SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Objective

Interleukin-21 (IL-21) is a T cell–derived cytokine that modulates T cell, B cell, and natural killer cell responses. In this study, the effects of blocking IL-21 were examined in 2 rodent models of rheumatoid arthritis (RA) to determine whether IL-21 contributes to their pathologic processes.

Methods

DBA/1 mice were immunized with bovine type II collagen and then treated with murine IL-21 receptor Fc fusion protein (IL-21R.Fc), which was initiated after the onset of arthritis symptoms in 10% of the cohort. The mice were assessed 3 times per week for signs of disease, including histologic features as well as serum cytokine, Ig, and cytokine messenger RNA (mRNA) levels in the paws. In a separate experiment, Lewis rats were immunized with Freund's complete adjuvant followed by administration of IL-21R.Fc at the peak of inflammation in the joints. Rats were assessed daily for histologic features and for scoring of arthritis severity. In addition, the effects of IL-21R.Fc on the production of interferon-γ (IFNγ) by T cells were examined.

Results

Treatment of DBA/1 mice with IL-21R.Fc reduced the clinical and histologic signs of collagen-induced arthritis. Nonspecific IgG1 levels were decreased in response to treatment. The levels of IL-6 mRNA in the paws and the serum IL-6 levels were decreased after treatment with IL-21R.Fc. IFNγ mRNA levels were increased in the paws, and the addition of IL-21R.Fc to collagen-activated lymph node cultures enhanced the levels of IFNγ. Collagen-specific spleen cell responses in IL-21R.Fc–treated mice were observed as reduced levels of IFNγ and increased levels of IL-6. Treatment of Lewis rats with IL-21R.Fc after induction of adjuvant-induced arthritis resulted in reversal of disease signs and improvements in histologic parameters.

Conclusion

These findings demonstrate a pathogenic role for IL-21 in animal models of RA, and support consideration of IL-21 as a therapeutic target in human RA.

T cells have been implicated in pathologic processes in animal models of rheumatoid arthritis (RA), by producing potentially inflammatory cytokines and by providing help for T cell–dependent autoantibody responses. Recent clinical data on the effects of abatacept, a reagent targeting the B7/CD28 costimulatory pathway, support the role of T cells in the pathogenesis of RA (1). T cells may contribute to human disease by driving pathogenic autoantibody responses, consistent with the success of B cell–depleting strategies in RA. In addition, T cells may contribute to disease by producing inflammatory cytokines.

Interleukin-21 (IL-21), a member of the type I cytokine family, is related most closely to IL-2 and IL-15 (2). IL-21 is made by activated CD4+ T cells and regulates the T cell, B cell, and natural killer (NK) cell responses (3, 4). Mice deficient in IL-21 receptor (IL-21R) have normal T cell and NK cell development and do not develop spontaneous autoimmune disease, suggesting a role for IL-21 that is distinct from that of IL-2 or IL-15 (5, 6). IL-21R–deficient mice have deficits in humoral immunity (6), consistent with a nonredundant role for IL-21 in B cell responses. In addition, multiple studies indicate that IL-21 can potentiate antitumor responses, supporting a potential role for IL-21 as a proinflammatory cytokine (7–9).

The receptor complex for IL-21 is composed of IL-21R, which specifically binds IL-21, together with the common cytokine receptor γ-chain (10, 11). In contrast to the cytokine, IL-21R is expressed on multiple cell types in the immune system (2, 5, 12), including T cells, B cells, NK cells, macrophages, and dendritic cells (13, 14).

In the present study, we evaluated the role of the IL-21 pathway in 2 animal models of autoimmune arthritis. Rat adjuvant-induced arthritis (AIA) is a T cell–dependent inflammatory arthritis (15, 16). The disease can be modulated by depletion of T cells or by the administration of antiinflammatory compounds. Mouse collagen-induced arthritis (CIA) is dependent on both T cells and B cells for pathologic development (17). Our results indicate that IL-21 contributes to the development of both AIA and CIA, and that therapeutic inhibition of IL-21 correlates with modulation of serum cytokine levels and improvements in disease scores. These findings suggest a novel mechanism by which T cells can contribute to pathogenesis in animal models of RA.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Animals.

