Upscaling surveillance of tick-borne pathogens in the French Caribbean islands

Background Among hematophagous arthropods, ticks transmit the greater variety of pathogens of public health and veterinary importance whose (re)-emergence is recognized worldwide. However, the epidemiological situation of the Caribbean area with regard to tick-borne diseases is poorly documented, mainly focusing on livestock pathogens such as Ehrlichia ruminantium, Babesia (bovis and bigemina) and Anaplasma marginale. These observations underline the need to undertake new large-scale epidemiological surveys to better assess the distribution of tick-borne pathogens and anticipate the risk of (re)-emergence of tick-borne diseases in these areas. To ease and reduce the cost of such large-scale surveys, the development of a fast and cheap high-throughput detection technics is desirable. Methods In this study, we first implemented a high-throughput microfluidic real-time PCR (BioMark™ dynamic arrays, Fluidigm Corporation) adapted for the large-scale screening of tick-borne pathogens. The system developed here includes 57 designs allowing both the screening of bacteria and protozoans potentially circulating in the West Indies (5 bacterial genera, 30 bacterial species, 1 protozoan phylum, 2 protozoan genera and 17 protozoan species), and the molecular identification of three tick species mainly involved in tick-borne pathogens transmission in the Caribbean (Amblyomma variegatum, Rhipicephalus microplus and Rhipicephalus sanguineus sensu lato). Then, using the new high-throughput microfluidic real-time PCR system, we performed an exploratory epidemiological study on 132 specimens of Amblyomma variegatum and 446 Rhipicephalus microplus collected in Guadeloupe and Martinique. Results We successfully detected tick-borne pathogens expected to be present in the area – Ehrlichia ruminantium, Rickettsia africae, Anaplasma marginale, Babesia bigemina, Babesia bovis, Theileria velifera and Theileria mutans – as well as unsuspected pathogens and microorganisms belonging to the genera Anaplasma, Ehrlichia, Borrelia, and Leishmania. Conclusions We demonstrated the ability of the Biomark system to give a rapid overview of the pathogens/microorganisms diversity present in ticks, thus opening new research perspectives on tick-borne pathogens epidemiology in the Caribbean. Our study demonstrated how high-throughput microfluidic real-time PCR technology is a major improvement in large-scale epidemiological studies, giving a rapid overview of the tick-borne pathogens and microorganisms present in ticks in a given area.

TBPs in the Caribbean is based on serological studies in animals or humans, or molecular biology testing (PCR, nested PCR). These studies did not survey nor discriminate the whole repertoire of infectious agents since they were often limited to the detection of some well-known pathogens (

Tick collected in Guadeloupe and Martinique
The ticks used in this study were collected within the framework of two distinct epidemiological surveys conducted in Guadeloupe (between February 2014 and January 2015) and Martinique (between February and March 2015), respectively. In Guadeloupe, adult ticks (any species, any sex) were collected on 40 bovines originating from 22 different herds that were sampled in nine localities situated in six different biotopes (urban area, dry coastal regions, valleys and hills, evergreen seasonal forest, sub-mountainous rain forest, swamp forest). In Martinique, engorged females of R. microplus only were collected from cattle in 29 farms participating in a study on acaricide resistance in ticks. All ticks were collected from cattle with the permission from farmers and cattle owners. Ticks were morphologically identified at the species level (Walker et al., 2003). A total of 578 adult ticks were included in the study: 132 A. variegatum and 165 R. microplus from Guadeloupe and 281 R. microplus from Martinique (see maps, Fig 2). GPS coordinates of tick collection sites are available in Additional file 1. All the ticks were partially engorged, and then conserved at -80°C.

