Persistent arthritis in Borrelia burgdorferiinfected HLADR4positive CD28-negative mice postantibiotic treatment.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 58, No. 12, December 2008, pp 3892–3901 DOI 10.1002/art.24028 © 2008, American College of Rheumatology Persistent Arthritis in Borrelia burgdorferi–Infected HLA–DR4–Positive CD28-Negative Mice Post–Antibiotic Treatment Bettina Panagiota Iliopoulou,1 Joseph Alroy,2 and Brigitte T. Huber1 Objective. The immunologic events that lead to persistent joint inflammation in certain patients with Lyme arthritis post–antibiotic treatment have been elusive so far. The prevalence of this condition is highest in individuals with rheumatoid arthritis–associated HLA–DR alleles. This study was undertaken to generate a murine model with persistent arthritis post–antibiotic treatment. Methods. We have previously shown that CD28ⴚ/ⴚ mice develop intermittent monarticular Lyme arthritis that is responsive to antibiotics. Since there seems to be a link in humans between persistent arthritic manifestations post–antibiotic treatment and the HLA–DR4 allele, we generated DR4ⴙ/ⴙCD28ⴚ/ⴚMHCIIⴚ/ⴚ mice, infected them with Borrelia burgdorferi, and subsequently treated them with antibiotics. Results. Thirty-eight percent of the B burgdorferi– infected DR4ⴙ/ⴙCD28ⴚ/ⴚMHCIIⴚ/ⴚ mice, but none of the B burgdorferi–infected CD28ⴚ/ⴚMHCIIⴚ/ⴚ mice, remained arthritic post–antibiotic treatment. A significant fraction (36%) of these mice, but none of the mice in which arthritis resolved, had serum antibodies to outer surface protein A of B burgdorferi. After abrogation of active B burgdorferi infection, the inflammatory reaction in mice with persistent joint inflammation was restricted to the joints, since their draining lymph nodes were no longer enlarged. Increased CD20 and interferon-␥ messenger RNA expression in the inflamed joints of these mice suggested a possible role of B cells and inflammatory cytokines in the pathogenesis of persistent arthritis post–antibiotic treatment. Conclusion. The establishment of this murine model allows, for the first time, the elucidation of the immunologic events that lead to persistent Lyme arthritis post–antibiotic therapy in genetically susceptible individuals. Lyme disease, caused by the tick-borne spirochete Borrelia burgdorferi, is the most common vectorborne illness in the US. After inoculation into the skin, B burgdorferi quickly disperses in the mammalian host by binding to components of the extracellular matrix (1). Three clinical stages of Lyme disease have been described in humans. Early infection consists of localized erythema migrans, followed within days or weeks by disseminated infection that affects the nervous system, heart, or joints, and subsequently by late or persistent infection (2). While the spirochetes can be eliminated from patients with Lyme disease by antibiotic treatment, chronic arthritis may persist, mainly in patients with rheumatoid arthritis (RA)–associated HLA–DR alleles, such as HLA–DRB1*0401 (DR4) and HLA– DRB1*0101 (3,4). Two basic hypotheses have been proposed to explain this phenomenon: persistent infection and infection-induced autoimmunity. The latter hypothesis is supported by the fact that manifestations of arthritis continue despite the absence of B burgdorferi DNA, documented by polymerase chain reaction (PCR) analysis of the synovial fluid (5–7). Interestingly, in 70% of the patients who continued to experience arthritis after antibiotic treatment, an antibody response to outer surface protein A (OspA) of B burgdorferi that seemed to parallel the severity and duration of arthritis was Supported by the NIH (grant R01-AR-45386). Tufts Medical Center’s Gastroenterology Research on Absorptive and Secretory Processes (GRASP) Center is funded by National Institute of Diabetes and Digestive and Kidney Diseases, NIH (grant DK-34924). 1 Bettina Panagiota Iliopoulou, PhD, Brigitte T. Huber, PhD: Tufts University School of Medicine, Boston, Massachusetts; 2Joseph Alroy, DVM: Tufts University School of Medicine, and Tufts Medical Center, Boston, Massachusetts. Address correspondence and reprint requests to Brigitte T. Huber, PhD, Department of Pathology, Tufts University School of Medicine, Jaharis 512, 150 Harrison Avenue, Boston, MA 02111. E-mail: firstname.lastname@example.org. Submitted for publication May 5, 2008; accepted in revised form August 1, 2008. 3892 HLA–DR4 AND PERSISTENT ARTHRITIS AFTER B BURGDORFERI INFECTION mounted during periods of maximal arthritis (8,9). In addition, an anti-OspA Th1 response has been documented in the synovial fluid of these patients (10–12). Murine systems have been developed to analyze the immunologic events that occur upon B burgdorferi infection in humans. It is well established that manifestations of arthritis depend on the administered dose of B burgdorferi and the age and genetic background of the mice (13–15). It has also been suggested that T cells, more specifically the CD4⫹ Th1 subset, as well as the proinflammatory cytokine interferon-␥ (IFN␥), are responsible for exacerbation of arthritis upon B burgdorferi infection (16–20). Murine Lyme arthritis peaks within the first 2 weeks after infection and then resolves spontaneously; very few individual mice continue to exhibit chronic arthritis (21,22). These manifestations are reminiscent of the acute phase of Lyme arthritis in humans (15). Detailed characterization of the chronic phase of Lyme arthritis has been elusive thus far. Through the study of the development of Lyme arthritis in different inbred mouse strains, it has become apparent that arthritis severity depends on a fine balance between proinflammatory factors and immunoregulatory mechanisms. We hypothesized that by interfering with this balance, such as in the CD28⫺/⫺ mouse, we could increase