Ethical clearance was obtained from the ethics committee of the Faculty of Veterinary Medicine and Animal Science at the University of Peradeniya (Peradeniya, Sri Lanka) and the Animal and Plant Health Agency (Addlestone, UK). Specimens were collected under permit no. WL/3/2/62/14 from the Sri Lanka Department of Wildlife Conservation.

During January 1, 2014–October 31, 2015, a total of 62 grounded bats were collected in an area inhabited by a long-established roost of ≈20,000 Indian flying foxes in Gannoruwa, Peradeniya, Sri Lanka (7°16′N, 80°36′E), which is located 600 m above sea level. Most bats were found dead. One bat (AK-42), which had clinical signs of illness, died shortly after capture (Table).

Figure 1

Figure 1. Negri bodies and lyssavirus antigens in brain tissue from an Indian flying fox, Sri Lanka. A) Degenerate Purkinje’s cell with an eosinophilic, intracytoplasmic inclusion body and a Negri body (arrow). Hematoxylin...

The first bat collected (AK-15) was a fresh carcass of a mature male that weighed 1.5 kg. A detailed necropsy showed that the animal had been healthy and had well-developed pectoral muscles. Except for a few multifocal hemorrhages in the lungs and mild, diffuse hyperemia and edema in the brain, gross pathologic findings were unremarkable. However, Negri bodies of various sizes were identified in the brain (Figure 1, panel A). Numerous aggregations of lyssavirus nucleocapsid antigen were observed in brain smears subjected to a direct fluorescence antibody test (dFAT) (Figure 1, panel B). Histopathologic examination of brain and spinal cord showed mild nonsuppurative lesions, leptomeningitis, and encephalomyelitis. Three additional dFAT-positive samples were identified from the 62 bats tested (Table). Subsequent virus isolation and molecular analysis were conducted for these 4 brain samples.

Virus was isolated by using N2A cells (9). After 5 days of incubation, 3 of 4 samples were positive for virus. Two of the isolates, RV3267 and RV3269, were subsequently cultured in BHK cells. RNA was extracted by using TRIzol reagent (Invitrogen, Paisley, UK). A pan-lyssavirus reverse transcription PCR yielded a specific 606-bp amplicon for the virus nucleoprotein gene (10). Results for a differential real-time reverse transcription PCR with a TaqMan probe specific for RABV showed no amplification for the 4 RNA samples. A specific 145-bp amplicon was visualized after electrophoresis on a 2% agarose gel. Thus, pan-lyssavirus primers used in the real-time assay detected this virus, but, the RABV-specific probe did not bind to the amplicon, which suggested presence of a non-RABV lyssavirus.

Figure 2

Figure 2. Phylogenetic relationships between representatives from all classified lyssaviruses and novel Gannoruwa bat lyssavirus (GBLV) on the basis of complete genome sequences. All 4 GBLV sequences form a monophyletic clade and are...

Complete genome sequences (GenBank accession nos. KU244266–9) were obtained from brain RNA samples by using next-generation sequencing according to previous methods (11,12). Phylogenetic analysis of complete genome sequences, including representatives of all lyssavirus species, showed that sequences of the new non-RABV lyssavirus clustered with each other and had a common ancestor with ABLV and RABV in phylogroup 1 (Figure 2). This novel virus was designated as Gannoruwa bat lyssavirus (GBLV).

Representative canine and golden palm civet RABV sequences from Sri Lanka were included in the dataset, but those sequences clustered with other RABVs, distinct from the GBLV sequence (nucleotide identity 78%). Nucleotide identity across the complete genome ranged from 61% (IKOV) to 76.5% (ABLV), which showed that GBLV is a member of the genus Lyssavirus but is distinct from viruses circulating in nonvolant mammals in Sri Lanka.

We report isolation of a novel non-RABV lyssavirus (GBLV) that is most closely related to RABV and ABLV. GBLV is pathogenic; it caused fatal disease in 4 Indian flying foxes, and clinical signs for these flying foxes were similar to those observed in other bat lyssavirus infections (Table). Diagnostic tests identified Negri bodies by staining with hematoxylin and eosin and lyssavirus antigens by dFAT in brain and spinal cord tissue (Figure 1). Molecular techniques identified lyssavirus nucleic acid, and full-genome analysis indicated that GBLV was divergent from known RABVs circulating in Sri Lanka (Figure 2).

Although rabies is prevalent in Sri Lanka, and a number of wildlife species have been confirmed as being rabid, most of the RABVs involved have not been genetically typed. Furthermore, over a 12-year period, only 1 bat tested for RABV was shown to be uninfected (13).

We report a novel non-RABV lyssavirus identified in Sri Lanka, which indicates that Indian flying foxes are a reservoir for lyssaviruses on the Indian subcontinent and nearby regions. Indian flying foxes are widespread in urban and rural areas and occasionally come in contact with humans and domestic dogs, which provides opportunities for virus spillover. Indian flying foxes are also reservoirs for Nipah virus in Bangladesh and India, where transmission to humans has resulted in outbreaks and human deaths. Other bat species, including insectivorous bats, might also be reservoir hosts for lyssaviruses in the study region. Thus, further surveillance is required to understand the role that bats play in the epidemiology of lyssaviruses in Asia.

Continued and extended surveillance of bats and other mammalian species is necessary to determine the distribution and prevalence of GBLV. Detailed phylogenetic analysis and monoclonal typing and antigenic mapping will help clarify the evolutionary relationship between GBLV and other lyssaviruses, in particular RABV and ABLV. In vitro and in vivo cross-neutralization and protection studies will elucidate properties of GBLV and provide information on protection from this virus by available prophylaxis.