DNA extraction
For 20 mg of ticks, 1 mL of recently prepared PBS 1X was added to the sample, then ticks were washed by gentle shaking over 2-3 min at 7 Hz/s in a Tissue Lyzer (Qiagen, Germany). After discarding the supernatant, ticks were frozen at -80°C during 15-20 min. Then a steel ball was added and samples were crushed twice during 2 min at 30 Hz/s with the Tissue Lyzer (Qiagen, Germany). 450 µL of fresh PBS 1X were added to the samples. Samples were vortexed during 10 s, and then centrifuged during 2-3 min at 8000 g. Finally, 20 µL of Proteinase K were added to 180 µL of crushed tick sample and DNA was extracted using the NucleoSpin® 96 virus Core kit (Macherey-Nagel, Germany) and the automatic platform Biomek4000 (Beckman Coulter). This protocol allows the simultaneous extraction of both DNA and RNA. Total nucleic acid per sample was eluted in 160 µl of rehydration solution and stored at −80°C until further use. A. variegatum ticks were individually extracted. R. microplus ticks were extracted both individually and in pools of two to four adult specimens when too small to be treated individually.

Assay design
The list of pathogens to be monitored, the sets of primers and probes required for their detection, as well as the targeted genes are shown on Table 1. Some of the oligonucleotides were specifically designed for the purpose of this study; the others came from Michelet et al., 2014

Pre-amplification of DNA samples
All the DNA samples were subject to preamplification in order to enrich the pathogen DNA content comparatively to tick DNA. The Perfecta® PreAmp SuperMix (Quanta Biosciences, Beverly, USA) was used for DNA pre-amplification, following the manufacturer's instructions. All the primers were pooled (except the ones targeting the tick species) with a final and equal concentration of 45 nM each.
Pre-amplification reaction was performed in a final volume of 5 µL containing 1 µL of PerfecTa PreAmp SuperMix (5X), 1.25 µL of pooled primers mix, 1.25 µL of DNA and 1.5 µL of MilliQ water with one cycle at 95°C for 2 min, 14 cycles at 95°C for 10 s and 60°C for 3 min. At the end of the cycling program the reactions were diluted 1:10. Pre-amplified DNAs were stored at −20°C until use.

High-throughput microfluidic real-time PCR
High-throughput microfluidic real-time PCR amplifications were performed using the BioMark™ realtime PCR system (Fluidigm, USA) and 96.96 dynamic arrays (Fluidigm, USA), that allows to perform up to 9,216 individual reactions in one run (Michelet et al., 2014). Real-time PCRs were performed using 6-carboxyfluorescein (FAM) and black hole quencher (BHQ1)-labeled TaqMan probes with TaqMan Gene Expression Master Mix (Applied Biosystems, USA) following the manufacturer's instructions. Cycling conditions were as follows: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 2-step amplification of 15 s at 95°C, and 1 min at 60°C. The BioMark™ Real-Time PCR system was used for data acquisition and the Fluidigm Real-time PCR Analysis software for Ct values determination.
Three kind of control were used per chip for experiment validation: a negative water control to exclude contaminations; a DNA extraction control, corresponding to primers and probe targeting a portion of the 16S rRNA gene of ticks; and an internal control to check the presence of PCR inhibitors made of a DNA of Escherichia coli strain EDL933 added to each sample with specific primers and probe targeting the eae gene (Nielsen and Andersen, 2003).

Infection rates of ticks from the French Antilles
According to the tick species and the island of origin, for each detected pathogen, infection rates (the proportion of infected ticks divided by the total number of ticks analyzed) were calculated. The majority of the samples were single specimens of ticks. When ticks were too small to be treated individually, they were pooled by two to four specimens. Thus, among the 523 samples analyzed, 47 consisted of a pool of two to four tick specimens. The final estimation of the infection rate also includes the pools and is therefore expressed as the minimal (assuming at least one positive ticks in the pools) and maximal (assuming all ticks positive in the pools) proportions of infected ticks out of the total number of ticks analyzed.

PCRs and sequencing for results confirmation
Conventional PCR/Nested PCR using primers targeting different genes or regions than those of the BioMark™ system were used to confirm the presence of pathogenic DNA in some field samples and positive controls (Table 2). Amplicons were sequenced by Eurofins MWG Operon (BIOMNIS-EUROFINS GENOMICS, France), and then assembled using BioEdit software (Ibis Biosciences, Carlsbad, CA, USA). An online BLAST (National Center for Biotechnology information) was used to compare results with published sequences listed in GenBank sequence databases.

