Prospective Identification of Enteroviruses Involved in Meningitis in 2006 through Direct Genotyping in Cerebrospinal Fluid

Abstract

Enterovirus infections were investigated with special emphasis on performing rapid molecular identification of enterovirus serotypes responsible for aseptic meningitis directly in cerebrospinal fluid (CSF). Enterovirus genotyping was carried out directly with specimens tested for the diagnostic procedure, using two seminested PCR assays designed to amplify the complete and partial gene sequences encoding the VP1 and VP4/VP2 capsid proteins, respectively. The method was used for identifying the enterovirus serotypes involved in meningitis in 45 patients admitted in 2005. Enterovirus genotyping was achieved in 98% of the patients studied, and we obtained evidence of 10 of the most frequent serotypes identified earlier by genotyping of virus isolates. The method was applied for the prospective investigation of 54 patients with meningitis admitted consecutively in 2006. The enterovirus serotypes involved were identified with the cerebrospinal fluid (CSF) of 52 patients (96%) and comprised 13 serotypes within the human enterovirus B species and 1 within the human enterovirus A species. The three most common serotypes were echovirus 13 (E13; 24%), E6 (23%), and coxsackievirus B5 (11.5%), a pattern different from that observed in 2005. Genotyping of virus isolates was also performed in 35 patients in 2006 (meningitis, n = 31; other diseases, n = 4). By comparison, direct genotyping in CSF yielded a more complete pattern of enterovirus serotypes, thereby allowing the detection of rare serotypes: three less common serotypes (CB2, E21, and E27) were not detected by indirect genotyping alone. The study shows the feasibility of prospective enterovirus genotyping within 1 week in a laboratory setting.

Enterovirus infections comprise a wide spectrum of clinical presentations and diseases (35). As with poliomyelitis, the most problematic clinical syndromes caused by nonpolio enteroviruses (NPEV)—meningitis, encephalitis, and acute flaccid paralysis—result from infection of the central nervous system (CNS) and are major causes of illness and morbidity in both children and adults (57).

Enterovirus meningitis (also known as aseptic meningitis) is the most commonly observed CNS infection, and its clinical presentation varies with the patient's age and immune status (38, 39). It occurs typically during outbreaks (in restricted geographical areas or communities) in the summer and autumn, leading to increased admissions to hospital wards for short periods. Unlike meningitis, enterovirus encephalitis is a more severe acute illness with more long-term sequelae (16) and, while it is less frequently observed than meningitis, its frequency should not be underestimated. For instance, in the California Encephalitis Project, an enterovirus infection was detected in 25% of patients with a viral etiology; 24% had a herpes simplex virus type 1 infection (20). NPEV can also be responsible for acute flaccid paralysis, a syndrome that mimics poliomyelitis. In Europe, 84 NPEV associated cases were registered in 2005 by the Polio Lab Network (3).

Enteroviruses have a positive single-stranded RNA genome subjected to high mutation rates and frequent genetic recombination events (1, 17, 30-33, 37, 46-48, 58, 60, 61) and thereby display a great diversity (10, 41, 52, 55, 59). More than 90 serotypes have been characterized among the human enteroviruses (HEVs), but only 68 are recognized in the current taxonomy (62). They are divided into four species (HEV-A to -D), and the three poliovirus serotypes are closely related to NPEV serotypes included within the HEV-C species (11). According to the serotyping results of strains recovered from the stool and throat specimens of patients with EV meningitis, most serotypes are suspected of being involved in meningitis (50), but only some enterovirus serotypes have been responsible for large nationwide epidemics (8, 14, 15, 25, 26, 29, 67). However, the true involvement of the different enterovirus serotypes responsible for CNS infection is difficult to determine because enteroviruses are only rarely isolated from cerebrospinal fluid (CSF) of patients and molecular typing is not performed to identify enteroviruses directly.

Although detection of the 5′ noncoding region of the enterovirus genome in CSF is the gold standard for the diagnosis of enterovirus CNS infection and provides results in a clinically relevant time frame (51, 53, 56, 66), the genome region is inappropriate for enterovirus identification (52). Several molecular typing methods based on the amplification and sequencing of part of the VP1 coding region have been developed (12, 13, 22, 36, 43, 45, 49), but few of them have been used to identify enteroviruses directly in the CSF of patients (13, 23, 42, 64) and, to our knowledge, none have been prospectively used in a clinical diagnostic laboratory.

Using a species-specific (HEV-B) method, we previously performed an indirect genotyping assay (molecular identification of clinical viruses isolated in vitro) and showed that it was suitable for prospective identification of enterovirus strains recovered in patients with various diseases, including meningitis (36). The assay relied on a set of two primers with degenerate nucleotide positions designed to amplify the complete VP1 sequence of B species enteroviruses and was carried out in 3 weeks. Even when performed prospectively, indirect enterovirus genotyping is not appropriate in clinical situations that require rapid identification, such as severe presentations especially in neonates (38) or immunosuppressed patients (5, 7), putative nosocomial infections (6), or epidemiological studies aimed at determining whether a meningitis outbreak is caused by a new variant or serotype.

The aim of the present study was to perform the prospective identification of enteroviruses involved in CNS infections directly with the CSF specimens used for diagnosis within 5 days versus the 3 weeks needed for indirect genotyping. Using a seminested PCR derived from the indirect genotyping method, we prospectively carried out direct identification in CSF samples of patients admitted to our local University Hospital (Clermont-Ferrand, France) in 2006 and compared the genotyping results with those obtained with indirect genotyping during the same period. Enterovirus genotyping was achieved directly in CSF samples in 96.3% of the patients with meningitis in 2006; 13 serotypes were identified within the HEV-B species, and 1 serotype was identified within the HEV-A species.

