The detection of 3 ambiguous type 2 vaccine-derived polioviruses (VDPV2s) in Uganda

VDPVs originate from OPV which has mutated in the gut. These mutant strains rarely cause outbreaks of paralytic poliomyelitis [1, 2] however such viruses can be culprits for major and prolonged outbreaks.

The poliovirus evolution usually occurs in unimmunized populations with low OPV coverage [3]. The poliovirus can continue to circulate uninterrupted for a long time. The lower the population immunity the longer the poliovirus is allowed to survive and the more it will replicate [2], mutate, and acquire wild poliovirus neuro-characteristics. Fully immunized populations are protected from VDPVs and wild poliovirus [4].

There are 3 serotypes of polioviruses; poliovirus type 1 (PV1), poliovirus type 2 (PV2), and poliovirus type 3 (PV3). The mutated OPV can be classified according to the diversity of the virus capsid protein 1(VP1). Poliovirus undergoes an evolution rate of approximately 1% per year, which results in ~ 10 nucleotide differences in the 903 nt. VP1. The poliovirus ribonucleic (RNA) polymerase enzyme has an error rate of 1 per 1000 nucleotides per round of replication [5, 6], and contributes to virus evolution. The normal period of poliovirus excretion is less than 3 months [7‐9].

The mutated polioviruses are classified as vaccine-related polioviruses (VRPVs) when there is a VP1 sequence divergence from the reference Sabin polioviruses of ≤1% for Sabin 1 and 3, and ≤ 0.6% for Sabin 2 [2]. Sabin polioviruses showing VP1 divergence of > 1% for Sabin 1 and 3, and > 0.6% for Sabin 2 are classified as vaccine-derived polioviruses (VDPVs) [2]. A lower number of nucleotide differences has been considered for the VDPV type 2 because such mutants behaved like the wild type in DRC and Nigeria [10, 11]. Wild polioviruses are distinct and are not genetically linked to vaccine polioviruses [12, 13].

VDPVs are classified into three categories; The first category are the circulating VDPVs (cVDPV) that are genetically linked and isolated from a) at least 2 individuals who may not be AFP cases but coming from different households or b) one individual and one or more environment samples or c) series of environmental samples from 2 or more sites or d) a single VDPV with documented evidence of circulation based on the genetic characteristics of the available VDPVs. The second category of VDPVs are the immune-deficiency-related VDPV (iVDPV). These are viruses isolated from persons with proven immunodeficiencies [14]. The third category are the ambiguous VDPV (aVDPV); these are isolates that are identified from individuals with or without AFP and who are not immune-compromised or from environmental samples without evidence of circulation. VDPVs reported as ambiguous VDPVs might later qualify to be iVDPV or cVDPV. The aVDPVs commonly occur in countries of low rates of polio vaccination coverage and could signal cVDPV emergence and gaps in immunization and surveillance activities [11].

cVDPVs are more transmissible and co-existence with OPV viruses may favor cVDPVs to become endemic while the OPV viruses get eliminated [15]. Without effective vaccination response to boost population immunity, VDPVs manifest with prolonged outbreaks [2, 16]. Sabin 2 is more transmissible than Sabin type 1 and 3 [2, 17] and has been reported to circulate longer.

From 2000 to 2015 cVDPV2 have been reported from 21 countries [18]. These were associated with low AFP surveillance and inadequate immunization coverage. Immunodeficiency-related VDPV2 and VDPV3 were detected in Iran in 1995, 2005, 2007 [19] and other several places [20‐22]. We report three aVDPVs type 2 that were sampled in 2014; two were detected during the routine AFP surveillance in 2014 and one was detected during a follow-up VDPV re-screening survey in 2016.

Two contacts of two AFP cases from Kamuli and Kween districts of Uganda were both confirmed VDPV2 on 03rd November 2014. In the manuscript, the viruses are identified as VDPV no1 and VDPV no2, respectively. A third VDPV2 from an AFP case was later confirmed on 04th May 2016 from Kisoro district located in South-western Uganda in a specimen from a one month old baby. A polio suspected virus was reported on virus culture. The subsequent ITD/VDPV assays displayed a VDPV result curve that was misinterpreted and scored a Sabin 2 poliovirus and not a ‘possible’ VDPV2. As a result the VDPV2 was missed at this stage. The first two contacts were detected during the routine national AFP surveillance activity using the WHO adapted AFP surveillance protocol whereas the third VDPV was detected during the retrospective VDPV screening performed in 2016. Stored frozen isolates of 2014 were investigated. These samples were obtained from cases that were fairly well distributed over the country (Fig. 1).

