Epidemiology of Group A rotavirus in rodents and shrews in Bangladesh

Rodents and shrews live in close proximity to humans and have been identified as important hosts of zoonotic pathogens. This study aimed to detect Group A rotavirus (RVA) and its potential risk factors in rodents and shrews in Bangladesh. We captured 417 small mammals from 10 districts with a high degree of contact between people and domestic animals and collected rectal swab samples between June 2011 and October 2013. We tested the swab samples for RVA RNA, targeting the NSP3 gene segment using real-time reverse transcription-polymerase chain reaction (rRT-PCR). Overall, RVA prevalence was the same (6.7%) in both rodents and shrews. We detected RVA RNA in 5.3% of Bandicota bengalensis (4/76; 95% CI: 1.4–12.9), 5.1% of B. indica (4/79; 95% CI: 1.4–12.4), 18.2% of Mus musculus (4/22; 95% CI: 5.2–40.3), 6.7% of Rattus rattus (6/90; 95% CI: 2.5–13.9), and 6.7% of Suncus murinus (10/150; 95% CI: 3.2–11.9). We found significantly more RVA in males (10.4%; OR: 3.4; P = 0.007), animals with a poor body condition score (13.9%; OR: 2.7; P = 0.05), during wet season (8.3%; OR: 4.1; P = 0.032), and in urban land gradients (10.04%; OR: 2.9; P = 0.056). These findings form a basis for understanding the prevalence of rotaviruses circulating among rodents and shrews in this region. We recommend additional molecular studies to ascertain the genotype and zoonotic potential of RVA circulating in rodents and shrews in Bangladesh.


Introduction
Group A rotavirus (RVA) causes acute dehydrating diarrhea (Wenman et al. 1979;Parashar et al. 2006) in humans, especially children, and animals worldwide. Annual 128,500 deaths and 258,173,300 cases of diarrhea in children <5 years of age are attributable to RVA infection (Troeger et al. 2018). RVA is commonly transmitted via the fecal and oral routes (Anderson and Weber 2004). However, water, food, fomites (de Wit et al. 2003), and flies (Musca domestica) (Tan et al. 1997) can also be a source of infection for humans. Common symptoms of RVA include vomiting and diarrhea in children but nausea, malaise, headache, abdominal cramping, diarrhea, and fever in adults (Anderson and Weber 2004).
Different animal species, including small mammals, are infected with RVA (Dhama et al. 2009). Rodents and shrews belong to a diverse group of small mammals and are found widely throughout the world (Meerburg et al. 2009). They often live in close proximity to humans and domestic animals. Small mammals have the highest capacity for successful adaptation. Twenty-two species of rodents are found in Bangladesh (Khan 2013). Among them, M. musculus (house mouse), R. rattus (black rat/house rat), Bandicota indica (greater bandicoot rat), and B. bengalensis (Indian mole rat) are very common . M. musculus and R. rattus are listed among the 100 of the world's most invasive alien species (Lowe et al. 2000). Of the 160 million people in Bangladesh, 13.6 million are exposed to rodents every 1 3 month, and 8.5% (95%CI: 7.9-9.1) of people have direct contact with rodents (Shanta et al. 2016).
