Molecular characterization of enteroviruses in patients with acute respiratory infection and influenza-like illness in the Central African Republic revealed mainly polioviruses and other enteric enterovirus types

Abstract

Enteroviruses (EVs) that infect humans cause a broad range of pathologies, notably neurological, respiratory, cutaneous and enteric diseases, leading to morbidity and mortality, especially in young children. These pathologies particularly affect countries in Asia and Sub-Saharan Africa. In the last decades, some EVs with a marked respiratory tropism have emerged as a source of acute respiratory illnesses: EV-D68 has been associated with outbreaks of severe respiratory distress and acute myelitis since 2014, while several respiratory genotypes belonging to the species Enterovirus coxsackiepol were discovered in the 2010s. Very little is known about the latter, and no data exist on their circulation or clinical impact in the Central African Republic. In an attempt to bridge this gap, a retrospective study was carried out on nasopharyngeal samples collected as part of influenza surveillance at five sentinel sites in the Central African Republic in 2019. A total of 178 samples negative for influenza viruses were pooled and screened for the detection of EVs (rhinoviruses excluded). A positive result was obtained for 17 out of 20 pools. Genomic amplification was attempted by using generic primers against EVs (rhinoviruses excluded), and the virus genomes were partly sequenced by Illumina technology. A total of 46 contigs covered the VP1 region. The corresponding viruses belonged to several virus types of three species: E. coxsackiepol (14/26, 58.3%), Enterovirus betacoxsackie (9/26, 34.6%) and Enterovirus alphacoxsackie (3/26, 11.5%). Unexpectedly, no respiratory virus types (such as EV-D68, EV-C104, EV-C105, EV-C109, EV-C117 or EV-C118) were detected. Of note, polioviruses from vaccine origin were detected in eight pools (40.0%), supporting the notion that respiratory specimens may serve as a source for poliovirus re-introduction after eradication.

In conclusion, we did not detect any respiratory EVs, a finding that most likely reflects a low baseline prevalence of these viruses during inter-epidemic periods. Moreover, we confirmed that several EV types, whose primary replication site is the gastrointestinal tract, are also capable of replicating in the oropharynx, polioviruses included, which constitutes a concern for mitigating the risk of poliovirus re-introduction in the post-eradication era.

Data Availability

The non-polio enterovirus nucleotide sequences generated during the current study have been deposited in the GenBank database with the accession numbers PX711930–PX711947. The nucleotide sequences of polioviruses were not submitted to GenBank since they have very few differences with those of the vaccine strains. They are available upon request.

Introduction

Respiratory infections cause high mortality and morbidity worldwide, especially among children under 5 years of age and individuals aged 60 years or older [1]. In 2021, there were an estimated 2.5 million deaths related to acute respiratory illnesses worldwide, with the highest mortality rates observed in sub-Saharan countries and especially in the Central sub-Saharan sub-region where the mortality rate reaches around 100 deaths/100,000 per year vs around 30/100,000 globally [2]. These infections took on an even greater significance with the arrival of COVID-19, which caused millions of deaths. In addition to very common respiratory viruses such as influenza viruses and respiratory syncytial virus, members of the genus Enterovirus (Family Picornaviridae) are also responsible for numerous respiratory symptoms, ranging from a simple irritant cough to a serious respiratory distress syndrome [3,4]. These viruses, generically named enteroviruses (EVs), are small, naked, positive-strand RNA viruses and have been sampled in humans and other mammals. EVs are phylogenetically classified into 15 species, each of which harbours a more or less broad repertoire of virus types. Most EVs that infect humans fell into two major groups depending on their main site of replication. Some EVs, such as polioviruses (PVs), have an enteric tropism: their particles resist acid pHs, which allows them to go through the stomach and reach the gut that is their main replication site. They are excreted in the faeces and are mainly transmitted by the faeco-oral route. Other EVs are respiratory viruses adapted to the replication in the respiratory tract. Rhinoviruses, which split into three different species (Enterovirus alpharhino, Enterovirus betarhino and Enterovirus cerhino), are etiologic agents of the common cold. Member of the species Enterovirus deconjunti, EV-D68, has caused severe forms of ARI around the world since 2014 [5], particularly in the USA [6,9]. More recently discovered, some members of the species Enterovirus coxsackiepol (generically named ‘enteroviruses C’, EV-Cs) are also respiratory viruses. Five EV-C virus types with respiratory tropism have been discovered to date (EV-C104, EV-C105, EV-C109, EV-C117 and EV-C118) [4,10], but little is known about these viruses that are refractory to isolation in cell cultures.

