Potential animal reservoirs (dogs and bats) of human visceral leishmaniasis due to Leishmania infantum in French Guiana

In French Guiana, cutaneous leishmaniasis is highly endemic, whereas no autochthonous case of visceral leishmaniasis have been reported so far. However, due to its proximity to Brazil which is highly endemic for visceral leishmaniasis, and the high transboundary population flow, an epidemiological challenge could arise at any time. As an overseas department and region and the largest outermost region of the European Union, epidemiological surveillance of visceral leishmaniasis is of great importance. Our study aimed to investigate the presence of Leishmania spp. in domestic (dogs) and sylvatic (bats) animals from French Guiana. Over the 2008–2018 period, samples from 349 animals were collected. They included blood from 179 autochthonous dogs and 59 bats, spleen samples from 33 bats and, blood from 78 military working dogs (MWD) collected before their departure from continental France and at the end of their four-month stay in French Guiana. Samples were screened using real-time polymerase chain reaction (qPCR) assays targeting Leishmania DNA followed by sequencing of 18S rRNA, kDNA and ITS2 genes. L. infantum was detected in 2.3% (8/349) of animals with 1.7% (3/179) of autochthonous dogs, 5.1% (4/78) of MWD returning from French Guiana, whereas they were negative before their departure. One of them dates back to 2012. All these dogs were positive for serological tests. In addition, L. infantum DNA was detectable in one bat spleen sample, belonging to Carollia perspicillata species. We report here for the first time an infection with L. infantum in dogs and bat from French Guiana. Our results suggest the existence of potential reservoir and transmission cycle for visceral leishmaniasis, at least since 2012, which was unknown in this territory until now. Further studies are needed to determine how these animals were infected and which vectors are involved in the transmission in this area.


Introduction
Leishmaniasis are among of the ''most neglected diseases" [1] from the group of vector-borne diseases. Leishmaniasis are caused by parasites belonging to the genus Leishmania (Trypanosomatida: Trypanosomatidae) with a worldwide distribution in large areas of the tropics, subtropics and Mediterranean basin, involving more than 98 countries [2]. The burden of leishmaniasis increased over the last decades, making them among tropical infections, the 2 nd and 4 th most common cause of death and disease, respectively [3]. Worldwide, the population of 350 million is at risk with an annual incidence of 0.7-1.2 million cases of cutaneous leishmaniasis (CL) and 0.2-0.4 million cases of visceral leishmaniasis (VL) [4]. VL is often caused by L. donovani and L. infantum (syn L. chagasi) and involves various mammals such as humans and dogs [3,5]. The parasite infects internal organs, such as spleen, liver and bone marrow. Dogs are the main reservoir and hosts of L. infantum causing canine leishmaniosis (CanL), a severe systemic disease reported in more than 70 countries and common in the Mediterranean region and in South America [5,6]. It is estimated that 2.5 million dogs are infected in the Mediterranean basin only [7]. Some VL cases, caused by L. tropica or L. amazonensis, have also been reported [8]. The main vectors belong to Euphlebotomus, Larroussius and Synphlebotomus sandfly subgenera in the Old World [9], and Lutzomyia subgenus in the New World [10,11]. Natural cycle of CL involves various vertebrate hosts (wild rodents and humans) and different sandfly species as vectors for its spread. Its etiological agents include L. tropica, L. major, and L. aethiopica transmitted especially by Phlebotomus and Paraphlebotomus sandflies in the Old World [12]. However, New World CL is caused by L. mexicana complex (L. mexicana, L. amazonensis, etc.) or the subgenus Viannia (L. braziliensis, L. guyanensis, etc.) and occurs especially in tropical and subtropical areas of Mexico and Central and South America. The main vectors are classified in the subgenera Nyssomyia, Psychodopygus, Lutzomyia s.str., and Verrucarum [11].
French Guiana is situated in the northern part of South America between Brazil and Suriname and extends on 84 000 km 2 . Its climate is equatorial (hot and humid) and the Amazonian forest covers 90% of its territory. Like the entire Amazon region, French Guiana hosts many wild and domestic animals that act as reservoirs for pathogens and the emergence of infectious diseases, amongst them, various zoonotic diseases. CL has been known here since 1943 [13] and cases have been reported regularly since then. The annual incidence rate was estimated between 15 and 20 new cases per 10,000 inhabitants between 1979 and 2012 [14,15]. L. guyanensis is undoubtedly the most common species, and other species implicated in CL has increased over the years: L. braziliensis, L. amazonensis, L. lainsoni, and L. naiffi [16,17]. Human and canine VL are widespread and are endemic in many areas of Latin America [18]. Although the principal foci of visceral leishmaniasis are located in the drier, poorly forested areas, a small number of human and canine infections have been recorded in the densely forested Amazon region, such as in the State of Para. In Roraima, a particularly large focus also extends to Venezuela and Guyana. In Venezuela, it occurs sporadically in almost every state of the country with a low endemicity. Till now, VL has been reported several times in Suriname [19] but never in French Guiana [20]. In 2006, two imported cases of canine VL in Cayenne, the main city of French Guiana, were reported; one infected dog was most probably imported from France. A second dog was then infected with Leishmania infantum in French Guiana [21].
Globally, Leishmania spp. natural infections have been repeatedly reported in domestic, peridomestic and wild animals, dogs and rodents being the most investigated animals and considered as reservoirs [22] The present study represents a ten-year surveillance (from 2008 to 2018) of the Leishmania circulation in dogs and bats in French Guiana using molecular and serological techniques.

