Mouse-to-human transmission of variant lymphocytic choriomeningitis virus

Emerging Infectious Diseases, March, 2007 by Sebastien Emonet, Karine Retornaz, Jean-Paul Gonzalez, Xavier de Lamballerie, Remi N. Charrel

Lymphocytic choriomeningitis virus (LCMV) belongs to the genus Arenavirus in the family Arenaviridae. Mus musculus mice constitute the reservoir of LCMV in nature (1), but hamsters can also carry the virus. Humans usually become infected through direct contact with infected rodents or by inhaling infectious rodent excreta or secreta during occupational exposure (laboratory workers, rodent sellers) or when caring for rodents as pets. Although LCMV infection is usually asymptomatic or mild and self-limiting, it can be severe and manifest itself as meningitis and encephalitis (2,3). Infection during pregnancy may cause abortion or congenital malformations (4). In 2003 and 2005, two episodes of infections occurred in 2 groups of 4 recipients of solid-organ transplants; 7 of the 8 patients died (5). Although laboratory evidence of LCMV infection was obtained from the 8 organ recipients and from the hamster handled by donor 2, no laboratory evidence of infection could be obtained from specimens from both donors. Therefore, evidence for transmission from donor to recipient was mainly based on epidemiologic data and on the genetic identity of the virus detected in transplanted organs.

The Study

A 5-year-old boy was admitted to the neuropediatric ward of a Marseille University hospital in August 2004 for aseptic meningitis. Forty-five days later, he was admitted again with fever and meningitis, and his condition rapidly deteriorated with encephalitis and hydrocephalus developing as described (6). LCMV infection was suspected on the basis of the clinical signs and a detailed interview of the mother, in which she indicated that "the house was invaded with mice." LCMV was diagnosed based on seroconversion and positive PCR results with confirmatory sequence data (6).

Because M. musculus is the natural host of LCMV, the Public Health Office of Marseille organized a trapping campaign in the vicinity of the patient's house, which resulted in the capture of 20 mice (M. musculus) with glue traps. Most were alive when they were received at the laboratory. They were humanely killed and dissected; kidneys, lungs, heart, spleen, and liver were placed individually in 1.8-mL tubes and stored at -80[degrees]C. For each animal, 1 kidney was homogeneized as previously described (7). Recovered material was used either for virus isolation on Vero cells (detailed protocol available on request) or for total RNA isolation with the EZ1 Virus Mini Kit v2.0 on the BioRobet EZ1 Workstation (QIAGEN SA, Courtaboeuf, France) and eluted in 75 [micro]L final volume. A total of 10 [mirco]L was tested for LCMV RNA as described herein. Three different PCR assays, targeting the nucleoprotein gene, were used. System 1 was a nested reverse transcription-PCR (RT-PCR) that used primers 1817V-LCM (5'-AIATGATGCAGTCCATGAGTGCA CA) and 2477C-LCM-3' (5'-TCAGGTGAAGGRTGGC CATACAT-3') for the first round and primers 1902V-LCM (5'-CCAGCCATATTTGTCCCACACTTT-3") and 2346C -LCM (5'-AGCAGCAGGYCCRCCTCAGGT-3') for the second round. These primers were derived from those reported by Bowen et al. (8,9) and designed from the alignment of LCMV sequence data retrieved from GenBank. System 2 was a real-time RT-PCR with primers LCM_TM_NP1 (5'-TCATGTGGCARRATGTTGTG-3') and LCM_TM_NP2 (5'-AAAAAGAAIAARGARAT CACCCC-3') together with a FRET probe LCM_ MAR_NP (5'-ATGATGCAATCCATAAGTGCGCAGT3'). System 3 was a SYBR Green real-time RT-PCR based on primers LCM_SG_NP1 (5'-TTRTCRTCYCTYYTYT CYTTYCTCAT-3') and LCM_SG_NP2 (5'-CAGGTRA CYTTYGARAAITRRAGRAA-3'). The 3 detailed RT-PCR protocols are available on request. Human cerebrospinal fluid (CSF) samples and mouse specimens were added to Vero cells. After incubation at 37[degrees]C for 7 days, cells were tested for LCMV RNA by PCR with PCR system 1.

Results of PCRs and virus isolation are presented in the Table. Criteria to consider that samples contained LCMV RNA or LCMV were 1) virus isolation or 2) positive PCR results for at least 2 systems. Genetic analyses and phylogenetic reconstruction were based on sequences flanked by primers 1902V-LCM and 2346C-LCM. These primers amplified a 445-bp PCR product (primers included) and provided a 400-nt sequence (primers excluded) used for analysis. Nucleotide alignments were performed by using ClustalX 1.81 with default parameters (10). Alignments included the 16 sequences determined in this study and homologous LCMV sequences retrieved from the GenBank database. Phylogenetic analysis was performed with the Jukes-Cantor algorithm for distance calculation and the neighbor-joining method for cluster reconstruction with the MEGA 2.0 program (11). The robustness of nodes was tested by 500 bootstrap pseudoreplications.

As shown in the Table, the 2 human CSF specimens and 14 of 20 mouse samples were PCR positive. The 2 sequences obtained from human CSF specimens were 100% identical to each other. The 14 sequences representing mouse kidney specimens were almost identical (98.5% nucleotide identity) (Figure). Comparison of human and mouse sequences showed genetic identity >98% at the nucleotide level (Figure). This high level of similarity suggests that human LCMV infection was caused by transmission from the mice. All 16 sequences determined in this study either from human or rodent material had 12%-13% nucleotide heterogeneity when compared with LCMV sequences deposited in the GenBank database and with the sequence of LCMV strain manipulated in the laboratory, thus excluding the possibility of laboratory contamination. Finally, a total of 12 strains were isolated from Vero cells; 1 was selected to be characterized by full-length genome sequencing (GenBank accession nos. DQ286931 and DQ286932 for S and L RNA sequences, respectively).