Analysis of Borrelia burgdorferi genotypes in patients with lyme arthritis: High frequency of ribosomal RNA intergenic spacer type 1 strains in antibiotic-refractory arthritis

/Home/Lyme Disease News/Lyme Disease Science Articles

Objective
Most of the Borrelia burgdorferi genotypes have been isolated from erythema migrans (EM) skin lesions in patients with Lyme disease. OspC type K strains, which are 16S–23S ribosomal RNA intergenic spacer type 2 (RST2) strains, are most commonly recovered, but a higher percentage of OspC type A strains (RST1), the next most commonly recovered type, is detectable in blood. The goal of this study was to determine the B burgdorferi genotypes in the joints of patients with Lyme arthritis.

Methods
Joint fluid samples from 124 patients seen over a 30-year period were analyzed for OspC types by semi-nested polymerase chain reaction (PCR) and sequencing, and for RSTs by nested PCR and restriction fragment length polymorphism analysis. These results were correlated with clinical outcome.

Results
OspC and RST genotypes were identified in 49 of the 124 joint fluid samples (40%). In these 49 samples, OspC type K strains (RST2) were identified in 21 samples (43%), OspC type A strains (RST1) were identified in 11 samples (22%), and 8 other OspC types and all 3 RSTs were identified among the remaining 17 samples (35%). However, among the 17 patients who had been treated with antibiotics according to current guidelines, all 7 patients who were infected with RST1 strains had antibiotic-refractory arthritis, compared with 4 of 6 patients infected with RST2 strains and only 1 of 4 infected with RST3 strains (P = 0.03).

Conclusion
Most of the B burgdorferi genotypes, particularly OspC type K (RST2), were identified in the joint fluid of patients with Lyme arthritis, and the genotype frequencies found in joints reflected those in EM skin lesions. However, RST1 strains were most frequent in patients with antibiotic-refractory arthritis. Our results help to further the understanding of the differential pathogenicity of strains of B burgdorferi.

Lyme disease in the US, which is caused by the tick-borne spirochete Borrelia burgdorferi, usually begins with an expanding skin lesion, called erythema migrans (EM) (1). Within days to weeks after initial infection, spirochetes often disseminate, particularly to the nervous system, heart, or joints. Within months, ∼60% of untreated individuals develop intermittent or persistent arthritis, which affects the knees in particular (2). All manifestations of Lyme disease usually respond well to appropriate oral or intravenous (IV) antibiotic therapy (3). In rare cases, however, Lyme arthritis can persist for months or even several years despite 2–3 months of oral and/or IV antibiotic therapy, a condition that has been termed antibiotic-refractory Lyme arthritis (1, 4). After such therapy, results of polymerase chain reaction (PCR) testing for B burgdorferi DNA in joint fluid are usually negative (4), which suggests that joint inflammation may persist after spirochetal eradication.

Two genetic markers of B burgdorferi, the OspC type and the 16S–23S ribosomal RNA (rRNA) intergenic spacer type (RST), have been used to correlate spirochetal strain variation with clinical outcomes (5–18). OspC typing divides B burgdorferi strains into 21 genetically distinct types, 16 of which have been identified in the northeastern US (5). This system allows for detailed classification of types, but the numbers in some groups are small (6). In contrast, the RST system divides B burgdorferi into only 3 groups (7), which allows larger numbers in individual groups but may miss differences within groups (8). Therefore, an analysis of strains using both typing systems may provide more accurate information.

Among B burgdorferi strains in the northeastern US, the plasmid-encoded gene encoding OspC is in strong linkage disequilibrium with rrs-rrlA, the chromosomal gene encoding the 16S–23S rRNA intergenic spacer region, suggesting a clonal structure of strains in this geographic region (13, 17, 19). In previous analyses, RST1 corresponded uniquely to OspC genotypes A and B; RST2 corresponded to OspC genotypes F, H, K, and N; and RST3 corresponded to the remaining 10 OspC genotypes, including D, E, G, and I (17, 18). Thus, in samples from this geographic region, the RST can be inferred accurately from the OspC sequence (18).

Previously, the 2 typing systems have been used primarily to determine the frequencies of isolates from EM skin lesions and to delineate strains associated with hematogenous dissemination. In an initial study of 132 isolates of skin, blood, or cerebrospinal fluid (CSF) (5), 15 of the 21 OspC types (most commonly, types K and A) were identified in EM skin lesions. However, only 4 types (A, B, I, and K), which represent all 3 RSTs, were found at sites of dissemination in blood or CSF. In a later study of 6 blood isolates, dissemination occurred not only with OspC types A, B, I, and K, but also with types H and N (16), suggesting that a wider range of strains may cause disseminated disease.

The findings of recent studies, employing both typing systems, revealed that most of the B burgdorferi genotypes characterized by either typing system could be found in EM skin lesions, and that all of the genotypes found in skin were sometimes detectable in blood (17, 18). Consistent with the results reported earlier, OspC type K (RST2) and OspC type A (RST1) strains were found most commonly. However, RST1 strains were overrepresented in blood, and RST3 strains were underrepresented (18), suggesting that B burgdorferi genotypes may have differential pathogenicity.

Only limited information is available regarding spirochetal strains that cause infection at sites of dissemination, where culturing has been difficult. Therefore, in this study, our goal was to directly identify the B burgdorferi OspC and RST types in the joints of patients with Lyme arthritis, using PCR, restriction fragment length polymorphism (RFLP) analysis, and sequencing techniques to analyze a unique collection of joint fluid samples from 124 patients seen over a 30-year period.

