Prevalence of Leptospira Infection in Rodents from Bangladesh

Worldwide, Leptospira infection poses an increasing public health problem. In 2008, leptospirosis was recognised as a re-emerging zoonosis of global importance with South-East Asia being one of the most significant centres of the disease. Rodents are thought to be the most important host for a variety of Leptospira serovars. Because Bangladesh offers a suitable humid climate for the survival of these pathogenic bacteria, the presence of rodents could be a serious risk for human infection, especially in peri-urban areas or locations where food is stored. In order to gain more understanding of the multi-host epidemiology, a prevalence study was conducted in Comilla, Bangladesh to determine the presence of pathogenic Leptospira species in rodents. Real-time Polymerase Chain Reaction (qPCR) and sequencing showed that 13.1% (61/465) of the trapped rodents were infected with pathogenic Leptospira. Sequencing of the qPCR products identified the presence of three species: Leptospira interrogans, Leptospira borgpetersenii, and Leptospira kirschneri. Rodents of the genus, Bandicota, were significantly more likely to be positive than those of the genus, Rattus and Mus. Our results confirm the importance of rodents as hosts of pathogenic Leptospira and indicate that human exposure to pathogenic Leptospira may be considerable, also in places where food (rice) is stored for longer times. This study emphasizes the need to improve rodent management at such locations and to further quantify the public health impacts of this neglected emerging zoonosis in Bangladesh.


Introduction
Commensal rodents are known to cause substantial pre-and postharvest losses. It is estimated that rodents contribute to 5% to 10% of the losses to rice production in Asia [1]. Besides causing direct loss to stored food, rodents also cause indirect loss: Their gnawing makes stored produce more prone to insect or fungal attacks and they contaminate a large percentage of produce with their droppings, urine, and saliva, which could possibly harbour pathogens [2][3][4][5][6]. A review by Meerburg et al. [3] points out the links between food security and rodents as rodents are potential reservoir hosts for over 60 zoonotic pathogens [7][8][9]. Asia is predisposed for (infectious) disease emergence [10] and there around food storages. The objective of this study was to assess the presence and infection rate of pathogenic Leptospira in commensal rodent species in rural Bangladesh.

Study Area
From March 2015 until March 2016, the first part of this study was conducted in a rice mill in the South-East region of Bangladesh. Two rice milling stations and 8 villages were selected for further research in 2016 and 2017. The villages were visited for 6 months per year: June, July, August, and October, November, December. Both rice mills were visited for a period of nine months, from August 2016 until March 2017. The selected rice mills and villages are situated in the Chittagong division, Comilla district (all within a 10 km circle from 23 • 27 23 N 91 • 10 20 E). Rodent trapping was conducted in the rice storage areas of the milling stations, where the paddy rice from the fields is stored in jute bags. The owners of the rice mills participated in the project and agreed to the use of their property and buildings for this study. In the villages, rodent trapping was conducted in the rice storage area of households. No ethics approval was required because the rodents that were trapped are pest species and the Bangladesh government has no regulations on animal ethics concerning pest species. However, all local staff were trained to work with the animals according to the Netherlands code of scientific practice and the participating researchers completed the Laboratory Science Course as required by European Directive 2010/63/EU and the revised Dutch Act on Animal Experimentation. Although the animals used were not laboratory animals, the NCad opinion on 'Alternative methods for killing laboratory animals' was followed, as provided by the Netherlands National Committee for the protection of animals used for scientific purposes. Moreover, the guidelines of the American Society of Mammologists were followed during the study. The above procedure was also mentioned in the original project proposal and approved by the donor (Netherlands Organization for Scientific Research, NWO-WOTRO).

Rodent Trapping
During the research period, every 14 days, rodent trapping was conducted for 3 consecutive periods of 24 h. Rodent trapping was performed by the use of 10 kill traps (14 × 7 cm; Big snap-e; Kness, Albia, IA, USA) and 10 life traps (purchased at local Bangladesh markets, Figure 1). Traps were placed during evening time at locations where rodent damage was observed and tracks were seen. Traps were checked by research staff the next morning. Rodents trapped in life traps were immediately euthanized by cervical dislocation and after identification of the species, gender, and life stage (juvenile/mature), each animal was dissected. Of each rodent, one kidney was randomly taken and stored in 98% ethanol and shipped to the OIE and National Collaborating Centre for Reference and Research on Leptospirosis (NRL), Amsterdam, The Netherlands for PCR testing.