The Institutional Animal Care and Use Committee approved all of the animal procedures performed. All animals were serologically negative for common pathogens. DBA/1LacJ mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Twelve-week-old male Lewis rats were obtained from Charles River (Wilmington, MA). Animals were housed according to standard procedures and maintained under a standard regimen of food and water ad libitum.

Rat IL-21 complementary DNA (cDNA).

IL-21 cDNA was amplified using KOD Hot Start (Novagen, Madison, WI) using PW485 (GTACAAAAAAGCAGGCTCCACCATGGAGAGGACCCTTGTCTGTC) and PW486 (ACTTTGTACAAGAAAGCTGGGTCTAGGAAAGATGCTGATGAATC). A gel-purified fragment was reamplified using PW469 (GGGGACAAGTTTGTACAAAAAAGCAGGCTCCACCATG) and PW482 (GGGGACCACTTTGTACAAGAAAGCTGGGT). The rat IL-21 coding region matched that of the rat Genomic Assembly (GenBank 26007405_16). The predicted amino acid sequence is MERTLVCLILIFLGTVAHKSSPQRPDHLLIRLRHLMDIVEQLKIYENDLDPELLTAPQDVKGQCEHEAFACFQKAKLKPSNTGNNKTFINDLLAQLRRRLPAKRTGNKQRHAMAKCPSCDLYEKKTPKEFLERLKWLLQKMIHQHLS.

Transfection of IL-21R Fc fusion protein (IL-21R.Fc).

Murine IL-21R.Fc was constructed by polymerase chain reaction (PCR), in which amino acids 1–235 of mouse IL-21R were linked, with a GSGS linker, to mouse IgG2a-Fc domains containing mutations, which allowed us to minimize Fc binding and complement fixation (18). This was followed by subcloning into a DHFR expression vector and transfection into CHO cells (19, 20). Peak fractions from a protein A affinity column were pooled and buffer exchanged into phosphate buffered saline. The protein was >95% pure as determined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

BAF-3 proliferation assays.

BAF-3 cells (2 × 104) expressing amino-terminal FLAG-tagged IL-21R were incubated for 3 days with murine IL-21 (594-ML-010 [R&D Systems, Minneapolis, MN] or recombinant murine IL-21) and murine IL-21R.Fc in 96-well plates, and pulsed with 0.5 μCi 3H-thymidine for 5 hours. The cells were then harvested and counted.

Analysis of cytokines in response to collagen.

Ten-week-old male DBA/1 mice were immunized with 100 μg/ml bovine type II collagen (Chondrex, Seattle, WA) in Freund's complete adjuvant (CFA), and draining lymph nodes were harvested on day 10. Lymph node cells (5 × 106 cells/ml) were cultured with 100 μg/ml heat-denatured bovine type II collagen and either 100 μg/ml isotype control IgG2a antibody or murine IL-21R.Fc. Spleens and draining lymph nodes were also harvested at the end of the studies and stimulated in culture with type II collagen. Cytokines were measured by multiplex array analysis (Pierce Biotechnology, Rockford, IL) after 72 hours.

Cytokine messenger RNA (mRNA) analysis of mouse paws.

RNA was isolated from individual homogenized mouse paws (n = 5) using RNeasy Mini kits (Qiagen, Chatsworth, CA) following the manufacturer's instructions. Purified RNA was treated with DNase (Ambion, Austin, TX) and adjusted to a concentration of 20 ng/μl before mRNA analysis by quantitative PCR on an ABI PRISM 7300HT sequence detection system (Applied Biosystems, Foster City, CA) using primer pairs and probes specific for GAPDH, IL-6, tumor necrosis factor α (TNFα), interferon-γ (IFNγ), and IL-21R, at optimized primer and probe concentrations. Expression of mRNA in individual paws was normalized to the expression of GAPDH in each sample, with results expressed as the relative units of target mRNA.

