Detection of Dengue Virus in Bat Flies (Diptera: Streblidae) of Common Vampire Bats,Desmodus rotundus,in Progreso,Hidalgo,Mexico

Worldwide, flaviviruses are responsible for numerous cases of human illnesses, and their emergence is a public health priority. Many arthropod species yet to be identified as disease vectors could potentially play a role in viral maintenance (Junglen and Drosten 2013). The presence of different flaviviruses in various arthropod species (including Uroteania and Culex spp.) has been reported, suggesting a complex epidemiology not fully described (Junglen and Drosten 2013). To date, there have been more than 70 flaviviruses isolated from over 50 reservoir species. The genome size of the Flaviviridae family is ∼9–11 kb; the NS5 gene is highly conserved, as it encodes a RNA polymerase, a protein responsible for the replication and capping in the family Flaviviridae (Kuno et al. 1998). Among the viruses associated with bats, Flaviviridae is the second most frequently detected after the Rhabdoviridae (Sotomayor-Bonilla et al. 2014).

The primary transmission cycle described for dengue virus (DENV) is the urban endemic–epidemic (between humans and Aedes aegypti). A second sylvatic-enzootic cycle has also been described (between nonhuman primates and wild Aedes mosquitoes), but it is less well understood (Rico-Hesse 2003, Sotomayor-Bonilla et al. 2014). We previously reported the presence of DENV in insectivorous and frugivorous bats from Mexico (Aguilar‐Setién et al. 2008, Sotomayor-Bonilla et al. 2014). In French Guiana, DENV has also been reported in bats, marsupials, and rodents (de Thoisy et al. 2009). Although the blood is an ideal medium for the dispersal of several pathogens, less is known about the potential for blood-feeding bats.

While the urban cycle DENV is maintained between humans and Aedes aegypti, the mechanisms for the maintenance of DENV in sylvatic cycle are still unknown, despite DENV having been detected by molecular methods (RT-PCR) in wildlife from different countries (Aguilar‐Setién et al. 2008, de Thoisy et al. 2009, Sotomayor-Bonilla et al. 2014). The objective of this study was to detect Flaviviruses in ectoparasitic flies and their common vampire bat (Desmodus rotundus) hosts in Mexico.

Common vampire bats were trapped, using mist nets, from a cave roost in Progreso, Hidalgo, Mexico (n = 160) (18° 54′ 43″ N; 98° 59′ 21″ W) during three sampling dates (November 2014, May 2015, and July 2015). Bat flies were collected using fine forceps. Sixteen bats (10% of the sample) randomly chosen were sacrificed to collect their tissues and screen for flaviviruses. All samples were conserved in dry ice until long-term storage at −70°C.

A total of 557 bat flies were collected, belonging to two different species: Strebla wiedemanni (N = 121) and Trichobius parasiticus (N = 436). Between five and nine flies were pooled, separated by fly species. RNA extraction from ectoparasites and bat tissues (liver or spleen) was performed by the TRIzol^® method. The SuperScript One-Step RT-PCR with Platinum Taq Kit (10928-042; Invitrogen) was used for NS5 gene RT-PCR amplification, according to the manufacturer's instructions. A primer set modified from Moureau et al. (2008) was used (Moureau et al. 2008). The FlaviPF1S (5′TGYRTBTAYAACATGATGGG3′) and FlaviPF2Rbis (5′GTGTCCCANCCNGCNGTRT3′) primers were used for the first amplification and FlaviPF2Rbis plus FlaviPF3S (5′GCNATHTGGTAYATGTGGYT3′) for the semi-nested reaction with KAPA Taq polymerase (KK1015; Kapa Biosystems). One fragment of 272-bp and another of 197-bp were obtained, respectively, and the DENV-1 (YUC18494) strain was used as a control. From the analyzed bat fly pools, 39.6% (38/96) were positive; 21.2% (7/33) were positive from November 2014, 6.2% (2/32) in May 2015, and 93.5% (29/31) in July 2015. The highest prevalence was observed during July, which coincides with the rainy season and increased mosquito populations, similar to that reported in Australia by Huang et al. (2013). Of the bat tissues screened for DENV, 50% (8/16) were positive.

