Sampling, virus isolation, ITD and shipping of poliovirus materials

AFP surveillance, well established in Uganda, was enhanced to ensure the effective detection of any potential polio paralytic cases. ES, operational in Uganda since May 2017, was expanded to include 11 sites sampled once a month. Following the cVDPV2 outbreak in the country, sampling frequency for Kampala sites was increased to twice a month to enhance sensitivity for PV detection. Stool samples were collected through the routine nationwide AFP surveillance programme, in which children <15 years old presenting to health facilities with weakness or paralysis are investigated regardless of vaccination status. Two stool samples, collected 24–48 h apart, were transported under cold chain to the laboratory for analysis. Grab sewage samples were collected at regular intervals from the 11 established ES sites and similarly transported. These collections occurred after nationwide nOPV2 campaigns targeting all children <5 years old.

Stool samples were processed for poliovirus isolation following the standard WHO protocol²³. Poliovirus isolation was performed using two WHO-recommended cell lines: L20B and RD. L20B cells, which are genetically engineered mouse L cells expressing the human poliovirus receptor CD155, provide a highly specific substrate for poliovirus replication and facilitate rapid differentiation from non-polio enteroviruses. RD cells (human rhabdomyosarcoma cell line), which support the growth of a broader range of enteroviruses, were used in parallel to ensure the sensitive detection of poliovirus. The cells were obtained from master cell banks stored at MHRA which are distributed through the World Health Organization’s Global Polio Laboratory Network (GPLN) and are used routinely in accredited national and regional polio laboratories. Sewage samples were collected using the grab method and concentrated overnight using the two-phase separation method²⁴, followed by poliovirus isolation using the cell culture algorithm following WHO guidelines for ES samples²⁵. All cell culture flasks with positive cytopathic effect (CPE) in the L20B cell line were subjected to ITD PCR²⁶ as independent specimens. The nOPV2 isolates were then inactivated by spotting the supernatant onto FTA cards (Whatman, Life Sciences), which contain chemicals that inactivate the virus while preserving the viral RNA. This method has been used in the WHO Global Polio Laboratory Network for several years and includes a ‘70 °C for 4 min’ heat treatment before spotting isolates on the FTA cards and shipment as non-infectious materials. The FTA cards were sent to the MHRA WHO Global Specialized Laboratory for Polio in the United Kingdom for whole genome sequence analysis. In addition, PV2+/nOPV2− virus isolates were shipped to MHRA for whole-genome sequencing. Due to the trigger of a level 2 response in relation to nOPV2 use, the original sewage concentrate plus previous and subsequent sewage concentrates from the same location (n = 5) were also shipped to MHRA for subsequent re-culture using WHO guidelines for ES samples and characterization of any polioviruses.

PV whole-genome and tiled PCR with reverse transcription amplification

Using disposable punches, 6 discs of 4 mm diameter were cut from each FTA card and viral RNA extracted using the Qiagen QIAamp Viral RNA kit (Qiagen, 52904). Viral RNA was purified from infected cell culture supernatant using the High Pure viral RNA kit (Roche, 11858882001) with Proteinase K pre-treatment. Whole-genome PV PCR with reverse transcription (RT–PCR) fragments were amplified from extracted viral RNA by one-step RT–PCR using a SuperScript III One-Step RT–PCR system with Platinum Taq High Fidelity DNA Polymerase (Invitrogen, 12574026) and primers PCR-F (5′-AGA GGC CCA CGT GGC GGC TAG-3′) and PCR-3′ (5′-CCG AAT TAA AGA AAA ATT TAC CCC TAC A-3′)²⁷. Amplification conditions were 50 °C for 30 min for the RT reaction plus 94 °C for 2 min plus 42 cycles of 94 °C for 15 s, 55 °C for 30 s and 68 °C for 8 min, with a final extension step of 68 °C for 5 min. Since the quality of RNA extracted from FTA cards is generally poor compared with RNA extracted from cell culture fluid, a custom-made tiled PCR approach with overlap between amplicons was used to generate sequence information spanning the whole genome of any poliovirus present in the FTA card. We followed a published protocol²⁸. Primers used to generate tiled PCR products are provided in Extended Data Table 1. Equimolar mixtures of the overlapping PCR products spanning the whole genome were used for sequencing, and additional PCRs were conducted to complete any sequencing gaps if required. Good laboratory practice was used in all molecular assays to prevent cross-contamination of samples. Positive and negative RNA extraction and PCR controls were included in every assay.

