Detection of Borrelia afzelii, Borrelia burgdorferi sensu stricto, Borrelia garinii and group VS116 by PCR in skin biopsies of patients with erythema migrans and acrodermatitis chronica atrophicans

Authors


Corresponding author and reprint requests: Sjoerd Rijpkema, Research Laboratory for Infectious Diseases, National Institute of Public Health and the Environment, P.O. Box 1, 3720BA Bilthoven, The Netherlands Tel: 31 30 274 2916 Fax: 31 30 274 4418 e-mail: Sjoerd.Rijpkema@rivm.nl

Abstract

Objective: To evaluate the diagnostic performance of two polymerase chain reaction (PCR) procedures using skin biopsies of 20 erythema migrans (EM) and 24 acrodermatitis chronica atrophicans (ACA) patients.

Method: One assay amplified a fragment of the outer surface protein (Osp) A gene. The second method amplified the spacer region between the 5S and 23S rRNA genes; hybridization of this fragment allowed identification of Borrelia burgdorferi sensu lato species.

Results: Among EM patients, both assays detected Borrelia DNA in 15 samples. Among ACA patients, the ospA PCR detected 15 positives and 10 samples were positive by 5S–23S PCR. In 19 samples one species was detected, 15 skin biopsies contained Borrelia afzelii, and Borrelia garinii was found in two patients. Group VS116 was detected in two EM patients, and therefore this group has pathogenic potential. Mixed infections of B. afzelii and B. garinii, group VS116 or B. burgdorferi sensu stricto were found in three EM and three ACA patients.

Conclusions: Diagnosis of EM and ACA by PCR is useful and knowledge of the presence of species may be used to predict the course of disease or the need for further antibiotics.

INTRODUCTION

In Europe, Lyme borreliosis (LB) is caused by the bite of Ixodes ricinus ticks, which transmit a group of diverse spirochetes called Borrelia burgdorferi sensu lato [1,2]. B. burgdorferi sensu lato is divided into several species on the basis of DNA relatedness as well as other genotypic and phenotypic characteristics [3–7]. In Europe, three species and two genomic groups of B. burgdorferi sensu lato have been isolated from Ixodes ricinus: B. burgdorferi sensu stricto, B. garinii, B. afzelii [3,8], PotiB2 and VS116 respectively [9,10]. Clinical manifestations of European LB have been associated with three species: acrodermatitis chronica atrophicans (ACA) is thought to be caused by B. afzelii, whilst B. garinii is associated with neurologic symptoms, and arthritis is attributed to B. burgdorferi sensu stricto [11,12]. These manifestations are often preceded by erythema migrans (EM), a skin rash at the site of tick bite [1]. Since clinical manifestations of LB are often unclear, laboratory confirmation is required, and for this purpose serology is used. However, in the early stage of disease false-negative results are frequently found [13], and the presence of antibodies is not always proof of active Borrelia infection [14]. The polymerase chain reaction (PCR) can detect DNA of Borrelia spp. in patient material and thereby establish the presence of the spirochete. In EM and ACA patients, the PCR has successfully been used to detect Borrelia DNA in blood [15] and in skin biopsies [16–18]. Cerebrospinal fluid, serum and urine are sources for detection of Borrelia DNA in neuroborreliosis patients [19–23]. In Lyme arthritis patients, Borrelia DNA has been detected in urine or synovial fluid [24,25]. Recently, species of B. burgdorferi sensu lato have been identified by PCRs which target the ospA gene. Mixed or single infections by different species have been detected by ospA PCR in ticks [21,26] and neuroborreliosis patients [20,21]. Three species were found in one ACA patient by analysis of single-strand conformational polymorphism of PCR products [27]. Previously, the highly polymorphic 5S–23S rRNA gene spacer region was used as target for the PCR [19,28]. Hybridization of the amplified intergenic spacer region in a reverse line blot (RLB) format allowed identification of B. burgdorferi sensu lato species and mixed infections of species in ticks [10]. The aim of the present study was two-fold: first, to compare the performance of the 5S–23S rDNA PCR with that of the ospA PCR; and second, to determine whether species other than B. afzelii can be detected in the skin lesions of patients with EM and ACA.