Patients.
Joint fluid samples were available from 124 patients with Lyme arthritis who participated in studies of Lyme disease at Yale–New Haven Hospital (1975–1987), Tufts Medical Center (1987–2002), or Massachusetts General Hospital (2002–2006). The Human Investigation Committees at each of these institutions approved the protocols, and all patients (or their parents, in the cases of patients who were minors) provided written informed consent. All 124 patients met the criteria of the Centers for Disease Control and Prevention (CDC) for the identification of Lyme disease (20). Patients had intermittent or persistent episodes of monarticular or oligoarticular arthritis, and a positive antibody response to B burgdorferi sonicates by enzyme-linked immunosorbent assay and Western blot, interpreted according to the CDC criteria (21). For clinical correlations, data from medical records were available for 119 of the 124 patients. The size of joint effusions was determined by joint aspiration, and the joint fluid white cell counts were determined in the hospitals' clinical laboratories.

During the initial study years (1975–1980), patients with Lyme arthritis were not treated with antibiotics (2). During the 1980s, various antibiotic regimens were tested (22–24), and by the 1990s, patients were given oral or IV antibiotic regimens that are now recommended by the Infectious Diseases Society of America (IDSA) (3). As physicians became more familiar with Lyme arthritis, patients were referred to us primarily because of incomplete responses to recommended antibiotic regimens. Therefore, our clinic population consisted of more patients with antibiotic-refractory arthritis, a rare manifestation of the disorder, than patients with antibiotic-responsive arthritis (4). For the current study, as in previous studies (4, 25, 26), antibiotic-responsive Lyme arthritis was defined as the resolution of arthritis within 3 months after the start of either ≤4 weeks of IV antibiotics or 8 weeks of oral antibiotics. Antibiotic-refractory arthritis was defined as persistent joint swelling for >3 months after the start of >4 weeks of IV antibiotics, >8 weeks of oral antibiotics, or both.

For the comparison of OspC and RST types in patients with joint or skin manifestations of the infection, we used data from previous studies of isolates from EM skin lesions (5, 17, 18). However, using the data from our own previous study (17), we recalculated the frequency of hematogenous dissemination based only on PCR evidence of B burgdorferi DNA in blood and not on clinical signs and symptoms, since culture or PCR testing of blood provides the most direct evidence of hematogenous dissemination.
Identification of OspC types and RSTs in joint fluid samples.
All 124 joint fluid samples were stored at −80°C. DNA isolation and extraction were carried out in a clean room that was not used for other purposes. After rapid thawing, DNA was isolated from 100 μl of joint fluid diluted 1:1 with phosphate buffered saline (pH 7.4; Fisher Scientific, Pittsburgh, PA) to reduce viscosity. The extraction was done using a QIAamp DNA Mini Kit (Qiagen, Valencia, CA) according to the Blood and Body Fluids Protocol, and DNA was eluted into a final volume of 50 μl of sterile water.

Joint fluid samples from all 124 patients were analyzed for the B burgdorferi ospC gene using semi-nested PCR and sequencing techniques, and for rrs-rrlA using nested PCR and RFLP analysis. The PCR amplifications for both genes used primers that have been described previously (6, 8, 17). Each 200-μl PCR mixture contained 5–20 μl of DNA extracted directly from joint fluid, 1× Qiagen PCR buffer, 0.2 mM dNTPs, 2.5 mM MgCl2, 0.25 μM each of forward and reverse primer, and 5 units of Qiagen HotStar Taq polymerase. A 1:10 dilution of the product from the first amplification was used in the second amplification. The cycling conditions for each PCR were identical: a polymerase activation step at 95°C for 15 minutes, followed by 42 cycles of denaturation at 94°C for 30 seconds, primer annealing at 45°C for 30 seconds, and extension at 72°C. PCR products were visualized on 2% agarose gels (Invitrogen, Carlsbad, CA) with ethidium bromide. The ospC PCR yielded a band of ∼300 bp (6), and the rrs-rrlA PCR yielded a band of ∼941 bp (8). When a PCR product of the correct size was present, 100 μl of each PCR product was eluted in 50 μl of distilled water following removal of dNTPs, enzymes, and PCR buffers, using the QIAquick PCR Purification Kit (Qiagen). For each assay, a negative control sample (dH2O) and a positive control sample (20 μl of DNA from an OspC type A, K, or E isolate of B burgdorferi) were placed side-by-side to the left of the molecular weight ladder and unknown samples.

In the 67 joint fluid samples with an ∼300-bp band, 10 μl of purified ospC second-round PCR product was then combined with 10 μl of the external (positive) primer and submitted for sequencing to the Massachusetts General Hospital DNA Core Facility. The OspC type was identified from these sequences in 49 of the samples by nucleotide-nucleotide BLAST (27, 28). Of the 46 samples with an ∼941-bp band, the RST was determined in 19 using RFLP techniques described previously (8, 17). In addition, 3 samples, 2 identified as RST1 (OspC types A and B) and 1 identified as RST2 (OspC type K), were sequenced at the Massachusetts General Hospital DNA Core Facility, and the results of sequencing confirmed the RFLP typing results. In the 30 samples with positive OspC, but negative RST determinations, DNA was extracted again from a second aliquot of joint fluid (100 μl) and reassessed by PCR and RFLP analysis in a second effort to determine the RST directly; however, an RST could be determined in only 4 additional samples. If the result was still negative, the RST type was inferred from the OspC sequence (17–19) for clinical correlations.

