1. Background

Ticks carry microorganisms that can be transmitted between humans and animals. In humans, some of these pathogens cause acute febrile illness (AFI) and form the largest proportion of emerging zoonotic infections [1] . The common tick-borne zoonoses (TBZ) are Anaplasma phagocytophilum, Ehrlichia chaffeensis, Borrelia species (spp), Babesia spp., Bartonella spp., Crimean Congo Hemorrhagic Fever virus and Coxiella burnetii. Multiple pathogens can be harbored by a single tick and can then be transmitted to susceptible hosts [2] [3] [4] . Surveillance studies for TBZ in Kenya and in Africa are scanty and when performed, they have tended to focus on single pathogens. For example, rickettsiae surveillance in ticks in Kenya showed a wide distribution of tick vectors, with 23.3% rickettsia infection [5] . Other tick studies have shown presence of bacteria that cause rickettsioses, bartonellosis, ehrlichiosis, Q fever, Lyme disease and Babesia [5] - [10] . A recent review by Prasad et al., identified lack of multi-pathogen surveillance approaches as one of the factors limiting fever surveillance efforts [11] . In order to start to close these diagnostic gaps, multiplex qPCR assays that are relevant to the local disease situations will be needed.

Multiplexing capabilities of Taqman qPCR provide a rapid, cost effective and high-throughput method for multiple pathogen detection in a single tube assay format and has been used in multiple studies [12] [13] [14] . Commercial test kits available for diagnosis of TBZ exist in the market but are expensive and generally contain targets that may not be of local relevance. In this study, we developed two multiplex Taqman qPCR assays for detection of the following common TBZ: A. phagocytophilum, C. burnetii, B. burgdorferi, non B. burgdorferi, Borrelia, E. chaffeensis, Rickettsia spp., Bartonella spp. and Babesia spp.

2. Material and Methods

2.1. Ethical Approval

Blood specimens used in this study were collected under an ongoing surveillance protocol approved by the Scientific and Ethical Review Committee of Kenya Medical Research Institute (SSC #1282) and the Human Subject Protection Branch (HSPB) of the Walter Reed Army Institute (WRAIR HSPB #1402). Patients were recruited to the study after signing informed consent.

2.2. Sources of DNA Samples for Assay Optimization

To develop and validate qPCR for detection of TBZ, we used artificial linearized plasmids containing the gene sequences targeted by the primers and probes listed in Table 1 that were commercially synthesized by Genscript (NJ, USA).

Full names of reporter and quencher fluorescent dyes on the probes: FAM: 6-carboxyfluorescein, CY5: cyanine-5, JOE: 6-carboxy-4’,5’-dichloro-2’,7’-dimethoxyfluorescein, VIC: 2’-chloro-7’phenyl-1,4-dichloro-6-carboxy-fluorescein, NED: Naphthyl Ethylene Diamine, BHQ: Black Hole Quencher, BBQ: Black Berry Quencher, MGB: Minor Groove Binder.

Validated biological samples for the targets were not available, thus necessitating synthesizes of plasmids. Plasmid DNA were supplied lyophilized and reconstituted to 200 ng/µL in nuclease free water. Plasmid copy numbers were computed using an online calculator that takes into account the average base pair weight (650 Daltons) and the length of the plasmid and target insert in base pairs [15] .

2.3. Human DNA Samples

Human whole blood samples (n = 512) that were collected between 2008 to 2016 from different counties in Kenya under an acute febrile illness surveillance (AFI) study were used. The aim of the AFI study is to identify pathogens associated with febrile illnesses in Kenya. To participate in the study, the patients had to be one year and above, willing to provide informed consent (if 18 years and older or by guardian if less than 18 years), able to provide assent for children 13 years or older and be febrile (fever of ≥38˚C) without a readily diagnosable source after routine clinical evaluation. Whole blood specimen is taken and frozen in Liquid Nitrogen dry shippers and shipped to the laboratory for testing.

2.4. Validation of the qPCR Assays Using Plasmids

The study aimed to develop and validate multiplex qPCR assays for detection of 8 TBZ following the guidelines for the Standards for Reporting Diagnostic Accuracy [16] .