Study Area
From March 2015 until March 2016, the first part of this study was conducted in a rice mill in the South-East region of Bangladesh. Two rice milling stations and 8 villages were selected for further research in 2016 and 2017. The villages were visited for 6 months per year: June, July, August, and October, November, December. Both rice mills were visited for a period of nine months, from August 2016 until March 2017. The selected rice mills and villages are situated in the Chittagong division, Comilla district (all within a 10 km circle from 23°27'23'' N 91°10'20'' E). Rodent trapping was conducted in the rice storage areas of the milling stations, where the paddy rice from the fields is stored in jute bags. The owners of the rice mills participated in the project and agreed to the use of their property and buildings for this study. In the villages, rodent trapping was conducted in the rice storage area of households. No ethics approval was required because the rodents that were trapped are pest species and the Bangladesh government has no regulations on animal ethics concerning pest species. However, all local staff were trained to work with the animals according to the Netherlands code of scientific practice and the participating researchers completed the Laboratory Science Course as required by European Directive 2010/63/EU and the revised Dutch Act on Animal Experimentation. Although the animals used were not laboratory animals, the NCad opinion on 'Alternative methods for killing laboratory animals' was followed, as provided by the Netherlands National Committee for the protection of animals used for scientific purposes. Moreover, the guidelines of the American Society of Mammologists were followed during the study. The above procedure was also mentioned in the original project proposal and approved by the donor (Netherlands Organization for Scientific Research, NWO-WOTRO).

Rodent Trapping
During the research period, every 14 days, rodent trapping was conducted for 3 consecutive periods of 24 h. Rodent trapping was performed by the use of 10 kill traps (14 × 7 cm; Big snap-e; Kness, Albia, IA, USA) and 10 life traps (purchased at local Bangladesh markets, Figure 1). Traps were placed during evening time at locations where rodent damage was observed and tracks were seen. Traps were checked by research staff the next morning. Rodents trapped in life traps were immediately euthanized by cervical dislocation and after identification of the species, gender, and life stage (juvenile/mature), each animal was dissected. Of each rodent, one kidney was randomly taken and stored in 98% ethanol and shipped to the OIE and National Collaborating Centre for Reference and Research on Leptospirosis (NRL), Amsterdam, The Netherlands for PCR testing.

Laboratory Diagnostics
Processing of the kidney samples for testing on pathogenic Leptospira was conducted at the NRL. From each kidney, a small transversal sample (up to 25 mg) was taken and processed for DNA extraction using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). A slightly adjusted protocol was applied; the amount of ATL buffer, proteinase K + , AL buffer, and absolute ethanol were doubled to ensure complete tissue lysis. Hereafter, DNA extraction took place according to the manufacturer's protocol. The extracted DNA samples were stored at −20 • C until tested by real-time quantitative polymerase chain reaction (RT-qPCR). DNA samples were 1:10 diluted using UltraPure DNase/RNase-free distilled water (Invitrogen, UK) and tested in triplicate using SYBR Green real-time qPCR, targeting the secY gene

Statistical Analysis
Descriptive statistics were used, and for comparisons between rodent species, Fisher's exact test was conducted. Results were considered statistically significant with a p-value of p < 0.05. Statistical analyses were performed by using SPSS, version 23 (IBM SPSS Statistics Inc, College Station, TX, USA).

Laboratory
No anomalies were found in the animals during the dissections. Out of 465 rodents, 177 rodents showed a positive qPCR result. Sequencing of the samples from the qPCR run revealed 61 samples which showed sequence data of the partial secY gene upon alignment. Table 1 shows these real-time PCR and sequencing results (more detailed information on all 61 positive samples can be found in Appendix A). In total, 13.1% (61/465; SD = 0.33) of the tested rodents were infected with pathogenic leptospires ( Table 2). Of the five Mus terricolor animals that were trapped, none were positive for Leptospira.
Significant differences were found when analysing for infection probability between species: B. indica showed significantly higher infection rates with Leptospira than all other species (p < 0.05, two tailed Pearson Chi-Square, Figure 2), and B. bengalensis and R. exulans both showed significant higher infection rates than M. musculus and R. rattus.
When considering the two seasons (wet and dry), a significant difference was found for the effect of season (wet/dry) on the chance of infection with Leptospira (p = 0.019, two tailed Pearson Chi-Square, Table 4), with a higher chance of infection in the dry season.