Previous investigations have characterized EVs circulating in different sub-Saharan countries, including the Central African Republic (CAR). These surveys have documented a broad spectrum of EV virus types, with several notable features. First, they have consistently reported a relatively high abundance of EV-C species when compared with data from other geographic regions [11,13]. Second, certain virus types or lineages appear to be confined to sub-Saharan Africa, for instance, EV-A71 strains belonging to genogroup E and EV-D94 [14,15]. A major limitation of these studies, however, is their frequent focus on EVs detected in stool and wastewater specimens. Although respiratory viruses can be molecularly identified in such matrices, the sampling strategies were not optimized for the detection of EVs that do not replicate efficiently in the gastrointestinal tract, especially when culture-based isolation protocols were employed. Consequently, EVs with a primarily respiratory tropism may have been under-represented in the existing datasets. In particular, data about respiratory EV-Cs circulating in sub-Saharan Africa consist of only a few specimens reported in Gambia and the Democratic Republic of the Congo [16,17]. Similarly, while EV-D68 has been the focus of many studies since its emergence in 2014, the genomic data available regarding this virus in Africa concern a limited number of countries [18,19]. The present study was designed to help bridge this gap by characterizing EVs in respiratory samples collected in the CAR, with rhinoviruses excluded. The study had two principal objectives: (i) to catalogue the EVs present in respiratory samples in CAR and (ii) to determine whether respiratory EV-Cs are circulating in this country. Because we were specifically searching for respiratory EV-Cs and EV-D68, rhinoviruses were not targeted. Besides, this study also aimed to assess the prevalence of PVs in respiratory specimens collected in a country that employs the oral poliovirus vaccine (OPV), thereby evaluating the biosafety risk posed by such samples in the context of PV eradication, which mandates the containment or destruction of any material potentially harbouring infectious PV [20]. For this purpose, pools of nasopharyngeal samples collected in CAR in the framework of influenza surveillance and previously tested negative for influenza viruses were screened with a commercial test towards EVs. After genomic amplification and Illumina sequencing, the EVs were molecularly typed.

Methods

Study design

We conducted a retrospective study on nasopharyngeal swabs sampled in CAR in the framework of the national influenza surveillance. This surveillance relies on respiratory samples collected from patients with influenza-like illness (ILI) or severe acute respiratory infection (SARI) as per WHO’s criteria [21]. Samples are collected in sentinel sites and sent to the WHO Collaborating Centre for Influenza at the Institut Pasteur de Bangui for screening. For this study, a total of 178 nasopharyngeal samples were selected from the database of the WHO Collaborating Centre for Influenza, by random selection of samples collected from January through December 2019 and found negative for influenza viruses [22]. The samples were from five sentinel sites. Two sites are in Bangui, the capital city: the Paediatric University Hospital Centre of Bangui and Saint-Paul Health Centre. Three sites are situated in three distinct administrative regions: the Pissa Health Centre (70 km east of Bangui), Bossembélé Hospital (160 km northeast of Bangui) and the Boali Health Center (90 km northeast of Bangui). The selected samples were pooled (~100 µl per sample) according to locality and month of collection prior to RNA extraction (Table 1).

Table 1. Composition of the pools with screening and amplification results.

