DBA/1 mice (Taconic Europe A/S, Ry, Denmark), 6 to 8 weeks old, were used for CIA experiments. For the delayed-type hypersensitivity (DTH) experiment, 6 to 8 week old mice were used. All of the mice were maintained in the animal facility of the Department of Rheumatology and Inflammation Research, University of Gothenburg, Sweden, in accordance with the local ethics board animal husbandry standards. Mice were housed up to 10 animals per cage under standard conditions of light and temperature and fed with standard laboratory chow ad libitum.

Chicken CII (Sigma-Aldrich, St. Louis, MO, USA) was dissolved at a concentration of 2 mg/ml in 0.1 M acetic acid and then emulsified in an equal volume of complete Freund's adjuvant (Sigma-Aldrich). Arthritis was induced by intradermal injection of DBA/1 mice at the base of the tails with 100 μL of the emulsion. Booster immunization containing 100 μg of CII in incomplete Freund's adjuvant (Sigma-Aldrich) was administered 21 days after the priming. The experiments were terminated in 6 to 8 weeks.

CIA was used to investigate the effect of DCA (sodium DCA 99% purity; BuyDCA, Sonora, CA, USA) on arthritis. DCA was administered by dissolving it in the drinking water. Control mice were given water only. The average amount of DCA per mouse taken was determined by measuring the volume of DCA solution that mice consumed in each cage. We calculated and adjusted the concentration of DCA required to achieve a daily dose of 0.3 mg or 3 mg DCA/mouse per day. DCA was provided in the drinking water from day 0 of all experiments. The experiment was repeated three times as outlined in Table 1. To study the impact of estrogen on DCA-mediated effects, endogenous estrogen production in female DBA/1 mice was blocked by ovariectomy (OVX). Sham-operated mice were used as controls. DBA/1 mice were regularly weighed from the day of priming and checked for the development of arthritis after booster immunization. When the experiments were terminated, blood was drawn for serological analyses. Paws were processed for histological analyses.

All of the DBA/1 mice were inspected every second or third day after booster to assess the presence of arthritis. To evaluate the intensity of arthritis, a clinical scoring system of 0 to 3 points for each paw was used: 0, no sign of inflammation; 1, mild swelling or erythema or both; 2, moderate swelling and erythema; and 3, marked swelling and erythema. The arthritic index for each mouse was constructed by summing up the scores of all four limbs.

To assess the impact of DCA on a T cell- and macrophage-dependent inflammatory response 1516, the DTH reaction was performed. Thirty female mice were divided into three groups (10 mice per group). Two groups were provided DCA in drinking water (0.3 and 3 mg/mouse per day, respectively). The control group was provided water only. After 2 days, all of the mice were immunized by epicutaneous application of 150 μL of a mixture of ethanol acetone (2:1) containing 3% (vol/vol) oxazolone (OXA) (Sigma-Aldrich) on the abdomen skin. One week after the priming, the right ears were challenged on both sides by topical application of 30 μL of 1% OXA, which was dissolved in olive oil. Thirty microliters of olive oil only was applied to the left ears as vehicles. The intensity of DTH reaction was examined as previously described 17.

Olive oil-induced skin inflammation is granulocyte-mediated but T cell- and monocyte-independent 18. A single intradermal injection of olive oil into mouse footpad induces massive infiltration of polymorphonuclear cells, which give rise to a localized footpad swelling. The thickness of footpad can be measured and relates to severity of the inflammatory process 19. Thirty microliters of olive oil (Apoteksbolaget, Göteborg, Sweden) was injected intradermally in a hind foot dorsum of mice. Footpads were measured before and 24 hours after injection using an Oditest spring caliper (Kröplin, Schluchtern, FRG). The footpad swelling was expressed as footpad increased thickness (in millimeters) after injection and was scored as described previously 20.

The level of serum hormones was analyzed by the following radioimmunoassays: insulin-like growth factor (IGF1) (Mediagnost, Reutlingen, Germany), testosterone (MP Biomedicals, Irvine, CA, USA), and cortisol (CIS Bio International, Marcoule, France). Uteri were weighed as an indirect indicator of estradiol level 21. Anti-CII antibody analyses were performed as previously described 22. IL-6 levels in sera were analyzed as described before 23.

All four paws from DBA/1 mice were excised, followed by routine fixation, decalcification, and paraffin embedding. Tissue sections were stained with hematoxylin/eosin. The sections were studied by a blinded examiner regarding synovitis and erosion of bone/cartilage. Synovial hypertrophy was defined as a membrane thickness of more than two cell layers 23. A histological scoring system was used: 1, mild; 2, moderate; and 3, severe synovitis or bone erosion 24. Knee joints, ankles, toes, elbows, wrists, and hands were examined. A mean score from all of the inspected paws for each animal was calculated 25.

The left femurs from DBA/1 mice were fixed in 4% (vol/vol) buffered formaldehyde for 3 days and then replaced by 70% (vol/vol) alcohol until analyses of bone mineral density (BMD) were performed. A peripheral quantitative computed tomography (pQCT) scan with a Stratec pQCT XCT Research M (Norland, Fort Atkinson, WI, USA) was used as previously described 26. Trabecular BMD was analyzed with a metaphyseal scan at a point located at a distance of 3% of the length of the femur from the distal growth plate. The inner 45% of the area was defined as the trabecular bone compartment. Cortical bone parameters were determined with a middiaphyseal scan, which contained only cortical bone.

