Clinical and diagnostic aspects of feline cutaneous leishmaniosis in Venezuela

Venezuela is an endemic area for human and canine leishmaniosis due to Leishmania infantum and parasites of the Leishmania braziliensis and L. mexicana complexes. Limited data are available on feline leishmaniosis (FeL) in this region. The aim of this study was to describe clinical and diagnostic aspects of FeL in Venezuela. Thirty-one domestic cats from urban areas of Lara State in Venezuela were enrolled. Twenty-five were healthy. Six other cats had solitary or multiple nodular lesions, which were located on the nose, ears, tail and lower limbs. Cutaneous lesions were characterized by diffuse pyogranulomatous infiltrate in all sick cats with numerous intracellular and extracellular amastigotes, and immunohistochemistry was positive for Leishmania in five sick cats. All healthy cats were seronegative for L. infantum and L. braziliensis antigens by ELISA. Two out of five sick cats yielded a positive ELISA result to both Leishmania antigens with higher antibody levels to L. braziliensis compared to L. infantum. Significantly higher antibody levels by ELISA as well as a higher number of bands by Western blot (WB) were found for L. braziliensis when compared to L. infantum antigens in all sera from Venezuelan sick and healthy cats. All healthy cats were blood Leishmania spp. qPCR negative while three out of six sick cats were blood qPCR positive. All paraffin-embedded skin biopsies (n = 4) as well as cutaneous cytology (n = 3) were positive by Leishmania spp. qPCR in sick cats. Leishmania speciation was obtained only from the cutaneous lesion samples from cytological preparations of two out of three sick cats which were identified as infected with L. mexicana or a closely related specie. Feline leishmaniosis should be included in the differential diagnosis list of nodular-ulcerative lesions. The most reliable diagnostic technique in sick cats is cytological or histopathological examination along with immunohistochemistry, since blood PCR and serology by ELISA might be negative. WB appears to be more sensitive in detecting infection. Cats with leishmaniosis from Venezuela are most likely infected with species of L. mexicana or a closely related species or the L. braziliensis species complex and not with L. infantum.


Background
Leishmaniosis is a parasitic disease caused by an obligate intracellular protozoan of the genus Leishmania (Kinetoplastida: Trypanosomatidae) [1]. In humans, it is manifested clinically in multiple forms including the cutaneous, mucosal and visceral leishmaniosis [2]. The cutaneous form of leishmaniosis in the Eastern Hemisphere is caused by Leishmania tropica, Leishmania  [3].
Cutaneous lesions are the most common clinical sign of leishmaniosis both in human [4] and canine patients [5] for certain Leishmania species such as L. infantum and the L. mexicana and L. braziliensis species complexes. Over 200,000 people develop dermal and mucosal leishmaniosis annually in Central and South America [6,7]. Venezuela is an endemic area of American human cutaneous leishmaniosis (ACL). A wide distribution of ACL has been observed, but the localized clinical form tends to concentrate in states with mountainous geography especially in the Andes (Trujillo, Mérida, Lara and Táchira). It is interesting to note that Lara and Mérida states are characterized by a fairly stable endemic situation that represents almost a third of all national cases [8]. Additionally, cutaneous lesions have been reported in association with L. venezuelensis, in the endemic focus of ACL both in humans and domestic animals such as cats in Barquisimeto, Lara State, Venezuela [9]. Moreover, description of human cutaneous leishmaniosis caused by L. infantum is also documented [10]. However, Viannia is the most relevant subgenus in this territory and is also responsible for metastatic human mucosal leishmaniosis, the severe form of tegumentary disease [11,12]. Feline leishmaniosis has been described with both visceral and cutaneous forms by demonstration of the presence of the parasite in cats. Several cases in domestic cats have been globally reported, also in America and especially in endemic areas of Central America [13], South America such as Brazil [14,15] and Paraguay [16] and also in the Mediterranean basin [17] and the Middle East [18]. Nonetheless, the real susceptibility of cats to infection by Leishmania spp., their role as reservoir hosts, and the outcome of leishmaniosis in these animals, are poorly understood [19]. Several Leishmania species such as L. mexicana [20], L. venezuelensis [21], L. braziliensis [22][23][24], L. amazonensis [25] and L. infantum [15,[26][27][28][29] have been identified to infect cats in Central and South America.
The most frequently described lesions in FeL are ulcerocrusting, nodular dermatitis, alopecia and scaling [30][31][32] while the visceral form of the disease involving the spleen, liver, lymph nodes, bone marrow, eye and kidney is less commonly diagnosed [33,34]. Although clinical cases of leishmaniosis have been reported in cats with coinfection of feline leukemia virus (FeLV) and feline immunodeficiency virus (FIV), the true association between FeL and retroviral infections remains unclear [32,35].
The laboratory tests recommended for diagnosis of FeL due to L. infantum include demonstration of the presence of the parasite by direct microscopic examination in stained smears, and/or culture, serological tests [indirect fluorescent antibody test (IFAT), enzyme-linked immunosorbent assay (ELISA) and Western blot (WB)], and molecular techniques such as the polymerase chain reaction (PCR) [36]. However, limited information is available regarding the diagnosis of other species of Leishmania such as L. braziliensis infection in dogs and cats [37].
Only very limited data are available on FeL in Venezuela [9]. The identification of clinical manifestations associated with Leishmania spp. infection in domestic cats in endemic areas as well as the best diagnostic techniques to be employed are crucial topics. Therefore, the aim of this study was to describe clinical and diagnostic aspects of FeL in an endemic area of American cutaneous leishmaniosis, the state of Lara in Venezuela.