the incidence of chronic Lyme arthritis. We have recently demonstrated that CD28⫺/⫺ mice, but not wild-type C57BL/6J (B6) mice, develop chronic Lyme arthritis upon B burgdorferi infection. The persistent episodes of arthritis observed in these mice lasted for ⬎6 months, but were sensitive to antibiotic treatment (23). Since a prerequisite for the development of persistent arthritis post–antibiotic treatment in humans is the presence of HLA–DR4 or related alleles, we introduced the CD28⫺/⫺ genotype onto the DR4⫹/⫹MHCII⫺/⫺ background. In the present study we showed that a significant fraction of B burgdorferi–infected DR4⫹/ ⫺/⫺ ⫹CD28 MHCII⫺/⫺ mice continue to manifest arthritis after antibiotic therapy. One-third of these mice, but none of the mice in which arthritis had resolved, maintained an OspA antibody titer after antibiotic treatment. Furthermore, we showed that CD20 and IFN␥ expression are increased in the joints of DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mice, which suggests that B cells and inflammatory cytokines may be involved in the perpetuation of inflammation. This animal system will allow, for the first time, the direct examination of the inflammatory events that lead to persistent Lyme arthritis after antibiotic treatment. 3893 MATERIALS AND METHODS Mice. CD28⫺/⫺ mice (B6.129S2-Cd28tm1Mak/J) were bred at the Tufts University Division of Laboratory Animal Medicine from breeding pairs (stock no. 002666) obtained from The Jackson Laboratory (Bar Harbor, ME) and confirmed to have been backcrossed for at least 10 generations to the B6 background. DR4⫹/⫹MHCII⫺/⫺ transgenic mice were a gift from T. Forsthuber (Case Western Reserve University, Cleveland, OH) and were bred in our facility. These mice were generated with HLA–DRA–IE␣ and HLA–DRB1*0401–IE␤ chimeric genes and were also on a B6 background (24). DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice were generated by breeding the CD28⫺/⫺ genotype onto DR4⫹/⫹MHCII⫺/⫺ mice. The F1 DR4⫹/⫺CD28⫹/⫺MHCII⫹/⫺ mice were backcrossed to the CD28⫺/⫺ mice, and the DR4⫹/⫺CD28⫺/⫺MHCII⫹/⫺ mice were selected by fluorescence-activated cell sorting (FACS) and then intercrossed to generate DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice. All offspring were screened by FACS analysis for surface expression of DR4, CD28, and I-Ab molecules (all antibodies were purchased from PharMingen, San Diego, CA). All animal experiments were approved by the Institutional Animal Care and Use Committee at Tufts Medical Center. B burgdorferi. Low-passage (passage-2) infectious B burgdorferi N40 clone D10E9A1-E (a kind gift from Dr. Jenifer Coburn, Medical College of Wisconsin, Milwaukee) (25,26) was used for all infections. B burgdorferi were cultured in complete BSK medium (Sigma, St. Louis, MO) at 34°C until mid-log phase (5 ⫻ 107 B burgdorferi/ml) and were counted by darkfield microscopy. Development of a murine model of persistent Lyme arthritis after antibiotic treatment. Sex-matched DR4⫹/⫹ CD28 ⫺/⫺ MHCII ⫺/⫺ , CD28 ⫺/⫺ , and DR4 ⫹/⫹ MHCII ⫺/⫺ mice (4–5 weeks old) were infected intradermally in the skin of the femoral area of both hind limbs with a total dose of 2 ⫻ 104 B burgdorferi per mouse (1 ⫻ 104 B burgdorferi in 50 l per hind limb). This protocol was used for all of the infections. No difference in arthritis development was observed between male and female mice; therefore, both male and female mice were included in the experiments. For each experiment, however, great care was taken that all of the mice in the experimental group (DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice) were sex- and age-matched with those of the control groups (CD28⫺/⫺ mice and DR4⫹/ ⫺/⫺ ⫹MHCII mice). Arthritis was assessed, based on edema formation, 2–3 times per week by an observer (BPI) who was blinded with regard to experimental group. The anteroposterior tibiotarsal joint thickness was measured using gauge calipers (Mahr Federal, Providence, RI). Upon establishment of chronic arthritis (within ⬃3 months after B burgdorferi infection), ceftriaxone (50 mg/kg/dose once a day for 5 days) was administered intraperitoneally. This treatment has been reported to be 100% effective in eradicating B burgdorferi (27,28). Mice were monitored for the presence of arthritis over a period of 2–5 months after antibiotic therapy. Since we previously established that when mice on the CD28⫺/⫺ background are not treated with antibiotics they continue to experience intermittent episodes of monarticular arthritis for at least 5 months (which is the longest they have been tested) (23), we did not include a group of mice that were not treated with antibiotics in this set of data. In addition, when DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice were infected with B burg- 3894 dorferi, but not treated with antibiotics, they developed chronic joint inflammation similar to that in CD28⫺/⫺ mice. At the end of the experiment, mice were killed by CO2 asphyxiation. Hearts and ankles were harvested and processed for histologic analysis. We also observed the development of anti-OspA antibody throughout the course of the experiment. Mice were bled weekly, and serum was subjected to enzyme-linked immunosorbent assay (ELISA) for the presence of anti-OspA antibodies. In order to assess the effectiveness of the antibiotic therapy, DNA was extracted from ear punch tissue before treatment with antibiotics and at the end of the experiment. Experiments were repeated 4 times. Histopathologic analysis. Ankles were harvested and decalcified overnight in Decalcifier I solution (SurgiPath, Richmond, IL). The hearts were cut in half through bisection across the atria and ventricles, and kept in formalin. Both ankles and hearts were processed for histologic analysis as previously described (23). Anti-OspA ELISA. Flat-bottomed