Molecular and parasitological detection of Leishmania spp. in dogs caught in Palmas, TO, Brazil

This study evaluated occurrences of Leishmania infantum in dogs in the municipality of Palmas, Tocantins, comparing diagnostic data obtained using the polymerase chain reaction (PCR) and parasitological diagnosis. Blood samples and lymph node aspirates were collected from 63 dogs of males and females and various ages and races, with or without owners, between August 2009 and June 2010. Slides containing smears of lymph node aspirates were stained with Giemsa stained. In PCR, the 145 bp target sequence of the LT1 fragment, located in the Leishmania donovani kDNA minicircle was detected using the RV1 and RV2 oligonucleotide primers. The chi-square test revealed that there was a significant relationship between the symptoms and dogs that were positive for visceral leishmaniasis (VL). The parasitological investigation showed concordance of 66.7% with PCR on blood and 84.1% with PCR on lymph node aspirate. In addition to these tests, evaluations of the diagnoses in parallel and in series were conducted, which showed concordances with the parasitological test of 76.2% and 74.6%, respectively. The results make it possible to suggest that PCR on lymph nodes should be used in evaluating large populations (surveys) and that the parasitological test should be used for initial clinical evaluations in veterinary consultation offices.


Introduction
Visceral leishmaniasis (VL), also called kala-azar, is an infectious disease with chronic progression (SHAW, 2006). The causal agent of VL is the protozoon Leishmania (Leishmania) infantum. In Brazil, as in other places, VL transmission takes place during the heteroxenous life cycle of the parasite. The major VL infection route for humans or animals is through the bites of female phlebotomine sand flies, particularly of the species Lutzomyia longipalpis (LAISON;SHAW, 1987). Dogs are considered to be the main domestic reservoirs of the parasite, playing an important role in the epidemiological cycle of VL transmission to human hosts (DANTAS-TORRES, 2007).
Different techniques may be used to diagnose VL in humans and dogs. However, these tests face limitations of various kinds, and differ in terms of sensitivity and specificity, practical applicability, requirements concerning field conditions, and availability of chemicals used (GONTIJO; MELO, 2004). PCR-based gene amplification techniques have gained increasing popularity for Leishmania diagnosis (CAMARGO; BARCINSKI, 2003), since these protocols are fast and sensitive, and do not require parasite culture growth (LACHAUD et al., 2002;ASSIS et al., 2010).
Nonetheless, parasitological methods carried out using biopsy specimens or even using material collected by puncture aspiration of the spleen, liver, bone marrow or lymph nodes, are still conventionally used to detect L. infantum parasitism. Although one of the possible disadvantages of this method is variable sensitivity, it is highly specific, and is considered to be the gold standard in diagnosing canine VL (BRASIL, 2003;GONTIJO;MELO, 2004).
In this light, the aim of the present study was to evaluate occurrences of the parasite L. infantum in blood and lymph node aspirates from dogs sampled in the municipality of Palmas, state of Tocantins, Brazil. The samples were analyzed by means of PCR and the results were compared with diagnoses made in accordance with conventional parasitological procedures.

Study area, canine population and collection of biological specimens
Blood samples and lymph node aspirates were collected from 63 male and female dogs of various ages and breeds that were either kept as pets or had been caught by the Zoonosis Control Center of the municipality of Palmas, state of Tocantins, Brazil. Sample collection took place between August 2009 and June 2010. The dogs were also evaluated in relation to coat length, presence of ectoparasites (Rhipicephalus sanguineus) and occurrences of visible signs like low weight, alopecia, scaliness, onychogryphosis and minor crusty injuries. The dogs were also clinically categorized as symptomatic, when they presented one or more clinical signs, and as asymptomatic, when no such signs were present, in accordance with previously defined criteria (MOLINA et al., 1994;MANCIANTI et al., 1988).
Blood samples (3 to 5 mL) were collected by puncturing the brachial or jugular veins. They were stored in sterile tubes containing anticoagulant (EDTA 27 mM) and transported to the laboratory. Lymph node aspirates were collected from the popliteal region using 10-mL syringes. Part of this material was smeared onto slides (in triplicate) and was used in the parasitological investigation, and part was sent to the laboratory for DNA analysis.

