A 6-year-old boy (patient 1) and his 22-month-old cousin (patient 2) were hospitalized with the same ANE-specific symptoms a few days apart in March–April 2005 in Centre Hospitalier Universitaire, Clocheville (Tours, France). Serum, urine, CSF, and throat swab specimens were collected from both children. An influenza-like syndrome developed simultaneously in the mother of patient 2.

To detect herpesviruses and enteroviruses, we used PCR or reverse transcription PCR (RT-PCR), respectively, with commercially available reagents (herpes consensus generic detection kit and enterovirus consensus detection kit [Argene, Verniolle, France]). Nasopharyngeal aspirates were tested by indirect immunofluorescence, by using specific monoclonal antibodies (Argene), for influenza viruses A and B, respiratory syncytial virus, parainfluenza viruses 1–3, and adenoviruses. Serologic assays (ELISA) with commercially available reagents (Behring, Paris, France) were conducted to detect immunoglobulin (Ig) G and IgM against herpes simplex viruses, Epstein-Barr virus, measles virus, and mumps virus. Antibodies to influenza viruses were tested by complement fixation assays with antigens derived from influenza viruses A and B. Serologic assays were also conducted to detect IgG and IgM against hantavirus, tick-borne encephalitis virus, dengue virus, and chikungunya virus (National Reference Center for Arboviruses, Institut Pasteur, Paris). In addition, because the 2 patients had been in contact with an uncle who had returned from Asia (Indonesia), the presence of Hendra virus and Nipah virus nucleic acid sequences in CSF was investigated.

Urine and throat swab specimens from each patient were added to MRC5, MDCK, and Vero cells. Early viral stocks were made from urine and throat specimens added to MRC5 cells as soon as cytopathic effects (CPEs) were observed. Late viral stocks were obtained after inoculation of BGM cells. Cells were harvested as soon as CPEs were observed and fixed by incubation for 48 h in 4% paraformaldehyde and 1% glutaraldehyde in 0.1 mol/L phosphate buffer, pH 7.2, as described (14). Cell pellets were embedded in Epon resin (Sigma-Aldrich, Saint-Quentin Fallavier, France), which was allowed to polymerize for 48 h at 60°C. Ultrathin sections were cut, stained with 5% uranyl acetate and 5% lead citrate, and deposited on electron microscopy grids that were coated with collodion membrane for examination under a Jeol 1010 transmission electron microscope.

Specific primers of each reovirus segment were constructed in the highly conserved regions (Table 1). The PCR products corresponding to large (L), medium (M), and small (S) segments were cloned and sequenced in both directions. Sequence analysis was conducted by using Vector NTI (Invitrogen, Carlsbad, CA, USA). To obtain the S1 segment, we extracted total RNA from infected BGM cells with TRIZOL reagent (Invitrogen). The double-stranded RNA was obtained as previously described (15) and separated through migration on a 6% precast Poly(NAT) gel (Elchrom Scientific AG, Cham, Switzerland). A ≈1,400-bp band was recovered from the gel and directly used for RT-PCR with random primers (Invitrogen and Roche [La Rochelle, France]). The PCR products were purified from the gel and cloned, and several clones were sequenced. To complete the S1 sequencing, a specific S1 primer was constructed and used for the RT step. The amplification was done by using the specific S1 primer and a random primer (Roche). The band corresponding to the expected size (684 bp) was purified from the gel, cloned, and sequenced. The new sequences enabled construction of specific primers at each extremity (Table 1) to amplify the entire S1 segment and sequence in both directions.

Sequence alignments (nucleic acid and amino acid) were constructed using ClustalW 1.74 (16) and refined by visual inspection with SEAVIEW; distance matrix and phylogenetic trees were computed with PHYLO_WIN (17). Distances between sequences were computed by observed divergence. Trees were built by using the neighbor-joining method, and tree topologies were tested with 1,000 bootstrap sampling replicates.

BGM cells infected with the MRV2Tou05 strain and noninfected cells were used for Western blot analysis with a 1:100 dilution of the 2 patients’ serum specimens. Thirty-eight serum specimens (supplied by the Établissement Français du Sang, Lyon, France) obtained from 38 healthy blood donors were tested by Western blot for antibodies against the MRV2Tou05, Lang (T1L), Jones (T2J), and Dearing (T3D) strains.