Immulon 2HB plates (Fisher Scientific, Pittsburgh, PA) were coated overnight with 5 g/ml of recombinant OspA or OspA fragments in binding buffer (0.1M Na2HPO4 [pH 9]). ELISA was performed as previously described (23). The cutoff used to calculate the end point titer was set as the lowest serum concentration, in which the optical density unit signal was twice that of the background wells (all reagents included, expect the serum) for each isotype. In vitro restimulation and interleukin-17 (IL-17) ELISA. Popliteal as well as inguinal lymph node cells were stimulated with anti-CD3 (145.2C11) at a 1:150 dilution for 48 hours in vitro, and IL-17 production was assayed by ELISA. IL-17 ELISA was performed as previously described (23), with the following modifications. Plates were coated overnight with 3 g/ml of capture anti-mouse IL-17 antibody (R&D Systems, Minneapolis, MN) in 1⫻ phosphate buffered saline (PBS). Coated plates were blocked with 2% bovine serum albumin and 5% sucrose in 1⫻ PBS at room temperature for 1 hour. Recombinant mouse IL-17 (standard curve) and the supernatant from the in vitro restimulation were added in duplicate to the ELISA plates and incubated for 45 minutes at 37°C. Plates were washed and incubated with biotinylated anti-mouse IL-17 (R&D Systems) for 1 hour at 37°C, followed by another wash and incubation with neutrAvidin–alkaline phosphatase for 30 minutes at room temperature. Plates were then developed as previously described (23). Determination of B burgdorferi burden. DNA was extracted from ear punch tissue and the B burgdorferi burden was determined by real-time quantitative PCR, as previously described (23,29). Real-time reverse transcriptase–PCR (RT-PCR) analysis of messenger RNA (mRNA) for IL-17, IFN␥, and CD20. Mouse ankles were harvested and immediately frozen in liquid nitrogen. Frozen tissue was pulverized using a mortar and pestle precooled in liquid nitrogen. RNA from pulverized ankles and popliteal as well as inguinal lymph node cells was extracted using the RNeasy Mini kit, according to the recommendations of the manufacturer (Qiagen, Valencia, CA). Complementary DNA (cDNA) synthesis and removal of genomic DNA were performed using the QuantiTect reverse transcription kit (Qiagen). The cDNA was then diluted 1:20 and was used as template in a 20-l reaction mixture contain- ILIOPOULOU ET AL ing primers and probes specific for IL-17 mRNA, IFN␥ mRNA, CD20 mRNA, and 18S. Primers and FAM-labeled probes specific for IL-17 mRNA (Mm00439619-m1), IFN␥ mRNA (Mm00801778-m1), and CD20 mRNA (Mm00545909m1), as well as VIC-labeled probes specific for murine 18S, were purchased from Applied Biosystems (Foster City, CA) and used as loading control. The PCR was carried out using iTaq Supermix with ROX (Bio-Rad, Hercules, CA) under the same cycling parameters as described above. The amount of template DNA was first normalized by the signal of the 18S housekeeping gene. Statistical analysis. Statistical analysis was performed using GraphPad Prism software (GraphPad Software, San Diego, CA). All data were tested for Gaussian distribution, using the Shapiro-Wilk normality test. Quantitative differences were assessed by Student’s 2-tailed t-test for comparisons of 2 groups, and by analysis of variance for comparisons of ⬎2 groups, for normally distributed data. For skewed data, the Mann-Whitney test was used for comparisons of 2 groups, and the Kruskal-Wallis test was used for comparisons of ⬎2 groups. Statistical differences in the proportion of mice with persistent joint inflammation post–antibiotic treatment as well as in the proportion of mice with anti-OspA antibodies were also determined by Fisher’s 2-tailed exact probability test. Correlation between 2 variables was assessed by calculating Pearson’s correlation coefficient for normally distributed data, or Spearman’s correlation coefficient for skewed data. P values less than 0.05 (2-tailed) were considered significant. RESULTS Persistent Lyme arthritis in DR4ⴙ/ⴙCD28ⴚ/ⴚ MHCIIⴚ/ⴚ mice after antibiotic treatment. DR4⫹/⫹ CD28⫺/⫺MHCII⫺/⫺ mice were generated (Figure 1) and subsequently infected with B burgdorferi, while DR4⫹/⫹ MHCII⫺/⫺ and CD28⫺/⫺ mice were used as controls. Upon establishment of chronic arthritis in these mice, intraperitoneal ceftriaxone, a treatment that has been reported to be 100% effective in eradicating B burgdorferi (27,28), was administered, and mice were monitored for the presence of arthritis for 2–5 months after antibiotic therapy. While joint inflammation was completely eradicated after antibiotic treatment in the CD28⫺/⫺ mice, a significant fraction of the DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mice (38%) remained arthritic (Figures 2A and B). The effectiveness of antibiotic therapy in all groups was shown by the fact that virtually none of the mice in the control groups (0 of 25 CD28⫺/⫺ mice and 1 of 31 DR4⫹/⫺MHCII⫺/⫺ mice) had joint inflammation after treatment with antibiotics, as opposed to the DR4⫹/⫺ CD28⫺/⫺MHCII⫺/⫺ group. If persistent spirochetes that would elicit an inflammatory reaction in the joints of these mice were present, the percentage of edema formation after antibiotic treatment should have been the same in all HLA–DR4 AND PERSISTENT ARTHRITIS AFTER B BURGDORFERI INFECTION Figure 1. Phenotype of the DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice. HLA– DR4 and I-Ab surface expression on B220⫹-gated DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mouse lymphocytes was determined by fluorescenceactivated cell sorting. For the CD28 staining, the B220⫹ lymphocytes were gated out. groups. The elimination of B burgdorferi in these mice systemically was also confirmed by real-time quantitative PCR (Figure 2A), indicating that chronic arthritis may occur in DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice in the absence of B burgdorferi. However, the possibility of persistent infection due to a small number of spirochetes that survived in the microenvironment of the joint could not be excluded. In accordance with observations made in human patients with Lyme disease who had persistent arthritis after antibiotic therapy, these mice continued to manifest mostly persistent and in some cases intermittent episodes of arthritis, which was monarticular in some mice, after antibiotic therapy (Figure 3A). 