Cats and sampling
Thirty-one outdoor domestic short hair cats from urban areas (Barquisimeto, Cabudare, Quibor) of Venezuela Lara State were enrolled. A full physical examination was performed, and breed, age and gender were recorded before sampling. Twenty-five cats were apparently healthy, and 6 cats presented cutaneous lesions. Blood samples (n = 31) were collected into ethylenediaminetetraacetic acid (EDTA) for DNA extraction and quantitative PCR (qPCR) and into plain tubes to obtain sera samples from 30 cats for Western blot (WB) and enzyme-linked immunosorbent assay (ELISA) and stored at -20°C before use. Unfortunately, hematological and biochemical profiles were not performed.

Diagnosis of Leishmania infection based on cytology, histopathology and immunohistochemistry of cutaneous lesions
Impression smears and fine needle aspirates from ulcerative nodular lesions from 6 sick cats were sampled and stained with a Romanowsky stain variant (Diff-Quick). Macroscopical skin lesions of solitary or multiple nodular and ulcerative areas were biopsied in 5 cats. Skin biopsies were fixed in 10% neutral buffered formalin. The dermal inflammatory pattern and the cell populations were evaluated histologically in hematoxylin and eosin (HE)-stained sections. A deparaffinization step was performed on the paraffin blocks of skin biopsies before Leishmania immunohistochemistry (IHC). Later, a standard staining protocol with AutostainerPlus (Dako Denmark A/S, Glostrup, Denmark) using rabbit polyclonal antibodies to L. infantum was followed. Sections were then counterstained with hematoxylin and coverslipped for their interpretation [38].
All plates included serum from a sick cat from Cyprus with a confirmed infection with L. infantum as a positive control [39] and serum from a healthy cat as a negative control. All samples were analysed in duplicate. The result was quantified as ELISA units (EU) related to a positive feline serum used as a calibrator and arbitrarily set at 100 EU [40] for both antigens.
The cut-off for L. infantum was established at 9.2 EU (mean + 4 SD of values from 80 cats from the UK, a non-endemic area). Sera were classified as being positive, when having a value equal or higher than 15.3 EU and negative with 9.2 EU. Values in between were considered doubtful.
The cut-off for L. braziliensis was established at 13.8 EU (mean + 4 SD of values from 80 cats from the UK, a non-endemic area). Sera were classified as positive when having a value equal or higher than 21.0 EU and negative with 13.8 EU. Values in between were considered doubtful.

