Seroprevalence and hematological abnormalities associated with Ehrlichia canis in dogs referred to a veterinary teaching hospital in..

Ticks are significant parasites of dogs in the tropics, where tick-borne pathogens are highly prevalent, especially in areas where tick control measures are frequently neglected. This study investigated the seroprevalence and hematological abnormalities associated with Ehrlichia canis in dogs referred to a veterinary teaching hospital in Central-western Brazil. Out of 264 dogs tested for anti-Ehrlichia canis antibodies by an indirect immunofluorescence assay (IFA), 59.1% (156/264) were positive. Seropositivity was significantly associated to anemia and thrombocytopenia, alone or in combination, and to leukopenia. Conversely, there were no differences in terms of seroprevalence according to sex, breed and age. This study demonstrated that dogs referred to a veterinary teaching hospital in Central-western Brazil are highly exposed to E. canis and that seropositive dogs are more likely to present hematological abnormalities, particularly anemia, thrombocytopenia and leukopenia. To our knowledge, this is the first study on detection of anti-E. canis antibodies by means of IFA among dogs in the state of Goiás. These findings highlighted the need for increasing awareness among dog owners regarding tick control measures in Central-western Brazil, ultimately to reduce the risk of exposure to E. canis and other tick-borne pathogens.


INTRODUCTION
Ticks are significant parasites of dogs in the tropics, where tick-borne pathogens are highly prevalent, especially in regions where tick control measures are frequently neglected (MAGGI & KRÄMER, 2019;DANTAS-TORRES et al., 2020). For instance, Ehrlichia canis is a tick-borne pathogen that infects predominantly cells of the mononuclear phagocyte system of canids (DUMLER et al., 2001). Paula et al. This pathogen is primarily transmitted by brown dog ticks (Rhipicephalus sanguineus sensu lato) and is the cause of canine monocytic ehrlichiosis (CME), a disease that affects dogs worldwide, but especially in tropical and subtropical regions (SAINZ et al., 2015). Indeed, CME is a multisystemic disease with variable clinical manifestations, ranging from subclinical to life-threatening disease (HARRUS & WANER, 2011;SAINZ et al., 2015;MAGGI & KRÄMER, 2019). Its severity depends on various factors, including the pathogen strain, immune response of the dog and presence of concomitant infections (SAINZ et al., 2015). Sick dogs may display unspecific clinical signs and hematological abnormalities, such as fever, weakness, lethargy, anorexia, lymphadenomegaly, splenomegaly, hepatomegaly, weight loss, pale mucous membranes, epistaxis, petechiae, ecchymoses, anemia and thrombocytopenia (SAINZ et al., 2015).
While E. canis presents a worldwide distribution, it is more prevalent in tropical and subtropical regions as compared with temperate regions. This is evident in Latin America, where several hotspots of E. canis infection have been detected (MAGGI & KRÄMER, 2019;BADER et al. 2020;DANTAS-TORRES et al., 2020). For instance, low seroprevalences (<10%) have been reported in countries located in the extreme south of Latin America, such as Argentina, Uruguay and Chile (MAGGI & KRÄMER, 2019). Conversely, moderate to high seroprevalences have been detected in tropical regions of Latin America (MAGGI & KRÄMER, 2019). This difference in terms of seroprevalence of E. canis is also evident in Brazil, a country with continental dimensions. In this country, seroprevalences ranging from 0.7% to 92.3% have been reported, with the lowest seroprevalences detected in the southern region and the highest from the southeastern region towards the north of the country (VIEIRA et al., 2011). The lower prevalence of CME in the southern Brazil has been be related to the geographical distribution of different lineages of R. sanguineus s.l. ticks (MORAES-FILHO et al., 2011). Nonetheless, differences in terms of prevalence may also be related to other factors, including dog owners' attitude towards the control of ticks. Indeed; although, there is a vast repertoire of veterinary products against ticks on dogs in Brazil, the use of such products as preventive tools varies widely from region to region (FIGUEREDO et al., 2017), thus facilitating the transmission of tick-borne pathogens such as E. canis (DANTAS-TORRES & OTRANTO, 2014).

