Prevalence of antibodies against Ehrlichia spp. and Orientia tsutsugamushi in small mammals around harbors in Taiwan

Tick-borne ehrlichiosis and mite-borne scrub typhus represent important emerging zoonotic rickettsial diseases. Although scrub typhus has been recognized by the Taiwanese public health system, information on ehrlichial infections is scarce in Taiwan. In this study, the risk of spread of ectoparasites on rodents through aerial and marine transportation was assessed in international and domestic harbors. Here, we report the first systematic surveillance of seroprevalence against Ehrlichia spp. in small mammals on the main island of Taiwan. In total, 1648 small mammals were trapped from 8 international ports, 18 domestic fishing harbors, and 7 local public health centers around Taiwan from November 2004 to December 2008. Sera were analyzed using indirect immunofluorescence assays to detect IgG antibodies against Ehrlichia chaffeensis and Orientia tsutsugamushi. A serum titer of ≧1:80 was considered positive. Antibodies against Ehrlichia spp. and O. tsutsugamushi were detected in 3.28 % and 4.92 % of small mammals active around harbors, respectively. The seropositive rate against Ehrlichia was higher in northern Taiwan from 2005 to 2008. However, O. tsutsugamushi infections increased in southern Taiwan during this period. The serological evidence of ehrlichial and O. tsutsugamushi infections in all international ports were included in the study. No significant differences were found among the seropositive rates of Ehrlichia spp. and O. tsutsugamushi in small mammals trapped between international and local harbors. The overall prevalence of Ehrlichia spp. and O. tsutsugamushi infections in small mammals active around harbors was 3.28 % and 4.92 %, respectively. The results provided serological evidence supporting the potential risks of transporting pathogens through air and maritime traffic. This study highlights serious issues of the emergence and spread of rickettsial diseases in Taiwan. The incidence of human ehrlichiosis requires further investigation.


Background
Rickettsial infections are caused by several genera of obligate intracellular Gram-negative bacteria with life cycles that involve vertebrate and invertebrate hosts. Different types of rickettsiosis are widely distributed worldwide, but the vector species and associated human illnesses are distinct depending on geographical locations. Rickettsioses are classically divided into the scrub typhus group, which is transmitted by mites, the spotted fever group, which is transmitted by ticks or mites, and the typhus group, which is primarily transmitted by lice or fleas. In Taiwan, notifiable rickettsial diseases include scrub typhus, epidemic typhus, and murine typhus. No epidemic typhus has been identified in Taiwan since World War II. Murine typhus, caused by Rickettsia typhi, is endemic to Taiwan, with approximately 20-60 cases each year. Scrub typhus group rickettsiosis is the most prevalent infection and causes 300-500 confirmed cases annually (http://www.cdc.gov.tw/). Scrub typhus is endemic to Asia, northern Australia, and the western Pacific regions [1,2]. Diverse strains of Orientia tsutsugamushi, the etiological agent of scrub typhus, have been isolated from humans and its vectors, larvalstage trombiculid mites (chiggers) [3,4]. Epidemiological surveys have suggested considerably high prevalence of O. tsutsugamushi infections in wild rodents, ranging from 69.1 % to over 90 %, on the offshore and main islands of Taiwan [5,6]. Human infections of scrub typhus are routinely diagnosed using molecular, serological and pathogen isolation with shell-vial methods performed at the Centers for Disease Control in Taiwan (Taiwan CDC).
Ehrlichia spp., another genus of bacteria responsible for rickettsial diseases in humans and wild or domestic animals, is less recognized in Taiwan. The first human case of ehrlichiosis was documented in the United States of America in 1986. Ehrlichiosis has since been reported worldwide [7]. The common pathogens that cause human infections are Ehrlichia chaffeensis and E. ewingii. The symptoms are moderate to severe and occasionally are fatal [8,9]. Enzootic cycles of Ehrlichia spp. transmitted between ticks and wild animals are widespread and have been reported in Korea, Japan, and southern China, which are in geographic proximity to Taiwan [10,11]. Previous studies have identified E. chaffeensis in Rhipicephalus haemaphysaloides and Ixodes granulatus ticks and rodents in Kinmen, one of Taiwan's offshore islands [12]. Here, we report ehrlichial infections in small mammal reservoirs on the main island. The potential impact of human ehrlichiosis in Taiwan has yet to be determined.
The impact of travel on the spread of infectious diseases is of great concern. Human migration has provided a shortcut for disease dissemination. When people travel, they often carry luggage, food and goods, and they are also accompanied by, as their biological microenvironments, microbes, animals, and parasites. To examine the risks arising from transportation of pathogenic Orientia and Ehrlichia through human activities, sera were collected from small mammals captured around harbors, and the prevalence of seropositivity against O. tsutsugamushi and Ehrlichia spp. was determined by immunofluorescence assay (IFA). To the best of our knowledge, this is the first systematic surveillance of ehrlichial and Orientia tsutsugamushi infections in their reservoirs on the main island of Taiwan.

