Laboratory diagnosis of nonpolio enteroviruses: A review of the current literature

Highlights

• •Laboratory diagnosis of non-polio enterovirus infections is important in determining a patient's prognosis and guiding clinical management, for surveillance and for the investigation of outbreaks.
• •Laboratory diagnosis of non-polio enteroviruses is mainly based on molecular techniques, but classical virus-isolation techniques are still used in references laboratories.
• •An important part of diagnosis and surveillance of enteroviruses infections include viral typing by VP1 gene sequencing and/or VP4-VP2 sequencing using conventional Sanger technique, and more recently, full-genome next-generation sequencing.

Abstract

Infections by nonpolio enteroviruses (EVs) are highly prevalent, particularly among children and neonates, where they may cause substantial morbidity and mortality. Laboratory diagnosis of these viral infections is important in patient prognosis and guidance of clinical management. Although the laboratory diagnosis of nonpolio EVs is mainly based on molecular techniques, classical virus-isolation techniques are still used in reference laboratories. Other techniques, such as antigen detection and serology, are becoming obsolete and rarely used in diagnosis. An important part of diagnosis and surveillance of EV infections is viral typing by VP1 gene sequencing using conventional Sanger technique and more recently, full-genome next-generation sequencing. The latter allows the typing of all EVs, better investigation of EV outbreaks, detection of coinfection, and identification of severity markers in the EV genome.

1. Introduction

Human enteroviruses (HEVs) constitute a large group of viral pathogens that cause diseases, which range from mild respiratory infections to paralysis or severe central nervous system infection [1]. Enteroviruses (EVs) are among the most common viruses infecting humans. They consist of ubiquitous, small (∼30 nm in diameter), non-enveloped, positive-sense, single-stranded RNA viruses belonging to the Picornaviridae family [1]. EVs have been recognized and characterized since the first electron microscopy images of the poliovirus, which were captured in 1952 by Reagan et al [2]. Coxsackievirus was the first human EV discovered after polioviruses, and it was named after the first geographical site of its isolation in New York. The name echovirus (enteric, cytopathogenic, human, and orphan virus) was selected for these viruses, which are associated with various clinical manifestations, such as gastroenteritis, meningitis, and respiratory illness [3].

The classification based on specific clinical manifestations becomes difficult with every newly discovered EV. Therefore, since 1974, novel EVs with serological properties have been numbered by their order of identification and more recently, by their molecular type [4].

Over a hundred types of EVs exist, and most of them are classified into four species based on molecular and biological characteristics: “enterovirus A (EV-A), enterovirus B (EV-B), enterovirus C (EV-C), and enterovirus D (EV-D)” (Table 1) [5].

Table 1.

Human enterovirus (EV) species and main types (adapted from Pons-Salort et al [5]).

Human enterovirus species	Types
Enterovirus A	Coxsackievirus A2-A8, A10, A12, A14, A16 Enterovirus A71, A76, A89-92, A114, A119, A120, A121
Enterovirus B	Coxsackievirus A9, B1-6 Echovirus 1–9, 11–21, 24–27, 29–33 Enterovirus B69, B73-75, B77-88, B93, B97, B98, B100, B101, B106, B107, B110-113
Enterovirus C	Coxsackievirus A1, A11, A13, A17, A19-A22, A24 Poliovirus 1–3 Enterovirus C95, C96, C99, C102, C104, C105, C109, C113, C116-118
Enterovirus D	Enterovirus D68, D70, D94, D111, D120

A single open reading frame encodes approximately 2,185 amino acids (range: 2,138–2,214) flanked by 5′- and 3′-non-translated regions (5′NTR and 3′NTR) of approximately 750 and 1,000 nucleotides, respectively (Fig. 1). The 5′NTR harbors secondary RNA structural elements needed for replication, and the 3′NTR is needed for replication control [6], [7]. All EVs have a similar genomic organization (7.2–8.5 Kb). The coding region is subdivided into the following three regions (from 5′ to 3′): the precursor 1 (P1) region that encodes structural capsid proteins (VP4, VP2, VP3, and VP1); the P2 and P3 regions that encode non-structural proteins, such as the RNA-dependent RNA polymerase, proteases, and other necessary proteins, needed for intracellular replication [1].

