Development of a Multiplex Droplet Digital PCR Assay for the Detection of Babesia, Bartonella, and Borrelia Species

We describe the development, optimization, and validation of a multiplex droplet digital PCR (ddPCR) assay for the simultaneous detection of Babesia, Bartonella, and Borrelia spp. DNA from several sample matrices, including clinical blood samples from animals and humans, vectors, in-vitro infected human and animal cell lines, and tissues obtained from animal models (infected with Bartonella and/or B. burgdorferi). The multiplex ddPCR assay was able to detect 31 Bartonella, 13 Borrelia, and 24 Babesia species, including Theileria equi, T. cervi, and Cytauxzoon felis. No amplification of Treponema or Leptospira spp. was observed. Sensitivity of 0.2–5 genome equivalent DNA copies per microliter was achieved for different members of the Bartonella and Borrelia genus, depending on the species or matrix type (water or spiked blood DNA) tested. The ddPCR assay facilitated the simultaneous detection of co-infections with two and three vector-borne pathogens comprising four different genera (Babesia, Bartonella, Borrelia, and Theileria) from clinical and other sample sources.

Due to characteristic ddPCR sample partitioning, this new technology provides key advantages over conventional and real-time PCR (qPCR): including increased sensitivity, unparalleled precision, and simplified absolute quantification (does not require the use of a standard curve). In addition, interference from inhibitory substances that may be present in different sample matrices is markedly reduced when using ddPCR as compared to qPCR [18]. Collectively, these features make this methodology an attractive alternative for the simultaneous amplification of vector-borne pathogen DNA, including vector-borne members of the genera Babesia, Bartonella, and Borrelia.
Babesia spp. are intraerythrocytic protozoan parasites of longstanding historical importance in human and veterinary medicine, which in recent years have been increasingly recognized as a cause of infectious disease in humans. Historically, bovine babesiosis was the very first disease proven to be tick-transmitted. Our research group was among the first to develop diagnostic PCR assays for documentation of Babesia spp. DNA in animal blood specimens [19][20][21][22]. Transmission of Babesia spp. occurs primarily by the bite of infected Ixodid ticks; however, other routes of transmission, including vertical (mother-child) and blood transfusion transmission have been reported [23][24][25]. Babesia microti is the causative agent for the majority of human babesiosis cases reported in the fastidious nature, complex growth requirements, cyclical, relapsing low bacteremia, and their ability to invade several cells types to subvert/evade the immune system (often leading to long delays in seroconversion and negative serology test results) [87][88][89][90][91][92][93][94][95][96][97][98], specialized diagnostic modalities, including a recently described Bartonella droplet digital PCR detection assay, are critically needed to improve diagnostic sensitivity [17,18,99].
We describe the development of a multiplex droplet digital PCR assay for the simultaneous detection of Babesia, Bartonella, and Borrelia species (BBB ddPCR) using the Bio-Rad QX One Droplet Digital PCR system. The QX ONE Droplet Digital PCR System integrates droplet generation, thermal cycling, droplet reading, and analysis into a single automated precision platform. The QX One improves upon the QX200 by providing an automated, high throughput testing platform, capable of simultaneous amplification of up to eight DNA/RNA targets with the same superior accuracy and absolute quantification achievable with qPCR. Several sample matrices were tested, including experimentally infected human and animal cell lines, spiked blood samples, animal (both domestic and wildlife) and human pre-characterized clinical samples (blood and tissue), and naturally infected sand-fly and tick species.

Sample Reference Types and DNA Extraction
DNA from previously characterized or as yet to be characterized Babesia (Piroplasma), Bartonella, and Borrelia spp., including positive and negative research and diagnostic samples from various host animals and humans submitted to the Vector-Borne Diseases Diagnostic Laboratory (VBDDL) and the Intracellular Pathogens Research Laboratory (IPRL), both at the College of Veterinary Medicine, North Carolina State University, were used to test the specificity of this BBB ddPCR assay. In addition to clinical human and animal blood and tissue samples, in vitro canine histiocytic and human epithelial experimentally infected cell lines, experimentally infected mouse tissues, naturally infected vectors (including Ixodes spp. ticks and sandflies), and spiked naïve human and dog blood DNA samples were tested to assess specificity and sensitivity of the assay. A descriptive list of host species and sample types are detailed for each pathogen group (Table 1: Babesia, Theileria, and Cytauxzoon; Table 2: Bartonella spp.; Table 3: Borrelia spp.; Table 4: Co-infections). Although our primary focus was ddPCR amplification of Babesia (Piroplasma), Bartonella, and Borrelia spp., the Piroplasma probe was designed to also amplify Theileria and Cytauxzoon spp., which are of veterinary medical importance.