Study area and small mammal capture
This study was conducted as part of the surveillance of the Hanta virus in rodents in Taiwan [13]. Small mammals were captured from 33 study sites, including 8 international harbors, 18 domestic fishing harbors, and 7 local public health centers around Taiwan from November 2004 to December 2008 ( Fig. 1 and Table 1). The methods for small mammal capture have been described in a previous report [14]. Briefly, approximately 20-30 traps were set up each month in areas with suspected rodent activities. Trapping at international ports took place for three days each month. Sweet potatoes, peanuts and/or sausages were used as bait to attract diverse species of small mammals. Blood was collected by cardiac puncture after the captured animals were anesthetized with Zoletil 50 (Fa. Virbac, Carros, France) and was left to stand for 1 h at room temperature. Sera were separated by centrifugation, aliquoted, and kept frozen at −20°C for later analysis.

Ethical considerations
The study was conducted following the regulations of Article 17 of Taiwan Wildlife Conservation Act and with the permission of Taiwan Centers for Disease Control. None of the trapped mammals belonged to Protected Species. All trapping and sampling procedures were performed according to U.S. CDC guidelines for sampling small mammals for virologic testing and met International Health Regulations (IHR2005). The use of animal sera for antibody detection was approved by National Taiwan

Serological analysis
Rodent serum samples were evaluated for reactivity to E. chaffeensis (prepared by Prof. Yu's laboratory, the University of Texas Medical Branch) and O. tsutsugamushi (Karp and Kato strains; prepared by Taiwan CDC) using IFA.
For detecting antibodies specific to Ehrlichia spp., antigen slides were prepared with E. chaffeensis Arkansasinfected DH82 cells. Briefly, E. chaffeensis-infected DH82 cells from a 150-cm 2 flask were collected when 100 % of the cells were infected. The cells were centrifuged at 200 × g for 10 min. Pellets were resuspended in 10 ml of phosphate-buffered saline (PBS) with 0.1 % bovine albumin. Sodium azide was added to a final concentration of 0.01 % and treated at 4°C overnight to inactivate Ehrlichia. Ten microliters of antigen was applied to each well of 12-well slides. The slides were air dried and fixed in acetone for 5 min before immunostaining [15].
For the detection of scrub typhus antibodies, O. tsutsugamushi (Karp strain)-infected L929 cells on chamber slides were fixed and permeabilized with ice-cold acetone/ methanol (1:1) at −20°C for 10 min. The slides were dried and blocked with PBS containing 1 % goat serum before further immunostaining [16].
Serum samples were centrifuged at 1730 × g at 4°C for 10 min and diluted to 1:80 using a sterile field diluent (PBS containing 1.5 % bovine albumin, 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 250 μg/mL of amphotericin B). A 20 μL volume of rodent serum was pipetted into each well, and the slides were incubated in a moist chamber at 37°C for 30 min. The slides were rinsed and washed in PBS for 10 min and then rinsed again with distilled water. The slides were air-dried, and the wells were incubated with 20 μL of fluoresce in isothiocyanate (FITC)-conjugated, affinity-purified, Fc fragment-specific, goat anti-rat IgG (Jackson Immuno-Research Laboratories, West Grove, PA) diluted to 1:100 in PBS containing Evans blue counterstain (Sigma Chemical Company, St. Louis, MO). After incubation in a moist chamber at 37°C for 30 min, the slides were rinsed and washed with PBS twice. Coverslips were mounted using mounting medium (PBS:glycerol = 3:7 v/v; pH 9.0), and the slides were examined at a magnification of 100× or 400× with a Zeiss fluorescent microscope. Sera with typical patterns of bright green fluorescence at titers of 1:80 for IgG were deemed positive [17,18]. Positive and negative control sera were included on each slide. All slides were read independently by two individuals who were blinded to specimen identity. Disagreements were resolved by consensus after re-examining the slides.