Fig. 1.

Organization of Enterovirus genome. Top: schematic of the viral RNA genome, with the genome linked protein VPg at the 5′ end, the 5′ non-translated region, the protein coding region, the 3′ non-translated and the poly(A) tail. Bottom: Processing pattern of Enterovirus protein: capsid proteins (VP1 – VP4), non-structural proteins 2A, 2B, and 3A, and conserved non-structural proteins polymerase 3D^pol, protease 3C^pro, helicase 2C. Amino acid sequence variation across the enterovirus genome are deduced from polyprotein amino acid sequences alignment from reference strains of human enteroviruses (darker boxes: more conservative; lighter boxes: more variable). The binding regions for the most commonly used primers are marked below (adapted from Racaniello et al. [1]).

EV infections are prevalent worldwide, although rates of infections vary by location and are seasonal. Individuals living in regions with temperate climates experience considerably higher rates of EV infections in the summer and fall months (from June to October) [8]. Meanwhile, transmission occurs throughout the year in tropical and subtropical areas [9]. Although EV infections occur in all age groups, the highest rates of infection and diseases are observed in infants and young children [8].

Human EV infections are usually acquired directly or indirectly by contact with the virus, which is shed in feces or the upper respiratory tract [6]. The incubation period for EV infections may vary based on the clinical syndrome but generally lasts for 7–14 days (range: 3–35 days) [10]. Reinfection with the same type of EV can occur despite previous immunity [10].

Most EV infections observed in clinical practice are asymptomatic/benign or result in mild symptoms, such as uncomplicated Hand, Foot, and mouth disease (HFMD), herpangina, and pleurodynia [6]. The most common symptomatic manifestation of EV infection is an acute, non-focal febrile illness that mainly affects infants aged < 1 year [11], [12], [13]. The same viruses causing these common syndromes have been incriminated in severe and life-threatening infections, such as aseptic meningitis, encephalitis, acute myocarditis, myositis, neonatal sepsis, and acute flaccid paralysis [14], [15]. Nonpolio EV infections have also been implicated in cases of Guillain–Barre syndrome, acute transverse myelitis, and cerebellar ataxia [16], [17], [18].

Protective immunity against EV infection is type specific [10]. Although humoral- and cell-mediated immune response mechanisms occur following EV infection, a type-specific antibody is the most important factor in limiting diseases and eradicating the virus [6]. Frequent recombination and mutations in EVs are the primary mechanisms for their high rate of evolution. This condition has contributed to the rapid response and adaptation of EVs to new environmental challenges. The molecular epidemiology of EVs is well studied compared with other enteric viruses because they are easy to isolate in cell culture [7]. Recent advances in molecular biology have led to improved diagnostics of EV and typing of EV strains [19], [20], [21], [22]. This review covers the currently available techniques and advances in the diagnosis and typing of EV. Rhinovirus species and “non-human” EV species E–J are outside the scope of this review.

2. Laboratory diagnosis on collected clinical samples

Diagnosis of EV infection can be complicated and challenging. Such difficulty is due to the biology and epidemiology of EV infections and limitations of current diagnostic techniques [6]. Diagnosis of a specific EV infection requires the detection of the virus itself, specific antibodies, or viral genome in patient samples [10].

Given the myriad of clinical manifestations of EV infections, confirmation of diagnosis may be necessary to reduce hospitalization, antibiotic use, and additional expensive diagnostic tests, which are often performed to exclude or treat other conditions [23].

