The prevalence of Borrelia miyamotoi infection, and co-infections with other Borrelia spp. in Ixodes scapularis ticks collected in Canada

Background Blacklegged ticks, Ixodes scapularis are vectors of the tick-borne pathogens Borrelia burgdorferi, Anaplasma phagocytophilum and Babesia microti. Recently, the I. scapularis-borne bacterium Borrelia miyamotoi has been linked to human illness in North America. The range of this tick is expanding in Canada which may increase the potential for human exposure to these agents. Methods In this study, 4938 I. scapularis ticks collected in 2012 were tested following a newly developed PCR-based testing protocol to determine the prevalence of infection with B. miyamotoi and other pathogens in I. scapularis in Canada. Results Borrelia miyamotoi was detected in blacklegged ticks from all provinces except Newfoundland, although the infection prevalence was low (<1%). There was significant variation among provinces in the prevalence of infection of ticks with B. burgdorferi and A. phagocytophilum, but not with B. miyamotoi. Conclusions Given the widespread distribution of B. miyamotoi, infection due to this agent should be considered in patients who have been exposed to blacklegged ticks in Canada.


Background
Borrelia miyamotoi was first described in Ixodes persulcatus ticks and in the blood of rodents collected in Japan in the early 1990s [1]. Subsequently, B. miyamotoi was detected, for the first time in North America, associated with blacklegged ticks, Ixodes scapularis in several states in the Northeastern United States [2]. Infection rates in field-collected nymphal I. scapularis were 1.9-2.5% [2] and unlike the agent of Lyme disease, Borrelia burgdorferi, B. miyamotoi is transmitted vertically from infected female I. scapularis to a variable proportion of larval progeny [2,3]. Initially the public health significance of B. miyamotoi was poorly understood; however, recent studies in Russia demonstrate that Old World strains of B. miyamotoi, transmitted by I. persulcatus cause an influenza-like illness with relapsing fever [4]. In North America, meningoencephalitis was recently described in an elderly immunocompromised patient [5] and results of a serosurvey of patients from southern New England and New York demonstrate that B. miyamotoi infection can cause a viral-like illness [6]. These studies support the contention that B. miyamotoi is yet another of the guild of pathogens, which includes the agents of Lyme disease, anaplasmosis, babesiosis, Powassan virus and the Ehrlichia muris-like agent, associated with blacklegged ticks in North America [7]. The discovery of DNA of B. miyamotoi in ticks during a study of B. burgdorferi diversity signaled the possible occurrence of B. miyamotoi in Canada [8].
Passive surveillance for blacklegged ticks, which involves the submission of ticks collected by the general public and participating medical and veterinary clinics, has been conducted across Canada (excluding British Columbia) since the 1990s [9,10]. Ticks identified as I. scapularis have been routinely tested for infection with B. burgdorferi and Anaplasma phagocytophilum using first a multiplex real-time PCR assay [11], followed by an ospA real-time PCR to confirm B. burgdorferi infection [9]. Most I. scapularis are submitted from locations where reproducing populations of I. scapularis occur (southern Manitoba, southern and eastern Ontario, southern Quebec, and locations in New Brunswick and Nova Scotia). Some I. scapularis are also submitted from locations where populations are not known to occur (e.g. Alberta, Saskatchewan, Prince Edward Island and Newfoundland) and it is thought that these ticks are 'adventitious' ticks dispersed from tick populations in Canada and the USA by migratory birds [9]. Each year a small number of I. scapularis submitted in passive surveillance test positive for Borrelia species infection in the 23S rRNA real-time PCR screening assay but are negative for B. burgdorferi infection in the confirmatory ospA real-time PCR assay. Subsequent testing by nested PCR (nPCR) and sequencing indicate that some of these extracts are positive for infection with B. miyamotoi. This supports earlier Multilocus Sequence Typing (MLST) analysis in which a small number of B. burgdorferi-infected ticks were found to be co-infected with B. miyamotoi [8].
We have developed and evaluated molecular assays to identify B. miyamotoi and a PCR-based testing protocol or diagnostic approach for testing I. scapularis ticks collected in surveillance for tick-borne agents. Data on the western blacklegged tick I. pacificus, and B. miyamotoi prevalence were not addressed by this study. Here we undertake a systematic analysis of ticks recently collected in surveillance in Canada to i) better understand the possible geographic range of B. miyamotoi in Canada; ii) estimate the prevalence of B. miyamotoi infection in I. scapularis ticks in Canada; and iii) investigate the frequency of co-infections with B. miyamotoi, B. burgdorferi and A. phagocytophilum in I. scapularis in Canada.