Western blot (WB)
Sera from 25 apparently healthy cats and five cats with lesions compatible with cutaneous leishmaniosis from Lara State in Venezuela were assessed by WB. In addition, sera from 8 cats from the Queen Mother Hospital at the Royal Veterinary College (RVC), University of London, were used as negative controls from cats living in a non-endemic area of leishmaniosis. Sera from 8 cats from Catalonia in Spain, of which 6 cats were seropositive to L. infantum by ELISA, one presenting a doubtful result and one being negative, were also evaluated. Cats from Catalonia (a non-endemic area for L. braziliensis infection) were included to compare the pattern of WB with cats from Venezuela. The seropositive cats to L. infantum from Catalonia were diagnosed with clinical leishmaniosis and used as positive controls for L. infantum antigen.
Western blot was performed with L. infantum (MHOM/MON-1/LEM 75) and L. braziliensis (MHOM/ BR/88/BCN-25) promastigotes as antigens [41,42]. Sera from cats with leishmaniosis that reacted with polypeptides of low molecular mass (< 36 kDa) of L. braziliensis or L. infantum antigens were considered positive for WB due to the fact that these antigens are the most specific fractions in diagnosis of FeL [17,43].
DNA extraction from blood, paraffin-embedded skin biopsies and cytology from skin lesions Blood DNA purification DNA was extracted using the Gen Elute blood genomic DNA kit (Sigma-Aldrich) from 31 blood samples. Blood from a non-infected clinically healthy cat was included as negative control every time that DNA extraction was performed [44].
Purification of genomic DNA from formalin-fixed, paraffinembedded skin biopsies A deparaffinization step was performed on the paraffin blocks of skin biopsies from 4 sick cats (FeV2, FeV3, FeV5 and FeV6) using buffer (20 mM Tris-HCL Ph 8.5; 1 mM EDTA), heating for 10 min at 95°C and centrifuging for 20 min at 12,000×g. Then, DNA extraction was performed using the QIAamp® DNA FFPE (Qiagen, Hilden, Germany) in accordance with the manufacturer's recommendations.
Purification of genomic DNA from cytological slides from skin lesions DNA extraction was performed from cutaneous lesions from cytological slides from 3 sick cats (Fev3, Fev4 and Internal transcribed spacer 1 (ITS1) restriction fragment length polymorphism (RFLP), quantitative PCR (qPCR), sequencing and phylogenetic analysis The species identification of the Leishmania isolates was performed on DNA from cutaneous lesions (cytological preparations) from 3 sick cats from Venezuela (FeV3, FeV4 and FeV5) and on DNA from cutaneous lesions (skin paraffin-embedded biopsies) from 4 sick cats (FeV2, FeV3, FeV5 and FeV6). Two different techniques were performed.
Polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis of amplified ITS-1 sequences braziliensis were used as positive controls. A non-template control with the same reagents described above but without DNA was added to PCR to rule out contamination.
The PCR products, previously digested with the restriction enzyme BsuRI (HaeIII), were separated by electrophoresis in 2% wide-range agarose (Sigma) at 150 V in SGTB 1× buffer (GRISP LDA, Research Solutions, Porto, Portugal). A solution of SYBR safe DNA gel stain (Invitrogen Ltd., Paisley, UK) was used to visualize the separated DNA fragments under UV light [47].
Quantitative PCR high-resolution melting (qPCR-HRM) Leishmania genotyping based on a ITS1, sequencing and phylogenetic analysis A fragment of ITS1 region of the leishmanial ribosomal RNA operon was amplified (265-288 bp) by real-time PCR using primers ITS-219F (5'-AGC TGG ATC ATT TTC CGA TG-3') and ITS-219R (5'-ATC GCG ACA CGT TAT GTG AG-3') and then evaluated by high resolution melting (HRM) analysis as previously reported [48]. DNA samples extracted from promastigote cell cultures of L. infantum, L. tropica and L. major were used as positive controls for each corresponding PCR reaction and DNA from colony-bred dogs negative by PCR for vector-borne pathogens was used as a negative control. A non-template control (NTC) with the same reagents described above but without DNA was added to each PCR to rule out contamination.
All positive PCR products were sequenced using the BigDye Terminator v.3.1 Cycle Sequencing Kit and an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems), at the Center for Genomic Technologies, Hebrew University of Jerusalem, Israel. DNA sequences were evaluated with the ChromasPro software version 2.1.1 (Technelysium Pty Ltd., South Brisbane, Australia) and compared for similarity with sequences available on the GenBank, using BLAST program (https://blast.ncbi.nlm.nih.gov/ Blast.cgi).
Phylogenetic analysis was performed by MEGA6 [49] using the Maximum Likelihood method based on the Tamura 3-parameter model [50]. Initial phylogenetic trees for the heuristic search were obtained by applying the Neighbor-Joining method to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach. The bootstrap consensus tree inferred from 1000 replicates was taken to represent the evolutionary history of the taxa analyzed [51] and branches corresponding to partitions reproduced in less than 60% bootstrap replicates were collapsed.

