Development and application of the technique for identification of Borrelia miyamotoi surface antigens

Full Text

Introduction

Borrelia miyamotoi is a pathogen causing erythema-free ixodid tick-borne borreliosis (ITBB). The first cultures of B. miyamotoi were obtained from the Ixodes persulcatus tick (strain HT31) and the Apodemus argenteus mouse (strain FR64b) in Japan in 1992 [1–3]. The detection of B. miyamotoi DNA in the blood of patients with erythema-free ITBB by scientists of the Central Research Institute of Epidemiology in Udmurtia in 2003 triggered studies of the biological properties of the pathogen, the specific features of the clinical course and epidemiology of ITBB caused by B. miyamotoi (ITBB-BM) [3–10].

The main problem in studying of the ITBB-BM pathogenesis was associated with the failing attempts to cultivate the B. miyamotoi obtained from a sick person, thus making it difficult to perform a whole genome analysis and full-featured genetic assessment of pathogens. Eventually, the problem was successfully solved [11–13]; by the time of the study, there were complete nucleotide sequences of chromosomes and sequences of all or some plasmids of 10 strains of B. miyamotoi: Strain FR64b, 3 strains CT13–2396, CA17–2241 and LB-2001 isolated from ticks in the United States [14][15] and 6 strains Izh-4, Izh-5, Izh-14, Izh-16, Yekat-1, and Yekat-6 isolated from blood of ITBB-BM patients in Russia [11][12][16][17].

B. miyamotoi express variable major proteins (Vmps), against which the affected host develops protective IgM and IgG antibodies during the infection. Based on the homology of amino acid sequences, Vmps are divided into two families: variable large proteins (Vlps) and variable small proteins (Vsps) approximately 330 and 200 amino acids in length, respectively. Vlps are further divided into 4 subfamilies or clusters: alpha (α), beta (β), gamma (γ), and delta (δ). Russian strains demonstrated several variants of expression of Vmps and their combinations [14][17][18]. Further on, we will refer to these 5 protein variants (Vlp-δ, Vsp, Vlp-γ, Vlp-α, and Vlp-β) as Vmp types. Note that previously, some authors referred to 4 types (or their representatives) as Vlp15/16, Vsp1, Vlp5, and Vlp18, respectively. The presence of the Vlp-β gene in the B. miyamotoi genome was first mentioned in the study [17]; we named it similarly to the homologous protein expressed by B. hermsii and other pathogens of tick-borne relapsing fevers. The mean difference in terms of amino acid sequences within the Vmp type is 20–40% [19], between the types — 55–80% (see Fig. 4 and Fig. 5 [17]).

Vmp genes are not located on the chromosome; they are located on plasmids, the number of which exceeds ten in the genome of the studied strains of B. miyamotoi [17]. It is assumed that at a particular time, a single borrelia expresses only one Vmp gene present at the expression site in the immediate vicinity to the bacterial promoter. However, it has a "stock" of Vmp genes (from 30 to 50), which, under certain conditions, can replace the former one at the expression site and, therefore, change the antigenic properties of the strain [14][17].

In epidemiology, the identification of a specificclass protein inherent in a particular strain is instrumental for evaluation of the intraspecific diversity of pathogens circulating at different times within different territories and in different sources (reservoirs). Information about antigenic properties of pathogens can be used in clinical practice for a prognosis of the course and outcome of the infection process as well as for development of preventive immunization products to enhance the epidemiological control of infectious diseases. In case of B. miyamotoi and other borrelia causing tick-borne relapsing fevers and Lyme disease, the major proteins can become the target for antigenic typing.

The direct serotyping of borrelia causing ITBBBM offers poor information so far, due to the limited number of isolated strains of B. miyamotoi. Identification of antigen specific antibodies produced during the disease is theoretically possible [20, 21], though it requires creating a set of antigens, which may be not commercially available so far, as well as additional equipment when using planar protein-based immunochips; such equipment is not available in most of the diagnostic laboratories [7][22–26]. Furthermore, in some cases, antibodies against Vmp can produce cross-reactions; for example, antibodies against Vlp-δ can partially interact with antigens of Vlp-γ and vice versa.