Development of B. miyamotoi-specific IGS real-time PCR
DNA from 25 ticks collected in surveillance prior to 2012 that tested positive with the screening 23S rRNA real-time PCR, but negative with the confirmatory ospA real-time PCR was tested with a nPCR specific to the genus Borrelia [12] which amplifies 587 bp of the 16S-23S IGS region. For this PCR, 5 μl DNA template was added to 95 μl master mix containing 0.2 mM each dNTP, 0.5 μM forward and reverse primers, 5 Units of AmpliTaq Gold® polymerase and 1.5 mM MgCl 2 (Life Technologies, Carlsbad, CA). The thermocycler conditions used were as follows: denaturation at 94°C for 4 minutes, 35 cycles of amplification at 94°C for 1 minute, 50°C for 1 minute and 72°C for 1 minute, followed by a 10 minute extension phase at 72°C for both stages of the nested PCR reaction. Amplification products were analyzed by ethidium bromide-stained 2% agarose gels. All nPCR products were purified using Montage®PCR filter units (Millipore) and sequenced on an ABI 3130xl Genetic Analyzer using BigDye™ Terminator version 3.1 cycle sequencing kits. Sequence data was analyzed using DNASTAR Lasergene 9 Software and multiple alignments were performed using Clustal W. Sequences were compared to those in GenBank and BLAST results indicated that 8 of the 25 tick extracts were positive for B. miyamotoi. Subsequently, B. miyamotoi-specific and B. burgdorferi-specific primers and FAM-labeled probes annealing to the 16S-23S IGS were designed from these sequences for real-time IGS PCR ( Table 1).

Validation of real-time PCR
Validation of the species-specific IGS real-time PCR assays was performed using DNA from the eight 23S PCR-positive and ospA PCR-negative ticks mentioned previously, DNA from 72 ticks (collected from 2008 to 2012) that were positive on both the 23S and the ospA real-time PCR confirming B. burgdorferi infection, and DNA from 9 ticks that were 23S PCR-positive and ospA PCR-negative and had been confirmed as being infected with B. bissettii by flagellin nPCR. DNA from cultures of B. garinii strain ATCC® 51991™ and B. afzelii strain ATCC® 51567™, and B. hermsii DNA from a clinical sample was used to evaluate the specificity of the assays. Once validated, the species-specific real-time PCR assays were applied to DNA from an additional 39 23S PCRpositive and ospA PCR-negative ticks collected during 2008-2012.
Reaction mixtures were prepared in 2x TaqMan® Universal Mastermix (Applied Biosystems, Life Technologies) to contain 300 to 600 nM of each primer and 200 nM probe. Amplification was carried out on either an ABI 7500 Real-time PCR System, ABI 7900HT or ABI ViiA7 using 96 well optical plates. Thermocycling conditions consisted of: activation of AmpErase at 50°C for 2 minutes, 10 minutes at 95°C for denaturation of AmpErase and activation of AmpliTaq Gold® Polymerase, followed by 40 cycles of amplification with denaturation at 95°C for 15 seconds and annealing at 58°C for 1 minute. Following amplification and real-time data acquisition, analysis was performed using the Sequence Detection System software. A second real-time PCR assay targeting B. miyamotoi glpQ [13] was performed as a confirmatory assay, using an annealing temperature of 50°C.
DNA from B. miyamotoi-positive, B. bissettii-positive and most B. burgdorferi-positive ticks was tested using the nested 16S-23S IGS PCR to generate products for sequencing. This sequence data was considered the gold standard and provided validation data for the B. miyamotoi real-time PCR assays.

Analysis of infections and co-infections in I. scapularis ticks collected in passive surveillance
A total of 4938 I. scapularis ticks collected in passive surveillance in 2012 (excluding the 68 used in validation) were tested using the developed testing protocol ( Figure 1). The screening 23S and confirmatory ospA real-time PCR were used to assess B. burgdorferi infection as described above. The screening 23S PCR is a multiplex assay that also detects the presence of A. phagocytophilum DNA using primers specific for the msp2 gene (ApMSP2f and ApMSP2r) [14]. Confirmation of infection with A. phagocytophilum is achieved by an in-house real-time PCR assay targeting 16S rRNA. The Borrelia miyamotoi-specific IGS real-time PCR was used to detect B. miyamotoi infections that were subsequently confirmed by B. miyamotoi glpQ real-time PCR. All real-time PCR assays were conducted using the conditions described above for B. miyamotoi IGS. Tick extracts that were positive for Borrelia spp. infection in the screening 23S real-time PCR, but negative for B. burgdorferi in the ospA real-time PCR, and negative in the B. miyamotoi IGS assay, were tested by 16S-23S IGS nPCR with the aim of sequencing products to identify other infecting Borrelia species.
Associations of infections and co-infections in ticks with province of origin, level of engorgement of the tick, host of origin, and tick instar were investigated by logistic regression in STATA version 11.0 for Windows (STATACorp, College Station, TX, USA). The most parsimonious multivariable model was created by backwards and forwards elimination and substitution of variables. Logistic regression models were used to investigate whether or not there were significant associations between infections of ticks with different pathogens. The level of significance throughout was P < 0.05.

