As a member of the genus Enterovirus of the family Picornaviridae, human enterovirus 71 (EV-71) is one of the most prevalent pathogens infecting young children and may cause outbreaks of hand-foot-mouth disease (HFMD), herpangina, and severe neurological complications [1]. More than one hundred serotypes have been identified in the genus Enterovirus [2], and classified into twelve species. EV-71 and other 23 serotypes are now members of the species Enterovirus A (http://www.picornaviridae.com/enterovirus/enterovirus.htm).

The EV-71 genome possesses approximately 7,500 nucleotides and encodes four structural capsid proteins (VP4, VP2, VP3, and VP1), and seven nonstructural proteins (2A, 2B, 2C, 3A, 3B, 3C, and 3D). Untranslated regions are located at both ends of the EV-71 genome. The capsid proteins contain antigenic sites that correspond to virus serotyping and receptor binding [1,3,4]. In previous studies, the VP1 protein was identified to induce human EV-71-specific CD4⁺ T-cell proliferation and was capable of eliciting neutralizing antibodies against EV-71 [5,6]. In addition, a VP1 mutation was shown to be the determinant of mouse adaptation and virulence [7,8]. Generally, sequencing of VP1 has been used for genotyping and phylogenetic analysis, and three genogroups (A, B, and C) were classified in EV-71, although the use of a combination of VP1 and 3D RNA polymerase gene sequences was also proposed for initial genotyping of EV-71 isolates [9]. Genogroup A contains only one strain, the prototype strain BrCr, while genogroups B and C are each divided into five subgenogroups (B1-B5 and C1-C5, respectively) [10,11]. Furthermore, some rare genogroups were identified sporadically, including subgenogroups B0, C0, D, and C2-like [9,12-15]. Although molecular biological methods have generally been used for the rapid detection and characterization of EV, virus isolation is still considered the gold standard in EV identification, and neutralization tests are used for serotyping and antigenic grouping.

EV-71 was first isolated in California in 1969, and it has been responsible for several outbreaks in different countries, most recently in Cambodia [16]. Since the discovery of genogroup A in the USA, that prototype strain was recently reported in China in 2008 [17]. Subgenogroups B1 and B2 circulated in the USA, Japan, Australia, and other countries between 1970 and 1990 [12,18]. Subgenogroups B3 to B5 have been described in Australia and the Asia-Pacific region since 1997 [10,19,20]. Subgenogroups C1 and C2 have been dominant in the USA and Europe since 1980s, C3 was first described in Korea in 2000, and C4 and C5 have been observed in the Asia-Pacific region since 1997 [10,11,14,21-23]. These epidemic patterns show that one subgenogroup may circulate in the same region for a long period (e.g., subgenogroup C4 in China) or cause outbreaks in different countries (e.g., subgenogroup C1 in the USA and Europe).

In 1998, a large HFMD outbreak caused by subgenogroups C2, B4, and C4 was reported in Taiwan. Subsequently, the subgenogroup B4 resulted in major outbreaks from 1999 to 2003, followed by subgenogroup C4 from 2004 to 2005, and subgenogroup B5 in 2008 [14,24,25]. In addition, a small number of subgenogroups, the C5, C4 and C2-like strains, were also observed in 2008 [14]. A rapid change of subgenogroups in Taiwan resulted in the possibility of severe outbreaks. Because of disease control and vaccine development needs, we analyzed genetic and antigenic diversity of the EV-71 isolates collected by the surveillance system from 2009 to 2012 and of reference strains by phylogenetic analyses and neutralization tests.

EV-71 infection, one of the common enterovirus infections in the Western Pacific region, was associated with HFMD, herpangina, and neurological diseases. Based on the surveillance system in Taiwan, there were several EV-71 outbreaks caused by different subgenogroups in recent years, for instance, subgenogroup B4 from 1999 to 2003, subgenogroup C4 from 2004 to 2005, and subgenogroup B5 in 2008 [14,24]. In addition, there were minor outbreaks of subgenogroup C4 from 2010 to 2011 and subgenogroup B5 from 2011 to 2012. The rapid changes of dominant subgenogroups and potential for outbreaks made it important to monitor the genetic and antigenic variations.

In this research, we reported the genetic and antigenic diversity of EV-71 subgenogroups in Taiwan from 2009 to 2012, including the disappearance of subgenogroup C5, an epidemic of subgenogroup C4, sporadic cases of subgenogroup C2, and genetic changes of subgenogroup B5. This type of circulation was not similar to that in other countries; for example, genogroup C was the main genogroup in the United Kingdom and mainland China for years [23,33].

Subgenogroup C5 was first described in southern Vietnam in 2005, and resulted in an outbreak with neurological manifestations [11]. In 2006, subgenogroup C5 was first isolated in Taiwan [19], and it was found in a small number of cases until 2010, based on our surveillance results. Although subgenogroup C5 was not observed in Taiwan in 2011 and 2012, the possibility of C5 re-emerging could not be ignored because B5 caused outbreaks in 2008 and re-emerged in 2011-2012.

Subgenogroup C2, the main genogroup of most EV-71 isolates from the 1998 epidemic in Taiwan, was detected in a small number of cases after 10 years. Interestingly, the C2 isolates in 2009 were clustered with those in 1998, but the isolates in 2012 were clustered with recent isolates, for example, isolates from Japan in 2010 and those from the United Kingdom in 2010. This clustering suggests that these C2 isolates evolved and spread worldwide. Additionally, the C2 isolates from Taiwan in 2012 were collected from different patients at different locations in different months, demonstrating that this subgenogroup was not a cluster event but had circulated for a period of time in a small number of cases. Moreover, a closely related phylogeny between these recent C2 isolates was suspected because of the high similarity (>98%) of the partial VP1 region.

