SYSTEMATIC DETECTION OF MYCOPLASMAS BY CULTURE AND POLYMERASE CHAIN REACTION (PCR) PROCEDURES IN 209 SYNOVIAL FLUID SAMPLES

Transcription

1 British Journal of Rheumatology 1997;36: SYSTEMATIC DETECTION OF MYCOPLASMAS BY CULTURE AND POLYMERASE CHAIN REACTION (PCR) PROCEDURES IN 209 SYNOVIAL FLUID SAMPLES T. SCHAEVERBEKE, H. RENAUDIN,* M. CLERC,* L. LEQUEN, J. P. VERNHES, B. DE BARBEYRAC,* B. BANNWARTH, Ch. BE BE AR* and J. DEHAIS Service de Rhumatologie and *Service de Bacte riologie, Centre Hospitalier Pellegrin, Universite de Bordeaux II, 6 place Amélie Raba-Le on, Bordeaux cedex, France SUMMARY The objective was to investigate the presence of mycoplasmas in rheumatoid arthritis (RA) and other chronic arthritides. Samples of synovial fluid (SF) were systematically collected from all patients presenting with an articular effusion. Each sample was divided into three parts. The first was kept for cytological count and culture on standard media for pyogens and mycobacteria, the second was cultivated on specific media for mycoplasmas and the third frozen for subsequent study by polymerase chain reaction (PCR). A total of 209 samples were studied. Half of the patients had inflammatory rheumatic diseases: RA (27), spondyloarthropathy (28), connective tissue disease (5), unclassified arthritis (45). The remaining suffered from other conditions, including osteoarthritis (60), gouty arthritis (19), haemarthrosis (5), post-traumatic effusion (2). Eight samples were positive by culture, two for Mycoplasma hominis, three for M. fermentans, one for M. salivarium, one for M. orale and one for Ureaplasma urealyticum. All the patients concerned had an inflammatory rheumatic disease: five had RA, one had psoriatic arthritis and two had unclassified arthritis. These results were confirmed by PCR in two cases (one M. fermentans, one U. urealyticum). The lack of sensitivity of the conventional PCR assay on SF is discussed. Mycoplasmas were mainly detected in SF of RA patients. These results raise the question of the possible role of mycoplasmas in the triggering and maintenance of inflammatory rheumatic diseases, especially RA. KEY WORDS: Rheumatoid arthritis, Mycoplasma, Synovial fluid. THE hypothesis that rheumatoid arthritis (RA), like other autoimmune diseases, could, in fact, be caused by an infectious agent, virus [1] or slow-growing bacteria [2, 3], has been widely discussed. Among the infectious agents suggested, mycoplasmas seem to be especially good candidates. They are known to be a major cause of acute and chronic arthritis in animals [4]. They frequently induce an immune disturbance in the host organism [5, 6]. Furthermore, 20 yr ago, a few cases were reported of mycoplasmas, especially M. fermentans, isolated from joint fluid in RA and other inflammatory arthritides [7 9]. Recently, M. genitalium has been detected by polymerase chain reaction (PCR) assay in synovial specimens from a patient with Reiter s disease and from another with chronic seronegative polyarthritis following a juvenile arthritis [10]. However, as many other attempts to detect the presence of these fastidious organisms in the joints of patients with such arthritides have failed, the question of their possible role in the pathogenesis of human rheumatic diseases remains controversial and needs to be re-examined. The aim of this study was to evaluate systematically the presence of mycoplasmas in synovial fluid (SF), using culture and PCR assay in patients with RA and other chronic arthritides. Submitted 31 May 1996; revised version accepted 3 September Correspondence to: J. Dehais, Service de Rhumatologie, Centre Hospitalier Pellegrin, 6, place Ame lie Raba-Le on, Bordeaux cedex, France. PATIENTS AND METHODS Patients Over a 2.5 yr period (30 months), 209 samples of SF were collected from 201 patients: 32 SF from 27 patients with RA, 29 SF from 28 patients with spondyloarthropathy (six of