Lewis (LEW/Hsd) (RT.1^l) and Wistar-Kyoto (WKY/NHsd) (RT.1^l) rats were purchased from Harlan Sprague-Dawley (HSD) (Indianapolis, IN, USA and Madison, WI, USA, respectively). Male, 4 to 6-week-old rats were used in this study. These rats were housed in the vivarium of the University of Maryland School of Medicine, Baltimore, MD, USA (UMB) and were treated as per the guidelines of the institutional animal care and use committee (IACUC) of UMB (protocol no. 0206011).

Mycobacterial hsp65 (Bhsp65) peptides 177 to 191 (B177) and 333 to 347 (B333), and HEL peptide 65 to 78 (HEL65) were obtained from Macromolecular Resources and Global Peptide Services (both at Fort Collins, CO, USA) 2122. The recombinant Bhsp65 was expressed and purified, as well as rendered free of endotoxin as described elsewhere 2122. Hen egg white lysozyme (HEL) and Concanavalin A (Con A) were purchased from Sigma-Aldrich Co. (St Louis, MO, USA), whereas purified protein derivative (PPD) was obtained from Mycos Research (Fort Collins, CO, USA). Recombinant rat TNFα was purchased from R&D Systems (Minneapolis, MN, USA), and its endotoxin content was below 1 endotoxin unit (EU)/μg. Units of TNFα were determined as ED₅₀ (1 U) = 15 pg.

LEW rats were immunised subcutaneously at the base of the tail with heat-killed M. tuberculosis H37Ra (Mtb) (Difco, Detroit, MI, USA) (1 mg/rat) suspended in oil (Sigma-Aldrich). Beginning on day 7 after Mtb challenge, these rats were observed and graded regularly for the severity of arthritis on the basis of erythaema and swelling of the paws on a scale of 0 to 4 as described elsewhere 1222. The highest arthritic score was 4 for each paw, with a maximum score of 16 per rat. Different phases of AA were labelled as follows: incubation (Inc), onset (Ons), peak (Pk), and recovery (Rec) phase.

Arthritic LEW rats were killed at different phases of AA (Inc, Ons, Pk, and Rec) and their draining lymph nodes (para-aortic, inguinal, and popliteal) were harvested post-Mtb challenge. For comparison, LNC of WKY rats immunised with Mtb were harvested at the time points corresponding to different phases of AA in LEW rats. Thereafter, a single-cell suspension of LNC was prepared, and the cells were washed three times with Hank's balanced salt solution (Invitrogen, Frederick, MD, USA) 1222. These LNC were cultured (2.5 × 10⁵ cells/well) for 4 days with or without antigen at 37°C in an atmosphere of 95% air and 5% CO₂ in a flat-bottomed 96-well plate in HL-1 serum-free medium (Ventrex Laboratories, Portland, ME, USA), which was supplemented with 2 mM L-glutamine, 100 U/ml penicillin G sodium, and 100 μg/ml streptomycin sulfate. HEL, HEL65, or B333 served as negative control antigens, whereas Con A or PPD was used as a positive control. The antigens were used at a pre-titred final concentration of 25 ug/ml that was determined to be optimal for comparison through pilot experiments. After 4 days of culture, the cells were pulsed with 1 μCi/well of [³H]-thymidine (International Chemical and Nuclear, Irvine, CA, USA) and then harvested after 16 to 18 h. The results were expressed either as counts per minute (cpm) or as a stimulation index (SI = cpm of cells cultured with antigen/cpm of cells in medium alone).

The LNC harvested from Mtb-immunised LEW and WKY rats were cultured in a 96-well plate as described above. These LNC were then re-stimulated in vitro for 48 to 72 h with the appropriate antigen, and the culture supernates were collected thereafter 22. These supernates were then tested by ELISA using commercially available kits for the detection of TNFα, IFN-γ and IL-10 (all from Biosource, Camarillo, CA, USA), with lower detection limits (pg/ml) of 4, 13, and 10, respectively. The results were expressed as pg/ml. For comparison of different groups, the background cytokine level was deducted from the antigen-specific cytokine secretion (pg/ml of cytokine from cells cultured with antigen – pg/ml of cytokine from cells in medium alone; also referred to as Δ pg/ml) 2223.

TNFα was injected intraperitoneally daily into naive LEW rats at 1 × 10⁵ U/ml per injection beginning 3 days before immunisation subcutaneously with Mtb on the fourth day. TNFα treatment was continued through 6 days post Mtb injection for a total of 10 injections. Control rats received equal number of injections of phosphate-buffered saline (PBS) following the same protocol as that used for TNFα injections, including Mtb injection after 3 days of starting PBS injection. Thereafter, all rats were observed regularly for signs of arthritis, and the severity of the disease was scored as described above.