Ethics statement
All dogs sampled in this study were examined with the assistance and acceptance of their owners. Blood samples were collected by veterinarians according to the good practices of veterinary medicine. Article R.214-88 of the French Rural Code and Sea Fishing (Decree No. 2013-118 of 1 February 2013 on the protection of animals used for scientific purposes) excludes these acts from the scope of applications for authorization granted by the Minister responsible for research. The protocol for capturing and sampling bat specimens (N˚1688) was approved by the Animal Ethics Committee of Marseille (C2EA14) and by the French authorities.

Animals
This study was conducted between 2008 and 2018. One hundred dogs were studied in Cayenne (4˚56 0 4.6@ N, 52˚19 0 49.19@ W), the capital of Guiana with 55,000 residents. Some 60 km northwest of Cayenne, in Kourou, we collected samples from 79 dogs (5˚9 0 34.92@ N, 52˚39 0 1.08@ W) (Fig 1). Dogs came from dog shelters from both cities and from private dog owners in Cayenne, who gave their consent. Overall, blood samples were taken from 179 adult dogs, all apparently healthy, including 107 females and 72 males, of which 26 were sampled in 2008, 55 in 2014 and 98 in 2016. In addition, between 2012 and 2018 period, pairs of blood samples have been collected from 78 military working dogs (MWD) before their departure from metropolitan France and after their return from a four-month stay in French Guiana. All of them were males and the majority was Berger Belge Malinois. Most of them stayed in Cayenne but some of them were involved in missions in the deep forest. We sampled 16 dogs in 2012, 26 in 2016, 20 in 2017 and 16 in 2018. All MWD were apparently healthy, except two of them: one (MWD1) had an ulcer on the hock and one (MWD2) had crusts at the elbows and skin lesions, after their return from French Guiana in 2018. For this last (MWD2), a skin swab and a bone marrow aspiration have been implemented for further Leishmania DNA detection. For dogs, blood collection was performed by cephalic venipuncture using EDTA tubes; all samples were stored at + 4˚C until being transported to our laboratory. Ethical aspects relating to dog sampling was made in accordance to the French law.

DNA extraction
We extracted DNA from 200 μL blood of dogs and bats after digestion with 15μL of proteinase K (20mg/mL) at +56˚C overnight. DNA from bat spleens was purified from approximately 25 mg of starting material mixed in 200 μL of buffer, 20 μL proteinase K and �20 mg of glass powder, followed by 30 sec lyse in FastPrep-24 device (Sample Preparation system) from MP Biomedicals, USA and incubation at +56˚C overnight. DNA was extracted using a commercial DNA extraction kit (QIAamp DNA Mini Kit, [Qiagen, Courtaboeuf, France]) on BIOROBOT EZ1 (Qiagen, Qiagen, Courtaboeuf, France) according to the manufacturer's instructions. It was then eluted in 200 μL of distilled water and stored at -20˚C under sterile conditions to farther use for PCR-based amplification.