Site	Nasopharyngeal swabs	Pools	Positive EV assay	Positive EV genome amplification
Paediatric hospital	24	3	2	2
Saint-Paul	45	5	5	3
Pissa	52	6	5	4
Bossembélé	28	3	2	2
Boali	29	3	3	1
Total	178	20	17	12

Molecular screening

RNAs from pooled nasopharyngeal samples were extracted using the QIAamp Viral RNA Mini kit (Qiagen, USA) according to the manufacturer’s instructions. EV genomic RNAs were detected using Fast Track Diagnostics Respiratory pathogens 21 (FTD21, Siemens Healthineers GmbH, Germany). The panel comprises 5 multiplex PCR assays detecting 18 viruses; however, only the EV-targeting assay was executed, as stipulated by the project’s ethical clearance. Real-time PCR was performed using 10 µl of RNA in a final volume of 25 µl following the manufacturer’s instructions. The reaction was carried out on an ABI 7500 Real-Time PCR System (Applied Biosystems). Following the indications for interpretation of results, samples with a cycle threshold (Ct) ≤35 were considered positive, samples with a Ct between 35 and 37 were considered weakly positive only in the presence of a clear amplification curve, and samples with no Ct or a Ct >37 were considered negative.

EV genome amplification and sequencing

The capsid-encoding region was amplified by using a strategy already described [23,24]. This strategy relies on primers that target conserved genetic sequences in the 5′UTR and in the cis-replicating element (cre) located within the 2C-encoding region. First, virus RNA was turned into cDNA by an RT step using the Maxima H Minus First Strand cDNA Synthesis Kit (Thermo Scientific, ref K1652) with an Oligo(dT)18 primer: 4 µl of extracted RNA was mixed with 1 µl of dNTPs (10 mM each), 1 µl of primer (0.5 µg µl⁻¹) and 14 µl of nuclease-free water. The mixture was incubated at 65 °C for 5 min and then on ice for 2 min; after adding 4 µl of RT buffer (5X) and 1 µl of Maxima H Minus Enzyme Mix, incubation was performed at 50 °C for 30 min and 85 °C for 5 min. The cDNAs were used to amplify the 5′ half of the genomes (~4,300 nt) by using the Phusion High-Fidelity DNA polymerase (Thermo Scientific) and the primer pair UG52 (5′-CAAGCACTTCTGTTTCCCCGG-3′) [25]/EV-cre-R (5′-TCAATACGGTGTTTGCTCTTGAACTG-3′) [24]. PCR amplification was performed using a thermocycler with the following protocol: 98 °C for 30 s, 35 cycles at 94 °C for 15 s, 55 °C for 30 s, 72 °C for 3 min and a last step at 72 °C for 10 min, ending with 2 min at 4 °C. When no amplicon at the expected size was observed on BET-stained agarose gels, a semi-nested PCR was performed using the same thermocycler profile with the primer pair UG53 (5′-TGGCTGCTTATGGTGACAAT-3′) [25]/EV-cre-R. The products of the PCR or semi-nested PCR were purified on 96-well ultrafiltration plates (NucleoFast) and their concentration measured by the Nanodrop ND 1000 spectrophotometer. The concentration of the purified PCR products was adjusted to 20 ng µl⁻¹. The sequencing libraries were created using 1 ng of DNA with the Nextera XT DNA Library Preparation Kit in a SureCycler 8800 thermocycler (Agilent). Following purification on AMPure beads (Beckman), the libraries were controlled using the High Sensitivity D1000 assay (Agilent) on a TapeStation 2200. The products were sequenced by paired-end 150 bp Illumina sequencing using Illumina NextSeq HiSeq. All kits were used following the manufacturer’s instructions. At least one million reads were generated for each sample.

Sequence analysis

De novo assembly was conducted with CLC Genomics Workbench v22.0 (Qiagen) using parameters previously optimized [23] to improve the resolution of reads from closely related viruses in mixtures (trim quality score limit=0.01; trim ambiguous nucleotides with maximum number of ambiguities=1, mismatch cost=2; insertion cost=2; deletion cost=2; length fraction=0.5; similarity fraction=0.95). Contigs that correspond to EV genomic sequences were flagged through a blast analysis. EV strain typing was established through VP1 nucleotide sequence comparison, wherein a sequence identity of >75% with the corresponding homotypic prototype strain was considered indicative of the same virus type. Sequence alignments were performed with CLC Main WorkBench version 24 (Qiagen), and phylogenetic trees were produced using mega X software [26]. The evolutionary history was inferred using the Minimum Evolution method; the evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site. The minimum evolution tree was searched using the Close-Neighbour-Interchange algorithm at a search level of 1. The neighbour-joining algorithm was used to generate the initial tree. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test was calculated from 1,000 replicates.