Detection of FeLV antigen and FIV antibody
In order to evaluate retroviral infections to rule out concomitant infections the same 30 cats from Venezuela described above (5 sick cats and 25 apparently healthy cats) were tested serologically for FeLV antigen and FIV antibody. Detection of FeLV p27 antigen and anti-FIV antibodies was performed by a commercial ELISA (INGEZIM FeLV and INGEZIM FIV®, Ingenasa, Madrid, Spain) according to the manufacturer's protocol.

Statistical analysis
Statistical analysis was performed using the SPSS 17.0 software for Windows (SPSS Inc., Chicago, USA). A non-parametric Mann-Whitney U-test was used to compare groups. A non-parametric Wilcoxon signed-rank test was used to compare paired continuous variables. Differences were considered significant with a 5% significance level (P < 0.05. The descriptive statistical analysis was conducted with R project software (2017).

Cats
Physical examination of all cats included in this study determined that 24 were adults, 5 were old cats and 2 were young cats. Moreover, the distribution according to gender was 13 females and 18 males. The majority of cats did not present any systemic clinical sign or dermatological lesions compatible with leishmaniosis and were classified as apparently healthy (25/31, 80.6%). All apparently healthy cats were owned cats from Barquisimeto and Cabudare cities from Lara State. They were 9 females (n = 2, old cats; n = 7, adults) and 16 males (n = 3, old cats; n = 12, adults; and n = 1 young cat). On the other hand, 6 of the 31 cats (6/31, 19.3%) presented dermatological clinical signs. Signalment, geographical location and clinical description are summarized in Table 1. Sick cats were all stray cats living in cat colonies. They were 4 females and 2 males all adults, except for one 8month-old young male. Skin lesions consisted of solitary or multiple nodular lesions (Fig. 1) which were located on the nose (n = 3) (Fig. 1d-f ); ears (n = 1) (Fig. 1c); ears and nose (n = 1) (Fig. 1a, b); and nose, ears, tail and lower limbs (n = 1). Cats did not show other clinical signs. Those cats did not receive any treatment and were humanely euthanized.

Cytology, histopathology and immunohistochemistry
Cytology was performed from cutaneous lesions in 4 sick cats. In the majority of cases, mixed inflammation with predominance of macrophages and neutrophils was found. Numerous intracellullar and extracellular Leishmania amastigotes were also noted ( Fig. 2a, b). Histologically, cutaneous lesions from 4 sick cats were characterized by epidermal hyperplasia and hyperkeratosis. Diffuse infiltrate with predominance of macrophages and plasma cells with numerous intracellular and extracellular amastigotes were observed (Fig. 2c). Occasionally, mast cells, lymphocytes and eosinophils were also encountered. Additionally, one sick cat presented crusting and necrosis, and presence of amastigotes was not observed. Immunohistochemistry was positive for Leishmania spp. in the 5 biopsied cats (Fig. 2d).
The cats that were seropositive to L. infantum antigen from Catalonia were diagnosed with clinical leishmaniosis. When those animals were tested serologically with L. braziliensis antigen, 5 out of 8 presented negative ELISA results, one had a doubtful result and two showed positive results.

