Currently, two epidemiological types of VL, anthroponotic (AVL) and zoonotic (ZVL), exist in western China [19, 41]. AVL is endemic only in the oases of the plains of Kashi prefecture, Xinjiang Uygur Autonomous Region. Most cases occur in juvenile and adult people. ZVL can be divided into two subtypes, namely mountainous and desert sub-type . The desert sub-type is endemic in the northwestern desert regions of China, including Xinjiang and western Inner Mongolia . These regions were uncultivated deserts for a long time before being populated by immigrants who introduced agricultural activities, consequently, autochthonous infantile kala-azar occurs. This region is considered to be a natural nidus of kala-azar-infected wild animals that are presumably considered to be the source of human infection. The wild species, Phlebotomus wui and Phlebotomus alexandri, are the vectors infesting different specific landscapes, dry desert region and the stony desert, respectively [22, 41]. The mountainous sub-type occurs in the western mountainous and hilly regions of Gansu, Sichuan, Shaanxi, and Shanxi provinces of China. Patients are mostly children younger than ten years old, and infants are commonly infected. These patients are distributed sporadically. The vector of this form is wild P. chinensis [22, 41]. Elimination and prohibition of dogs in the endemic area markedly reduced the number of human cases, thus dogs are likely the principal source of infection for the mountainous sub-type . Therefore, evaluation of prevalence of Leishmania infection in dogs is of great importance to understand the epidemiology of the mountainous sub-type of ZVL and to prevent transmission of human VL. In this study we collected blood samples from dogs in an endemic area of Jiuzhaigou County, China and evaluated the prevalence of L. infantum infection in these dogs using PCR, ELISA, and immunochromatographic dipstick tests.
According to our results, among 106 dogs studied, 6 (5.66%) dogs presented CanL signs and 100 (94.34%) dogs were asymptomatic. The seroprevalence of dog infection in Jiuzhaigou County is 36.79% by ELISA, higher than those obtained by various authors throughout the Mediterranean basin [42–44]. However, the positive rate detected by dipstick test (9.43%) is much lower than that detected by ELISA in this study. The highest positive rate (51.88%) was found in the PCR detection of parasitic DNA in the blood samples of dogs with significant higher than that of ELISA (p = 0.001) and dipstick (p < 0.001), indicating the PCR method is the most sensitive to detect Leishmania infection in dogs. The result is consistent with that found by other studies [25, 45, 46].
In epidemiological studies on CanL, serological and parasitic DNA detection are sensitive and specific to detect Leishmania infection in dogs [24, 25, 31, 42]. For detection of symptomatic dogs, the two methods obtained comparable sensitivity in this study, while in detection of asymptomatic dogs PCR method was more sensitive than serological methods. The probable reason is that asymptomatic infection dogs may have low infective burden and therefore the lower levels of anti-Leishmania antibodies  while PCR can detect as few as a single parasite theoretically. In this study, most of dogs (94.34%) were asymptomatic, resulting in lower positive rates detected by serological methods than that by PCR. In addition, asymptomatic but seropositive dogs may indicate previous contact with parasite with some of these seropositive dogs being PCR-negative . The biopsy samples of lymph node, bone marrow, skin or conjunctive were usually taken to detect Leishmania infection by PCR and can yield excellent sensitivity [40, 42]. Lachaud L. et al demonstrated that PCR method targeting kDNA (K13A-K13B) using peripheral blood effectively detected the parasite in symptomatic or asymptomatic infected dogs , while other study indicated that blood sample reduces the sensitivity of the test . However, the PCR test based on blood is advantageous than tissues because blood sample can be obtained safely and less invasively, and relatively easy to process. In diagnosis of symptomatic dogs with leishmaniasis rk39 dipstick gave reliable results [49, 50], however, our results demonstrated that the test is not suitable to detect asymptomatic infection of dogs with Leishmania. The probable reason is that k39 antigen is only expressed during active infections [51, 52].
The prevalence of Leishmania infection in dogs that we found in Jiuzhaigou County is very high (total 59.43%), especially in asymptomatic dogs up to 57% (57/100). Some studies indicated that asymptomaticlly infected dogs with Leishmania were not infective to sandflies [53, 54], while some others demonstrated that asymptomatically infected dogs were the sources of the parasite for vector sandflies and as a consequence play an active role in the transmission of L. infantum to human [13–17]. The high prevalence of Leishmania infection in dogs in regions where leishmaniasis is endemic has to be taken into account in any campaign aiming at controlling CanL. In fact, some authors have demonstrated that removing seropositive dogs is an insufficient method to eradicate CanL [55, 56]. In this study we found that Leishmania DNA could even be detected in serologically negative dogs. These seronegative dogs are also possible sources of Leishmanial transmission. These data provide a better estimation of transmission level of the parasite in the endemic area, therefore are essential for designing and implementing appropriate control measures.
The results of this study demonstrated that the amplification of 116-bp fragment in the conserved region of Leishmania kDNA minicircles could be used for detection of parasite DNA in the blood samples of infected dogs. The sequence data of the PCR products also showed the significant diversity among the samples isolated from different dogs and from lab maintained strain of L. infantum (MCAN/CN/90/SC). More than 20% of the amplified 116-bp fragments showed two stretches of sequence AAAA mutated to TTTT and therefore named as Class II compared to the regular Class I type with no mutations. Based on our findings and the findings of others [32, 57], it suggests that a certain percentage of sequence divergence existed in the constant region of the minicircle classes not only among species but also within the species or even within the same parasite. It could be due to recombination or mutation events occurring in the kDNA molecules . Based on the divergence existing in the constant region of the minicircle classes among Leishmania spp., in this study we constructed a phylogenetic tree revealing that L. infantum isolate MCAN/CN/90/SC used in this study was closely related to three other existing L infantum isolates, and more genetically related to other visceralizing species such as L. donovani than to those Leishmania species associated with cutaneous disease. Therefore, the sequence data of Leishmania kDNA minicircles could possibly be used for distinguishing species or strains of Leishmania. It is the plan to collect parasite samples from different endemic areas of China to genetically type the strains or isolates of L. infantum based on the diversities of the Leishmania kDNA minicircles