Small mammals' species richness, density, and diversity are an indicator of a healthy and stable ecosystem (Avenant 2003). Although rodents and shrews play an essential role in ecology, they act as reservoirs of zoonotic pathogens (Meerburg et al. 2009) like Hantavirus (Wang et al. 2000;Radosa et al. 2013), Yersinia pestis, Rickettsia (typhus), Leptospira (Weil's disease), Toxoplasma, Trichinella, Hepatitis E (Favorov et al. 2000), Bartonella (Ellis et al. 1999), Borrelia, Babesia, Anaplasma and Ehrlichia (Tadin et al. 2016). Rodents can transmit more than 20 diseases to humans through blood-sucking parasites (fleas, ticks, and mites) (Singla et al. 2008). Rodents and shrews are usually found near human and animal habitations at high densities as well as in other habitats like woodland and abandoned warehouses (Veciana et al. 2012). Due to urbanization and deforestation, wild small mammal populations come to human localities. Human-animal contact frequently occurs, which boosts the chance of cross-species transfer of zoonoses, including rotavirus (Sumangali et al. 2007), and poses a health risk to humans (Paramasvaran et al. 2009). Currently, zoonotic rotavirus infection in humans is more frequent than in the past, and various studies have reported the interspecies transmission of rotavirus from animals to humans (Doro et al. 2015;Martella et al. 2010). There is no way to rule out the possibility of rotaviruses becoming more pathogenic and increasing their transmissibility to humans through reassortment with other genotypes of RVA, similar to influenza viruses (Cowley et al. 2013;Li et al. 2016). Small mammals pose a significant risk to human health, especially those that have direct or indirect contact with animals. However, viruses carried by small mammals have not been well investigated in Bangladesh. The evolution and emergence of rotaviruses in the context of human health is an enigma, and it is public health to better understand the diversity, evolution, and origins of rotaviruses in small mammals.
In Bangladesh, one study reported a 64% prevalence of rotavirus among hospitalized children 5 years of age and younger admitted with acute gastroenteritis (Satter et al. 2017). There is an opportunity for cross-species transmission of zoonotic pathogens, including rotaviruses, because of the higher percentage of human cases, the high density of the human population, frequent contact with animals, including small mammals, and lack of awareness regarding hygiene and sanitation.
Data on rotavirus in small mammals are limited in the Indian subcontinent. RVA was isolated from urban wild rats (R. norvegicus) in Germany (Sachsenröder et al. 2014) and Brazil (Tonietti et al. 2013). Common shrews (Sorex araneus) (21.7%) in Germany were found to be infected with RVA (Johne et al. 2019). The prevalence of RVA in different rodents and shrew species in China was found to be 1.06% and 2.67%, respectively, by RT-PCR (Li et al. 2016), whereas rotaviral antigen was detected in 6.7% of samples from wild tree shrews (Tupaia belangeri Chinese) using ELISA (Wang et al. 2011). Due to the segmented nature of the viral genome, animal RVA can easily infect and adapt to humans through reassortment with other RVAs (Ianiro et al. 2017). Similarly, diverse strains of human RVA have been found that share genetic and antigenic features with animal RVA strains (Ianiro et al. 2017). The G3 genotype frequently infects humans, pigs, dogs and cats, horses, bats, and murine rodents (Geletu et al. 2021;Simsek et al. 2021). Murine RVA strains prolifically infect and multiply in mice only. Even though cross-species transmission of RVA from mice to humans has never been documented, animal-associated RVA is prevalent. However, within the last few years, interspecies transmission and genetic assortment between human and animal rotaviruses from cows, pigs, cats, and dogs have been reported (He et al. 2017;Sawant et al. 2020). Hence, it can be speculated that the cross-species transmission of RVA might occur between humans and rodents in Bangladesh. However, there is no published literature on RVA in rodents and shrews in Bangladesh to our knowledge. Therefore, the study aimed to determine the prevalence of RVA in the rodent and shrew population of Bangladesh. This will help conduct further research on their zoonotic potential in the future.

Study sites and duration
From June 2011 to October 2013, we captured rodents and shrews from 10 different districts (Faridpur, Mymensingh, Rajbari, Rangamati, Khagrachhari, Dinajpur, Maulvibazar, Rangpur, Cox's Bazar, and Joypurhat) in Bangladesh (Fig. 1). We selected sampling sites that were high-risk interfaces where frequent animal-human contact occurs. We categorized the selected study sites based on different land gradients as i) urban areas having a high population density and built environment infrastructure; and ii) Rural areas having a comparatively low population density with more agricultural land and less infrastructure. We captured the rodents and shrews from human dwellings, agricultural fields, and bushland in both urban and rural areas.