Primers and probes
Leishmania DNA detection and species identification were performed by PCR/sequencing. In order to detect the Leishmania DNA, we performed the qPCR-based screening of samples with qPCR assay using primers and probe targeting a portion of small subunit of the 18S ribosomal gene (18S rRNA) to detect the presence of the genus Leishmania [2]. Positive samples were confirmed by a conventional PCR (PCR) targeting 550 bp fragment length from the same gene followed by sequencing, allowing the identification of Leishmania at the complex level (Medkour et al., submitted). Samples were also analyzed by qPCR using primers and probe for detection and quantification of L. infantum DNA with high sensitivity, targeting a conserved region of the kinetoplast minicircle DNA (kDNA) (several 1000's-fold repeated sequence), as described previously [29,30,31]. Positive samples were confirmed by PCRs/sequencing using primers RV1/RV2 targeting 140 pb of kDNA gene, primers LGITSF2/LGITSR2, targeting of 370 to 450 bp fragment of internal transcribed spacer 2 gene (ITS2); to identify Leishmania species. Primers and probes, their conditions and sources are listed in Table 1.

Polymerase chain reaction amplification, sequencing and phylogeny
For all DNA samples, qPCR assays based on 18S rRNA and kDNA were prepared in a final volume of 20 μL with 10 μL of Eurogentec Master Mix Roche (Eurogentec, Liège, Belgium), 3 μL of distilled water DNAse and RNAse free, 0.5 mM of each primer and 0.5 mM of the FAMlabeled probes (Table 1), 0.5 μL UDG and 5 μl of the DNA template. The amplification was performed in a CFX96 Real-Time system (BioRad Laboratories, Foster City, CA, USA) using the following thermal profile: one incubation step at 50˚C for two minutes and an initial denaturation step at 95˚C for three minutes, followed by 40 cycles of denaturation at 95˚C for 15 seconds and annealing extension at 60˚C for 30 seconds. DNA from cultured L. infantum and L. donovani were included as positive controls and master mixtures as a negative control for each assay. Results were considered positive when the cycle threshold (Ct) was lower than 38 Ct and 35 Ct for 18S rRNA-and kDNA-based qPCRs, respectively.
For the kDNA qPCR assay, the standard curve of amplification was 8 folds-serial dilution of plasmid of 10 8 copy of DNA/mL from the kDNA region, equivalent of 10000 parasites/mL; 5 μl of serial dilutions ranging from 10000 to 0.001 parasites/mL, was introduced into reaction tubes. For blood, results were expressed as the number of Leishmania parasites present in 1 ml of blood, taking in account the volume (200 μl of blood) and the elution (200 μl) introduced during the extraction process. For bone marrow, skin lesion and spleen, number of Leishmania LGITSR2 GGCCAACGCGAAGTTGAATTC parasites present in 1 g of tissue taking in account the quantity (20 mg environ) and the elution (200 μl) volume introduced during the extraction. Samples positive by qPCR were amplified in conventional PCR assays mentioned above (Table 1). PCR reactions contained 5 μl of DNA template, 25 μl AmpliTaq Gold 360 Master Mix from Applied Biosystems (Thermo Fisher Scientific), 1 μl of each primer and water to create a final reaction mixture volume of 50 μl. Amplifications were performed in a Peltier PTC-200 model thermal cycler (MJ Research Inc., Watertown, MA, USA) under the thermal cycling conditions: one incubation step at 95˚C for 15 minutes, 40 cycles of one minute at 95˚C, 30 seconds annealing at the corresponding temperature (Table 1) and one minute at 72˚C followed by a final extension for five minutes at 72˚C. Visualization was performed in electrophoresis on 2% agarose gels. Master mixture and DNA from cultured L. infantum and L. donovani were included as negative and positive controls, respectively, in each assay. Purification of PCR products was performed using NucleoFast 96 PCR plates (Macherey-Nagel EURL, Hoerdt, France) according to the manufacturer's instructions. The amplicons were sequenced using the Big Dye Terminator Cycle Sequencing Kit (Perkin Elmer Applied Biosystems, Foster City, CA) with an ABI automated sequencer (Applied Biosystems). The obtained electropherograms were assembled and edited using ChromasPro software (ChromasPro 1.7, Technelysium Pty Ltd., Tewantin, Australia) and compared with those available in the GenBank database by NCBI BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). All sequences were deposed in the GenBank database. Molecular Phylogenetic analysis was performed with MEGA version 7 software [34].