Statistical analyses were performed by using the Mann-Whitney U test and the chi-square test. Values are reported as median ± 10% to 90% range. A P value of less than 0.05 was considered significant.

To evaluate whether DCA had an effect on the development of CIA, male and female DBA/1 mice were provided drinking water with or without DCA from the priming day until the experiment was terminated. At a dose of 3 mg DCA/mouse per day, for female DBA/1 mice, none of the 22 mice that drank DCA had signs of arthritis 37 days after the priming with CII, whereas most of the control mice (16 of 22, or 73%) (P < 0.0001) already had ongoing arthritis (Figure 1a). Female mice that drank DCA had a much lower severity of arthritis (Figure 1b). For male mice, there was no difference between the DCA-drinking group and the water group in regard to the onset of arthritis and its severity. In the dose of 0.3 mg DCA/mouse per day, we did not find any impact of DCA on the development and course of CIA (data not shown).

Histological sections from female mice confirmed that the DCA group had a lower severity of arthritis. Notably, the destruction of bone and cartilage was significantly diminished in the DCA-drinking group compared with the control group (Figures 1c and 2). In contrast, there was no significant difference in male mice between the two groups. Importantly, DCA-treated mice did not show decreased weight gain as compared with control mice (Figure 1d), indicating that DCA was not toxic. In fact, female mice receiving DCA gained slightly more weight than their controls.

A potential mechanism by which DCA could suppress CIA is the blocking of anti-CII antibody production. Anti-CII antibody levels and serum IL-6 were analyzed in serum obtained at the termination of experiments. In female mice, but not in male mice, circulating anti-CII IgG antibody levels were significantly decreased in the DCA-drinking group compared with the water group (P = 0.04) (Figure 3a). The serum IL-6 level in the DCA-drinking group was lower than in the water group, but the data did not reach statistical significance (P = 0.06) (Figure 3b). Thus, the anti-inflammatory effect of DCA on arthritis is accompanied by lower levels of anti-CII antibodies.

DTH is a T cell-mediated immune reaction. To test the effect of DCA on DTH, DCA-treated mice and water controls were epicutaneously immunized and challenged with OXA. DTH reactivity was registered by measuring the increase in ear thickness 24 hours after the challenge. We did not find any significant differences in regard to the severity of DTH between DCA-treated mice and their controls (data not shown). Likewise, DCA had no effect on the granulocyte-dependent olive oil-induced inflammation (data not shown).

To evaluate whether the ameliorative effect of DCA on CIA also was reflected in the protection of arthritis-induced bone loss, the left femur of each mouse was subjected to a pQCT scan at the termination of CIA. Mice immunized with CII and treated with DCA displayed significantly higher cortical bone mineral content than did the water-drinking group (P = 0.001) (Figure 4a). Likewise, thickness of cortical bone (P = 0.039) (Figure 4b) and cortical bone area (P = 0.01) (Figure 4c) were significantly higher in the DCA-treated group. No differences were observed with respect to total BMD and trabecular BMD between the DCA-drinking group and their water controls (Figure 4d, e).

Because only female mice responded to DCA therapy, we hypothesized that the beneficial effect of DCA on arthritis was related to estrogen. To evaluate the importance of estrogens, one group of DBA/1 mice was ovariectomized (OVX) and another group had sham surgery. After 1 week, both groups were primed and booster - immunized with CII and treated with DCA (3 mg DCA/mouse per day) as previously described. On day 30, 12 of 31 mice in the OVX group had signs of arthritis (38.7%) compared with only 1 of 29 mice in the sham controls (3.4%) (P = 0.001). On day 34, 17 of 31 in the OVX group had arthritis (54.8%) but only 5 of 29 in the sham-operated group did (17.2%) (P = 0.003). When the experiment was terminated on day 38, 19 of 31 mice in the OVX group had developed arthritis (61.3%). In contrast, only 10 of 29 mice in the sham-operated group (34.5%) had signs of arthritis (P = 0.04) (Figure 5a). The sham-operated group also had much less severe arthritis than the OVX group (P = 0.006) (Figure 5b). Histological analysis showed that the OVX group had synovitis (P = 0.004) and bone erosion (P = 0.01) that were significantly more severe compared with sham-operated controls (Figure 5c). But no difference was found between the OVX group and the sham-operated group in regard to weight changes (data not shown).

Although OVX mice and sham-operated mice had different arthritis responses following DCA treatment, we did not find any difference between the OVX group and the sham group in regard to level of IL-6 (OVX: median 76 pg/mL, range 20 to 1,032 pg/mL versus sham: median 111.2 pg/mL, range 30 to 348.2 pg/mL) and anti-CII antibodies (optical density: OVX: median 0.91, range 0.37 to 1.32 versus sham: median 0.84, range 0.30 to 1.69). As an indirect indicator of estrogen levels, the weight of uteri from DCA- and water-treated intact female mice was recorded. No difference was found between the DCA group and their controls (DCA: median 50 mg, range 26 to 111 mg, n = 8 versus water: median 49 mg, range 15 to 104 mg, n = 8). We also measured the level of testosterone. There was no significant difference between the DCA group and the control group (Data not shown).