Singleplex assay optimization: The amplification reactions were performed on Applied Biosystems 7500 fast Real time PCR system in a total of 10 μL reaction that contained 5 µL of Sensifast probe Lo-ROX PCR master mix (Bioline, CA, USA), 0.5 μM of each forward and reverse primers, 0.4 μM of the probe and 3.0 μL of plasmid (6.15 × 10⁶ copies/μL) and nuclease-free water. Except for Bartonella genus, all other primers and probes were selected from published assays and only those with data on specificity [12] [17] [18] [19] [20] [21] . For Bartonella, the primers and probe used were redesigned to target a short region (200 bp) of the citrate synthase gene using primer express software (Applied Biosystems, CA, USA) from a published conventional PCR assay [22] . The specificity of all primers and probes was tested by homology searches in the NCBI nucleotide database [23] . Probes were labeled at the 5’ end with a fluorescent reporter dye and an appropriate non fluorescent quencher at the 3’ end. Reporter dyes combinations were selected to allow multiplexing of four targets in a ABI 7500 fast Real-Time PCR System (Applied Biosystems, CA, USA). Primers and probes used for the assays are listed in Table 1.

Duo 4 plex assay optimization: Following the initial optimization of singleplex assays, the primers and probes were pooled in a duo 4-plex assay. The 5^th channel of the 7500 fast Real time PCR machine was reserved for the ROX reference dye. The reaction mix for assay one was optimized to detect C. burnetii, A. phagocytophilum, B. burgdorferii and E. chaffeensis at the following cycling conditions: a 50˚C hold for 2 minutes, 95˚C initial denaturation for 10 minutes followed by 45 cycles of 95˚C denaturation for 15 seconds and 60˚C annealing for 1 minute. Assay two was optimized to detect Rickettsia spp., Babesia spp., Bartonella spp. and other Borrelia spp. at the following cycling conditions: 50˚C hold for 2 minutes, 95˚C denaturation for 10 minutes followed by 45 cycles of 95˚C denaturation for 15 seconds and 56˚C annealing for 1 minute. After the run, each target result was analyzed separately to set appropriate threshold level which is the amplification cycle where fluorescence increases significantly above the background fluorescence. For both assays, qPCR was performed in a total of 10 μL reaction volume using 2× Sensifast probe Lo-ROX PCR master mix (Bioline, CA, USA) that contained 5 μL of master mix, 0.5 µM of each primer, 0.4 µM of each probe and 3.0 μL of plasmid, nuclease-free water was added to a final volume of 10 μL. For each assay a no template control was included. A reaction was scored positive for any of the targets if there was amplification resulting to fluorescence that generated a threshold cycle (ct) of less than 40.

2.5. Determining Assay Sensitivity for Singleplex and 4-Plex Assays

Each plasmid was diluted by 10 fold over a range of 6.15 × 10⁶ to 6.15 × 10⁻¹ copies/μL. Each concentration was tested eight times with both singleplex and 4-plex assays. For 4-plex assays, plasmids were mixed together at a concentration of 6.15 × 10⁶ copies/μL each to mimic co-infections, and then diluted in 10-fold increments. Limit of detection (LOD) for each assay was determined as the lowest plasmid concentration that was amplified for all targets. LOD was then reported as equivalent pathogen detectable by the assay depending on copy number of each target gene present in its genome. B. burgdorferi has 2 copies of 23S rRNA gene [24] , A. phagocytophilum has 1 copy of the MSP2 gene [25] , E. chaffeensis has 1 copy of the 16S rRNA gene [18] , Coxiella burnetii has 2 to 20 copies of the IS111 gene [26] , Rickettsia has 1copy of 17 kDa gene [27] , Borrelia has 1 copy of the flagelin gene [28] , Bartonella has 1 copy of citrate synthase gene (gltA) [29] and Babesia has 2 copies of 18S rRNA gene [30] . Sensitivity of co-amplifications was also tested using lowest concentrations (6.15 × 10⁻¹ copies/μL) of each target in presence of high concentrations (6.15 × 10⁶ copies/μL) of othernon-targets.

2.6. Determining Efficiency, Linearity and Precision of Singleplex and 4-Plex Assays

To determine linearity and efficiency of the assays, standard curves of threshold cycle (Ct) versus copy numbers were generated over the 6.15 × 10⁶ to 6.15 × 10⁻¹ copies/μL test range. qPCR reaction efficiency was calculated from the slope of the standard curve according to the equation $\log E={10}^{{\left(-1/slope\right)}^{-1}}$ [31] . To determine the reproducibility (precision) of the assays, three concentrations (6.15 × 10⁶, 6.15 × 10³ and 6.15 × 10⁻¹ copies/μL) of plasmid were assayed in triplicates five times on separate days. Coefficient of variation (CV) and Standard Deviation (SD) were calculated at each concentration for all targets.

2.7. Pairwise Comparison between Singleplex and Multiplex Assays

Bland-Altman analysis was used to depict the magnitude of disagreement between singleplex and multiplex assays. In Bland-Altman analysis, the Ct value difference (∆Ct) between singleplex and multiplex assays are plotted against the mean of the Ct values of the two assays [32] . Pairwise comparison was performed using Ct values from singleplex and 4-plex assays at various dilutions.