Development and validation of B. miyamotoi-specific IGS real-time PCR
Primers were developed using a sequence with 100% similarity to a sequence from B. miyamotoi  Table 2) serve as the basis for our testing protocol to detect the suite of Borrelia species found in blacklegged ticks collected in Canada.

Analysis of infections and co-infections in I. scapularis ticks collected in passive surveillance
Of the 4938 ticks tested (  other explanatory variables were not significantly associated with B. miyamotoi infection. There were significant variations amongst provinces in the likelihood that a tick was positive for B. burgdorferi. The provinces could be simplified into three groups within which the prevalence of infection was similar: i) Alberta and Manitoba, ii) New Brunswick and Prince Edward Island, and iii) ticks from all other provinces (Ontario, Quebec, Nova Scotia and Newfoundland) (χ 2 = 4.8, df = 5, P > 0.1). Ticks from New Brunswick and Prince Edward Island were significantly less likely to be infected with B. burgdorferi than ticks from Ontario, Quebec, Nova Scotia and Newfoundland combined (OR = 0.55, 95% CI = 0.39 -0.78, P < 0.01), and less likely to be infected than ticks in Alberta and Manitoba (by Wald test of final model parameters: χ 2 = 14.7, df = 1, P < 0.001). Ticks from Alberta and Manitoba were less likely to be infected with B. burgdorferi than ticks from Ontario, Quebec, Nova Scotia and Newfoundland combined (OR = 0.64, 95% CI = 0.41 -0.98, P < 0.05). Ticks were significantly less likely to be infected if they fed on humans (OR = 0.59, 95% CI = 0.46 -0.75, P < 0.001), and were less likely to be infected if they were slightly, partially or fully engorged than if they were unfed (ORs = 0.44, 0.30, 0.21; 95% CIs = 0.33-0.60, 0.24-0.39, 0.10-0.44; P < 0.001 for all).
No immature ticks were infected with A. phagocytophilum but there was significant variation amongst province of origin in the proportion of adult ticks infected with A. phagocytophilum. Adult ticks collected in Alberta and Manitoba were significantly more likely to be infected than ticks from other locations (OR = 4.5, 95% CI = 2.0-10.4, P < 0.001). Ticks from Quebec were significantly more likely to be infected than ticks from Ontario, New Brunswick, Nova Scotia and Newfoundland combined (OR = 3.7, 95% CI = 1.7-7.7, P < 0.001). There were no significant variations in prevalence of A. phagocytophilum infection associated with state of engorgement or host of origin.
Co-infections were detected in 19 ticks (15 being adult females, one being an adult male and 3 having the instar unrecorded), of which 11 (0.23% of adult ticks) were coinfected with A. phagocytophilum and B. burgdorferi, and 8 (0.17% of adult ticks) were co-infected with B. burgdorferi and B. miyamotoi (Tables 4 and 5). Consequently statistical analysis was limited to adult ticks. Adult ticks were significantly more likely to be infected with B. miyamotoi if they were infected with B. burgdorferi (OR = 3.5, 95% CI = 1.5 -8.4, P < 0.01). This relationship remained significant when other variables (province of origin, level of engorgement of the tick, host of origin, and tick instar) were included in the model. There was no significant association between

