Background
Enteroviruses (EVs) are members of the genus
Enterovirus in the family Picornaviridae
, order Picornavirales. There are 13 species within the genus, four (EV-A to EV-D) of which (alongside the Rhinoviruses) have been repeatedly found to infect humans. The enterovirus capsid is a non-enveloped icosahedron with a diameter of 20–30 nM. It encloses a positive sense, single stranded RNA genome of ~7500 nt. The genome has untranslated regions (5′ and 3′ UTRs) flanking a single one open reading frame (ORF), whose polyprotein product is auto-catalytically cleaved into eleven proteins; four structural (VP1 – VP4) and seven non-structural (2A – 3D) proteins. Amplification of the 5′-UTR and/or the VP1 region can be used to detect the presence of enteroviruses [
1‐
5], and the nucleotide sequence of the VP1 gene is used to identify enterovirus isolates [
1‐
6].
Poliovirus, the best studied member of the genus
Enterovirus and the etiologic agent of poliomyelitis belongs to species C within the genus. Humans remain the only known host of poliovirus, thus suggesting feasibility of its eradication. Consequently, in 1988, the World Health Assembly (WHA) resolved to eradicate poliomyelitis by the year 2000 [
7], and the Global Polio Eradication Initiative (GPEI) was established. Courtesy of the GPEI’s activities, by year 2015 indigenous poliovirus transmission has been interrupted globally except in Pakistan and Afghanistan (
www.polioeradication.org). This success has been the result of effective vaccination and intensive surveillance. Two poliovirus vaccine preparation (Oral Polio Vaccine [OPV] and Inactivated Polio Vaccine [IPV]) are currently being used by the GPEI and both are very effective [
8]. However, due to the reversion of OPV, as part of the ‘end game’ strategy, the GPEI is tilting towards IPV as we approach the final phase of polio eradication [
9].
Coupled with the vaccination effort is a very effective surveillance network that looks for polioviruses in both sewage-contaminated water (Environmental Surveillance [ES]) and children below the age of 15 years diagnosed with AFP. The ES strategy searches for enteroviruses in sewage-contaminated water due to the fact that all enterovirus infected individuals shed the virus in large amounts in faeces for several weeks and in turn into sewage and/or sewage contaminated water [
10] . Therefore, ES is very sensitive and can detect enterovirus isolates from both symptomatic and asymptomatic individuals. The demerit of ES based strategy is that, alone, it cannot differentiate which isolates are associated with clinical manifestations and hospitalisations. On the other hand, AFP surveillance finds enteroviruses associated with a clinical manifestations and consequent hospitalisation. However, considering that AFP surveillance detects only the <10% of enterovirus infections with clinical manifestations, the inability of AFP surveillance to see beyond the tip of the iceberg is the strength of the ES surveillances strategy. Therefore, combining both strategies better illuminates our understanding of the epidemiological and evolutionary trajectory of enteroviruses, particularly with respect to pathogenicity. Hence, the reason the ES-AFP surveillance strategies are combined by the Global Polio Eradication initiative (GPEI) in some countries [
11,
12].
As part of the surveillance network are about 150 laboratories globally (Global Polio Laboratory Network [GPLN]) that use the RD-L20B isolation protocol [
11,
12] for poliovirus isolation. The RD-L20B isolation protocol is predicated on two different cell lines, RD (from a striated muscle cancer) [
13] and L20B (a murine transgenic L cell that expresses the poliovirus receptor and is consequently permissive and susceptible to polioviruses) [
14,
15]. As a by-product of the poliovirus surveillance programme, the GPLN recovers several non-polio enteroviruses (NPEVs) on the RD cell line.
The earliest molecular epidemiology study documenting NPEV from AFP cases in Nigeria was carried out using the RD-L20B protocol between 2002 and 2004 [
16,
17]. Subsequent studies [
18‐
22] generating nucleotide sequence data on enterovirus diversity in Nigeria have been largely ES based. The only exception is a recent study [
23] on enterovirus diversity in healthy Nigerian children, in which a cell-culture independent enterovirus detection strategy was employed. Consequently, in this study we revisit NPEVs from AFP cases in Nigeria in an attempt to revise their diversity in individuals with this clinical condition in the region.