Western blot
As expected, sera from cats from the UK did not react with any polypeptides from both antigens. Bands recognized for L. braziliensis and L. infantum antigens by cat sera from Venezuela and Catalonia are described in The majority of sick cat sera from Venezuela recognized variable patterns of polypeptides with molecular masses ranging between 14-70 kDa for L. braziliensis antigen while they recognized polypeptide from L. infantum antigen less frequently (Table 1).
There was a statistically significant predominance of bands specific for L. braziliensis antigen in cats from Venezuela when compared to Catalonian cats ( Table  2) (Mann-Whitney U-test: Z = -4.03, P = 0.0001). Also, when sick Venezuelan cats were compared to seropositive cats from Catalonia, a high number of bands for L. braziliensis antigen was observed (Mann-Whitney U-test: Z = -2.55, P = 0.01) ( Table 3).
The Venezuelan cats showed a significantly higher number of bands for L. braziliensis antigen when compared to L. infantum antigen (Wilcoxon signed-rank test: Z = -3.15, P = 0.02) ( Table 2). Additionally, a higher number of bands for L. brazilienis antigen were also found when compared with L. infantum antigen in Venezuelan sick cats (Wilcoxon signed-rank test: Z = -3.58, P = 0.0001) ( Table 3). No statistical differences were observed within Catalonian cats when L. braziliensis and L. infantum bands were compared.  Four out of five sick cats from Venezuela resulted positive for L. braziliensis [low molecular mass (< 36 kDa)]. In addition, positive results to L. braziliensis WB were found in 6 out of 25 (24%) apparently healthy cats from Venezuela and also in 2 sick cats from Catalonia. In the case of cats from Venezuela, 3 sick cats and 7 apparently healthy cats resulted positive for L. infantum WB. Five out of eight Catalonian cats presented compatible results for L. infantum antigens. In general, the intensity of bands in sick cats increased with an increase in the antibody level. Blood and cutaneous lesions Leishmania kinetoplast qPCR, ITS1 RFLP and qPCR for identification, sequencing and phylogenetic analysis All twenty-five clinically healthy cats were blood qPCR negative (25/31 cats, 90.3%). Additionally, when blood from sick cats was analysed, 3 out of 6 cats were Leishmania kinetoplast qPCR positive (3/31 cats, 9.7%) while 3 were negative, including 1 seropositive sick cat by ELISA. DNA extraction was performed from paraffinembedded skin biopsies of 4 sick cats and all samples were positive by kinetoplast qPCR (Table 1). There was no correlation between the pattern of bands recognized by WB and PCR results.
DNA samples from cytological preparations of cutaneous lesions from 3 cats was extracted and all samples were positive by kinetoplast qPCR (Table 1). Positive DNA samples from cutaneous lesions paraffinembedded skin biopsies (n = 4) and cytological preparations (n = 3)] were submitted to parasite species identification by PCR amplifying a fragment of the ITS1 region. Only samples from cytological preparations of 2 cats were positive by ITS1 qPCR-HRM (FeV3 and FeV4) while samples from FeV5 were negative. In addition, the same DNA from the cytological preparation from FeV3 was also confirmed as positive by PCR-RFLP, but FeV4 and FeV5 were PCR-RFLP-negative (Table 1). All DNA samples from paraffin-embedded skin biopsies were negative by ITS1 qPCR-HRM (Table 1) and PCR-RFLP. The DNA sequence of cat FeV3 was 100% identical to a partial 18S rRNA ITS1 sequence of L. mexicana (Gen-Bank: AB558250.1) over 210 bp, as found by BLAST analysis. In contrast, the DNA sequence of cat FeV4 was  only 93% identical to a partial 18S rRNA ITS1 sequence of L. mexicana (GenBank: AB558250.1). A phylogenetic tree with the two Venezuelan cats' results is presented in Fig. 3; in this tree, the DNA sequences from cats FeV3 and FeV4 clustered together with L. mexicana sequences from other sources deposited in GenBank.

Detection of FeLV antigen and FIV antibodies and relationship with Leishmania infection
All cats tested (n = 30) were negative for FeLV antigenemia. FIV-specific antibodies were found in 2 out of 25 apparently healthy cats tested (6.6%), both of them seronegative for L. infantum and L. braziliensis-specific antibodies based on ELISA and also negative by kinetoplast qPCR, but when WB positivity was studied, both cats resulted positive to L. braziliensis-specific bands and one was positive to L. infantum-specific bands.