MATERIALS AND METHODS

Patients and samples

From January 1989 to July 1994, EM (n=20) and ACA (n=24) were diagnosed in patients who were seen by one of us (D.T.) at a dermatologic and phlebologic practice in a regional hospital located in the province of Friesland, The Netherlands (Tables 1 and 2). None of the patients had received antibiotic treatment prior to clinical diagnosis. EM had a size of 5 cm or more at presentation. The diagnosis of ACA was made on the bases of clinical presentation, histopathologic examination of skin biopsies and serologic data. Clinical and histopathologic data are described in detail elsewhere [29].

Table 1.  Detection of infection by Borrelia spp. in EM patients by immunoassays, PCR and RLB
Patient No.SexAge (yr)History of tick biteSite and duration of EMResult in ELISAospA PCR5S–23S PCR and RLB
  1. IgM+=IgM detected; IgG/M+=IgM and IgG detected; ND=not done; IgG+=IgG detected.

1F25YesChest, 3 weeks+
2F58YesChest, 4 days
3F35YesLeft knee, 4 weeksIgM++Group VS116
4M54NoRight knee, 3 weeksIgM++B. afzelii
5M43NoLeft ankle, 2 weeksIgM++Group VS116
6F52YesLeft shoulder, 6 weeksIgG/M++B. afzelii
7F49NoRight hip, 3 months+B. afzelii/B. garinii/Group VS116
8F52YesRight hip, 6 weeksIgG/M++B. afzelii
9M50YesRight hip, 3 weeks+B. afzelii
10F50NoLeft hip, 4 monthsND+B. afzelii
11M23YesLeft calf, 3 daysIgG+
12F48NoRight hip, 4 weeksIgM++B. afzelii
13F70NoLeft calf, 1 weekNDB. garinii
14F57NoRight ankle, 4 weeksIgM++B. afzelii
15F51YesRight ankle, 2 weeks+B. afzelii
16F57NoLeft calf, 3 monthsIgG+
17F62YesLeft knee, 4 weeks+B. afzelii/B. garinii/B. burgdorferi sensu stricto
18F63NoLeft knee, 3 weeksIgM++
19F21NoLeft calf, 5 monthsIgG/M+B. afzelii/B. garinii/Group VS116
20F60NoRight buttock, not knownIgM++B. afzelii
Table 1.  Detection of Borrelia infection in ACA patients by immunoassays, PCR and RLB
Patient No.SexAge (yr)History of tick bite/EMSite and duration of ACAResult in IFAELISAospA PCR5S–23S PCR and RLB
  1. IgG+=IgG detected; ND=not done; IgG/M+=IgM and IgG detected; Ig+=Ig detected, no differentiation between IgM and IgG; IgG±=IgG level borderline.

1F73Yes/NoBoth legs, 18 monthsIgG+ND+
2F71No/NoBoth legs, 6 monthsIgG/M+ND+B. afzelii/B. garinii
3F17Yes/YesFour extremities, 6 monthsNDIg++B. garinii
4F54No/NoOne leg, 12 yearsIgG+ND+B. afzelii
5F48No/NoBoth legs, 10 yearsIgG+ND+
6M60No/NoOne leg, 6 monthsIgG/M+ND
7M34No/NoOne leg, 9 monthsIgG±ND+B. afzelii
8F69No/NoOne leg, 7 yearsIgG+ND
9F34Yes/YesOne leg, 9 monthsNDIg++B. afzelii
10F66No/NoOne leg, 3 yearsNDIg+
11M50Yes/YesOne arm, chest, 3 yearsNDND
12F83Yes/YesFour extremities, 8 yearsNDIgG+
13M38No/NoFour extremities, 15 yearsNDIg+
14F64Yes/NoOne leg, 8 yearsNDIg++B. afzelii
15M51Yes/YesOne leg, 18 monthsNDIg++
16M58No/NoOne leg, 2 yearsNDIg+B. afzelii/Group VS116
17F70No/NoTwo legs, 2 yearsNDIg+
18M24No/NoOne leg, two arms, 1 yearNDIg++B. afzelii
19M44No/NoOne leg, 3 yearsNDIgG++
20F41No/NoOne leg, 6 yearsIgG+ND+
21F63No/YesOne leg, 6 yearsIgG+ND+B. afzelii
22F79No/NoOne leg, 3 yearsIgG+ND+B. afzelii/Group VS116
23F75No/NoOne leg, one hand, 9 monthsNDIgG+
24F40No/NoOne leg, 3 monthsNDIg++