Viral RNA was extracted from infected cells and supernatant and from patients’ samples by using the Nuclisens EasyMAG Kit (bioMérieux). Mock cells and supernatant from uninfected cells were used for RNA extraction as negative controls. A molecular test was set up with the L3 segment as target with primers enabling detection of the 3 reovirus serotypes (Table 1).

Patient 1, a previously healthy 6-year-old boy, was admitted to Centre Hospitalier Universitaire (Tours) on March 29, 2005, with fever, vomiting, and impaired consciousness. Results of his neurologic examination were normal, except for palsy of the left facial nerve. The Glasgow Coma Scale score was 7. He could respond to simple verbal commands, such as requests to open eyes or make some other movements, but he was lethargic (no articulated language) and had epilepsy-like abnormal movements of the face or distal muscles. For 2 days previously, he had experienced a prodromal illness with high temperature (40°) and headache. Results of an analysis of CSF on the first day of hospitalization were within normal limits and showed a protein level of 0.81 g/L and 4 leukocytes/mm³ on day 5. An electroencephalogram showed generalized slow-wave activity. Computed tomography of the brain showed a low-density change in the thalami. Brain magnetic resonance imaging (MRI) performed on the first day of hospitalization showed multiple symmetric lesions, with high signal intensity on T2-weighted images and low signal intensity on T1-weighted images, mostly involving the thalami bilaterally but also the brainstem tegmentum (Figure 1, panel A) and cerebral white matter in external capsules. He received treatment in the intensive care unit and was given acyclovir, methylprednisolone, and clonazepam intravenously. Beginning 2 days after he arrived at the hospital, he showed improvement, and he recovered completely in a few weeks. He was discharged on the 25th day of illness. A follow-up brain MRI, 1 year later, showed no abnormal findings. Three years later, his development and neurologic examination results were within normal limits.

Patient 2, a 22-month-old girl, was the cousin of patient 1. She was admitted 4 days later, on April 3, 2005, with drowsiness and fever after 1 day of high fever, asthenia, and rhinorrhea. A neurologic examination showed brisk deep tendon reflexes and extensor plantar responses; Glasgow Coma Score was 8. CSF initially showed a slightly elevated protein level (0.55 g/L, reference <0.30 g/L) with no cells, and findings were within normal limits on day 4 of illness (0.38 g/L protein, no cells). Radiologic examinations of the brain showed similar lesions to those of her cousin (Figure 1, panel B), and she received the same treatment. She experienced mild hepatomegaly on day 11 of illness. Her condition began to improve on day 7 of illness, and she was discharged from the hospital on day 18, April 20, with marked hypotonia, responsiveness but absence of language, and pyramidal tract signs. One year later, she remained easily tired, with uncoordinated movements. MRI of the brain found that the damage was less visible and the signal was much less marked.

Two other family members (the mother of patient 2 and the grandmother of both children) also had influenza-like symptoms (headache, fever, vomiting) at the same period for a few days without neurologic signs. The past family history and medical history did not appear relevant.

Results of molecular tests for herpesviruses and enteroviruses performed initially on the CSF specimens from both children were negative. The results were also negative for infection of respiratory viruses (influenza viruses A and B, respiratory syncytial virus, parainfluenza viruses 1–3, and adenoviruses). Serologic test results for HIV were negative. Serologic assays for IgG and IgM against herpes simplex virus, Epstein-Barr virus, measles virus, and mumps virus did not show IgM. Serologic assays were negative for both IgG and IgM against hantavirus, tick-borne encephalitis virus, dengue virus, and chikungunya virus. The results were also negative for nucleic acid sequences of Hendra virus and Nipah virus (F. Wild, National Reference Center for Measles and Other Paramyxoviruses, pers. comm.). All serum specimens collected during the acute and convalescent phases were negative for antibodies against influenza viruses.