3895 More specifically, 8 of 11 mice had prolonged and persistent joint inflammation (ankle width 2.9–3.1 mm) that did not resolve with time. In addition, in the 3 remaining mice, we never observed a remission phase during which inflammation was completely gone. Instead, we observed a fluctuation of ankle width that ranged from 2.7 to 3.1 mm. These results imply that the presence of HLA–DR4, or a related RA-associated DR allele, is necessary for the development of persistent Lyme arthritis after antibiotic treatment in a murine model. Joint inflammation over the course of arthritic disease in mice is routinely measured with calipers, which allows monitoring of edema formation. To examine whether cellular infiltration occurred in the inflammatory process that was established in these arthritic mice after antibiotic treatment, we performed histologic analyses of their ankles at the end of the experiment, 4 months after antibiotic therapy. Of 11 mice with persistent joint inflammation post–antibiotic treatment, 6 (55%) showed cellular infiltration, mostly of neutrophils, as opposed to the synovial infiltrate in human patients, which contains mainly lymphocytes and macrophages and very few neutrophils (2) (Figure 3B). This observation is consistent with the data obtained using calipers, since not all of the mice had continuous joint inflammation; instead, some exhibited recurring ankle swelling over time that could be attributed only to Figure 2. Persistence of Lyme arthritis in DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice post–antibiotic treatment. DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺, CD28⫺/⫺, and DR4⫹/⫹MHCII⫺/⫺ mice were infected with Borrelia burgdorferi (Bb), and arthritis was assessed by measuring the ankles using calipers. On day 55 after infection, mice were treated with antibiotics. Mice were monitored for arthritis for 2 months. A, Arthritis incidence and B burgdorferi burden in the 3 mouse strains. A significant fraction of DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice (11 of 29) remained arthritic post–antibiotic treatment, compared with CD28⫺/⫺ mice (0 of 25) and DR4⫹/⫹MHCII⫺/⫺ mice (1 of 31). All mice were negative for the B burgdorferi recA gene post–antibiotic treatment. The median and 25th to 75th percentile B burgdorferi burden before antibiotic treatment for each mouse strain is shown. Data are the pooled results of 4 independent experiments. Differences in B burgdorferi burden were not statistically significant (P ⫽ 0.6 by Kruskal-Wallis test). ⴱ ⫽ P ⫽ 0.004; ⴱⴱ ⫽ P ⫽ 0.0004, by Fisher’s 2-tailed exact probability test. B, Mean ⫾ SEM ankle width of 3 representative mice per group over 120 days. ⴱ ⫽ P ⬍ 0.001 versus mock-infected mice, by Student’s 2-tailed t-test. 3896 ILIOPOULOU ET AL Figure 3. Development of persistent or intermittent arthritis, which was monarticular in some cases, in antibiotic-treated DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mice. A, Width of the left (L) and right (R) ankles of 3 representative mice per group before and after antibiotic treatment. Red, green, and blue lines represent individual mice. Shaded areas represent mock-infected mice. B, Representative hematoxylin and eosin–stained joint sections from a CD28⫺/⫺ mouse (a and b) and a DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mouse (c and d) post–antibiotic treatment. b and d are higher-magnification views of the boxed areas in a and c, respectively. Mice were killed 4 months after antibiotic therapy. Heavy infiltration, mostly of neutrophils, occurred in the joints of DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice. edema formation. Besides cellular infiltration, we did not detect any significant changes in the synovial tissue of these mice. Another manifestation of Lyme disease in B burgdorferi–infected mice is carditis, an inflammatory response characterized by cellular infiltration of macrophages (30). To exclude the possibility that the edema in the joints of these mice was due to obstruction of lower limb circulation caused by carditis, we performed histologic analysis of their hearts. None of the mice examined showed cellular infiltration in the heart, suggesting that edema formation was due to continued inflammation of the joints, even after elimination of B burgdorferi with antibiotics (results not shown). Incidence of anti-OspA antibodies in DR4ⴙ/ⴙ ⴚ/ⴚ CD28 MHCIIⴚ/ⴚ mice with persistent joint inflammation after antibiotic treatment. As mentioned above, in 70% of patients with chronic Lyme arthritis, an anti-OspA immune response is mounted that seems to correlate with periods of maximal arthritis (8,9). Therefore, we monitored the OspA serum antibody titers in our murine model before and after treatment. AntiOspA antibodies were observed in the majority of B burgdorferi-infected DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice before antibiotic therapy. Following antibiotic treatment, anti-OspA was retained in a significant proportion of the mice with persistent joint inflammation (36%), as opposed to none of the mice that no longer had inflammation in the joints (Figure 4). In addition, none of the Figure 4. Incidence of anti–outer surface protein A (anti-OspA) antibodies (Ab) in DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation post–antibiotic treatment. Anti-OspA antibody levels in DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice were determined by enzyme-linked immunosorbent assay before and after antibiotic treatment. Squares represent individual mice. Horizontal lines represent the median. Numbers in parentheses are the absolute numbers of DR4⫹/⫹ CD28⫺/⫺MHCII⫺/⫺ mice with anti-OspA antibodies. ⴱⴱ ⫽ P ⫽ 0.003 by Mann-Whitney test; ⴱ ⫽P ⫽ 0.014 by Fisher’s 2-tailed exact probability test. HLA–DR4 AND PERSISTENT ARTHRITIS AFTER B BURGDORFERI INFECTION 3897 Figure 5. Absence of active immune response in the draining lymph nodes of mice with persistent joint inflammation post–antibiotic treatment. A, Mean and SEM absolute number of popliteal lymph node cells in mock-infected mice, CD28⫺/⫺ mice actively infected with Borrelia burgdorferi, and antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation. Data are the pooled results of 3 independent experiments. B, Mean and SEM expression of interleukin-17 (IL-17) mRNA in draining popliteal and in inguinal lymph node cells in B burgdorferi– infected CD28⫺/⫺ mice (left) and in antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice (right), as assessed by real-time reverse transcriptase– polymerase chain reaction. Data were normalized to the 18S housekeeping gene and are representative of 3 independent experiments. C, Mean and SEM expression of IL-17 cytokine in draining popliteal and in inguinal lymph node cells in B burgdorferi–infected CD28⫺/⫺ mice (left) and in antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice (right), upon anti-CD3 stimulation for 48 hours, as assessed by IL-17 enzyme-linked immunosorbent assay. Data are the pooled results of 3 independent experiments. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.001, by Student’s 2-tailed t-test. CD28⫺/⫺ or DR4⫹/⫹MHCII⫺/⫺ mice had an antibody response to OspA post–antibiotic treatment (data not shown). No significant differences were observed between mice that had anti-OspA antibodies and those that did not. Since mice with persistent joint inflammation were bled weekly, they were further analyzed post–antibiotic therapy to determine whether there was a correlation between arthritis development and anti-OspA antibody titers. Consistent with the observations made in patients with chronic Lyme arthritis, there was a significant correlation over time between anti-OspA antibodies and development of arthritis post–antibiotic treatment in these mice (r ⫽ 0.53, P ⬍ 0.0001) (results not shown). In addition, we determined the isotype of the OspA-specific antibodies. We observed IgM and IgG3, as well as Th1-dependent IgG2c (results not shown). Based on the notion that IgG3 class switching can occur in the presence of IFN␥ (31), we speculated that the inflammatory milieu was biased toward a Th1 response. The complete absence of Th2-dependent IgG1 antibodies in all 4 mice with anti-OspA antibodies and persistent joint inflammation may also be indicative of a Th1/Th2 imbalance. Absence of active immune response in the draining lymph nodes of mice with persistent joint inflammation after antibiotic treatment. To investigate the cellular mechanisms that contribute to persistent inflammation post–antibiotic treatment, we assessed the dynamics and the cytokine profile of the draining popliteal, as well as the inguinal, lymph node cells in B burgdorferi– infected, antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice. B burgdorferi–infected CD28⫺/⫺ mice were used as a positive control. The size and absolute number of popliteal lymph node cells was much larger in the B burgdorferi–infected CD28⫺/⫺ mice with active arthritis compared with the B burgdorferi–infected CD28⫺/⫺ mice that no longer had inflammation in the joints. In contrast, the absolute number of popliteal lymph node cells in the antibiotic-treated DR4 ⫹/⫹ CD28 ⫺/ ⫺/⫺ ⫺MHCII mice with persistent arthritis was identical to that in the DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice in which arthritis was resolved (Figure 5A). A similar pattern was observed in the inguinal lymph nodes (data not shown), 3898 ILIOPOULOU ET AL this observation in an animal model in vivo. To assess CD20 and IFN␥ mRNA expression in antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation, we isolated the arthritic and nonarthritic joints from each mouse and performed RT-PCR. There was a significant increase in both CD20 and IFN␥ mRNA expression in the inflamed joints, compared with the uninflamed ones, in each DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mouse examined (Figure 6). This observation is indicative of a possible role of B cells, as well as inflammatory cytokines, in the manifestation of persistent inflammation after antibiotic therapy. No difference in IL-17 mRNA expression was observed in joints from DR4⫹/ ⫺/⫺ ⫹CD28 MHCII⫺/⫺ mice with persistent arthritis and those from mice in which arthritis was resolved (data not shown). DISCUSSION Figure 6. Mean and SEM CD20 (top) and interferon-␥ (IFN␥) (bottom) mRNA expression in arthritic joints from 3 antibiotic-treated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation. CD20 and IFN␥ mRNA expression was assessed by real-time reverse transcriptase–polymerase chain reaction. Data were normalized to the 18S housekeeping gene. P values were determined by Student’s 2-tailed t-test. providing independent confirmation that arthritis persists in these mice in the absence of systemic B burgdorferi infection. The presence of IL-17–producing T cells has been associated with development of inflammation in a wide range of infectious, as well as autoimmune, diseases (32–34). In this model of arthritis, IL-17 mRNA expression was observed only in the lymph nodes of the B burgdorferi–infected CD28⫺/⫺ mice (Figure 5B), even though an IL-17 memory