Molecular detection of Ehrlichia canis in dogs from three districts in Punjab (Pakistan)

Abstract Canine monocytic ehrlichiosis is a tick‐borne disease caused by an intracellular alpha‐proteobacterium, Ehrlichia canis, which replicates within mononuclear cells in the host. This study was designed to use a polymerase chain reaction (PCR) protocol for the molecular detection of E. canis by the amplification of a portion of its 16S rRNA gene, as well as the effects of this alpha‐proteobacterium on the haematological parameters of the sampled dogs and the risk factors associated with E. canis infection. A total of 151 blood samples were collected from dogs of various breeds at three sampling sites (Lahore, Rawalpindi/Islamabad and Multan) in Punjab, Pakistan. Data regarding the epidemiological factors (including age, gender, breed, body temperature, deworming, vaccination, mucous membrane status, hydration status, the presence of haematuria and tick infestation) were collected through a questionnaire at the time of sample collection. A 400 bp DNA fragment of the 16S rRNA gene of E. canis was amplified from 42 dog blood samples (28% of the total), [Lahore (N = 24), Rawalpindi/Islamabad (N = 13) and Multan (N = 05)] through PCR. Data analysis revealed that the character of the animals (age, sex and breed) had no significant association (P > 0.05) with the presence of E. canis. Various haematological parameters were also compared, and the results revealed that all of the parameters remained unaffected, except significantly lower white blood cell counts (P = 0.004) in E. canis‐positive blood samples, as compared with the control group. We concluded that this is the first molecular confirmation of canine infection by E. canis using PCR. Moreover, no specific epidemiological parameter was found associated with the prevalence of E. canis in dogs.


Introduction
Dogs are among the most common pets, and as their population has increased tremendously, parasitic diseases are a major health concern (McBride et al. 1996). Among these parasitic and infectious diseases, canine monocytic ehrlichiosis is a tick-borne disease caused by an obligate intracellular alpha-proteobacterium, Ehrlichia canis (E. canis), which replicates within mononuclear cells in the host (Harrus & Waner 2011). Canine ehrlichiosis has been reported from subtropical and tropical areas of the world where it is one of the very common disease reported during veterinary practice in dogs (Rani et al. 2011). One of the major reasons for the common occurrence of E. canis is the tick vector, Rhipicephalus sanguineus sensu lato, which is common in rural and urban areas in tropical and subtropical areas of the world (Aguiar et al. 2007;Dantas-Torres 2008;Aktas 2014;Aktas et al. 2015). 1 These two authors have contributed equally to the manuscript. Three consecutive phases of canine ehrlichiosis have been reported in the literature; the subclinical, acute and chronic phase. Thrombocytopenia, variable leucopenia and anaemia are commonly found to be associated with ehrlichiosis (Greene 2013). The acute phase of this disease is characterized by symptoms of fever, depression, dyspnoea, anorexia, lymphadenopathy and slight weight loss (Iqbal et al. 1994). In the chronic disease, dogs have been reported suffering from haemorrhages, epistaxis, peripheral oedema, emaciation and hypotensive shock leading to death (Ristic & Woldehiwet 1993).
To our knowledge, no information is available in literature regarding the prevalence of E. canis in dogs in Pakistan, the present study aimed to perform a molecular detection of canine infection by E. canis in Pakistan. As this pathogen is found in the blood, we are also reporting the comparison of various haematological parameters between E. canis-positive and -negative animals.

Sample and data collection
Blood samples from 151 dogs were randomly collected from pet clinics in Lahore (N = 50); Rawalpindi/Islamabad (N = 50) and Multan (N = 51), districts of Punjab, during June-December 2014 All of the dogs included in the study were pets, except for stray dogs (N = 2), and informed consent was obtained from pet owners before enrolling them in the study. Sampled dogs included healthy animals as well as those having clinical symptoms including fever (diagnosed with thermometer) pale mucus membrane, haematuria and vomiting (physical examination). Blood samples were collected from the saphenous vein of the animals and immediately preserved in EDTA tubes. A questionnaire was filled at the sampling site in order to gather data (including age, sex, breed, body temperature, deworming, vaccination, mucous membrane status, hydration status, the presence of haematuria and tick infestation) and to report risk factor(s) associated with the prevalence of E. canis DNA in dogs, if any (de Castro et al. 2004;Shipov et al. 2008). The body of each dog, with special attention to the ears, was examined for the presence of ticks. If present, they were removed with forceps and placed in bottles with moistened cotton wool and were then transferred to the laboratory for identification using taxonomic keys (Aktas 2014).

DNA extraction
The inorganic method of DNA extraction was used, following Shahnawaz et al. (2011). The quality of the DNA extracted was assessed with optical density counts at 260/280 nm and submerged gel electrophoresis to determine its purity and integrity.

