Genetic landscape and macro-evolution of co-circulating Coxsackieviruses A and Vaccine-derived Polioviruses in the Democratic Republic of Congo, 2008-2013

Abstract

Enteroviruses (EVs) are among the most common viruses infecting humans worldwide but only a few Non-Polio Enterovirus (NPEV) isolates have been characterized in the Democratic Republic of Congo (DR Congo). Moreover, circulating vaccine-derived polioviruses (PVs) [cVDPVs] isolated during multiple outbreaks in DR Congo from 2004 to 2018 have been characterized so far only by the sequences of their VP1 capsid coding gene. This study was carried to i) investigate the circulation and genetic diversity of NPEV and polio vaccine isolates recovered from healthy children and Acute Flaccid Paralysis (AFP) patients, ii) evaluate the occurrence of genetic recombination among EVs belonging to the Enterovirus C species (including PVs) and iii) identify the virological factors favoring multiple emergences of cVDPVs in DR Congo. The biological material considered in this study included i) a collection of 91 Sabin-like PVs, 54 cVDPVs and 150 NPEVs isolated from AFP patients between 2008 and 2012 in DR Congo and iii) a collection of 330 stool specimens collected from healthy children in 2013 in the Kasai Oriental and Maniema provinces of DR Congo. Studied virus isolates were sequenced in four distinct sub-genomic regions 5'-UTR, VP1, 2CATPase and 3Dpol. Resulting sequences were compared through comparative phylogenetic analyses. Virus isolation showed that 19.1% (63/330) healthy children were infected by EVs including 17.9% (59/330) of NPEVs and 1.2% (4/330) of type 3 Sabin-like PVs. Only one EV-C type, EV-C99 was identified among the NPEV collection from AFP patients whereas 27.5% of the 69 NPEV isolates typed in healthy children belonged to the EV-C species: CV-A13 (13/69), A20 (5/69) and A17 (1/69). Interestingly, 50 of the 54 cVDPVs featured recombinant genomes containing exogenous sequences in at least one of the targeted non-structural regions of their genomes: 5'UTR, 2CATPase and 3Dpol. Some of these non-vaccine sequences of the recombinant cVDPVs were strikingly related to homologous sequences from co-circulating CV-A17 and A20 in the 2CATPase region as well as to those from co-circulating CV-A13, A17 and A20 in the 3Dpol region. This study provided the first evidence uncovering CV-A20 strains as major recombination partners of PVs. High quality AFP surveillance, sensitive environmental surveillance and efficient vaccination activities remain essential to ensure timely detection and efficient response to recombinant cVDPVs outbreaks in DR Congo. Such needs are valid for any epidemiological setting where high frequency and genetic diversity of Coxsackieviruses A13, A17 and A20 provide a conducive viral ecosystem for the emergence of virulent recombinant cVDPVs.

Fig 1. Summary of the studied collections of stools, virus isolates and overall results of cell cultures and molecular typing.

Circulating vaccine-derived polioviruses (cVDPVs) were obtained from Acute Flaccid Paralysis (AFP) cases originating all over the territory. They were isolated and typed using the same techniques as for Sabin-like polioviruses. ITD, intratypic differentiation; rRT-PCR, real-time RT-PCR.

Fig 2. Phylogenetic relationships of the full-length VP1 sequences of the studied EV-A and -D.

Maximum likelihood trees were inferred from the alignments of full-length VP1 sequences of EV-A76 using the GTR+G+I model of nucleotide substitution (A) and partial VP1 sequences of EV-D111 strains (nt 1–471 according to the VP1 sequence of the EV-D111 prototype strain KK2640) using the T92+G model of nucleotide substitution (B). Studied isolates are specified in bold red while isolates previously reported in DR Congo are highlighted by circles in bold blue. The year and country of isolation of each reference isolate are indicated, when known (BGD: Bangladesh; CAF: Central African Republic; CMR: Cameroon; CHN: China; COD: Democratic Republic of the Congo; EGY: Egypt; FRA: France; JPN, Japan; KAZ: Kazakhstan; USA: United-States of America). Prototype strains are highlighted by black triangles. For clarity, bootstrap values less than 70% have been omitted and the scale bars indicate nucleotide distance as substitutions per site. Isolates belonging to specific lineages commented in the main text are gathered in grey-shaded boxes.

Fig 3. Phylogenetic relationships of the newly sequenced EV-C isolates from DR Congo.