MATERIALS AND METHODS

Patients.

All CSF specimens were collected from patients admitted to the University Hospital of Clermont-Ferrand, France. Throughout the text, the designation of the CSF specimens is given in the following format: number of the CSF sample-year of collection. For the virus isolates, the number is preceded by CF for Clermont-Ferrand.

Diagnosis of enterovirus meningitis was based on signs indicative of meningitis (fever, headache, photophobia, and/or nausea/vomiting, and/or stiffness of the neck), negative CSF culture for bacteria, and either a positive reverse transcription-PCR (RT-PCR) assay result (see below) in the CSF or the isolation of an enterovirus in cell culture from a peripheral specimen (throat swabs or feces).

From January to December 2005, 77 inpatients were diagnosed as having enterovirus meningitis. The diagnostic RT-PCR was positive for 70 of 77 patients. Overall, 56 enterovirus isolates were recovered and identified prospectively in a previous study (36). In 45 of 77 patients with meningitis, the viral RNAs extracted from CSF specimens were stored at −80°C for future retrospective genotyping investigations.

In 2006, 61 inpatients received a clinical diagnosis of enterovirus meningitis and were enrolled consecutively during the study period from February to December (see Table 3). Detection of the viral RNA in the CSF was positive in 54 of 61 patients (88.5%), and in 24 of 54 patients an enterovirus was isolated from throat swabs (n = 20) or feces (n = 4). In 7 of 61 patients (11.5%), the diagnostic RT-PCR assay was either negative (n = 5) or not done (n = 2). During the study period, an enterovirus was also recovered in four patients with clinical signs of pneumopathy (n = 1) or diarrhea (n = 2) and in one healthy boy whose brother (patient P33 in Table 3) was included in the study with proven meningitis.

TABLE 3.

Prospective enterovirus genotyping in CSF specimens and isolates in 61 patients with meningitis in 2006^a

Patient	CSF specimens						Enterovirus isolates
	Designation	Highest scoring strains				Accession no. (CSF)
		Accession no.	Designation	% nt sequence identityb	Typec		Designation	Type	Accession no.
P1	045067-06		No amplification (PCR A and PCR B)
P2	089047-06	AY919480	Isolate 10496	91 (310)	E27	AM711037
P3	105003-06	AJ537606	CF1393-00	96 (859)	E13	AM711038
P4	142037-06	AY896760	EV6-10887-99	86 (907)	E6	AM711039
P5	149036-06	AY896760	EV6-10887-99	85 (849)	E6	AM711040
P6	150001-06	X90723	isolate M1262	83 (480)	E25	AM696268	CF151101-06	E25	AM711079
P7	150005-06	AF295487	764/85	86 (571)	E21	AM711041
P8	157038-06	AY896761	EV6-14103-00	85 (422)	E6	AM696269
P9c							CF167067-06	CB5	AM711079
P10	168005-06	AJ537606	CF1393-00	96 (841)	E13	AM711042
P11	168007-06	AY186745	pMP1-24	82 (800)	CB1	AM711043	CF168014-06	CB1	AM711081
P12	177089-06	AJ537606	CF1393-00	96 (842)	E13	AM711044
P13	181093-06	AY896760	EV6-10887-99	86 (937)	E6	AM711045	CF180083-06	E6	AM711082
P14	184041-06	AY896760	EV6-10887-99	86 (864)	E6	AM711046
P15	184046-06	AY896760	EV6-10887-99	86 (908)	E6	AM711047
P16	186055-06	AY268575	KOR-E13-02-34	95 (847)	E13	AM711049
P17	189004-06	DQ092797	Human CB-5	92 (847)	CB5	AM711050
P18	186037-06	AY896760	EV6-10887-99	86 (906)	E6	AM711048	CF189033-06	E6	AM711083
P19	192009-06	AJ537606	CF1393-00	96 (841)	E13	AM711051
P20	193076-06	DQ092797	Human CB-5	91 (857)	CB5	AM711052
P21	194031-06	DQ092797	Human CB-5	90 (858)	CB5	AM711053
P22	196007-06	AJ417361	00CF1532/P17	97 (855)	E30	AM711054	CF196006-06	E30	AM711084
P23	196008-06	AY896760	EV6-10887-99	86 (947)	E6	AM711055
P24	196009-06	AY896761	EV6-14103-00	85 (410)	E6	AM696270
P25	200001-06	AB239987	501-5R/Mog/03	87 (404)	E25	AM696271
P26	200065-06	AF081621	NC95-2135	96 (390)	CB2	AM711056
P27	202075-06	AB167991	03-158FCC2	91 (686)	CB5	AM711057	CF202075-06	CB5	AM711085
P28	205064-06	AB239987	501-5R/Mog/03	87 (377)	E25	AM696272	CF205083-06	E25	AM711086
P29	206013-06	AY896761	EV6-14103-00	85 (421)	E6	AM696273	CF206019-06	E6	AM711087
P30	207033-06	AJ537606	CF1393-00	97 (835)	E13	AM711058	CF207006-06	E13	AM711088
P31	207034-06	AJ537606	CF1393-00	97 (835)	E13	AM711059
P32	208033-06	AJ537606	CF1393-00	96 (861)	E13	AM711060	CF209038-06	E13	AM711088
P33	209006-06	AB126217	HO786/Kanagawa/2003	96 (544)	EV71	AM696274
P34	209061-06	AY896763	CBV3-18219-02	91 (980)	CB3	AM711061	CF209062-06	CB3	AM711090
P35	211014-06	AY896761	EV6-14103-00	84 (386)	E6	AM696275	CF212044-06	E6	AM711091
P36	214014-06	AB178768	OC/01397	96 (884)	E13	AM711062	CF214017-06	E13	AM711092
P37	214068-06	DQ092797	Human CB-5	90 (851)	CB5	AM711063	CF216024-06	CB5	AM711093
P38	214088-06	AF160018	p234-pak92	93 (954)	CB4	AM711064
P39							CF219051-06	CB5	AM711094
P40	220005-06	AY896761	EV6-14103-00	85 (410)	E6	AM696276	CF219087-06	E6	AM711095
P41	221007-06	AJ537606	CF1393-00	97 (835)	E13	AM711065
P42	229025-06	AY896760	EV6-10887-99	83 (975)	E6	AM711066	CF230052-06	E6	AM711096
P43							CF235029-06	E13	AM711097
P44	235076-06	AJ537606	CF1393-00	97 (841)	E13	AM711067	CF235044-06	E13	AM711098
P45	237015-06	AY896765	EV7-15936-01	83 (484)	E7	AM696277
P46							CF246026-06	E13	AM711099
P47	247016-06	AY208092	99/00010-41	95 (972)	E18	AM711068
P48	263048-06	AY896760	EV6-10887-99	86 (917)	E6	AM711069	CF262045-06	E6	AM711100
P49	271062-06	AY208092	99/00010-41/99	94 (934)	E18	AM711070	CF272084-06	E18	AM711101
P50	280036-06	AB178768	OC/01397	95 (891)	E13	AM711071	CF282003-06	E13	AM711102
P51							CF287012-06	E18	AM711103
P52	293003-06	AY919558	Isolate 10569b	88 (885)	E9	AM711072	CF293042-06	E9	AM711104
P53	303018-06	AY208092	99/00010-41/99	95 (981)	E18	AM711073	CF303024-06	E18	AM711105
P54	312001-06	AJ417361	00CF1532/P17	92 (860)	E30	AM711074
P55	312082-06	DQ251322	Glas12	93 (215)	Untyped	AM696278	CF314038-06	E25	AM711106
P56	326004-06	AJ537606	CF1393-00	96 (839)	E13	AM711075	CF327002-06	E13	AM711107
P57							CF327034-06	EV71	AM696279
P58	345042-0	AF085363	Strain Ohio-1	84 (956)	CB2	AM711076
P59	347020-06	AF524866	Strain Barty	83 (1033)	E9	AM711077
P60	364004-06	AF160018	9105net93	92 (951)	CB4	AM711078
P61							CF356006-06	CB4	AM711108

