Background
Coronaviruses (CoVs) are classified into four genera:
Alphacoronavirus (αCoV),
Betacoronavirus (βCoV),
Gammacoronavirus, and
Deltacoronavirus [
1]. CoVs infect wide variety of mammals and birds, causing upper and lower respiratory, hepatic, enteric and neurological illnesses with varying severity. Bat CoVs (BtCoVs) are likely the gene source of αCoV and βCoV, while avian CoVs are sources of
Gammacoronavirus, and
Deltacoronavirus [
2]. Although there is single lineage in αCoV, βCoVs are further separated into four lineages (A – D) [
3]. Lineage A βCoV, including bovine CoVs, human CoV (HCoV)-OC43 and related viruses, have been detected in various mammals such as cows, horses, deer, antelopes, camels, giraffes, waterbucks, dogs, and humans worldwide, but not in bats. Lineages B-D βCoVs have been detected in bats worldwide [
4].
Currently, six CoV strains are known to cause human infection; four CoVs cause mild respiratory illness, including two αCoVs: HCoV-NL63 and HCoV-229E, and two βCoVs: HCoV-HKU1and HCoV-OC43 [
5]. The other two βCoVs cause severe respiratory tract infection with high-fatality rates, such as severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS), belonging to lineages B and C, respectively. Bat related MERS-CoVs phylogenetically-related to humans have been previously discovered in
Tylonycteris pachypus (BtCoV-HKU4) and
Pipistrellus abramus (BtCoV-HKU5) in Hong Kong, in 2006 [
6],
Neoromicia capensis (NeoCoV) from South Africa, in 2011 [
7], and
Pipistrellus cf.
hesperidus (PREDICT/PDF-2180 CoV) from Uganda, in 2013 [
8]. SARS-like BtCoV was initially identified from the genus
Rhinolophus in 2005, after the SARS outbreak in humans in 2002–2003, and identification of virus in palm civets (
Paguma larvata) from live animal market in Guangdong, China in 2003 [
9,
10].
BtCoVs have been identified in many insectivorous and frugivorous (family Pteropodidae) bats on many continents: America, Europe, Africa, and Asia [
4]. Different species of Pteropodidae have been identified as a major source of lineage D βCoV (HKU9) in Africa (
Rousettus aegyptiacus, Kenya [
11],
Pteropus rufus and
Eidolon dupreanum, Madagascar [
12]), and Asia (
R. leschenaulti, China [
13],
Cynopterus brachyotis, Philippines [
14],
Ptenochirus jagori, Philippines [
15],
Pteropus giganteus, Bangladesh [
16],
Cynopterus sphinx, Thailand [
17].
Thailand is home to 146 bat species (125 insectivorous and 21 frugivorous) [
18]. The prevalence and diversity of BtCoVs in Thailand has been studied in the last decade [
17,
19,
20]. CoVs were found in 11 insectivorous bat species and in 2 frugivorous bat species. However, data from
Pteropus bats have been lacking despite
Pteropus being the biggest colony of Pteropodidae in Thailand. Three species
(P. lylei,
P. vampyrus and
P. hypomelanus) are reservoirs of Nipah virus (NiV) in Thailand [
21]. The prevalence of NiV RNA in urine of
P. lylei has been seasonally detected during the months of May and June [
22].
P. lylei (Lyle’s flying fox (LFF)) ranges from Yunnan in China, and extends to Cambodia, Thailand, and Vietnam [
23]. Up to 20 colonies have been identified in Thailand [
24] and the largest known colony comprises of about 10,000 individuals [
22]. It shares foraging areas with other frugivorous bats in fruit trees, from which the fruits are also shared by humans. Moreover, trees in populated temple grounds and cultivated land are common roosting sites for LFF. Thus, consumption of partially eaten fruit, uncooked meat, or contact with saliva, urine or faeces, which can be contaminated with bat viruses, poses a risk of viral transmission from LFF to humans or domestic animals.
The potential for emergence of zoonotic viruses into the human population depends on the prevalence of the virus in its host species, host range mutations within viral quasispecies, and the degree to which the reservoir host interacts with humans [
25]. To better understand the prevalence, persistence, phylogeny, and potential for interaction with humans, here we describe a comprehensive longitudinal study to detect CoV in LFF, and factors influencing infectivity. Bat rectal swabs were collected monthly from their roosting area and from two human dwellings (foraging sites) nearby. Individual bats were weighed and forearm (FA) lengths were measured for further characterization on its body composition index (BCI). Our results demonstrated for the first time that α- and β-CoVs are endemically circulating in LFFs in Thailand, and that age and BCI are significantly different between infected and uninfected bats.
Discussion
This is the first longitudinal study of CoV infection in wild bats in Thailand, where 367 LFF bats were captured monthly for one year at one roosting site and two foraging sites close to the bat roost. One fourth of bats were juvenile, and 59.9% were male.