Leishmania antibodies detection
Leishmania antibodies were also tested on the PCR positive blood samples by Rapid immunomigration (RIM) using Witness Leishmania commercial test (Zoetis, Lyon, France), which uses an antigen from L. infantum to quickly identify antibodies in blood, sera or plasma from Leishmania-infected animals. When available, indirect immunofluorescence antibody test (IFAT) was performed for some MWD and Leishmania antibodies were then quantified [35].

Statistical analysis
Data were collected and described in XLSTAT version 2018.7.  Table 2). These three PCR positive samples were also found positive by serology using the RIM test (Table 3). In addition, four MWD (4/78, 5.1%) were diagnosed infected upon their return to France after a four-month stay in Guiana. One MWD was diagnosed in 2012 and another in 2016. In 2018, 2 other MWDs from the group of 16 dogs (12.5%) were infected during their stay in French Guiana ( Table 2). All the MWDs were negative for both PCR and serology before their stay in French Guiana. The two infected MWDs in 2018 showed presence of L. infantum antibodies by RIM and by IFAT at dilution 1/200 for one (MWD1) and 1/3600 for the other (MWD2). These two dogs presented signs of CanL (Fig 1), so, among the 7 positive dogs (PCR and/or serology), two were symptomatic and 5 asymptomatic. Symptomatic dogs were examined and treated by military veterinarians using the classical treatment (Glucantime + Allopurinol). Overall, the mean parasite load in dog's blood was 10921 (min: 3.7; max: 42390) parasites/mL. We detected 33,940 parasites/mL blood on the MWD1. In the sick MWD2, we detected: 4.210 4 parasites/mL of blood, 1.310 6 and 6.6 10 6 parasites/g from skin lesion and bone marrow, respectively (Table 3).

Molecular
Blood samples were collected from 59 bats in 2013, according to the following species distribution: 10 Eumops auripendulus, 8 Artibeus planirostris, 22 Noctilio albiventris, 2 Molossus barnesi, 14 Pteronotus parnellii, 1 Phyllostomus elongatus and 2 Carollia perspicillata. No blood samples taken from these species were found positive. In addition, 33 spleen samples were collected from Carollia perspicillata in 2014 and Leishmania DNA was detected in one of them   (3%, 1/33). The total prevalence of L. infantum infection in bats observed in this study was 1.1% (1/92) and the parasite load was 55.6 parasites/g in spleen. The infected bat, Carollia perspicillata (Fig 1), did not show any lesions and no apparent macroscopical lesions were found in its organs.

Sequencing
Of the 7 (6 from dogs and one from bat) amplicons obtained for the 18S rRNA, kDNA and ITS2 genes, 6 (5 from dogs and one from bat) had been sequenced. Sequencing analysis of these genes concludes for L. infantum species. The 18S rRNA obtained sequences of 520 to 537-bp fragment length were almost identical and had 99-100% similarity with L. donovani complex available sequences in GenBank database such as: L. donovani (a.n. GQ332356) and many other L. infantum/chagasi sequences: L. infantum JPCM5 isolate (a.n. XR 001203206) and L. chagasi C29 isolate (a.n. KT762398) (Supplementary Fig 1). Four sequences of 140-bp of the kDNA gene obtained from two local dogs and two MWDs were identical and showed 97% similarity with L. infantum MCAN/ES/98/10445 clone LinGpja 8 detected on a Spanish dog (a.n. EU437406.1). Further amplicons of 120-bp of the kDNA gene from a local dog and a bat showed 92% identity with L. chagasi AJS-PPECO isolate (a.n. Z35276) detected in a human in Brazil and 97% identity with L. infantum AJS-IPTRS isolate (a.n. Z35274) from a human in Tunisia, respectively (Fig 2). For the ITS2 gene sequencing, two amplicons of 390 from a local dog and a bat and two others of 420 bp from two local dogs were closely identical to each other and with many L. infantum/chagasi ITS2 gene published sequences in GenBank database, i.e. L. infantum (a.n. AJ634339) detected in a human in France and L. chagasi isolate 20 clone 1 (a. n. 1GU045591) from Brazil. Two MWDs ITS2 amplicons of 360 bp were identical and exhibited 99% with L. infantum/chagasi cited above (Fig 3).