whom had reactive arthritis and five psoriatic spondyloarthropathy), five SF from five patients with connective tissue disease, 46 SF from 45 patients with unclassified arthritis, 60 SF from 60 patients with osteoarthritis, 20 SF from 19 patients with gouty arthritis or chondrocalcinosis, one SF from a patient with villonodular synovitis, two SF from two patients with amyloidosis arthropathy, five SF from five patients with haemarthrosis, two SF from two patients with post-traumatic joint effusion, two SF from two patients with effusion due to a neighbouring metastasis, two SF from two patients with neuroarthropathy, two SF from two patients with arthropathy associated with dialysis and one SF from a patient with hypertrophic osteoarthropathy. SF specimens Each sample of SF was divided into three parts. The first part was used for cytological count and culture on standard media for pyogens and mycobacteria. The second, 1 ml SF, was added to 1 ml 2SP (sucrose phosphate) transport medium [11], and immediately brought to the laboratory to be cultured on specific media for mycoplasmas. The third was frozen ( 80 C) for subsequent study by PCR British Society for Rheumatology

2 SCHAEVERBEKE ET AL.: MYCOPLASMAS AND ARTHRITIS 311 TABLE I Results of the cultures and data of the corresponding patients Disease Age duration Cell count Neutrophils No. Organism Culture PCR Sex (yr) Diagnosis Features (months) (/mm 3 ) (%) 1 M. hominis + F 59 RA Symmetrical polyarthritis RF M. hominis + M 21 PsA Monoarthritis HLA B27 + psoriasis M. fermentans + F 69 Unclas A Monoarthritis juv.a in past history M. fermentans + + F 55 RA Destructive polyarthritis RF M. fermentans + inh F 48 Unclas A Non-erosive oligoarthritis RF-ANA 1/ M. salivarium + NA M 69 RA Chronic erosive polyarthritis RF M. orale + NA F 60 RA Symmetrical polyarthritis RF U. urealyticum + + M 72 RA Symmetrical polyarthritis RF RA = rheumatoid arthritis. PsA = psoriatic arthritis. Unclas A = unclassified arthritis. Juv.A = juvenile arthritis. RF = rheumatoid factor. ANA = antinuclear antibodies. inh = inhibition. NA = not available. Mycoplasma culture conditions The mixture of SF and 2SP was incubated on different specific media: 200 l in 2 ml of liquid Edward modified medium supplemented with arginine, 100 l in 1 ml of liquid Edward modified medium supplemented with glucose and 100 l on agar Edward modified medium for mycoplasmas [12], 100 l in1ml of liquid Shepard medium and 100 l on agar Shepard medium for ureaplasmas [13]. These media were incubated for 1 month at 37 C under a carbon dioxide atmosphere. Species were identified according to the appearance of colonies, their biochemical properties, growth-inhibition tests [14] and, when possible, by PCR. Preparation of sample for PCR After thawing, 500 l of SF samples were centrifuged for 60 min at g. The pellet was resuspended in TE buffer [10 mm Tris HCl (ph 8.0), 1 mm EDTA], incubated for 1 h at 56 C with 1% sodium dodecyl sulphate (SDS) and 100 g/ml proteinase K (Sigma), and then heated at 95 C for 15 min to inactivate proteinase K. Nucleic acids were extracted once with phenol chloroform isoamyl alcohol (25:24:1), and once with chloroform isoamyl alcohol (24:1). The aqueous phase was precipitated at 20 C for 18 h by the addition of 2 vols of 95% ethanol and 0.1 M sodium acetate. DNA was recovered by centrifugation at g for 30 min and the pellet resuspended in 100 l distilled water. A volume of 10 l of the DNA preparation was used for the PCR assay. PCR assay PCR reactions were performed with an automated thermocycler (Perkin-Elmer Cetus 480, Norwalk, CT, USA). Specific oligonucleotide primers were used for the different species of mycoplasma affecting humans: M. hominis, M. fermentans, M. pneumoniae, M. genitalium and Ureaplasma urealyticum. These were chosen in published nucleotide sequences: in the 16S rrna gene for M. hominis [15], in an insertion sequence for M. fermentans [16], in the adhesin gene for M. pneumoniae and M. genitalium [17], and in the urease gene for U. urealyticum [18]. To evaluate the presence of PCR inhibitors, each sample was also tested with primers settling to a fragment of the DQ human leucocyte antigen gene [19]. All these oligonucleotides were synthesized by the methoxyphosphoramidite method on an Applied Biosystems Model 38SA DNA synthesizer (Applied Biosystems, San Jose, CA, USA) at the Institut de Biochimie Cellulaire et de Neurochimie (Bordeaux, France). Amplification was performed in a final volume of 50 l containing 5 l of 10 assay buffer, 1.5 mm MgCl 2, 200 M each of datp, dctp, dttp and dgtp, 1 M of each primer and 1.5 U of Taq DNA polymerase (Promega Corporation, Madison, WI, USA). Amplification consisted of a 4 min thermal delay step at 94 C followed by 35 cycles comprising a 1 min denaturation step at 94 C, a 1 min annealing step at 55 C and a 1 min elongation step at 72 C. The products of PCR were analysed by 2% agarose gel electrophoresis and ethidium bromide staining. Samples containing a band of the expected size were controlled by hybridization. After Southern blotting, filters were hybridized with a 32 P-radiolabelled internal specific probe at 55 C overnight, then washed under highly stringent conditions before autoradiography. Statistical analysis Fisher s exact test was used to compare proportions and Student s t-test for the comparison of means between the two groups. RESULTS Results are summarized in Tables I and II. TABLE II Comparisons between the mycoplasma-positive and -negative groups of patients with inflammatory rheumatic diseases Disease Cell Age duration ESR count Neutrophils (yr) (months) (mm/h) (/mm 3 ) (%) Positive group Mean S.D Negative group Mean S.D P

3 312 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 36 NO. 3 Standard cultures All standard cultures remained negative. Culture on specific media for mycoplasmas Eight cultures were positive: two for M. hominis, three for M. fermentans, one for M. salivarium, one for U. urealyticum and one for M. orale. In each case, the culture was confirmed on both solid and liquid media. PCR study The presence of inhibitors of DNA polymerase was demonstrated in 30% of the samples studied by a negative reaction of the PCR assay with human HLA primers. PCR with HLA primers were positive for all but one (no. 5) of the eight SF from which mycoplasmas had been cultured. Only two fluids were positive, thus confirming two of the culture results: one sample positive for U. urealyticum, one for M. fermentans. Regarding the latter case, three SF samples had been collected at 6 monthly intervals from the same patient. Interestingly, the first SF was culture positive and PCR negative, the second culture negative and PCR positive, and the third negative for both culture and PCR. The patient in question, who had classic RA, was given no antibiotics during this period but was treated with methotrexate and low-dose prednisone. Relationship between the presence of mycoplasmas and the different diseases All the mycoplasmas were detected in the SF of patients (five women and three men) who had an inflammatory rheumatic disease: RA (n = 5), psoriatic spondyloarthropathy (n = 1) and unclassified arthritis (n = 2). Comparison between mycoplasma-positive and -negative patients No significant difference was found between the age of the eight patients whose SF was culture positive (positive group) and that of the other patients with inflammatory rheumatic conditions (negative group). There was also no difference regarding the duration of illness between the positive and the negative groups. In four of the eight positive cases, however, onset of the illness was recent, with the first articular manifestations dating from less than 6 months before SF was collected. The erythrocyte sedimentation rate (ESR), as well as cell and neutrophil counts in SF, did not differ significantly between the two groups (see Table II). Antibiotic treatment Amongst the eight patients from whose SF a mycoplasma was isolated, six were treated with cycline once the results of the culture study were known (doxycycline 200 mg/day for 2 months), in addition to their previous treatment: NSAID or prednisone and a slow-acting anti-rheumatic drug (SAARD). Only one patient, with unclassified arthritis, improved significantly. Twenty months after the cycline treatment was stopped, she was still in subcomplete remission. She is still receiving hydroxychloroquine. DISCUSSION The possibility of mycoplasmas being isolated from synovial specimens of patients with various rheumatic disorders, including RA, was largely debated in the 1960s and the early 1970s, but, as a matter of fact, very few teams reported such isolation, and these results were sometimes contested. As mycoplasmas are common cell culture contaminants, the studies in which cell cultures were used to isolate mycoplasmas from synovial specimens were not reliable [20 23]. Like Williams et al. [7, 8], Ma rdh et al. [9] and Jansson et al. [24], we used cell-free media, thus excluding theoretically the principal risks of contamination. Jansson et al. could isolate mycoplasmas from synovial samples of patients with rheumatic disorders, but could not clearly identify these isolates, concluding that they were serologically related to M. arthritidis, which is known to be a pathogen in rodents [24, 25]. We were able to identify all strains isolated from our clinical samples: according to the appearance of colonies, biochemical properties and growth-inhibition test for all species, and with confirmation by PCR assays when available (for M. fermentans, M. hominis and U. urealyticum). Williams et al. claimed to have isolated M. fermentans in 40%, using a sucrose density gradient to separate the organisms from the cells contained in the synovial effusion before cultivating them in conventional liquid medium. However, this technique has been criticized because the sucrose density gradient was charged with inducing colour changes of the culture medium that could have been considered as a positivity of the culture. Williams admitted that the organisms could not be grown in all cases on solid medium after the change in colour was observed in liquid medium [26]. Moreover, such isolations were not confirmed by numerous later studies [27 29], despite using 12 different culture media in one of them [29]. So, our culture study confirmed that different human mycoplasmas can occasionally be cultured from the SF of patients with RA and other inflammatory rheumatic arthritides. Our results cannot be explained by contamination. First, as mentioned above, we used cell-free media. Second, mycoplasmas were only recovered from SF of patients with inflammatory rheumatic diseases. No mycoplasma was isolated from the SF of patients suffering from other conditions, including osteoarthritis, gouty arthritis, chondrocalcinosis and post-traumatic joint effusion. A contamination should have led to a random distribution of the results. Moreover, two of the positive cultures were confirmed by PCR. In one of these patients, the sample positive by PCR was different from the one which was culture positive. Furthermore, the recovered mycoplasmas have been undoubtedly identified as M. fermentans (n = 3), M. hominis (n = 2), M. salivarium (n = 1), M. orale (n = 1) and U. urealyticum (n = 1), which are all human mycoplasmas. Our PCR results are perhaps more surprising. Since PCR is considered to be a very sensitive method, we should have obtained better results with PCR than with

4 SCHAEVERBEKE ET AL.: MYCOPLASMAS AND ARTHRITIS 313 cultures. In fact, we simply obtained confirmation of two positive culture results. We did not, however, use specific primers for M. salivarium and M. orale which were cultured from two patients, and PCR was inhibited (PCR with HLA primers negative) in one of the SF from which M. fermentans had been isolated (no. 5). Nevertheless, the sets of primers we used were considered to be specific and sensitive, as they were validated on other kinds of clinical samples (urine, genital swabs, endotracheal aspirates, etc.). Moreover, the DNA was extracted according to standard protocols and the conditions of amplification were those commonly used in the laboratory for the identification of culture products as well as detection in other clinical samples. However, PCR sensitivity would appear to be relatively poor in SF, due to the presence of many inhibitors of the reaction (DNases, proteases, hybridization with partially complementary DNA facilitated by the disproportion between human and bacterial DNA, etc.). By voluntary contamination of SF samples with different concentrations of bacterial cultures, we have been able to evaluate the sensitivity of our PCR protocol at colour change units per millilitre (personal unpublished data). This is less than that attributed to the culture assay. Thus, we may have failed to detect low amounts of mycoplasma DNA. The limits of the PCR technique in some clinical samples have already been reported [30, 31]. Moreover, after we did this study, we tried a semi-nested PCR assay developed for M. fermentans [32] in another series of synovial specimens. With this assay we could detect the presence of M. fermentans in synovial specimens of eight (21%) patients with RA, two (20%) patients with spondyloarthropathy, one (20%) patient with psoriatic arthritis and four (13%) patients with undifferentiated arthritis [33]. Thus, conventional PCR is probably not sensitive enough in synovial specimens. Other PCR assays, such as nested PCR or reverse transcriptase PCR (RT-PCR), would probably improve the sensitivity of detection studies in synovial specimens, but still have not been developed for most human mycoplasmas. So, we have shown that mycoplasmas were present in the SF of eight patients presenting with several inflammatory rheumatic diseases of unknown cause, including RA. Such results raise the question of the possible role of mycoplasmas in the triggering and maintenance of these diseases. Further studies are required to answer this question. Improvement of the PCR procedures and adaptation to synovial specimens should facilitate the task. ACKNOWLEDGEMENT We are grateful to the Centre Hospitalier Universitaire of Bordeaux for its financial support. REFERENCES 1. Silman AJ. Is rheumatoid arthritis an infectious disease? Br Med J 1991;303: Rook GAW, Lydyard PM, Stanford JL. A reappraisal of the evidence that rheumatoid arthritis and several other idiopathic diseases are slow bacterial infections. Ann Rheum Dis 1993;52:S Saag MS, Bennet JC. The infectious etiology of chronic rheumatic diseases. Semin Arthritis Rheum 1987;17: Cole BC, Washburn LR, Taylor-Robinson D. Mycoplasma-induced arthritis. In: Razin S, Barile MF, eds. The mycoplasmas, Vol. 4. Mycoplasma pathogenicity. New York: Academic Press, 1985: Cole BC, Ward JR. Mycoplasmas as arthritogenic agents. In: Tully JG, Whitcomb RF, eds.the mycoplasmas, Vol. 2. Human and animal mycoplasmas. New York: Academic Press, 1979: Biberfeld G. Infection sequelae and autoimmune reactions in M pneumoniae infections. In: Razin S, Barile MF, eds. The mycoplasmas, Vol. 4. Mycoplasma pathogenicity. New York: Academic Press, 1985: Williams MH. Pathogenic mycoplasma in rheumatoid arthritis? In: Pathogenic mycoplasma a Ciba Foundation symposium. Amsterdam: Associated Scientific Publishers, 1972: Williams MH, Brostoff J, Roitt IM. Possible role of Mycoplasma fermentans in pathogenesis of rheumatoid arthritis. Lancet 1970;ii: Ma rdh PA, Nilsson FJ, Bjell A. Mycoplasmas and bacteria in synovial fluid from patients with rheumatoid arthritis. Ann Rheum Dis 1973;32: Taylor-Robinson D, Gilroy CB, Horowitz S, Horowitz J. Mycoplasma genitalium in the joint of two patients with arthritis. Eur J Clin Microbiol Infect Dis 1994;13: Smith TF, Weed LA, Petterson GR, Segura JW. Recovery of chlamydia and genital mycoplasma transported in sucrose phosphate buffer and urease colour test medium. Health Lab Sci 1977;14: Edward DG. Selective medium for pleuro-pneumoniaelike organism. J Gen Microbiol 1947;1: Shepard MC, Lunceford CD. Differential agar medium (A7) for identification of Ureaplasma urealyticum (human T-mycoplasmas) in primary culture of clinical material. J Clin Microbiol 1976;3: Clyde WA. Mycoplasma species identification based upon growth antisera. J Immunol 1964;92: Blanchard A, Dybvig YK, Watson HL, Griffiths G, Cassell GH. Evaluation of intraspecies genetic variation within the 16S rrna gene of Mycoplasma hominis and detection by polymerase chain reaction. J Clin Microbiol 1993;31: Wang RYH, Hu WS, Dawson MS, Shih JWK, Lo SC. Selective detection of Mycoplasma fermentans by polymerase chain reaction and by using a nucleotide sequence within the insertion sequence-like element. J Clin Microbiol 1992;30: de Barbeyrac B, Bernet-Poggi C, Febrer F, Renaudin H, Dupon M, Be be ar C. Detection of Mycoplasma pneumoniae and Mycoplasma genitalium in clinical samples by polymerase chain reaction. Clin Infect Dis 1993;17(suppl. 1):S Blanchard A, Hentschel J, Duffy L, Baldus K, Cassell GH. Detection of Ureaplasma urealyticum by polymerase chain reaction in the urogenital tract of adults, in amniotic fluid, and in the respiratory tract of newborns. Clin Infect Dis 1993;17(suppl. 1):S Scharf SJ, Horn GT, Erlich HA. Direct cloning and sequence analysis of enzymatically amplified genomic sequences. Science 1986;233:

5 314 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 36 NO Bartholomew LE. Isolation of mycoplasmas (PPLO) from patients with rheumatoid arthritis, systemic lupus and Reiter s syndrome. Arthritis Rheum 1965;8: Bartholomew LE. Characterization of mycoplasma strains and antibody studies from patients with rheumatoid arthritis. Ann NY Acad Sci 1967;143: Barnett EV, Balduzzi P, Vaughan JH, Morgan HE. Search for infectious agents in rheumatoid arthritis. Arthritis Rheum 1966;9: Fraser KB, Shirodaria PV, Haire M, Middelton D. Mycoplasmas in cell cultures from rheumatoid synovial membrane. J Hyg 1971;69: Jansson E, Vainio U, Snellman O, Tuuri S. Search for mycoplasma in rheumatoid arthritis. Ann Rheum Dis 1971;30: Jansson E, Miettinen A, Hakkarainen K, Backman A, Seniusova B, Makisara P et al. Cultivation of fastidious mycoplasmas from human arthritis. Z Rheumatol 1983;42: Williams MH. Pathogenic mycoplasma in rheumatoid arthritis? Discussion. In: Pathogenic mycoplasma a Ciba Foundation symposium. Amsterdam: Associated Scientific Publishers, 1972: Cole BC, Ward JR, Smith CB. Studies on the infectious etiology of rheumatoid arthritis. Arthritis Rheum 1973;16: Stewart SM, Duthie JJR, Mackay JMK, Marmion BP, Alexander WRM. Mycoplasmas and rheumatoid arthritis. Ann Rheum Dis 1974;33: Barile MF, Yoshida H, Roth H. Rheumatoid arthritis: new findings on the failure to isolate or detect mycoplasmas by multiple cultivation or serologic procedures and review of the literature. Rev Infect Dis 1991;13: Leng Z, Kenny GE, Roberts M. Evaluation of the detection limits of PCR identification of Mycoplasma pneumoniae in clinical samples. Mol Cell Probes 1994;8: He Q, Marjama ki M, Soini H, Mertsola J, Viljanen MK. Primers are decisive for sensitivity of PCR. Biotech 1994;17: Katseni VL, Gilroy CB, Ryait BK, Ariyoshi K, Bieniasz PD, Weber JN et al. Mycoplasma fermentans in individuals seropositive and seronegative for HIV-1. Lancet 1993;341: Schaeverbeke T, Gilroy CB, Be be ar C, Dehais J, Taylor-Robinson D. Mycoplasma fermentans in joints of patients with rheumatoid arthritis and other joint disorders. Lancet 1996;347:1418.