Mammalian Cell Line In Vitro Infection
The ddPCR detection of intracellular Bartonella spp. or Borrelia spp. was performed on DNA extracted from experimentally infected canine histiocytic (DH82) and human epithelial (MCF10A) cell-lines cultures. Briefly, confluent DH82 (kindly provided by Henry Marr, College of Veterinary Medicine, North Carolina State University) and MCF10A (kindly provided by Dr. David Alcorta, Department of Pharmacology and Cancer Biology, Duke University, Durham, North Carolina, USA) cells were infected in vitro with either Bartonella henselae San Antonio 2 (strain B. henselae SA2) or B. burgdorferi clonal strain B31-MI 16 at a multiplicity of infection (MOI) of 10:1 and incubated at 37 • C with 5% CO 2 for 24-72 h. Prior to DNA extraction, cells were subjected to a gentamicin elimination assay to eliminate extracellular bacteria [100]. Briefly, cells were treated with 150 µg/mL gentamicin for 4 h at 37 • C with 5% CO 2 and gently washed three times with sterile PBS. Cells were harvested from each well via pipetting up and down following the addition of 1 mL of ice-cold PBS-5mM EDTA and 15-min incubation at 4 C. Cellular DNA was extracted, as described below, at 24, 48, and 72 h post-infection.

Experimentally Infected Mouse, Rabbit and Hamster Tissue
DNA of skin, heart, and ear tissues from experimentally and uninfected infected mice [101], kindly provided by Dr. Spector research group (Department of Pharmacology and Cancer Biology, Duke University, Durham, NC, USA), were used to assess the efficiency of Borrelia burgdorferi B31 detection. Similarly, blood samples collected from experimentally infected rodents, kindly provided by Dr. Sam Telford from the Department of Infectious Disease and Global Health, Tufts University, were used to assess the detection of Babesia divergens (rabbits), B. duncanii (hamsters), and B. microti (hamsters).

Spiked Blood Samples
Naïve human and dog blood was spiked with different Bartonella (Table 2) and Borrelia species (Table 3) at a final concentration of 5 genome copies per microliter for each studied pathogen. To assess the simultaneous detection of co-infections, Babesia gibsonii naturally infected dog blood samples were spiked with Bartonella spp. (B. henselae SA2, B. quintana, B. vinsonii subsp. berkhoffii TI) and/or Borrelia burgdorferi B31 DNA (Table 4). Similarly, naïve human blood samples spiked with B. henselae SA2, B. burgdorferi B31, and/or with B. microti DNA, were used to assess the detection of any combination of the three pathogens (Table 4).

Previously Characterized Borrelia spp. DNA Samples
A total of 94 samples, kindly provided by Drs. Volker Fingerle and Reinhard Straubinger (Laboratory Medicine, Region Jönköping County, Jönköping, Sweden) containing DNA extracted from 11 cultured Borrelia species at concentrations ranging from 0.1 to 2000 genome equivalents per microliter of sample (Table 5) were tested to assess ddPCR assay specificity and sensitivity. This panel was previously used for an analytical comparison of real-time PCR protocols from five different Scandinavian laboratories [102].

DNA Extraction
DNA extraction from all sample types was performed using either the Qiagen DNeasy Kit (for tissue DNA extraction) or QIAsymphony ® SP robot (QIAGEN, Valencia, CA, USA) and QIAsymphony ® DNA Mini Kit (for blood and blood culture DNA extractions), per the manufacturer's recommended protocols. DNA quality and concentration for blood, tissue, and spiked blood samples was assessed by 260/280 nm OD measurement, as described previously [18].