Since the RST type of OspC type J strains has not been reported previously, the sequences of type J strains (GenBank accession nos. DQ437444, EF537359, EF537360, EF53736, and EF537366) were determined by BLAST (27), and were used in a search of GenBank (28) to establish the type J RST sequences. The RST sequences were then analyzed with the NEBcutter V2.0 shareware program (29), which predicts restriction digestion patterns. The predicted RFLP pattern for the 5 OspC type J strains most closely resembled those of RST3 strains.
Statistical analysis.
Categorical variables were compared by chi-square test. The distribution of values between groups was analyzed first by three-way analysis of variance, and if significant differences were shown, differences between groups were assessed using the Mann-Whitney rank sum test. All P values are 2-tailed. P values less than or equal to 0.05 were considered significant.

B burgdorferi OspC and RST types in joint fluid.
As determined by PCR, 74 of the 124 joint fluid samples (60%) had an OspC or RST band, or both, on agarose gels. An OspC type was then identified in 49 samples (40%) using sequencing, an RST was identified in 23 samples (19%) using RFLP analysis, and both an OspC type and an RST were determined directly in 22 samples (18%). In each of these 22 samples, the RST matched that predicted from the ospC gene sequence, as shown previously with B burgdorferi isolates from the northeastern US (17, 18). For the remaining 27 samples in which it was possible to determine the OspC type, but not the RST, the RST was inferred based on the ospC gene sequence. In the 1 sample in which only the RST could be determined, the OspC type was not known because a single RST includes multiple OspC types. Altogether, the OspC and RST types were determined in 49 samples (40%), and the RST alone was identified in 1 additional sample.

Among the 49 joint fluid samples in which both the OspC type and the RST were known, 10 of the 16 OspC types found in the northeastern US and all 3 RSTs were identified. Of these 49 samples, OspC type K strains (RST2) were identified in 21 samples (43%), OspC type A strains (RST1) were identified in 11 samples (22%), and 8 other OspC types and all 3 RSTs were identified among the remaining 17 samples (35%) (Table 1). The greater frequency of OspC type K strains than type A strains was statistically significant (P = 0.05). The distribution of RSTs was similar to that in the samples in which the RST type was determined directly or inferred from the ospC gene sequence (data not shown). Infection with mixed OspC or RST types was not observed. Altogether, nearly half of the patients with Lyme arthritis in whom B burgdorferi genotypes could be determined were infected with OspC type K (RST2) strains, and nearly one-quarter had OspC type A (RST1) strains.

Table 1. OspC and RST types in patients with EM skin lesions from 3 previous studies compared with genotypes in joint fluid samples from patients with Lyme arthritis in the current study* RST type (1, 2, or 3), OspC type Previous studies of EM skin isolates Current study of joint fluid, 1976–2006 (n = 50)‡
Ref.5, pre-1999 (n = 118)† Ref.18, 1991–2005 (n = 290) Ref.17, 1998–2001 (n = 90)
*Values are the number (%) of patients. There were no significant differences among the 4 study groups in the frequencies of any OspC type, as calculated in 2 × 4 tables by chi-square analysis. However, the frequency of ribosomal RNA intergenic spacer type 3 (RST3) strains differed significantly among the 4 groups (P = 0.005). EM = erythema migrans.
†Only OspC typing was done, and the RST was inferred from the OspC sequence.
‡One RST3 strain yielded no usable OspC sequence.
§P = 0.05 by chi-square test for the comparison of OspC types K and A in joint fluid.

1
A 23 (19) 46 (16) 27 (30) 11 (22)§
B 19 (16) 37 (13) 11 (12) 5 (10)
Total 42 (36) 83 (29) 38 (42) 16 (33)
2
F 0 9 (3) 1 (1) 1 (2)
H 6 (5) 13 (4) 4 (4) 5 (10)
K 32 (27) 86 (30) 25 (27) 21 (43)
N 3 (3) 17 (6) 9 (10) 1 (2)
Total 41 (35) 125 (43) 39 (43) 28 (57)
3
C 3 (3) 2 (0.7) 0 1 (2)
D 1 (1) 4 (1.4) 1 (1) 0
E 1 (1) 14 (5) 3 (3) 1 (2)
G 7 (6) 14 (5) 2 (2) 2 (4)
I 9 (8) 20 (7) 7 (8) 0
J 7 (6) 3 (1) 0 1 (2)
M 3 (3) 11 (4) 0 0
O 1 (1) 1 (0.3) 0 0
T 1 (1) 2 (0.7) 0 0
U 2 (2) 11 (4) 0 0
Total 35 (30) 82 (28) 13 (14) 6 (12)

Genotype frequencies in joint fluid were similar to those reported in 3 previous studies of EM skin lesions (5, 17, 18) (Table 1). When the frequencies of OspC types in these previous studies and in the current study were compared, there were no significant differences among the groups. Because stratification by RST leads to larger groups since there are only 3 RSTs, the frequencies of RST3 strains in joint fluid were fewer than in EM skin isolates (P = 0.005), but this difference was seen only in comparison with the studies by Seinost et al (5) and Wormser et al (18), in which larger numbers of unusual OspC types were identified in a few patients each. Thus, we concluded that the frequencies of B burgdorferi genotypes in joint fluid reflected those in EM skin lesions.