PCR amplification
A set of previously reported oligonucleotide primers, Fwd ECA: 5 0 -AAC ACA TGC AAG TCG AAC GGA-3 and Rev HE3: 5 0 -TAT AGG TAC CGT CAT TAT CTT CCC TAT-3, was used to amplify the 16S rRNA gene sequences of E. canis, as previously reported (Wen et al. 1997). E. canis-positive samples, kindly donated by Prof. Dr. Diego Fernando Eiras from National University of La Plata Facultad de Ciencias Veterinarias Argentina, and negative controls (polymerase chain reaction (PCR) mixture without DNA) were amplified during each PCR as positive and negative controls, respectively.

Haematological analysis
Various haematological parameters in blood samples from dogs, i.e. white blood cells, platelets, red blood cells, haemoglobin, mean corpuscular volume and packed cell volume, were determined in blood samples from E. canis-positive and -negative dogs by an automated haematology analyser (Abbott Cell-Dyn 3700), Illiois USA).

Statistical analysis
All of data are expressed as mean AE standard deviation. Minitab Statistical package (Pennsylvania, USA) was used for the statistical analysis of the results. Animals were divided into two age groups, animals up to 1 year (juvenile) and more than 1 year (mature). The association between the presence of E. canis and the various risk factors, e.g. the sex and age of the animal, was evaluated by contingency table analysis using the Fisher's exact test (for 2 x 2 tables). Two sample t-tests were calculated to compare the various studies haematological parameters between E. canis-positive and -negative blood samples.

Results
All dogs were examined for the presence of ixodid ticks. Of the 151 dogs, 11 (7.3%) were infested with adult and nymphal Rhipicephalus sanguineus s.l. Forty-eight ticks (29 nymphs and 19 adults) were removed from dogs.
PCR amplification revealed that 42 out of 151 (28%) blood samples were positive for E. canis, as they amplified a 400-bp amplicon using the set of primers targeting the 16S rRNA gene of E. canis. The prevalence of E. canis varied significantly (P < 0.001) between the three sampling sites. Maximum prevalence of E. canis was observed in the Lahore district (48%) followed by Rawalpindi/Islamabad (26%) and Multan (10%) ( Table 1).
The one-way ANOVA was calculated to report the prevalence of parasite in various dog breeds enrolled in present study and the results indicated that prevalence of E. canis was not restricted (P > 0.05) to a particular breed. The highest prevalence of E. canis was detected in stray dogs (100%; N = 2) while the lowest was observed in Labradors (15.38%; N = 26). E. canis was not detected in Shih Tzu, Spaniel, Pug and Rottweiler breeds during this study ( Table 2). Analysis of data, collected through a questionnaire at the sampling site, revealed that none of the studied epidemiological parameters were found associated (P > 0.05) with the prevalence of E. canis in dogs (Table 3).
Comparison of the haematological parameters between E. canis-positive (N = 42) and -negative blood (N = 109) samples from Lahore, Rawalpindi/ Islamabad and Multan revealed that most of the studied parameters varied non-significantly (P > 0.05) when compared between the two groups except white blood cell count that was significantly lower (P = 0.004) in E. canis positive than in blood samples where parasite was not detected (Table S1).