In addition to their improved sensitivity and quantitative results, molecular tests can be used to identify specific EV types. Sequencing of the EV VP1 capsid coding region correlates well with the neutralization serotyping and represents an ideal target for EV detection and identification tests [24]. Specific EVs, notably EV-A71 and EV-D68, have been associated with outbreaks, and they occasionally result in significant morbidity and mortality in Asia, Europe, and Northern America [25], [26], [27], [28]. EV-A71 has been involved in severe neurological manifestations, ranging from aseptic meningitis to acute paralysis and brainstem encephalitis associated with systemic features, such as severe pulmonary edema and shock [25]. Meanwhile, EV-D68 has been implicated mainly in severe respiratory illnesses, including severe asthma and bronchiolitis, in children [26], [27].

2.1. Clinical samples collection

Samples should be collected immediately after the onset of symptoms. The most common clinical samples include stool, cerebrospinal fluid (CSF), nasopharyngeal aspirate/swab, vesicular fluid, bronchoalveolar lavage, blood, conjunctival swab, biopsy specimen, and urine [28].

Multiple samples are collected from different sites in case of a suspected EV infection.[28]. As severe infection has been described in children, EVs can be detected in the plasma, stool, respiratory secretions, CSF, and biopsies [8]. Further, viral loads are usually higher in stool and respiratory specimens compared with those in CSF [29]. Therefore, these samples are recommended for EV identification [28].

2.2. Direct methods for primary diagnosis

2.2.1. Viral antigen detection tests

Immunohistochemistry detection has been described primarily not only for the diagnosis of myocardial infections but also for other tissues, such as the spleen, lung, kidney, intestine, bone marrow, and pancreas [30], [31], [32]. Detecting viruses in tissues can be important for understanding viral pathogenesis. Most of these detection tests locate the conserved region of the VP1 peptide [30], [31], [32]. Direct detection of EV antigens in stool samples by enzyme immune-assay (EIA) is as sensitive as traditional cell culture but only achieves 58% sensitivity compared with reverse-transcription polymerase chain reaction (RT-PCR) [33]. Available commercialized tests include either immunofluorescence tests (IMAGEN^TM Enterovirus, Thermo-fisher) or rapid POCT chromatographic immunoassays (CDIA Enterovirus Antigen Test, Creative Diagnostics). The main problem with these assays is their failure to type EVs; thus, they are rarely used in common practice [10].

2.2.2. Nucleic acid amplification tests (NAATs)

The application of molecular biology techniques has significantly changed approaches to EV diagnostics. The most common use of PCR for EV diagnostics is the direct detection of EV genome in cell cultures, clinical specimens, and biopsy tissues [24]. In contrast to antigen detection and viral culture, NAATs have been applied successfully in the diagnosis of EV infections. These tests are extremely sensitive, and they attain 100 % specificity in the absence of laboratory cross-contamination. RT-PCR techniques for EV detection in clinical samples are more sensitive than cell culture [34], [35], [36]. These assays are mainly based on the detection of viral RNA, which targets the highly conserved 5′NTR region in EVs [36], [37], [38], [39]. Different primers targeting this region have been used with various sequences and product sizes. Given the genetic conservation of this region, sequencing prevents the accurate identification of EV type [40] and should not be used for typing [28].

Correct sample preparation prior to nucleic acid extraction is primordial, especially for fecal specimens. To avoid PCR inhibitors, medical experts should disaggregate fecal samples in normal saline and clarify them by centrifugation [28]. The addition of an internal control during extraction is important to identify the presence of these inhibitory compounds (blood and dietary supplement) [41].

Several commercial real-time PCR kits are available for EV detection in CSF (Xpert EV, Cepheid; NucliSENS EasyQ Enterovirus, BioMérieux) and respiratory specimens (FilmArray Respiratory Panel EV/HRVb, BioFire; xTAG Respiratory Viral Panel EV/HRV; Luminex). However, more data are needed to confirm the sensitivity of commercial 5′NTR-based assays for the possible detection of all EV strains [10], [39], [42].