Rat T cell proliferation.

Splenic T cells from male Sprague-Dawley rats were purified to 91% CD3+ using rat T cell–enrichment columns (RTCC-5; R&D Systems) and incubated on plates coated with 1 μg/ml anti–rat CD3 antibody (554829; BD PharMingen, San Diego, CA) in 0.1% conditioned medium from mock-transfected or rat IL-21 cDNA–transfected COS cells. Murine IL-21R.Fc or isotype control was titered into the cultures. On day 3, 0.5 μCi of 3H-thymidine was added for 5 hours, and cells were harvested and counted.

Rat spleen cell cytokines.

Spleen cells (8 × 106) from Sprague-Dawley rats were grown in Dulbecco's modified Eagle's medium plus 10% fetal calf serum in 9.1-cm2 wells for 3 days, and then incubated in 4 μg/ml concanavalin A (Con A) (C-2010; Sigma, St. Louis, MO) with either 10 μg/ml or 1 μg/ml murine IL-21R.Fc or 10 μg/ml control Ig. The levels of IFNγ were measured by enzyme-linked immunosorbent assay (ELISA) (R1F00; R&D Systems).

Induction of AIA in rats.

Immunization with CFA (lot 084H8800; Sigma-Aldrich, Bornem, Belgium) was used to induce AIA in rats (21, 22). The study treatment was then initiated on day 9 after induction of arthritis, when the joints were maximally inflamed. Rats were administered a mouse IgG2a control (anti–Elmeria tenella, HB 8389 2.03.7; American Type Culture Collection, Rockville, MD), murine IL-21R.Fc, murine TNF receptor type II Fc fusion protein (TNFRII.Fc), or saline on alternate days. Arthritis severity was scored daily (21, 22) for 9 days, after which the rats were killed and the hind limbs fixed in 10% buffered formalin. For histopathologic analysis, the tarsal joints were decalcified, embedded in paraffin, and stained with hematoxylin and eosin or Saffranin O–fast green. Each stained section was evaluated for synovial abnormalities and articular cartilage abnormalities and scored as previously described (21–24).

Induction of CIA in mice.

Bovine type II collagen (Chondrex) was dissolved in 0.01N acetic acid and emulsified in an equal volume of CFA containing 1 mg/ml of heat-killed Mycobacterium tuberculosis (Sigma). Arthritis was induced in 10-week-old male DBA/1LacJ mice (25). Semitherapeutic treatment was started in the mice after priming and boosting with type II collagen. The treatment, either murine IL-21R.Fc, murine TNFRII.Fc, or control mouse IgG2a (anti–E tenella, HB 8389 2.03.7; American Type Culture Collection), was initiated when 10% of the mice exhibited clinical signs of disease. Experiments contained 14 mice per group and were performed 3 times.

Arthritis severity was scored visually 3 times per week, beginning 3 weeks after the primary type II collagen immunization (25). Paws were processed for histologic assessment and stained with hematoxylin and eosin (Sigma). Each stained section was evaluated for synovial hyperplasia, inflammatory cell infiltrates, cartilage damage, pannus formation, bone erosion, fibrillation, and ankylosis. Histopathologic scores were assigned using previously described methods (25). The arthritis severity score for each paw was weighted based on the number of joints per paw receiving a specific score.

IgG antibody levels against the immunogen were measured by ELISA (25). The anti–type II collagen concentrations (in units/ml) were determined by reference to a standard curve generated from 1:2 serial dilutions of a standard CIA serum.

Statistical analysis.

All histologic features between groups were analyzed using SuperANOVA (Abacus Concepts, Berkeley, CA) with Duncan's new multiple-range post hoc testing. Disease severity, histopathologic scores, and serum anti–type II collagen IgG levels were compared between groups, using Student's t-test. Disease incidence was analyzed with Fisher's exact test. Results are expressed as the mean ± SD. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Neutralization of IL-21 bioactivity and modulation of IFNγ production by IL-21R.Fc.