2.8. Use of Duo 4-Plex qPCR for Detection of Tick Borne Zoonoses in Human Samples

The duo 4-plex qPCR was then used to screen 512 blood samples randomly selected using Microsoft Excel 2007 from a repository of over 7000 clinical samples. Nucleic acids were extracted from 200 µL whole blood using ZR-Duet DNA/RNA mini prep kit (ZYMO Research, CA USA) and eluted in 50 µL. Three microliters of nucleic acid was used as template in each of the assays.

2.9. Data Analysis

To determine the linearity, precision and efficiency of the assays, standard curves of threshold cycle versus copy numbers were generated using an Applied Biosystems real time PCR software version 2.3. Graphpad Prism version 5 was used for prevalence calculations, plotting graphs and Bland-Altman Pairwise comparison.

3. Results

3.1. Performance of Duo 4-Plex qPCR Assays

Assay specificity: The primer and probe specificities were first evaluated in silico on the NCBI Blast [33] . No matches to the primer sequences were found other than those corresponding to the target genes. To validate the in silico results, the specificity of the primers and probes were evaluated by qPCR in the presence targeted and non-target plasmids. As shown in Table 2, there was no cross-amplification or interference in product formation (Ct values were fairly similar between corresponding singleplex and 4-plex targets). High target plasmid concentration did not affect detection level of the other targets.

-Target not detected by the assay.

Assay sensitivity and dynamic range: As shown in Figure 1, in 8 repeat experiments, all targets in assay one were detected between 6.15 × 10⁶ to 6.15 × 10¹ copies/μL corresponding to Ct values of 12 to 40 and linear range for this assay was 6.15 × 10⁶ to 6.15 copies/μL (Figure 1). For assay two in 8 repeat experiments all targets were detected between 6.15 × 10⁶ to 6.15 × 10⁰ copies/μL corresponding to Ct values of 11 to 36 for Rickettsia and Borrelia, 11 to 38 for Bartonella and 23 to 37 for Babesia. The linear range for assay two varied per target with Rickettsia and Borrelia ranging between 6.15 × 10⁶ to 6.15 copies/μL, 6.15 × 10⁵ to 6.15 copies/μL for Bartonella and 6.15 × 10³ to 0.615 copies/μL, for Babesia (Figure 2).

Efficiency, linearity and precision of optimized multiplex assays: As shown in Table 3, the duo 4-plex assays had strong linear correlation to input DNA template. The assay’s efficiencies ranged from 94 to 100 with correlation coefficient (R²) values of 0.99 to 0.999 indicating high linearity (Table 3). All assays had a standard deviation of one cycle or below indicating high precision for all targets at different concentrations. Coefficients of variation for replicate Ct values at each concentration for all targets were below 0.1 with a median value of 0.01 for assay one and 0.02 for assay two for all the concentration tested.

3.2. Comparison of Multiplex and Singleplex qPCR Assays

Bland-Altman analysis was used to depict the magnitude of disagreement between multiplex and singleplex assays. The difference in Ct values (∆Ct) between multiplex and singleplex assays were plotted against the mean of the Ct values of the serially diluted plasmid for multiplex and singleplex assays. As shown in Figure 3, the ∆Ct for the assays were within a range of −1.7 to 0.8 for Coxiella,

−1.8 to 2.9 for A. phagocytophilum, 03 to 3.4 for B. burgdorferi, −1.3 to 2.9 for E. chaffeensis, −1.3 to 0.4 for Rickettsia, −0.8 to 0.1 for Bartonella, −0.3 to 2.3 for Babesia and −1.7 to 1.7 for other Borrelia. Mean ∆Ct are very close to zero indicating small differences between the multiplex and singleplex assays.

3.3. Use of Duo 4-Plex qPCR in Clinical Specimens

A total of 512 blood samples that were collected from patients with acute febrile illness were tested for TBZ using the duo 4-plex qPCR. A total of 512 whole blood samples collected between the period of February 2008 and December 2016 were tested for presence of TBZ using the duo 4-plex qPCR. Of 512, 194 (37.9%) were males while 208 (40.6%) were females.. The age ranged from 1 to 75 years with a median of 35 years.

Of the 512 samples, 102 (20%) had at least one of the 8 tick-borne pathogens: C. burnetii (5%), A. phagocytophilum (6%), E. chaffeensis (0.8%), B. burgdorferi (7%), non B. burgdorferi Borrelia spp. (4%), Babesia spp. (0.4%), Rickettsia spp. (2%), and Bartonella spp. (0.8%). 86 of the 102 specimens (84.3%) had mono-infections and 16 (15.7%) had two or more TBZ pathogens present of which 14 (13.7%) had two infections and 1 specimen (<1%) had three or four targets as shown in Table 4.