Discussion
Pathogenic L. interrogans, L. borgpetersenii, and L. kirschneri species were identified in the rural Bangladesh rodent population. Almost 1 out of every 7 (61/465) rodents trapped had Leptospira bacteria in their kidneys and where thus potentially capable of shedding leptospires in and around food storage. In our study, no damage or anomalies were found during the dissections, which confirms the role of the animals as a natural reservoir. These results indicate that the risk of human exposure to pathogenic Leptospira is likely to be substantial for the workers of the Rice Milling Station of Comilla, and also for local people, since people can acquire leptospirosis via direct or indirect contact with the urine of an

Discussion
Pathogenic L. interrogans, L. borgpetersenii, and L. kirschneri species were identified in the rural Bangladesh rodent population. Almost 1 out of every 7 (61/465) rodents trapped had Leptospira bacteria in their kidneys and where thus potentially capable of shedding leptospires in and around food storage. In our study, no damage or anomalies were found during the dissections, which confirms the role of the animals as a natural reservoir. These results indicate that the risk of human exposure to pathogenic Leptospira is likely to be substantial for the workers of the Rice Milling Station of Comilla, and also for local people, since people can acquire leptospirosis via direct or indirect contact with the urine of an infected host, which in this case can also lead to a risk for the consumers of rice. However, although handling the rice may be dangerous, the risk of contracting leptospirosis via food consumption is limited if the rice is properly cooked.
The prevalence of leptospirosis in humans in Thailand, Malaysia, and India has been reported for some areas with infection rates between 15% and 35% [55][56][57][58][59][60][61][62][63][64]. Despite the occurrence of leptospirosis in South-East Asia, only a few studies have been performed on Leptospira prevalence in rodents, and to our knowledge, no studies from Bangladesh have been published before. Moreover, there are only scarce and frequently dated reports about the epidemiology of leptospirosis among citizens in Bangladesh. Research from 1994 showed a human seroprevalence of 38% (n = 89) in a rural district of Bangladesh close to rivers that regularly flood [39]. However, no link with risk factors (such as rodents) was made. In 2001, febrile patients (n = 1297) from two hospitals in Dhaka, Bangladesh were tested on leptospirosis and 63 patients (4.8%) were found to be positive [23]. More recent research from Bangladesh showed that over 13% of febrile people (n = 584) were serologically positive for Leptospira organisms [65]. One study on the prevalence of Leptospira in Bangladesh looked at dairy cows in Chittagong and showed that almost 50% of the samples were positive for Leptospira organisms [66], which poses a high infection risk to people working on cattle farms.
A study from Cambodia (2012) on Leptospira in rodents showed an overall infection of 11.1% (n = 642) [67] and a study from Malaysia found 11.0% (n = 357) [68], which correlates with our findings (13.1%), although we did not use a serological assay but a molecular assay. A serological study from Vietnam showed that 22% of trapped rodents host L. interrogans and all rodents were trapped in urban areas close to the South-China Sea and in Hanoi City, a city along the Red River [69]. Research from 2003 conducted on the Andaman Islands showed a seroprevalence of 7.1% in R. rattus (n = 85) [70], which again is in line with our findings (7.9%). Research on rodents from a suburban area in India showed 14.3% of the trapped R. rattus (n = 28), and 16.1% of the B. bengalensis (n = 58) were serological positive for Leptospira [71]. The findings specifically on R. rattus infections in India by Saravanan et al. were higher than the 7.9% found in the R. rattus (n = 191) from the current study in Bangladesh. This difference could be caused by the sensitivity and specificity of the used testing methods (serology vs. molecular diagnostic), or differences in the environment, to the ratio in which rodent species occurred and possibly also to the trapping easiness (or shyness) of each species. Meta-studies on Leptospira prevalence in rodents in South-East Asia (Table 5) showed that B. bengalensis, B. indica, B. savilei, and R. exulans were infected with Leptospira, whereas investigated species of the Mus genus appeared to be uninfected [60,61,72,73]. In contrast, we found a Leptospira prevalence of 5.5% in trapped M. musculus.
For B. bengalensis and R. rattus, the ratio of males and females trapped was almost evenly distributed (≈1:1). For R. exulans (although we trapped only 23 animals of this species), more males than females were trapped (18 out of 23). For the other three species from this study, B. indica, M. musculus, and M. terricolor, more female specimens were collected. Research in 2006 from Pakistan found 40% of the trapped B. bengalensis to be female (n = 167). This male:female ratio can, in comparison to our results, be explained due to the fact that rodents trapped in this study in Bangladesh were trapped at indoor locations only, whilst in the study of Rana et al. [74] they were trapped over multiple crop fields, which reflects the difference between male and female behaviour more. Males are more active in their explorations and depredation on food crops, whereas females may be more (trap) shy or are more concealed, or were retracted prior to breeding periods [74]. Moreover, in our study, there was no correlation between Leptospira prevalence and the gender of the animals for five of the six rodent species, which was also demonstrated in other earlier studies [75,76]. Some studies have shown that in Norway rats (R. rattus), both in South-East Asia and in the United States of America, male specimens have a higher infection rate than females [77][78][79]. However, in the current study, it was found that female R. rattus were significantly more prone to a Leptospira infection than males. Apparently, in our situation, the habitat's use of rats creates a mechanism where both sexes were more or less equally exposed to infection.
