DBA/1 and C57BL/6 mice (Thy1.2 or Thy1.1 background) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained under specific pathogen-free conditions at the Yale University School of Medicine. DBA/1 background IL-10 receptor dominant-negative Tg mice were obtained by backcrossing C57BL/6 Tg mice 21 with wild-type (Wt) DBA/1 mice for nine generations. The H-2^q/q and Tg genotypes were screened by PCR using tail DNA. The IL-10 GFP knockin mouse, designated IL-10-internal ribosomal entry site (IRES)-GFP-enhanced reporter (Tiger), and the Foxp3 red fluorescent protein (RFP) knockin mouse, designated Foxp3-IRES-mRFP (FIR) mice, were described previously 2223. All the animal procedures were performed with the approval of the Institutional Animal Care and Use Committee of the Yale University School of Medicine.

Bovine CII for immunization, T-cell proliferation and ELISA, as well as Complete Freund's Adjuvant (CFA) and Incomplete Freund's Adjuvant (IFA), were all purchased from Chondrex (Redmond, WA, USA). BD GolgiPlug and antibodies for flow cytometry were purchased from BD (San Diego, CA, USA). A Foxp3 intracellular staining kit was purchased from eBioscience (San Diego, CA, USA). Cytokines and antibodies for T-cell differentiation in vitro were obtained from R&D Systems (Minneapolis, MN, USA). Phorbol 12-myristate 13-acetate (PMA) and ionomycin were purchased from Sigma-Aldrich (St Louis, MO, USA).

The CIA model was employed as described previously 24. Briefly, 100 μg of bovine CII dissolved in 0.1 M acetic acid was emulsified with 100 μl of CFA containing 2 mg/ml inactivated Mycobacterium tuberculosis. Mice were sensitized by intradermal injection at the base of the tail and boosted at day 21 with the same dose of CII and IFA intradermally. Mice were scored blindly for disease severity every other day after the boost using a grading scale from 0 to 3 that was described previously 24. Each paw was graded with a maximum score of 12 per mouse.

Splenocytes from CIA mice were cultured with different concentrations of CII for 3 days. After pulsing with 1 μCi of [³H]thymidine per well for the last 18 hours, proliferation was measured as radioactivity incorporation (counts per minute (cpm)).

Splenocytes from CIA mice were stained with anti-CD3, anti-CD4, anti-CD8, anti-CD44 and anti-CD62L antibodies, and the surface markers for activation in CD4⁺ and CD8⁺ T cells were analyzed using an LSR II flow cytometer (BD).

Serum samples were collected from different groups of immunized mice at different days postimmunization (days 0, 14, 28, 35 and 42), the levels of anti-CII antibody in different isotypes were measured by ELISA as described in our previous studies 25. Briefly, ELISA plates were coated with bovine CII (0.5 μg/well). After the plates were blocked with 5% fetal bovine serum-PBS, 1:800 diluted (with blocking buffer) mouse sera were added to duplicate wells and incubated for 2 hours at room temperature. The plates were then washed, and biotin-conjugated goat anti-mouse immunoglobulin G (IgG), IgG1 and IgG2a were added at a dilution of 1:2,000 for 1 hour, followed by the addition of streptavidin-horseradish peroxidase (Sigma-Aldrich). The plates were then developed, and antibody levels were measured as described previously 24.

Total RNA was isolated from joints by using the RNeasy Mini Kit (QIAGEN, Valencia, CA, USA). Genomic DNA was removed by digestion with DNase I. cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA), and gene expression was examined using the Stratagene Mx3005P QPCR System with Brilliant SYBR Green QPCR Master Mix reagent (both from Agilent Technologies, Inc, Santa Clara, CA, USA). The data were normalized to a gapdh reference 26. The forward and reverse primers for il17 and gapdh were as follows 27: il17, TTTAACTCCCTTGGCGCAAAA, CTTTCCCTCCGCATTGACAC; and gapdh, TTCACCACCATGGAGAAGGC, GGCATGGACTGTGGTCATGA.

Single-cell suspensions were prepared from spleens, lymph nodes or thymi and stained with surface CD4 and CD25. After being fixed with freshly prepared fixation/permeabilization solution at 4°C for at least 30 minutes in the dark, cells were blocked with Mouse BD Fc Block (BD) in permeabilization buffer at 4°C for 15 minutes and stained with anti-Foxp3 antibody at 4°C for 30 minutes. For intracellular cytokine staining, cells were stimulated with PMA plus ionomycin in the presence of brefeldin A for 4 hours. Cytokine production was measured as described previously.