Clinical correlations with positive typing results.
Clinical information was available on 119 of the 124 patients, and in 48 of the 119 patients (40%), the OspC type and RST were known. To identify clinical factors associated with positive typing results, clinical data on the 48 patients for whom the OspC type and RST were known were compared with data on the 71 with negative OspC and RST typing results (Table 2). Age, sex ratio, duration of arthritis prior to the date of joint fluid collection, and total duration of arthritis were similar in the 2 groups. However, prior to the date of sample collection, patients with positive results more often received oral or intraarticular steroids and had less often been given antibiotics. In addition, on the date of sample collection, these patients tended to have larger joint effusions, and they had significantly higher joint fluid white cell counts.

Table 2. Demographic and clinical characteristics of the patients with Lyme arthritis, according to positive or negative results on OspC and RST typing of joint fluid samples* Positive OspC and RST typing (n = 48)† Negative OspC and RST typing (n = 71)
*Except where indicated otherwise, values are the number (%) of patients. RST = ribosomal RNA intergenic spacer type.
†No clinical information was available on 2 of the 50 patients whose OspC and RST typing yielded positive results, and on 3 of the 74 patients whose OspC and RST typing yielded negative results.
‡P ≤ 0.05 versus patients with negative typing results.

Age, median (range) years 35.5 (7.0–68) 34.0 (9.0–79)
Male 36 (75) 46 (65)
Year of infection
1976–1985 27 (56)‡ 25 (35)
1986–1995 4 (8)‡ 17 (24)
1996–2006 17 (35) 29 (41)
Duration of arthritis prior to date of sample collection, median (range) months 2.0 (0–19) 3.0 (0–24)
Size of joint effusion, median (range) ml 50 (15–200) 40 (4–190)
Joint fluid white cell count, median (range) 21,500 (4,550–110,000)‡ 12,600 (60–230,000)
Total duration of arthritis, median (range) months 11 (1.0–71) 14 (0–102)
Oral or intraarticular steroid therapy prior to sample collection 27 (56)‡ 24 (34)
Antibiotic therapy prior to sample collection 17 (35)‡ 41 (58)

Primarily during the initial study years (1976–1985), before the cause of the disease was known and before antibiotic trials were completed (22), patients were treated with intraarticular steroids. During this period, joint fluid samples from the majority of patients yielded positive typing results, whereas after this period, samples from the majority of patients yielded negative results (Table 2). Thus, it appears that larger numbers of spirochetes were present in the joint fluid of patients who received intraarticular steroids, and larger numbers of organisms resulted in larger joint effusions and higher white blood cell counts.
B burgdorferi genotypes and clinical course.

To determine whether the B burgdorferi genotype influenced the clinical features of Lyme arthritis, the 48 patients for whom the OspC type and RST were known were stratified into 3 groups according to results of RST typing (Table 3). During the first 10-year period (1976–1985), OspC type K (RST2) strains were most often identified, whereas during the most recent 10-year period (1996–2006), there was a more equal distribution among the 3 RSTs. Age, sex ratio, duration of arthritis prior to the date of sample collection, and treatment prior to that date were similar among the 3 groups. However, the size of the joint effusion tended to be larger and the total duration of arthritis tended to be longer in the 42 patients infected with RST1 or RST2 strains than in the 6 patients infected with RST3 strains, and the total duration of arthritis was significantly longer in those infected with RST1 strains than in those infected with RST3 strains (P = 0.05).

Table 3. Demographic and clinical characteristics of the patients with Lyme arthritis, according to Borrelia burgdorferi genotype* RST1 (OspC A, B) (n = 16) RST2 (OspC F, H, K, N) (n = 26)† RST3 (OspC C, E, G, J) (n = 6)
*Except where indicated otherwise, values are the number (%) patients. RST = ribosomal RNA intergenic spacer type.
†No clinical information was available on 2 of the 21 patients with the OspC type K or RST2 genotype.
‡P = 0.05 versus RST1 group.

Age, median (range) years 34 (7–67) 31 (9–68) 47 (37–57)
Male 11 (69) 21 (81) 4 (67)
Year of infection
1976–1985 8 (50) 17 (65) 2 (33)
1986–1995 1 (6) 3 (12) 0
1996–2006 7 (44) 6 (23) 4 (67)
Duration of arthritis prior to date of sample collection, median (range) months 3.0 (0–18) 2.0 (0–19) 1.0 (0–4.0)
Size of joint effusion, median (range) ml 53 (15–100) 48 (15–200) 30 (15–80)
Joint fluid white cell count, median (range) 17,446 (6,877–110,000) 23,250 (4,550–98,000) 24,600 (17,000–40,556)
Total duration of arthritis, median (range) months 13 (2.0–64) 12 (1.0–71) 5.25 (1.5–14)‡
Oral or intraarticular steroid therapy prior to sample collection 6 (38) 17 (65) 4 (67)
Antibiotic therapy prior to sample collection 7 (44) 7 (27) 3 (50)
B burgdorferi genotypes and antibiotic-refractory arthritis.
Of the 48 patients for whom OspC and RST type was known, 17 seen after 1986 received antibiotic therapy according to the current guidelines of the IDSA (3), which allowed these patients to be categorized into either an antibiotic-responsive or an antibiotic-refractory group (Table 4). Among the 17 patients, all 7 infected with RST1 strains had antibiotic-refractory arthritis, compared with 4 of 6 patients infected with RST2 strains and only 1 of 4 patients infected with RST3 strains (P = 0.03 for the comparison of all 3 groups). Patients in the 3 groups had arthritis for a median duration of 1–3 months prior to diagnosis. However, the median duration of arthritis after the start of therapy was significantly longer in those with RST1 strains (10 months), intermediate in those with RST2 strains (6 months), and shortest in those with RST3 strains (2 months) (P = 0.03) (Table 4). All 17 patients had positive PCR results for B burgdorferi DNA in joint fluid prior to or during antibiotic treatment. However, in the 6 patients with antibiotic-refractory arthritis in whom joint fluid was obtained at least once during the postantibiotic period (5 of whom were infected with RST1 strains), PCR results were negative.