As with all PCR-based techniques, the sensitivity of RNA amplification from biological specimens is extremely variable and depends on the nature of the specimen. Some specimens have modest sensitivity (e.g., CSF), and others exhibit excellent sensitivity (e.g., stool) [6]. Extensive studies of the capsid region of EV isolates have led to the design of primers that selectively target single types. This progress was accomplished using inosine in the primer synthesis despite the high rate of analogy in nucleotide substitutions [19], [24].

PCR is useful for the diagnosis of pediatric EV meningitis, non-specific fever illnesses, and neonatal infections. In related studies, PCR testing of CSF, serum, and throat specimens from children with acute illnesses showed > 90% sensitivity and high specificity; meanwhile, urine testing showed less sensitivity [10]. Patients suffering from aseptic meningitis, encephalitis, or myelitis have detectable virus in stool and respiratory secretions for periods longer than those for CSF, where virus remains undetectable on occasions. EV RNA can be detected in CSF by PCR in the majority of meningitis cases but inconsistently in encephalitis. EV excretion in throat and stool samples is prolonged, but detection is not a confirmation of infection [28]. Any detection in stool or throat samples should be carefully interpreted because it does not apply causation and may simply reflect prolonged excretion or carrier status [43]. Droplet digital PCR is a useful tool for the identification and detection of EV 71 [44] and quantification of EVs in untreated sewage waters [45].

2.3. Viral isolation and identification

The isolation of EV in cell culture has historically been the primary method for diagnosis mainly because most EV strains grow readily [10]. Traditional methods for EV detection and characterization rely on time-consuming and labor-intensive procedures of viral isolation in cell culture and neutralization by reference antisera [6].

The viral culture procedure involves the inoculation of appropriate specimens onto susceptible cell lines. The World Health Organization (WHO) suggested L20B, RD, and former HEp-2 cell lines for poliovirus isolation in a global poliovirus laboratory network. At least three different cell lines are often used for EV culture; these cell lines include monkey kidney cells, A549 (human lung carcinoma), buffalo green monkey kidney (BGMK), human embryonic lung fibroblasts, epithelial, and rhabdomyosarcoma cell lines [6], [15], [46]. Transgenic and co-cultured cell lines can increase the sensitivity of EV recovery and eliminate the need for multiple shell vials. BGMK cells transfected with human decay-accelerating factor have been co-cultured with Caco-2 cells (human colon-adenocarcinoma cell line) [47]. These mixed cell cultures showed higher sensitivity for recovery of EV from clinical samples than single cell lines [47], [48].

In routine diagnostic testing, EV growth is detected by its cytopathic effects, which include visible rounding, shrinking, nuclear pyknosis, refractility, and cell degeneration [6]. The cytopathic effect is confirmed as that of a specific EV by neutralization with type-specific antisera, immunofluorescence with type-specific monoclonal antibodies, or polymerase chain reaction (PCR) coupled with sequencing [6].

In general, virus-isolation techniques possess high sensitivity and specificity. The highest yield in neonatal EV infections can be obtained using viral culture specimens, such as those from the rectum or stool (91%–93% positive), CSF (62%–83% positive), and nasopharynx or throat (52%–67% positive) [49]. Serum and urine cultures provide lower yields (24%–74%) [10].

Viral isolation faces significant limitations. First, this process requires expertise and is relatively expensive. Second, no single optimal cell line can be used to grow all EVs. Most EVs can be isolated in cell cultures of mammalian cell lines; however, some EVs exhibit poor growth [50]. Most coxsackie A viruses, except CVA9 and CVA16, grow poorly in cultures [51]. Third, the sensitivity of culture is further limited by neutralization of antibodies in diagnostic specimens and inadequate specimen collection or handling. Lastly, growth in culture can be slow, lasting for 3–8 days [6].