Serology

Indirect immunofluorescence (IFA) [30] was used to detect immunoglobulin (Ig) G and IgM in ACA patients who were diagnosed from 1989 to the beginning of 1993 (Table 2). Since 1993, an indirect enzyme-linked immunosorbent assay (ELISA) (Lyme disease microassay, Diamedix Co., Miami, Florida) has been used for detection of total Ig to the Borrelia genus, and a μ-capture flagellum ELISA (Dako A/S, Glostrup, Denmark) was used for detection of Borrelia antibodies in serum of ACA and EM patients [13]. Test procedures and interpretation of results were carried out as described by the manufacturer.

Preparation of skin biopsies for PCR

Skin biopsies were taken near the border of the EM, and from ACA patients at a site where manifestations of disease were evident. Skin was disinfected with an aqueous solution of 2% polyvinylpyrrolidone iodine. Biopsies were obtained by punch (3 mm), and immediately frozen in liquid nitrogen and stored at −80°C. For the PCR, skin biopsies were processed as described by Bergmans et al. [31] and DNA was extracted by the method of Boom et al. [32]. Briefly, 100 μL of lysate was mixed with 1 mL of guanidine thiocyanate buffer and 40 μL of silica particles. DNA was adsorbed to silica particles by continuous mixing for 30 min, and then the particles were washed extensively and dried. DNA was eluted from silica with 100 μL of H2O at 60°C and supernatant was stored at −20°C. A negative sample was included for every five patient samples, and samples of phosphate-buffered saline (pH 7), spiked with 10, 100 and 1000 spirochetes of B. burgdorferi sensu stricto strain B31 (American Type Culture Collection no. 35210), were included in each run to determine the efficiency of the DNA extraction procedure. Immediately before use in the PCR, supernatant was incubated for 10 min at 60°C and spun for 2 min to remove remaining silica particles. Five microlitres of the supernatant were used in the PCR.

Detection of B. burgdorferi sensu lato by ospA PCR

Primer sequences and positions have been derived from the ospA gene of B. burgdorferi sensu stricto B31 [33], and are given in Table 3. The ospA gene was detected by a nested PCR [15], and the specificity of the PCR product was determined by hybridization. For the first PCR, 5 μL of extracted DNA was added to 20 μL of reaction mix, which contained Super Tth polymerase (0.125 U; HT Biotechnology, Cambridge, UK), uracyl-DNA-glycosylase (UDG, 0.05 U, Life Technologies, Gibco BRL, Breda, The Netherlands), Super Tth reaction buffer (HT Biotechnology) supplemented with 2.5 mM MgCl2, dNTPs (100 μM dTTP, 100 μM dUTP, and 200 μM dATP, dGTP and dCTP; Boehringer Mannheim, Mannheim, Germany), and 5 pmol each of primers (Isogen, Amsterdam, The Netherlands) OspA-N1 and OspA-C1. PCR was performed in a thermal cycler (Thermocycler 480, Perkin-Elmer, Gouda, The Netherlands). The initial incubation of 5 min at 37°C removed any contamination by previous PCR products, and was followed by denaturation (10 min at 94°C) and 30 rounds of temperature cycling (94°C for 1 min, 45°C for 1 min, and 72°C for 1 min, the DNA amplification step being extended for 5 s with each cycle). Samples were kept at 72°C and finally put on ice. Two microlitres of the first PCR were added to the mix for the second PCR. This reaction mix contained the same ingredients as the first PCR mix except for UDG and the primers, which were replaced by ospA-N2 and ospA-C2 (Isogen). DNA was amplified by 35 rounds of temperature cycling and the annealing temperature was set at 43°C. Samples were kept at 72°C for at least 6 min, and finally samples were stored at −20°C. Efficiency of DNA amplification was checked with three controls which contained 1, 10 and 100 fg of DNA of B. burgdorferi sensu stricto B31. DNA amplification by the second PCR was also monitored separately by controls which contained 400, 40 and 4 pg of B31 DNA. Ten microlitres of the second PCR mixture were examined by electrophoresis on ethidium-bromide-stained 2% agarose gels and Southern blotting. The specificity of the 352-bp PCR product was confirmed by overnight hybridization at 50°C with a digoxigenin-labeled probe BB6 (Isogen), and bound probe was detected using the DIG nucleic acid detection kit (Boehringer Mannheim). Samples from patients who were negative by PCR were spiked with 50 fg of B31 DNA (approximately 10 genome equivalents) to determine whether PCR was inhibited by components in the DNA extract of the skin biopsy. The ospA genes of representatives of B. afzelii, B. garinii, B. burgdorferi sensu stricto and group VS116 were detected by this PCR (results not shown).