Table 4. Clinical characteristics of the 17 Lyme arthritis patients who received currently recommended antibiotic regimens, according to Borrelia burgdorferi genotype* RST1 (OspC A, B) RST2 (OspC F, H, K, N) RST3 (OspC C, E, G, J)
*Currently recommended antibiotic regimens are described in ref.3.
†P = 0.03 for the comparison of all 3 groups; P = 0.025, ribosomal RNA intergenic spacer type 1 (RST1) group versus RST3 group.
‡P = 0.03 for the comparison of all 3 groups.

No. of patients with antibiotic-responsive arthritis 0 2 3
No. of patients with antibiotic-refractory arthritis† 7 4 1
Duration of arthritis, median (range) months
Prior to antibiotic therapy 1 (0.5–12) 1.5 (0.5–8) 3 (1–7)
After the start of antibiotic therapy‡ 10 (4–30) 6 (1–11) 2 (1–4)

As described previously (7, 17, 18) and as shown in Figure 1 using data from our previous study (17), patients with EM skin lesions who were infected with RST1 strains, compared with those infected with RST2 or RST3 strains, were more likely to exhibit evidence of spirochetes in the blood (Figure 1A). Similarly, in the current study, a larger number of patients with Lyme arthritis who were infected with RST1 strains had antibiotic-refractory arthritis than did patients who were infected with RST2 or RST3 strains (Figure 1B). Thus, OspC type K (RST2) strains were most often found in the joints of patients with Lyme arthritis, but RST1 strains were most frequent in those with antibiotic-refractory arthritis.

Figure 1. Distribution of ribosomal RNA intergenic spacer types (RSTs) in patients with erythema migrans (EM) skin lesions or Lyme arthritis. A, Patients with EM skin lesions (n = 90; see ref.17) stratified according to those without a positive test result (open bars) and those with a positive test result (solid bars) for Borrelia burgdorferi DNA in the blood. B, The subgroup of Lyme arthritis patients from the present study who received currently recommended antibiotic regimens, stratified according to those with antibiotic-responsive arthritis (open bars) and those with antibiotic-refractory arthritis (solid bars). OspC types corresponding to each RST are shown in parentheses.

In the current study, 10 of the 16 B burgdorferi OspC types found in the northeastern US and all 3 RST types were identified in the joint fluid of 49 of 124 patients (40%) with Lyme arthritis seen during a 30-year period. Previous OspC or RST typing was done using culture isolates obtained from EM skin lesions, blood, or CSF samples, in which unlimited amounts of spirochetal DNA were available (5, 6, 11, 16). However, it has been difficult to culture B burgdorferi from joint fluid in patients with Lyme arthritis (30, 31), and the OspC type of only 1 joint fluid isolate (type A) has been reported (5). For this study, B burgdorferi OspC and RST types were determined directly from joint fluid samples rather than culture isolates, using PCR and sequencing or RFLP analysis. Because only very small amounts of spirochetal DNA were present, nearly half of the samples in which typing could be done were obtained prior to the use of antibiotics for the treatment of Lyme arthritis (1975–1980), when therapy with intraarticular steroids, without antibiotics, presumably increased the spirochetal burden.

In our 1994 study (32), B burgdorferi DNA was detected, using PCR, in joint fluid samples from 75 of 88 patients (85%). In the current study, in which many of the same samples were tested, the samples from 74 of 124 patients (60%) had positive PCR results, and in 50 samples (40%), sequencing or RFLP analysis of the PCR product yielded an OspC type or RST. The 1994 study included a larger percentage of samples obtained in the period before antibiotic treatment of Lyme arthritis (1975–1980) and used more primer sets and a radiolabeled internal probe for the detection of gene sequences. In addition, there may have been some degradation of spirochetal DNA during the 15-year period since the 1994 study was completed, as suggested by the small number of samples with positive results from the middle years (1986–1995), when currently recommended antibiotic regimens were first used.

Were the results of the current study skewed by the fact that it was possible to determine the B burgdorferi genotypes in only 40% of the 124 joint fluid samples? We think that this result was caused primarily by the fact that very little or no spirochetal DNA was present in many joint fluid samples, particularly those obtained after the start of antibiotic therapy. However, in the samples in which typing could be done, 10 of the 16 OspC types and all 3 RST types were identified, and the distribution of types in joint fluid was similar to that reported in previous studies of EM skin lesions (5, 7, 14, 16, 17). Thus, we feel confident that the distribution of B burgdorferi genotypes in joints reflects that in skin.

In a recent analysis of B burgdorferi isolates, it was postulated that the spread of a high-virulence OspC type A clone may have contributed to the rise in the incidence of Lyme disease (33). In the current study, OspC type K strains originally comprised a large proportion of the B burgdorferi genotypes in joints, whereas in recent years, the proportion of OspC type A strains seemed to increase. However, in the context of the current study, this apparent trend could have also resulted from changes in the referral of patients. Increasingly, we were referred patients with persistent arthritis after antibiotic therapy (4), and these patients were more often infected with RST1 strains.