However, cell culture is still applied in many applications and can be used to confirm new variants of EV, detect EV coinfection (more than one EV strain), type EVs by neutralization, and assess the efficacy of antiseptic antiviral agents.

2.4. EV typing

2.4.1. Sequencing by Sanger technique

Single detection of an EV during an outbreak, even in a patient with typical clinical manifestations, does not ensure an epidemiological link with an outbreak without molecular typing [7]. EVs are among the most rapidly changing viruses [52]. A goal in virus identification is determining the viral genome sequence. Originally, EV typing was based on sero-neutralization. However, nowadays, the sequencing part of the VP1 capsid protein gene has been the pillar for EV typing and is widely used worldwide [53]. This step is performed directly by PCR in conjunction with nucleic acid sequencing of a full (about 900 nucleotides) or partial VP1 genome region (a minimum of 350 nucleotide sequences are required) and is recommended by WHO [42]. The type of an unknown isolate is inferred by comparing the VP1 sequence with a nucleotide sequence database, e.g., GenBank and RIVM [6], [54]. The VP1 gene encodes an important type-specific neutralization epitope, and therefore, its sequence correlates very well with the classical type of classification [55], [56].

VP1 nucleotide identity of > 75% (amino acid sequence>85%) between an EV isolate and a prototype strain may be used to establish the isolate type [6], [19]. The best-match nucleotide sequence identity of < 70% may indicate a distinct new type. Any sequence identity between 70% and 75% requires further characterization before final identification [6]. This approach reduces the time required to type EV strains and can be used to type difficult or untypable isolates using sero-neutralization. This method also helps in the rapid determination of the epidemiological relationship of viruses isolated during an outbreak [6]. The complete VP1 sequence (∼900 nucleotides) is used for formal type identification and required for the characterization of new types [28].

Given a good reference dataset, the VP1 genome region is commonly used for phylogenetic analysis. The VP2 genome region is also applied for typing and phylogenetic analysis, although less frequently [7]. Assays targeting other genomic regions (mainly capsid proteins VP4/VP2) can be used if VP1 sequencing fails; some of these tests display increased sensitivity [6], [28], [56], [57], [58], [59], [60]. However, VP4/VP2-based tests can limit the comparability of data between laboratories because of a small database [28]. In addition, the VP4 region may include a recombination hotspot at the 5′NTR-VP4 junction, and recombination events within the VP4 domain may compromise phylogenetic analysis [7]. Other genome regions are less suitable for phylogenetic analysis due to ubiquitous recombination outside the VP2/VP1 region [61] and are used to highlight the circulation of the recombinant forms of EVs [62].

According to the WHO EV surveillance guidelines, Sanger sequencing technique for EV typing is based on a semi-nested PCR (snPCR) amplification of the VP1 region using specific primers [53] (Supplementary Material 1). Subsequently, the Nix et al.'s protocol was improved by conducting snPCR separately for all EV, EV-A, EV-B, EV-C, and EV-D [59], [62], [63]. These techniques include other sets of primers [59], [62] or their combination [63], [64] (Table 2).

Table 2.

Primers used for amplification and sequencing of VP1 (in pairs, where first primer is forward).

