TEXT

Top Abstract Text References

Several hypotheses, reviewed in reference 13, have been proposed to explain the chronic course of Lyme arthritis, a feature of Borrelia burgdorferi infection. However, if a pathogenic role of viable spirochetes has been proven in acute Lyme arthritis (12), no such data are available from chronic stages of the human disease. Although the presence of borrelial DNA has been reported in a mouse model to correlate with that of viable B. burgdorferi (8), DNA detection by PCR does not currently provide formal evidence of the bacterial viability in human disease. This property could be assessed with more accuracy by detection of specific mRNA because of the very short half-life of these molecules. Due to the small number of spirochetes present in infected tissues, reverse transcription of bacterial mRNA followed by DNA amplification (reverse transcriptase PCR [RT-PCR]) could be of major interest to assess the presence or absence of viable bacteria in chronic Lyme arthritis.

The aim of this work was to evaluate the feasibility of applying RT-PCR techniques to synovial tissue samples from the C3H mouse model of Lyme borreliosis. Since the standard two-step RT-PCR method is labor intensive and carries a high risk of contamination, we developed a one-step RT-PCR method and compared the two techniques.

Three-week-old C3H/HeNCrlBR (C3H) mice were purchased from Charles River Laboratories (Saint-Aubin-Les Elbeuf, France). These inbred mice develop polyarthritis and carditis within 3 weeks of B. burgdorferi inoculation (1). Mice were initially infected by intradermal inoculation of 10⁴ B. burgdorferi sensu stricto PBre organisms at passages 7 to 9 in 0.1 ml of BSK-H medium (Sigma, Saint Quentin Fallavier, France). Since transplant of ear punch biopsies from infected mice has been found to be a reliable way to maintain virulent spirochetes in vivo (2), this method was employed for further infections. Briefly, 15 days after inoculation, infected mice were anesthetized with isoflurane (Abbott, Rungis, France), and after local disinfection, ear biopsies were performed by using a 3-mm-diameter sterile biopsy punch (Stiefel, Nanterre, France). Half of each biopsy was placed into a BSK-H medium tube to monitor skin infection, and 0.25-mm² parts of the second half were used for xenografts. Tissues were transplanted into anesthetized recipient mice in a shaved part of the dorsal lumbar region. After disinfection of the skin, a 1-mm-long incision of the epidermis was made with a sterile blade, and then an infected ear piece was sterilely inserted into the subcutis of each recipient mouse. This procedure did not require any suture, and local infection was never noticed.

Ten mice were tested: 4 infected for 1 month, 2 infected for 2 months, 1 infected for 3 months, and 3 uninfected as controls. Infection was confirmed by macroscopic examination which showed arthritis of the rear ankles and by culture of ear biopsies performed 15 days after transplant challenge, which were positive for the seven infected mice and none of the three control mice. Synovial tissues were collected under strictly aseptic conditions, as described by Malawista et al. (8). Control and infected mice were killed with isoflurane and rinsed externally with 60% ethanol, and after sampling, the two tibio-tarsal joints from each mouse were aseptically washed in BSK-H medium to eliminate contaminating blood. One joint was put directly into a sterile Eppendorf microtube, frozen on dry ice, and stored at 80°C, while the other one was placed in a culture tube containing 0.3 ml of rabbit serum (Sigma) in 6 ml of BSK-H medium at 34°C. Heart samples were also collected and processed in the same manner. Likewise, positive B. burgdorferi cultures were obtained from synovial and heart samples from seven of seven (100%) infected mice but never from the three control mice.

DNA extraction and PCR were performed with one aliquot of synovial or heart tissue from each mouse as previously described (7). The B. burgdorferi target chosen for DNA amplification was a specific segment from the flagellin gene (fla) of B. burgdorferi (5), previously reported for the detection of B. burgdorferi in synovial tissues (7) and cerebrospinal fluids (6) by using the Bb1 and Bb2 primers. After PCR or RT-PCR, amplified products were electrophoretically separated, and Southern blot analyses were then performed with the Bb3 oligonucleotide as an internal probe with a sensitivity previously evaluated as 10 bacteria (7). The fla gene was detected by PCR in synovial and heart samples from seven of seven infected mice up to 3 months after inoculation, but not in samples from the three control mice.

Total RNA was isolated from a second aliquot of synovial or heart tissue from each mouse by using Tri-Reagent (Molecular Research Center, Cincinnati, Ohio). In each RNA sample, the RNA/DNA ratio was controlled by spectrophotometric absorption (GeneQuant; Pharmacia, Saclay, France). Each sample used for further analysis had an A₂₆₀/A₂₈₀ ratio higher than 1.8, so treatment with DNase was not necessary.