Discussion
The present study describes a clinical case series of solitary or multiple ulcerative nodular dermatitis due to Leishmania spp. in cats from Lara State, Venezuela. Unfortunately, so far, Leishmania identification has only been possible from the skin lesion of two cats and identified as L. mexicana. To the best knowledge of the authors, we report the first feline case of cutaneous lesions due to L. mexicana infection from Lara State, Venezuela. Interestingly, so far cats have only been described to be  infected with L. mexicana in Texas in the USA [13,20,52]. In agreement with our findings, L. mexicana infection was previously diagnosed in humans with cutaneous leishmaniosis from Lara State, Venezuela [53] as well as in pools of Lutzomyia sand flies from Sucre State in Venezuela [54]. Unfortunately, Leishmania identification was not obtained from the remaining sick cats from Venezuela described in the present study. Since formalin-fixation of histological specimens causes partial DNA destruction, which may hamper diagnostic PCR analysis [55,56], we believe that there was inhibition of DNA PCR-based analyses of histological specimens. Therefore, the amount of DNA amplified was reduced and this did not enable the identification of Leishmania. However, based on clinical and serological findings and the geographical distribution of FeL [36], it is likely that species of L. mexicana and/or L. braziliensis complexes were the cause of infection of the remaining cats described here. As with the cases presented here, clinical disease in cats caused by natural infection with species other than L. infantum is typically reported as nodular or ulcerative dermatitis with no systemic clinical signs. Skin lesions are often single but they can metastasize [36].
Solitary cutaneous lesions have been reported in association with L. venezuelensis infection, in the endemic focus of ACL both in humans and domestic animals in Barquisimeto, Lara State, Venezuela [9,53]. Interestingly, in this previous study, four cats were observed with cutaneous nodules on the nose and smaller nodules on the ears, and diffuse nodular lesions on the tail and legs [9]. This clinical presentation was similar with the findings obtained in the present study from cats from the cities of Quíbor, Cabudare and Barquisimeto in Lara State, Venezuela. It is also important to highlight that feline cutaneous lesions described in the present study are similar to the ones described also in humans. It is likely that cats might only be an accidental host of L. venezuelensis infection [36]. It is also important to remark that the grouping of L. mexicana species complex is still controversial [53]. Leishmania venezuelensis was originally described on the basis of distinguishable multilocus enzyme electrophoresis (MLEE) patterns as a species independent of other members of the L. mexicana complex [57]. Furthermore, a monoclonal antibody specific for L. venezuelensis was developed for identification using immunological methods [58]. However, there are limited molecular data regarding L. venezuelensis [59] and some authors strongly suggested that L. venezuelensis is a variant of L. mexicana [53].
A study carried out in Cojedes State of Venezuela revealed human, dog and equine populations with ulcers, other active lesions, skin scars and mucosal alterations, due to Leishmania (Viannia) braziliensis characterized by zymodeme and serodeme typing [60]. In addition, natural [23,24] and experimental [19] L. Fig. 3 Phylogenetic analysis constructed based on 210 bp DNA sequences of the ITS1 locus of Venezuelan cats (FeV3 and FeV4). Sequences from this study were compared to other sequences deposited in GenBank. Phylogenetic analysis was inferred by using the Maximum Likelihood method based on the Tamura 3-parameter model. The number of bootstrap replicates are 1000 and branches corresponding to less than 60% bootstrap replicates are collapsed. Bootstrap values higher than 60% are indicated next to the branches. GenBank accession number, the strain, country of origin and host from which the sequences were derived are included for each sequence braziliensis infections have been described in domestic cats. Leishmania braziliensis natural infection in cats has been described in Brazil [60], French Guiana [23] and in northern Argentina [61]. Cutaneous lesions previously described were also similar to the ones reported in the present study. However, the finding of cats with cutaneous leishmaniosis does not reflect an important role of these domestic animals in the natural transmission of the disease in these areas, and these animals probably represent accidental hosts [21].
In the present study, cats presented ulcerative nodular dermatitis mainly on the face. Cutaneous lesions in cats naturally infected with Leishmania spp. occur mainly on the nose, followed by the ears or at both sites [2] and also occasionally on the limbs. The skin alterations in FeL are unspecific and can be associated with other clinical conditions [25]. The commonly seen cutaneous nodular form in FeL cases should be distinguished from nodules caused in cats with sterile or eosinophilic granuloma, cryptococcosis, sporotrichosis, histoplasmosis, mycobacterioses, and a wide variety of cutaneous neoplasms, e.g. feline sarcoids, mast cell tumor, fibrosarcoma, basal cell carcinoma, bowenoid in situ carcinoma and lymphoma [36]. The main differentials of the ulcerative lesions include squamous cell carcinoma, idiopathic ulcerative dermatitis, herpes virus dermatitis, mosquitobite dermatitis, atypical mycobacteriosis and feline leprosy, cutaneous vasculitis, erythema multiforme and cold-agglutinin disease. Interestingly, squamous cell carcinoma may co-exist with Leishmania infections as clinical case reports due to L. infantum have been documented in cats in Europe [62,63]. In the present study, concomitant disease was not diagnosed in the sick cats studied.