Most
studies investigating the seroprevalence of E. canis in dogs in Brazil have been conducted in the southeastern region of the country (VIEIRA et al., 2011). For instance, limited information is available for the Central-west region of Brazil. In the same way, information on the seroprevalence of E. canis in hospital population in Brazil is fragmentary. Considering that the tropical lineage of R. sanguineus s.l. is present in Centralwestern Brazil, we hypothesized that even dogs with access to healthcare services are highly exposed to E. canis in this region. To investigate this hypothesis, we evaluated the seroprevalence and hematological abnormalities associated with E. canis in dogs referred to a veterinary teaching hospital in Centralwestern Brazil.

Hematological evaluation
Blood samples (3 mL) from dogs were withdrawn from the cephalic or jugular veins into Vacutainer ® tubes containing EDTA (Becton Dickinson, New Jersey, USA) and sent to the Laboratory of Clinical Pathology of the Veterinary Hospital of the School of Veterinary Medicine and Animal Science of the Federal University of Goiás. Red blood cell, leukocyte and platelet counts were performed using a hematological analyzer (Celltac Alpha/MEK-6550 ® , Nihon Kohden). Blood smear examination and hematocrit was determined by the microhematocrit method.

Indirect immunofluorescence assay (IFA)
Serum samples were tested by an IFA using crude antigens of E. canis (strain Cuiabá #1), prepared in a cell culture, as previously described (AGUIAR et al., 2007(AGUIAR et al., , 2008. Reactions were performed with FITC-labeled anti-dog IgG (Sigma-Aldrich) previously titrated to the best working dilution (1:1000), as described elsewhere (RISTIC et al., 1972;AGUIAR et al., 2007). On each slide, a nonreactive and a reactive serum sample (endpoint titer of 1,280) were included as negative and positive controls, respectively. Control serum samples derived from studies by AGUIAR et al. (2007) and KRAWCZAK et al. (2012). A serum sample was considered to contain antibodies to E. canis if reactive at the 1:40 dilution. Samples reacting at the screening dilution (1:40) were then titrated using serial fourfold dilutions to determine endpoint titers. For comparison purposes, antibody titers were classified as low-medium titer (<10,240) or high titer (10,240).

Statistical analyzis
The minimal sample size (n = 264) was calculated using the software WinEpi (http://www. winepi.net/uk/index.htm), considering a confidence level of 90%, unknown population size, expected proportion of 42% (based on DUARTE et al., 2013) and margin of error of 5%.
Hematological parameters were compared between three groups of dogs [seronegative, seropositive with low-medium titers (<10,240) and seropositive with high titer (10,240)], using chisquare test or Fisher's exact test, with Bonferroni correction. The association of serological status (negative and positive) with sex, age and breed were also performed with chi-square or Fisher's exact tests. The 95% confidence interval (95% CI) of seroprevalences was also calculated. All statistical analyzes were performed using R program (version 3.6.1) and a P <0.05 was considered as statistically significant.
By comparing hematological abnormalities between seronegative dogs and seropositive dogs with moderate and high antibody titers, anemia and leukopenia were significantly more frequent in dogs with high antibody titers as compared with seronegative dogs (X 2 = 11.05, df = 1, p<0.001 and X 2 = 4.19, df = 1, p-value = 0.0404, respectively). No statistically significant differences were observed for other hematological abnormalities in relation to antibody titers (Table 2).