Primers and Probes for Bartonella, Borrelia, and Babesia DNA Detection
Bartonella DNA detection, targeting a 90-120 bp segment of the intergenic spacer region (ITS) located between the 16SrRNA-23SrRNA, was performed as described previously [18]. Briefly, oligonucleotides BsppITS325s and BsppITS543as were used as sense and antisense primers respectively, and oligonucleotide BsppIT500 was used as the fluorescent probe for detection of Bartonella DNA amplification within channel 1 of the QX One Droplet Digital PCR system (Table 6). Table 6. ddPCR primers and probes for Babesia, Bartonella, and Borrelia amplification and detection.
Babesia DNA detection, targeting a 125-138 bp segment (depending on species) of the 18SrRNA gene, was performed using oligonucleotides Piro18S-238s and Piro18S-380as as forward and reverse primers, respectively (Table 6). Oligonucleotide Piro18S-340 was used as the fluorescent probe for Babesia (Piroplasma) DNA detection within channel 3.

DNA Amplification
The ddPCR reaction consisted of a 20 µL final PCR reaction consisting of: 5 µL of ddPCR™ Multiplex Supermix (Bio-Rad, Hercules, CA, USA); 0.2 µL of 100 µM of each Taqman probe, forward, and reverse primer, (IDT ® DNA Technology, Coralville, IA, USA); 7.5 µL of molecular-grade water; 5 µL of sample DNA; 0.27 µL of 300 mM Dithiothreitol (DTT); and 1 µL of HindIII DNA restriction enzyme. The ddPCR analysis was performed using a QX One Droplet Digital PCR (Bio-Rad, Hercules, CA, USA) system under the following amplification conditions: a single hot-start cycle at 95 • C for 10 min followed by 40 cycles of denaturing at 94 • C for 30 s and annealing at 62.9 • C for 1 min. A final extension at 98 • C was performed for 5 min. Fluorescent droplet detection and distribution readings were recorded in channel 1 (for Bartonella DNA), channel 2 (for Borrelia DNA), channel 3 (for Babesia DNA), and channel 4 (for housekeeping DNA). Naïve blood and tissue DNA from dogs, cats, humans, and vectors (sand-flies, ticks) were used as negative controls and were assayed alongside the positive control samples within each run.

Naive Blood, Tissue and In Vitro Cultivated Samples
Negative controls tested to assess assay specificity included over 250 pre-characterized uninfected mammalian cell culture (DH82 nor MCF10A cells) samples, uninfected human and dog blood and blood culture samples, and animal tissues. Bartonella (channel 1), Borrelia (channel 2), or Babesia (channel 3) DNA was not amplified from any sample. DNA amplification of the corresponding housekeeping gene (HMBS for humans, BRAF for animals) within channel 4 occurred in all uninfected-host samples (naïve cell-lines, clinical blood samples, or tissues from negative control animals). Housekeeping DNA was not amplified from the DNA extraction mixture or the negative water control (data not shown).

Detection of Babesia, Theileria and Cytauxzoon Species
Clinical Samples from Naturally and Experimentally Infected Animals When testing pre-characterized Piroplasma positive DNA within animal samples (blood from naturally and experimentally infected animals, Table 1), all 23 Babesia species (including nine as yet to be characterized species) [20], T. equi, T. cervi, and C. felis spp. DNA were amplified ( Table 7). Examples of reference signals of Babesia species DNA amplification (Channel 3) in naturally infected blood samples from dogs and other animals are represented in Figure 1.

Detection of Babesia, Theileria and Cytauxzoon Species
Clinical Samples from Naturally and Experimentally Infected Animals When testing pre-characterized Piroplasma positive DNA within animal samples (blood from naturally and experimentally infected animals, Table 1), all 23 Babesia species (including nine as yet to be characterized species) [20], T. equi, T. cervi, and C. felis spp. DNA were amplified ( Table 7). Examples of reference signals of Babesia species DNA amplification (Channel 3) in naturally infected blood samples from dogs and other animals are represented in Figure 1.

Detection of Babesia, Theileria and Cytauxzoon Species
Clinical Samples from Naturally and Experimentally Infected Animals When testing pre-characterized Piroplasma positive DNA within animal samples (blood from naturally and experimentally infected animals, Table 1), all 23 Babesia species (including nine as yet to be characterized species) [20], T. equi, T. cervi, and C. felis spp. DNA were amplified ( Table 7). Examples of reference signals of Babesia species DNA amplification (Channel 3) in naturally infected blood samples from dogs and other animals are represented in Figure 1.