T cell response was detectable in the draining popliteal lymph nodes of antibiotictreated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation (Figure 5C). Increased CD20 and IFN␥ mRNA expression in the arthritic joints of DR4ⴙ/ⴙCD28ⴚ/ⴚMHCIIⴚ/ⴚ mice with persistent joint inflammation. The presence of B cells, as well as of the cytokine IFN␥, has been associated with persistent manifestations of arthritis in humans (10,11,35,36), providing a mechanism for the perpetuation of inflammation in the absence of B burgdorferi. However, it has not been possible thus far to confirm The immunologic mechanisms that trigger persistent joint inflammation after elimination of B burgdorferi infection have been elusive thus far, due to the absence of an animal model of this inflammatory disease (37). Here we describe the development of a murine system in which we observed self-perpetuating arthritis upon systemic eradication of B burgdorferi, as assessed by PCR analysis. Clinical data have shown that persistent arthritis is manifested in a small percentage of antibiotictreated patients with Lyme arthritis who had experienced intermittent inflammation in the joints before treatment. These patients had a preponderance of RAassociated HLA–DR alleles, such as HLA–DR4 (1,3,4). The intriguing phenomenon of persistent arthritis in patients with Lyme disease who were treated with antibiotics has been extensively studied. Two main hypotheses have been formulated to explain its pathogenesis, namely, persistent B burgdorferi infection after antibiotic treatment and infection-induced autoimmunity (1). The fact that spirochetal DNA is no longer detectable in the inflamed joints after antibiotic treatment seems to rule out the first hypothesis (5–7). However, the strong HLA–DR restriction observed in patients with persistent arthritis post–antibiotic treatment favors the autoimmunity-based hypothesis. The importance of understanding the mechanism(s) that trigger this chronic arthritic disease is 2-fold. It contributes to the general understanding of chronic inflammatory arthritis, and it may provide a model for other infection-induced autoimmune diseases. There seems to be a well-defined linkage between certain HLA–DR haplotypes and various autoimmune HLA–DR4 AND PERSISTENT ARTHRITIS AFTER B BURGDORFERI INFECTION diseases. HLA alleles are responsible for the presentation of autoantigens, as well as the positive and negative selection of autoreactive T cells in the thymus (38,39). Patients with Lyme arthritis that is refractory to treatment have a high frequency of HLA–DR4–related alleles compared with patients with treatment-responsive disease (1,3,4). Interestingly, the same HLA–DR alleles have also been associated with susceptibility to RA (40,41). Our murine studies indicate that the presence of HLA–DR4 or a related allele accentuates persistent Lyme arthritis post–antibiotic treatment, since a significant fraction of antibiotic-treated DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mice, but none of the CD28⫺/⫺ mice, showed persistent joint inflammation after antibiotic treatment. Interestingly, it has been shown that there is an increased frequency of CD4⫹CD28⫺/⫺ T cells in patients with RA, which correlates with the severity of the disease (42,43). This observation validates the use of DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice in developing a murine model of persistent Lyme arthritis post–antibiotic treatment. The humoral response against B burgdorferi antigens in antibiotic-treated patients with Lyme arthritis with persistent joint inflammation is of particular interest. Clinical studies have shown that antibodies against OspA mark the initiation of prolonged episodes of arthritis, as opposed to other B burgdorferi proteins that are associated with mild arthritis of short duration (3,9). While the majority of B burgdorferi–infected DR4⫹/⫹ CD28⫺/⫺MHCII⫺/⫺ mice with arthritic symptoms developed anti-OspA antibodies prior to antibiotic treatment, the anti-OspA humoral response was detectable in some, but not all, antibiotic-treated mice with persistent joint inflammation. These results provide evidence that in our murine model of persistent joint inflammation post–antibiotic treatment, anti-OspA antibodies play an important, but not necessarily exclusive, role in the pathogenesis of the disease. However, there was a significant correlation between anti-OspA antibody titers and development of arthritis in the mice that developed an anti-OspA humoral immune response after antibiotic treatment. This is consistent with clinical data that indicate that the levels of anti-OspA antibodies parallel the severity and duration of arthritis (3,9). B burgdorferi has been shown to induce IL-17 production in murine T cells in vitro, as well as in synovial fluid T cells isolated from patients with Lyme arthritis (44,45). In addition, B burgdorferi–induced arthritis was ameliorated by administration of anti–IL-17 (46). However, it has not yet been demonstrated whether Th17 cells participate in the pathogenesis of 3899 persistent joint inflammation post–antibiotic therapy. In the present study we did not detect any IL-17 mRNA expression in the inflamed mouse joints, although the B burgdorferi–infected mice showed increased IL-17 mRNA production in the inguinal and popliteal lymph nodes before antibiotic treatment, consistent with previous reports. It is possible that the number of infiltrated Th17 cells is small, rendering the IL-17 mRNA negligible in the context of an entire joint. Alternatively, IL-17 may not play an active role in the pathogenesis of persistent inflammation upon antibiotic treatment. A memory IL-17 response, however, was detectable in the lymph nodes of the DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice in which disease was