Direct Diagnosis of Echovirus 12 Meningitis Using Metagenomic Next Generation Sequencing

Abstract

The current point-of-care diagnosis of enterovirus meningitis does not identify the viral genotype, which is prognostic. In this case report, more than 81% of an Echovirus 12 genome were detected and identified by metagenomic next-generation sequencing, directly from the cerebrospinal fluid collected in a 6-month-old child with meningeal syndrome and meningitis: introducing Echovirus 12 as an etiological agent of acute meningitis in the pediatric population.

1. Introduction

The highly diverse viral genus Enterovirus encompasses more than 300 genotypes [1]. Some Enterovirus members are responsible for central nervous system (CNS) infections, which clinical and epidemiological characteristics and prognosis, vary according to the precisely identified enterovirus, and some genotypes have been associated with a particular clinical severity and mortality [2]. The Enterovirus genotype is not routinely determined by molecular diagnosis assays detecting enterovirus RNA at the core and point-of-care (POC) laboratories [3,4,5,6]. Echovirus strains belonging to Enterovirus B species preferentially infect infants and young children [1,4] and are frequently involved in aseptic meningitis and encephalitis. Indeed, Enterovirus genotyping is most commonly based on partial VP1 gene Sanger sequencing using a generic standard protocol which is not commonly applied during the time of care [7].

In order to challenge Enterovirus genotyping at the POC laboratory, we herein developed a unique protocol to diagnose and genotype Enterovirus CNS infection directly from the cerebrospinal fluid (CSF) using metagenomic Next-Generation Sequencing (mNGS). This diagnosis approach is here illustrated by the diagnosis of Echovirus 12 meningitis in a child, a rarely reported situation in such setting [8,9].

2. Case Presentation and Methods

A 6-month-old girl born from twin pregnancy was admitted at the emergency department with fever (37.7 °C), cough and meningeal syndrome. She had an history of Respiratory Syncytial Virus bronchiolitis that led to hospitalization one month prior to meningitis. At the admission, CSF analysis after lumbar puncture showed a leukocyte count of 1 cell/mm³, protein at 0.14 g/L and glucose at 3.41 mmol/L. Microscopic analysis after Gram staining was negative. Investigation of the CSF at the POC laboratory using the Biofire FilmArray Meningitis/Encephalitis panel (bioMérieux, Marcy-l, Etoile, France) [6] was positive for Enterovirus.

In parallel, total RNA was manually extracted from 200 µL of CSF, following an in-house developed protocol, using QIAamp Viral RNA Mini Kit solutions (Qiagen, Hilden, Germany) for lysing, and washing steps, then total RNA was then purified using RNA specific magnetic Dynabeads (Life technology, Oslo, Norway). Briefly, 200 µL of CSF were incubated with 50 µL proteinase K (Qiagen) for 5 min at room temperature, then 300 µL AVL lysis buffer (Qiagen) were added and incubated for 15 min at room temperature. A 150-µL volume absolute ethanol (99%) were added to the lysis mix, 50-µL Dynabeads (40 mg/µL) were added and incubated for 15 min at room temperature. The Dynadeads/RNA complex washed twice with 850 µL AV1 solution (Qiagen), then two times with 450 µL AV2 solution (Qiagen) in the presence of 70% ethanol. After the second wash, Dynabeads were dried for 15 min at room temperature and eluted in a 60 µL-volume, then incubated for 3 min at 70 °C, followed by magnetic separation. Finally, 3 µL of RNaseOUT™ Recombinant Ribonuclease Inhibitor (Invitrogen, Illkirch, France were added to the purified RNA was and stored at −70 °C.