Table 3.  Primers and probes used for the detection and identification of B. burgdorferi sensu lato species
DesignationNucleotide sequenceNucleotide position on target DNA
  1. aPositions and sequences of the primers and probe of the B. burgdorferi sensu stricto B31 ospA gene are derived from Bergstrom et al. [33].

  2. bPositions and sequences of the primers and probe positions on the 5S–23S intergenic spacer are derived from Schwartz et al. [28].

  3. cSequences of the probes were obtained from Postic et al. [9].

  4. da=aminolink spacer.

Primers for ospAa PCR I  
OspA-N15′GAGCTTAAAGGAACTTCTGATAA334–356
OspA-C15′GTATTGTTGTACTGTAATTGT894–874
Primers for ospAa PCR II  
OspA-N25′ATGGATCTGGAGTACTTGAA362–381
OspA-C25′CTTAAAGTAACAGTTCCTTCT713–693
Probe for hybridization of ospAa PCR product  
BB65′Digoxigenin-ACTAATGTTTTGCCATCTTCTTTGAA449–424
Primers for 5S–23S intergenic spacerb PCR I  
23SN15′ACCATAGACTCTTATTACTTTGAC469–446
23SC15′TAAGCTGACTAATACTAATTACCC92–115
Primers for 5S–23S intergenic spacerb PCR II  
23SN25′ACCATAGACTCTTATTACTTTGACCA469–444
5SCB5′Biotin-GAGAGTAGGTTATTGCCAGGG243–263
Probes for hybridization of 5S–23S intergenic spacerb,c PCR product  
B. burgdorferi sensu lato5′a-CTTTGACCATATTTTTATCTTCCAd453–430
B. burgdorferi sensu stricto5′a-AACACCAATATTTAAAAAACATAA322–299
B. garinii5′a-AACATGAACATCTAAAAACATAAA322–298
B. afzelii5′a-AACATTTAAAAAATAAATTCAAGG305–278
Group VS1165′a-CATTAAAAAAATATAAAAAATAAATTTAAGG303–278

Identification of B. burgdorferi sensu lato species by hybridization of the amplified intergenic spacer region between the 5S and 23S rRNA genes

Primers and probes and the sensitivity and specificity of the nested 5S–23S PCR and the RLB for the detection and identification of genomic groups of B. burgdorferi sensu lato have been described previously[9,10,28]. Sequences and positions of the primers and the probes are given in Table 3. In this study, we modified the PCR procedure. Five microlitres of the first PCR were added to 20 μL of the reaction mixture of the second PCR, which consisted of 0.625 U of Taq DNA polymerase (Perkin-Elmer), Saiki buffer (Perkin-Elmer) supplemented with 2.5 mM MgCl2, dNTPs, and 5 pmol each of biotin-labeled primer 5SCB and primer 23SN2 (Perkin-Elmer). The sensitivity of the 5S–23S PCR in patient material was checked with a DNA dilution range (1 to 100 fg) which was also used to determine the sensitivity of the ospA PCR. Samples from patients who were negative by PCR were spiked with 50 fg of B31 DNA (approximately 10 genome equivalents) to determine whether PCR was inhibited by components in the DNA extract of the skin biopsy. Samples were analyzed by agarose gel electrophoresis, and, when amplified DNA was present, the sample was hybridized in the RLB to identify the species or genomic group of B. burgdorferi sensu lato [10].

Statistical analysis

The Fisher's exact test (two-tailed) of the package GraphPad Instat (version 1.14, GraphPad Software, San Diego, CA) was used to compare results between the PCRs.

RESULTS

Sera and skin biopsies of patients with EM or ACA were investigated by PCR, RLB and immunoassays. The results are summarized in Tables 1 and 2. A selection of PCR products obtained from skin biopsies and hybridized in the RLB is shown in Figure 1.