Results

Sociodemographic data

Our study was carried out on samples from 178 patients from five sentinel sites (Table 1). The sex ratio was 1.12 (94/84). The age ranged from 0 through 75 years, with a median age of 36 months and a mean age of 10 years (Fig. 1). The dominant clinical picture was ILI (75.0%), while 25% of patients displayed SARI. Fever was present in almost all patients, with a mean temperature of 38.7 °C, ranging from 37.6 to 40.7 °C. Cough and rhinorrhoea were present in 97.7 and 71.13% of patients, respectively, while 14.0% had dyspnoea. The average consultation time was 4 days after the onset of the symptoms. Because the samples were pooled, no association between the presence of EV and neither the sociodemographic data nor the symptoms could be conducted.

Fig. 1. Age distribution of the patients.

Enterovirus typing

Samples were pooled according to the sampling site and date of collection. Of 178 samples, 20 pools were constituted (Table 1), of which 17 were positive for EV detection. Upon amplification, 12 out of 17 pools yielded bands corresponding to the size expected for EVs (Table 1). After Illumina sequencing and de novo assembly, the number of EV capsid sequences per pool of nasopharyngeal swabs ranged from 1 to 4. The corresponding viruses were molecularly typed by considering the VP1-encoding region [27]. Overall, 26 EVs were identified, of which 14 belonged to the species E. coxsackiepol, 9 to the species Enterovirus betacoxsackie (generically named EV-Bs) and 3 to the species Enterovirus alphacoxsackie (generically named EV-As). PV of serotype 3 (PV3) was found in eight pools, of which one also contained PV2 (Table 2). Besides PVs, we found a few viruses belonging to three EV-C virus types: two EV-C99s, one coxsackievirus (CV) A11 and two CVA24s (Table 3). The other EVs belonged to 9 virus types, 8 of which are enteric virus types frequently described in biological and wastewater samples: CVs – CVA4, CVA5, CVA16, CVB4 and CVB5 – and echovirus (E) 7, E1 and E29 (Table 3); the last one, EV-B83, is more rarely reported, with less than 50 specimens available in public databases. Unexpectedly, the dataset did not contain any EV with a clear respiratory tropism, such as EV-D68 or respiratory EV-Cs.

Table 2. Detection of PVs and non-polio EVs in pools of samples.

	Non-polio enterovirus		Total for PV
	Negative	Positive
PV
Negative	8	4	12
Positive	3	5	8
Total for Non-polio enterovirus	11	9	20

Table 3. EVs detected in this study.

Virus type	No.
Coxsackievirus A4	1
Coxsackievirus A5	1
Coxsackievirus A16	1
Total EV-A	3
Coxsackievirus B4	2
Coxsackievirus B5	1
Echovirus 1	1
Echovirus 7	3
Echovirus 29	1
Enterovirus B83	1
Total EV-B	9
Coxsackievirus A11	1
Coxsackievirus A24	2
Enterovirus C99	2
Poliovirus 2	1
Poliovirus 3	8
Total EV-C	14
Total	26

All PVs featured zero to three nucleotide differences with the corresponding vaccine strain in the VP1, making them ‘Sabin-like strains’ according to the Global Polio Lab Network’s classification [28], i.e. viruses likely excreted by vaccine recipients. No recombination event was detected in their genomes. The VP1 sequences of all non-polio EVs were compared to that of viruses of the same types available in public databases: no notable characteristics were observed in these viruses, except for the EV-B83 identified in this study, which is relatively distant from all previously described EV-B83 strains (nt similarity <82% in the VP1) and therefore forms a separate phylogenetic branch (Fig. 2). Of note, one of the CVA24s was a CVA24 variant known to be associated with large outbreaks of acute haemorrhagic conjunctivitis due to its marked ocular tropism (Fig. 3) [29].