Enriched CD4⁺ T cells were sorted into a CD4⁺CD25⁺ subset as the regulatory cells and a CD4⁺CD25^- subset as the responder cells. T-cell-free splenocytes treated with mitomycin C were used as antigen-present cells (APCs). The responder cells were cultured with the regulatory cells at ratios of 1:1, 1:0.5, 1:0.25 and 1:0.125 or without the regulatory cells as the positive control in the presence of 2 μg/ml anti-CD3 antibody for 3 days in 96-well U-bottom plates, and APCs were added with the responder cells at a 5:1 ratio. After pulsing with 1 μCi of [³H]thymidine per well, proliferation was measured as radioactivity incorporation. The suppressive efficiency was calculated using the following formula: percentage suppression = (cpm of positive control) - (cpm of experiment)/(cpm of positive control) × 100.

CD4⁺RFP⁺ Tregs were sorted from the spleens of B6 Thy1.1 IL-10 dominant-negative receptor Tg and Wt mice and were transferred into B6 Thy1.2 Wt mice at equal cell numbers. One day after transfer mice were immunized with CFA and bovine CII as described above. One week after immunization Foxp3 expression in transferred cells was analyzed by checking the levels of RFP.

On day 58 after CIA induction, mice were killed and their paws were dissected free and used for histologic examination. The joints were immediately fixed in 10% buffered formalin and decalcified in a decalcifying solution for about 3 to 5 days 24. The tissues were then processed and embedded in paraffin. Five-micrometer tissue sections were prepared and stained with H & E using standard methods. All sections were stained at the Yale Pathology Laboratory. Sections were blindly observed and scored by histologists. Five arthritis severity factors were assessed using a grading scale from 0 (normal) to 5 (severe), based on a previously described scoring system 28.

Data presented are means ± SEM. Statistical testing was two-tailed with a 5% level of significance. A two-tailed nonparametric paired test (Wilcoxon signed-rank test) was used to analyze the arthritis and histologic scores. Student's paired or unpaired t-test was used for all other statistical analyses (InStat version 2.03 software; GraphPad Software, San Diego, CA, USA).

To define the role of IL-10 signaling in T cells in the pathogenesis of CIA, we employed newly generated IL-10 receptor dominant-negative Tg mice in which T cells expressed high levels of Tg IL-10Ra that were truncated to remove its signaling domain 21. Tg mice were backcrossed with DBA/1 mice 10 times to develop Tg mice on the CIA-susceptible background (DBA/1, H-2^q). Four groups of DBA/1 mice (male Wt or Tg and female Wt or Tg) from the same cohort of littermates at age 8 weeks were immunized with bovine CII to induce arthritis using the standard CIA protocol described in our previous studies. It has been observed that even in the susceptible strain of mice, such as DBA/1 (H-2^q), female mice are more resistant than male mice, with a lower incidence and less severity of CIA 28. As expected, our experiment showed that male mice were highly susceptible to and females relatively resistant to CIA (Figure 1A). All the male Wt mice developed progressive disease (with 100% incidence) and had a mean arthritis score of 8.25 ± 2.88 (n = 10) on day 35 postimmunization. In contrast, all the female Wt mice developed visible arthritis (also 100% incidence) with a significantly lower mean arthritis score of 5.00 ± 4.10 (n = 12) on day 35 postimmunization. Interestingly, female Tg mice developed severe progressive disease and had a mean arthritis score of 8.78 ± 2.33 (n = 10), slightly higher than those parameters in male Wt mice and male Tg mice (7.89 ± 2.49, n = 12). Further evidence in support of increased arthritis severity in female Tg mice was obtained by histopathologic analysis. On day 50 postimmunization, female Tg mice exhibited robust pannus formation and significant articular cartilage erosion, with pannus eroding and replacing the articular cartilage overlying the bone. In contrast, Wt female mice exhibited well-preserved joint spaces and articular cartilage surfaces, with minimal pannus formation. Among male mice, both Wt and Tg exhibited extensive pannus formation and destruction of bone and cartilage. One representative example is shown in Figure 1B. At 50 days postimmunization, the histologic score of Wt female mice was 3.8 ± 1.4 of a possible score of 25. All the other groups showed similar scores (Figure 1C) (female Tg, 13.7 ± 0.6; male Wt, 14.7 ± 0.8; male Tg, 13.3 ± 1.0, respectively; n = 4 mice in each group). Because the major difference in CIA disease severity was observed between female Wt and Tg mice rather than male Wt and Tg mice, only female mice were used for the rest of our study. Thus our data suggest that IL-10 signaling in T cells plays a critical role in suppressing the pathogenesis of CIA.