The characteristics relating to the three aVDPV type 2 from Uganda are shown in Table 1. The 3 aVDPVs were sampled in the 2nd and 3rd quarter of 2014. These aVDPVs type 2 were confirmed from Kamuli and Kween districts in 2014 and from Kisoro district in 2016. Kamuli and Kween are located in the Eastern Uganda 167 and 299 km from Kampala and 130 km apart whereas Kisoro is located in the South-western Uganda 481 km from Kampala. Kampala is the capital city of Uganda characterized with busy travel. The vaccination coverage during 2011, 2012, 2013, and 2014 was not satisfactory for Kamuli, Kween and Kisoro districts. Furthermore the poliovirus AFP surveillance indicators of the non-polio AFP rate and stool adequacy were not satisfactory for the 3 districts (Table 1).

The Kamuli, Kween and Kisoro VDPVs did not share three mutations in the VP1 region. The VDPVs from Kamuli and Kween both contained 6 mutations with a VP1 variation from the reference Sabin of 6.6%. Though these viruses had the same VP1 variation rates they only shared two mutations at the nucleotide level (G308A and T428C) and two mutations at the amino acid level (R103K and I143T). The VDPV from from Kisoro contained 7 mutations with a VP1 variation from the reference Sabin of 7.8%. It shared two mutations at the amino acid level R103K and K160E with Kamuli and Kween and K160E with Kamuli. At the nucleotide level is shared mutations at G308A, and A505G (Tables 2 and 3). Thus these VDPVs have been categorized as ambiguous VDPVs and not cVDPVs. During 2015 one National Immunization Day (NID) of house to house (HTH) approach and two Sub National Immunisation days (SNID) of HTH were conducted. Furthermore during 2016 one NID and one SNID rounds of OPV immunization were conducted in Kamuli, Kween and Kisoro and other high risk districts [32] to interrupt virus circulation (Fig. 2).

A review of the Sabin poliovirus 2 characteristics showed that out of the 19 cases; 21.4% were VDPVs, 42.9% were VRPVs and 35.7% were non-mutant Sabin polioviruses 2. The point mutations that were observed for Sabin 2 were: C156T, G165A, C180A, C279T, G308A, T316G, T428C, A505G, T819C, and A828C (Table 2). Substitutions G308A, T428C, and A505G were common in VDPVs. The mutation sites G308A and T428C were common to the Kamuli and Kween VDPVs. The Kisoro VDPV contained mixed bases at 7 positions (Table 2).

The amino acid substitutions in the VP1 region of the Sabin mutants are further shown in Table 3.

All the VDPVs contained the G308A and R103K nucleotide and amino acids substitutions respectively (Tables 2 and 3). The common amino acid mutations for the VDPV were K169E for Kamuli and Kisoro and I143T together with R103K for Kamuli and Kween districts. The phylogeny of the aVDPV type 2 reported for Democratic Republic of Congo (DRC), Nigeria and Madagascar was constructed and shown (Fig. 3). The Uganda (UGA) VDPVs were less divergent representing a short evolution period.

Ninety-nine percent, (92 out of 93) of the specimens from the AFP cases that were re-screened for Sabin 2 and VDPV2 in the retrospective VDPV study and the 4 and 5 secondary contacts of VDPVs no1 and no2 were negative for the serotype 2 polioviruses. One specimen was collected from the original AFP cases on follow-up and also one specimen from the secondary contacts which could have compromised on virus recovery. Only one virus out of the 93 was a type 2 poliovirus that had not been detected initially. Results from poliovirus dual-stage rRT-PCR ITD/VDPV kit that was used for the index testing and the ITD rRT-PCR 4.0 kit that was used for the re-screening were concordant; previously detected Sabin 2 polioviruses were not re-run on ITD but were sequenced to screen for VDPV2s. Only one VDPV type 2 from the previously detected Sabin 2 polioviruses was identified on the second screening.