Animal capture, sample collection, and data recording
We captured live rodents and shrews using locally made steel wire traps (27 cm × 13 cm × 13 cm) that have proven efficacy in sampling medium-and large-sized small mammals. We baited the traps with ghee-smeared biscuits and dried fish. We set the traps in human dwellings, agricultural fields, or bushland at dusk after obtaining verbal permission from the owners and collected them at dawn the next day.
As rodents and shrews are considered pests, there was no objection from the owners of the properties, and we got their full consent and cooperation. We anesthetized the trapped animals using isoflurane following the procedures described elsewhere (Shafiyyah et al. 2012;Rahman et al. 2018). We collected rectal swabs and/or feces from each captured animal. Our research protocol and methods were carried out in accordance with the national institutional ethics committee of the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b) (reference number: 2008-074) and the international ethics committee of the University of California, Davis Institutional Animal Care and Use Committee (IACUC #16048). After sampling, we released all the animals at the sites of capture. We identified the species of small mammals based on their morphological characteristics as described by Aplin et al. (Aplin et al. 2003). We recorded the lengths of the head, body, tail, hindfoot, and ear and the bodyweight; these parameters helped determine the species and age classes of the animals. We captured a total of 417 small mammals consisting of B. Indica (N = 79), B. bengaleensis (N = 76), M. musculus (N = 22), R. rattus (N = 90), and S. murinus (N = 150). We placed the swab samples into 0.5 mL lysis buffer (NucliS-ENS Lysis Buffer, BIOMERIEUX, France) in a 1.8 mL cryotube (Corning, USA). We then stored the cryotubes in a liquid nitrogen dewar (Princeton Cryogenics, NJ, USA) immediately after collection in the field until their transfer to a -80 °C freezer at the icddr,b laboratory. We used a data sheet to record information like location, habitat, gradient of sampling sites, season and prominent anthropogenic changes, species, age, sex, morphometric measurements, Body Condition Score (BCS), and health status. We categorized the age groups as i) juvenile: weaned, independent from parental nursing, and without developed secondary sexual characteristics like descended testicles in males and ii) adults: matured in size and weight, with developed secondary sexual characteristics (Rahman et al. 2019). We categorized the land gradient as urban or residential, rural, and agricultural land (crop or pastureland). During data collection, we classified the BCS according to Hickman and Swan (2010), as emaciated (BCS-1), under conditioned (BCS-2), well-conditioned (BCS-3), over-conditioned (BCS-4), and obese (BCS-5). Finally, we regrouped the BCS-1 and BCS-2 as poor, BCS-3 as fair, and BCS-4 and BCS-5 as good.

RNA extraction, and PCR
According to the manufacturer's instructions, we extracted viral RNA from 200 μL rectal swab samples using the magnetic particle-based InviMag Virus DNA/RNA Mini kit (STRATEC Molecular GmbH, Germany); the final elution volume was 100 μL. We tested rectal swabs for RVA RNA by rRT-PCR using NSP3-specific primers and probes using the AgPath-ID One-Step RT-PCR system (Ambion Inc. Austin, USA) (Table 1) as described by Jothikumar et al. (2009) and Islam et al. (2020a). We performed conventional RT-PCR using the QIAGEN® One-Step RT-PCR kit (QIAGEN, Germany) to amplify the VP7 and VP4 gene fragments using consensus primer pairs Beg9/End9 and Con2/Con3 for identifying the G and P genotype, respectively (Table 1), as described by Rahman et al. (2007) and Islam et al. (2020b). We used MOCK (only lysis) and known RVA positive samples during the extraction process to ensure proper nucleic acid extraction.

Statistical analysis
We entered the data into MS Excel-2013 (Microsoft office excel-2013, USA) and imported it to STATA-13 (StataCorp, 4905, Lakeway Drive, College Station, Texas 77,845, USA) for analysis. We performed descriptive statistics for different variables using Fisher's exact test. Then, we forwarded the variables (p < 0.2) to multivariable logistic regression and checked for confounding. We tested the model's goodness of fit using the Hosmer-Lemeshow test. Additionally, the predictive ability of the model was determined using the Receiver Operating Characteristic (ROC) curve (Dohoo et al. 2003). The area under the curve (AUC) was categorized as acceptable (AUC = 0.7 to 0.8), excellent (AUC = 0.8-0.9), and outstanding (AUC = 0.9 to 1.0) (Dw 2000;Sayeed et al. 2017). We considered differences among the variables to be significant if p < 0.05.