In evaluation of the antigenic diversity of B. miyamotoi, new avenues can be opened up by PCR-based molecular techniques targeted at detection or sequencing of nucleotide sequences encoding antigens. Similar approaches have been developed and are successfully used in Russian practical epidemiology for other pathogenic microorganisms [27], for example, pneumococci [28].

Considering the above, the purpose of this work was to create a PCR-based technique and to use it for identification of variable major proteins expressed by B. miyamotoi strains circulating in Russia.

Materials and methods

DNA sequences

To analyze sequences of Vmp gene fragments and to select oligonucleotides for a real-time PCR (rtPCR), we used nucleotide sequences of strains Izh-4, Izh-5, Izh-14, Izh-16, Yekat-1, and Yekat-6 deposited in the GenBank (accession numbers CP024390-CP024407, CP024205-CP024222, CP024371-CP024389, CP024351- CP024370, CP024333-CP024350, and CP024316- CP024332, respectively) [16, 17] along with the analysis of primary sequences of Vmp gene promoters [14, 18].

DNA extraction, PCR, and sequencing

The DNA was extracted from biological samples collected from ITBB-BM patients (leukocyte residue in 100 μl of the supernatant, i.e. blood plasma) by using a RIBO-prep kit in accordance with the instructions applicable to the AmpliSens® Borrelia miyamotoi-FL reagent kit (registration certificate RZN 2018/7316 dated 3/7/2018). The samples from ticks collected in the Novosibirsk Region and in the Far East were provided by E.I. Bondarenko. The presence of the B. miyamotoi DNA (the glpQ gene region) was confirmed with a RealBest DNA Borrelia miyamotoi diagnostic PCR test (Vector-Best JSC, Novosibirsk), registration certificate No. RZN 2014/1405 dated 26/5/2017 [29]. S.Yu. Kovalev provided DNA-containing B. miyamotoi samples from ticks collected in the Sverdlovsk and Tomsk Regions.

In the PCR identifying Vmps, the reaction mixture in a total volume of 25 μl contained primers, probes, and dNTPs (0.44 mM) — 10 μl; Polymerase TaqF reagents — 0.5 μl and RT-PCR-mix-2 FEP/FRT – 4.5 μl; the extracted DNA — 10 μl. The multiplex rtPCR was performed by using a Rotor-Gene Q cycler (Qiagen) in accordance with the following protocol: 95ºC — 15 min (1 cycle); 95ºC — 5 sec/60ºC — 20 sec/72ºC — 10 sec (45 cycles with detection of a fluorescent signal at 60ºC). When analyzing the amplification results, the threshold line was set at 10% of the maximum value of the fluorescent signal for each of the 4 detection channels. All reagents for PCR were made at the Central Research Institute of Epidemiology.

The PCR products were sequenced by using a 3500xL genetic analyzer and the respective reagents from Applied Biosystems in accordance with the manufacturer’s instructions.

Biological material (sources of B. miyamotoi DNA)

To check the efficiency of the new technique, we used biological samples collected from patients diagnosed with erythema-free ITBB (172 samples from 172 patients) [6–8][29–31] and ixodid tick suspensions (109 samples) [32][33]; the ticks were collected in different regions of Russia for the purpose of conducting studies focused on the epidemiology and clinical presentation of the infection caused by B. miyamotoi (Table 1). All patients signed an informed consent form in accordance with, and approval from, the institutional review boards of the Council on Bioethics of the Central Research Institute of Epidemiology, Moscow (Minutes no. 83 of 26.06.2018) and the Council on Bioethics of the Izhevsk State Medical Academy (Minutes no. 17 of 24.12.2012). Blood samples were generally collected from patients at fever peak, i.e. at the peak of spirochetemia. All samples contained B. miyamotoi DNA; its presence was confirmed by tests employing an AmpliSens® Borrelia miyamotoi-FL reagent kit.