^aResults obtained with VP4-VP2 sequences are indicated in italics. Results for patients with a negative (or not done) diagnostic RT-PCR in the CSF are indicated in boldface.

^bNumbers in parentheses indicate the length of the nucleotide (nt) sequence investigated by BLAST analysis.

^cE, echovirus; CB, coxsackievirus B; EV, enterovirus.

Diagnostic RT-PCR assays.

In 2005 and from January to May 2006 (the first seven patients hospitalized during the period of the prospective study), the positive detection of the enterovirus genome was obtained with the in-house real-time one-step RT-PCR assay described previously (4). In June 2006, a new procedure, NucliSens-EasyMag/EasyQ (bioMérieux, Marcy l'Etoile, France), was adopted in the laboratory for the diagnosis of enterovirus infections and used in 47 of 54 positive CSF specimens. The method allows automated extraction and a real-time NASBA detection of the viral RNA (19).

Virus isolation in cell culture.

Human lung embryonic fibroblasts (MRC5 cells; bioMérieux) and/or human epidermal carcinoma of the mouth (KB) cell lines were used for virus isolation from throat swabs and/or feces. Primary identification was based on the appearance of an enterovirus-like cytopathic effect and confirmed by indirect immunofluorescence assay as previously described (36). For patient P57 (Table 3), confirmation was made with the diagnostic RT-PCR because the indirect immunofluorescence assay result was negative.

Genotyping RT-PCR assays and nucleotide sequencing.

Synthesis of the cDNA was performed using Superscript III reverse transcriptase (Invitrogen, Cergy Pontoise, France) with random hexanucleotides at 2.5 ng/μl as primers, according to the manufacturer's instructions. Each reaction was carried out in a 20-μl reaction mix containing 8 μl of viral RNA, 0.5 mM concentrations of each deoxynucleotide triphosphate (Eurogentec, France), 2.5 μl of 5× reaction buffer, 0.01 M of dithiothreitol, 2 U of RNAseout (Invitrogen), and 10 U of reverse transcriptase. A total of 5 μl of the RT reaction mixture was used in subsequent assays.

Two different seminested PCR assays (PCR A and PCR B) were used in the study. PCR A and B were directed at the VP1 encoding sequence and at the genome segment comprising the VP4 and partial VP2 (5′ end) encoding genes, respectively. Primers used in the study are listed in Table 1. Primers HEVBS1695 and HEVBR132 were derived from the oligonucleotides ecox1 and ecox2 (6) previously tested on 43 EV isolates representing 18 different serotypes of the HEV-B species (accession numbers AJ131523, AJ241422 to AJ241456, AJ276624 to AJ276626, and AJ276812 to AJ276815) and were used in earlier studies (5, 8). Primers HEVBS1695 and HEVBR132 were first described for successful genotyping of 73 enterovirus isolates and amplified six additional serotypes in the HEV-B species (36).

TABLE 1.