Activation and cytokine production of CD4⁺ T cells play an important role in the pathogenesis of CIA. To define the underlying mechanisms of increased arthritis severity in female Tg mice, T-cell activation and proliferation in CIA mice were first analyzed. Cells from draining lymph nodes on day 50 postimmunization were stained with anti-CD3, anti-CD4, anti-CD8, anti-CD44 and anti-CD62L antibodies. After gating on CD3⁺ T cells, the CD44^hi T cells were recognized as activated cells. Activated CD4⁺ was found to be significantly higher in Tg mice than in Wt mice (Figure 2A, left), whereas activated CD8⁺ in Tg and Wt mice appeared to be comparable (Figure 2A, right), implying that IL-10 signaling in T cells controlled CD4⁺ T cell activation in inflammatory response. One representative dot plot is shown (Figure 2B). To test the impact of IL-10 signaling in antigen-specific T-cell proliferative response in immunized mice, cells recovered as described above were cultured in 96-well U-bottom plates in the presence of different concentrations of CII for 3 days, followed by pulsing with 1 μCi of [³H]thymidine per well for the last 18 hours, and proliferation was measured as radioactivity incorporated (in cpm). Cells from Tg mice showed significantly increased proliferative responses to specific antigen (Figure 2C), suggesting a critical role of IL-10 signaling in T cells in controlling lymphocyte cell proliferation in CIA. To further test whether IL-10 signaling in T cells would affect T-cell differentiation in CIA, we also analyzed IFN-γ-producing CD4⁺ T cells in draining lymph nodes upon immunization and found elevated percentages of IFN-γ⁺CD4⁺ cells in Tg mice (Figure 2D), suggesting that IL-10 signaling suppressed the pathogenic Th1 response. One representative dot plot is shown (Figure 2E). Interestingly, no significant differences were observed in terms of the numbers and the activation states of γδ T cells in the spleens and draining lymph nodes of Wt and Tg mice (data not shown).

It has been demonstrated that CD4⁺Foxp3⁺ Tregs play important roles in the suppression of CIA in animal models and human studies 910. To define the role of IL-10 signaling in T cells in the development of naturally occurring Tregs (nTregs), the percentage and absolute number of CD4⁺CD25⁺Foxp3⁺ T cells in DBA/1 and B6 Tg and Wt mice were analyzed. Very similar percentages and numbers of CD25⁺Foxp3⁺ T cells were found in both Wt and Tg mice (Figures 3A and 3B), indicating that IL-10 signaling in T cells was not required for the development of nTregs. Similar results were observed in DBA/1 mice upon immunization (data not shown). To further define the function of Tregs, a standard in vitro T-cell suppression assay was performed. CD4⁺CD25⁺ T cells were sorted from B6 Wt or Tg mice as Tregs and CD4⁺CD25^- T cells from B6 Wt mice as responders. Tregs from both Wt and Tg mice were anergic to TCR activation (data not shown). Interestingly, the suppressor function of Tregs from Tg mice was significantly impaired in comparison to those from Wt mice (Figures 3C and 3D). To rule out the contamination of activated CD4⁺ T cells in the Treg population when using CD25 as the marker of Tregs, RFP-Foxp3 reporter mice 23 were used to sort Foxp3⁺ regulators and Foxp3^- responders and the suppress assay mentioned above was repeated. Similar results were obtained (Figure 3E). Our results demonstrate that IL-10 signaling in T cells is not required for Treg development, but rather for T cells' suppressive function.