Two RT-PCR methods were compared: a one-step procedure and a two-step procedure. (i) In the two-step RT-PCR assay, the murine leukemia virus RT kit (Promega) was used according to the manufacturer's instructions. Reverse transcription was carried out for 15 min at 42°C. The resultant cDNA was amplified by PCR. (ii) In the one-step RT-PCR assay, reverse transcription and PCR were performed in a single reaction tube by using a thermostable rTth DNA polymerase for both reactions (Perkin-Elmer, Courtaboeuf, France) according to the manufacturer's instructions. Reverse transcription was carried out at 60°C for 30 min and directly followed by 40 cycles of cDNA amplification under the same conditions as described above.

In each PCR or RT-PCR run, two negative controls were included: sterile water to control the purity of the reagents and DNA or RNA from synovial tissues of age-matched uninfected control mice to confirm the reaction specificity and the absence of cross-contamination during sample processing. Amplification efficiency was monitored by using DNA and RNA extracted from B. burgdorferi cultures as positive controls. In RNA samples, the absence of DNA contamination was checked by submitting each sample to a PCR run using Taq polymerase. No amplified products were detected in this procedure, thus confirming the purity of the RNA preparations.

The sensitivities of the two RT-PCR methods were compared by analysis of 10-fold serial dilutions of RNA from B. burgdorferi cultures. After gel electrophoresis and Southern blotting, the one-step RT-PCR assay gave stronger signals and a 10-fold lower detection limit than the two-step assay (Fig. 1). RNA of the fla gene was detected by gel electrophoresis (Fig. 2A) and Southern blotting (Fig. 2B) in all synovial and heart samples from the seven infected mice, whereas none of the samples from the three control mice were positive for fla RNA.

View larger version (57K): [in this window] [in a new window]   FIG. 1.   Comparison of the sensitivities of the two RT-PCR methods by gel electrophoresis (A) and Southern blotting (B) with serial dilutions of RNA from culture of B. burgdorferi sensu stricto PBre. Lanes: 1 to 6, one-step RT-PCR protocol; 8 to 13, two-step RT-PCR protocol; 1 and 8, sterile water (negative controls); 2 and 9, 104-fold dilution; 3 and 10, 103-fold dilution; 4 and 11, 102-fold dilution; 5 and 12, 10-fold dilution; 6 and 13, no dilution; 7, 100-bp DNA ladder (Gibco-BRL, Cergy-Pontoise, France).

View larger version (69K): [in this window] [in a new window]   FIG. 2.   Gel electrophoresis (A) and Southern blot (B) analysis of one-step RT-PCR products from synovial and heart samples from four mice infected with B. burgdorferi sensu stricto PBre. Lanes: 1, sterile water (negative control); 2, 4, 6, and 8, synovial samples from infected mice; 3, 5, 7, and 9, heart samples from the corresponding infected mice. 10, synovial sample from a control mouse (uninfected); 11, 30 fg of RNA of B. burgdorferi PBre; 12, 100-bp DNA ladder (Gibco-BRL, Cergy-Pontoise, France).

In order to assess the absolute sensitivity of the one-step RT-PCR procedure in mouse samples, DNA from synovial tissue was amplified in the same run as DNA extracted from 10-fold dilutions of cultured spirochetes. Comparison of the signals obtained after Southern blot analysis gave an estimated number of spirochetes of 400 per synovial sample. With one-step RT-PCR, a signal was still detected in 10-fold dilutions of RNA from the same synovial extracts, so the sensitivity of this RT-PCR method may be considered to be about 40 spirochetes.

Other groups (4, 9, 14) have previously reported the use of RT-PCR to detect B. burgdorferi RNA in infected C3H mice. However, these studies were carried out with heart and spleen samples, which are known to contain larger numbers of bacteria (15), whereas in human (11) or murine (3) chronic Lyme arthritis, only a few bacteria are present in synovial tissue. This emphasizes the requirement for a highly sensitive method of detecting B. burgdorferi RNA. Thus, the greater sensitivity of the one-step RT-PCR assay compared to that of the two-step assay is of considerable interest. Our one-step RT-PCR procedure could derive its higher sensitivity from the ability of rTth DNA polymerase to efficiently reverse transcribe RNA templates at elevated temperatures (10), at which RNA secondary structures are unstable and do not affect the enzyme activity.