Discussion
The objective of this study was to develop and implement a systematic approach using real-time PCR assay to detect B. miyamotoi, B. burgdorferi and A. phagocytophilum infections and co-infections in ticks collected in surveillance. In doing so, we were able to assess the prevalence of infection of ticks collected in Canada, with the newly-recognized pathogen B. miyamotoi. Results of our study identified B. miyamotoi-infected ticks at low (<1%) prevalence in most provinces. Few ticks were coinfected, however a third of B. miyamotoi-infected ticks and a quarter of A. phagocytophilum-infected ticks were also infected with B. burgdorferi and co-infections of B. miyamotoi and B. burgdorferi occurred more frequently than would be expected by chance. It is increasingly recognized that I. scapularis ticks transmit a range of bacteria including the Lyme diseasecausing B. burgdorferi, B. miyamotoi, the E. muris-like agent [15], and A. phagocytophilum [16] as well as bacteria such as B. carolinensis [17] and B. bissettii [18] whose pathogenicity has not yet been determined. Here we have developed new assays and combined them with existing ones to create a PCR testing protocol, similar to that of Ullman et al. [13], which allowed us to detect and identify infections and co-infections of ticks with different Borrelia species and A. phagocytophilum. The new 16S-23S IGS real-time PCR assays were robust showing 100% concordance between positive results and sequence analysis indicating high specificity. There was slight reactivity of B. hermsii DNA in the B. burgdorferi IGS real-time assay, but this is of little consequence on test outcomes as B. hermsii is transmitted by Argasid ticks and would rarely, if ever, be encountered in blacklegged ticks obtained in surveillance. Future refinements of this testing protocol will include implementation of a duplex real-time PCR assay for B. burgdorferi ospA and B. miyamotoi 16S-23S IGS to reduce PCR steps, development of a real-time PCR assay to detect B. bissettii and incorporation of PCR assays for nonbacterial I. scapularis-borne pathogens such as Powassan encephalitis virus.
By implementation of this new testing protocol for the detection of selected species of Borrelia, we have expanded on the findings of Ogden et al. [8] who first detected B. miyamotoi in blacklegged ticks collected in Canada. The prevalence of B. miyamotoi in blacklegged ticks in our study (<1%) was lower than the 1 -5% reported in the eastern USA [2,3], and this difference may suggest that B. miyamotoi transmission cycles are at an early stage of becoming established amongst resident tick and rodent populations in Canada compared to localities in the USA [19]. Infection prevalence of tickborne pathogens in ticks and hosts may take some years to rise to an equilibrium level, particularly if the ticks are at low densities, being themselves at an early stage of becoming established [20]. However, our study also confirms that as in the US, B. miyamotoi can be detected across the geographic range of I. scapularis in Canada   [20,23]. Higher infection A. phagocytophilum prevalence in ticks in some locations in Quebec has been detected possibly associated with founder events in naïve host populations [24,25]. In contrast to B. burgdorferi, there was no evidence of geographic variation (in the presented analysis as well as in cluster analysis not described here) in the prevalence of B. miyamotoi infection of ticks, which is consistent with more simultaneous introduction of B. miyamotoi with I. scapularis. As B. miyamotoi is transovarially and transtadially transmitted in ticks, this bacterium can be imported in host-dispersed infected engorged nymphal ticks as well as larval ticks, while only imported engorged larvae can efficiently introduce B. burgdorferi [22]. Variations in prevalence of B. burgdorferi infection with stage of engorgement are consistent with our findings in previous studies; B. burgdorferi multiplies in the tick as it feeds [21]. Variations in prevalence of B. burgdorferi infection with host of origin have also been observed in our passive surveillance data [21] although here ticks collected from humans were less likely to be detected as infected. The underlying reason for this variation is unknown; however an analysis of the quality of the submitted ticks did not indicate significant differences between ticks removed from humans or companion animals (data not shown). It is possible that ticks from dogs had higher infection prevalence than ticks from humans because some of the dogs were infected and consequently infective for ticks that fed on them [26]. Some I. scapularis ticks were co-infected with B. miyamotoi and B. burgdorferi or with A. phagocytophilum and B. burgdorferi however, co-infection with B. miyamotoi and B. burgdorferi occurred more frequently than by chance, which is consistent with shared reservoirs for these species [20,27]. The implications of these observations for disease in humans are at present unknown and require further investigation, as does the occurrence of B. miyamotoi in I. pacificus ticks, the other main vector of tick-borne zoonoses that occurs in British Columbia.

Conclusions
The relatively limited (though expanding) distribution of blacklegged tick populations in Canada [22,28,29] and the lower prevalence of B. miyamotoi infection in these ticks means that at present the risk of infection of humans in Canada would be lower than in parts of the USA [6]. Nevertheless, our study indicated that B. miyamotoi is present across a wide geographic range in Canada, and clinicians should consider B. miyamotoi infection as a possible diagnosis, alongside Lyme disease, Anaplasmosis, Ehrlichiosis, Babesiosis and arboviral infections, in patients suffering from suspected infectious disease who have potentially been exposed to ticks in Canada. Our findings underline the need for improved diagnostics for B. miyamotoi and other tick-borne pathogens, and ongoing exploration for novel tick-borne pathogens.