Phylogenetic sequence analysis
Alignments were performed using ClustalW (Thompson et al., 1994). Maximum Likelihood trees were generated by 1,000 bootstrap repetitions under the Tamura-Nei model (Tamura and Nei, 1993

Results
Implementation of the high-throughput microfluidic real-time PCR system for tickborne pathogens survey The high-throughput microfluidic real-time PCR system developed for the screening of known and potential TBPs included 61 sets of primers and probes. Among them, 49 designs were developed for the detection of bacterial (n=32) and protozoan (n=17) species, bacterial (n=5) and protozoan genera/phylum (n=3) ( Table 1). Three sets of primers and probes were developed for the molecular identification of the three tick species found in the Caribbean, including A. variegatum, R. microplus and R. sanguineus sensu lato( Table 1). Finally, we developed a design targeting a conserved region of the 16S rRNA genes within ticks, called Tick spp., used as a control of DNA/RNA extraction ( Table 1).
The detection ability of each designs and the effect of pre-amplification on detection signals were first checked by Taqman real-time PCRs on a LightCycler 480 apparatus using dilution range of positives controls (Table 1). Three kind of positive controls were used, including bacterial or protozoan cultures when available, DNA from infected ticks or blood samples, and plasmidic constructions as a last resort (Table 1). Except the design targeting Borrelia burgdorferi sensu stricto which never succeeded in detecting the positive controls even after a preamplification step, the remaining 60 designs targeting TBPs and ticks species were able to detect their target with Ct values between 6 to 38 (data not showed). Pre-amplification improved the quality of the detection and was therefore validated as part of the screening protocol (see Additional file 2).
Then, the specificity of the designs was evaluated using the Biomark system and a total of 62 positive controls (Table 1, Fig 1).

Figure 1. BioMark TM dynamic array system specificity test (96.96 chip).
Each square corresponds to a single real-time PCR reaction, where rows indicate the pathogen in the sample and columns represent the target of the primers/probe set. Ct values for each reaction are represented by a color gradient; the color scale is shown on the right y-axis. The darkest shades of blue and black squares are considered as negative reactions with Ct > 30.
42 primers/probe sets were able to detect and amplify specifically their target using a Ct cut-off value of 30 and were then directly validated (Fig 1). The design Tick spp., used as tick acid nucleic extraction control, was able to detect A. variegatum and R. sanguineus s.l. samples, as well as the DNA of R. sanguineus s.l. tick present in the Rickettsia conorii positive control as expected (Fig 1). However the DNA of ticks from the R. microplus control sample and others positive controls including tick DNA (such as the controls Borrelia lonestari, Anaplasma phagocytophilum, etc.) were not detected (Fig 1).
The detection ability of this designed has been corrected by the addition of the Tick spp. primers during the preamplification step, which were initially excluded since the objective was to enrich pathogen DNA content comparatively to tick DNA (data not shown).
The remaining designs gave false-positive results in outgroup controls in addition to their target detection. Interestingly, two kind of unsuspected signals were observed, some related to cross-reactions with closely related species and some related to potential co-infections in controls corresponding to field samples (Fig 1). Eight designs -Rickettsia massiliae, Rickettsia conorii, Bartonella henselae, Bartonella bacilliformis, Babesia canis vogeli, Babesia microti, Theileria parva, Hepatozoon americanum -gave positive results in outgroup controls, revealing cross-reactions with one to two closely related species (Fig 1). Babesia ovis and Rickettsia rickettsii designs gave multiple cross-reactions with closely related species or distant outgroups and thus were considered as non-specific and removed from the rest of the study (Fig 1). Finally, the seven remaining designs -Rickettsia spp., Rickettsia felis, Rickettsia africae, Apicomplexa, Babesia bigemina, Hepatozoon spp., Hepatozoon canis -gave positive results in outgroup controls linked to potential co-infection in controls corresponding to DNA from ticks collected in the field or DNA from infected blood (Fig 1). As co-infections may occur in natural tick or blood samples, these unexpected detections in biological samples could be due to natural (co-)occurrence of microorganisms rather than cross-reactions (see details of co-infection analysis in Additional file 3).
To conclude, with the exception of the sets of primers and probes targeting Borrelia burgdorferi sensu stricto, Babesia ovis and Rickettsia rickettsii that were finally removed from the study, the 58 remaining designs were validated for the high-throughput screening of pathogens in Caribbean ticks, taking into account the notified cross-reactions.