Parasitological investigation
The smears of lymph node aspirates were fixed in methyl alcohol and dried. Next, the smears were stained using Giemsa. After 30 minutes, the slides were washed in buffered water and left to dry at room temperature. Following this, they were inspected under a microscope. The smears were considered to be positive for VL when at least one parasite in the amastigote stage was present.

Molecular diagnosis
Initially, DNA was extracted in accordance with the phenol/ chloroform method (SAMBROOK, 2001). DNA was then electrophoresed and quantified by means of spectrophotometry. The DNA specimens were stored in a freezer at -20 °C until PCR analysis.
The reactions were conducted using a 20 µL final volume containing 1.5 U/µL of Taq DNA polymerase (Invitrogen), 20 mM of tris-HCl buffer (pH 8.4), 50 mM of KCl (Invitrogen), 1.5 mM of MgCl 2 (Invitrogen), 0.2 mM of deoxyribonucleotide triphosphate (dNTPs) (Invitrogen), RV1 and RV2 primers (10 pmol each), DNA from the individuals (between 80 and 150 ng/µL) and ultrapure water (12.7 µL) to complete the desired volume. Ultrapure water was used instead of the DNA sample in the negative control. The L. infantum DNA strain MHOM/BR2000/MER2 (FIOCRUZ/BA; Merivaldo strain, IOC-LC2455), which had been isolated from a VL patient in an endemic area of the town of Jequié, state of Bahia, Brazil, was used as the positive control (PARANHOS-SILVA et al., 2001). The reactions were carried out in a thermal cycler (PxE 0.2, Thermo Electron Corporation) with the following cycling parameters: 94 °C for 5 minutes; 35 cycles at 94 °C for 45 seconds, 58 °C for 45 seconds and 72 °C for 45 seconds; and an annealing cycle at 72 °C for 10 minutes and 4 °C for 10 minutes. Amplicons were electrophoresed in agarose gels stained with ethidium bromide (0.02 µL/mL). The gels were inspected under UV light and photographed. A 100-bp ladder (Amresco) was used for comparison. Samples that were diagnosed negative at the first examination were retested following the same routine as described above.

Data analysis
The characteristics of the canine population sampled and the respective positive results for VL were statistically analyzed using the chi-square (χ 2 ) test. The degree of agreement between the results obtained from PCR and parasitological investigation was analyzed using the kappa coefficient. Both of these tests were available through the freeware BioEstat version 5.0, and were used with 5% statistical significance. Co-positivity was calculated as a/(a + c), co-negativity was expressed as d/(b + b) and raw concordance was described as (a + d)/(a + b + c + d) (TÁVORA et al., 2007).

Clinical signs and characteristics of the canine population
As a rule, the characteristics of the canine population investigated varied considerably (Table 1). The signs most commonly noticed were alopecia, low weight, onychogryphosis and minor crusty injuries.