refractory to treatment, indicating that Th17 cells were induced during the B burgdorferi infection phase in these mice, prior to antibiotic treatment. Increased CD20 mRNA expression was found in the joints of antibiotic-treated DR4⫹/⫹CD28⫺/⫺ MHCII⫺/⫺ mice with persistent joint inflammation, suggesting that inflammation is locally perpetuated in the microenvironment of the joint. Taken together, these data and the presence of anti-OspA antibodies in the peripheral immune system of these mice are indicative of a possible role of B cells in the murine model, consistent with findings in previous clinical studies (3,9). In addition to the presence of a humoral response in the inflammatory milieu of the joints, a critical role of T cells, in particular of Th1 cells, has been documented in patients with Lyme disease who have persistent arthritis after antibiotic therapy. IFN␥ production has been associated with increased severity of Lyme arthritis in mice (20), and IFN␥-producing cells were identified in the synovial tissue of patients with Lyme disease who had persistent arthritis post–antibiotic treatment (10,11,47,48). In this regard, it is of special interest that we observed increased IFN␥ mRNA expression in the joints of antibiotictreated DR4⫹/⫹CD28⫺/⫺MHCII⫺/⫺ mice with persistent joint inflammation, suggesting that IFN␥ plays a pathogenetic role in the murine model. Interestingly, the presence of IFN␥ in the inflamed joint may explain the absence of IL-17 transcription, since IL-17 production is negatively regulated by IFN␥ in vitro (49,50). In conclusion, we have described a mouse model with persistent joint inflammation post–antibiotic treatment. Although the possibility of residual infection due to the presence of a small number of spirochetes in the joints cannot be excluded, it is apparent from this model that the expression of the HLA–DR4 allele accentuates chronic inflammation in B burgdorferi–infected CD28⫺/⫺ mice after antibiotic therapy. The persistent joint inflam- 3900 ILIOPOULOU ET AL mation observed in these mice may be due to an autoimmune response that is linked to the presence of the HLA–DR4 allele. Therefore, the murine model developed in the present study will allow, for the first time, precise definition of the mechanism involved in the development of persistent Lyme arthritis post–antibiotic treatment. Unraveling its manifestation represents a unique opportunity, because it provides insights into other chronic inflammatory arthritides with a possible autoimmune etiology, such as RA, in which the infecting agent is unknown. ACKNOWLEDGMENTS We are grateful to Dr. Jenifer Coburn for the gift of B burgdorferi and to Dr. Jeffrey Bluestone for guiding us to the CD28⫺/⫺ system. We thank members of the Huber laboratory for critical reading of the manuscript and Lin Miao and Francesca Chang for technical help. AUTHOR CONTRIBUTIONS Dr. Huber 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 design. Iliopoulou, Huber. Acquisition of data. Iliopoulou. Analysis and interpretation of data. Iliopoulou, Huber. Manuscript preparation. Iliopoulou, Huber. Statistical analysis. Iliopoulou. Scoring of hematoxylin and eosin–stained sections. Alroy. REFERENCES 1. Steere AC, Glickstein L. Elucidation of Lyme arthritis. Nat Rev Immunol 2004;4:143–52. 2. Steere AC. Lyme disease. N Engl J Med 2001;345:115–25. 3. Kalish RA, Leong JM, Steere AC. Early and late antibody responses to full-length and truncated constructs of outer surface protein A of Borrelia burgdorferi in Lyme disease. Infect Immun 1995;63:2228–35. 4. Steere AC, Klitz W, Drouin EE, Falk BA, Kwok WW, Nepom GT, et al. Antibiotic-refractory Lyme arthritis is associated with HLADR molecules that bind a Borrelia burgdorferi peptide. J Exp Med 2006;203:961–71. 5. Nocton 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. 6. Carlson 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. 7. Steere AC, Angelis SM. Therapy for Lyme arthritis: strategies for the treatment of antibiotic-refractory arthritis [review]. Arthritis Rheum 2006;54:3079–86. 8. Kalish RA, Leong JM, Steere AC. Association of treatmentresistant chronic Lyme arthritis with HLA-DR4 and antibody reactivity to OspA and OspB of Borrelia burgdorferi. Infect Immun 1993;61:2774–9. 9. Akin E, McHugh GL, Flavell RA, Fikrig E, Steere AC. The 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. immunoglobulin (IgG) antibody response to OspA and OspB correlates with severe and prolonged Lyme arthritis and the IgG response to P35 correlates with mild and brief arthritis. Infect Immun 1999;67:173–81. Gross DM, Steere AC, Huber BT. T helper 1 response is dominant and localized to the synovial fluid in patients with Lyme arthritis. J Immunol 1998;160:1022–8. Shin 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. Yssel H, Shanafelt MC, Soderberg C, Schneider PV, Anzola J, Peltz G. Borrelia burgdorferi activates a T helper type 1-like T cell subset in Lyme arthritis. J Exp Med 1991;174:593–601. Barthold SW, Beck DS, Hansen GM, Terwilliger GA, Moody KD. Lyme borreliosis in selected strains and ages of laboratory mice. J Infect Dis 1990;162:133–8. Ma Y, Seiler KP, Eichwald EJ, Weis JH, Teuscher C, Weis JJ. Distinct characteristics of resistance to Borrelia burgdorferiinduced arthritis in C57BL/6N mice. Infect Immun 1998;66:161–8. Wooten RM, Weis JJ. Host-pathogen interactions promoting inflammatory Lyme arthritis: use of mouse models for dissection of disease processes. Curr Opin Microbiol 2001;4:274–9. McKisic MD, Redmond WL, Barthold SW. Cutting edge: T cellmediated pathology in murine Lyme borreliosis. J Immunol 2000; 164:6096–9. Keane-Myers A, Nickell SP. Role of IL-4 and IFN-␥ in modulation of immunity to Borrelia burgdorferi in mice. J Immunol 1995;155: 2020–8. Kang I, Barthold SW, Persing DH, Bockenstedt LK. T-helper-cell