Blood from LEW rats treated with TNFα in vivo as described above along with that from control rats was collected either from tail vein or via cardiac puncture. The serum was separated from the clotted blood and tested in ELISA for the presence of sTNFR-I or anti-TNFα antibody. ELISA for sTNFR-I (R&D Systems, Minneapolis, MN, USA) was performed following the manufacturer's instructions, and the results were expressed as pg/ml. ELISA for anti-TNFα antibody was set up and optimised in-house. The ELISA plate (Greiner, Monroe, NC, USA) was coated with 100 μl (0.1 μg/well) of TNFα (Biosource, Camarillo, CA, USA) overnight at 4°C. After washing with PBS containing 0.05% Tween-20 (PBST), the wells were blocked with 200 μl/well of 10% bovine serum albumin (BSA) in PBST. Thereafter, the plate was washed, and 100 μl of diluted rat sera (1:50, 1:100, 1:200, and 1:400) were added per well and incubated at room temperature for 1 h. After washings, 100 μl of horseradish peroxidase (HRP)-conjugated polyclonal anti-rat antibody (BD PharMingen, San Diego, CA, USA) (1:2,500) was added per well. After 1 h at room temperature, the plate was washed, and the colour was developed by adding 30 μl/well of ABTS substrate (Bio-Rad, Hercules, CA, USA) and incubating for 15 min. The colour reaction was then stopped with 50 μl/well of 0.5 M H₂SO₄. The OD₄₅₀ was measured using a Vmax microplate reader (Molecular Devices, Sunnyvale, CA, USA).

The draining LNC were harvested from TNFα- or PBS-treated and Mtb immunised rats, and cultured for 48 h in the presence or absence of the appropriate antigen. Total RNA was prepared from 1 × 10⁶ cells and reverse-transcribed using the iScript cDNA synthesis kit (Bio-Rad Laboratories). The cDNA thus obtained was amplified using an ABI Prism 7900HT cycler (Applied Biosystems, Foster City, CA, USA) 24. The primers used in the assay for the detection of mRNAs for IL-17, IDO, TTS, and hypoxanthine-guanine phosphoribosyl transferase (HPRT) were designed using the Primer Express 2.0 program (Applied Biosystems) and were synthesised at the UMB Biopolymer Core Facility. The mRNA levels of each entity tested were normalised to the HPRT gene, and the relative gene expression levels were determined 24. The results were expressed as 'fold increase' over mRNA levels of cells cultured in medium alone. We also confirmed the IDO mRNA expression results in splenic adherent cells (macrophages and dendritic cells).

The Student t test assuming equal or unequal variance (determined by the F test) was used as appropriate for the data to test the statistical significance of the differences observed among various test and control groups. A non-parametric Wilcoxon rank sum test was employed to compare the arthritic scores of any two groups of rats over the entire disease course. The results were considered significant at p < 0.05.

The results of ex vivo TNFα secretion (cytokine secretion without any exogenously added antigen; Figure 1) showed that there was a gradual increase in levels along with the progression (time post-Mtb injection) of AA in LEW rats with the highest level observed during Rec phase, while an opposite pattern was observed in WKY rats. However, following re-stimulation with Bhsp65, TNFα secretion was at a high level in both LEW and WKY rats without significant changes during the course of AA (Figure 1). Importantly, the level of TNFα secreted in response to the pathogenic epitope B177 of Bhsp65 was significantly increased at the Rec phase of AA in the LEW rats, but at Inc phase in WKY rats. Overall, the highest level of TNFα secretion was observed during Rec phase in LEW rats, but at Inc phase in WKY rats.

The above results showed that high TNFα levels correlate with recovery from acute AA in LEW rats, and with resistance against AA in WKY rats. To further examine this correlation, we tested the effect of TNFα treatment on AA in the LEW rats. Naïve LEW rats were given a total of 10 injections intraperitoneally of TNFα (10⁵ U/day) in PBS with three doses given before Mtb-injection and then continued on the day (fourth day) of Mtb injection and for 6 more days thereafter. After Mtb injection, rats were observed regularly for signs of arthritis. The control rats received 10 injections of PBS and were injected with Mtb at the same time as the experimental rats. The results (Figure 2) revealed that the TNFα-treated rats had a significantly reduced severity of AA compared to that of the control rats. This suppression of AA in the experimental group of rats was evident before the peak of AA, and it continued for an average of 7 days. Thus, treatment with TNFα, a pro-inflammatory cytokine, significantly attenuated the severity of AA in the LEW rats.