Oligonucleotide primers used in this study

Primera	Sequence (5′-3′)b	Locationc	Tm (°C)
HEVBS1695*	CTTGTGCTTTGTGTCGGCRTGYAAYGAYTTYTCWG	2375-2409	59.5
HEVBR132*	GGTGCTCACTAGGAGGTCYCTRTTRTARTCYTCCCA	3467-3432	61.2
P1S1695S*	CTTGTGCTTTGTGTCGGC	2375-2392	48.4
P2R132S*	GGTGCTCACTAGGAGGTC	3467-3450	50.2
EV2C	CAATACGGCATTTGGACTTGAACTGTATG	4443-4415	55.1
5NC63B**	CGGTACCYTTGTRCGCCTG	64-82	52.6
5NCS663	GCGGAACCGACTACTTTGGGTGTCCGTGTTTC	538-568	61.4
HEVR436**	CCCATGTCAGTCAGCGCATCIGGIARITTCCAIYACCAICC	1210-1191	61.4

^aPrimers previously described by Mirand et al. (36) or Bailly et al. (9) are indicated by one or two asterisks, respectively.

^bSequences are written using standard nucleotide codes. R = A or G; Y = C or T; W = A or T; I = A, T, G, or C.

^cThe positions are given relative to the genome of echovirus 30 strain Bastianni (AF311938).

All of the reaction mixtures had a final volume of 50 μl and contained 320 μM each of the deoxynucleotide triphosphates, 5 μl of the 10× reaction buffer, 400 nM each of the two oligonucleotide primers (see below), and 1.25 U of Taq DNA polymerase (Qiagen S.A., Courtaboeuf, France).

First-round PCR A assays were performed with the reaction mixtures, including primers HEVBS1695 and EV2C (Table 1), and were run under the following conditions: 2 min at 94°C for denaturation, followed by 41 cycles of 15 s at 94°C, 20 s at 58°C, and 1.5 min at 68.5°C, with a final 10 min at 68.5°C. The second-round PCR A contained 5 μl of the first-round PCR A amplicons and oligonucleotides HEVBS1695 and HEVBR132 as primers; it was performed according to the conditions described above except for the annealing temperature, which was set at 58°C instead of 55°C (36). Gene amplifications with primer pairs HEVBS1695-EV2C and HEVBS1695-HEVBR132 produced two amplicons of about 2,070 and 1,100 bp, respectively.

First-round PCR B assays were performed with the reaction mixtures, including primers 5NC63B and HEVR436 (Table 1). The second-round PCR B assays were done with primers 5NCS663 and HEVR436. The samples were subjected to an initial denaturation step at 94°C for 2 min and then to 41 and 38 cycles of denaturation at 94°C for 15 s, annealing at 54°C, and 53°C for 20 s, respectively, for the first- and second-round assays, and extension at 72°C for 20 s. Gene amplification with primer pairs 5NC63B-HEVR436 and 5NC663-HEVR436 produced two amplicons of about 1,200 and 700 bp, respectively.

PCR products were examined on standard 1% agarose gel stained with ethidium bromide and were visualized under UV light. Negative (water) and positive (strain CF2660-01, an echovirus 6 strain cultured in MRC5 cells) controls were included in all PCR assays, as from the cDNA synthesis step, and were run in parallel with test samples. Standard precautions were undertaken to avoid the risks of contamination. Visible PCR products after gel electrophoresis were purified and subjected to nucleotide sequencing with the primers P1S1695S and P1R132S (PCR A products; Table 1) and 5NCS663 and HEVR436 (PCR B products), as described elsewhere (36). A contig was assembled with forward and reverse nucleotide sequences for PCR products.

Assay sensitivity.

The sensitivity relative to the results of cell culture infectivity was measured by using a titrated viral culture supernatant from the echovirus 6/CF2660-01 strain. It was serially diluted from 10 to 10⁸ times in a pool of CSF specimens that tested negative for enteroviruses by both diagnostic procedures. Viral RNA of two aliquots of 200 μl of each dilution was extracted with the NucliSens-EasyMag platform. The two eluates of 25 μl were pooled and tested (three replicates) both with the two diagnostic RT-PCR methods (in-house real-time RT-PCR and the NucliSens-EasyQ assays) and with molecular typing assays. The diagnostic RT-PCR assays yielded a positive signal down to dilution 10⁻⁶, indicating a detection limit of a 0.05 50% tissue culture infective dose per 5 μl of RNA extract. The limits of detection of genotyping assays were 0.8 and 0.08 50% tissue culture infective doses per 8 μl of RNA extract with the PCR A (VP1 sequence) and B (VP4/VP2 sequences) assays, respectively.

Nucleotide sequence and phylogenetic analyses.

The nucleotide sequence obtained for PCR A and PCR B products were analyzed according to a previously described method (36). Briefly, identification was carried out by comparative analysis of contig sequences with a local database constructed with all of the enterovirus sequences available in GenBank. Analysis was performed by the BLAST method (2), and each sample was assigned the serotype of the strain that gave the highest identity score (45). The result was confirmed by phylogenetic analysis. The sequences representing the VP1 and VP4/partial VP2 capsid proteins were selected from contigs and compared to one another and with the homologous sequences of all of the prototype strains included in the HEV-B species. Multiple sequence alignments were constructed and edited by using the CLUSTAL W (65) and Genedoc (Nicholas and Nicholas Jr available at http://www.psc.edu/biomed/genedoc/) computer programs, respectively. MEGA version 3 was used for phylogenetic analysis (28). Genetic distances were calculated by Tamura Nei's model of evolution (63), and the phylogenetic trees were constructed with the neighbor-joining method. The reliability of the branching patterns was determined by the bootstrap resampling test with 1,000 replicates (18).

RESULTS

Retrospective enterovirus genotyping in 45 patients hospitalized in 2005.