Discussion
The Amazon rainforest, with luxurious vegetation and high rainfall (> 3 m / year) is an environment that promotes both multiplication of vectors and mammalian reservoir hosts. The demographic development of the population and the anthropisation of the environment facilitate the Leishmania cycle development. These features are found in almost all the Guianas. So far, two parasitic cycles of leishmaniasis have been described in French Guiana. L. guyanensis occurs in the canopy with the arboreal sand fly Lutzomyia umbratilis as a vector, the two-fingered sloth, Choloepus didactylus, as main reservoir and the opossum Didelphis marsupialis as secondary reservoir. The second cycle concerns especially L. amazonensis, evolving at the ground level with Lutzomyia flaviscutellata as vector and the spiny rat Proechymis cuvieri as reservoir [36]. In this study, and for the first time, we report the presence of L. infantum in dogs and bat in French Guiana, suggesting the existence of a potential new domestic and wild reservoir of VL in this area where no indigenous human cases have yet been reported. Our ten-year longitudinal study (2008-2018) showed L. infantum infection in surveyed animals. In French Guiana, literature reports no case of L. infantum infection since the imported canine case (from Spain) in 2006 [21]. According to the French reference center for leishmaniasis in 2017, an average of 180 cases of cutaneous leishmaniasis are reported in French Guiana each year: 85% due to L. guyanensis and 10% caused by L. braziliensis [19]. Leishmania species known so far in Guiana are Leishmania guyanensis, L. braziliensis, L. amazonensis, L. naiffi and L. lainsoni [21]. Another study conducted in French Guiana included 1017 new diagnosed cases of leishmaniasis records between 2006 and 2013 and showed 86.2% for L. guyanensis; 9.7% for L. braziliensis; 2.8% for L. amazonensis; and 1.3% for L. lainsoni [16]. Recently and for the first time in America, a case of human visceral leishmaniasis caused by Leishmania siamensis, acquired in Guyana, has been reported [37]. Another publication describes a clinical case in a cat from French Guiana with a nasal ulcer and ear nodules; molecular analysis identified L. braziliensis [38]. In our study, no L. infantum infection was detected in autochthonous French Guianan dogs sampled in 2008 and 2014, while 3.1% (3/98) were infected in 2016, all of them coming from Cayenne. It may suggest a recent emergence of this infection in this area. Dog is known to be the main reservoir of L. infantum, the causative agent of human and canine VL in South America [6]. The disease is also well known in the neighboring countries, such as Brazil, Suriname, Guyana, Venezuela, Colombia and Lutzomyia longipalpis is the main vector [39][40][41]. A study conducted in Northern Brazil, reports a prevalence of 13.1% in domestic dogs using IFAT [42].
Because French Guiana is a French territory and is part of the European Union, military dogs have come from metropolitan France (endemic territory for CanL) to Guyana for missions sometimes lasting a few months. During their stay in French Guiana, dogs can sometimes be involved in military missions taking place in the deep forest. As part of the follow-up  [43]. Consequently, the question of an emergence or of a newly recognized infection remains opened. Among infected dogs, only two MWDs presented clinical signs. CanL is a good example for the Iceberg phenomenon, since almost 50% of the affected canine population does not exhibit clinical signs [44]. Moreover, sick dogs manifest a variable and non-specific clinical spectrum [45], because CanL is a chronic and multisystemic disease that may potentially involve any organ [46]. In our study, apparently healthy infected dogs from Cayenne had a low or medium parasitemia (mean: 34 parasites/mL blood). By contrast, parasitemia in Leishmania-infected MWDs was low in two dogs and very high in two others, the higher parasitemia level being found in the sick dogs. It has been shown that dogs with medium to high parasitemia are sick or will eventually develop overt disease [47].