Parasitological findings
The parasitological examinations revealed that 26 dogs (36%) were positive for VL, of which 13 (50%) were males and 13 (50%) were females. However, the χ 2 test did not show any statistically significant association between the majority of the characteristics evaluated and positive results from the parasitological examination. This finding is in agreement with the results from Saridomichelakis et al. (2005), who also used the technique to analyze lymph node and bone marrow smears, and Azevedo et al. (2008), who equally did not observe any statistically significant difference between the sexes regarding positive canine VL diagnoses using serological methods. Likewise, the authors of that study did not find any significant association between positive VL results and the dogs' sex. Concerning infestation by ectoparasites, in spite of the potential role that ticks (R. sanguineus) may play as vectors in leishmaniasis, which was previously investigated in an interesting paper by Coutinho et al. (2005), no statistically significant difference in VL detection was observed in the present study between dogs with and without tick infestation.
The associations shown by clinical signs with positive VL results in dogs are presented in Table 2. A statistically significant association was observed between the signs and infection by L. infantum, though this result disagrees with the serological findings reported for canine VL in a previous study (SANTOS et al., 2010). Nevertheless, it is important to stress that not every symptomatic dog was diagnosed as VL-positive, which agrees with the conclusions reached by Osman et al. (1997) and Dantas-Torres et al. (2006). Azevedo et al. (2008) obtained a somewhat disturbing and opposite result, reporting that 55.2% of VL-positive dogs in their study were asymptomatic. Their findings are a matter of concern, because they suggested that VL could not be easily associated with infected dogs, since those animals did not show signs of the disease, even though they played an important role in VL transmission (AZEVEDO et al., 2008). These results indicate that characterization of VL-positive or negative status based on clinical examination should be made with caution (OSMAN et al., 1997). Table 3 shows the comparison between the parasitological findings and PCR analysis relating to VL in 63 dogs. Identical parasitological findings and PCR analyses of blood samples was observed for 42 dogs, thus representing a concordance rate of 66.7%. However, 21 dogs presented discordant parasitological and PCR diagnoses for L. infantum. For the lymph node samples, concordance was observed for 53 dogs (84.1%), while for 10 animals, the two techniques produced discordant results.  As expected, the concordance between parasitological findings and blood PCR results was lower than for the lymph node PCR results. These differences may be explained in the light of the fact that infection by Leishmania starts in blood, a medium in which immune system cells are present, and quickly proceeds to infect the viscera, preferentially in the macrophagic system of the spleen and in the liver, bone marrow and lymphoid organs, thus demonstrating the influence of the VL stage on detection of the protozoan (REY, 2008). These differences were also observed when the PCR results from blood samples were compared to those from lymph node aspirates, with low percentage concordance (54%; kappa = 0.0833; p-value = 0.2183). Such divergences may be explained through considering that in lymph nodes, the parasite is subject to confinement during the dog's immune response, or to different stages of the infection, when the parasite may or may not be present in the blood flow.

Comparison between parasitological findings and molecular results
Another reason for the divergence in the results may be the low sensitivity of PCR on blood samples, possibly due to the low numbers of parasites usually detected in these samples (FISA et al., 2001;MANNA et al., 2004;NUNES et al., 2007). On the other hand, other biological materials like skin, lymph node aspirates, bone marrow aspirates and leukocyte cream (also called buffy coat) present higher sensitivity in the PCR technique (REITHINGER et al., 2000;MANNA et al., 2004). Furthermore, there is evidence that inhibitors (heme groups) of Taq play an interfering role during PCR on blood samples (ABU AL-SOUD; RÂNDSTROM, 1998;REITHINGER et al., 2000). Nevertheless, use of blood as a biological analysis material is justified given the simplicity of collecting it, with less invasiveness, especially when a large number of animals are analyzed. However, independently of the biological analysis material used, it is important to stress that PCR does not produce false positive diagnoses (FALLAH et al., 2011).
In addition to these analyses, we also analyzed diagnoses in parallel. In this, the dogs that produced positive PCR results in at least one of the types of sample used (blood or lymph node aspirates) were considered to be VL-positive and any dog with negative PCR results in both samples at the same time was considered to be VL-negative. Using this method, 41 (65.1%) of the 63 dogs evaluated were VL-positive in PCR analyses on blood or lymph nodes and, among these, 15 (36.6%) had negative parasitological results. In turn, 26 animals (41.7%) had positive parasitological findings which were not confirmed by PCR using either blood or lymph node aspirates. In this regard, the concordance between the techniques was 76.2%.
We also analyzed these results serially, i.e. the animals were considered to be infected by L. infantum when both blood and lymph node aspirates produced positive PCR results. Here, the concordance with parasitological findings was 74.6% (Table 3). Analysis of the results in parallel did not show any difference of enough importance to justify its use, since the PCR results from lymph nodes were similar and therefore the technique can be used alone.
The results from the present study make it possible to suggest that diagnostic procedures like analysis of lymph nodes through PCR can be used to investigate large populations (surveys), while parasitological examinations for initial clinical evaluation can be used in veterinary consultation offices. PCR protocols are more automated, do not produce false positive results and allow analysis on large numbers of samples. In turn, parasitological investigation is technically easier, demands less training, is less costly and may be useful for examinations on individual animals, even though it is more labor-intensive.