Detection of Bartonella spp.
Similar to the previously reported detection using the Bio-Rad QX200 system [18], Bartonella DNA was amplified from a wide variety of sample matrices when the BBB ddPCR assay was conducted on the Bio-Rad QX One Droplet Digital PCR system. Bartonella DNA was detected within samples from naturally infected mammals ( Figure 2) and from experimentally infected cell lines (dog histiocytic DH82 and human epithelial MCF10A cells), spiked naïve human and dog blood samples, and from sand-fly and tick vector tissue (results not shown). A total of 31 Bartonella spp. from pre-characterized human and animal clinical cases [18,[103][104][105][106][107][108][109][110][111][112][113][114][115][116] were amplified using the QX-one BBB ddPCR assay described in this study (Table 8).

Detection of Bartonella spp.
Similar to the previously reported detection using the Bio-Rad QX200 system [18], Bartonella DNA was amplified from a wide variety of sample matrices when the BBB ddPCR assay was conducted on the Bio-Rad QX One Droplet Digital PCR system. Bartonella DNA was detected within samples from naturally infected mammals ( Figure 2) and from experimentally infected cell lines (dog histiocytic DH82 and human epithelial MCF10A cells), spiked naïve human and dog blood samples, and from sand-fly and tick vector tissue (results not shown). A total of 31 Bartonella spp. from pre-characterized human and animal clinical cases [18,[103][104][105][106][107][108][109][110][111][112][113][114][115][116] were amplified using the QX-one BBB ddPCR assay described in this study (Table 8).

Detection of Borrelia spp.
A total of 13 Borrelia species (Table 9) [120][121][122]; and B. turcica, a member of the Borrelia "reptile group" [123].  (Table 1) was amplified when spiked into naïve dog or human blood, and they were compatible with the one found previously reported [18]. Amplification detection limits ranged between 0. Similarly, Borrelia DNA from seven species (Table 1) was amplified when spiked into naïve dog or human blood (Figure 3). Amplification was not observed when naïve blood DNA or water were used as templates (results not shown). The level of detection was dependent upon the Borrelia sp. tested. DNA from B. burgdorferi B31 and B. hermsii were detectable at 0.5 genome equivalents per microliter, whereas 5 genome copies per microliter was the lowest detectable level for B. coriaceae, B. turicatae, B. garinii, B afzelii, and B. lusitaniae. The level of detection for B. bissetii was not assessed within spiked blood, as isolates from this species were unavailable.

Previously Characterized Borrelia spp. DNA Samples
DNA from 12 Borrelia species previously used in a multi-laboratory validation panel ( Table 5) that was designed to comparatively assess analytical performance of real-time PCR protocols from five different Scandinavian laboratories [102] was amplified using the BBB ddPCR assay. Detection limits of DNA spiked in water varied from 0.2 to 2 genome copy equivalents per microliter of sample (Table 10) depending on the species and strain tested. No Borrelia DNA was detected in negative control samples (blood samples, tissues, cell-lines, or water) nor when other non-Borrelia spirochetal DNA, such as from T. phagedenis or Leptospira at concentrations of 2000 spirochetes per microliter [102], were assayed at the same time, under the same conditions.

Tissues from Experimentally Infected Mice and Cell Lines and from Naturally Infected Vectors
Borrelia spp. DNA was amplified from skin, heart, and ear tissues obtained from B. burgdorferi B31 experimentally infected mice [101] but not from tissues samples from negative control animals. Similarly, Borrelia spp. DNA was amplified within intracellular fractions of experimentally infected human (MCF10A) and dog (DH82) cell lines but not within uninfected cells processed in an identical manner, and at the same time points (results not shown).
Borrelia spp. DNA was also amplified from naturally infected North Carolina ticks previously characterized in a separate study from our laboratory [124]. Ticks from that study assayed herein included a single I. scapularis tick infected with B. burgdorferi and single I. affinis ticks infected with B. burgdorferi, B. bissettii, or both Borrelia species [124]) (Figure 3b).