Fig. 2. Phylogenetic relationships of EV-B83 strains based on the VP1 nucleotide sequence. The strain identified in this study is indicated by a circle. The other sequences are named by using their respective GenBank accession number, country (ISO 3166-1 alpha-3 codes) and year of sampling. The sequence of the EV-B73 prototype strain has been introduced for rooting of the tree. The bootstrap values are indicated if higher than 90%.

Fig. 3. Phylogenetic relationships of CVA24 strains based on the VP1 nucleotide sequence. Circles indicate the strains identified in this study. The other sequences are named by using their respective GenBank accession number, country (ISO 3166-1 alpha-3 codes) and year of sampling. Triangles indicate the prototype strains. The sequence of the EV-C99 prototype strain has been introduced for rooting of the tree. The bootstrap values are indicated if higher than 90%.

Discussion

This study aimed to characterize respiratory EVs that circulate in CAR, in particular, respiratory EV-Cs that are poorly described, especially in sub-Saharan Africa. For this purpose, we chose to screen samples from patients with SARI or ILI, given that respiratory EV-Cs have been associated with such symptoms. To increase the number of samples that could be screened, pools were constituted. Besides reducing the costs compared to individual screening, sample pooling streamlines ethical approval by precluding the need for individual informed consent, thereby circumventing the necessity for individual patient recontact. To ensure robust performance metrics, screening was carried out using a commercially available assay that has been formally validated for human diagnostic use and already used to detect EV-D68 in respiratory samples [30,31]. An in-house assay was deliberately avoided because its sensitivity and specificity could not be objectively assessed under the study conditions. The positivity rate in this study (85%) was relatively high compared to previous studies conducted elsewhere in patients with respiratory symptoms. For instance, the detection rate of rhinoviruses/EVs was 27% in Singapore [32], 37% in Italy [33], 20% in Turkey [34] and 45% in Thailand [35]. The high positivity rate in our study is likely due to the fact that we screened pools while other studies screened individual samples, but it could also result from differences in the inclusion criteria and the respective sensitivity of the detection methods. Nonetheless, it is noteworthy that in a study carried out in Cameroon with a very close design (same pool sizing, similar screening assay), the positivity rate was particularly low, with only 2 positive out of 23 pools (8%) [36]. EV genomic amplification failed for five pools that were found positive through screening, maybe because of low concentrations of virus RNA. Our study was intended to particularly focus on EVs with a respiratory tropism, such as EV-D68, a virus known since the 1960s that suddenly emerged as a virus of concern in 2014 [18], and some EV-C virus types more recently discovered that were associated with severe respiratory symptoms [10,37]. A small number of works reported the detection of EV-D68 in Western Africa [38,40], Central Africa [36], Eastern Africa [41] and South Africa [42]. Among the five known respiratory EV-C virus types (EV-C104, EV-C105, EV-C109, EV-C117 and EV-C118), two have already been sampled in Africa: EV-C104 in Gambia [17] and EV-C105 in Burundi [16] and the Democratic Republic of the Congo [43]. The absence of respiratory EV-Cs in our results could be due to the lack of sensitivity of the screening assay towards these viruses. To the best of our knowledge, the panel of viruses used by the developers to validate its sensitivity is not publicly available. The genomic regions targeted by the assay are also unknown. Pan-EV assays generally target the 5′ non-coding region of the genome, which is highly conserved among EVs. Nonetheless, respiratory EV-Cs differ markedly from other EVs in this region [44], a discrepancy that may allow them to evade detection by standard assays. Several studies have evaluated the commercial assay employed in our work using panels of clinical specimens; however, the proportion of EV-positive samples in these investigations was modest, and the presence of respiratory EV-Cs within the panels was not documented [45,47]. Nonetheless, one study reported the detection of EV-C117 clinical strains with this assay [48], which suggests that it efficiently targets respiratory EV-Cs. Furthermore, the primers used for genomic amplification (UG52, UG55 and cre-R) were designed to efficiently target a broad range of EV-Cs, including respiratory EV-Cs. Their ability to amplify EV-D68 genomes was experimentally validated in our laboratory. Because respiratory EV-Cs do not replicate in established cell lines, no viral isolates were available to directly assess primer performance for these viruses. Nevertheless, the limited number of mismatches between the primers and respiratory EV-C sequences was unlikely to compromise amplification efficiency. Therefore, the absence of respiratory EV sequences in our dataset likely reflects a low baseline prevalence of these pathogens during inter-epidemic periods [49,50]. This hypothesis is in accordance with a previous study that did not report any respiratory EV-Cs in wastewater sampled in Nigeria [12]. Previous large-scale screenings of respiratory specimens conducted in other continents reported a relatively low number of EV-Cs. In a 22-month Danish survey, 6 EV-Cs were detected among 221 EV-positive samples [51], whereas a Dutch 8-year study found 20 EV-Cs in 14,565 respiratory samples [52]. Nasopharyngeal swabs may not be suitable if these viruses multiply mainly in the lower respiratory tract. A study carried out in Belgium found EV-C104 in bronchial aspirates [53], but a study conducted in Italy on EVs in the upper and lower respiratory tracts did not show a significant difference in the results obtained [37].