Bats are nocturnal and are the only mammals able to fly (sometimes seasonal migration), an important trait that can result in the dissemination of parasite species [48]. They are commonly infected by several trypanosomatid species, mainly from Trypanosoma genus: T. cruzi, T. vespertilionis, and T. (Megatrypanum) sp., among others [49] and by numerous Leishmania species [24]. It has been reported that bats, particularly numerous in French Guiana, are not reservoirs of Leishmania [50]. This may be due to the fact that French Guiana is a territory rich in large mammals, Leishmania parasites vectors are more attracted to these species than bats. In this study, 1.1% (1/92) of Guianan bats were infected. All blood samples from bats captured in 2013 were negative. L. infantum DNA was detected in a spleen of a short-tailed fruit bat Carollia perspicillata. To the best of our knowledge, there is only one report of the isolation of Leishmania (L. infantum) from the blood of this species, in Venezuela [51]. Two Leishmania species were identified in spleen and liver from Brazilian bats: Molossus molossus and Glossophaga soricina were found to be infected with L. infantum and L. amazonensis, and the latter was also found in Molossus rufus, Nyctinomops laticaudatus, Eumops glaucinus, E. auripendulus, Artibeus literatus, Sturnira lilium and Myotis nigricans [52]. Leishmania (Viannia) sp. was detected in a skin lesion from G.soricina and blood from M. molossus [53]. Recently, 6.4% (3/ 47) of Brazilian bats have been detected infected with L. brazilinensis, two Platyrrhinus lineatus and one Artibeus planirostris [54]. Leishmania spp. DNA was detected also in 8% of the 448 bat blood samples from a non-endemic region of leishmaniasis in Brazil, 41.6% of which were L. infantum, 38.9% L. amazonensis and 19.4% L. braziliensis [55]. Also, 9.8% (41/420) bats from Mexico were found infected by L. mexicana [56]. On the African continent, Leishmania kDNA was detected in 4.9% (8/163) of bats; Leishmania isolates from two bats were confirmed by ITS1 PCR to be L. tropica and L. major, isolated from two individual bats, Cardioderma cor and Nycteris hispida, respectively [57]. The low prevalence of Leishmania spp. infection in bats revealed in all these studies were in line with our investigation. By contrast, Shapiro et al. reported a 40% higher prevalence among bats from an endemic area of Brazil [54]. These results indicate that further studies are needed to assess the role of bats in maintaining the life cycle of leishmaniasis, especially in areas where these diseases are endemic.
The detection of the infection in autochthonous dogs, bats and in MWD in French Guiana suggests the possible existence of L. infantum-transmission cycle in this area. Many VL forms are asymptomatic in healthy humans, but symptomatic forms appear mainly in cases of malnutrition or the development of an immune deficiency (HIV in particular). Possible transmission cycle of L. infantum in French Guiana is therefore a threat for the Guyanese civil population, whose HIV prevalence is very high (1.0-1.5%) and is among the highest in the American continent [58]. Other investigations are required to fully understand the epidemiology.

Conclusions
The detection of L. infantum in local dogs and bats from French Guiana, and in initially noninfected dogs coming from metropolitan France, suggests the possible existence of an autochthonous transmission cycle for visceral leishmaniasis. It highlights the need for active surveillance in domestic and wild animals, especially the potential reservoirs identified in this study and implementation of control measures. Competent vectors in this region remain to be identified.
Supporting information S1 Fig. Phylogenetic tree constructed based on the sequences of 18S rRNA gene for isolates in this study and isolates of the Leishmania complex from GenBank database. Neighbor-joining tree was constructed from 18S rRNA partial gene using MEGA 7.0 software. The Kimura-2-parameter method was used. Numbers above branches correspond to bootstrap values based on 1,000 replicates. The analysis involved 37 nucleotide sequences. All positions containing gaps and missing data were eliminated. There were a total of 527 positions in the final dataset. Isolates were designated by their accession numbers in the beginning and their names.