PV whole-genome next-generation sequencing

PV whole-genome and tiled PCR products were sequenced using MiSeq Illumina technology. Sequencing libraries were constructed using the Nextera DNA Prep kit (Illumina, 20060059) and dual-indexed using Nextera DNA Indexes (Illumina, set A 20091660). These libraries were pooled in equimolar concentrations and sequenced with 250-bp paired-end reads on MiSeq v2 (500 cycles) kits (Illumina, MS-102-2003) or with 150-bp paired-end reads on NextSeq 2000 (Illumina, MS-20100984). Initial demultiplexing was performed on board using bcl2fastq v.2.2 (MiSeq) and DRAGEN BCL Convert v.3.8.4 (NextSeq 2000) software. FASTQ sequencing data were adapter and quality trimmed using Cutadapt v.2.10 for a minimum Phred score of Q30, minimal read length of 75 bp and 0 ambiguous nucleotides.

In addition, selected PV whole-genome and tiled PCR products were also sequenced using Oxford Nanopore Technologies (ONT). PCR products underwent a standard preparation for Nanopore sequencing with the SQK-NBD110-9/SQK-NBD112.96 kit together with the barcoding expansion EXP-PBC096 kit. Briefly, PCR products were sheared using the NEBNext Ultra II End Repair/dA-Tailing Module (E7546L), and the native barcode adapters were ligated to the PCR product using Blunt/TA Ligase Master Mix (M0367). Individual barcoded samples were pooled into one tube and the barcoded libraries were purified using AmPure XP beads (A63882). Next, the sequencing adapters supplied in the kit were ligated to the pooled library using NEBNext Quick Ligation Module (E6056L). The flow cells were primed using flow cell priming kit (EXP-FLP002) and the DNA libraries were loaded along with sequencing buffer and loading beads. Sequencing was performed on an ONT MinION Mk1B sequencer using R9.4, R10.3 or R10.4 flow cells. Nanopore data were drawn from several individual MinION sequencing runs. Sequencing run duration ranged between 4 h and 16 h depending on the number of samples included in that run. Only flow cells with a minimum of 600 active pores on flow cell check (pre-test) were used. Fast5 files were base called with the high-accuracy basecalling model using Guppy v.5 with graphics processing unit (GPU) acceleration on a Precision 7540 Dell laptop running Ubuntu 18 LTS with 32 GB RAM, a 16 core Intel Xeon central processing unit (CPU) and Nvidia RTX3000 GPU. FASTQ files were generated using MinKNOW software and stored for further processing.

Fastq files from both Nanopore and Illumina sequencing were further processed and analysed using Geneious 10.2.6 software. Raw sequence data were imported into Geneious 10.2.6 and sequence contigs were built by reference-guided assembly as previously validated^{27,29,30,31,32}. Fastq reads were initially mapped simultaneously to Sabin 1, nOPV2 and Sabin 3 whole-genome sequences, and contig sequences were generated. Potential nOPV2 recombinant sequences were identified in contigs showing partial sequence coverage across the genome. Filtered reads were then iteratively reassembled to consensus sequences covering the nOPV2 capsid region using stringent assembly conditions to build whole-genome contig sequences. Assembly conditions were as follows: minimum of 50 base overlap, minimum overlap identity of 98%, maximum of 2% mismatches per read and both end-pair reads mapping for Illumina and minimum of 2,000 base overlap, minimum overlap identity of 95%, maximum of 5% mismatches per read for Nanopore. Final consensus sequences were obtained by assigning the most common nucleotide sequence to each nucleotide position within each contig, requiring a minimum of 20-nt sequence coverage, trimming to the reference sequence and removing the primer sequences from the final consensus sequence. When necessary and to confirm the results, fastq reads were also independently assembled de novo, using the same stringent assembly conditions. In addition, the options to produce scaffolds and ignore words repeated more than 100–1,000 times, available in the Geneious assembler, were selected to improve the quality of the assembly process. When sequencing whole-genome PCR products with Nanopore, assembly and contig generation was much more straightforward as it simply required a single step mapping whole-genome reads simultaneously to Sabin 1, nOPV2 and Sabin 3 whole-genome sequences, and generating consensus sequences as above. Results using different PCR amplification strategies and assembly approaches were identical. Manual analyses for visualizing and quantifying assembly results were performed throughout the process. Final consensus sequences are available from GenBank under accession numbers listed in Supplementary Table 1. nOPV2 genome sequences were aligned using the programme MUSCLE (v.3.8.425), and phylogenetic trees were constructed using the FastTree algorithm (v.2.1.11), both within Geneious.