The relapsing fever spirochete, Borrelia turicatae, has a structural homolog of OspC called variable surface protein (Vsp), which seems to have a role in tissue tropism (34, 35). In a mouse model, VspA strains were neurotropic (34, 36), whereas VspB strains were arthritogenic and were associated with greater numbers of spirochetes in blood (34, 37). Because B burgdorferi OspC type A (or RST1) strains were identified more often in the blood of patients with early, disseminated disease (5, 7, 16, 17), and because OspC type K strains were found slightly more often in CSF isolates from 20 patients with neuroborreliosis (5), we initially hypothesized that OspC type A strains might be found more often in the joints of patients with Lyme arthritis, analogous to observations with B turicatae (34–37). Instead, OspC type K strains were found more frequently in both skin and joint fluid. Thus, our data do not suggest that OspC plays a role in tissue tropisms at different sites of dissemination in Lyme disease.

Because RST1 strains are overrepresented in blood early in the course of infection, and RST3 strains are underrepresented (7, 18), it is thought that B burgdorferi genotypes vary in their capacity to disseminate. This observation is supported by the fact that mice infected with RST1 strains by needle inoculation, compared with those infected with RST3 strains, had greater densities of spirochetes in their blood and organs, the spirochetes were present for longer durations, and the mice had more severe organ system involvement (9, 10). However, when mice were infected by tick bite, as in the natural infection, RST1 isolates displayed higher densities in blood, but the number of spirochetes in the heart or bladder was similar for RST1 and RST3 strains (14). This suggests that in the natural infection, smaller numbers of RST3 organisms, which may not be detectable in blood, are still able to spread to the joints and cause infection there. Consistent with this observation, in the current study, the frequencies of B burgdorferi genotypes in human patients were similar in EM skin lesions and in joints. Thus, it seems that all 3 RSTs have a similar predilection for dissemination, but larger numbers of RST1 organisms are more often detectable in blood.

Why are RST1 strains most often associated with antibiotic-refractory Lyme arthritis? We hypothesize that RST1 strains are more virulent, leading to larger numbers of organisms in blood, but they are also more inflammatory. In a recent study, patients with antibiotic-refractory arthritis had significantly higher levels of proinflammatory cytokines and chemokines in joint fluid, including interleukin-6 (IL-6), tumor necrosis factor α, IL-1β, interferon-γ, CXCL9, and CXCL10, than patients with antibiotic-responsive arthritis (25). However, the strong immune response induced by RST1 strains may limit their persistence in the blood or joints. In B burgdorferi–infected C57/BL6 mice not treated with antibiotics, DNA from an RST1 strain was initially found in higher numbers in joints than was DNA from an RST3 strain, but DNA from the RST3 strain persisted longer (Bockenstedt L: personal communication). In the current study, the duration of arthritis after antibiotic therapy when PCR results were negative (4, 32, 38) was significantly longer in patients infected with RST1 strains than in patients infected with RST2 or RST3 strains. Thus, a marked inflammatory response, such as that induced by RST1 strains, may set the stage for joint inflammation that persists after spirochetal killing, particularly in patients with the HLA–DRB1*0401 or *0101 alleles (26) and immunoreactivity with OspA (39–42).

In summary, our results add to the emerging literature concerning the differential pathogenicity of strains of B burgdorferi (5, 7, 16–18). In the current study, most of the B burgdorferi genotypes infected the joints of patients with Lyme arthritis, and genotype frequencies in joints reflected those in EM skin lesions. However, RST1 strains, which appear to be more virulent, were most common in patients with antibiotic-refractory arthritis.

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Steere had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Jones, Steere.

Acquisition of data. Jones, McHugh.

Analysis and interpretation of data. Jones, Glickstein, Steere.

We thank Dr. Linda Bockenstedt for sharing information about RST types in a murine model of Lyme disease, and Colleen Squires for help with preparation of the manuscript.