Mouse CIA is a model of RA that is mediated by both cellular and humoral immune responses (17). Immunization of susceptible DBA/1 mice with bovine type II collagen results in an RA-like disease of the joint, characterized by inflammatory cell infiltrates, swelling, and, ultimately, cartilage and bone damage. Th1 cells have been shown to play a role in driving the initial immune response, which is followed by pathogenic anti–type II collagen antibody production. Cytokines, such as TNFα and IL-6, also play a role in CIA, underscoring the shared mechanisms with human RA (26, 27).

To examine the effects of blocking IL-21–IL-21R interactions on lymphoid cell responses, we used a fusion protein (mouse soluble IL-21R.Fc) containing the extracellular domain of mouse IL-21R fused to the constant region of mouse IgG2a, which bears a mutation that prevents binding to Fc receptors. BAF-3 cells, which contain endogenous γ–common chain, were engineered to express the IL-21R. As shown in Figure 1A, IL-21R–expressing BAF-3 cells exhibited vigorous proliferation in response to increasing doses of IL-21. Treatment with the IL-21R.Fc fusion protein, but not an isotype-matched control antibody, effectively neutralized the IL-21–mediated proliferation of IL-21R–expressing BAF-3 cells (Figure 1B).

thumbnail image

Figure 1. Inhibitory actions of mouse interleukin-21 (IL-21) receptor Fc fusion protein (mIL-21R.Fc) on IL-21–dependent proliferation and on interferon-γ (IFNγ) production. A, Proliferation of BAF-3 cells transfected with mIL-21R cDNA, in response to increasing doses of mIL-21. B, Neutralization of mouse IL-21 bioactivity by mIL-21R.Fc. IL-21R–expressing BAF-3 cells were cultured with increasing levels of soluble IL-21R.Fc or control antibody in the presence of 80 pM mouse IL-21. The calculated 50% inhibitory concentration of mIL-21R.Fc is 2.9 nM. C, Enhanced production of IFNγ from collagen-primed lymph node cells by blockade with mIL-21R.Fc. The cells were isolated from DBA/1 mice primed for collagen-induced arthritis on day 10 and restimulated in vitro with collagen in the presence of mIL-21R.Fc or control Ig, followed by measurement of IFNγ levels in the conditioned media. D, Neutralization of proliferation of rat T cells in response to rat IL-21 and anti-CD3 by mIL-21R.Fc. An IgG2a antibody (anti–Elmeria tenella) was used as control. Results were compared with the proliferative responses to anti-CD3 plus 0.1% conditioned media from COS cells transfected with rat IL-21 (rIL21 0.1% CM const). E, Enhancement of production of IFNγ from concanavalin A (Con A)–stimulated rat splenocytes by mIL-21R.Fc. Results are the mean ± SD.

Download figure to PowerPoint

In experiments utilizing DBA/1 mice immunized with bovine type II collagen in CFA to induce arthritis, the draining lymph nodes were harvested on day 10. Lymph node cells were restimulated in vitro with type II collagen in the presence of IL-21R.Fc or isotype control antibody. No changes in the proliferative responses to type II collagen were observed (results not shown).

IL-21 has been shown to specifically modulate the production of IFNγ from naive T cells (28). Since IFNγ modulates the severity of disease in CIA, conditioned media from these cultures were assayed for production of IFNγ. Specific enhancement of IFNγ production was observed in cultures in which IL-21R.Fc was added during antigen restimulation (Figure 1C). These results indicate that IL-21 can modulate IFNγ production by T cells primed with collagen in vivo.