4. Discussion

Worldwide ticks transmit the widest diversity of pathogens to humans and domestic animals, compared to other vectors. As a consequence, TBZ form the largest proportion of emerging infections [34] [35] . In Kenya, many studies have reported presence of Rickettsia, Borrelia, Coxiella, Ehrlichia, Babesia, Anaplasma and Crimean-Congo Hemorrhagic fever virus [5] [7] [8] [9] [36] [37] [38] . Although these reports evaluated the presence of these pathogens in different studies, they serve as pointers to the variety of tick borne pathogens in Kenya. Lack of multi-pathogen diagnostics has been cited as a major limitation to surveillance efforts in Africa that has led to poor epidemiological information and a

narrow list of differential diagnosis [11] . As has been reported earlier [39] [40] assay multiplexing simplifies workflow, is amenable to high throughput, conserves specimen volumes, reduces reagent costs and person time. In the current study, multiplex assays were designed for 8 TBZ pathogens and thereafter used to evaluate their presence in clinical samples obtained from febrile illness patients.

Real time PCR instruments have limited number of acquisition channels which limit the number of targets that can be multiplexed in an assay. Therefore, for the 8 target pathogens in this report, the 5 channel 7500 Fast Real time PCR system (Applied Biosystems. Foster City CA USA) could only accommodate 4 targets at a time. The duo 4-plex assays performed as well as the singleplex assays qualifying them for simultaneous detection of four TBZ pathogens each in a sample which reduces time to result (Table 2 and Figure 3).

Many of the tick-borne bacteria targeted by the duo 4-plex assays do not generate high bacteremia and have fastidious growth requirement in culture, thus making convectional diagnostic methods less sensitive and time consuming [41] . An assay that can detect low copies of bacterial DNA in samples during early stage of the infections is critical for early diagnosis and targeted treatment [42] . Our duo 4-plex qPCR assays have a limit of detection of 1.8 gene copies for assay one and 18.45 gene copies for assay two (Figure 1 and Figure 2). The optimized assays are robust with high linearity and efficiency and they can be used for quantification of TBZ pathogens (Table 3).

The high analytical sensitivity of the assays demonstrated the potential for use of the duo 4-plex qPCR in detection of the 8 common TBZ. To validate these assays, we evaluated 512 clinical blood samples collected from febrile patients attending hospitals in different regions of Kenya. A 20% prevalence of tick-borne pathogens comprising B. burgdorferi (7%), C. burnetii (5%), A. phagocytophilum (6%), non B. burgdorferi Borrelia spp. (4%), Rickettsia (2%), E. chaffeensis (0.8%), Bartonella spp. (0.8%), and Babesia spp. (0.4%) was detected. Majority were mono-infections (84.3%) and declined progressively to 13.7% for dual infections, and <1% for three or four targets (Table 4). As has been pointed out, risk of co-infections is dependent on the prevalence of the pathogens in the host tick [43] or antagonistic effect of one bacterium on another that may inhibit transmission of a second agent [44] .

One limitation of this study is the lack of confirmed positive or negative clinical samples for evaluation of analytical sensitivity of the assays. Because clinical samples have a complex matrix, analytical sensitivities based on artificial plasmids could differ from those of vectors, patients or host animals.

As far as we are aware, the tick borne pathogens identified in the study patients who had febrile illnesses are neither routinely diagnosed in Kenyan public hospitals nor are they in the differential diagnosis list of clinicians.

In conclusion, the duo 4-plex qPCR assays that were developed and used in this study identified a variety of tick borne pathogens in patient with febrile illnesses. Additional work to link pathogens detected and clinical description is needed to move such diagnosis from the research environment to hospitals. It is recommended that TBZ be included in differential diagnosis of AFI and that the laboratory capability for their diagnosis be enhanced.

Acknowledgements

Funding for this study was obtained from the Armed Forces Health Surveillance Branch (AFHSB) and GEIS (Global Emerging Infections Surveillance and Response) Section, Global Emerging Infectious Surveillance and Response (GEIS).

Ethical Statement

The study protocol was approved by the Scientific and Ethical Review Committee of Kenya Medical Research Institute (SSC #1282) and the Human Subject Protection Branch (HSPB) of the Walter Reed Army Institute (WRAIR HSPB #1402).

Disclaimer

Material has been reviewed by the Walter Reed Army Institute of Research. There is no objection to its presentation and/or publication. The opinions or assertions contained herein are the private views of the author, and are not to be construed as official, or as reflecting true views of the Department of the Army or the Department of Defense. The investigators have adhered to the policies for protection of human subjects as prescribed in AR 70-25.