Dual Detection of Bartonella and Borrelia spp.
Simultaneous amplification of Borrelia and Bartonella DNA was observed for all species (B. burgdorferi B31, B. henselae, B. quintana and B.v. berkhoffii genotype II) in each respective channel when the DNA of each target organism was spiked into naive human blood samples at a concentration of 50 genome copies per microliter (Figure 4). Simultaneous detection of both Bartonella and Borrelia spp. DNA was also achieved with concentrations down to 0.01 pg/µL for each organism (equivalent to 5 bacteria genome copies per microliter of blood). Although the limit of detection for Bartonella DNA within dual spiked samples was 0.5 genome copies per microliter, as previously reported for spiked samples containing only Bartonella DNA [125], amplification of Borrelia DNA was not achievable within dual spiked samples at 0.5 genome copies per microliter (results not shown). Simultaneous Borrelia and Bartonella DNA amplification was observed within previously characterized naturally co-infected ticks [124,126,127] (results not shown).

Dual Detection of Bartonella and Babesia
The BBB ddPCR assay was able to detect natural, dual infection with B. vinsonii and Babesia vulpes within blood samples obtained from gray foxes from Portugal ( Figure 5). Consistent mean fluorescent intensities and droplet distribution patterns were obtained for each pathogen within their respective channels.

Dual Detection of Bartonella and Babesia
The BBB ddPCR assay was able to detect natural, dual infection with B. vinsonii and Babesia vulpes within blood samples obtained from gray foxes from Portugal ( Figure 5). Consistent mean fluorescent intensities and droplet distribution patterns were obtained for each pathogen within their respective channels.

Dual Detection of Bartonella and Babesia
The BBB ddPCR assay was able to detect natural, dual infection with B. vinsonii and Babesia vulpes within blood samples obtained from gray foxes from Portugal ( Figure 5). Consistent mean fluorescent intensities and droplet distribution patterns were obtained for each pathogen within their respective channels.

Babesia, Bartonella, and Borrelia Spiked Naïve Blood Samples
DNA from all three species, Babesia (Babesia microti), Bartonella (B. henselae), and Borrelia (Borrelia burgdorferi B31), were detected when spiked into DNA extracted from naïve human blood ( Figure 6) or when Bartonella (B. henselae) and Borrelia (Borrelia burgdorferi B31) DNA were spiked (at concentrations of 5 genome equivalent copies per microliter) into DNA extracted from a naturally B. gibsonii infected dog blood (results are not shown).

Babesia, Bartonella, and Borrelia Spiked Naï ve Blood Samples
DNA from all three species, Babesia (Babesia microti), Bartonella (B. henselae), and Borrelia (Borrelia burgdorferi B31), were detected when spiked into DNA extracted from naïve human blood ( Figure 6) or when Bartonella (B. henselae) and Borrelia (Borrelia burgdorferi B31) DNA were spiked (at concentrations of 5 genome equivalent copies per microliter) into DNA extracted from a naturally B. gibsonii infected dog blood (results are not shown).

Discussion and Conclusions
We developed and optimized a multiplex droplet digital PCR assay, using the Bio-Rad QX One System that can simultaneously amplify DNA from Babesia, Bartonella, and Borrelia species. The assay (BBB ddPCR) amplified DNA from 31 Bartonella species, 13 Borrelia species (from the Lyme, relapsing fever, and reptile complex), and 24 Babesia species. The assay also amplified two Theileria spp. (T. equi and T. cervi), as well as C. felis DNA from naturally infected clinical animal blood specimens. The multiplex BBB ddPCR assay presented herein reliably detected single and co-infections involving vector-borne pathogens from the genera Babesia, Bartonella, Borrelia, and Theilaria, using a variety of animal and human clinical samples, vectors, and experimentally infected tissues and cell-lines. The assay did not amplify Babesia, Bartonella, or Borrelia species DNA (no droplets observed) in multiple negative control samples (tissues, naïve cell-lines, naïve and clinical blood specimens, or water), tested at the same time and under the same conditions. The ability to co-amplify multiple vector-borne pathogens within a single sample with high sensitivity will greatly enhance the efficiency and efficacy of clinical diagnostic testing, particularly of volume-limited or otherwise hard to obtain sample matrices.
Despite the high analytical specificity and low limit of detection measured for the Bartonella and Borrelia spp. tested with the BBB ddPCR assay, one limitation of the current Bio-Rad QX One System is the inability to concentrate amplified DNA for genus and species confirmation by DNA sequencing. In addition to other diagnostic and epidemi-