To study the role of IL-21 in rat AIA, cDNA was isolated for assessment of rat IL-21 and used to express recombinant rat IL-21 protein in COS cells. Proliferation of rat splenic T cells in response to plate-bound anti-CD3 was enhanced by addition of rat IL-21, in that the addition of 0.1% conditioned medium from transfected 293 cells augmented the anti-CD3–induced baseline proliferation from 2,250 counts per minute (mock transfection) to 150,000 cpm. Thus, rat IL-21 is a potent costimulus for the proliferation of unprimed rat T cells, which are characterized by suboptimal stimulation through the T cell receptor (TCR). Addition of the mouse IL-21R.Fc fusion protein to this assay effectively neutralized the enhanced proliferation of rat T cells occurring in response to rat IL-21 (Figure 1D).

We also examined the effect of IL-21R.Fc on rat T cells stimulated with Con A. As observed with primed mouse T cells, incubation of Con A–stimulated rat T cells with IL-21R.Fc resulted in no alteration in proliferation (results not shown). However, measurement of IFNγ production in these cultures revealed that treatment with IL-21R.Fc resulted in enhanced IFNγ production (Figure 1E).

Reduction in the severity of CIA following IL-21R.Fc administration.

We examined the effects of IL-21R.Fc in a semitherapeutic model of CIA in which animals were treated after priming and boosting with type II collagen so that both cellular and humoral immune responses were ongoing. Our rationale was to model an early stage of RA in which the disease process had already been initiated. DBA/1 mice were immunized with type II collagen in CFA. Twenty-one days later, the mice were boosted with type II collagen in Freund's incomplete adjuvant. IL-21R.Fc was administered to the experimental cohort of mice when 10% of the mice began to exhibit clinical signs of disease.

Administration of IL-21R.Fc resulted in reduced disease severity scores (Figures 2A and B). Moreover, we observed a greater reduction in disease severity with the higher dose of IL-21R.Fc (400 μg), such that the animals receiving the higher dose showed significantly less disease on several days compared with the 400 μg control–treated animals. The lower dose of IL-21R.Fc (200 μg) produced a significant difference in disease severity scores only at the last time point of the study, as compared with the respective control group. As a positive control, mouse soluble TNFRII.Fc was administered, and the results indicated a profound amelioration of disease, which is consistent with the role of TNFα as an effector-phase cytokine (Figure 2A).

thumbnail image

Figure 2. Reductions in disease severity scores in mouse paws with collagen-induced arthritis via blockade of the interleukin-21 (IL-21) pathway. DBA/1 mice were immunized with type II collagen/Freund's complete adjuvant and boosted on day 21 with type II collagen/Freund's incomplete adjuvant. Murine IL-21 receptor Fc fusion protein (IL-21R.Fc), murine soluble tumor necrosis factor receptor type II Fc fusion protein (m.s.TNFRII.Fc), or Ig control was administered at 200 μg/mouse or 400 μg/mouse 3 times per week, beginning on day 0, which was the time at which 10% of the mice exhibited disease signs. Mice were dosed for 4 weeks. Results are the mean ± SEM arthritis severity scores, representative of 3 independent experiments, for A, the 200 μg treatment group and B, the 400 μg treatment group (n = 14 in each). ∗ = P ≤ 0.05 versus the murine isotype control, by 2-tailed t-test. IP = intraperitoneal.

Download figure to PowerPoint

Histologic scoring of the paws at the end of the experiment reflected the clinical findings, in that the histologic severity score was reduced significantly after administration of soluble mouse IL-21R.Fc, although the analysis did not reveal a dose-response effect (Table 1). A treatment regimen with mouse soluble TNFRII.Fc also resulted in a significant reduction in histologic severity scores (Table 1).

Table 1. Histologic assessment of paws and serum interleukin-6 (IL-6) levels after treatment in mice with collagen-induced arthritis*
Treatment groupHistologic severity score, mean ± SEMPaws with grade 0 arthritis, %Serum IL-6, mean ± SEM pg/ml
  • *

    Front and rear paws were harvested from mice at the end of the experiment for histologic evaluation. The histologic severity score, ranging from 0 to 4, was assessed for each paw as detailed in Materials and Methods. TNFRII = tumor necrosis factor receptor type II; IL-21R = interleukin-21 receptor.