 

Statistical methods

The lower and upper limits of the confidence interval for samples of a specific type were estimated by using the Wilson method [34] at 0.95 CI¹. The IBM SPSS Statistics 19 program and Fisher’s exact test were used to estimate the significance of differences in distributions of qualitative variables in the groups [35].

Results

The rtPCR technique for antigenic evaluation of B. miyamotoi

The genome of B. miyamotoi contains a considerable number of sequences homologous to Vlp and Vsp genes; the sequences are located on the expression plasmid or on storage plasmids. For the purpose of detection of the sequences, which presumably have corresponding protein products, the annealing site for a forward primer was located in the promoter region of the lpB plasmid specific in sequence [14]. The annealing sites for reverse primers and probes were located on the sequences corresponding to Vlp and Vsp genes. The selected oligonucleotides are shown in Table 2.

 

The annealing site for the common VMP-F2 primer is located in the promoter region. The annealing sites for other primers and probes detected in green, yellow, and red channels are located in the regions corresponding to Vlp-δ, Vsp, and Vlp-γ proteins. The reaction mixture contained an internal positive control for detection of the B. miyamotoi DNA via the orange channel — diagnostic primers and a probe used in the AmpliSens® Borrelia miyamotoi-FL reagent kit.

To validate the technique, we used DNA samples of Russian B. miyamotoi isolates. In all cases, the identification revealed only those types of expressed Vmps, which were inherent in these strains according to the data of whole-genome sequencing [16, 17], i.e. Vlp-δ in strains Izh-4 and Izh-5, Vlp-γ in Izh-14, Vsps in Yekat-1 and Yekat-6 as well as both Vlp-δ and Vsp in Izh-16.

Identification of expressed Vmp B. miyamotoi in biological samples

The results of identification of B. miyamotoi Vmp types in the biological samples collected from ITBB-BM patients and ticks are shown in Table 3. Out of 281 tested samples, 51 (18%) failed to demonstrate a potentially expressed Vmp gene: the signal was detected only in the orange channel. In other 230 samples, the positive signal was detected; the median values of the threshold cycles and their dispersion were as follows: for the green channel — 31 and 19-42, for the yellow channel — 30 and 21-42, for the red channel — 30 and 21-41. These values correlated with the amplification parameters of the internal positive control – the diagnostic target in the B. miyamotoi genome, which was detectable in the orange channel: the median value of the threshold cycles — 28; the dispersion — 18-37.

 

Negative results (Table 3, line 7) are much less often observed among samples from patients than among samples from ticks (p = 0.0002; two-sided Fisher’s exact test). At the same time, the proportion of samples where Vmps of only one type are detected (line 10) is almost identical among the clinical samples and ticks (p = 0.7), while the proportion of the samples where two types of Vmps (line 11) are concurrently detected is significantly higher among the clinical samples (p = 0.0003), which explains the decrease in the number of negative samples. The more thorough examination of the types of the Vmp gene and their combinations detected at the expression site (lines 1–6) spotlights the Vlp-δ and Vsp combination (line 4) frequently detected in the clinical samples (p = 0.00003). Any other differences in the occurrence of Vmp genotypes in the samples from ITBB-BM patients and ticks are not significant.

Lines 1–6 in Table 3 clearly show that the occurrence of Vmp types ranges from higher to lower rates: Vlp-δ → Vsp → Vlp-γ. However, the occurrence assessment based on positive signals in each of the detection channels (lines 12–15) will be more visible and reliable; the signal typical of the Vmp type is recorded any time, no matter whether it is detected as a single signal or together with the signal from another Vmp type. Such analysis confirms the Vlp-δ → Vsp → Vlp-γ ranking and highlights the fact that the occurrence of the most widely spread Vmp types is identical among patients with acute ITBB-BM infection and in tick vectors (the probability that the null hypothesis implying there are no differences is true is 0.94).

