At autopsy, brain samples were obtained from the patient. Presence of RV antigen on brain tissue was confirmed by the direct fluorescent antibody test (www.cdc.gov/ncidod/dvrd/rabies/Professional/publications/DFA_diagnosis/DFA_protocol-b.htm
) and direct rapid immunohistochemical test (3
) (). Antigenic typing was performed with a panel of anti-RV nucleoprotein (N) monoclonal antibodies as described (4
). Comparison of the human sample with established reaction patterns of RV variants showed that the closest antigenic match was among insectivorous bat patterns, specifically a unique RV variant in Colorado Myotis
sp. and several Tadarida brasiliensis
variants (Appendix Table
). Patterns obtained from hematophagous bat-, canine-, and terrestrial carnivore–associated RV variants were not consistent with the pattern obtained from the patient.
Figure 1 Detection of rabies virus antigen in brain impressions of the patient by direct fluorescent antibody test counterstained with Evans blue, 200× total magnification: A) positive control; B) negative control. Direct rapid immunohistochemistry test (more ...)
Total RNA was extracted from infected tissue, and the entire N gene was amplified by reverse transcription–PCR in 2 overlapping amplicons, as described (7
). Phylogenetic analyses were conducted by comparing full and partial RV N sequences with sequences derived from major extant rabies enzootics in both dogs and vampire bats in Mexico, as well as sequences associated with RVs maintained by other bat species and wild terrestrial carnivores from the United States and the Americas (8
) (). MEGA and BioEdit software were used to perform the phylogenetic reconstructions and sequence analyses (11
Figure 2 Phylogenetic tree of complete lyssavirus nucleoprotein genes, comparing the patient isolate with representative rabies virus variants associated with common New World animal reservoirs. The map shows the locations of representative samples associated (more ...)
Rabid dogs and vampire bats are the most common sources of exposure for humans in Mexico (www.salud.gob.mx/unidades/cdi/documentos/rabia.pdf
). Residual canine rabies enzootics persist in central and southeastern Mexico, whereas vampire bat rabies is found throughout a wide geographic focus, particularly in the tropical and subtropical areas (8
). In addition, at least 20 different lineages of RV that are associated with at least 9 bat species have been described in Mexico (8
). Nevertheless, phylogenetic analyses of the RV obtained from the brain sample of the patient did not support a close relationship with any of the RV variants previously described. The isolate was found to be most closely related with those from Mexican free-tailed bats (T. brasiliensis
); overall average identity was 95% but clearly segregated in an independent lineage (). Given that the average percentage of genetic divergence among previously sequenced members of the T. brasiliensis
RV clade ranges between 2.4% for the full N over a 20-year period and 1% for the partial N over a 40-year period, the extent of genetic divergence (5%) between the Oaxaca sample and the T. brasiliensis
RV clade suggests that this isolate represents a new RV variant. Additionally, although the Oaxaca sample shares a distinctive molecular signature with the T. brasiliensis
RV clade (i.e., conserved amino acid sequence alanine, aspartic acid, and threonine located at positions 377–379 within the N gene), the histidine at position 321, which is unique and highly conserved in members of the T. brasiliensis
RV lineage, was changed to glutamine in the Oaxaca patient.
Although this patient’s history indicated that he had been bitten near his home (San Vicente Coatlan, district of Ejutla, Oaxaca) by a costoche
) (gray fox, Urocyon cinereoargenteus
), the genetic and phylogenetic analyses did not support a close relationship to any known RV associated with terrestrial carnivores. RVs of major rabies epizootics associated with dogs and other terrestrial carnivores in Mexico and the United States are genetically distinct (average genetic distance 14%–16%) from those in bats throughout North America. Also, although the RV associated with the human case was nested within the monophyletic assemblage of bat RV variants, RV variants phylogenetically closest to this case were still genetically distant. RV variants associated with North American Tadarida
and vampire bat rabies in Mexico were from 5% to 7% divergent from that of the human case.
Results of partial RV N gene sequence analyses indicated that at least 2 other human rabies cases—one in California in 1995 and the other in Nuevo Leon, Mexico, in 1999—were associated with the T. brasiliensis variant. These cases segregated within the monophyletic assemblage that includes enzootic rabies in T. brasiliensis bats collected over a period of ≈40 years in the United States and Mexico; the Oaxaca human case sample fell outside the T. brasiliensis clade, forming an independent lineage that was statistically supported in both the partial and full N phylogenetic reconstructions. These results plus the amino acid change found at position 321 in the RV associated with this case reinforced the concept of a new RV lineage associated with an unknown animal reservoir.