Discussion
In this study, we document nucleic acid sequence data for twenty-seven (27) different enterovirus types circulating and particularly present in children below the age of 15 years diagnosed with AFP in Nigeria in 2014. It is important to mention that prior this study, no nucleic acid sequence data existed in Genbank for nine (9) of these enterovirus types from Nigeria. To be precise, to the best of the authors’ knowledge, nucleic acid sequence data for Nigerian strains of E17, CV-B2, CV-B4, EV-B97, EV-B80, EV-B73, EV-B93, EV-C99 and EV-A120 are being reported for the first time. It is however, our opinion, that the fact that these molecular sequence data are being reported for the first time does not imply new introduction of these types into the region. Rather, we believe that these types had probably been around for a long time. Hence not detecting them might reflect the lack of interest in NPEVs because the global effort focused on eradicating polioviruses. However, as the goal of poliovirus eradication nears [
9], there might be an upsurge of interest in NPEVs in a bid to better understand enterovirus biology and association with varying clinical conditions.
In this study, of the 69 isolates sequenced, 95.65% (66/69) belong to HEV-B representing 88.9% (24/27) of enterovirus types identified. (Table
2). This is consistent with the findings of other studies from the region [
16,
17,
27] predicated on the RD-L20B algorithm, irrespective of whether healthy children [
27] or those diagnosed with AFP were investigated [
16,
17]. However, it was recently shown that when cell-culture bias is bypassed by the direct detection of enteroviruses from stool specimen, EV-As appear to be the most preponderant in the stool of healthy children from the region [
23]. Furthermore, recent studies [
20,
21] showed that most CV-A13s (EV-Cs) circulating in the region selectively replicate with evident cytopathology, and can be isolated on MCF-7, but not in RD cell line. Also, when the same sample was simultaneously inoculated into RD and MCF-7 cell lines, EV-Bs and EV-Cs respectively are specifically recovered on the two cell lines [
21]. In fact, we recently observed [
28] that when AFP samples that are negative for enteroviruses using the RD-L20B enterovirus detection algorithm [
11] are subjected to the recently recommended [
6] direct detection RT-snPCR algorithm described by Nix et al., [
5], EV-Cs form the majority of enteroviruses detected. In addition, we also recently observed (unpublished) that about 50% of stool suspension from children with AFP in the region that yield EV-Bs in RD cell line also have an EV-C member present in the faecal suspension that was not detectable by the cell line. Put together, the preponderance of EV-B depicted by the results of this study and previous RD-L20B based studies from the region and globally should not be interpreted as an unbiased picture of the diversity landscape of enteroviruses in the AFP samples that yielded the isolates analyzed. Rather, it should be correctly viewed as the landscape as seen through the bias of RD cell culture.
Though in this study we describe enteroviruses present in the stool of children diagnosed with AFP, the findings of this study show the significance of merging ES and AFP data. Phylogenetic analysis (Figs.
3,
4 and
5) suggests that representatives of some of the lineages of EV-B75, E19 and E7 detected in the AFP cases had been previously detected through environmental surveillance. Interestingly, all the sequences from Nigeria in Figs.
4 and
5 except those from this study and the 2002–2004 sequences were from environmental surveillance. The ES data rightly show that more lineages are circulating in the population than depicted by the AFP data reported in this study. This thereby emphasizes the power of ES to illuminate our understanding of the diversity of enterovirus types and lineages circulating in the population. This will, consequently, further enable us to better understand the evolutionary trajectory of these enterovirus types and detect their silent circulation especially in the absence of clinical manifestations.
Surveillance of enterovirus diversity among healthy children can also increase the power of the ES-AFP surveillance strategy. For instance, it was previously shown [
23] that other enterovirus types not detected in this study (e.g. EV-A71 and several CV-A types) were also present and circulating in Nigeria in the same year, 2014. In addition, we had shown [
23] that the EV-B80 detected in a healthy Nigerian child belonged to the same lineage as those detected in children diagnosed with AFP in this study. We further showed [
23] that the EV-C99 lineage found in a healthy Nigerian child is different from that described in this study. Thus, though not suggested or included in the GPEI surveillance algorithm, enterovirus surveillance in healthy children is useful and can significantly complement the ES-AFP strategy of the GPEI.
The baseline nucleotide sequence data for CV-B3 and E19 circulating in the region were generated in 2002 and 2003 respectively [
16,
17]. Hence, the data presented in this study is a re-sampling of circulating strains of these types after over a 10-year period. For both types, the strains circulating when the baseline data was generated appear to have been completely replaced (Figs.
2 and
4). On the other hand, E7 and EV-B75 for which baseline sequence data from the region was generated in 2010 [
18] and 2012 [
22] respectively, members of the baseline lineage were still detected, however, alongside new lineages.
In some instances, it appears the variation in the population is seeded from another population. For example, as shown for CV-B3 and E7 (Figs.