Leishmaniosis is diagnosed by demonstration of the parasite by direct microscopic examination in stained smears, and/or culture of skin lesions, lymph node aspirates, peripheral blood, bone marrow aspirates or indirectly by serological techniques [41]. A clinical form characterized by a very low number of intralesional parasites can be detected by an immunohistochemical technique. This technique is a highly sensitive and specific tool for the diagnosis of canine and feline leishmaniosis [36,64]. In this study, different techniques were used in the cats with cutaneous lesions, such as cytology, skin biopsy and immunohistochemistry, by which the diagnosis of infection by Leishmania spp. was made. Besides that, serological and molecular diagnostic techniques were also used in sick and apparently healthy cats. One sick cat and 25 clinically healthy cats were negative by serology and blood qPCR. The sick cats did not always give a positive result to blood qPCR and serology. The present findings might indicate that clinically healthy cats are not carriers of Leishmania spp. present in Venezuela.
Moreover, the ELISA and qPCR discordant results can be attributed to the inherent differences between serological testing and molecular methods. PCR is a very sensitive technique. However, intermittent parasitemias are very likely in cats as described in dogs [65] and therefore, PCR from blood might not be very sensitive. The parasite load in blood was quite variable in the cats studied and poorly correlated with the degree of antibody levels [17]. The present findings suggest that more than one technique should be used for detection of feline cutaneous leishmaniosis in South America.
Leishmania species such as L. infantum and L. braziliensis among others co-exist in South America and both species can infect cats [36]. However, there are limited studies regarding serological tests that will distinguish between L. infantumand L. braziliensis-specific feline antibodies in these regions. Here, we report an ELISA that combines both antigens. Interestingly, higher antibody levels were found using L. braziliensis antigen in Venezuelan cats when compared with L. infantum antigen. This quantitative in-house ELISA appears to help establishing what Leishmania species or closely related Leishmania species are most likely infecting cats in the respective endemic areas. These findings are extremely important in areas where several parasite species might co-exist, therefore, this type of ELISA should be used to determine the most likely Leishmania parasite infecting dogs and cats in South America. Unfortunately, we did not have antigen of L. mexicana to perform WB.
Furthermore, the WB analysis also revealed that Venezuelan cat sera recognized a significantly higher number of L. braziliensis polypeptides when compared to L. infantum antigen. In addition, the intensity of bands increased with an increase in the antibody level. There was no correlation between the pattern of bands recognized and PCR results. Antigens of low molecular weight (12-14 and 14-18 kDa), seem to be very specific and their recognition in the immunoblot is highly sensitive in the diagnosis of subclinical Leishmania infection in dogs and cats [43]. In the cats analysed, similar results were observed. In experimentally infected dogs, antibodies specific for low molecular weight fractions are the first to appear following infection [41]. Based on the results of this study, it appears that the WB enables detection of early phases of infection in apparently healthy cats with negative antibody levels by ELISA or PCR results. It also important to highlight that WB seems to be the best serological technique to be used when testing sick and apparently healthy cats from Venezuela as previously reported for European cats [43]. In agreement with the present study, another study reveals that the use of WB with whole antigen or antigenic Fe-SODe (iron superoxide dismutase) fraction was an optimal method for the detection of FeL [66]. The use of antigenic fractions of cultures of L. mexicana, L. braziliensis and L. infantum showed satisfactory results with high sensitivity, specificity and efficacy for the detection of this disease in cats [66]. Therefore, WB should be widely used in the clinical setting for diagnosis clinical FeL as well as for detection of subclinical infections.

Conclusions
We conclude that leishmaniosis should be included in the differential diagnosis list of nodular-ulcerative lesions in cats, mainly on the nose and ears. In addition, to the best of our knowledge, we described for the first time, cutaneous lesion associated with L. mexicana infection from two Venezuelan cats. The most reliable diagnostic technique in sick cats is cytological or histopathological examination along with immunohistochemistry, since blood PCR and serology by ELISA might be negative. However, WB appears to be more sensitive in detecting infected cats. Based on molecular and serological findings, cats from Venezuela are most likely infected with species of L. mexicana or L. braziliensis species complexes rather than L. infantum. Finally, the present findings might indicate that clinically healthy cats are not carriers of Leishmania spp. present in Venezuela.