Introduction

Human infection by Leishmania parasites can cause leishmaniasis, a complex of diseases presents in areas of tropical and subtropical countries. Endemic transmission occurs in 98 countries and 3 territories, where official records estimate the occurrence of 0.2 to 0.4 million cases of visceral leishmaniasis (VL) and 0.7 to 1.2 million cases of cutaneous leishmaniasis (CL) each year (Alvar et al. 2012). As several species may be pathogenic to humans, leishmaniasis can be considered a polymorphic group of diseases, in which clinical and epidemiological characteristics, response to treatment and prognosis, can differ (Pérez-Ayala et al. 2009).

The clinical presentation depends on several factors, among which the infecting species of the parasite is quite relevant (Goto and Lauletta-Lindoso 2012). In consequence, species identification has become an essential requirement for choosing the best treatment with the least side effects and late complications (de Vries et al. 2015; Hodiamont et al. 2014); in research concerning the differential response of Leishmania species to drugs (Fernandez et al. 2012; Obonaga et al. 2014); and in epidemiological studies involving vectors and reservoirs (Özbel et al. 2016; Córdova et al. 2011).

According to the current report on worldwide leishmaniasis incidence, a considerable number of CL cases are reported in the American region, comprising Colombia, Honduras, and Guatemala (Alvar et al. 2012). Some of these countries are also enrolled in investigations concerning the differential effect of drugs against Leishmania in relation to the infecting species (Reveiz et al. 2013; Obonaga et al. 2014), which makes the identification step a necessity for some laboratories.

At least six species of Leishmania have been reported in Colombia: Leishmania panamensis, Leishmania braziliensis, Leishmania guyanensis, Leishmania mexicana, and Leishmania amazonensis, causing both cutaneous and mucocutaneous diseases, and Leishmania infantum (chagasi), mainly associated with the visceral form (Corredor et al. 1990). Different protocols have been used for molecular identification of the species present in clinical isolates using several targets: the genes coding for the heat-shock protein HSP70 (hsp70) and cytochrome B (cyt B), as well as kinetoplast DNA (kDNA) (Montalvo et al. 2016; Martínez et al. 2010; Urbano et al. 2011).

In northern Guatemala, cutaneous leishmaniasis is considered endemic, being caused by L. mexicana and L. braziliensis (Zeledón et al. 1993; Herwaldt et al. 1992). Molecular detection of parasites in clinical samples was reported in this country using a real-time PCR targeting a region of the 18S rRNA gene (Wortmann et al. 2007); and more recently, a conventional PCR was applied to diagnose an imported case of Chiclero’s ulcer coming from Guatemala (Blaylock and Wortmann 2012).

In Honduras, L. braziliensis, L. panamensis and L. mexicana have been involved as the etiological agents of the cutaneous disease, and L. infantum (chagasi) causes visceral and cutaneous lesions (Grimaldi et al. 1987; Ponce et al. 1991; Noyes et al. 1997). No recent records about the use of molecular approaches to characterize parasites from clinical samples in this country were encountered.

The gene coding for the HSP70 protein has been largely used in the last years for Leishmania detection and identification. Compared to the first PCR-RFLP report using this gene (García et al. 2004), three protocols were shown to have superior sensitivity and specificity (Montalvo et al. 2012, 2014; Fraga et al. 2012, 2013a, b). In the current work, the two more appropriated variants for New World species discrimination, PCR-F and PCR-N, and their respective RFLPs, are applied to the analysis of clinical samples from cutaneous leishmaniasis cases from Colombia, Honduras, and Guatemala, in order to establish their usefulness for Leishmania spp. typing in those settings.

Materials and methods

Clinical samples

Lesion lancet scrapings or biopsies were obtained from 93 patients coming from different endemic areas of cutaneous leishmaniasis, where transmission occurs. In some cases, lesion aspirates were obtained as well. They originated from Colombia (n = 73), Guatemala (n = 10), and Honduras (n = 10). Of all samples, 63 were positive for Leishmania according to clinical and parasitological criteria (direct smear and/or culture). The remaining 30 had clinical symptoms of CL, and lived, worked, or visited areas where transmission occurs. All samples were recruited and analyzed at the “Programa de Estudio y Control de Enfermedades Tropicales” (PECET, Universidad de Antioquia, Colombia). An approval from a review board was obtained from PECET (F-CM-0001, 26/07/2016) as well as an informed consent from all patients, allowing the use of their clinical samples in research projects related to diagnosis.

Parasitology

Direct microscopic examination was done on lesion samples fixed with methanol and stained with Giemsa. Slides were analyzed with optical microscopy (×100 magnification). Lesion aspirates were taken with a tuberculin syringe and gently put in a tube containing biphasic NNN medium (Novy, Mc Neal, Nicolle) using a salt solution as the liquid phase. The tubes were incubated at 27 °C for promastigote growth. Samples were considered positive when either parasites could be detected microscopically, or could be isolated by culturing, or both. After a month, a new passage from culture was made and the primary one was considered negative.