Discussion
The study detected RVA in different species of rodents and shrews in several areas in Bangladesh. Previous studies identified RVA in other species of rodents and shrews using RT-PCR in Germany (Sachsenröder et al. 2014), China (Li et al. 2016), Brazil (Tonietti et al. 2013), and New York (Williams et al. 2018). In contrast, studies in Australia (McInnes et al. 2011), West Indies (Boey et al. 2019), and Western Europe (Mähler and Köhl 2009) reported rotavirus antibodies in rodents. However, published literature on RVA in small mammals is scarce worldwide. To the authors' knowledge, the present study is the first to report RVA in rodents and shrews in different habitat types and land gradients in Bangladesh. Previous studies reported P and G genotypes of RVA infection in humans and animals in Bangladesh, strongly suggesting the possibility of zoonotic transmissions  . 2 The plot of sensitivity versus 1-specificity for a receiver operating characteristic curve (ROC) of various parameters of the logistic model of RVA in small mammals in Bangladesh 1 3 Mazid et al. 2020;Sharif et al. 2020). Thus, the current study findings extend the known host range of RVA in Bangladesh and will be helpful for the rest of the world prevent and control the spillover of RVA from rodents and shrews. We detected RVA in Rattus rattus, but in Brazil and Germany, RVA was seen in R. norvegicus (Tonietti et al. 2013;Sachsenröder et al. 2014). Similarly, we identified RVA in M. musculus, whereas studies in Australia reported it in Mus musculus domesticus, a subspecies of M. musculus Smith et al. 1993). However, our study did not genotype R. rattus and Mus musculus mice. Nevertheless, earlier studies reported the presence of both R. norvegicus and M. musculus doemesticus from Bangladesh (Adhikari et al. 2018; Barman et al. 2020). The prevalence of RVA in our studied rodents was higher than in previous studies (Li et al. 2016;Ianiro et al. 2017), but in the case of shrews, our results are similar to those of a former study from China that found a 6.7% seroprevalence in wild tree shrews (Tupaia belangeri Chinese) using antigen-capture ELISA (Wang et al. 2011). A study in Germany reported a 21.7% prevalence of RVA in common shrews, which is much higher than in our study (Johne et al. 2019). The higher prevalence may be due to the differences in host species (Sorex araneus) and the wide circulation of RVA in shrews from different regions in Germany.
A comparatively high prevalence of RVA was found in rodents and shrews in urban areas. Rodent and shrew density is higher in urban and peri-urban areas compared to rural areas in Bangladesh (Shanta et al. 2016). Moreover, more than 90% of urban and peri-urban household members observed rodents and shrews on their premises, where 8.5% of respondents reported direct contact with them. A previous study from Bangladesh reported a high density of rodents and shrews in urban areas, making them more susceptible to infectious agents either by functional suppression of the immune system caused by a viral infection, malnutrition, or the stress induced by overcrowding . Besides, urban areas are densely populated with humans where rodents and shrews are considered pests. These small mammals live in human houses (Veciana et al. 2012). They collect food at night and pick up contaminated food from human sources, which may also expose them to RVA. Moreover, RVA is circulating in domestic and wild animals in developing countries like Bangladesh, where humans and animals live in proximity and have frequent interactions ). This also indicates the possible transmission of RVA from humans to small mammals (Ianiro et al. 2017), which may explain the greater positivity rate of samples near human dwellings.
