key: cord-0319094-8ikyxx9o
authors: Reyes, A.; Carbo, E. C.; van Harinxma thoe Slooten, J.; Kraakman, M. E. M.; Sidorov, I. A.; Claas, E. C. J.; Kroes, A. C. M.; Visser, L. G.; de Vries, J. J. C.
title: Viral metagenomic sequencing in a cohort of international travellers returning with febrile illness
date: 2021-05-17
journal: nan
DOI: 10.1101/2021.05.13.21257019
sha: 1f2eebe157915c2b753794b686627669cdb4dc42
doc_id: 319094
cord_uid: 8ikyxx9o

Background Diagnosis of infections in returning international travellers can be challenging because of the broad spectrum of potential infectious aetiologies potentially involved. Viral metagenomic next-generation sequencing (mNGS) has the potential to detect any virus present in a patient sample and is increasingly being used for difficult to diagnose cases. The aim of this study was to analyse the performance of mNGS for viral pathogen detection in the clinical setting of international travellers returning with febrile illness. Methods Thirty-eight serum samples from international travellers returning with febrile illness and presenting at the outpatient clinic of the Leiden University Medical Center in the Netherlands in the time period 2015-2016 were selected retrospectively. Samples were processed for viral metagenomic sequencing using a probe panel capturing all known vertebrate viruses. Bioinformatic analysis was performed using Genome Detective software for metagenomic virus detection. Metagenomic virus findings were compared with viral pathogen detection using conventional methods. Results In 8 out of the 38 patients (21%), a pathogenic virus was detected by mNGS. All viral pathogens detected by conventional assays were also detected by mNGS: dengue virus (n=4 patients), Epstein-Barr virus (n=2), hepatitis B virus (n=1). In addition, mNGS resulted in additional pathogenic findings in 2 patients (5%): dengue virus (n=1), and hepatitis C virus (n=1). Non-pathogenic viruses detected were: GB virus C (n=1) and torque teno viruses (n=3). High genome coverage and depth using capture probes enabled typing of the dengue viruses detected. Conclusions Viral metagenomics has the potential to assist the detection of viral pathogens and co-infections in one step in international travellers with a febrile syndrome. Furthermore, viral enrichment by probes resulted in high genome coverage and depth which enabled dengue virus typing.

Background Diagnosis of infections in returning international travellers can be challenging because of the broad spectrum of potential infectious aetiologies potentially involved. Viral metagenomic next-generation sequencing (mNGS) has the potential to detect any virus present in a patient sample and is increasingly being used for difficult to diagnose cases. The aim of this study was to analyse the performance of mNGS for viral pathogen detection in the clinical setting of international travellers returning with febrile illness.

Thirty-eight serum samples from international travellers returning with febrile illness and presenting at the outpatient clinic of the Leiden University Medical Center in the Netherlands in the time period 2015-2016 were selected retrospectively. Samples were processed for viral metagenomic sequencing using a probe panel capturing all known vertebrate viruses. Bioinformatic analysis was performed using Genome Detective software for metagenomic virus detection. Metagenomic virus findings were compared with viral pathogen detection using conventional methods.

In 8 out of the 38 patients (21%), a pathogenic virus was detected by mNGS. All viral pathogens detected by conventional assays were also detected by mNGS: dengue virus (n=4 patients), Epstein-Barr virus (n=2), hepatitis B virus (n=1). In addition, mNGS resulted in additional pathogenic findings in 2 patients (5%): dengue virus (n=1), and hepatitis C virus (n=1). Non-pathogenic viruses detected were: GB virus C (n=1) and torque teno viruses (n=3). High genome coverage and depth using capture probes enabled typing of the dengue viruses detected.

Accurate diagnosis of travel-associated febrile illness in the returning traveller can be challenging, because of the broad spectrum of viral aetiologies potentially involved (1, 2) . Identification of potential viral pathogens is important for clinical management and epidemiological reasons.

Metagenomic next-generation sequencing (mNGS) has the potential to detect any known or new pathogen in one single run, in contrast to conventional targeted methods such as PCR. In addition, molecular techniques targeting specific pathogens are dependent upon matching specific primers, leaving variant pathogens unidentified. Finally, emerging pathogens that have not associated with a specific clinical syndrome before, for example encephalitis caused by astroviruses, will be included in the metagenomic width of detection (3) .

Since the amount of pathogen in a sample is relatively low and human background is high, several strategies have been applied to increase the sensitivity of mNGS based on physical or enzymatic preprocessing of samples for human DNA depletion (3) (4). Another strategy is hybridization enrichment, with the application of a virome probe panel that targets all known vertebrate viruses to increase the sensitivity of virus detection and characterization (5, 6) . Viral enrichment by capture probes increased viral sequence read counts in cerebrospinal fluid samples 100 -10.000 fold, compared to unenriched sequencing (7) . These probe capture panels can also be helpful in the detection of genetic variants and novel viruses up to approximately 40-58% different at nucleotide level from the genome references used in the probe library design (5) (6) (7) (8) .