Statistical analysis
All statistical analyses were performed using SAS v9.1.3 software (SAS Institute, Cary, NC). Fisher's exact test was used for comparisons of seroprevalence in rodents between different sampling sites. A twotailed p-value of 0.05 or less was considered to be statistically significant. The trend across time was examined with the Cochran-Mantel-Haenszel Chi-squared test.

Field collection of small mammals
A total of 1648 small mammals were caught from 33 study sites in Taiwan from November 2004 to December 2008, including 762, 252, 575 and 59 animals in northern, central, southern, and eastern Taiwan, respectively (Table 1). Figure 1 summarizes the numbers of trapped small mammals at the different sites. The seroprevalence of ehrlichial and O. tsutsugamushi infections in 4 districts of Taiwan was determined. Six species of rodents, Rattus norvegicus (brown rat), Bandicota indica (Taiwan bandicoot rat), R. losea (brown country rat), Mus musculus (house mouse), R. tanezumi (oriental house rat), and Apodemus agrarius (striped field mouse), and a species of insectivore, Suncus murinus (Asian house shrew), (EHR%, STG%)     Fig. 1).
In this study, 1202 and 446 small mammals were trapped at the international and domestic harbors (including local public health centers), respectively ( Table 3). The prevalence of antibodies specific to O. tsutsugamushi was slightly higher than the prevalence of antibodies specific to Ehrlichia spp. Moreover, a higher but not significantly greater number of animals expressed antibodies against the rickettsial diseases around the international ports. Most importantly, serological evidence of infections was detected in all international harbors (Table 1).

Seroprevalence in small mammals captured at different time points
The seroprevalence of O. tsutsugamushi infections in small mammals increased in the harbors of southern Taiwan during the year of 2004 and 2008 (p < 0.01) ( Table 4). Antibodies against Ehrlichia spp. and O. tsutsugamushi were detected year-round (Fig. 2). In trapped small mammals, the seropositive rates were highest in November and December for Ehrlichia spp., whereas the positive rates were higher in January and October for O. tsutsugamushi.