We found the highest odds ratio for RVA presence in M. musculus but could not find any previous study to compare our results. However, some studies have evaluated the presence of antibodies against RVA in different animals. One study estimated the seroprevalence of RVA in laboratory mice and rats in Western Europe and found the highest seroprevalence in mice (Mähler and Köhl 2009). We found RVA in a higher percentage of small mammals sampled during the wet season. A study conducted on M. musculus domesticus by Singleton et al. (1993) also found similar trends of seroprevalence during the wet season (April to September) in Southern Australia. The authors also reported increased mouse densities over time. Though we did not record any animal density-related information, this may be the case in our study.
Additionally, serology does not confirm the presence of infection, instead indicating recent infection. All the positive samples from sick animals were collected during the wet season. From April to September (wet season), comparatively little food is available for the animals. Moreover, the quality of food available is not adequate, resulting in malnutrition among small mammals. September to October is also the breeding season, creating social stress on male mice ). All these factors may influence viral prevalence and persistence in the studied animals.
Sex and BCS also influenced the presence of RVA in our study, but the reason behind this is not clear. One explanation may be that male mammals are more active than females, which predisposes them to infection from various sources. Sometimes, BCS provides more precise information to assess the health status of animals (Hickman and Swan 2010). Usually, animals with poor body conditions are immune-compromised and have limited access to food, which makes them susceptible to different infectious diseases ). However, we admit that the observed relationship between health conditions and RVA presence in the animals may be due to a sampling artifact. We captured a disproportionately higher percentage of rodents and shrews in the wet seasons, corresponding to a study in Uganda (Ssuuna et al. 2020). Environmental factors influence the composition and abundance of rodent species, and rapid growth of vegetation occurs during the rainy season, providing shelter and food for rodents and shrews. Besides, human activities differ concerning different months and seasons, influencing the capture of small mammals (Mulungu et al. 2003).
Human infections associated with group A, B, and C rotavirus are common in Bangladesh ), but the rodent-borne RVA is unknown as data regarding the incidence of zoonotic human RVA infections in Bangladesh are not available. We found RVA in rodents and shrews in Bangladesh, with some critical factors related to RVA infection in peri-domestic species. However, it is crucial to establish longitudinal surveillance across multiple regional animal populations to detect and genetically characterize the RVA. This study did not focus on proving human infection from rodents and shrews. However, as rodents and shrews often live in close contact with humans, although the presence of RVA in rodents and shrews in this study was low (6.7%), it would not be rational to overlook the risk of small mammal borne RVA infection in humans. On the other hand, human RVA strains cannot infect mice under laboratory conditions. Therefore, it is also unlikely that human RVA will infect rodents and shrews (Ciarlet et al. 2002).
The study has some limitations. The primers and probes used were not designed for rodents. However, these primers and probes have a wide range and are used for detecting RVA in clinical and environmental samples. We detected RVA in macaques, bats, and domestic animals using the same primers targeting the NSP3 gene in earlier studies Islam et al. 2020a;Islam et al. 2020b). In the current study, we did not have access to rodent-specific internal positive controls (IPC). Using IPCs would have excluded the possibility of getting false-negative results. We also could not successfully amplify and sequence the VP4 and VP7 genes from RVA-positive samples. This may be due to lower nucleic acid content in the swab samples.
Additionally, we cannot ignore the genetic diversity of RVA in rodents, which may impact the performance of the primers used in this study (Čolić et al. 2021). Nevertheless, next-generation sequencing and/or primer-independent approaches are needed to improve the characterization of RVAs in the fecal samples of rodents and shrews in Bangladesh. We recommend establishing longitudinal surveillance to detect and genetically characterize RVA among multiple host populations, conduct DNA barcoding to identify spatial patterns in these species.

Conclusions
Taken together, the findings of this study suggest that RVA is circulating in rodents and shrews in Bangladesh. Similar to other wildlife species, it is difficult to detect and sequence RVA in small mammals. As rodents and shrews often live in close contact with humans as well as with farm and pet animals, they form a distinct and significant nexus between wildlife communities and human populations. We recommend further studies on the molecular characterization of RVA in rodents and shrews, their epizootiology, and possible risks to humans at different land gradients in Bangladesh.