Nonoriented virus isolation was attempted by inoculation of MRC5, MCDK, BGM, and Vero cell cultures with urine collected from each patient on April 7 and with a throat swab specimen from patient 2 on April 11. A CPE was observed on day 7 in MRC5 cells inoculated with urine from both patients, and on day 7 in MCR5 cells and day 10 in MDCK cells inoculated with a throat swab specimen from patient 2. A discrepant result for enterovirus was obtained when identification was attempted. Indeed, indirect immunofluorescence on fixed cells was positive for enterovirus (Pan-Enterovirus Blend, Light Diagnostics [Millipore, Molsheim, France]), but results of RT-PCR for detection of enteroviruses, performed on the cell culture supernatant, were negative. The virus isolates were then identified as reovirus type 2 by seroneutralization at the National Reference Center for Enteroviruses (Lyon, France). The discrepancy initially observed between immunofluoresence and RT-PCR results may be attributed to the cross-reactivity of the Pan-Entero Blend reagent toward reoviruses, as mentioned in the manufacturer’s description. Microscopic examination after staining of infected cells showed voluminous cytoplasmic inclusions characteristic of CPEs induced by reoviruses; electron microscopy showed the accumulation of virions in formation (Figure 2).

Reovirus sequences were searched for retrospectively by RT-PCR in the patients’ available specimens. A specific 512-bp fragment corresponding to the L3 expected region was obtained from urine specimens from both patients and from 1 serum specimen that was obtained 21 days after the onset of symptoms from patient 1 (data not shown). All amplified fragments were sequenced and showed identical profiles.

The complete sequence of the MRV2Tou05 genome was determined in both directions. Nucleotide and deduced amino acid sequences obtained for each segment were analyzed in the National Center for Biotechnology Information database (www.ncbi.nlm.nih.gov/Database/) to determine the percentage of identity with mammalian reoviruses. Sequence analysis of the 3 large segments, the 3 medium segments, and the S2, S3, and S4 segments indicated overall a close genetic relationship between MRV2Tou05 and swine SC-A viruses and a relationship with the 2 reovirus prototypes 1 and 3 (Table 2). The sequences of the MRV2Tou05 S1 gene showed great diversity at the nucleotide and amino acid levels, compared with the S1 gene from mammalian reoviruses and SC-A strains (Table 2, Table 3). Otherwise, the S3 gene showed high identity with type 2 human reovirus strains (Table 3). The highest identities (83% and 99%) were with type-2 human isolates 302I and 302II from the People’s Republic of China (18) and human isolates from the Netherlands (18), respectively, for S1 and S3 genes (Table 3). Nucleotide sequences of the prototype T2J and the novel MRV2Tou05 S1 genes shared 62% of positional identity, which provided sequence confirmation that this new isolate was a type 2 strain.

To establish the evolutionary relationship of MRV2Tou05 with the known mammalian reoviruses, we constructed phylogenetic trees on the basis of the nucleotide sequences of the S1 and S3 segments (Figure 3). The S1 phylogenetic tree shows that the MRV2Tou05 S1 sequence is more closely related to the S1 genes of other type 2 reovirus strains and most divergent from the type 3 reovirus strains, including the SC-A strain. In contrast to the great diversity of the MRV2Tou05 S1 gene sequence compared with the other reovirus strains, similar trees obtained with sequences of the MRV2Tou05 S2, S3, and S4 genes showed a randomized genetic clustering between the different reovirus types for these genes (data not shown).

The 4 S segments, S1–S4, of MRV2Tou05, isolated from the throat swab specimen of patient 2 and the urine specimen of patient 1, were entirely sequenced. The nucleotide sequence was identical for 62 of 81 clones, and <3 point mutations were observed among the whole RNA segments for the remaining 19 clones, indicating that the 2 children had been exposed to the same novel isolate.

Specific antibodies against MRV2Tou05 were detected in serum specimens from patient 2; reactivity was higher in the 2 specimens collected in the convalescent phase, i.e., 13 and 19 days after symptom onset (Figure 4, panels B, C) than in specimens collected in the acute phase (6 days after symptom onset) (Figure 4, panel A). Similarly, antibodies against MRV2Tou05 were detected in the serum specimens from patient 1 and the mother of patient 2 (data not shown). The molecular mass of 145 kDa corresponded to λ proteins encoded by the large segments, with a size of 142–145 kDa. None of the 38 serum specimens from healthy blood donors was positive for MRV2Tou05 reovirus (Figure 4, panel D) but 20 of 38 serum specimens from donors were positive for the reovirus prototype strains, with the highest prevalence for T1L and T3D (data not shown). In the donors’ blood, we observed a signal directed against the λ proteins as well as 2 additional bands corresponding to the μ proteins (75 kDa and 100 kDa) from the human reovirus prototypes. Antibodies against T1L and T3D strains were also found in serum specimens from patient 2, but not in specimens from patient 1 and from the mother of patient 2 (data not shown).