Co-infection with different pathogens in captured animals
Of the 1648 small mammals tested in the study, 5 (0.30 %) were seropositive against both Ehrlichia spp. and O. tsutsugamushi, including one R. norvegicus in Keelung Harbor, 2 R. norvegicus in Mailiao Harbor, one M.   [7,[19][20][21][22][23][24][25]. These large wild animals are less common in Taiwan. Conversely, shrews and rodents are more abundant and often serve as reservoirs for many communicable diseases transmitted by their ectoparasites. Previous studies have discovered E. chaffeensis DNA in 16 of 108 (14.8 %) spleen or liver samples from shrews and rodents on one of Taiwan's offshore islands, Kinmen, by polymerase chain reaction (PCR) and sequencing [14].
Our results further showed that 3.28 % (54/1648) of captured animals were seropositive against ehrlichial infections, and more infected animals were found in the north and central parts of the main island of Taiwan. During this study, ectoparasites collected from murine-like animals included two fleas: X. cheopis and Nosopsyllus nicanus, three ticks: Ixodes granulatus, Haemaphysalis bandicota, and Rhipicephalus haemaphysaloides, three mites: Laelaps nuttalli, L. sedlaceki, and L. echidninus, and lice. Among these ectoparasites, 45.04 % were X. cheopis, and 29.25 % were H. bandicota. In terms of the ectoparasites, X. cheopis was the most common species. X. cheopis carries R. typhi and serves as a major vector to cause human murine typhus. Haemaphysalis bandicota, the second common ectoparasite, involves in disease transmission among rats. However, H. bandicota has never been found to be associated with any vectorborne communicable diseases in humans [26]. Information on trombiculid mites was limited in this study, but these mites may be associated with small mammal species. It is worth noting that serological cross-activities have been addressed among Ehrlichia spp., such as E. chaffeensis, E. ewingii, E. canis and E. ruminantium [27][28][29][30]. The enzootic cycle of E. canis has been documented in dogs in Taiwan [31,32]. The positive reactions observed by IFA in the current study therefore may have been confounded by infections with other Ehrlichia species. In North America, the seroprevalence of E. canis, E. chaffeensis and E. ewingii in dogs was 0.8 %, 2.8 %, and 5.1 %, respectively [33]. In China, the overall seroprevalence of E. chaffeensis was 9.8 % in rural residents and 2.4 % in urban residents [34]. Nevertheless, our results suggested that ehrlichial infections are prevalent on the island. The implications for human ehrlichiosis should be scrutinized and re-evaluated because it is not a commonly observed disease in Taiwan. Several studies have evaluated the prevalence of scrub typhus infections in animal reservoirs. A seropositive rate of over 90 % has been reported in rodents in Kinmen [5]. A prevalence of 70 % has been observed in small mammals captured in abandoned agricultural fields in central Hualien County in eastern Taiwan [35]. Our results showed that 4.92 % of the small mammals caught around the harbors expressed antibodies against O. tsutsugamushi. The prevalence was relatively low compared with those in previous reports and may reflect the different ecosystems of our sampling areas. Different rodent hosts affect the seropositive rates of rickettsial infections [5,12,13]. Our study focused on small mammals active around harbors, which have, to a certain degree, much more human activities than present in the study sites selected in other studies. Furthermore, the rodent species in these harbors differed from those on small islets or in traditional agricultural environments. Indeed, the dominant species among captured animals varies in the literature. R. norvegicus was the dominant rodent species in this study, whereas R. losea and Apodemus agrarius are abundant in Kinmen and Hualien, respectively [5,35].
The distributions of the rodent species were similar among harbors in Taiwan, with the exceptions of Suao Harbor and Taoyuan Airport. On the whole, R. norvegicus was the most frequent species followed by S. murinus and B. indica. However, the Taiwan bandicoot rat and brown country rat were the primary rodents captured in Taoyuan Airport. This may be related to the trap sites and the behaviors of B. indica, which are usually active in nonresidential areas. Haemaphysalis bandicota was also primarily found at Taoyuan International Airport [26]. For Su-ao Harbor, the only species captured were R. norvegicus and M. musculus (n = 1).
Together, the geographic distribution, activities of the vectors and reservoirs, and human activities determine the prevalence of rickettsial disease infections. In this study, B. indica appeared to be most influenced by E. chaffeensis (10.44 %), and R. tanezumi had the highest rate of seropositivity against O. tsutsugamushi (23.08 %). Although seropositive rates were highest in November and December for Ehrlichia spp., whereas the positive rates were higher in January and October for O. tsutsugamushi. The numbers of captured animals did not differ significantly between seasons, and the antibodies were detected year-round. In general, our results revealed that small mammals, which are common in residences, were infected by Ehrlichia and O. tsutsugamushi. The frequent association between these mammals and human activities may increase the risk of rickettsial diseases.
Our results indicated that the rates of seropositivity against O. tsutsugamushi were higher in the eastern and central parts of Taiwan, which correlated with the high incidence of human scrub typhus in these regions [3]. Multiple environmental factors and vector species have been shown to affect the prevalence of vector-borne infectious pathogens in animal hosts [36][37][38]. A recent article has also reported a positive correlation between the rate of seropositivity in rodents and the incidence of human scrub typhus; however, the major captured animal in that study, S. murinus, was not tested due to a lack of antiserum [38]. FITC-conjugated goat anti-rat antibodies were used to detect antibodies from captured small mammals in our study. Similar strategies have been applied in other experiments when no antiserum was commercially available [39,40]. Although we cannot guarantee that anti-rat secondary antibodies react with rodent and shrew sera equally, the shrew was a minor species among all trapped small mammals (17.23 %). Zero and 4 (1.41 %) shrews tested positive for antibodies against E. chaffeensis and O. tsutsugamushi, respectively.
Travel and trade have provided a source for disease dissemination throughout recorded history. Plague represents one of the most striking examples of this phenomenon. Human population movement has been crucial to the dispersal of Yersinia pestis and rodents and their fleas to new territories [41]. Moreover, international trade of pet reptiles and amphibians has been shown to have a potential to cause emerging infections of Rickettsia and Ehrlichia spp. in Japan [42]. In our study, animals were found to harbor antibodies against Ehrlichia and O. tsutsugamushi around ports and harbors. The prevalence of seropositivity was slightly higher in the samples collected near international harbors; however, no significant differences were observed. Nevertheless, serological evidence of infections was detected in all international harbors studied. These findings highlight the risk of spreading pathogenic Ehrlichia and Orientia associated with ectoparasites on rodents through aerial and marine transportation.

Conclusions
Our study provided serological evidence of Ehrlichia spp. infection in small mammals on the main island of Taiwan. The implications for human ehrlichiosis should be re-evaluated because it is not a commonly observed disease in Taiwan. Small mammals can serve as useful indicators and/or sentinels for epidemiological surveillance, disease control and prevention of vector-borne infectious diseases. The prevalence of infected animals captured around the harbors should raise awareness of the potential risks of disseminating pathogens by domestic and international aerial and marine transport.