The Maximum likelihood phylogram was inferred from the alignment of full-length VP1 sequences of EV-C strains using the most complex GTR+G+I model of nucleotide substitution. Previously described isolates originating from DR Congo are indicated in bold blue while isolates from this study are highlighted by red circles. Isolates from healthy children are named with a 3 letter code indicating the province of origin (MAN: Maniema and KOR: Kasai Oriental) whereas the unique EV-C99 recovered from a paralyzed child is R08-3798. The year and country of isolation of each reference isolate are indicated, when known (ARG: Argentina; AUS: Australia; BGD: Bangladesh; CAF: Central African Republic; CMR: Cameroon; CHN: China; COD: Democratic Republic of the Congo; NIC: Nicaragua; MEX: Mexico; MDG: Madagascar; OMN: Oman; TCD: Chad; USA: United-States of America; ZAF: South Africa). In addition to the sequences of prototype strains, sequences displaying highest similarities with the studied isolates were retrieved from databases using NCBI BLAST search and included as references in the analyzed dataset. Prototype strains are highlighted by triangles. For clarity, most bootstrap values less than 60 have been omitted. The scale is shown at the bottom, as substitutions per site. Isolates belonging to the virus types and lineages commented in the main text are gathered in color-shaded boxes.

Fig 4. Phylogenetic trees depicting the genetic relationships between the nucleotide sequences derived from sub-genomic regions of species C enteroviruses.

Maximum likelihood trees were inferred from specific alignments of partial nucleotide sequences of the VP1, 2C^ATPase and 3D^pol coding genes [nucleotide positions 2482 to 2953 for VP1 (A); 4123 to 4922 for 2C^ATPase (B); 6166 to 6837 for 3D^pol (C) according to the complete genome sequence of Sabin 2]. Branch lengths were calculated using the best-fit model of nucleotide substitution estimated with Smart Model Selection based on the Bayesian Information Criterion: GTR+G for the VP1 (A), and 3D^pol (C) regions and GTR+G+I for the 2C^ATPase (B) region. Phylogenetic lineages of CV-A13 are indicated on the VP1 (A) and 2C^ATPase (B) derived trees using letters A to F. The genetic distance is indicated with a scale bar at the bottom. Numbers at nodes correspond to the percentage of 1,000 bootstrap replicates supporting the distal cluster. Since databases available sequences of wild polioviruses as well as most non-polio enteroviruses species C originating from other epidemiologic context featured no peculiar relationships with the studied isolates, most of them were omitted from the final datasets. However, selected poliovirus strains isolated during type 1, 2 and 3 cVDPV-associated outbreaks in Egypt (Sab2.Egy88, Sab2.Egy93), Madagascar (MAD04 and MAD29), Haiti (Sab1.HAI03, Sab1.HAI07), Dominican Republic (Sab1.Dor13), Philippines (Sab1.Phi01) and Cambodia in 2002 (Sab3.Cam02) were included. Trees were oriented using the nucleotide sequences of Enterovirus D70 (EV-D70) as outgroup. cVDPV isolates obtained from paralyzed patients from 2008 to 2012 and Coxsackievirus A viruses isolated from healthy children in 2013 in the Maniema (MAN) and Kasai Oriental (KOR) provinces of DR Congo are highlighted with circles color-coded according virus types: red for cVDPV, blue for CV-A13, green for the unique CV-A17 and purple for CV-A20 isolates. cVDPV identifiers include: “R” standing for DR Congo; “two digit code” indicating the year of isolation (08, 2008; 09, 2009; 10, 2010; 11, 2011 and 12, 2012) and “four digit code” representing the serial number. For clarity, some clusters defined by homotypic virus isolates were collapsed: Phi., Philippines; Dom. Rep., Domican Republic; Eq. and ka., Equateur and Katanga provinces of DR Congo. EV-C types and/or cVDPV sequences featuring peculiar relationships discussed in the text are gathered in grey-shaded boxes.

Fig 5. Phylogenetic trees depicting genetic relationships between nucleotide sequences derived from the 5’utr genomic regions of species C enteroviruses.

The maximum likelihood tree was based on the alignment of partial nucleotide sequence of the 5’UTR region of the genome (nt 183 to 573 according the complete genome of the Sabin 2). Branch lengths were calculated using the best-fit model of nucleotide substitution GTR+G+I estimated with Smart Model Selection based on the Bayesian Information Criterion. Numbers at nodes correspond to the percentage of 1,000 bootstrap replicates supporting the distal cluster. For comparison, lineages depicted on the VP1-based reference phylogeny (Fig 4A), were highlighted here despite the fact that they were not supported by well-defined and bootstrap supported groups (Fig 4A). For comparison, the cluster defined by the 41 circulating vaccine-derived polioviruses featuring Sabin 2 related sequences were collapsed. Details about isolates’ labeling and the dataset considered are the same as specified in the legend of Fig 4.