C4, the main subgenogroup in China, has circulated and spread in recent years. There were two clusters of C4 (C4a and C4b) from 1998 to 2010, with a shift from cluster C4b to C4a in 2003–2004 [21,33]. The C4b cluster was first isolated in Taiwan in 1998, and the C4a cluster caused outbreaks in 2004-2005 [24]. In 2008, several C4 isolates (C4a) identified in Taiwan were thought to have been transmitted from China because these isolates correlated genetically with C4 strains from China but not with those isolated in Taiwan in 2004-2005 [14]. Furthermore, this C4 subgenogroup caused small outbreaks in 2010-2011 based on the present surveillance results. Additional studies are needed to clarify why subgenogroup C4 caused different epidemic patterns. For example, several large outbreaks occurred in China in specific years, but only small ones occurred in Taiwan. However, we could not exclude the effects due to the diversity of EV-71 circulation and herd immunity in Taiwan.

The subgenogroup B5 has circulated in Taiwan since 2007 and caused outbreaks in 2008, although there were two isolates sporadically detected in 2003. Towards the end of 2009, a genogroup shift (B5 to C4) and drift (B5b to B5c) of circulating EV-71 was observed. Isolate 2009-10657 was thought to be the intermediate between B5b and B5c. Although the epidemics in 2011-2012 may be partially attributed to the emerging subgenogroup B5c of EV-71, no evident divergence of antigenic relationship between subgenogroup B5b and B5c was displayed in this study. In addition to the viral factor, herd immunity and the susceptible population (infants, children, and adolescents) may be important factors in EV-71 epidemics because these groups may not have experienced the previous outbreak in 2008 [14] or produced the necessary antibodies yet. However, the genogrouping of EV-71 does not absolutely reflect the antigenicity of the strains examined [34].

An amino acid mutation at position 145 of VP1 was associated with increased virulence in mice [8]. In addition, the E145Q substitution was observed in more virulent strains [35], and this residue of genogroup B was under positive selection based on the results of evolutionary analysis with single-likelihood ancestor counting and fixed-effects likelihood methods [36]. In this study, we described three subgenogroup B5 isolates with substitutions at position 145 that were isolated from patients with AFP. These results emphasize the importance of substitution at position 145, and additional studies of receptor binding and virulence correlation with this substitution are needed. However, another genetic determinant (K244E) of mouse adaptation and virulence was not observed in this study [7].

Few studies of cross-protection among different genogroups of EV-71 were available [14,37,38]. Furthermore, the antigenicities of EV-71 strains varied based on different reports. Different genogroups displayed less than 4-fold differences in neutralizing antibody titers based on a serological survey in Japan [38]. However, antigenic differences were detected between genogroup B and C, and also between B5 and B4 viruses [36]. By using rabbit antiserum, anti-B2 could cross-neutralize B0, B1 and B2 viruses but not C1 or C2 viruses. On the other hand, anti-C1 rabbit serum could neutralize both genogroup B and C viruses [39]. Mao et al. detected different cross-neutralization capacities against different strains with the same subgenogroup (C4a) in animal sera [40]. The variations observed between different reports might be due to the different sample sizes, virus strains, sera, experimental methods, and cell lines used in the neutralization tests. However, standard protocols, including standard sera and viruses, may be used to minimize variations.

In this study, there were 35 strains and corresponding serum samples used to analyze the cross-protection and antigenic diversity among different subgenogroups (C4a, C5, B4, B5b, B5c, and C2-like) of EV-71. The subgenogroup C2-like strain exhibited evident antigenic diversity. All sera were positive against the 35 EV-71 strains, but the titers of neutralization antibody produced varied. Most of the patients infected with subgenogroup B4, C4, B5, and C5 had lower neutralizing antibody titers against the subgenogroup C2-like strain than against other strains, whereas the C2-like infected patient had broad cross-reactivity against all strains with different subgenogroups. These results were not unexpected because C2-like was a recombinant strain originating from subgenogroup C2 and B3 [14]. Since the C2-like serum was collected in the acute-phase while most other serum samples were collected in the recovery-phase, we cannot rule out the possibility that the serum collected at different stages may influence the neutralization antibody titers. However, the rabbit antisera against C2-like strain, prepared by our laboratory, were used for conducting neutralization tests for EV-71 viruses in the study, and showed similar results as using patients’ sera (Figure S1). Although C2-like viruses did not spread widely, the antigenic diversity of these viruses may lead to concern for the development of EV-71 vaccines. The key factors for selecting candidate vaccine strains include high virus titers, the ability to induce strong neutralization antibody titers, and broad cross-neutralization effects. These serum samples and the serological data obtained from the present study may be helpful to select a suitable vaccine strain and to recognize antigenic variants of EV-71.

In conclusion, first, the genetic changes of circulating EV-71 in Taiwan in recent years were described, including the re-emergence of subgenogroup B5, the short circulation of C4, the disappearance of C5, and the sporadic occurrence of subgenogroup C2. Second, the antigenicities of different subgenogroups of EV-71 were evaluated by using serum samples obtained from patients with EV-71 infections. These results may be useful for the surveillance and prevention of EV-71, and the development of vaccines.