DNA extraction

DNA was isolated using the QIAamp DNA Mini Kit (n = 70) or the DNeasy Blood & Tissue Kit (n = 23), both from Qiagen (www.qiagen.com), according to manufacturer’s instructions. The elution volume varied in relation to the sample size (100–150 μL).

PCR amplification

A specific hsp70 gene fragment was amplified using the PCR-F (1286 bp) and PCR-N (593 bp) protocols described by Montalvo et al. 2012. The primers were: PCR-F (F25: GGACGCCGGCACGATTKCT and R1310: CCTGGTTGTTGTTCAGCCACTC); PCR-N (F25 and R617: CGAAGAAGTCCGATACGAGGG). The amplification conditions, for both PCR variants, were: denaturation at 95 °C for 5 min; followed by 35 cycles of denaturation at 94 °C for 40 s, annealing at 61 °C for 1 min, extension at 72 °C for 1 min (except PCR-F: 2 min); and a final extension step of 10 min at 72 °C. Five microliters of DNA obtained from each clinical sample were used as template in most cases. Negative controls were always included, along with a positive one consisting of 100 fg DNA from L. braziliensis strain MHOM/BR/75/M2903. Thermal cycling was performed in a MyCycler™ (Bio-Rad, www.bio-rad.com). Analysis on a 2% agarose gel was used to verify the presence and size of amplified product. The results, expressed as parasite detection sensitivity, represent the total of positive samples obtained by each PCR.

RFLP analysis

For identification of Leishmania species, enzymatic restrictions of PCR-F and PCR-N amplified products were performed following the schemes of Montalvo et al. (2012) and Fraga et al. (2013a). The restriction enzymes used were: HaeIII, RsaI, BsaJI, HindII, SduI (MBI Fermentas, St. Leon-Rot, Germany), and BccI (New England Biolabs, Ipswich, MA, USA). A sequential approach was followed: PCR-F was digested with HaeIII to separate groups of species, after which it was digested with RsaI, SduI, and BccI to further discriminate within these groups. For PCR-N, the first separation was done with BsaJI, followed by RsaI, HindII, and SduI. Enzymatic digestion of strong PCR amplicons was performed in a total volume of 10 μL, while weak amplicons were digested in 20 μL. Each digestion contained ×1 optimal buffer (recommended by the manufacturer), 7 or 17 μL of unpurified PCR product respectively, and 2 U of enzyme. The reactions were incubated for 3 h at 37 °C (55 °C for BsaJI). Restriction patterns were analyzed by electrophoresis in a 3% of small fragment agarose gel (Gentaur, Brussels, Belgium), running at 3.5 V/cm for 3 h. The GeneRuler™ 100 bp DNA Ladder (MBI Fermentas, St. Leon-Rot, Germany) was used as a size reference marker. In case of weak PCR amplicons, some samples were only partially identified as the restriction algorithm could not be completed due to insufficient amounts of PCR amplified products (second or third digestion not done). The results are expressed as species identification efficiency, which represents the number of samples that could be typed or partially typed.

Results

In samples from 63 patients that were positive by parasitological methods, the detection sensitivity of PCR-N was 89%, and of PCR-F 71%. Only two cases (PE-C17 and PE-C23) were encountered positive by PCR-F while negative by PCR-N. RFLP-N could be performed on 71% of the samples, and RFLP-F on 44%.

In all the remaining 30 suspected, but non-confirmed leishmaniasis cases, PCR-N detected Leishmania DNA in the samples, while PCR-F only detected 50% as positive. Species identification could be done for 57% of the samples with RFLP-N, but only for 13% with RFLP-F.

The species identified with each PCR-RFLP protocol are shown in Fig. 1, with stratification per country. A detailed listing of individual results for all samples is available in supplementary Table ESM1. When samples were typed with both protocols, a complete correspondence in the results obtained by both of them was observed. In addition, the species found agreed with those described previously in the countries of origin, with one notable exception. The particular exception was a sample from Colombia (PE-C27) showing an RFLP-N pattern corresponding to L. naiffi (Fig. 2).

Fig. 1
figure 1

Leishmania species identification, after RFLP-F and RFLP-N protocols. The number of species or group of species discriminated is shown according to the country of origin. A species complex (L. mexicana complex: L. mexicana, L. amazonensis, L. garnhami); individual species (L. infantum (chagasi), L. braziliensis, L. panamensis, L. braziliensis, L. naiffi); and common patterns among two or more species are shown (L. pan: L. panamensis, L. guy: L. guyanensis, L. peruviana, L. nai: L. naiffi, L. braziliensis, L. per: L. peruviana)

Full size image
Fig. 2
figure 2

Restriction patterns obtained after RFLP-N. MM molecular weight marker GeneRuler™ 100bp Plus bp DNA Ladder. L. nai RFLP-N/SduI pattern (525 and 68bp). L. bra RFLP-N/SduI pattern of L. braziliensis (411, 114, 68bp) (according to Fraga et al. 2013a, b). The RFLP-N/SduI pattern of sample PE-C27 is shown on the right