The ratios between captured juvenile and adult bats were different at the bat roost and foraging sites. Only 13% of juvenile (30/224) bats were captured at the bat roost in the year of the study, whereas half of the juvenile bats were captured from both foraging sites (24/44, 49/99, from S1 and S2 respectively). The maximum linear distances between roosts and foraging areas of LFF at this site varied from 2.2–22.3 km [
30]. Foraging sites near roost, even with limited food sources, may be practical for young or unhealthy bats that are unable to fly far.
CoV RNA was detected in approximately 18% of all bats sampled, which is in the same range as the study in China (16%, [
31]; 15.8%, [
32]), and Germany (9.8%, [
33]). The prevalence of CoV infection in
Pteropus bats (
P. rufus) from Madagascar was similar to this study (17.1%, 13/76) [
12]. On the other hand, the prevalence in this study was higher than the two previous studies in Thailand by Wacharapluesadee et al. (6.7%, 47/626) [
17] and Gouilh et al. (10.5%, 28/265) [
20]. This may be the result of a bias from the cross sectional study of these two previous studies or an indication of difference in prevalence rate in different bat species.
Ratios of captured bat genders in this study were roughly similar at foraging sites. At the roost, male bats were predominantly captured. CoV infection was not correlated with sex of bat, neither at the roost nor at the foraging sites. This finding is similar to the studies from Germany [
33] and Colorado, USA [
25].
In our study, CoV infection was found to be associated with younger ages; 39.8% of juvenile bats versus 10.2% adult bats were positive for CoV RNA. Similar findings have been reported from the study in insectivorous bats from USA (19% juvenile versus 9% adult bats positive for CoV) [
25] and Vespertilionid bats in Germany (23.7% juvenile versus 15.9% sub-adult versus 8.5% adult bats positive for CoV) [
33]. These findings support the hypothesis that young bats may be more susceptible to CoV infection, and serves to propagate and play an important role in maintaining the virus within bat colonies. The divergence in rate of CoV infection from different study sites (Table
1) was likely to be influenced by the age and body condition of bats.
Three of 9 unweaned pups were CoV RNA positive, while their mothers and all lactating female carrying pups were negative for CoV. It may be possible there was a placental transmission, after which the virus was then cleared from adult female bats. Another possibility is that the unweaned bats acquired infection from contaminated secretion of other bats hanging from the same tree. However, the study by Gloza-Rausch et al. 2008 [
33], where 54 of 178 (30%) of studied female bats were lactating, found higher rate of CoV infection in lactating bats (22.4%) than in non-lactating bats (9.7%) which supports the first scenario. It is to be noted that limited number of lactating bats were included in our study (9 of 147, 6.1%). Targeting mother-pup pairs in future studies would be required to confirm the vertical (placental) transmission of CoV in LFF.
Seasonal prevalence was mostly related to the number of juvenile bats captured for testing in each month (Table
2), except in January when all four CoV positive bats were adult. Notably, these positive adult bats had lower BCI (2.83, 1.77, 1.84, 2.75) than the mean uninfected adult bats (2.88). Three of the 4 infected adult bats had lower body mass (444, 429, 258, 276 g) than mean uninfected adult bats (439 g). The mean body mass of infected bats was significantly lower than in uninfected bats (Table
3). This is similar to the study where
Hipposideros pomona bats in Hong Kong with HKU10 CoV infection had lower body mass than uninfected bats, even though they appeared to be healthy [
34]. These bats seemed to be in poor condition, serving as the other group in addition to juvenile bats that further maintained the virus within the population.
Sixty eight CoVs were detected from this study, forming 2 genetically distinct strains. Sixty four belonged to βCoV (SARS-related group) with relatively close homology to the reference virus, BtCoV-HKU9 [
6]. Four belonged to αCoV, and their sequences related to CoVs previously detected in insectivorous bats in Thailand such as
H. lekaguli,
H. armiger and
Taphozous melanopogon [
17]. This supports the possibility of interspecies transmission, rather than virus-host specific sharing, between bats of different suborder (
Pteropus in Pteropodidae, Hipposideridae and Emballonuridae) that do not share food, foraging sites, or roosts, similar to the earlier HKU10 CoV study between
R. leschenaulti and
H. pomona bats [
15]. The evolution of CoVs in different host species-order should be further studied in order to understand the route of spillover and transmission.
Bats from different species-genus that share foraging sites may also share infections and particular CoV strains, for example βCoV from LFF (this study),
C. sphinx,
S. Heathii, and
S. kuhlii [
17] (Fig.
4a-b). βCoV from same bat genus in different geographic region displayed distinct clusters (Fig.
4b), for example
P. rufus from Madagascar (cluster 2–3) [
12] and LFF from this study (cluster 4). This demonstrated the βCoV inclines interspecies sharing rather than virus-host specific sharing.
Given the mobility of LFF in Thailand, where the maximum linear distance between day roosts and foraging areas for LFF is 23.6 km [
29], and its tendency for sharing habitat with other colonies, the detected strains of CoVs from this study may be found in LFFs all over the region. The high prevalence of CoV in this study suggests circulation of infection within the bat colony. Study of CoV diversity from other LFF colony in Thailand and region is required to improve our understanding of the evolution and spillover patterns of CoV.