When looking at the results from the rodent population from our study sites, it showed R. exulans (n = 23) were present in lower numbers compared to both B. bengalensis (n = 140) and R. rattus (n = 191), while the population of R. exulans consisted mainly of males. This raises the question of whether R. exulans competes with one of the other rodent species present. It is not unusual that species compete with each other. An early report from Bombay and Calcutta (1966) showed that B. bengalensis increased in population size enormously and displaced the R. rattus in urban areas [80]. One of the underlying mechanisms was the high reproductive capacity of B. bengalensis [81][82][83], and the aggressive behaviour that dominant males exhibit [82,83]. Other research from India reported B. bengalensis to be the most aggressive field rodent [84] and that the females confine most of their time to hoarding food and are less active than males [83]. Furthermore, although bandicoots are generally nocturnal, the B. bengalensis is known to also become active during daytime when conditions are undisturbed [83]. This could be advantageous in comparison to other rodent species. Unfortunately, almost no scientific studies are published on the competition of B. bengalensis with other rodent species. We found only one report from India that claimed that female R. rattus are submissive to B. bengalensis [85]. Sridhara et al. [85] postulate that there is less aggression in closely related rodent species (e.g., within species from the Rattus genus) compared to the violent interaction of rodents more distantly related from each other (e.g., Rattus genus vs. Bandicota genus). Unfortunately, rodents of the Rattus exulans species were not researched. This finding might explain the higher number of Bandicota specimens and the higher Leptospira infection rate compared to Rattus specimens trapped in our study. R. rattus is amongst the most omnipresent rodents in the world, and has a strong potential to displace other (native) rodent species [86,87]. In Australian ecosystems, it was shown that the invasive R. rattus was dominant over the native R. fuscipes [88,89]. On Madagascar, R. rattus competes for resources with the native rodents and replaces native rodent species [90]. Because B. bengalensis and R. rattus are larger than R. exulans [91], we expect B. bengalensis and R. rattus to be superior to R. exulans. In New Caledonia, Perez et al. [92] most frequently trapped R. Rattus specimens (>60%, n = 140) and very rarely trapped R. exulans rodents (<5%, n = 11), which supports our hypothesis that R. rattus is dominant over R. exulans.
In Bangladesh, Leptospira prevalence (in %) in rodents was significantly higher in the dry season than in the rainy season. These findings are not in line with the findings from Malaysia and also from Cambodia, where rodents showed a lower infection rate in the dry (6.3%) than in the wet (26.7%) season [67,68]. In a study on febrile patients (n = 1297) from Dhaka, a peak in the occurrence of Leptospira in patients was found in October and November, shortly after the monsoon season in Bangladesh [23]. These findings confirm the relation between floods or excessive rainfall and Leptospirosis outbreaks [21,32,70]. Besides this seasonal influence on human infection, the risk of Leptospira infection also depends on the geographic location, as well as on other risk factors, such as the risk of flooding, contaminated surface waters, and proximity to rubbish dumps (attractive for rodents) [93][94][95]. Easterbrook et al. [77] stated that seasonal fluctuations in Leptospira infections in rodents do not occur due to the fact that once infected, the antibodies remain in the animal and the animal will test positive. This can explain our results and findings from other countries in South-East Asia, which show that rodent species living in households have a stable infection level, regardless of the geography and season [67,78].
Our study has some limitations. The extracted DNA from the kidneys was diluted before being added to the PCR mix in order to obtain better results by reducing potential inhibitors present in the samples. Inhibitors reduce the activity of the DNA polymerase enzyme and as a result a false negative result will be observed. As pathogenic Leptospira colonize the kidney of the rodent, using this tissue as a source for leptospires DNA is critical; however, very high concentrations of host DNA are present. The abundant DNA of the host in the PCR reaction could lead to false-positive results as SYBR green dye is able to bind to any double-stranded DNA present in a very high concentration, thus rodent DNA can also be bound. By sequencing all samples, it was possible to filter out all false positive qPCR results. Furthermore, the effect of potential residual inhibitors in the samples could possibly result in false negative qPCR results. Because of this, we consider the 13.1% prevalence to be a prudent estimate of the actual prevalence in the rodent population in the study areas in Comilla. However, the asset of this study is that molecular diagnostics were used instead, which gives an indication of the carriership of the animal. Serologic methods used by most of the other studies conducted on Leptospira in South-East Asia indicate that the animal did have contact with leptospires, but it remains unclear whether the animal is still a possible carrier and therefore a potential reservoir [27,60].