Primer	Use fora	Specificityb	Sequence (5ˊ-3ˊ)c	Reference
AN32	RT	All EV	GTY TGC CA	Nix et al., 2006 [19]
AN 33	RT	All EV	GAY TGC CA
AN 34	RT	All EV	CCR TCR TA
AN 35	RT	All EV	RCT YTG CCA
224	1st round	All EV	GCI ATG YTI GGI ACI CAY RT	Nix et al., 2006 [19]
222	1st round	All EV	C ICC IGG IGG IAY RWA CAT	Oberste et al., 2003 [89]
HEVBS1695	1st round, 2nd round	EV-A	CTTGTGCTTTGTGTCGGCRTGYAAYGAYTTYTCWG	Mirand et al., 2016 [59]
EV2C	1st round	EV-A	CAATACGGCATTTGGACTTGAACTGTATG
EV-OS	1st round	All EV	TTA AAA CAG CCT GTG GGT TG	Leitch et al., 2009 [62]
EV-OAS	1st round	All EV	ATT GTC ACC ATA AGC AGC CA
EV-A-OS	1st round, 2nd round, SR	EV-A	CCN TGG ATH AGY AAC CAN CAY T
EV-A-OAS	1st round	EV-A	GGR TAN CCR TCR TAR AAC CAY TG
EVB-OS	1st round	EV-B	GGY TAY ATN CAN TGY TGG TAY CAR AC
EV-B-OAS	1st round	EV-B	GGT GCT CAC TAG GAG GTC YCT RTT RTA RTC YTC CCA
AN89	2nd round, SR	All EV	CCA GCA CTG ACA GCA GYN GAR AYN GG	Nix et al., 2006 [19]
AN88	2nd round, SR	All EV	TAC TGG ACC ACC TGG NGG NAY RWA CAT
HEVBR132	2nd round	EV-A	GGTGCTCACTAGGAGGTCYCTRTTRTARTCYTCCCA	Mirand et al., 2016 [59]
5NCS663	2nd round, SR	EV-B	GCGGAACCGACTACTTTGGGTGTCCGTGTTTC
HEVR436	2nd round, SR	EV-B	CCCATGTCAGTCAGCGCATCIGGIARITTCCAIYACCAICC
EV-A-IAS	2nd round, SR	EV-A	GAN GGR TTN GTN GKN GTY TGC CA	Leitch et al., 2009 [62]
EV-B-IS	2nd round, SR	EV-B	CTT GTG CTT TGT GTC GGC RTG YAA YGA YTT YTC WG
EV-B-IAS	2nd round, SR	EV-B	TCY TCC CAC ACR CAV TTY TGC CAR TC
EV-IS	2nd round, SR	All EV	GGN AYY YTW GTR CGC CTG TTT T
EV-IAS	2nd round, SR	All EV	CAC CCA AAG TAG TCG GTT CCG C
189	2nd round, SR	EV-A	CAR GCI GCI GAR ACI GGN GC	Oberste et al., 2000 [90]
187	2nd round, SR	EV-B	ACI GCI GYI GAR ACI GGN CA
188	2nd round, SR	EV-C&D	CAR GCI GCI GAR ACI GGN
AN232	SR	All EV	CCAGCACTGACAGCA	Nix et al., 2006 [19]
AN233	SR	All EV	TACTGGACCACCTGG
P1S1695S	SR	EV-A	CTTGTGCTTTGTGTCGGC	Mirand et al., 2016 [59]
P1R132S	SR	EV-A	GGTGCTCACTAGGAGGTC

a) Abbreviations: RT, reverse transcription; 1^st round, first round of nested polymerase chain reaction (PCR); 2^nd round, second round of nested PCR; SR, sequencing reaction.

b) Abbreviations: EV-A, enterovirus A; EV-B, enterovirus B, EV-C, enterovirus C, EV-D, enterovirus D.

c) Sequences are written using the IUPAC nucleotide ambiguity codes, where I = deoxyinosine; N = G, A, T or C; Y = C or T; W = A or T; R = A or G.

Despite its sensitivity, when Nix et al.'s protocol is applied directly to clinical samples, non-typed EV isolates can subsequently be detected after the samples are subjected to cell culture using a susceptible and permissive cell line [63]. This approach was used by Faleye et al. to improve protocol performance. Consequently, these authors added species-specific primers in the second round PCR and used an RD cell line additionally; they noticed the following results. First, Nix et al.'s protocol, although extremely sensitive for detecting EV genomes, will often mask the presence of more than one EV isolate per sample. Second, even when a clinical sample is negative for EVs by pan-EV VP1 screen-based primers, the designed species-specific screen (using the first-round product as a template) still types EVs successfully at times. Finally, non-detected EV types in native clinical samples will emerge after cell culture. Thus, the pan-EV VP1 screen (Nix et al.'s protocol) is fallible, and virus concentration significantly influences the detectability of virions present in samples.