Direct enterovirus genotyping was first evaluated retrospectively in 45 inpatients with enterovirus meningitis that were hospitalized in 2005 (Table 2). Amplification of the VP1 sequence with the seminested PCR A assay was carried out directly from stored nucleic acids extracted from CSF specimens and achieved in 42 samples (93.3%). Analysis of nucleotide sequences determined from PCR A products by a gapped BLAST assigned viruses present in the samples to eight different serotypes. Each sample was assigned the type that gave the highest VP1 identity score and, as expected from the specificity of subgeneric primers, all of the serotypes were included within the HEV-B species (Table 2). Echovirus 30 was the most frequently identified serotype, accounting for 66.6% (28 of 42) of the samples tested, followed by echovirus 18, coxsackievirus B5 and B3 (7.1% each [3 of 42]), and echovirus 13 (4.7% [2 of 42]). The viruses detected in three samples were assigned to the echovirus 9, 11, and 33 enterovirus serotypes.

TABLE 2.

Retrospective enterovirus genotyping of 45 CSF specimens obtained in patients with meningitis in 2005

CSF sample tested	Highest scoring strain				Accession no. (CSF)c	Virus isolated
	Accession no.	Designation	% nt sequence identitya	Typeb		Designation	Type
159064-05	AJ295172	ITA97-002	93 (988)	E30	AM710995	CF160077-05	E30
175081-05	AY695108	Zhejiang	98 (849)	CB5	AM710996	CF175089-05	CB5
181054-05	AJ417361	00CF1532/P17	95 (854)	E30	AM710997
185099-05	DQ936620	304037	93 (854)	CB3	AM710998	CF185104-05	CB3
192020-05	AJ417361	00CF1532/P17	96 (856)	E30	AM710999
192022-05	AJ537606	CF1393-00	97 (835)	E13	AM711000	CF193056-05	E13
194070-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711001	CF196050-05	E30
196046-05	DQ092796	Human E11	96 (955)	E11	AM711002	CF196051-05	E11
197025-05	AJ417361	00CF1532/P17	96 (855)	E30	AM711003	CF197006-05	E30
199006-05e	DQ345771	Constant	86 (388)	E6	AM696266
203029-05	DQ936620	304037	93 (851)	CB3	AM711004	CF203101-05	CB3
204013-05	AJ295172	ITA97-002	93 (981)	E30	AM711005
206006-05	AJ295172	ITA97-002	93 (992)	E30	AM711006
206007-05	AJ295172	ITA97-002	93 (986)	E30	AM711007
207049-05	AJ417361	00CF1532/P17	95 (854)	E30	AM711008
216056-05	AJ537606	CF1393-00	97 (841)	E13	AM711009	CF216077-05	E13
218022-05	AY695108	Zhejiang 12/02	97 (852)	CB5	AM711010
220018-05	AJ295172	ITA97-002	95 (855)	E30	AM711011	CF220062-05	E30
230016-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711012	CF230033-05	E30
232001-05	AJ417361	00CF1532/P17	96 (855)	E30	AM711013	CF232022-05	E30
234024-05	AY167774	WK-18-04-00	90 (837)	E33	AM711014	CF235069-05	E33
235066-05	AJ295172	ITA97-002	93 (975)	E30	AM711015	CF237031-05	E30
236023-05	AY695108	Zhejiang	97 (849)	CB5	AM711016	CF238048-05	CB5
241017-05	AJ417361	00CF1532/P17	95 (857)	E30	AM711017
242044-05	AJ295172	ITA97-002	92 (986)	E30	AM711018
257023-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711019
258066-05	AF524867	DM	95 (928)	E9	AM711020
264080-05	AB188509	80365/Hiroshima	94 (767)	E18	AM711021	CF265112-05	E18
265024-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711022	CF265113-05	E30
265094-05		No amplification
272012-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711023	CF272058-05	E30
278024-05	AB167999	03-3003FCR2	98 (724)	E18	AM711024	CF279084-05	E18
283003-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711025
298098-05	DQ093621	302784	94 (853)	CB3	AM711026
302018-05	AJ417361	00CF1532/P17	95 (854)	E30	AM711027	CF302019-05	E30
302022-05	AB188509	Hiroshima-JP-04	94 (767)	E18	AM711028
302024-05	AJ417361	00CF1532/P17	95 (855)	E30	AM711029
304075-05	AJ295172	ITA97-002	92 (986)	E30	AM711030	CF306033-05	E30
308026-05	AJ314833	sh/Roma99	92 (307)	CA9	AM696267
312098-05	AJ417361	00CF1532/P17	95 (850)	E30	AM711031	CF313076-05	E30
313072-05	AJ417361	00CF1532/P17	97 (854)	E30	AM711032
316003-05	AJ295172	ITA97-002	92 (977)	E30	AM711033
344017-05	AJ417361	00CF1532/P17	97 (856)	E30	AM711034
346022-05	AJ417361	00CF1532/P17	97 (855)	E30	AM711035
350034-05	AJ295172	ITA97-002	92 (991)	E30	AM711036	CF350107-05	E30

^ant, nucleotide. Numbers in parentheses indicate the length of the nucleotide sequence investigated by BLAST analysis.

^bE, echovirus; CB, coxsackievirus B; CA, coxsackievirus A.

^cThat is, the accession number of the nucleotide sequence determined in the CSF specimens tested.

^dThat is, the corresponding virus isolates recovered in peripheral specimens, genotyped prospectively in 2005 (36).

^eBLAST searches with VP4-VP2 sequences are indicated in boldface.

A PCR B assay directed at the 5′ part of the enterovirus genome (amplification of complete VP4 and partial VP2 encoding sequences) was performed in three samples for which PCR A assay did not detect any viral genomic sequence. Amplification was achieved in two specimens, which were assigned the echovirus 6 (199006-05) and coxsackievirus A9 (308026-05) enterovirus serotypes. In the third specimen (265094-05), no amplification product was detected, suggesting an amount of viral RNA below the detection limit, mispairing of any of the different oligonucleotides, or RNA degradation during storage. Altogether, using PCR A and PCR B assays, enterovirus genotyping was achieved directly in CSF samples in 44 of 45 patients studied (97.8%).