Transfection of RNA extracted from FTA cards

Transfection experiments were carried out to show viability of the recombinant virus genetic information obtained from the FTA cards by sequencing on the Illumina platform. Fresh RNA was extracted as described above on the day of transfection. Transfection was carried out using Hank’s balanced salt solution and the DEAE-dextran transfection method³³. The flasks with RNA and the control were incubated at 35 °C in a CO₂ incubator. CPE (4+) was observed for flasks transfected with RNA extracted from FTA cards (n = 3) on day 3. No CPE was observed in the control flasks at 5 days post transfection. Flasks showing CPE were then subjected to RNA extraction and whole-genome pan-PV PCR amplification, and sequenced using Illumina and Nanopore technology as described above²⁷.

Direct sequencing of enteroviruses from sewage concentrates

A retrospective investigation was conducted to determine the EV distribution in sewage concentrates collected from January 2022 to April 2022 from the Kisenyi site, that is, before and after the point detection of the double recombinants (n = 5). RNA was extracted from the sewage concentrates using the High Pure viral RNA kit (Roche, 11858882001) with Proteinase K pre-treatment, and pan-EV entire capsid-coding region RT–PCR templates were generated and sequenced using the MiSeq Illumina platform as described here and elsewhere³². The closest virus relatives to the Uganda EV sewage strains were identified using the RIVM and BLAST online sequence analysis tools^34,35, and EV serotypes were assigned on the basis of their VP1 sequence. EV genome sequences were aligned using the programme MUSCLE (v.3.8.425), and phylogenetic trees were constructed using the FastTree algorithm (v.2.1.11), both within Geneious.

Transgenic mouse neurovirulence test

Tg66-CBA transgenic mice expressing the human poliovirus receptor (50% male, 50% female, 6–8 weeks old) were used for these experiments. Tg66-CBA mice are the product of crossing Tg66 mice with CBA/J mice several times then selecting for homozygous CBA MHC genes and homozygous PVR. Animal experiments were performed at the MHRA, with ethics approval from MHRA’s Ethics and Human Materials Advisory Committees and the Animal Welfare and Ethical Review Body. All procedures were conducted under UK Home Office Procedure Project Licence Number PPL PP6108158. The mice were housed in individually ventilated cages or conventional cages under controlled environmental conditions in accordance with the UK Animals (Scientific Procedures) Act 1986 and the Home Office Code of Practice for the Housing and Care of Animals Bred, Supplied or Used for Scientific Purposes³⁶. Ambient temperature was maintained between 20–24 °C with relative humidity maintained between 45–65%. A daily 12 h/12 h light/dark cycle with half an hour of half-light to mimic dawn and dusk was provided to regulate circadian rhythms. Stocking density was carefully considered to ensure that mice were provided with sufficient floor space and to allow for provision of environmental enrichment in line with legislative standards. Environmental enrichment, including nesting material, refuges, wooden and disposable enrichment, was provided during weekly cage cleaning to encourage natural behaviours that are crucial for maintaining the health and wellbeing of the mice. Diet and water were provided ad libitum, and environmental parameters were continuously monitored to meet legislative and welfare requirements. Mice were routinely handled using refined handling techniques, including tunnel handling and cupping, to minimize stress and ensure high standards of animal welfare. The mice were inoculated via the intraspinal route with 5 µl of tenfold serial virus dilutions (a minimum of 8 mice per dilution) and monitored for clinical signs for 14 days (ref. ³⁷). The cell culture infectious dose (CCID₅₀) required to paralyze 50% of the mice (PD₅₀) was calculated using the Spearman–Karber method³⁸. While the transgenic mouse neurovirulence test (TgmNVT) does not reproduce natural infection through the oral route, it has been proven to accurately measure the neurovirulence of PV isolates and hence their potential for causing paralytic disease, showing very good correlation with the results using the gold standard monkey neurovirulence test and is recommended by the World Health Organization.

Statistical analysis

Simple linear regression and molecular clock-based inference of the nucleotide sequence data were done using GraphPad Prism (v.10.1.2) software. Regression analyses are exploratory/descriptive (mutation vs time), hence no inferential statistics (p values) were the focus. The dose required to cause paralysis in 50% of transgenic animals (PD₅₀) was calculated using the Spearman–Karber method using Microsoft Excel software. Poliovirus sequencing data were processed and analysed using Geneious (v.10.2.6).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.