  • For each treatment group, 56 paws were scored, except for the 200 μg mouse soluble IL-21R.Fc group, in which 52 paws were scored.

  • P < 0.05 versus the same dose of Ig control.

Untreated3.25 ± 0.8417.9
Ig control   
 200 μg3.29 ± 1.0916.1917.3 ± 132.2
 400 μg3.46 ± 1.0112.5846.1 ± 113.9
Mouse soluble TNFRII.Fc (200 μg)1.98 ± 1.4050.0204.8 ± 21.7
Mouse soluble IL-21R.Fc   
 200 μg2.18 ± 1.2844.6182.1 ± 34.1
 400 μg2.89 ± 0.9226.8311.8 ± 55.1

At the end of each experiment, we examined serum cytokine levels in mice primed for CIA and subsequently treated with soluble mouse IL-21R.Fc or control Ig in the semitherapeutic regimen. We observed statistically significant decreases in IL-6 levels (Table 1). In addition, we observed increased serum levels of IFNγ in some of the mice primed for CIA and treated with IL-21R.Fc, consistent with our in vitro results, although these changes did not reach statistical significance (results not shown).

The levels of mRNA for several inflammatory cytokines were determined by quantitative PCR of homogenized paws harvested at the end of the study. Treatment with IL-21R.Fc resulted in significantly lower levels of IL-6 mRNA when compared with the levels in control-treated animals. Expression of mRNA for IL-21R and TNFα also appeared lower, but the changes were not statistically significant. In contrast, levels of IFNγ mRNA in the paws were significantly increased in response to IL-21R.Fc treatment (Figure 3A). We were unable to detect mRNA for IL-21 itself (results not shown).

thumbnail image

Figure 3. Alterations in cytokine mRNA and collagen-specific antibody levels in mouse paws with collagen-induced arthritis, and spleen cell cytokine levels at the end of the study, after treatment with murine interleukin-21 (IL-21) receptor Fc fusion protein (mIL-21R.Fc). A, Messenger RNA was isolated from homogenized paws of mice treated with 400 μg of control Elmeria tenella or mIL-21R.Fc or 200 μg of murine soluble tumor necrosis factor α receptor type II Fc fusion protein (msTNFRII.Fc). Levels of each indicated cytokine were determined by quantitative mRNA analysis, with results expressed as the mean ± SD of 5 mice per group. B, Collagen-specific IgG1and IgG2a levels were determined in individual mice from both the 400 μg E tenella control and IL-21R.Fc groups, by enzyme-linked immunosorbent assay (ELISA). Total serum IgG1 levels were also measured. Bars show the mean. C, Collagen-induced cytokines from spleen cells were determined by harvesting spleens at the end of the study from the 400 μg E tenella control, IL-21R.Fc, and 200 μg msTNFRII.Fc groups, which were then cultured with collagen and assessed by multiplex ELISA. Results are the mean ± SD of 5 mice per group. ∗ = P < 0.05 versus 400 μg isotype control antibody, by 2-tailed t-test. RU = relative units; IFNg = interferon-γ; GMCSF = granulocyte–macrophage colony-stimulating factor.

Download figure to PowerPoint

Development of CIA is dependent on the generation of antibodies to bovine type II collagen that are cross-reactive to mouse type II collagen, and IL-21 is important in the regulation of Ig production, specifically IgG1 and IgE responses (5, 6). We therefore examined the levels of serum IgG1 and IG2a antibodies to type II collagen and the total IgG1 levels in the mice after treatment with IL-21R.Fc (Figure 3B). Levels of anti–collagen IgG1 and anti–collagen IgG2a were slightly lower, but not significantly different, in the IL-21R.Fc–treated mice when compared with the control-treated mice. Similarly, IgG2b levels were not significantly different between control and treated mice (results not shown). However, total IgG1 levels were significantly lower in the IL-21R.Fc–treated group. We could not detect antigen-specific IgE in the serum of these mice (results not shown).

