High frequency of Polio-like Enterovirus C strains with differential clustering of CVA-13 and EV-C99 subgenotypes in a cohort of Malawian children

Genus Enterovirus is a member of the family of Picornaviridae and consists of 13 species of which 7 classify as viruses which infect humans, i.e. Enterovirus A-D and Rhinovirus A-C [1]. Enteroviruses (EV’s) are known to cause a wide variety of clinical symptoms, ranging from mild respiratory infections to invasive disease such as meningitis, encephalitis and acute flaccid paralysis (AFP). Systematic surveillance programs have led to an increased knowledge on EV circulation in large parts of the world. In Europe and the United States, EVs are found in 5-12% of clinical samples [2‐6]. Most of the EV strains found in Europe and the USA belong to species B, while species A is dominant in Asia [2, 3, 5, 7‐12]. Regular outbreaks of EVs causing severe disease and complications, such as EV-A71 and EV-D68, have been reported in various countries in North-America, Europe and Asia, as well as in Australia [13, 14]. The status of EV prevalence in sub-Saharan Africa remains largely unknown, due to incomplete sampling or data collection. The few data available report a high EV prevalence – up to 50% – with an EV-C proportion of up to 76% amongst circulating EV strains [15‐28].

Poliovirus (PV) is the most well-known EV causing AFP [29, 30]. Global vaccination programs have significantly reduced the incidence of PV infections, with PV now being endemic in only three countries (Afghanistan, Pakistan and Nigeria) [31]. PV belongs to species EV-C, and attenuated PV, as administered in the oral polio vaccine (OPV), can recombine with other strains belonging to EV-C to form vaccine-derived poliovirus (VDPV) [32‐35]. Outbreaks of such VDPVs causing polio-like symptoms have been reported in the Philippines, Madagascar, the Dominican Republic, Haiti, Cambodia, Nigeria and Egypt [32, 36‐40]. A high prevalence of EV-C in sub-Saharan Africa could increase the chances of VDPV’s arising in this continent.

The aim of this study was to provide further insights into the prevalence of EVs in children in sub-Saharan Africa, to assess the distribution of species EV-A, -B and –C, and finally to examine the genetic variability within the circulating species and genotypes. For this, we used fecal samples obtained in a case-control study conducted on children in Malawi. We report a high frequency of EVs classifiable as subgroup C of species C, a group that also contains PV.

Our study contains new information about EV epidemiology and genetic diversity within sub-Saharan Africa, which contributes to our knowledge on EV-C circulation. Since epidemiological data from sub-Saharan Africa are scarce and the circulation of EV-C is focused upon in light of the PV eradication campaign, data from older cohorts like ours are still highly relevant. We detected EV in 89.9% of fecal samples collected from children between 2002 and 2004 in two hospitals in southern Malawi. This EV frequency is higher than in previous studies from sub-Saharan Africa, that reported EV prevalence numbers ranging from 1.5% to 50% [15‐28]. Furthermore, it exceeds the 50% EV prevalence that has previously been reported in Malawi [26]. This difference might be partially explained by several factors. Firstly, in our study, we used real time PCR for detection of EV in fecal samples, whereas until recently, cell culture and virus isolation was the method most often used to detect EVs. 5’UTR PCR has been shown to detect EV from clinical specimens with a higher sensitivity than cell culture, resulting in higher yields especially for the non-B viruses [4, 6, 47]. Secondly, we hypothesize that our high EV prevalence is further explained by our relatively young population. While other studies often focus on a broader age group, EV’s are more prevalent in young age groups, when compared to older children and adults [15, 19]. Thirdly, the inclusion criteria of the SevAna study led to a higher number of participants included during the rainy season, in which malaria, a well-known cause of anemia, is highly prevalent. Possible seasonal variation in EV prevalence, much like in the Western world, might therefore have led to a higher detected prevalence in our study.