We found L. interrogans as well as L. borgpetersenii DNA in the samples from all five rodent species that tested positive (B. bengalensis, B. indica, M. musculus, R. rattus, R. exulans). Other studies confirm the relation between R. exulans and L. interrogans [78,92], and in Thailand, R. exulans was also found to be infected with both L. interrogans and L. borgpetersenii [60,96]. These findings are in line with the fact that Leptospira species, borgpetersenii and interrogans, contribute a great deal to human disease in Asia [49,61,78,79]. Also, in Europe, L. borgpetersenii and L. interrogans are the most observed Leptospira genomospecies present in rodents; however, in Europe, a third genomospecies is also commonly found in rodents: L. kirschneri [97,98]. It is interesting to note that we also found one rodent sample infected with L. kirschneri from a R. exulans sample. Our findings are the first to confirm the presence of L. kirschneri in the rodent species, R. exulans, in Bangladesh. Mus terricolor 0% (n = 5) x Bangladesh † * All isolates were tested with Microscopic Agglutination Test (MAT) assay on a reference panel of 21 well-known serovars, from the L. interrogans and L. borgpetersenii group, but also on 1 serovar from L. noguchii, L. biflexa, and L. satarosai groups-unfortunately, no specification on which serogroups tested positive is mentioned. † This study.
Bandicota is an Asian genus of rodents, consisting of three species: B. bengalensis, B. indica, and B. savilei [103][104][105][106][107]. The few reports on Leptospira prevalence in the Bandicota genus indicate that all three species are potential carriers of the same leptospira species (L. interrogans, L. borgpetersenii, L. weilli, L. inadai). Our study found that the probability of an infection with Leptospira was significantly higher for rodents of the Bandicota genus. Thus, Bandicota rats could be an important host in the epidemiology of leptospirosis in Comilla. Previous studies on urban rodents identify R. rattus to be the main reservoir host for human pathogenic Leptospira [94,108,109]. However, due to the limited information available, it is not possible to link the strain or serovar infection to a specific host species. This is unfortunate, as such information could give insight into a possible co-evolution of serovars with specific rodent species. Other studies from countries with a similar climate and cropping season as Bangladesh have used mostly serological and culturing methods, and unfortunately no specification is made into specific serogroups for each rodent species. Therefore, the only comparison possible is to see which strains are found and if this correlates with the findings in Bangladesh (Table 5).
Cosson et al. [78] postulate that Leptospira species show an ecological niche; they found L. borgpetersenii to be more abundant in rodents from dry habitats (non-floodable lands) than L. interrogans, which implies that the infection of rodents can be linked to ecology. The B. indica (n = 172) is more common in the field (74%) than in or around houses [110,111], and in Vietnam, the B. bengalensis was found only in grass habitats (1969). The fact that the Bandicota species from Bangladesh were found to be infected with L. interrogans (Table 5) is thus in line with expectations.
Our study provides new data on rodent species as carriers of pathogenic Leptospira in South-East Asia. From our results and the literature research (Table 5), we can state that L. interrogans and L. borgpetersenii are the most common species found in rodents in South-East Asia. However, to find out whether specific strains/serovars adapt to specific reservoir hosts in specific habitats, more in-depth research with different diagnostics needs to be conducted. Although our results confirmed the importance of Bandicota spp. and Rattus spp. as hosts of leptospires for human health and our findings indicate that human exposure to pathogenic Leptospira may be considerable in Comilla, the impact of leptospires on human health continues to be under recognised. In many Asian human populations, including populations in Bangladesh, the burden of undifferentiated feverish illness is substantial [26]. One way to minimise this problem of recognising the disease is to conduct broader diagnostic tests to determine the cause of these illnesses and to inform people on the preventive measures they can take to prevent leptospirosis. Our findings highlight the complex multi-host epidemiology of leptospirosis and the importance of considering the role of rodents, and other animal hosts in the maintenance and transmission of infection when evaluating human risks. One of the key actions to minimise the public health impacts of leptospirosis in Bangladesh is to improve rodent management. A key question is the percentage to which the rodent population should be reduced to and which species should be diminished to prevent infection. In any case, preventive measures should be taken for rodent control, such as preventing rodents from accessing domestic areas and food storage to prevent pathogen transmission to humans. Rodents could be denied access to food and drinking water by taking rodent-proofing measures to existing buildings or by constructing rodent-proof warehouses. Furthermore, rodents could be discouraged from visiting domestic areas by keeping the environments clean, removing potential nesting sites, and by installing adequate sanitation and waste disposal.
With more knowledge about the rodent species present, leptospirosis, and its consequences, local people can be informed about the need for better rodent management practices. This could lead to a reduction of the impact of rodent-borne zoonotic disease in Bangladesh.