Large scale TBPs detection survey in ticks from Guadeloupe and Martinique
A total of 578 adults ticks were collected from cattle in Guadeloupe and Martinique. A total of 523 were tested using the BioMark TM system developed in this study. The number of positive ticks and the corresponding infection rates for each pathogen detected were calculated for 132 A. variegatum, 165 and 281 R. microplus from Guadeloupe and Martinique, respectively ( Fig 2). As some of the R.
microplus samples corresponded to pool of two to four adult specimens, we reported the minimum and maximum infection rates (see Material and methods). Conventional PCR/Nested PCR followed by amplicon sequencing were performed on several tick samples to confirm some of the results of the newly designed BioMark TM system ( Table 2).
Percentage of identity of the sequences obtained with reference sequences available in NCBI are presented in Table 3.  Fig 2).
Rickettsia spp. were only detected in ticks collected in Guadeloupe (Fig 2). R. africae was identified in 95.6% of the A. variegatum samples (Fig 2). In contrast, Rickettsia spp. detected in 15.7-23.5% of the R. microplus samples from Guadeloupe were not directly identified as R. africae with the Biomark TM system (Fig 2). Thus, fourteen samples of A. variegatum (6/14) and R. microplus (8/14) positive for Rickettsia spp. were tested by nested PCR with primers targeting the ompB gene, followed by sequencing ( Table 2). All the sequences recovered were identical, and displayed 100% of identity with R. africae ( Table 3). The consensus sequence was deposited under the name Rickettsia africae Tick208 (accession number MK049851).
E. ruminantium was identified in 5.1% of the A. variegatum ticks from Guadeloupe (Fig 2). We confirmed the presence of E. ruminantium nucleic acids by testing one sample of A. variegatum by conventional PCR targeting the 16S rRNA genes followed by amplicon sequencing (  Fig 3). An. marginale was identified in R. microplus ticks from the both islands, with infection rates reaching 3.6-4.8% and 39.5-41.3% of specimens from Guadeloupe and Martinique, respectively ( Fig   2). We confirmed the detection of An. marginale by testing two samples of R. microplus by conventional PCR targeting the 16S rRNA genes followed by amplicon sequencing (  4). Guadeloupe and Martinique respectively (Fig 2). B. bovis was only detected in tick from Martinique, with an infection rate of 0.7% in R. microplus samples (Fig 2). As conventional and nested PCR did not succeed in detecting those parasites, we directly sequenced amplicons obtained with the B. bigemina and B. bovis designs developed here and corresponding sequences were identified (accession numbers MK071738 and MK071739 respectively) ( Table 3).
T. velifera and T. mutans were detected in both tick species and in the two islands. Detection of unexpected microorganisms in Caribbean ticks. Unexpected signals were obtained during the screening of microorganisms in ticks from Guadeloupe and Martinique, including the first detection of untargeted species belonging to the genera Anaplasma, Ehrlichia, Borrelia and Leishmania (Fig 2).
Ehrlichia spp. were detected in R. microplus ticks from both islands, with infections rates reaching 4.2-6.6% and 47.7-49.1% in Guadeloupe and Martinique, respectively (Fig 2). We tested two of the Ehrlichia spp. positive R. microplus samples by conventional PCR targeting the 16S rRNA genes (Table 2). We obtained two identical sequences, deposited under the name Ehrlichia sp. Tick428 (accession number MK049849) ( Table 3). Phylogenetic and genetic distance analyses were performed using fragment of the 16S rRNA genes of several Ehrlichia species (Fig 3). The Ehrlichia sp. Tick428 sequence was found within a cluster including various uncharacterized Ehrlichia species detected in ticks from Asia and Africa.
In addition, in around 50% and 18% of the R. microplus specimens positive for Anaplasma spp., none of the Anaplasma species targeted by the Biomark TM system gave signals, suggesting the presence of an unexpected or a new Anaplasma spp. (Fig 2). We tested two of the Anaplasma spp. positive R.
microplus samples by conventional PCR targeting the 16S rRNA genes ( Table 2). We obtained two identical sequences, deposited under the name Anaplasma sp. Tick314 (accession number MK049845) (Table 3). This sequence displayed 100% of sequence identity with Candidatus Anaplasma boleense.
Phylogenetic and genetic distance analyses were performed using fragment of the 16S rRNA genes of several Anaplasma species (Fig 4). Anaplasma sp. Tick314 sequence was found in a cluster including Candidatus Anaplasma boleense, Anaplasma platys and Anaplasma phagocytophilum.
Borrelia spp. was detected in both tick species from both islands (Fig 2). Infection rates reached 5.1% in A. variegatum, and 0.6% and 4.3% in R. microplus from Guadeloupe and Martinique, respectively (Fig 2). None of the specific Borrelia species targeted causing Lyme disease (Borrelia burgdorferi sensu lato) or Borrelia Relapsing Fever group, gave any positive results suggesting the occurrence of a new or unexpected Borrelia spp. in our samples (Fig 2). We tested 30 of the Borrelia spp. positive ticks by nested PCR targeting the flaB genes (  (Table 3). Phylogenetic and genetic distance analyses were performed using fragment of the flaB gene of several Borrelia species (Fig 5). Surprisingly, the Borrelia sp. Tick7 sequence recovered from the A. variegatum sample, and found closely related to Bo. anserina, displayed an intermediate position, sharing homology with both the relapsing fever and Lyme disease groups (Fig 5). Finally, the Borrelia sp. Tick457 sequence recovered from the R. microplus samples confirmed the previous observations, forming a cluster with various Relapsing Fever Borrelia species encountered in hard ticks, including Bo. lonestari and Bo. theileri (Fig 5). Finally, 0.7% of R. microplus ticks from Martinique were positive for Leishmania spp. (Fig 2).
We tested two of the Leishmania spp. positive ticks by nested PCR targeting the small sub-unit rRNA gene ( Table 2). We obtained one sequence from one sample, deposited under the name Leishmania martiniquensis Tick389 (accession number MK049850) ( Table 3). This sequence displayed 100% of identity with both Leishmania martiniquensis and Leishmania siamensis sequences (Table 3).