Figure 1.

Identification of species of B. burgdorferi sensu lato in skin biopsies of EM and ACA patients. Isolates of Borrelia spp. (positive controls) in lanes 1 to 4, and patient samples in lanes 5 to 26. Lane 1: B. afzelii, strain A39S. Lane 2: B. burgdorferi sensu stricto, strain B31. Lane 3: group VS116, strain AR-2. Lane 4: B. garinii, strain AR-1. Lanes 5 to 20: ACA patients. Lanes 21 to 26: EM patients. Lanes 5, 15 and 24: skin biopsies which contain more than one species.

Comparison of the ospA PCR and the 5S–23S PCR

The sensitivities of the ospA PCR and the 5S–23S PCR were determined with a dilution range of purified B31 DNA which originated from one batch. Both PCRs detected one genome equivalent (1 to 5 fg) diluted in water (results not shown). The ospA PCR and the 5S–23S PCR both detected Borrelia DNA in 15 of 20 EM patient samples (75%). Discrepant PCR results were found. Two EM patients were positive by 5S–23S PCR only, and two were positive by ospA PCR only. In four seronegative EM patients, the presence of Borrelia spp. was confirmed by PCR.

The ospA PCR detected Borrelia DNA in skin biopsies of 15 of 24 (63%) ACA patients. Ten (42%) ACA patients were positive by 5S–23S PCR. In one ACA patient Borrelia DNA was detected only by the 5S–23S PCR, whereas six ACA patients were positive in the ospA PCR and not in the 5S–23S PCR. The difference in diagnostic sensitivity between the PCRs was not statistically significant (P= 0.25). Samples of skin biopsies which were negative in the 5S–23S PCR (n=19) or the ospA PCR (n=14) were spiked with 50 fg of Borrelia DNA and were subsequently found positive by PCR (results not shown).

Identification of B. burgdorferi sensu lato species in skin biopsies

Hybridization of 5S–23S PCR products showed that one species of B. burgdorferi sensu lato was present in a total of 19 samples (Tables 1 and 2). In 15 of these samples B. afzelii was detected; nine samples originated from EM patients. B. garinii was detected in one EM patient and in one ACA patient, and group VS116 was found in two EM patients. Mixed infections were found in three EM and three ACA patients. Among EM patients, mixed infections consisted of three species, and ACA patients had infections with two species. The duration of ACA in the patients who had mixed infections ranged from 6 months to 3 years. Mixed infections always contained B. afzelii and either group VS116 or B. garinii. B. burgdorferi sensu stricto was detected once as part of a mixed infection in an EM patient.

Serology

Borrelia antibodies were detected in 12 of 18 EM patients (67%), and nine sera contained IgM. Borrelia-specific antibodies were detected in 22 of 23 ACA patients (96%); one serum sample (Table 2, patient 7) contained borderline levels of anti-Borrelia IgG, and IgM reactivity was found in two of 12 patients (17%).

DISCUSSION

We investigated 44 skin biopsies of EM and ACA patients to evaluate the performance of the 5S–23S rDNA spacer PCR and the ospA PCR. Both PCRs detected Borrelia DNA in an equal number of EM patients, whereas among ACA patients diagnostic sensitivity of ospA PCR (63%) was higher than that of the 5S–23S PCR (42%). In comparison with serology, both PCRs had additional value for the confirmation of patients diagnosed as having EM, whereas serologic analysis confirmed nearly all cases of ACA. In this study we report one case (Table 2, patient 7) where diagnosis of ACA was confirmed by PCR only. This patient had a bluish discoloration with slight edema around the ankle and leg, and histopathologic examination of the subepidermis showed a characteristic band-like infiltration of lymphocytes [30]. All clinical signs disappeared after antibiotic treatment.