cytokines in the early evolution of murine Lyme arthritis. Infect Immun 1997;65:3107–11. Matyniak JE, Reiner SL. T helper phenotype and genetic susceptibility in experimental Lyme disease. J Exp Med 1995;181:1251–4. Crandall H, Dunn DM, Ma Y, Wooten RM, Zachary JF, Weis JH, et al. Gene expression profiling reveals unique pathways associated with differential severity of Lyme arthritis. J Immunol 2006;177:7930–42. Barthold SW, de Souza MS, Janotka JL, Smith AL, Persing DH. Chronic Lyme borreliosis in the laboratory mouse. Am J Pathol 1993;143:959–71. Barthold SW, Persing DH, Armstrong AL, Peeples RA. Kinetics of Borrelia burgdorferi dissemination and evolution of disease after intradermal inoculation of mice. Am J Pathol 1991;139: 263–73. Iliopoulou BP, Alroy J, Huber BT. CD28 deficiency exacerbates joint inflammation upon Borrelia burgdorferi infection, resulting in the development of chronic Lyme arthritis. J Immunol 2007; 179:8076–82. Ito K, Bian HJ, Molina M, Han J, Magram J, Saar E, et al. HLA-DR4-IE chimeric class II transgenic, murine class II-deficient mice are susceptible to experimental allergic encephalomyelitis. J Exp Med 1996;183:2635–44. Coburn J, Barthold SW, Leong JM. Diverse Lyme disease spirochetes bind integrin ␣IIb␤3 on human platelets. Infect Immun 1994;62:5559–67. Coburn J, Leong JM, Erban JK. Integrin ␣IIb␤3 mediates binding of the Lyme disease agent Borrelia burgdorferi to human platelets. Proc Natl Acad Sci U S A 1993;90:7059–63. Moody KD, Adams RL, Barthold SW. Effectiveness of antimicrobial treatment against Borrelia burgdorferi infection in mice. Antimicrob Agents Chemother 1994;38:1567–72. Pavia C, Inchiosa MA Jr, Wormser GP. Efficacy of short-course ceftriaxone therapy for Borrelia burgdorferi infection in C3H mice. Antimicrob Agents Chemother 2002;46:132–4. Morrison TB, Ma Y, Weis JH, Weis JJ. Rapid and sensitive quantification of Borrelia burgdorferi-infected mouse tissues by HLA–DR4 AND PERSISTENT ARTHRITIS AFTER B BURGDORFERI INFECTION 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. continuous fluorescent monitoring of PCR. J Clin Microbiol 1999;37:987–92. Ruderman EM, Kerr JS, Telford SR III, Spielman A, Glimcher LH, Gravallese EM. Early murine Lyme carditis has a macrophage predominance and is independent of major histocompatibility complex class II-CD4⫹ T cell interactions. J Infect Dis 1995;171: 362–70. Stavnezer J. Immunoglobulin class switching. Curr Opin Immunol 1996;8:199–205. Lubberts E, Koenders MI, Oppers-Walgreen B, van den Bersselaar L, Coenen-de Roo CJ, Joosten LA, et al. Treatment with a neutralizing anti-murine interleukin-17 antibody after the onset of collagen-induced arthritis reduces joint inflammation, cartilage destruction, and bone erosion. Arthritis Rheum 2004;50:650–9. Stockinger B, Veldhoen M. Differentiation and function of Th17 T cells. Curr Opin Immunol 2007;19:281–6. Veldhoen M, Hocking RJ, Flavell RA, Stockinger B. Signals mediated by transforming growth factor-␤ initiate autoimmune encephalomyelitis, but chronic inflammation is needed to sustain disease. Nat Immunol 2006;7:1151–6. Ghosh S, Steere AC, Stollar BD, Huber BT. In situ diversification of the antibody repertoire in chronic Lyme arthritis synovium. J Immunol 2005;174:2860–9. Ghosh S, Seward R, Costello CE, Stollar BD, Huber BT. Autoantibodies from synovial lesions in chronic, antibiotic treatmentresistant Lyme arthritis bind cytokeratin-10. J Immunol 2006;177: 2486–94. Benoist C, Mathis D. Autoimmunity provoked by infection: how good is the case for T cell epitope mimicry? Nat Immunol 2001;2: 797–801. Marrack P, Kappler J, Kotzin BL. Autoimmune disease: why and where it occurs. Nat Med 2001;7:899–905. Ridgway WM, Fathman CG. MHC structure and autoimmune T cell repertoire development. Curr Opin Immunol 1999;11: 638–42. Buckner JH, Nepom GT. Genetics of rheumatoid arthritis: is there a scientific explanation for the human leukocyte antigen association? Curr Opin Rheumatol 2002;14:254–9. 3901 41. Pascual M, Mataran L, Jones G, Shing D, van der Slik AR, Giphart MJ, et al. HLA haplotypes and susceptibility to rheumatoid arthritis: more than class II genes. Scand J Rheumatol 2002;31:275–8. 42. Pawlik A, Ostanek L, Brzosko I, Brzosko M, Masiuk M, Machalinski B, et al. The expansion of CD4⫹CD28⫺ T cells in patients with rheumatoid arthritis. Arthritis Res Ther 2003;5:R210–3. 43. Schmidt D, Martens PB, Weyand CM, Goronzy JJ. The repertoire of CD4⫹ CD28⫺ T cells in rheumatoid arthritis. Mol Med 1996;2:608–18. 44. Knauer J, Siegemund S, Muller U, Al-Robaiy S, Kastelein RA, Alber G, et al. Borrelia burgdorferi potently activates bone marrow-derived conventional dendritic cells for production of IL-23 required for IL-17 release by T cells. FEMS Immunol Med Microbiol 2007;49:353–63. 45. Infante-Duarte C, Horton HF, Byrne MC, Kamradt T. Microbial lipopeptides induce the production of IL-17 in Th cells. J Immunol 2000;165:6107–15. 46. Burchill MA, Nardelli DT, England DM, DeCoster DJ, Christopherson JA, Callister SM, et al. Inhibition of interleukin-17 prevents the development of arthritis in vaccinated mice challenged with Borrelia burgdorferi. Infect Immun 2003;71:3437–42. 47. Gross 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. 48. Yin Z, Braun J, Neure L, Wu P, Eggens U, Krause A, et al. T cell cytokine pattern in the joints of patients with Lyme arthritis and its regulation by cytokines and anticytokines. Arthritis Rheum 1997; 40:69–79. 49. Harrington LE, Hatton RD, Mangan PR, Turner H, Murphy TL, Murphy KM, et al. Interleukin 17-producing CD4⫹ effector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. Nat Immunol 2005;6:1123–32. 50. Park H, Li Z, Yang XO, Chang SH, Nurieva R, Wang YH, et al. A distinct lineage of CD4 T cells regulates tissue inflammation by producing interleukin 17. Nat Immunol 2005;6:1133–41.