Amplicons of six samples (three bat fly pools and three bat tissues) were sequenced directly by the Sanger method in a Genetic Analyzer 3100 (Applied Biosystems). The recovered NS5 sequences were aligned using ClustalX ver. 2.1 with sequences of other viruses in the same family available in GenBank (NCBI site). A phylogenetic inference analysis was performed using the maximum likelihood algorithm implemented in PhyML (PhyML 3.0 2017). The most appropriate nucleotide substitution model was determined in jModelTest ver. 3.7 using Akaike Information Criterion (Posada and Crandall 1998). The best model was TrN+I+G (AIC = 13698.9335, −lnL = 6698.4667, I = 0.18, G = 0.4940, k = 151, where–lnL is the value of the negative natural logarithm of maximum likelihood, I is the proportion of invariant site, G is the gamma shape parameter, and k is the number of parameters). The topology of the tree was optimized rather than branch length. To estimate the reliability of each node, a bootstrap test using 1000 pseudoreplicates was performed. The Batu cave virus sequence (accession no. AF013369.1) was used as outgroup.

Phylogenetic analysis showed that the obtained sequences (Progreso1S, Progreso2T, Progreso3S, Progreso4D, Progreso5D, and Progreso6D) were branching (integrated) with other DENV sequences, which are independent from other viral genera (bootstrap 53.5%) (Fig. 1). To confirm the presence of DENV, an additional RT-PCR described by Lanciotti et al. was performed (Lanciotti et al. 1992), and the same six samples were positive for DENV-2. From these, the isolation on C6/36 cells and a RT-PCR (Lanciotti et al. 1992) from cellular lysates were performed, and just one sample was positive. This may be because the quantity of inoculum was very small, and therefore, insufficient to infect cells in the other samples. Our results show that DENV circulates in the region and potentially more broadly depending on transmission between different species.

OPEN IN VIEWER

FIG. 1.

Maximum likelihood phylogenetic tree of dengue virus sequences detected in ectoparasitic flies (S. wiedemanni and T. parasiticus) and the common vampire bat (D. rotundus) in Progreso, Hidalgo, Mexico. Several sequences of other flavivirus obtained from GenBank are included. The best model was TrN+I+G (I = 0.18, G = 0.4940), and Bootstrap values are shown near to nodes.

Other studies have reported the presence of flaviviruses in various arthropods, particularly in hippoboscid flies: during the 1999–2004 West Nile Virus (WNV) outbreak in Ontario, Canada, 88.9% of owl (Icosta americana) ectoparasites were positive for WNV (Gancz et al. 2004). Alkhurma hemorrhagic fever virus has also been detected in camel ticks, and there is a report of a new flavivirus sequence, NOUV, in Uranotaenia mashonaensis mosquitoes (Abdel-Shafy and Allam 2013, Junglen and Drosten 2013). This supports the possibility of identifying new or known flaviviruses in arthropods not been previously screened for these pathogens.

According to the study by Rico-Hesse (2003), some DENV strains are limited in wild systems, where several mosquito species, which mostly do not feed on humans (species other than A. aegypti or A. albopictus) cohabitate in the canopy, and besides the author considers that some viral strains have low transmissibility or virulence for humans (Rico-Hesse 2003). Further research is needed to determine if the DENV detected in S. wiedemanni and T. parasiticus during this study is the result of a DENV strain adaptation to this particular ecological niche. Currently, the Asian genotype (DENV-2) is widely present in human populations and expanding worldwide. Therefore, we hypothesize that the role of these bat ectoparasites in the DENV epidemiology may be similar to the role of mosquito species in the forest canopy, where they do not directly transmit to humans but rather act as a DENV reservoir.

DENV sequences obtained from bat flies (S. wiedemanni and T. parasiticus) and their common vampire bat (D. rotundus) hosts clustered with other dengue sequences based on analysis of the NS5 gene. Besides, a high prevalence of positive samples was found (39.6%). Although this study supports other reports of DENV detection in bats and arthropods other than Aedes mosquitoes, the role of these ectoparasitic flies and of hematophagous bats in the epidemiology of DENV still warrants further investigation. In addition, it is important to know if the isolates of DENV obtained from these sources are feasible or can develop an infection in hematophagous bats, which would help to the understanding of the presence of such viruses in these ectoparasites of bats.