loop in Clade B, and one amino acid deletion (N692) was observed in the DE loop in Clade A and Clade D. In addition, a two-amino-acid insertion (arginine-leucine) was also observed, but only at the 859-860 aa position in Clade D strains (Fig 4). Specific amino acid differences between clades or subclades in the same clade were also observed in the BC and   Re-emergences of EV-D68 in Osaka City occurred mainly from June to October (summer to autumn) in 2013 and 2015, similar to its first appearance in 2010 [17]. Other epidemiological data on EV-D68 in Japan, Yamagata Prefecture, showed an upsurge of EV-D68 between summer and autumn in 2005-2010 [12]. Based on the available data, summer through autumn is believed to be the EV-D68 season in Japan. Multi-year studies on EV-D68 suggest that the EV-D68 epidemic season differs by country, with peak seasons of August-January in China [23], July-December in Australia [24], October-December in Germany [25], and July-October in the Netherlands [26], and variable seasons in Thailand and Hong Kong [15,27]. Data from a multi-year study conducted worldwide are needed to precisely define EV-D68 seasonality. Re-emergence of enterovirus D68 in Osaka, Japan The phylogenetic analysis of complete or nearly complete genomes and VP1 sequences revealed that Osaka strains detected in 2010, 2013, and 2015 belonged to three genetic clades. Earlier reports describing receptor binding activity of EV-D68 strains in all genetic clades showed that distinct genetic clades have different antigenicities [28]. Therefore, differences in the prevalent genetic clade of EV-D68 each year in Osaka City might have contributed to its re-emergence in this limited geographic area. There were high similarities in VP1 nucleotide and amino acid sequences of Osaka strains detected in 2010, 2013, and 2015. However, lower similarities were observed for strains detected in 2010 vs. 2013, 2013 vs. 2015, and 2010 vs. 2015. This low similarity might be associated with differences in EV-D68 antigenicity. Lower herd immunity against strains in specific genetic clades of EV-D68 might have facilitated the rapid spread of a clade of EV-D68 that was not reported to be prevalent in the limited geographic area. Intriguingly, co-circulation of all three clades of EV-D68 was observed in Yamagata Prefecture, Japan, in the 2010 epidemic [12]. Based on these results, EV-D68 dynamics might differ in the same period in different parts of the country. One might infer that herd immunity against different clades of EV-D68 is associated with their prevalences in a limited geographic area.
Certain amino acid changes and deletions in the BC and DE loops of the VP1 protein were observed between the clades of EV-D68. Moreover, characteristic amino acid substitutions were observed in a comparison of strains in subclades of the same clade. Considering the effects of the BC and DE loops on the surface of the VP1 protein, EV-D68 strains belonging to different subclades of the same clade might have different antigenicities.
A partial deletion in the 5 0 UTR was first reported based on our previous work [17]. Some characteristic partial deletions, relative to the prototype Fermon strain, were observed in the 5 0 UTRs of EV-D68 strains in three clades. However, we did not analyze functional differences or the physiological importance of the deletion on EV-D68. To date, there has been no published report on the function of the EV-D68 5 0 UTR. The 5 0 UTR of an enterovirus contains an internal ribosome entry site (IRES) that is associated with translational efficiency and virulence [29][30][31]. Deletions in EV-D68 strains appeared to be in the region between the IRES and the open reading frame. Analysis of this deletion may yield important information about EV-D68. The 5 0 UTR of EV-D68 strain CA/AFP/11-1767 in clade B showed a deletion similar to that of strains in Clade A. Du et al. analyzed 2014 North American EV-D68 outbreak using VP1, which revealed that CA/AFP/11-1767 strain was phylogenetically distant from other subclade B1 strains [32]. Their findings might partially explain one amino acid substitution (T650) in the BC loop of VP1 in strain CA/AFP/11-1767. However, the reason for the Clade A-like nucleotide deletion in the 5 0 UTR of strain CA/AFP/11-1767 remains unknown.
Tan et al. reported intersubclade recombination events in the VP2 region between EV-D68 strains in clades B1 and B2 [33]. However, no intersubclade recombination was observed in EV-D68 strains from Osaka City in 2015 (data not shown).
During the study period, four EV-D68-positive patients without signs of ARI were also found. These patients showed the following: AFP (n = 1 in 2013), fever (n = 1 in 2015), unknown fever (n = 1 in 2015), and rash (n = 1 in 2015) from respiratory specimens (Table 1). Although EV-D68 was detected in specimens from several individuals without ARI symptoms, whether EV-D68 was directly associated with specific clinical signs remains uncertain.
EV-D68 in cases of AFM were reported in the US from 2012 to 2014. All of these strains belonged to Subclade B1 [5]. In addition, EV-D68 in cases of AFP/AFM were reported in Japan in 2015 [34,35]. However, no EV-D68-positive AFP/AFM cases were observed in Osaka City in 2015. Only one Clade A strain in 2013 was detected from a case of AFP (692-Osa-kaC-JPN-2013). A few reports describing EV-D68 from patients with AFP/AFM were published in 2010 and 2013 in Japan, but genetic clades of the infecting strains were not determined [34]. Two AFM cases in France and Norway were associated with Subclade B2 EV-D68 infection [6,9], as were two neurological cases in Kenya with Subclade A1 [36] described in an earlier report [33]. Considering previous reports of neurological conditions associated with the detection of strains in various clades of EV-D68, only Subclade B1 strain was found predominant in AFM cases. It is not a major cause of AFM.
Overall, results of the analysis of complete or near-complete genome sequences suggest that the presence of strains in distinct lineages of EV-D68 in 2010, 2013, and 2015 might have contributed to the re-emergence of EV-D68 in this limited region because of lack of herd immunity. To analyze EV-D68 further, near-complete genome sequences of earlier reported strains from cases of ARI, AFM, encephalitis, central nervous system disorders, and other diseases must be determined. Recently, a novel genetic lineage of Subclade B3 was identified that now includes 2015 Osaka strains. Based on our phylogenetic analysis of VP1 (Fig 2), there are some strains (2013-0720-6-CHN, 2011-21186-CHN, CA/AFP/v12T00346-USA-2013, CU171-THA-2011, CU134-THA-2011) that do not clearly cluster in the currently recognized genetic subgroups in Clade B. Therefore, novel genetic subclades may be identified in the future.
To date, it is uncertain whether specific clades of EV-D68 are associated with specific clinical signs. Accumulating epidemiological information and analyzing sequences of EV-D68 strains are important to understanding the recent upsurge in cases of EV-D68 infection and might suggest associations between specific genetic clades and clinical signs.
Supporting information S1 Table. Nine primer pairs used to amplify the EV-D68 genomes. (XLSX)