The three VDPVs shared the R103K mutation that could be linked to a natural selection process. The mutation I143T was common to Kamuli and Kween and has been reported in all the polioviruses that were detected from stool specimens of AFP cases in an Afghanistan study [5]. The nucleotide mutation T428C was the most frequent mutation among the studied viruses, which is consistent with the previous reports in the scientific literature. The T428C mutation may have a pivotal role in VDPV2 evolution. The VDPV cases from Kamuli and Kween originated from immune-competent persons with normal CD4 cell counts compatible with the aVDPV classification. Vaccine-derived polioviruses have been reported elsewhere in African countries namely DRC, Angola, Egypt, Nigeria, and Sudan [16, 33] and now VDPVs identified in Uganda for the first time.

The wild poliovirus 2 was eradicated ahead of wild poliovirus 1 and 3. Wild Poliovirus 1 (WPV1) and Wild Poliovirus 3 (WPV3) are being eradicated later because the OPV1 and 3 are less immunogenic than OPV2 [34]. In order to overcome the challenge of a lower immunogenicity, a more immunogenic bivalent vaccine has been developed for OPV 1 and 3. It has already been introduced in the routine OPV immunization schedule and OPV 2 has been withdrawn from the trivalent OPV. Trivalent OPV (tOPV) previously contained OPV type 1, OPV type 2 and OPV type 3 and now a bivalent OPV (bOPV) which contains OPV type 1 and OPV type 3. Since 1999, no cases of Wild Poliovirus 2 (WPV2) have been reported globally and OPV2 included in tOPV has continued to cause VDPV type 2 infections. A switch from tOPV2 to bOPV and Inactivated Polio Vaccine (IPV) has been adopted as a mitigation measure against VDPV type 2 infections [35, 36].

The third VDPV was from a baby in Kisoro district and contained mixed bases characteristic of iVDPV [37, 38]. Unfortunately, the immune status of the child could not be established because she died before the follow-up visit following a respiratory disease. The evolution of VP1 accumulate at a rate of 1% per year therefore this VDPV could have been from an exogenous source. It could not be established whether the VDPV came from a family member or not because the virus was confirmed late. Confirmation by performing a repeat stool test was also not possible because the sample had been destroyed in accordance with the WHO programmatic recommendations for type 2 containment. The poliovirus immunization and the surveillance indicators for Kisoro district were not satisfactory. Kisoro has a steep mountainous landscape and a constrained accessibility to the provider services could explain the inadequate performance. The DRC that has reported VDPVs borders Uganda. Kisoro is closest to DRC in southwestern Uganda and thus Kisoro is a high risk area by vicinity and restricted accessibility In view of the high risk exposure for Kisoro the national EPI program enhanced the polio vaccination in Kisoro district to safe guard against virus importation. It is commendable that by the time of reporting this VDPV, the high risk for VDPVs in Kisoro district had been pre-determined and corrective actions had already been taken. The 3 districts and country at large need to be closely monitored to enhance population immunity and prevent re-occurrences. VDPVs can be identified in low risk areas among subpopulations with low immunity and in recipients with no known immune deficiencies [33]. Eradicating of VDPVs remain a priority in both the high and low risk populations.

In the 2014–2015 VDPV rescreening, a survey was conducted in 33 district and 93 AFP cases were tested; nineteen (20.4%) were Sabin poliovirus 2, 28 (30.1%) Sabin poliovirus 1 and 46 (49.5%) were Sabin poliovirus 3. Sabin 2 was least prevalent which is not surprising because OPV2 is more immunogenic than OPV1 + 3. However what is surprising, VDPV2s are more common than VDPV1 or 3. The Sabin 2 poliovirus could be associated with sub-optimal secondary excretion or spread which result in partial secondary immunization giving a chance to the virus to evolve. Understanding more about the Sabin poliovirus 2 evolution remains an important area for investigation and could enhance the understanding of the pathogenesis of the live attenuated vaccines [39].

The emergence of VDPVs under ideal good hygienic conditions, high IPV coverage and high population immunity has not been demonstrated [40] however, these parameters differ especially in the developing countries and need to be considered. In Uganda the risk of VDPVs has been attributed to the inadequate vaccination coverage and surveillance system together with possible proximity to an endemic area. The national immunization program responded exceptionally well, with targeted vaccination and intensified AFP surveillance.

One additional isolate was confirmed an aVDPV during the VDPV survey. The VDPV survey enhanced VDPV detection in the country.