DISCUSSION
In the present study, we reported a high seroprevalence of E. canis in dogs referred to a veterinary teaching hospital in central-western Brazil, confirming our initial hypothesis. The seroprevalence of E. canis in dogs from Brazil varies widely, according to studies conducted in different geographical regions of the country (VIEIRA et al., 2011;KRAWCZAK et al. 2012;GOTTLIEB et al. 2016;FIGUEREDO et al., 2017;DANTAS-TORRES et al., 2018. However, some of these  et al., 2004;TRAPP et al., 2006;CARLOS et al., 2007;DINIZ et al., 2007;CARVALHO et al., 2008;NAKAGHI et al., 2008;ORIÁ et al., 2008;BORIN et al., 2009;SANTOS et al., 2009). In hospitalbased studies, the seroprevalence reached as high as 70.1% (138/198) in sick dogs referred to a veterinary teaching hospital in Botucatu, São Paulo, southeast Brazil (DINIZ et al., 2007). While many studies have been conducted in other Brazilian regions, information on the seroprevalence of E. canis in dogs from Centralwestern Brazil is scarce. Our study is the third largest hospital-based study conducted in Brazil, in terms of sample size. In a study conducted in a veterinary teaching hospital in Londrina, southern Brazil, authors tested 381 dogs using a point-of-care ELISA and 87 (23%) of them were positive (TRAPP et al., 2006). In another study carried out in a veterinary teaching hospital Santa Maria, southern Brazil, 316 dogs were evaluated and 14 (4.4%) were seropositive (KRAWCZAK et al., 2012). Our results showed a much higher seroprevalence (i.e., 59.1%), suggesting that even privately-owned dogs that have access to healthcare services in centralwestern Brazil may be highly exposed to E. canis. This indicate that tick control measures to reduce the risk of tick-borne pathogen transmission are not frequently applied or, if applied, not being effective in the studied population.
The high seroprevalence of E. canis in the Central-western Brazil may also be partly explained by the presence of the tropical lineage of R. sanguineus s.l. in this region of the country, rather than R. sanguineus sensu stricto (=temperate lineage) (NAVA et al., 2018). In fact, an experimental study demonstrated that R. sanguineus s.s. ticks from southeastern Brazil, which belong to the tropical lineage, were competent vectors of E. canis (MORAES-FILHO et al., 2015). yet, R. sanguineus s.s. ticks from southern Brazil, Argentina and Uruguay were not (MORAES-FILHO et al., 2015). While further studies with other tick populations and  (DUARTE et al., 2013). These dogs presented variable clinical signs, including fever, pale mucous membranes, inappetence, prostration, occasionally associated to blood coating abnormalities and tick infestation. These findings along with data presented herein indicate that dogs exposed to E. canis are more likely to present clinical signs and/or clinicopathological abnormalities.
In this regard, anemia, leukopenia and thrombocytopenia were significantly more frequent in seropositive dogs. Thrombocytopenia in the acute phase may be accompanied by anemia and reduced leukopenia, but in the chronic phase they usually present more marked values (HARRUS & WANER, 2011), leading to pancytopenia, usually due to bone marrow hypoplasia, a severe sign observed in the chronic phase of CME (HARRUS & WANER, 2011). In our study, pancytopenia was observed in 3.8% (6/156) of the dogs, but a much higher frequency has been reported elsewhere in a small group of sick dogs (NAKAGHI et al., 2008). These hematological abnormalities have also been reported in previous studies carried out in Brazil (OLIVEIRA et al. 2000;BULLA et al., 2004;MACIEIRA et al., 2005;NAKAGHI et al., 2008;ORIÁ et al 2008;SANTOS et al., 2009;UENO et al. 2009;FONSECA et al., 2017). For instance, the prevalence of E. canis in dogs with thrombocytopenia ranged from 26.8% (30/112) to 45.0% (66/146) (BULLA et al., 2004;MACIEIRA et al., 2005;SANTOS et al., 2009). Nonetheless, while hematological abnormalities such as thrombocytopenia are commonly seen in CME (SAINZ et al., 2015), these are unspecific and may be present in other tick-borne diseases (DE TOMMASI et al., 2013). Thus, while platelet count is a good indicator of CME, thrombocytopenia is not a specific finding and should not be used alone to establish a diagnosis (HARRUS &WANER, 2011).
It is worth mentioning that the IFA is considered the 'gold standard' test for detecting anti-E. canis IgG antibodies, indicating exposure to E. canis. However, a definitive diagnosis of E. canis infection requires molecular techniques, such as PCR (HARRUS & WANER, 2011), which is a limitation of the current study.

CONCLUSION
This study demonstrated that dogs referred to a veterinary teaching hospital in Central-western Brazil are highly exposed to E. canis and that seropositive dogs are more likely to present hematological abnormalities, particularly anemia, thrombocytopenia and leukopenia. To our knowledge, this is the first study on detection of anti-E. canis antibodies by means of IFA among dogs in the state of Goiás. Our results highlight the need for increasing awareness among dog owners regarding tick control measures Central-western Brazil, ultimately to reduce the risk of exposure to E. canis and other tickborne pathogens.

DECLARATION OF CONFLICT OF INTEREST
Authors declare they have no competing interest.

BIOETHICS AND BIOSSECURITY COMMITTEE APPROVAL
Collection of serum samples from dogs was previously authorized by the Institutional Animal Care and Use Committee (IACUC) of the Universidade Federal de Goiás (protocol 102/17).