REFERENCES
1Steere AC. Lyme disease. N Engl J Med 2001; 345: 115–25.CrossRef,PubMed,CAS,Web of Science® Times Cited: 4072Steere AC, Schoen RT, Taylor E. The clinical evolution of Lyme arthritis. Ann Intern Med 1987; 107: 725–31.PubMed,CAS,Web of Science® Times Cited: 4003Wormser GP, Dattwyler RJ, Shapiro ED, Halperin JJ, Steere AC, Klempner MS, et al. The clinical assessment, treatment, and prevention of Lyme disease, human granulocytic anaplasmosis, and babesiosis: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis 2006; 43: 1089–134.CrossRef,PubMed,Web of Science® Times Cited: 1384Steere AC, Angelis SM. Therapy for Lyme arthritis: strategies for the treatment of antibiotic-refractory arthritis. Arthritis Rheum 2006; 54: 3079–86.Direct Link:AbstractFull Article (HTML)PDF(101K)References5Seinost G, Dykhuizen DE, Dattwyler RJ, Golde WT, Dunn JJ, Wang IN, et al. Four clones of Borrelia burgdorferi sensu stricto cause invasive infection in humans. Infect Immun 1999; 67: 3518–24.PubMed,CAS,Web of Science® Times Cited: 986Wang IN, Dykhuizen DE, Qiu W, Dunn JJ, Bosler EM, Luft BJ. Genetic diversity of ospC in a local population of Borrelia burgdorferi sensu stricto. Genetics 1999; 151: 15–30.PubMed,CAS,Web of Science® Times Cited: 877Wormser GP, Liveris D, Nowakowski J, Nadelman RB, Cavaliere LF, McKenna D, et al. Association of specific subtypes of Borrelia burgdorferi with hematogenous dissemination in early Lyme disease. J Infect Dis 1999; 180: 720–5.CrossRef,PubMed,CAS,Web of Science® Times Cited: 708Liveris D, Varde S, Iyer R, Koenig S, Bittker S, Cooper D, et al. Genetic diversity of Borrelia burgdorferi in Lyme disease patients as determined by culture versus direct PCR with clinical specimens. J Clin Microbiol 1999; 37: 565–9.PubMed,CAS,Web of Science® Times Cited: 659Wang G, Ojaimi C, Iyer R, Saksenberg V, McClain SA, Wormser GP, et al. Impact of genotypic variation of Borrelia burgdorferi sensu stricto on kinetics of dissemination and severity of disease in C3H/HeJ mice. Infect Immun 2001; 69: 4303–12.CrossRef,PubMed,CAS,Web of Science® Times Cited: 4610Wang G, Ojaimi C, Wu H, Saksenberg V, Iyer R, Liveris D, et al. Disease severity in a murine model of Lyme borreliosis is associated with the genotype of the infecting Borrelia burgdorferi sensu stricto strains. J Infect Dis 2002; 186: 782–91.CrossRef,PubMed,CAS,Web of Science® Times Cited: 4811Lagal V, Postic D, Ruzic-Sabljic E, Baranton G. Genetic diversity among Borrelia strains determined by single-strand conformation polymorphism analysis of the ospC gene and its association with invasiveness. J Clin Microbiol 2003; 41: 5059–65.CrossRef,PubMed,CAS,Web of Science® Times Cited: 2712Lagal V, Portnoi D, Faure G, Postic D, Baranton G. Borrelia burgdorferi sensu stricto invasiveness is correlated with OspC-plasminogen affinity. Microbes Infect 2006; 8: 645–52.CrossRef,PubMed,CAS,Web of Science® Times Cited: 1113Bunikis J, Garpmo U, Tsao J, Berglund J, Fish D, Barbour AG. Sequence typing reveals extensive strain diversity of the Lyme borreliosis agents Borrelia burgdorferi in North America and Borrelia afzelii in Europe. Microbiology 2004; 150: 1741–55.CrossRef,PubMed,CAS,Web of Science® Times Cited: 4914Dolan MC, Piesman J, Schneider BS, Schriefer M, Brandt K, Zeidner NS. Comparison of disseminated and nondisseminated strains of Borrelia burgdorferi sensu stricto in mice naturally infected by tick bite. Infect Immun 2004; 72: 5262–66.CrossRef,PubMed,CAS,Web of Science® Times Cited: 715Ojaimi C, Mulay V, Liveris D, Iyer R, Schwartz I. Comparative transcriptional profiling of Borrelia burgdorferi clinical isolates differing in capacities for hematogenous dissemination. Infect Immun 2005; 73: 6791–802.CrossRef,PubMed,CAS,Web of Science® Times Cited: 1116Alghaferi MY, Anderson JM, Park J, Auwaerter PG, Aucott JN, Norris DE, et al. Borrelia burgdorferi ospC heterogeneity among human and murine isolates from a defined region of northern Maryland and southern Pennsylvania: lack of correlation with invasive and noninvasive genotypes. J Clin Microbiol 2005; 43: 1879–84.CrossRef,PubMed,CAS,Web of Science® Times Cited: 1917Jones KL, Glickstein LJ, Damle N, Sikand VK, McHugh G, Steere AC. Borrelia burgdorferi genetic markers and disseminated disease in patients with early Lyme disease. J Clin Microbiol 2006; 44: 4407–13.CrossRef,PubMed,CAS,Web of Science® Times Cited: 918Wormser GP, Brisson D, Liveris D, Hanincova K, Sandigursky S, Nowakowski J, et al. Borrelia burgdorferi genotype predicts the capacity for hematogenous dissemination during early Lyme disease. J Infect Dis 2008; 198: 1358–64.CrossRef,PubMed,Web of Science® Times Cited: 819Hanincova K, Liveris D, Sandigursky S, Wormser GP, Schwartz I. Borrelia burgdorferi sensu stricto is clonal in patients with early Lyme borreliosis. Appl Environ Microbiol 2008; 74: 5008–14.CrossRef,PubMed,CAS,Web of Science® Times Cited: 420Centers for Disease Control and Prevention. Case definitions for infectious conditions under public health surveillance. MMWR Recomm Rep 1997; 46: 1–55.21Centers