As TNFα treatment decreased the severity of AA, we tested whether the suppression of AA involved any major changes in the immune responsiveness to antigenic challenge. LEW rats were treated with TNFα using the protocol described above, including immunisation with Mtb or a control antigen (HEL/IFA). After 9 days of antigenic challenge, the draining LNC of these rats were harvested and tested for proliferative and cytokine response using Bhsp65, HEL, using their peptides as recall antigens. We obtained comparable (p > 0.05) numbers of LNC from TNFα-treated and PBS-treated rats in both Mtb-immunised and HEL-immunised groups (data not shown), suggesting that, at the dose used, the injected TNFα did not lead to a significant change in the number of cells (for example, via apoptosis) in the draining lymph nodes. In the cohort of Mtb-immunised rats, the LNC recall response to Bhsp65 and B177 in TNFα-treated rats was comparable to that of PBS-treated control rats (Figure 3a). Similar results were obtained in the two groups of rats that were immunised with HEL (Figure 3b) instead of Mtb. Furthermore, the results of cytokine testing showed that IFN-γ secretion by LNC of TNF-treated, Mtb-immunised LEW rats after B177 recall decreased significantly compared to that of the PBS-treated control rats (Figure 3c). This decrease in IFN-γ secretion in Mtb-immunised LEW rats was specific to B177 as IFN-γ response to the control antigen (HEL) in PBS-treated, HEL-immunised rats was comparable to that of TNFα-treated, HEL-immunised rats (Figure 3d). As there was no difference in the level of IL-10 secretion between the two Mtb-immunised groups (Figure 3e,f), the decrease in IFN-γ in response to B177 in TNFα-treated, Mtb-immunised LEW rats steered the overall cytokine response towards a T_h2 type. Thus, TNFα treatment neither resulted in a general non-specific enhancement of antigen-specific proliferative T cell response, nor induced a generalised immunosuppression. Instead, the AA-protective effect of TNFα involved a decrease of IFN-γ in response to the pathogenic epitope (B177) of Bhsp65.

In another set of experiments, we examined whether TNFα treatment had any effect on IL-17 production by Bhsp65- or B177-reactive T cells. We tested IL-17 by qRT-PCR because of rather limited reagents for the newer rat cytokines, including IL-17. The level of IL-17 in TNFα-treated, Mtb-immunised rats was comparable (p > 0.05) to that of PBS-treated, Mtb-immunised rats (data not shown).

We examined two other parameters that might contribute to TNFα-mediated protection against AA. First, the excessive shedding of sTNFR-I 2526, and second, the generation of anti-TNFα antibody that might neutralise TNFα in vivo 27. The levels of sTNFR-I (Figure 4A) as well as anti-TNFα antibodies (Figure 4B) in sera of TNFα-treated, Mtb-immunised LEW rats were comparable to that in sera of PBS-treated, Mtb-immunised LEW rats.

We also tested whether intraperitoneal injection of TNFα might deviate the migration of T cells away from the joints into the peritoneal cavity. Our results showed no difference in the number/proportion of T cells (CD3) or macrophages/neutrophils (CD11b/c) infiltrating into the peritoneal cavity after PBS treatment vs TNFα treatment intraperitoneally (Figure 5). These results suggest the absence of a major shift in the migration of T cells into the peritoneal cavity following TNFα treatment.

To gain further insights into the mechanisms by which TNFα treatment might suppress AA, we compared the relative levels of components of the two immunosuppressive pathways, the IDO-TTS by qRT-PCR and the Treg by flow cytometry. IDO is predominantly expressed in myeloid cells, and it catabolises tryptophan 192028. By contrast, TTS binds to tryptophan and makes it available for protein synthesis 192028. The IDO-induced deprivation of tryptophan has been invoked in T cell tolerance and suppression of T cell response. Similarly, Treg can suppress the activity of pathogenic effector T cells via cell-cell contact and immunomodulatory cytokines, TGF-β and IL-10 1718. Our results show that the levels of IDO mRNA in Bhsp65-restimulated LNC of TNFα-treated rats (1.51 fold compared to LNC in medium) were comparable (p > 0.05) to that of PBS-treated rats (2.26 fold). Similarly, the TTS mRNA levels in TNFα-treated versus PBS-treated rats were 1.69 fold versus 2.6 fold, respectively, and this difference was not significant (p > 0.05). The results for IDO mRNA testing using splenic adherent cells (data not shown) were similar to that obtained with LNC. In regard to Treg frequency, the levels (mean ± SEM) were slightly lower in TNFα-treated rats (8.5% ± 0.4) than that of PBS-treated rats (10.2% ± 0.3), but this difference was not statistically significant (p > 0.05).