Clinical Characteristics of Patients with EV-D68 Infection
Among 3071 specimens, ten EV-D68 positive specimens are detected (4 cases from Zhejiang province and 6 cases from Ningxia province). The annual positive specimens of EV-D68 are 1/485 in 2017, 9/1518 in 2018 and 0/1068 in 2019. Four EV-D68 strains (YC17106/NX/CHN/2017, YC18116/NX/CHN/2018, YC18137/NX/CHN/2018, WZ17226/ZJ/CHN/2018) are isolated. Ten VP1 sequences and 4 complete genomes are successfully obtained.
The demographic and clinical characteristics of 10 EV-D68 patients are shown in Table 2. The median age of 10 EV-D68 patients is 32.7 months old (5-92m). Nine patients are male and one patient is female. Nine patients were admitted to the hospital in the summer and autumn.
Among 10 EV-D68 patients, 8 patients were diagnosed with pneumonia and 2 patients with bronchiolitis. No case was diagnosed with severe pneumonia or admitted to the ICU ward. No neurological symptoms and signs were found. All cases had no history of asthma or immunodeficiency. All patients were cured and discharged from hospital.
The patients had common clinical respiratory symptoms, such as cough (100%), dyspnea (50%), stuffy nose (10%). One case with bronchiolitis had diarrhea. Two patients co-infected with 2009H1N1 and respiratory syncytial virus, respectively.
Phylogenetic Analysis
One hundred and sixteen complete genomes of EV-D68 were downloaded from the GenBank database. Based on VP1 sequence typing, 116 VP1 sequences were obtained to construct the phylogenetic tree with 10 VP1 sequences in this study (Accession nubmer: MZ824215-MZ824224).
Ten VP1 sequences from this study share 95.7-99.5% nucleotide identity and 98.0-100.0% amino acid identity with each other. The phylogenetic tree of VP1 sequences (933bp) show that EV-D68 strains could be divided into 4 clades (A-D) and clade B could be further separated into 3 subclades (B1, B2, B3). All 10 VP1 sequences in this study belong to subclade B3 (Figure 1). Moreover, we found in VP1 sequence that clade A and D are characterized by 3-nt deletions at position 418-420 and clade D has another 6-nt insertions at position 916 (Table 3).
Four complete genomes from this study share 96.6-99.1% nucleotide identity in complete genome and 99.3-99.6% amino acid (AA) identity in coding region with each other. Three complete genomes (WZ17226, YC17106, YC18116) are composed of 7335-nt including the 5’-UTR of 701-nt, the coding region of 6567-nt and the 3’-UTR of 67-nt. The strain YC18137 has 4 more nucleotides deletions in the 5’-UTR than three genomes which is similar with the deletions of 5’UTR in the strain USA/FL/2016-19504 (Acession number: KX675261).
Compared with EV-D68 Fermon prototype strain, the nucleotide identity of four complete genomes in this study ranges from 87.2-87.5%. The nucleotide and amino acid identities ranges from 87.3-87.5% and 95.9-96.1% in the coding region respectively. The detailed results show that the nucleotide and AA identity of the structural protein region are more variable than that of the functional protein region. In the structural protein region, the identity of VP1 sequence is the lowest showing the high variabilities of VP1 sequence. In the functional protein region, the AAs of protein 2A is most variable and the AAs of protein 3B is absolutely conserved (AA identity:100%).
In order to verify whether these findings are applicable among different subclades of EV-D68 and other serotypes of enterovirus species D, four complete genomes from this study compared with 8 subclades of EV-D68 and 4 serotypes of enterovirus species D. The results are consistent with the genetic analysis compared with EV-D68 Fermon prototype strain, except that the AA identity of protein 3B presents higher variabilities in different serotypes of enterovirus species D (Table 4).
Recombination analysis
Recombination events easily occurr in enteroviruses among different subclades and serotypes. To explore it, the phylogenetic trees of each protein (VP2-4, 2A-2C, 3A-3D) were constructed. The sequences comprised each subclade of EV-D68 and other serotypes of enterovirus species D such as EV-D70/90/111/120. The result shows that 4 sequences from this study are always in the same cluster with EV-D68 subclade B3 (Figure 2A-J). To confirm this finding, the recombination analysis in Simplot was conducted. The results of similarity plot and bootscan analysis also suggest that no recombination events are found (Figure 2K-L).
Infection in Cell Line SH-SY5Y
David et al. found that EV-D68 subclade B1 strains related to AFP could infect “neuronal-like” cell line SH-SY5Y, but Fermon prototype strain couldn’t [19]. However, the subclade B3 strains which circulated dominantly nowadays were not tested the ability of infecting cell line SH-SY5Y in their study. Four EV-D68 isolates obtained in this study belonged to subclade B3 and share high identity with each other. So the strain WZ17226 was selected as a representative strain for the experiment. The titer of strain WZ17226 was first determined in RD cells in a standard 50% tissue culture infective dose (TCID50) assay and calculated by the Reed-Muench method. The results showed that the viral titer of strain WZ17226 in RD cells was ~106.0TCID50/ml by 72h post-infection. Then, the strain WZ17226 was used to infect “neuronal-like” cell line SH-SY5Y and cytopathic effect was observed. The viral titer was ~104.4 TCID50/ml in SH-SY5Y cells by 72h post-infection (Figure 3). The titer difference between the two cells was statistically significant. This results indicated that the EV-D68 isolate in this study was able to infect neurocyte, but had lower replication ability in SH-SY5Y than RD cells.
Genetic Variation Analysis
The infection in SH-SY5Y cells indicated that the strains in this study has the ability of infecting neurocyte . But the patients in this study didn’t present with any neurological symptoms and signs. In order to investgate whether the sequence differences determin the neurotropism, the complete genome sequences of 9 strains related to AFP (AFP strain) in the USA were downloaded (including 6 subclade B1 strains and 3 subclade B3 strains) and used to analyze the genetic variation of amino acids. Amino acid substitutions in the coding region may have a greater impact on the virulence of the virus. Firstly, the results of the coding region show that there are 31 AA substitutions, including 15/862 AA substitutions in structural protein sequences and 16/1331 in functional protein regions. The AA substitutions in capsid proteins are mostly distributed in VP1 and VP2, while the functional protein 2B and 3B have no AA substitution (Table 5). There was a study showed that 6 AA substitutions (M291T, V341A, T860N, D927N, S1108G and R2005K) related to neurovirulence were found in subclade B1 AFP strains [23]. In this study, these 6 AA substitutions were indeed found in subclade B1 AFP strains in 2014 but not in our four subclade B3 strains, three subclade B3 AFP strains in 2016 as well as one subclade B1 AFP strain in the USA in 2013. Secondly, the noncoding region was analyzed. There were three mutations positions in 5’UTR (127T, 262C, 339T) of subclade B1 strains and other AFP-causing enteroviruses such as EV-D70 [24]. But these three nucleotide substitutions are also not observed in our four strains and subclade B3 AFP strains. The 3’UTR may form a pseudoknot structure which may influence the replication of EV-D68. This structure appeared in all 2014 Chinese strains but not in 2014 American strains [6]. We found that this pseudoknot structure could also be observed in our four subclade B3 strains in 2018 and American subclade B3 AFP strains in 2016. These results suggested that these existing differences in complete genome may not be enough to explain the neurovirulence of EV-D68.