for Disease Control and Prevention. Recommendations for test performance and interpretation from the Second National Conference on Serologic Diagnosis of Lyme Disease. MMWR Morb Mortal Wkly Rep 1995; 44: 590–1.PubMed22Steere AC, Green J, Schoen RT, Taylor E, Hutchinson GJ, Rahn DW, et al. Successful parenteral penicillin therapy of established Lyme arthritis. N Engl J Med 1985; 312: 869–74.CrossRef,PubMed,CAS,Web of Science® Times Cited: 21623Dattwyler RJ, Halperin JJ, Volkman DJ, Luft BJ. Treatment of late Lyme borreliosis: randomized comparison of ceftriaxone and penicillin. Lancet 1998; 1191–4.24Steere AC, Levin RE, Molloy PJ, Kalish RA, Abraham JH III, Liu NY, et al. Treatment of Lyme arthritis. Arthritis Rheum 1994; 37: 878–88.Direct Link:AbstractPDF(954K)References25Shin JJ, Glickstein LJ, Steere AC. High levels of inflammatory chemokines and cytokines in joint fluid and synovial tissue throughout the course of antibiotic-refractory Lyme arthritis. Arthritis Rheum 2007; 56: 1325–35.Direct Link:AbstractFull Article (HTML)PDF(325K)References26Steere AC, Klitz W, Drouin EE, Falk BA, Kwok WW, Nepom GT, et al. Antibiotic-refractory Lyme arthritis is associated with HLA-DR molecules that bind a Borrelia burgdorferi peptide. J Exp Med 2006; 203: 961–71.CrossRef,PubMed,CAS,Web of Science® Times Cited: 3027Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol 1990; 215: 403–10.PubMed,CAS,Web of Science® Times Cited: 2710928Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Wheeler DL. GenBank. Nucleic Acids Res 2006; 34: D16–20.CrossRef,PubMed,CAS,Web of Science® Times Cited: 19829Vincze T, Posfai J, Roberts RJ. NEBcutter: a program to cleave DNA with restriction enzymes. Nucleic Acids Res 2003; 31: 3688–91.CrossRef,PubMed,CAS,Web of Science® Times Cited: 7130Snydman DR, Schenkein DP, Berardi VP, Lastavica CC, Pariser KM. Borrelia burgdorferi in joint fluid in chronic Lyme arthritis. Ann Intern Med 1986; 104: 798–800.PubMed,CAS,Web of Science® Times Cited: 12231Schmidli J, Hunziker T, Moesli P, Schaad UB. Cultivation of Borrelia burgdorferi from joint fluid three months after treatment of facial palsy due to Lyme borreliosis. J Infect Dis 1988; 158: 905–6.PubMed,CAS,Web of Science® Times Cited: 11332Nocton JJ, Dressler F, Rutledge BJ, Rys PN, Persing DH, Steere AC. Detection of Borrelia burgdorferi DNA by polymerase chain reaction in synovial fluid from patients with Lyme arthritis. N Engl J Med 1994; 330: 229–34.CrossRef,PubMed,CAS,Web of Science® Times Cited: 33633Qiu WG, Bruno JF, McCaig WD, Xu Y, Livey I, Schriefer ME, et al. Wide distribution of a high-virulence Borrelia burgdorferi clone in Europe and North America. Emerg Infect Dis 2008; 14: 1097–104.CrossRef,PubMed,CAS,Web of Science® Times Cited: 634Cadavid D, Pachner AR, Estanislao L, Patalapati R, Barbour AG. Isogenic serotypes of Borrelia turicatae show different localization in the brain and skin of mice. Infect Immun 2001; 69: 3389–97.CrossRef,PubMed,CAS,Web of Science® Times Cited: 2635Cadavid D, Pennington PM, Kerentseva TA, Bergstrom S, Barbour AG. Immunologic and genetic analyses of VmpA of a neurotropic strain of Borrelia turicatae. Infect Immun 1997; 65: 3352–60.PubMed,CAS,Web of Science® Times Cited: 4636Pennington PM, Allred CD, West CS, Alvarez R, Barbour AG. Arthritis severity and spirochete burden are determined by serotype in the Borrelia turicatae-mouse model of Lyme disease. Infect Immun 1997; 65: 285–92.PubMed,CAS,Web of Science® Times Cited: 4337Pennington PM, Cadavid D, Barbour AG. Characterization of VspB of Borrelia turicatae, a major outer membrane protein expressed in blood and tissues of mice. Infect Immun 1999; 67: 4637–45.PubMed,CAS,Web of Science® Times Cited: 3338Carlson D, Hernandez J, Bloom BJ, Coburn J, Aversa JM, Steere AC. Lack of Borrelia burgdorferi DNA in synovial samples from patients with antibiotic treatment–resistant Lyme arthritis. Arthritis Rheum 1999; 42: 2705–9.Direct Link:AbstractPDF(125K)References39Drouin EE, Glickstein LJ, Steere AC. Molecular characterization of the OspA(161-175) T cell epitope associated with treatment-resistant Lyme arthritis: differences among the three pathogenic species of Borrelia burgdorferi sensu lato. J Autoimmun 2004; 23: 281–92.CrossRef,PubMed,CAS,Web of Science® Times Cited: 1040Chen J, Field JA, Glickstein L, Molloy PJ, Huber BT, Steere AC. Association of antibiotic treatment–resistant Lyme arthritis with T cell responses to dominant epitopes of outer surface protein A of Borrelia burgdorferi. Arthritis Rheum 1999; 42: 1813–22.Direct Link:AbstractPDF(194K)References41Gross DM, Forsthuber T, Tary-Lehmann M, Etling C, Ito K, Nagy ZA, et al. Identification of LFA-1 as a candidate autoantigen in treatment-resistant Lyme arthritis. Science 1998; 281: 703–6.CrossRef,PubMed,CAS,Web of Science® Times Cited: 282,ADS42Drouin EE, Glickstein L, Kwok WW, Nepom GT, Steere AC. Searching for borrelial T cell epitopes associated with antibiotic-refractory Lyme arthritis. Mol Immunol 2008; 45: 2323–32.CrossRef,PubMed,CAS,Web of Science® Times Cited: 6