Routine PV immunization in CAR relies on a combination of inactivated polio vaccine (one dose at 14 weeks) and bivalent OPV (one dose at birth and three subsequent doses at 6, 10 and 14 weeks). This explains the detection of PV vaccine strains in eight pools of respiratory samples. Although the gut constitutes the main replication site of PVs, it is known that PVs can also replicate in the respiratory tract for a short period [54,55]. Because the samples were pooled, it is impossible to infer the proportion of respiratory samples that were positive for PVs. Furthermore, it is impossible to determine whether the original samples contained infectious particles or virus RNA only. Nonetheless, these observations support the inclusion of nasopharyngeal swabs among PV-infected materials, as recommended by the WHO guidelines for PV containment [56].

In addition to vaccine PVs, the EVs found in our study are similar to those often seen via polio surveillance because they have an enteric tropism. The ability of echoviruses of different types to replicate in the respiratory tract and to induce respiratory illnesses has been established a long time ago [57,60]. Similarly, CVs members of species E. alphacoxsackie and E. betacoxsackie are frequently detected in throat swabs of patients displaying respiratory diseases [61,63]. By contrast, to the best of our knowledge, all EV-B83s previously reported had been detected in faecal samples or in wastewater. Non-polio EV-Cs with no specific respiratory tropism (such as CVA11, EV-C99 and CVA24 found in this study) are able to replicate in airway cells [64], but they are less frequently reported in respiratory samples than EV-Bs [65] either because they replicate less efficiently in the respiratory tract or because they are not associated with severe symptoms, with the exception of CVA21, which is responsible for outbreaks of severe respiratory disease [66,67] but was not found among our dataset. It is important to note that, owing to the pooled screening strategy (rather than individual sample testing), minority EVs may have been obscured during genome amplification and sequencing. Consequently, our findings should be interpreted as reflecting the composition of sequences recovered from pooled samples, rather than providing a direct estimate of viral prevalence in individual patients.

In conclusion, no genuine respiratory EVs were identified, but the data confirm that a broad spectrum of enteric-tropic EVs are capable of infecting the pharyngeal mucosa. Among these, PVs were relatively frequent, highlighting that respiratory samples constitute a potential source of PV re-introduction in the post-eradication era.

Abbreviations

CAR

Central African Republic

cre

cis-replicating element

Ct

cycle threshold

E

echovirus

EV-Cs

enteroviruses C

EVs

enteroviruses

ILI

influenza-like illness

OPV

oral poliovirus vaccine

PVs

polioviruses

SARI

severe acute respiratory infection