General characteristics of patients included in the prospective study.

In 2006, 61 patients received a clinical diagnosis of enterovirus meningitis and were included in the study for the prospective identification of the enterovirus involved either directly in CSF specimens (n = 54) or indirectly after isolation in cell culture (n = 31). There were 20 adults (32.8%) and 41 children (67.2%); the median patient age was 7.7 years (age range, 2 days to 71 years). The male/female ratios were 1.5:1 and 1.7:1 in adults and children, respectively. The seasonal distribution of enterovirus meningitis in 2006 showed that 54% of cases occurred in July (n = 22) and August (n = 11), with further cases being recorded up to December 2006 (data not shown).

Prospective enterovirus genotyping with CSF.

Direct genotyping in CSF samples was prospectively performed in 54 consecutive patients with enterovirus meningitis in 2006 with the viral RNA used for the diagnostic RT-PCR assay. An amplicon was successfully obtained with the PCR A assay in 42 (77.8%) samples, and all of the VP1 sequences were determined. The highest VP1 identity scores determined by the gapped BLAST analysis are presented in Table 3. The nucleotide identity values ranged from 83 to 97%, and most of the sequences (28 of 42) displayed at least 90% nucleotide similarity with an enterovirus sequence in the database. A nucleotide similarity value lower than 90% was observed for 14 sequences, 9 of which had the highest VP1 identity score with the same echovirus 6 strain (AY896760) isolated in Russia in 1999 (27). The other five sequences displayed highest identity scores with sequences assigned to the serotypes coxsackievirus B1 and B2 and echovirus 9 (n = 2) and 21.

In 12 CSF samples, no amplification was obtained with the PCR A assay, and the alternative PCR B assay was used to detect the viral genome. The expected amplification product was not obtained in only one sample (045067-06), which suggests an extremely low viral load in the P1 patient. In seven CSF samples, the analysis of the nucleotide sequence determined assigned the enteroviruses to only three serotypes: echovirus 6 (n = 5), echovirus 25 (n = 1), and enterovirus 71 (n = 1), which belongs to the HEV-A species. With four samples, the identification was not straightforward. For samples 200001-06, 205064-06, and 312082-06 (patients P25, P28, and P55, respectively), the highest scoring strain (AY875692) was reported as coxsackievirus B5 in the database, while the second and third highest scoring strains were reported as echovirus 25. When BLAST analysis was done with a selected segment corresponding to the only VP4-VP2 sequence, the first two samples (200001-06 and 205064-06) were assigned to serotype echovirus 25 (AB239987). With the third CSF sample (312082-06), the highest scoring strain (DQ251322) detected in the database had been reported as an untyped strain, while the second highest scoring strain (AB239987) had been identified as an echovirus 25. In addition, identification based on the VP1 sequence of the virus isolate obtained from a throat swab paired with the CSF sample assigned the virus to the echovirus 25 serotype (see Table 3). Finally, with the 237015-06 sample (patient P45), the highest scoring strain (AY896765) determined by genotyping had been reported as an echovirus 7, but all of the first four highest scoring strains had a different serotype, which makes it impossible to unequivocally identify the virus.

Overall, direct genotyping with both assays (PCR A and PCR B) allowed the identification of the involved enterovirus in 52 (96.3%) of the 54 consecutive patients with a positive diagnosis of meningitis. Thirteen different serotypes were within the HEV-B species, and one serotype was within the HEV-A species.

Prospective genotyping of enterovirus isolates.

Genotyping was also performed with isolates recovered in 31 of 61 inpatients with enterovirus meningitis, including 24 inpatients with paired CSF samples. All of the enterovirus isolates were genotyped with either the PCR A assay (n = 30) or the PCR B assay (n = 1). The nucleotide identity values determined by a gapped BLAST analysis (Table 3) ranged from 82 to 97%, and most (19 of 31) were greater than 90%. No discordance was observed between the serotype identified with the CSF specimens and that determined with the isolates recovered from throat swabs or feces. Moreover, the PCR A-positive CSF samples and the virus strains from paired specimens were assigned the same highest scoring strains. In four patients in whom genotyping was determined with the VP4-VP2 sequence detected directly in CSF samples, the results were also confirmed by genotyping with the VP1 sequence in paired virus isolates. Ten different serotypes were identified within the HEV-B species (PCR A assay), and one serotype was identified within the HEV-A species (PCR B assay) (Table 3).

Overall, enterovirus genotyping with CSF specimens and enterovirus isolates allowed the identification of the virus strain involved in 59 of 61 (96.7%) patients (Table 4). Echovirus 13 and echovirus 6 were responsible for a meningitis syndrome in 24.6% (15 of 61) and 23% (14 of 61) of the patients, respectively. They were followed by coxsackievirus B5 (7 of 61 [11.5%]); echovirus 18 and 25 (each 4 of 61 [6.6%]); coxsackievirus B4 (3 of 61 [4.9%]); and echovirus 9 and 30, coxsackievirus B2, and enterovirus 71 (each 2 of 71 [3.3%]). Serotypes coxsackievirus B1 and B3 and echovirus 21 and 27 were each identified once. The causative enteroviruses remained unidentified in two patients (P1 and P45).

TABLE 4.