Viral mNGS is increasingly being applied directly on different types of samples from patients for pathogen detection in undiagnosed cases, both in retrospective studies (9) and prospective ones (10) with a wide range of proportion of additional findings. A striking example is the rapid and impactful metagenomic analysis of SARS-CoV-2 in 2019 (11) . The aim of the current study was to investigate the utility of viral enhanced metagenomic sequencing (mNGS) as a diagnostic tool for viral infections in the returning traveller with febrile illness.

Retrospectively, a cohort of international travellers with febrile illness upon their return was studied. 

The procedure for metagenomic detection using a viral probe capture panel for clinical samples has been validated previously (7) . Prior to nucleic acid extraction, serum samples were spiked with fixed amounts of non-human pathogenic viruses as internal sequencing RNA and DNA controls: Equine Arteritis Virus (EAV) and Phocid alpha-herpesvirus (PhHV-1). Nucleic acid extraction was performed using MagNApure 96 DNA and Viral NA Small volume extraction kit on the MagNApure 96 instrument (Roche, Germany), with 200 µl of serum sample input and 100 µl output eluate. STAR Buffer was used as negative control for the entire workflow from nucleotide isolation throughout sequencing. Library preparation was carried out with the NEBNext® Ultra II Directional RNA Library Prep kit for Illumina® . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. and NEBNext® Multiplex Oligos for Illumina® (unique dual index primers pairs, E6440) with 10 ul of pre-concentrated eluate and following a modified version of the protocols for use with "purified mRNA or rRNA Depleted RNA" as described previously (13, 14) . Nuclease free water was used as a library preparation control (upstream negative control). Libraries were combined in pools of three libraries from samples plus one negative control for viral capture probe enrichment. Workflow User's Guide (Roche) was followed with several in-house adaptations to the manufacturers protocol as described previously (7) . Subsequently, the target regions were captured by hybridization each pool of four sample libraries with the SeqCap EZ probe pool (6) overnight. The HyperCap Target Enrichment kit and Hyper Cap Bead kit were used for washing and recovery of the captured DNA.

Finally, post-capture PCR amplification was performed using KAPA HiFi HotStart ReadyMix (2X) and

Illumina NGS primers (5 uM), followed by DNA purification using AMPure XP beads. Final products were sequenced using the Novaseq6000 platform (Illumina, San Diego, California, USA), obtaining up to 10 million of 150bp paired-end reads per sample (GenomeScan B.V. Leiden, the Netherlands).

Primary data analysis, bcl conversion and demultiplexing was performed with bcl2fastq (Illumina).

After quality pre-processing, sequencing reads were taxonomically classified and subtyped with . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Table 1 shows the number of reads, genome coverages and dengue genotypes found in relation with the conventional diagnostic test performed. Genome coverage bars are shown in Figure 1 A.

Using mNGS, two additional pathogenic viruses were detected (Table 2, Figure 1 B) : one additional dengue virus and one hepatitis C virus. These infections were confirmed as true positives by qPCR. In the patient in whom the dengue virus infection was diagnosed mNGS, the original dengue NS1 antigen screening test was negative at the time of the visit at the outpatient clinic.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. ; https://doi.org/10.1101/2021.05.13.21257019 doi: medRxiv preprint

The following non-pathogenic viruses were detected: GB virus C (one patient, 1.821.303 sequence reads) and torque teno viruses in three patients: type 6 (147 reads), 24 (892 reads) and 18 (139 reads), coverage bars are represented in Figure 1B . The GB virus C was found in co-infection with dengue virus.

The four dengue virus infections could also be typed and phylogenetic analysis was performed ( Figure   2 ). Three of the infections were classified as serotype 1 (genotype I, IV and V) and one as serotype 3 (genotype I). The HCV positive finding was classified as genotype 2 by alignment with reference NC_009823.1 strain.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. ; https://doi.org/10.1101/2021.05.13.21257019 doi: medRxiv preprint

In the current study, viral mNGS was successful in detecting six previously diagnosed infections and revealed two new findings (5%) in 38 serum samples from returning travellers with febrile illness. In similar studies using mNGS in returned travellers, new and diverse findings have been reported but none of them used a capture panel for pathogen detection (1, 17) . Application of capture probes results in higher coverage of the genomes detected and in more reliable sequences because of an increased sequencing depth (6, 7). As a result, subsequent typing and phylogenetic analysis could be performed using the consensus genome sequences after de novo genome assembly.

The diagnosis dengue virus was rejected in one patient after a negative dengue NS1 antigen rapid test upon outpatient visitation, whereas dengue sequences were detected by mNGS and afterwards confirmed by PCR. Dengue virus antigen tests are known for their lower sensitivity after one week of onset of disease and potentially in case of secondary infections with lower loads (18) .

Serotypes and subsequent genotypes of dengue virus were available after sequence analysis using a dengue virus typing tool based on E gene sequences (1,485bp) for phylogenetical classification.