Conclusions
The objective of this study was to assess the presence and infection rate of pathogenic Leptospira in commensal rodent species in rural Bangladesh. In order to do so, 465 rodents were collected from a total of six different species, in descending order of appearance: Bandicota bengalensis, Rattus rattus, Mus musculus, Rattus exulans, Bandicota indica, and Mus terricolor.
Pathogenic Leptospira was found in 13.1% of all rodents, and three Leptospira species were identified: Leptospira interrogans, Leptospira borgpetersenii, and Leptospira kirschneri. Significant differences were found for the infection probability between species: B. indica showed significantly higher infection rates with Leptospira than all other species, and B. bengalensis and R. exulans both showed significantly higher infection rates than M. musculus and R. rattus.
Rodents trapped carrying Leptospira bacteria in their kidneys (13.1%) are potentially capable of shedding leptospires in and around food storage. These findings indicate that the risk of human exposure to pathogenic Leptospira is likely to be substantial for local people, since people can acquire leptospirosis via direct or indirect contact with the urine of an infected host, which in this case can also lead to a risk for the consumers of rice. We can conclude that our findings highlight the complex multi-host epidemiology of leptospirosis and the importance of considering the role of rodents, and other animal hosts, in the maintenance and transmission of infection when evaluating human risk. One of the key actions to minimise the public health impacts of leptospirosis in Bangladesh is to improve rodent management.

Conflicts of Interest:
The authors declare no conflict of interest.