Discussion
In this study, a high-throughput microfluidic real-time PCR system based on the use of multiple primers/probes was developed for large scale surveys of bacteria and protozoans potentially transmitted by ticks from the Caribbean area.
The association of genus and species primer/probe designs targeting TBPs improved the screening capacity of the technology, allowing not only the identification of infectious agents known to circulate in the studied area, but also the detection of unsuspected TBPs or new microorganisms belonging to the main bacterial and protozoan genera/phylum involved in TBD worldwide.
Nevertheless, as some endosymbiotic microorganisms could belong to known TBPs genera, such as While analyzing the specificity of the microfluidic real-time PCR system, cross-reactions have been observed for some designs targeting closely related species that must be taken into account when interpreting results. Due to high design constraints and a lack of available sequences in public databases, the improvement of such cross-reacting oligonucleotides remain challenging. Here, the concomitant use of the bacterial and protozoan genera can assist non-specific signals identification.
In addition to microorganism's detection, we developed sets of primers and probes allowing the  africae in R. microplus may be due to bacteria circulating in cattle blood which would have been picked up by engorging ticks, or to a cross-contamination during R. microplus ticks co-feeding next to infected

Conclusion
Our study demonstrated the high ability of microfluidic real time PCR technology to give a rapid overview of the diversity of TBPs of veterinary and medical importance present in ticks from the Caribbean. This innovative high-throughput tool is promising and could improve significantly TBPs surveillance and study, allowing a rapid screening of multiple especially in regions where few epidemiological data are available and TBD are numerous.

Declarations Ethics approval and consent to participate
Not applicable

Consent for publication
Not applicable

Availability of data and material
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The