However, a unique set of primers is hardly sufficient for the amplification of all EV types, and the use of degenerate inosine primers as the ones in Nix's protocol may carry non-specific amplification [19], [55].

2.4.2. Next-generation sequencing (NGS)

Conventional Sanger sequencing has been the gold standard of the analysis of pathogens for over three decades. Sequencing of the whole EV genome (∼7,800 nucleotides) by the Sanger method is time consuming, difficult to scale-up, and a laborious procedure because the use of multiplex samples is not possible. However, with the expansion of NGS techniques, the demand for low-cost high-throughput sequencing has increased. NGS has high sensitivity and specificity, and a great sequencing depth as a single nucleotide is sequenced for at least 50 times. The PCR cycle threshold value of 30 is an assay limit required for NGS, and samples with low viral loads are not usually recommended [65].

NGS has been successfully applied in poliovirus typing and identification [66]. This success led to the use of NGS techniques in EV typing. Viral RNA is obtained either from purification and extraction from cell culture supernatant [67] or direct isolation from clinical samples [68]. Different strategies for reverse transcription of viral RNA into cDNA fragments, which can be subsequently sequenced by high-throughput techniques, have been reported. Some strategies are based on random amplification using non-specific primers [69], [70]. These strategies generally lead to the generation of DNA libraries, which mainly consist of non-viral sequences, thus decreasing the number of relevant reads obtained by sequencing. Alternate strategies are based on primers that specifically target conserved genomic regions of the viral RNA to be sequenced [71] or the whole-genome sequencing approach, which directly targets native RNA [72]. Such strategies have been used for EV-A71 [73], [74] and EV-C species [71]. These specific primers are usually degenerated and target either the full EV genome [68], [71], [72], [75], [76] or parts of the viral genome as the VP1 gene [77], [78].

After quality assessment of the prepared library, sequencing is performed on the NGS platform. To date, three different platforms are used for EV typing: MinION sequencing (Oxford Nanopore technologies), MiSeq sequencing (Illumina), and ion torrent sequencing (Thermo-Fisher). Sequence assembly is performed by mapping on a reference sequence or de novo assembly to obtain a full or partial EV genome [71].

The minimum batch size for NGS is eight samples, with typical sets of 24–96 samples [66], [75]. The processing time from extraction to nearly complete sequencing for a typical set of 24–96 specimens ranges from 3 days [66] to 4–5 days [65], [68], [75]. The total reagent costs for NGS technology per sample range from 70 USD to 120 USD and vary significantly when samples are multiplexed in one run [65], [67]. In addition, the number of reads for each sample varies from 0.3 million reads to 1.5 million reads [67]. The total cost per run, together with hardware-associated costs combined with bioinformatic challenges, may remain as major obstacles preventing Illumina-based sequencing.

NGS has become a useful tool for sub-typing, particularly in outbreak investigations [79] or in situations when the infectious agent is difficult to identify [80]. Other studies reported the successful application of NGS-based assays to typing of the HFMD-causing strain EV-A71 [57], [58], [65], [68] and severe respiratory disease-causing strain EV-D68 [81].

NGS of EV has several advantages over single-gene analysis.

• i.Detection of recombination events

  NGS techniques showed good correlation with classic Sanger techniques for the analysis of serologically untypable EV [67]. This approach has been used to detect 23 genome variants, which are potential genetic recombinants [67].