Dengue serotypes differ in more than 30% in their amino acid sequences (19) . Dengue virus typing is not relevant for pathogen detection but it is for epidemiological surveillance of infecting strains. Furthermore, dengue virus typing can be of use for differentiating isolates in case of secondary infections, while severity of primary or secondary infections are related to serotypes (20) . In our study, all the dengue virus serotypes could be properly identified as a result of increased genome coverage and depth due to effective enrichment.

The detection of hepatitis C virus in serum of a patient highlights the potential of mNGS to diagnose infectious diseases beyond the differential diagnostic list of expected pathogens. The patient was a 60-69 years-old person presenting with fever at the outpatient clinic three days after the return from Suriname and French Guiana. She was tested for dengue antigenemia, and Zika and Chikungunya antibodies which were all negative. There was no suspicion of HCV infection since signs and symptoms . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. ;  were not consistent and transaminases levels were in the range of normality. Although HCV infection is not a cause of febrile illness in the returning traveller, the detection of unsuspected pathogens is crucial for clinical management, therapy administration and prevention of transmission. GB virus C, formerly known as hepatitis G virus, is a lymphotropic RNA virus of the Pegivirus family, it is related to hepatitis C virus and was thought to cause chronic hepatitis in the past, however nowadays is considered non-pathogenic. It has been suggested to have a controversial protective role in HIV-1 infected patients, of which the underlying mechanisms are not fully known (21) . A recent study associated GB virus C metagenomic detection with an attenuation of disease caused by Ebola virus through immunomodulation (22) . We did not find any report about GBV-C and dengue coinfection or association with dengue pathogenicity. Our finding may rise questions about the possible immunomodulating role of GBV-C virus in dengue when the two viruses are present in blood simultaneously since dengue virus can replicate in lymphocytes.

The use of a virus capture panel has been reported to increase significantly the number of reads and coverages generated in sequencing platforms (7) . The high sensitivity of NGS when combined with the viral capture panel does not only enable the finding of clinically unsuspected pathogens, but also may provide a negative predictive value within the range of use for clinical practice (7, 14) . The increase in viral sequence reads also enables subsequent typing or detection of potential antiviral resistance. A recent review highlighted the sensitivity of hybridization-based enrichment techniques for viral genomes screening and proposed its deployment as an alternative diagnostic tool when traditional methods fail to detect a pathogen, even when viral genomes differs <40% from probe sequences (23) .

In conclusion, the application of viral metagenomics in this study provided the additional detection of two unsuspected viral pathogens and one first report coinfection. Metagenomic sequencing has the potential to diagnose a viral febrile syndrome in the returned traveller with the use of a single test.

The use of a broad viral capture panels makes this method more sensitive and generates enough reads and coverages for reliable pathogen identification and typing.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. ; https://doi.org/10.1101/2021.05.13.21257019 doi: medRxiv preprint

We want to thank the Spanish Society of Microbiology and Infectious Diseases (SEIMC) for the mobility grant provided for contributing to this work in the clinical microbiology laboratory of LUMC, The Netherlands.

None . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 17, 2021. ; . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The last dengue virus infection was not detected by means of conventional lateral flow assay. The letters correspond to mNGS findings # 1, 2, 3 and 7 respectively in Tables 1 and 2) .

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 17, 2021. ; https://doi.org/10.1101/2021.05.13.21257019 doi: medRxiv preprint

Travelassociated infection presenting in Europe (2008-12): an analysis of EuroTravNet longitudinal, surveillance data, and evaluation of the effect of the pre-travel consultation

Approach to Fever in the Returning Traveler

Astrovirus VA1/HMO-C: an increasingly recognized neurotropic pathogen in immunocompromised patients

Recommendations for the introduction of metagenomic high-throughput sequencing in clinical virology, part I: Wet lab procedure

Enhanced virome sequencing using targeted sequence capture

Virome Capture Sequencing Enables Sensitive Viral Diagnosis and Comprehensive Virome Analysis

Improved diagnosis of viral encephalitis in adult and pediatric hematological patients using viral metagenomics

Coronavirus discovery by metagenomic sequencing: a tool for pandemic preparedness

Clinical metagenomics

Clinical Metagenomic Sequencing for Diagnosis of Meningitis and Encephalitis

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Sentinel surveillance of imported dengue via travellers to

Retrospective Validation of a Metagenomic Sequencing Protocol for Combined Detection of RNA and DNA Viruses Using Respiratory Samples from Pediatric Patients

Genome Detective: an automated system for virus identification from high-throughput sequencing data

Centrifuge: rapid and sensitive classification of metagenomic sequences

Metagenomic nextgeneration sequencing aids the diagnosis of viral infections in febrile returning travellers

Performance of Dengue Diagnostic Tests in a Single-Specimen Diagnostic Algorithm

The global distribution and burden of dengue

Meta-Analysis of Dengue Severity during Infection by Different Dengue Virus Serotypes in Primary and Secondary Infections

GB virus C: the good boy virus?

GB virus C coinfections in west African Ebola patients

Current Trends in Diagnostics of Viral Infections of Unknown Etiology