Full size image

Discussion

The higher sensitivity of PCR-N compared to PCR-F agrees with previous evaluations of the same protocols, on clinical samples from Peru, where PCR-N reached the highest sensitivity regardless of the clinical presentation or the type of clinical sample (Fraga et al. 2012). This was expected because PCR-N produces a 593 bp amplicon, half of the size of the one obtained using PCR-F (1286 bp) and which is therefore easier to amplify. Apart from this, PCR-N was also more sensitive on biopsies from Old World cutaneous leishmaniasis (Montalvo et al. 2014). However, two samples in this study (PE-C17 and PE-C23) were found positive by PCR-F and not by PCR-N. A probable explanation to this could be the amount of DNA used for the PCR-N reaction in those cases (3 μl), instead of the 5 μl used in PCR-F, although in other samples this difference did not affect the amplification (PE-C 3, 6, 15, 38, 58). If those two particular cases had a considerable small amount of Leishmania DNA in the solution, probably the difference in the volume was crucial in the results obtained. Also, PE-C23 produced a PCR-F weak amplicon, which support this possibility.

PCR-N also detected parasite DNA in clinical samples in which no parasites from smear examination or culture could be demonstrated, probably due to a low parasite load and/or the time of evolution of the lesion. Given the previously reported 100% specificity (analytical and diagnostic) of the here used PCRs (Montalvo et al. 2012; Fraga et al. 2012), and the fact that RFLP resulted in Leishmania-specific patterns, we can assume that these cases are true positives. Hence, both PCRs are complementary to parasitology in identifying leishmaniasis cases. Even though they have the additional advantage over parasitology methods of enabling species typing, they cannot replace these methods because PCRs remained negative in 10–30% of samples in which parasites were found. Nevertheless, if only parasite detection at genus level is desired, without the identification step, multicopy targets (kDNA, rDNA) would fit better for diagnosis.

When deciding between RFLP-F and RFLP-N, two aspects need be considered. On the one hand, typing efficiency is higher with RFLP-N, as a higher number of clinical samples could be typed. On the other hand, RFLP-F allows discriminating all species, whereas for instance L. panamensis and L. guyanensis (both endemic in Colombia) cannot be distinguished with RFLP-N.

In the samples infected with species from the L. mexicana complex, typing could not reach the species level because current hsp70 RFLP protocols do not allow discriminating L. mexicana from L. amazonensis (syn. L. garnhami) (Montalvo et al. 2012). However, phylogeny of Leishmania species based on either HSP70 gene sequences alone (Van der Auwera et al. 2013) or in combination with HSP20 genes (Fraga et al. 2010, 2013b), supported their division into different species, Using a PCR-RFLP protocol that targets the internal transcribed spacer (ITS) of the ribosomal RNA gene array, Berzunza-Cruz et al. 2002 succeeded in differentiating L. amazonensis from L. mexicana for diagnosis.

Leishmania (V.) naiffi was previously not reported in Colombia. It was originally described as a parasite of small mammals in the Amazonian area of Brazil (Lainson and Shaw 1989). Human infection with this species resulting in CL were reported in Brazil (Naiff et al. 1991; Felinto de Brito et al. 2012) but also in French Guiana, Martinique, Guadeloupe, Surinam, and more recently, in Ecuador (Grimaldi et al. 1991; Darie et al. 1995; Pratlong et al. 2002; Fouque et al. 2007; van der Snoek et al. 2009; Kato et al. 2013). Although sequencing is needed to verify this case, the potential presence of this species in the Colombian territory (Guaviare) must be highlighted, to foresee cases of mild cutaneous infections due to L. naiffi in remote areas of the country.

Whereas the number of samples analyzed from Honduras was very low, 9 out of 10 were identified as L. infantum (chagasi). This agrees with previous reports involving this species not only as the responsible for visceral leishmaniasis, but also causing cutaneous disease in Honduras and other Central-American countries (Grimaldi et al. 1987; Ponce et al. 1991; Noyes et al. 1997). In the case of Guatemala, only five samples were studied and four of them had patterns related to the L. mexicana complex and L. braziliensis, which have been recognized as the etiological agents of the disease (Herwaldt et al. 1992; Zeledón et al. 1993).

The here used RFLP protocols can be adapted to different settings, depending on the sensitivity required and the species that must be separated. As such, RFLP-F could be more useful for typing Leishmania spp. from Colombia if identification of all circulating species is needed. On the other hand, because of its sensitivity PCR-N could be the method of choice in Guatemala and Honduras, where less species are thought to circulate.

Many studies have documented differential response to a variety of treatments in relation to the species of Leishmania (Llanos-Cuentas et al. 2008; Obonaga et al. 2014). A recent global analysis recommended a species-related therapy for travelers coming from endemic areas, including those from Latin America (Hodiamont et al. 2014). Taking into account that tools based on hsp70 gene are, in parallel, among the most used and best validated typing assays (Van der Auwera and Dujardin 2015) we sustain that molecular tools applied here for typing New World Leishmania parasites could be very valuable for diagnostics, case management and therapeutic studies focused on differential response of treatments according to the species of Leishmania causing the disease.