• ii.Capability to identify mutations and virulence markers across the genome

  Huang et al. conducted a retrospective study of the genomic diversity of EV-D68 strains collected during an outbreak in the late summer and fall of 2014. Their analysis showed that most EV-D68 strains (26 out of 29) differed significantly from prior ones. Two functional mutations in the protease cleavage sites may have been responsible for the increased rate of viral replication and transmission [81]. Routine sequencing of longer genome fragments or the complete genome is required to obtain a better image. These genetic determinants can be located anywhere in the genome but not in the highly exploited VP1 genome region.

• iii.Identification of potential new biomarkers throughout the genome

  In a study by Grädel et al., direct RNA sequencing using nanopore technology for fast whole-genome sequencing of viruses from clinical samples showed a good correlation with cDNA Illumina MiSeq Sequencing. This technology correctly identifies EV genotypes from stool samples and provides rich meta-transcriptomics information [72].

• iv.More sensitive and specific tracking of viral evolution and further identification and investigation of EV outbreaks and coinfection:

  An Australian study by Stelzer-Braid et al. investigated and sequenced the near full-length genome of 23 EV strains identified during a known outbreak of EV-A71 in 2012/2013. The identified EV-A71 strains were remarkably similar to those circulating in Asia during the same period [82]. In addition, Issacs et al. similarly characterized 58 clinical stool samples and discovered coinfection in four clinical samples, with each containing two distinct EV genotypes [83]. Similar results from the work of Chien et al. showed coinfection with two virus genomes in eight sequenced samples [67].

• v.Detection of EV circulation in the human population and sewage

  Recent studies have described the isolation of EVs in sewage by NGS. This process allowed the monitoring of EV diversity in wastewater and monthly analysis of concentrated samples [77], [84], [85], [86]. A genotype C1-like EV-A71 variant was recovered from wastewater in Arizona, and the strain was suggested to be circulating for more than two years [78]. Such a wastewater-based surveillance will help in outbreak prediction before the onset of cases in the community [78].

3. Indirect methods by serology

As in classical virology, serological diagnosis of EV infections can be made by comparing titers in acute and convalescent-phase serum specimens [6]. This process has a limited clinical value in the routine diagnosis of EV because of the time needed to obtain the samples, large number of existing types, and lack of a common antigen for serological assays. In general, EV serodiagnosis is more suitable for epidemiological studies than clinical diagnosis [6]. A more than fourfold increase in EV-specific antibodies between acute and convalescent samples is suggestive of a recent infection and may be useful in epidemiological investigations, such as studies of outbreaks caused by a specific serotype [10], [28].

Detection of neutralizing antibodies is used to quantify neutralizing antibodies against selected EV types [38]. Several commercial enzyme-linked immunosorbent assay (ELISA) tests are available on the market for EV-specific IgM [6].

ELISA tests have been successfully applied for epidemiological investigators of outbreaks and specific diagnostic use. As with neutralization IgM assays, EIAs possess exceptionally good sensitivity and specificity. In most cases, IgM tests are not serotype specific, and sensitivity is widely variable. A total of 10% to nearly 70% of serum samples showed a heterotypic response caused by other EV infections [87], [88].

4. Conclusion and perspectives

As technology advances, laboratories progress toward the implementation of novel state-of-the-art sequencing techniques to rapidly characterize and identify EVs in clinical specimens. NGS of pooled samples provides advantages as a screening tool to complement standard molecular assays during the surveillance and diagnosis of EV infection. In addition, NGS is a broad-spectrum approach that is not limited to pathogen-specific primers and probes. However, the cost of NGS technology is still an issue, and the provided results are complex and thus require careful pre-analytical phase and sample preparation, quality assessment of DNA libraries, and exhaustive data processing by experienced bioinformaticians to manage the large, generated databases. Global cooperation between experts in EVs is highly desirable to provide sufficient sample coverage and a good reference dataset. However, specific virologic diagnostic skills, including virus-isolation and histopathological investigations, should be continually implemented in reference laboratories to investigate newly emerging, re-emerging, and less-characterized or rare EV strains.

Supplementary data

The following are the Supplementary data to this article: