Eight-week old male Lewis rats were purchased from Harlan Laboratories (Indianapolis, IN, USA). They were maintained in microisolators under specific pathogen-free conditions in the animal care facility of the Toronto Western Hospital, University Health Network. All studies were conducted with the approval of the Animal Care Committee of the University Health Network.

Arthritis was induced in the rats by the intra-articular injection of live Chlamydia packaged in Lewis rat synovial fibroblasts as previously described 23. Briefly, C. trachomatis serotype L2 was inoculated into monolayers of the fibroblast lines in culture. After overnight incubation, the cells containing chlamydia were harvested and injected into the knee joint of each rat at 5 × 10⁵ colony forming units (CFU)/joint. Rats were assessed on a daily basis and then sacrificed four days after injection. At necropsy, their knee joints were either processed for histopathology or for quantitation of intra-articular Chlamydia.

For histology evaluations, joints were fixed in formalin. They were measured with a caliper and then decalcified as described previously 23. They were sectioned and stained with H & E. Sections were evaluated and scored according to the system mentioned by us before.

We have adapted a clinical use ELISA kit, the IDEIA PCE Chlamydia from OXOID-Dako (Basingstoke,, UK) for the quantitation of Chlamydia in tissues 6. Synovial tissues from the joints were carefully dissected out and homogenized. They were suspended in the transportation medium provided by the ELISA kit in a fashion as to normalize samples to the same wet weight per volume. All samples were frozen and stored at -70°C until assayed.

NM was purchased from Sigma (St. Louis, MO, USA). The treatment scheme of Li et al. 7 for mice was followed. Daily intraperitoneal injections at a dose of 10 mg/kg body weight were made. NM was dissolved fresh in water each day just before use and then filtered sterilized before injection. The animals were briefly anesthetized with Isofurane (Zenoca Pharma, Missisauga, Ont., Canada), weighed and then the precise amount injected intraperitoneally. The injection schedule began on the day before Chlamydia infection, and then on a daily basis until the rats were sacrificed. Sterile water was injected into the untreated control rats.

Lewis rat synovial fibroblast monolayers were set up on six well tissue culture plates. One hour prior to inoculation with Chlamydia, NM was added to the wells to give a range of concentrations from 0 to 200 μg/mL in duplicate wells. One hour after exposure to NM, the monolayers were infected at 10⁵ CFU/well. The plates were spun down at 2 000 × G for 20 minutes to impact the bacteria onto the cells. Thereafter, plates were incubated at 37°C in a 5% CO₂ incubator for 24 hours.

At the end of the incubation period, the monolayers were examined under phase contrast microscopy for Chlamydia inclusion bodies. Individual wells were harvested using cell scrapers and the contents frozen at -70°C until assayed.

The above IDEIA PCE Chlamydia kit from OXOID-Dako was also used for the quantitation of Chlamydia growth. Since the read out for this ELISA was in O.D., in order to combine different experimental runs, percent inhibition was calculated. The O.D. values from wells without NM was set at 0% inhibition.

The percentage difference from test wells with decreased O.D. values was calculated as percent inhibition.

Twenty four hour after fibroblast monolayers were inoculated with 10⁵ CFU/well of Chlamydia, large inclusion bodies could be seen under phase contrast microscopy within the majority of the cells. When NM was introduced into the culture media, obvious effects could be observed at 5 μg/mL concentration. The sizes of the inclusion bodies became more variable and smaller ones could be seen (Figure 1). With increasing concentrations of NM the number of inclusions decreased and also became smaller. At 50 μg/mL and above, no inclusion bodies could be found.

Under phase contrast microscopy, the number of inclusion bodies in 100 high-power (40x) fields per well in duplicate wells for each NM concentration were carefully enumerated. Results are plotted in Figure 2 which represents data averaged from three separate experiments. Without NM in the media, the average number of inclusion bodies per well was 17,886 (± 1415). At 5 μg/mL, there were 8,490 (± 756) inclusions. At 25 μg/mL 35 inclusions were found and no inclusion bodies were observed at 50 μg/mL.

The chlamydial antigens in each well were assayed by ELISA along the concentration gradient (Figure 3). Results are expressed as the average percent inhibition of three separate experiments with duplicate wells for each concentration. It can be seen that NM is a very powerful inhibitor of Chlamydia proliferation. At 25 μg/mL inhibition was almost complete. This mirrored the inclusion body counts above.

As expected from our model of CtIA, the non-treated control rats began to experience symptoms of joint inflammation beginning on the second day after intra-articular infection with synoviocyte-packaged Chlamydia. By the end of the experiment four days later, all the injected joints were severely swollen. Animals experienced gradual weight lost for the period of the experiment beginning on Day 3 (Figure 4). On the other hand, the body weights of similarly infected rats receiving daily treatments with NM remained constant. The infected knee joints of these treated rats only experienced minimal swelling. At necropsy, Chlamydia-infected joints of the rats were measured. The average lateral width of the infected joints for the NM treated animals was 8.55 mm (sd ± 0.6578, n = 10) while that of the untreated control rats was significantly larger, 11.18 mm (sd ± 0.5672, n = 10) (P < 0.001).

The knee joints of untreated control rats all had severe inflammatory changes with heavy infiltration of leukocytes, predominately neutrophils, consistent with our established experimental model (Figure 5). Soft synovial tissues were heavily hypertrophic with foci and necrosis, and there were extensive erosive changes. Most of the cartilage tissues showed structural damage with active panus formation. Panus could be seen invading through into the bone marrow spaces. Using the Yang-Hamilton grading scale, control knees had a mean score of 15.9 (± 1.45, n = 10). In contrast, infected knee joints from NM-treated rats showed much milder histopathology. The leukocyte infiltrations were moderate and the majority of the cells were mononuclear in nature. In areas where the infiltration is light, synovial fatty tissues could be seen. The synovial spaces were clearly defined and there were only minor erosive changes. The average Yang-Hamilton score for the test group is only 10.9 (± 2.45, n = 11). The difference between test and control scores is very significant at 9.79 × 10^-6.

The amounts of chlamydial antigen within the synovial tissues of the infected knee joints were assayed by ELISA. As shown in Figure 6, it is obvious that there is significantly more Chlamydia in the synovial tissues of the non-treated control knees than in the NM treated joints. The mean O.D. values for the controls is 0.18 (± 0.05, n = 4) and for the NM-treated knees is 0.05 (± 0.02, n = 4). The difference is significant (P < 0.001).