Interestingly, while we included one sample for each study participant, high EV incidence numbers have been found in studies that included multiple samples from children followed over a longer time-period. One study in Kenya showed that in a group of HIV-positive children 92% had at least one EV positive fecal sample during a 1-year study period [21]. A study conducted in Norway found that 90% of healthy children had shed EV at least once during a two-year follow-up [48].

In our study, 34% of the typed EV strains belonged to species EV-C. Although EV-C is a rather rare species in most of the world [3, 7, 8, 49‐51], it accounts for up to 76% of typed EV strains found in African populations [18, 20, 22, 25‐28]. The high proportion of EV-C subgroup C as found in our study is in accordance with findings in Cameroon and Madagascar. Furthermore, the types within EV-C that were most frequently detected in our study (CVA-13, CVA-20, EV-C99 and CVA-24) are found in approximately the same proportions in Cameroon and Madagascar [27, 52].

We saw a vast genetic diversity within EV-C subgroup C, especially within serotype CVA-13. It has been reported by others that CVA-13 strains group together in clusters (A-E) [27, 46]. In our phylogram, we could see this clustering, although cluster B and E were supported by low bootstrap values (73% and 33% respectively). Cluster D seems genetically distinct from the other clusters, with the maximum nt identity percentage compared to the other clusters falling below 75% (73.3%, table 4). Furthermore, several of the CVA-13 strains in our study did not fall within any of the clusters. Four of those strains grouped together, possibly forming a new cluster ‘F’.

For EV-C99, cluster B in our phylogenetic tree is merely supported by a bootstrap value of 5%. This is most likely a result of several of our strains grouping close to cluster B. Even so, the joint group of cluster B and our strains is supported by a bootstrap value of merely 25%.

We found PV in two of our samples (one strain PV-2 and one strain PV-3). Since the oral polio vaccine is administered at birth, this low prevalence is in accordance with our participants being between 6 and 60 months of age. The PV strains found in our study are most likely derived from children who had received a boost dose, or by secondary spread of the vaccine.

We found several strains of recently discovered genotypes EV-A119, EV-A120 and EV-C116. The prototype strains of these genotypes are derived from samples obtained years after the collection date of samples analyzed in our study [53‐55]. We found eight strains of EV-A119, whereas the oldest known reference strain dates back to 2008 [52]. EV-A119 has only been detected in three children in Cameroon, Côte d’Ivoire and Nigeria [16, 18, 56]. The large proportion of EV-A119 in our database is therefore remarkable. In contrast, EV-A71, circulating widely in Asia and Europe, and also reported in several studies in sub-Saharan Africa, was not detected in our population [7, 27, 57]. Furthermore, we found several EV-B genotypes – Echo 1, Echo 15 and types EV-B69-100 – that are rarely found in Asia, Europe and the US, but seem to be rather prevalent in sub-Saharan Africa [7, 18, 27, 28, 57].

The major limitation of our study is the sample collection taking place between 2002 and 2004. Over time, the circulation and distribution of genotypes might have changed. However, the high diversity within CVA-13 found in our study and the repeated isolation of this type throughout the whole study period is interesting and suggests continuous circulation. Furthermore, making use of sequence based typing, which was not available at the time of sample collection, gives a unique insight into an older sample set, e.g. revealing circulation of EV-A119 before the first strain was even identified.

In conclusion, we found high rates of EV prevalence in young children in Malawi and high rates of EV-C – specifically of subgroup C within EV-C. High EV-C circulation is worrying, as strains belonging to this species are able to recombine with PV, giving rise to virulent VDPV strains. Furthermore, we saw a vast genetic diversity within CVA-13. Further studies using full length sequences of our study strains should reveal whether and to what scale recombined EV-C strains containing PV fragments are circulating within our population. Moreover, these future studies will also show the exact genetic diversity within CVA-13 – focusing on the genetic variety of cluster D when compared to the other clusters, as well as the genetic diversity of cluster F.