The regional differences in terms of expression of different Vmp types normally do not demonstrate statistical significance, as only 10–20 samples from multiple regions have been studied so far (Table 1). Nevertheless, it should be noted that the samples from ITBBBM patients from the Sverdlovsk Region were negative less frequently (2.4%) than the samples from the Novosibirsk Region (15.9%), p = 0.008. At the same time, among the positive samples from ITBB-BM patients from the Sverdlovsk and Novosibirsk Regions, the relative proportions of Vlp-δ, Vsp, and Vlp-γ do not differ (p = 0.71) and are close to the values shown in lines 12–14 in Table 3. On the other hand, the proportion of negative samples from ticks from the Sverdlovsk Region (39%) and from the Novosibirsk Region (29%) and relative proportions of Vlp-δ, Vsp and Vlp-γ in ticks do not differ significantly (p > 0.45).

Confirmation of the reliable identification of expressed B. miyamotoi Vmps

Among some samples, nucleotide sequences of Vmp gene fragments located at the expression site were also identified by sequencing of the respective genomic fragments. Based on the rtPCR data, the Vsp gene was expressed in 19 out of the sequenced samples, while the Vlp-δ – in 23, and the Vlp-γ — in 11 samples.

In all cases, the sequencing confirmed the reliability and accuracy of the rtPCR results in DNA samples from ticks and patients’ blood as well as the match between the primer and probe biding sites in the target nucleotide sequences and the sequences of primers and probes, which were initially selected with reference to the data of whole-genome sequencing of 6 Russian isolates.

Discussion

The new technique demonstrated sufficient efficiency in the analysis of B. miyamotoi DNA-containing biological samples; it can be extensively used as it is; it also has good prospects for improvement and upgrading. There are two main paths of development: To achieve the similar efficiency in analysis of other widely spread B. miyamotoi genotypes – American and European — and to explore the possibility to identify and assess the expression of less common Vmp types — Vlp-α and Vlp-β. The work in this field has started.

The research of the fundamentally new aspects of genetics and epidemiology of B. miyamotoi raised a number of new questions; however, at this stage, any possible answers can be only hypothetical. Firstly, why do some samples of B. miyamotoi DNA from patients’ blood (11%) remain negative? The preliminary results show that when additional Vlp-α and Vlp-β genes are  ncluded in the identification system, the percentage of negative results decreases slightly. Therefore, they can be explained by the limited sensitivity of rtPCR or by a low bacterial load in blood samples or by the presence of unknown mutations at the binding sites for primers and probes. The first explanation looks more plausible, as the thorough prospective study of ITBB-BM patients in Ekaterinburg [7][11] demonstrated that the timely selection, storage, and transportation of blood and/or DNA samples following the cold chain requirements can lead to a 3% decrease in the proportion of negative samples. The second question: Why is the proportion of negative B. miyamotoi DNA samples from ticks (29%) is significantly higher? It can be assumed that B. miyamotoi present in ticks do not need Vmp expression, at least, the expression of the most widespread Vmp types: Vlp-δ, Vsp, or Vlp-γ. Such phenomenon is typical of B. burgdorferi sensu lato, which start expressing OspC (outer surface protein C) only in human blood where borrelia OspC performs protective functions or, more specifically, inhibits the activation of the complement cascade on their surface [36]. Recently, it has been found that B. miyamotoi Vlp-δ and Vlp-α also inhibit the bypass activation of the complement cascade [37].