Number of enterovirus serotypes identified in patients with meningitis in 2006 through direct (CSF samples) and indirect (virus isolates) genotyping

Identified serotype	No. (%)a
	CSF samples (n = 54)	Virus isolates (n = 31)	Total meningitis cases (n = 61)
Echovirus 13	13 (24.1)	8 (25.8)	15 (24.6)
Echovirus 6	14 (25.9)	7 (22.6)	14 (23)
Coxsackievirus B5	5 (9.3)	4 (12.9)	7 (11.5)
Echovirus 25	4 (7.4)	3 (9.7)	4 (6.6)
Echovirus 18	3 (5.6)	3 (9.7)	4 (6.6)
Echovirus 30	2 (3.7)	1 (3.2)	2 (3.3)
Coxsackievirus B2	2 (3.7)	0	2 (3.3)
Coxsackievirus B4	2 (3.7)	1 (3.2)	3 (4.9)
Echovirus 9	2 (3.7)	1 (3.2)	2 (3.3)
Echovirus 27	1 (1.9)	0	1 (1.6)
Echovirus 21	1 (1.9)	0	1 (1.6)
Coxsackievirus B1	1 (1.9)	1 (3.2)	1 (1.6)
Coxsackievirus B3	1 (1.9)	1 (3.2)	1 (1.6)
Enterovirus 71	1 (1.9)	1 (3.2)	2 (3.3)
Unidentified	2 (3.7)	0	2 (3.3)

^an, Total number.

Prospective enterovirus genotyping in patients with other clinical syndromes.

In the prospective study, an enterovirus isolate was genotyped in four inpatients admitted with clinical manifestations other than those associated with meningitis. Echovirus 6 and 13 were identified in two patients presenting with diarrhea. Another echovirus 6 was isolated from a bronchoalveolar fluid specimen in an immunocompromised 71-year-old woman who died during hospitalization. Finally, enterovirus 71 was isolated in the throat of a healthy 1-year-old boy. His brother (patient P33) was included in the study with proven enterovirus meningitis, and enterovirus 71 was identified directly from the CSF sample.

Phylogenetic confirmation of genotyping results obtained with the BLAST analysis.

Phylogenetic analysis was performed with the 14 VP4/VP2 sequences (obtained from 13 CSF specimens and 1 isolate) with homologous sequences of all prototype strains to confirm the typing results. In the VP4-VP2 phylogenetic tree (Fig. 1), samples that were assigned the same serotype clustered together but not with their homologous prototype strain. Sequences in four samples (150001-06, 200001-06, 205064-06, and 312082-06) clustered with a 100% bootstrap value in accordance with their clustering observed in the VP1 sequence. Thus, according to the BLAST results, they should be considered as echovirus 25. The virus in one patient (237015-06) which could not be formally identified by the BLAST comparison clustered with the echovirus 17 prototype strain but with a bootstrap value lower than 70%. Phylogenetic analysis performed with the VP4 coding segment alone did not provide additional information (data not shown).

FIG. 1.

Phylogenetic tree based on the complete VP4 and partial VP2 encoding sequences (420 nucleotides) of enteroviruses obtained from CSF specimens and isolates and all HEV prototype strains. The tree was constructed by the neighbor-joining method and validated with 1,000 bootstrap pseudoreplicates. Only bootstrap values of greater than 70% are shown. Genetic distances were calculated with Tamura-Nei's model of evolution, and branch length is drawn to the indicated scale (proportion of nucleotide substitutions per site).

The 114 VP1 sequences (obtained from 84 CSF specimens and 30 isolates) were also compared to homologous sequences of all prototype strains (data not shown). Sequences that were assigned the same serotype by the BLAST analysis were all monophyletic with respect to the homologous prototype sequence in the phylogenetic tree. All of the genetic clusters corresponding to each serotype had a 100% bootstrap value, and therefore the phylogenetic analysis definitely confirmed genotyping with the VP1 sequences. Moreover, whenever the number of sequences made it possible, phylogenetic analysis evidenced reliable subclusterings. For instance, within the echovirus 6, coxsackievirus B5, and echovirus 18 serotypes, two or three distinct lineages were observed (data not shown). In contrast, within the cluster of the echovirus 13 serotype, all of the strains displayed little genetic diversity.

DISCUSSION

In particularly serious or atypical presentations of CNS infections or in the event of an epidemic of meningitis, identifying the causative agent rapidly is of practical importance because a number of different viruses can be involved (24). Since enteroviruses are the most commonly identifiable cause of aseptic meningitis (50), it would be an advantage if enterovirus identification could be achieved prospectively, directly within CSF, through rapid genotyping: the overall management of outbreaks would be improved (40).

Direct enterovirus genotyping in 99 patients with meningitis using a two-step method.

The two-step method described here was used in 99 patients with positive diagnosis of enterovirus meningitis, including 54 sequential patients enrolled in 2006 for the prospective study and 45 patients with a positive diagnosis of enterovirus meningitis made in 2005 (36) for the retrospective study. The enterovirus serotype was identified directly in CSF samples in 84 of 99 patients (84.8%) with the first assay (PCR A; VP1 sequence). In the 15 patients for whom serotype identification was not achieved, genotyping was done successfully in 12 with the second assay (PCR B; VP4/VP2 sequence). Overall, direct genotyping was achieved in 97% of the patients in the present study.

Long nucleotide segments are not readily amplified by consensus PCR, and the amplicon length may explain the insufficient sensitivity of PCR A assay. With a two-step approach, Thoelen et al. (64) achieved direct enterovirus identification in 65% of 122 patients using degenerate primers designed for amplifying and sequencing a VP1 segment smaller than in the present study. A second step with multiple serotype-specific primers was necessary for identifying enterovirus strains in 35% of the patients. Few other PCR assays have been developed to identify enterovirus serotype directly in CSF specimens (13, 23, 42, 64). The methods were based on sequencing of different partial VP1 segments after nested gene amplification and used either generic (13, 42) or species-specific (23, 64) primers. To our knowledge, none have been used prospectively in a clinical diagnostic laboratory to identify enteroviruses responsible for meningitis and other CNS diseases.