2 and
5), it appears that in both instances, strains from Asia were imported into Nigeria and subsequently detected in the faeces of children diagnosed with AFP. These suspected importations corroborate what is known about poliovirus global circulation [
8,
29]. What is not clear is why, as suggested in the regional confinement hypothesis [
30], most non-polio enterovirus lineages detected in sub-Saharan Africa are yet to be detected and described in other world regions. Also interesting is the observation that enterovirus lineages found in Nigeria are also present and circulating in Niger (Figs.
2,
4 &
5). Considering the apparent porosity of borders in the Chad basin, this is not surprising. It is however crucial to note that the lineages circulating in Nigeria and Niger appear to be different from those recently shown [
31] to be circulating in Gambia, Guinea and Senegal (Figs.
3,
4 &
5). Hence, though all the mentioned countries are in sub-Saharan Africa and more interestingly, West-Africa, it appears there might also be a level of sub-regional restriction of circulation. Characterization of more NPEV isolates will definitely show whether the observed is a true biological phenomenon or relic of paucity of data.
In this study, three (one 5′-UTR and two VP1) different enterovirus detection assays were used. The 5′-UTR screen is used to screen isolates in a bid to determine if they are enteroviruses. This is done because the 5’untranslated region (UTR) contains the internal ribosome entry site (IRES). On entry into the cytoplasm of any susceptible cell, to initiate translation of the single open reading frame (ORF) in the enterovirus genome, the ribosome must be assembled on the IRES in the 5′-UTR. The consequent conservation of this enterovirus genomic region is capitalized upon for enterovirus detection using the 5′-UTR assay. However, because of the conserved nature of the 5′-UTR its sequences do not provide the resolution needed to determine enterovirus type. On the other hand, the VP1 gene encodes one of the virus structural proteins. Studies [
1‐
6] have established a correlation between the nucleotide or amino acid sequence of VP1 and enterovirus types determined using neutralization assays. The fact that there are over two hundred and fifty (250) enterovirus types (
www.picornaviridae.com) gives a good impression of how variable the VP1 structural protein is. Hence, irrespective of how degenerate the primer combinations used, it might not be surprising if any one Panenterovirus RT-PCR screen fails to amplify the VP1 gene of any enterovirus of interest. Against this backdrop, enterovirus identification assays (including the poliovirus identification algorithm in use by the Global Polio Eradication Initiative [GPEI]) are built around first detecting the conserved enterovirus genomic region (5′-UTR) and subsequently subjecting isolates detected by this screen to VP1 assays [
4‐
6].
Overall, six (6/96) of the isolates analyzed were positive for the 5′-UTR screen but negative for the VP1 screen. These isolates might therefore be enteroviruses that have VP1 primer binding sites that are too divergent to be bound by the primers used in this study. Another seven (7/96) isolates on the other hand were negative for the 5′-UTR screen but positive for the VP1 assays. Should we have used the 5′-UTR result as the basis for selecting isolates subjected to the VP1 screens, all these isolates would have been missed. It is currently not clear why these isolates were negative for the 5′-UTR screen considering they replicated in culture and are consequently viable strains. However, it is important to mention that recently, a new enterovirus 5′-UTR region was described that has a large deletion overlapping the region amplified in this assay [
32‐
35]. Though described in EV-C isolates, it is not clear whether such constructs are present in members of other enterovirus species and whether such constructs would be functional when transferred to other species through recombination in the 5′-UTR region. Whatever be the case, the 5′-UTR primers used in this study could not detect these isolates. Consequently, the results of this study suggest that coupling the 5′-UTR and VP1 assays in a way that ensures both are independent and equally important for enterovirus identification might provide the added sensitivity required to find some divergent types. Particularly, it suggests that being positive for the 5′-UTR assay should not be the basis for subjecting isolates to the VP1 assays.
In this study, there were different groups of ‘untypable’ isolates (Table
1). The first group were those positive for the 5´-UTR screen but negative for the VP1 screen and have been addressed above. The second group were those positive for the VP1 assay but for which the eletropherogram could not be exploited due to multiple peaks. These are likely to come from cases where the children in question were co-infected with more than one type of enterovirus. This is not unusual and, as previously mentioned, we have more recently observed (unpublished) that in about 50% of cases, stool samples from children with AFP in Nigeria contain more than one enterovirus type and/or species.
The third group were those negative for both the 5′-UTR and the VP1 assays. Considering we detected isolates that were positive for the 5′-UTR screen but negative for the VP1 screen and vice-versa, it is not difficult to conceptualize the possibility that enterovirus isolates might exist that are negative for both assays. However, currently, this is only a conjecture. Since RD cell line can also support replication of other enteric viruses like the adenoviruses [
36], these third group of ‘untypables’ might not necessarily be enteroviruses but could be other viruses for which the RD cell line is both susceptible and permissive.