A 40-µL volume of total RNA was treated with ezDNase (Invitrogen, Illkirch, France) and concentrated with (Zymo Research, Irvine, CA, USA) kit, then eluted in 20 µL sterile water (Figure 1). The complementary DNA (cDNA) synthesis was performed using TaqMan kit according to the manufacturer protocol (Applied Biosystem, Foster City, CA, USA) in 50µL containing 19.25 µL eluted RNA, then 20 µL of the cDNA were used as a matrix for double strand synthesis, using 3 units of DNA Polymerase I, Large (Klenow) Fragment (BioLabs) in a 30 µL-volume. Double stranded DNA purified with Agencourt^® AMPure beads (Invitrogen) and eluted in 17 µL of 1x-sterile Tris-EDTA solution. Finally, 1 ng of cDNA was used for metagenomics Next-Generation Sequencing (mNGS) library preparation (Supplementary Materials 3), using Illumina Nextera XT paired-end protocol (Illumina, San Diego, CA, USA), as previously described [10,11] and sequenced on iSeq 100 instrument in a single 17.5-h run providing 2 × 150-base pair (bp) long reads. The NGS generated sequences were assembled by Spades on-line software available on Galaxy/Europe bioinformatics (https://usegalaxy.eu, accessed on 17 June 2020) and mapped with CLC Genomics Workbench software version 7.5.0 (Qiagen).

3. Results and Discussion

BLAST analysis of the contigs generated by mNGS after assembling reads with Spades (https://usegalaxy.eu, accessed on 17 June 2020), yielded as best match the Echovirus 12 strain Travis 2–85 gene (GenBank accession number AF295499.1). This strain was originally isolated from a 6-year-old healthy American male, caused cytopathic effect in tissue culture, was not neutralized by poliomyelitis antiserum, and failed inducing disease in infant mice [12]. As the complete genome of this strain was not available in the GenBank database, the Echovirus 12 complete genome, prototype Travis (X77708.1) was used as reference sequence for mapping of total reads by CLC Genomics Workbench software. The iSeq sequencing generated 114.818 reads, and 76.284 (66.4%) reads could be mapped on the Echovirus 12 genome, generating 7 contigs (GenBank accession number; PRJEB39568) covering 6,127 bp, hence 81.7% of this genome (Figure 1). Phylogenetic analysis based on VP1 gene and 3D polymerase encoding genes identified that these sequences belong to an Echovirus 12 as supported by bootstrap values of 98% and 97%, respectively (Figure 2, Supplementary Materials).

Using this approach of whole genome sequencing, Echovirus 12 was obtained in one-shot protocol directly from CSF sample. To confirm the NGS result, two Echovirus 12-specific primers were designated and used to target a 291-bp Echovirus 12 genome fragment. The same strain of Echovirus 12 was identified by sequencing of the amplified fragment at 98.44% sequence identity (not published data).

Echovirus 12 has been detected in patients with diarrhea and aseptic meningitis [9,13]. In this study, we described for the first time a clinical case of Echovirus 12 meningitis diagnosed by near whole genome sequencing of an Echovirus 12 directly from CSF by mNGS. This strategy proves high sensitivity in enterovirus detection, which warrants its introduction for routine diagnostic of enterovirus meningitis in addition to viral genomic surveillance and may even be considered for POC laboratories if the very fast Oxford Nanopore Technology is used. Current routine diagnostic targets a short sequence covering around 7% of the Enterovirus genome, so that is not indicative for Enterovirus genotyping [11]. The availability of the genome instead of a gene fragment necessarily provides improved information regarding typing of the viral strain and identifying mutations and recombinations, and correlating these genotypic features with epidemiological and/or clinical ones. Another important benefit of mNGS upon qPCR or Sanger sequencing is its versatility. Indeed, this is not a targeted approach but instead it is an opened approach that can potentially detect sequences from any virus or microorganism provided these are in sufficient amount; this is thus of particular interest in cases when no infectious agent could be diagnosed. In addition, NGS cost per clinical sample is currently in the same order of magnitude than that of Sanger sequencing. Indeed, we have estimated that current cost per nucleotide is EUR 0.3 for Sanger sequencing and EUR 0.03 for NGS sequencing with Illumina technology; such cost cannot be directly extrapolated to other laboratories due to highly variable cost components, among which the commercial policy of suppliers or the infrastructure of the laboratory. Overall, the present case exemplifies the powerfulness of mNGS in the setting of the routine diagnosis of meningitis.