To define the mechanisms underlying the reduced suppressive function of Tregs in Tg mice, we first analyzed IL-10 production in Tregs. For this purpose, we used IL-10-Foxp3 double-reporter mice 2223. The percentage of IL-10-expressing cells was similar between Wt and Tg mice in both unimmunized and immunized states (Figures 4A and 4B), indicating that IL-10 signaling in T cells was not required for IL-10 production by Tregs and that IL-10 was not related to the impaired suppressive function of Tg Tregs. We next analyzed the transforming growth factor (TGF)-β production by Tregs in Wt and the Tg mice. Lymphocytes from Wt and the Tg mice were stimulated with PMA and ionomycin for 4 hours in the presence of brefeldin A, and intracellular Foxp3 and TGF-β were stained in the CD4⁺ T cell population. No significant differences in TGF-β production between Tregs from Wt and the Tg mice were detected (Figure 4C), suggesting that IL-10 signaling in T cells was not involved in TGF-β production. We then investigated whether IL-10 signaling in T cells would affect CTLA-4 expression level in Tregs. Cells from the lymph nodes of unimmunized and immunized Wt and the Tg mice were stained with Foxp3 and intracellular CTLA-4. CTLA-4 was highly expressed in CD4⁺Foxp3⁺ T cells, but not in CD4⁺Foxp3^- T cells; however, no significant differences were found in the levels of CTLA-4 in Tregs from Wt and Tg mice (Figure 4D). Because Foxp3 has been reported to be the master transcription factor controlling Treg development and determining the suppressive function of Tregs 293031, we studied whether IL-10 signaling played any role in maintaining the stable expression of Foxp3. Similarly to the results shown in Figure 3A, the percentage of Foxp3⁺ T cells was very similar between Wt and Tg mice (Figure 4A, left). Interestingly, it decreased by 30% upon immunization (Figure 4A, right, and Figure 4E). A similar phenomenon was observed in DBA/1 mice using intracellular staining (data not shown). To define whether IL-10 signaling was critical for the maintenance of Foxp3 expression in Tregs, CD4⁺RFP⁺ T cells from Wt and Tg Thy1.1 mice were sorted and transferred to Thy1.2 Wt mice, followed by immunization with CII and CFA. One week later the transferred cells were analyzed for expression of RFP. Transferred Foxp3⁺ cells were observed to become RFP^- cells in both Tg and Wt Tregs (Figures 4F and 4G). However, Tregs from Tg mice showed significantly higher percentages of RFP^- cells than those from Wt mice, indicating that IL-10 signaling in Tregs was critical to maintaining the level of Foxp3 expression upon immunization.

IL-17 is a critical cytokine in the pathogenesis of CIA 32. To define the role of IL-10 signaling in T cells on IL-17 production by CD4⁺ T cells, we first compared the IL-17⁺CD4⁺ T cells from unimmunized mice. We found that Tg mice had increased Th17 cells in both peripheral and central lymph organs (Figure 5A), suggesting that IL-10 signaling negatively regulates naturally occurring Th17 cell development in steady state. To further define the role of IL-10 in Th17 cell differentiation, we sorted naïve CD4⁺ T cells (CD4⁺CD44^lo CD62L^hi cells) from Tg as well as IL-10-knockout (IL-10Ko) mice versus Wt controls cultured under Th17 differentiation conditions in the presence or absence of IL-10, and the cells were then used to analyze cytokine production. No significant difference was found in terms of the percentage of IL-17-producing cells among Tg, Ko and Wt mice with or without IL-10, indicating that IL-10 signaling had no effect on inducible Th17 cell differentiation (Figure 5B). To define the role of IL-17 in local inflammatory response, the level of IL-17 mRNA (il17) in the joints of CIA mice was analyzed using real-time PCR. We found that joint tissues from Tg mice expressed significantly higher levels of il17 than Wt mice did (Figure 5C), which was consistent with the severity of arthritis. To define the potential sources of IL-17 in the joints, cells from draining lymph nodes of CIA mice (Wt and Tg mice) were analyzed for IL-17A production. Interestingly, significantly higher percentages of γδ T cells from Tg mice produced IL-17 than those from Wt mice (Figure 5D), especially CD44^hi γδ T cells. In contrast, we did not find significant differences in the number of IL-17-producing CD4⁺ T cells (Figure 5E), suggesting γδ T cells as the primary source of IL-17 in the inflammatory joints.

Antibody production against CII is an important factor in the development of CIA 33. To define the impact of IL-10 signaling in T cells on collagen-specific antibody response, serum samples were collected from naïve mice as well as from different groups of immunized mice at different time points after immunization (days 0, 14, 28, 35 and 42), and the levels of anticollagen antibodies were analyzed by ELISA as described in our previous studies 24. The level of collagen-specific antibody in serum increased upon immunization and was maintained at high levels between days 14 and 42. Surprisingly, there was no difference in collagen-specific antibodies, including the total IgG, IgG1 and IgG2a isotypes, among Wt and Tg mice (Figure 6). These results suggest that blocking IL-10 signaling in T cells has no direct effect on collagen-specific antibody production by B cells in a CIA model.