Discussion and Conclusions
We developed and optimized a multiplex droplet digital PCR assay, using the Bio-Rad QX One System that can simultaneously amplify DNA from Babesia, Bartonella, and Borrelia species. The assay (BBB ddPCR) amplified DNA from 31 Bartonella species, 13 Borrelia species (from the Lyme, relapsing fever, and reptile complex), and 24 Babesia species. The assay also amplified two Theileria spp. (T. equi and T. cervi), as well as C. felis DNA from naturally infected clinical animal blood specimens. The multiplex BBB ddPCR assay presented herein reliably detected single and co-infections involving vector-borne pathogens from the genera Babesia, Bartonella, Borrelia, and Theilaria, using a variety of animal and human clinical samples, vectors, and experimentally infected tissues and celllines. The assay did not amplify Babesia, Bartonella, or Borrelia species DNA (no droplets observed) in multiple negative control samples (tissues, naïve cell-lines, naïve and clinical blood specimens, or water), tested at the same time and under the same conditions. The ability to co-amplify multiple vector-borne pathogens within a single sample with high sensitivity will greatly enhance the efficiency and efficacy of clinical diagnostic testing, particularly of volume-limited or otherwise hard to obtain sample matrices.
Despite the high analytical specificity and low limit of detection measured for the Bartonella and Borrelia spp. tested with the BBB ddPCR assay, one limitation of the current Bio-Rad QX One System is the inability to concentrate amplified DNA for genus and species confirmation by DNA sequencing. In addition to other diagnostic and epidemiological considerations, this limitation is of particular clinical relevance for Babesia species in veterinary medicine, where the treatment protocol varies depending upon the infecting Piroplasma (large versus small Babesia) spp. Sequence-based confirmation of pathogen identity is also critical in the context of chronic, stealth, and/or low-yield infections both to fulfill Koch's postulates, which stipulate identification/isolation of the disease-causing pathogen, as well as to ensure that the proper and most effective anti-microbial therapies are being administered to patients. This limitation is also critical for hard to obtain or volume-limited samples for which insufficient sample may remain for additional testing, or for situations where no secondary companion confirmatory test modality is available.
Compared to currently available molecular diagnostic modalities, it is anticipated that ddPCR will provide both exceptional sensitivity and specificity for the diagnosis of babesiosis, bartonellosis, and borreliosis within animal and human patients. In addition, the previously reported clinical enhanced sensitivity of ddPCR [9,13,16,18] will facilitate the discovery and subsequent characterization of novel organisms infecting animals, humans, and vectors. In contrast to serological assays, ddPCR will also enhance the capability of diagnostic laboratories to confirm a molecular diagnosis of co-infections by providing the ability to simultaneously assay multiple combinations of vector-borne pathogens and will shorten the sample to answer window for providers by reducing the number of tests to be performed on a single patient sample. Co-infections in animals and human patients induce increased clinical complexity, present more robust diagnostic challenges, and greatly influence and complicate treatment decisions. Future studies aimed at the addition of other vector-borne organisms such as Anaplasma, Ehrlichia, and Rickettsia species to the existing BBB ddPCR platform, without decreasing assay sensitivity, would be highly beneficial for clinical and research applications in human and veterinary medicine.  Institutional Review Board Statement: The study was approved by the Institutional Review Board NCSU IRB1960. Animal samples were acquired through the VBDDL from veterinarians submitting samples from animals suspected of vector-borne disease for testing. Veterinarians are informed through submission forms that the VBDDL reserves the right to use stored samples for research purposes, always respecting privacy rights of the contributing animal, owner and veterinarian.
Informed Consent Statement: Written informed consent has been obtained from the patient(s) to publish this paper.
Data Availability Statement: Data supporting reported results are available upon request. Please, contact rgmaggi@ncsu.edu. the co-director of the Vector-Borne Disease Diagnostic Laboratory, an animal diagnostic laboratory associated to the College of veterinary Medicine, North Carolina State University.