Discussion
Canine monocytic ehrlichiosis is a tick-borne disease of increasing importance in dogs, and has become an active area of research in recent years, although it remained unexplored in Pakistan. To our knowledge, this is the first report regarding the prevalence of E. canis DNA in dogs in Pakistan.
E. canis has been detected and reported in dogs from many parts of the world Rani et al. 2011;Sasaki et al. 2012;Ybañez et al. 2012;Nazari et al. 2013;Aktas et al. 2015;Inpankaew et al. 2016). An analysis of our results revealed a 28% overall prevalence of E. canis (N = 42) in the studied blood samples from the three districts, suggesting that E. canis is prevalent in Punjab province (Table 1) but the prevalence of E. canis varied significantly between three sampling areas indicating that geographical features and climatic conditions do affect the parasite prevalence (Rani et al. 2011;Mircean et al. 2012). In Turkey, a large study was conducted in the coastal provinces (Sakarya, Kocaeli, Mersin, Giresun and _ Izmir) as well as in land provinces (Elazig, Diyarbakır, Erzurum, Ankara, Nevs ßehir) in 2015 and it was observed through PCR and reverse line blotting (RLB) assays that 37/757 (4.9%) dogs were positive for E. canis (Aktas et al. 2015). In a similar study conducted in four different  (Rojas et al. 2014). An analysis of our results revealed a 28% overall prevalence of E. canis (N = 42) in the studied blood samples from the three districts, suggesting that E. canis is prevalent in Punjab province (Table 1). These sampling sites have different geographical and climatic conditions. Islamabad is located at 33.43°N 73.04°E at the northern edge of the Pothohar Plateau at the foot of the Margalla Hills. The city has a humid subtropical climate, with five seasons: winter (November-February), spring (March and April), summer (May and June), monsoon (July and August) and autumn (September and October). Lahore is located at 31°15 0 -31°45 0 N and 74°01 0 -74°39 0 E and has a semi-arid climate. The hottest month is June, with high temperatures routinely exceeding 40°C. Multan (30°11' 44" N, 71°28' 31" E) has an arid climate with hot summers and mild winters. The city experiences some of the most extreme weather in the country. Our results indicated that prevalence of E. canis varied significantly (P < 0.001) between three sampling site that can be attributed to different climatic conditions that effects the prevalence of vector ticks and hence parasite prevalence. These variation in the prevalence of E. canis among the various studies discussed here  Prevalence of the parasite is given in parenthesis. P-value represents the results of the Fisher's exact test calculated for each studied parameter. P > 0.05 = Non-significant.
could be due to many factors including the distribution and population density of the vector (Otranto et al. 2011), the sampling methodology and the characteristics of the targeted dog population (Gomes et al. 2010;De Miranda et al. 2014). An analysis of risk factors indicated that male dogs were more often infected with E. canis than the females, but this association was statistically non-significant (P = 0.15) confirming previous studies reporting that there is no correlation between the sex and the presence of E. canis infection in dogs (Rani et al. 2011;Aktas et al. 2015). Similarly, data regarding the age of the animals indicated that animals less than 1-year-old were more prone to E. canis infection compared with animals older than 1 year old, but this association was also statistically non-significant (P > 0.05) indicating that neither sex nor the specific stage of life made subjects more prone to E. canis infection. This observation is contradictory to the results of a study conducted in Mexico, as they had reported higher incidence of E. canis infection in dogs of an age range between 2 and 4 years (Rodriguez-Vivas et al. 2005).
The vectors of E. canis are reported to be R. sanguineus s.l. world wide. In addition, Haemaphysalis longicornis has been documented to be infected by E. canis in Korea (Kang et al. 2013). In this study, R. sanguineus s.l. ticks were collected and identified from 11 dogs in the three cities in Punjab. This observation confirms the previous findings that urban dogs are susceptible to R. sanguineus s.l. infestation (Melo et al. 2011;Rani et al. 2011). However, we did not find any correlation between presence of E. canis DNA with the presence of R. sanguineus s.l. on dogs (P = 1). This finding is not consistent with a previous report from Turkey, in which E. canis infection in dogs was positively correlated with the presence of ticks (Aktas et al. 2013). Most of the animals sampled in this study were pets and groomed by the owners and ticks were only found on 11 of 151 (7.2%) sampled dogs. Out of these 11 dogs having tick burden, only 3 were found infected with E. canis. Good care and management of dogs by owners could be the reason that most dogs were not found to be infested with ticks, and hence, we have not seen a positive correlation between the presence of ticks and prevalence of E. canis in the blood of enrolled dogs. For 39 dogs in which E. canis were detected but ticks were not found on their bodies, we assume that dogs may had came across ticks during their daily routines (for example while on a walk in grassy parks) but ticks were removed by their owners during routine washing and checkup or ticks may also detach spontaneously after blood meal.
We have also tried to correlate various clinical signs (including body temperature, mucous membrane status, hydration status and the presence of haematuria) observed in sampled dogs with the prevalence of E. canis, but our results indicated that none of them was found to be associated with the presence of E. canis in dog's blood (P > 0.05 for all parameters) (Table 3). Our results are in overall agreement with a recent report from India in which it has been documented that none of the studied epidemiological factors was found to be associated with the presence of E. canis in dogs (Rani et al. 2011).
Diagnosis of ehrlichiosis can be challenging due to its different phases and multiple clinical manifestations. Canine monocytic ehrlichiosis should be suspected when a compatible history (living in or travelling to an endemic region, previous tick exposure), typical clinical signs and characteristic haematological and biochemical abnormalities are present (Harrus & Waner 2011). As ehrlichiosis is a vectorborne disease, various haematological parameters were also compared between parasite-positive (N = 42) and -negative (N = 109) blood samples. Our results indicated that white blood cell counts were significantly increased (P = 0.004) in dogs where E. canis was detected in blood than E. canisnegative dogs (Table S1). Our results are in agreement with Shipov et al. (2008), as they had reported a higher number for WBC in dogs surviving with an E. canis infection (Shipov et al. 2008). Characteristic features of ehrlichiosis, like thrombocytopenia and pancytopenia, were not observed during present study. The potential reason for this could be the coinfection of multiple parasites masking the typical symptoms of a particular disease. Often, subclinical infection with E. canis will go unrecognized, as most dogs are thought to control the infection immunologically, and infected dogs appear healthy until late in the infection, when pancytopenia, uveitis, weight loss and haemorrhagic disorders arise, and a diagnosis of ehrlichiosis is made (Gaunt et al. 2010;Greene 2013).
In conclusion, we have used an already established PCR protocol for the detection of E. canis in apparently healthy dog blood samples and this is first report from Pakistan on this topic. We are recommending the use of this PCR based protocol to veterinary practitioners and pet owners for the detection and/or confirmation of E. canis infection in dogs to improve their health status.