Our findings are supported by studies of others who compared PCR and conventional methods for laboratory diagnosis of EM and ACA patients. Among EM patients, detection of Borrelia spirochetes by PCR proved to be more sensitive than culture [17,18], serology [17] and microscopy [16]. Moter et al. [17] reported a higher sensitivity for the ospA PCR for diagnosis of ACA patients and EM patients than we show here. The lower diagnostic sensitivity of the ospA PCR described in the present study may be attributed to differences in PCR components (e.g. origin of Taq DNA polymerase), procedures for sample preparation or selection of primers. Our primers were able to amplify the ospA gene of the same species as investigated by Moter et al., and spiking of PCR-negative samples did not reveal the presence of PCR inhibitors. Therefore, the procedure used for sample preparation may be an important factor. Prior to isolation of total DNA, Moter et al. enriched the skin biopsy for Borrelia spirochetes by a prolonged incubation in Barbour-Stoenner-Kelly II medium, while we isolated DNA directly from the skin biopsy, without enrichment, and we extracted the DNA by adsorption to a silica matrix. Nearly all DNA in a digested skin biopsy sample will be of human origin, and competition for a binding site on the matrix may result in under-representation of the Borrelia DNA in the processed sample. These differences in the DNA extraction procedure could lead to the lower sensitivity of our ospA PCR assay.

In the present study we observed a lower sensitivity for the 5S–23S PCR than for the ospA PCR among ACA patients. The copy numbers of both PCR targets, which reside on a linear plasmid and the chromosome respectively, are similar for low-passage B. burgdorferi sensu stricto isolate B31 [34]. The discrepancy between the outcome of both PCRs may therefore be explained by the use of different DNA polymerases and buffers.

Hybridization of 5S–23S PCR products showed that apart from B. afzelii, mixed infections of B. afzelii and B. garinii or group VS116 occurred frequently in EM and ACA patients. All species reported here have been found in Dutch Ixodes ricinus[10,12,35]. B. afzelii was present in 90% of PCR-positive skin biopsies from ACA patients. The association of B. afzelii and ACA has been noted previously [8,11]. The abundant presence of B. afzelii in skin biopsies of Dutch EM and ACA patients may also be explained by a predominance of B. afzelii in Dutch Ixodes ticks or a tropism of B. afzelii to low-temperature areas of the body. Mixed infections, which always contained B. afzelii, were found in 24% of PCR-positive patients. Previously, three species have been detected simultaneously in one ACA patient [27], and mixed infections of B. garinii and another species have been reported in neuroborreliosis patients [20]. The presence of multiple species in biopsies and biological fluids of LB patients is not surprising. Ixodes ricinus ticks[10,21,26,36] and competent reservoirs like small rodents [37] can carry infections with multiple species. The detection of mixed infections by PCR seems to be in contradiction to the observation that spirochetes isolated from ticks and LB patients by culture contain only one species [6]. However, culture procedures, which often include antibiotics, may inadvertently select for only one species [12], one species may dominate other species in patient material, or one species may adapt best to culture in vitro.

In one EM patient and one ACA patient only B. garinii was detected. B. garinii has been isolated from skin lesions of EM patients, and these lesions resolve more quickly than those caused by infection with B. afzelii [12]. Mild neurologic complaints are sometimes found in ACA patients [38]; however, the ACA patient described in this study did not have any symptoms of neuropathy. Detection of group VS116 alone in skin lesions of EM patients is intriguing and indicates that group VS116 has pathogenic potential.

In conclusion, rapid identification of B. burgdorferi sensu lato species in skin biopsies is a valuable tool for the laboratory diagnosis of EM or ACA. B. afzelii could be identified in nearly all ACA patients, which confirms the close association of this species with chronic cutaneous LB. Mixed infections were regularly detected in EM patients as well as in ACA patients. Although reinfections by separate tick bites cannot be excluded, the presence of multiple species in lesions of EM and ACA patients may reflect the occurrence of mixed infections in the local tick populations, and shows that infections with multiple species can persist for a prolonged period. Whether all species which are detected by PCR in mixed infections contribute substantially to the development or severity of clinical manifestations of LB remains to be clarified. In EM patients, the presence of different species may have an effect on the course and severity of disease. Thus identification of the species of B. burgdorferi sensu lato in patients who are in the early stage of disease can provide important information concerning the need for further antibiotic treatment.

Acknowledgment

We wish to thank Dr B. Wilske for providing ospA sequences of various B. burgdorferi sensu lato strains, Dr J. de Koning for storage of skin biopsies, Mr J. Hitzert for serodiagnosis of patients, and Dr A. van Dam for the gift of DNA of different Borrelia species.

Ancillary