In our study, when low amounts of viral RNA in CSF specimens prevented genotyping with the VP1 sequence, identification with the VP4/VP2 sequence was used as an alternative method, because of the simplicity of the amplification/sequencing stages, which are performed with only three primers, and its greater sensitivity. Using a two-step approach, Manzara et al. (34) showed that typing with the VP4/VP2 sequence was helpful, provided that amino acid sequence identity was not less than 90% with the highest scoring strain obtained by BLAST analysis. However, concerning the validity of enterovirus typing with the VP4/VP2 sequence, opinions are highly divergent. While the VP4 sequence was effective in identifying enterovirus serotypes (22), the 5′ part of the VP2 encoding sequence was reported not to be efficient at distinguishing between all enterovirus serotypes (13, 22, 44). In our experience, the interpretation of BLAST analyses of raw contigs that included a portion of the 5′ untranslated region and the VP4/VP2 segment posed problems in some patients. However, when the portions corresponding to the 5′ untranslated region were deleted from the contigs, identification was straightforward, and all of the viruses were assigned the echovirus 25 serotype. Overall, the BLAST analysis of the VP4/VP2 segment identified four serotypes (echovirus 6, 7, and 25 and coxsackievirus A9) within the HEV-B species and one serotype (enterovirus 71) within the HEV-A species. Although phylogenetic analysis of all of the VP1 sequences validated genotyping results obtained with BLAST comparisons, it was of little help in validating genotyping results obtained with the VP4/VP2 sequences. In contrast to the results obtained with the VP1 sequences, clustering with homologous prototype sequences was not observed with the VP4/VP2 sequences, with the exception of those assigned to the enterovirus 71 serotype. This is consistent with previous observations which indicated that this genome segment might not be appropriate for identifying all enteroviruses (13, 44). However, homotypic sequences clustered together with high bootstrap values, suggesting that this segment is suitable for detecting reliable phylogenetic relationships between circulating enteroviruses.

Direct genotyping versus indirect genotyping: implications for enterovirus epidemiology.

In the prospective study in 2006, a virus isolate was recovered in 31 of 61 patients (50.8%) with positive diagnosis of enterovirus meningitis, including seven patients for whom direct genotyping was not possible. Thus, an enterovirus was identified either with the direct or indirect method for 96.7% of the patients included. Although the overall frequencies of the most frequent serotypes were similar with both methods, three minor serotypes (coxsackievirus B2 and echovirus 21 and 27) were detected only by direct genotyping because no specimen was available for cell culture (Table 3). In 2005, an enterovirus isolate was obtained in only 41 of the 70 (58.5%) patients with meningitis. The retrospective analysis of 45 CSF specimens collected in 2005 allowed the identification of the involved enterovirus in 19 additional patients (for whom no enterovirus isolate was obtained; Table 2) and the detection of two additional serotypes (echovirus 6 and coxsackievirus A9), extending the results of indirect genotyping reported earlier (36). Since only CSF samples are provided for patients with meningitis, direct molecular typing allowed a more accurate and complete picture of the circulation patterns of involved enteroviruses in both prospective and retrospective studies.

During the prospective study, enterovirus meningitis was diagnosed in 61 inpatients and 14 distinct serotypes were identified (see Table 4). In 2005, by comparison, meningitis was evidenced in 70 inpatients, and 10 serotypes were identified (36). Although the numbers of patients and enterovirus serotypes were closely similar in the 2 years, the epidemiological patterns of EV serotypes determined were very different. Among the enterovirus strains identified in patients hospitalized in 2005, the most common serotype, echovirus 30, was replaced in 2006 by the echovirus 13 and 6 serotypes. This is not unexpected since these two serotypes were among the three serotypes most frequently isolated between 2000 and 2004 in France (54). The high immunization rate of the general population against echovirus 30 acquired during earlier meningitis outbreaks (2000 and 2005) may explain the provisional extinction of this serotype in France. However, in 2006, in Spain, echovirus 30 was the most frequently isolated serotype involved in meningitis (21). The coxsackievirus B5 serotype was the third most identified serotype both in 2005 and in 2006, accounting for 11.5% of the meningitis cases in the prospective study. This virus has been regularly isolated in France since 2000 and in 2003 was the most frequently isolated serotype (54). It was also the most frequently isolated serotype in Korea in 2005 (29). It has been suggested that the cyclic occurrence of coxsackievirus B5 infections is due to the reemergence of specific genotypes (27). This may explain the persistence of the serotype in the general population within the Clermont-Ferrand area. The pattern of enterovirus serotypes responsible for meningitis in 2006 in Clermont-Ferrand was also marked by the detection of enterovirus 71 in two patients. This is an unexpected finding because the virus was rarely detected in acute diseases during the period from 2000 to 2004 (54). According to a recent study in Norway, the prevalence of this enterovirus was 6.8% in a cohort of 113 asymptomatic children (69). This finding suggests that the circulation of enterovirus 71 might also be underestimated in France, and detection of this serotype in the present study underlines the necessity of molecular typing for enterovirus serotypes within the HEV-A species, because enterovirus 71 can be associated with severe brain stem encephalitis and pulmonary edema (68).

In conclusion, the present study demonstrates the feasibility of direct genotyping in CSF for prospective and rapid identification of enteroviruses involved in meningitis and provides a more complete picture of the circulation patterns of enteroviruses than indirect genotyping. Among the factors that can be involved in the varied pattern of CNS diseases caused by enteroviruses, differences in the enterovirus susceptibility of individuals and populations have been suggested, as have factors intrinsic to the virus itself. Analysis of CSF during acute and persistent infection phases could provide an insight into the pathogenesis of CNS infection by combining rapid procedures for clinical diagnosis, direct genotyping and, in the future, quantitative detection of enteroviruses.