Finally, the most intriguing questing is why in the blood samples from ITBB-BM patients (lines 11 and 4 in Table 3) our rtPCR technique quite often identifies the expression of two Vmp types (21%), most frequently, Vlp-δ and Vsp concurrently (17%). If we assume that "a single borrelia expresses only one Vmp gene present at the expression site", we should also assume that concurrently there are several B. miyamotoi subpopulations coexisting in the patient’s body. It is not a double infection (mixed infection) with two different borrelia strains; it is the result of quasispecies emerging from the same "ancestor" present in the infecting tick. Quasispecies can result from phase variations, i.e. protein synthesis switch, in part of the population of the pathogen colonizing the host. Phase variations are usually of adaptive significance. In case of surface proteins having antigenic properties, specifically Vmps, the alteration of the antigenic pattern may make borrelia less sensitive to bactericidal antibodies developed against the initial variants of antigens. Such phenomenon known as immune evasion helps the pathogen extend its existence in the host, at least, until antibodies against new variants of antigens are developed. Genetic mechanisms of phase variations are extensively explored for the VlsE protein of B. burgdorferi sensu lato [38]. The experiments on laboratory animals and/or borrelia culture showed the actual possibility of Vmp synthesis switch for B. hermsii [39][40] and B. miyamotoi [18]. With acute ITBB-BM disease, more than half of the patients produce antibodies against several Vmp types [7][24], thus implying the existence of immune evasion in vivo, similarly to the presence of several Vmp genotypes expressed in the same clinical sample, as it was found in this study. During the laboratory infection of small rodents with Russian strains of B. miyamotoi, we documented the manifestations of immune evasion by using direct genetic and serological methods [41]. Presumably, it is immune evasion that is responsible for fever relapses in ITBB-BM patients having no adequate antibiotic therapy [42]. The relapsing course is typical of tick-borne relapsing fevers and other recurrent fevers, for example, for such life-threatening disease as louse-borne relapsing fever or Borrelia recurrentis infection [2].

With the increasing number of clinical and biological field samples containing pathogenic DNA, the new technique can be used in further epidemiological studies in Russia and worldwide to produce a clearer picture of the antigenic spectrum of B. miyamotoi strains circulating in different regions and to compare it with the profile of more conserved regions of their genome, for example, by using the multilocus sequence typing method [27]. It will help understand the stages of evolution and spread of B. miyamotoi sensu lato – pathogens causing ixodid tick-borne borreliosis, which is very similar to the well-known Lyme disease in its epidemiological significance.

The second strand of research addresses the analysis of the relationship between the complex of clinical manifestations and complications of ITBB-BM and the Vmp type expressed at the onset of the disease, including the possibility of expression switch (immune evasion) during the infection process.

About the authors

K. O. Mironov

Central Research Institute of Epidemiology

Author for correspondence.
Email: mironov@pcr.ru
ORCID iD: 0000-0001-8207-9215

Konstantin O. Mironov — D. Sci. (Med.), Head, Scientific group of developing new genetic polymorphisms detection methods

Moscow

Россия

A. V. Titkov

Central Research Institute of Epidemiology

Email: fake@neicon.ru
ORCID iD: 0000-0001-7548-9267

Anton V. Titkov — researcher, Laboratory of zoonoses

Moscow

Россия

K. V. Kuleshov

Central Research Institute of Epidemiology

Email: fake@neicon.ru
ORCID iD: 0000-0002-5238-7900

Konstantin V. Kuleshov — Cand. Sci. (Biol.), senior researcher, Laboratory of molecular diagnostics and epidemiology of intestinal infections

Moscow

Россия

N. M. Kolyasnikova

Central Research Institute of Epidemiology; Chumakov Federal Scientific Center for Research and Development of Immune-and-Biological Products of RAS

Email: fake@neicon.ru
ORCID iD: 0000-0002-9934-2582

Nadezhda M. Kolyasnikova — Cand. Sci. (Med.), Head, Laboratory of tick-borne encephalitis and other viral encephalitis; researcher, Laboratory of zoonoses

Moscow

Россия

E. I. Bondarenko

JSC «Vector-Best»

Email: fake@neicon.ru
ORCID iD: 0000-0002-4699-9548

Evgeniy I. Bondarenko — Cand. Sci. (Med.), researcher, Laboratory of PCR

Novosibirsk

Россия

A. E. Platonov

Central Research Institute of Epidemiology

Email: fake@neicon.ru
ORCID iD: 0000-0001-7450-0081

Alexander E. Platonov — D. Sci. (Biol.), Prof., chief researcher, Laboratory of zoonoses

Moscow

Россия

References

Supplementary files