To determine the cytokine phenotype of lymphoid populations in response to treatment, we cultured draining lymph node cells or splenocytes with collagen, and examined the levels of several cytokines. No significant differences were observed in collagen-restimulated lymph node cell cultures in response to treatment (results not shown). In splenocyte cultures, however, we observed higher levels of IFNγ, IL-2, and granulocyte–macrophage colony-stimulating factor and lower levels of IL-6 and IL-17 in IL-21R.Fc–treated mice compared with control mice. Mice treated with murine soluble TNFRII.Fc secreted significantly lower levels of IL-6 compared with the control group (Figure 3C).

Reversal of rat AIA by blockade of the IL-21 pathway.

Rat AIA is a disease initiated by polyclonal T cell responses to heat-killed M tuberculosis in CFA in male Lewis rats (15, 16). The disease is characterized by a multiorgan inflammatory response, which includes an arthritic response in the paws that occurs 8–14 days after immunization. Inflammatory infiltrates are found in the spleen, liver, bone marrow, meninges, skin, and eyes. The disease resolves after several months. Examination of the joints during maximal disease reveals infiltration of neutrophils, monocytes, and lymphocytes. In addition, synovial pannus formation, proteoglycan depletion, and cartilage erosion are observed.

We evaluated the effect of administration of mouse soluble IL-21R.Fc on rat AIA. Arthritis was induced in male Lewis rats, and clinical signs of arthritis were allowed to develop. The tarsal joints were fully swollen by day 8. Starting on day 9, the rats were given IL-21R.Fc or IgG control 3 times per week. In the first experiment (Figure 4A), doses of 1 mg/kg and 3 mg/kg of IL-21R.Fc resulted in partial reduction of swollen joint scores over time, whereas no reduction in swollen joint scores was observed in the controls. The response was dose dependent, suggesting that additional reagent might be needed to fully abrogate disease.

thumbnail image

Figure 4. Effects of alternate-day treatment with vehicle, IgG, murine interleukin-21 receptor Fc fusion protein (IL-21R.Fc) or murine tumor necrosis factor receptor type II Fc fusion protein (muTNFR.Fc) on the clinical signs of joint inflammation in adjuvant-induced arthritis, beginning 9 days after adjuvant injection of rats. Results are the mean ± SD joint scores. A, In a first experiment, the dose-response effects of all 3 proteins were compared with those of saline vehicle injection. Clinical signs of joint inflammation were decreased in a dose-dependent manner by mouse IL-21R.Fc or muTNFRII.Fc, while IgG produced no discernible effect (n = 4 per group). B, In a second experiment, a higher dose of mouse IL-21R.Fc was compared with IgG or muTNFR.Fc. The effects of mouse IL-21R.Fc at this dose were similar to those of muTNFR.Fc (n = 6 per group).

Download figure to PowerPoint

In a second experiment (Figure 4B), the dose of IL-21R.Fc was increased to 6 mg/kg, and full amelioration of clinical signs was observed. After completion of the experiment, the rats were killed and the tarsal joints were obtained for histologic analysis. As shown in Table 2, features of synovitis and Mankin scores of articular cartilage damage were significantly reduced with IL-21R.Fc. These data indicate that IL-21 contributes to the pathologic features of AIA, even after the disease is fully established. Inhibition of IL-21 in this model resulted in a reduction in inflammatory cell infiltrates and a decrease in cartilage erosion in the joints.

Table 2. Histologic scores for rat AIA*
Histologic feature (score range)PBS vehicle, 1 ml/kgIgG, 6.0 mg/kgMurine TNFR.Fc, 3.0 mg/kgIL-21R.Fc, 6.0 mg/kg
  • *

    Values are the mean ± SD histologic scores for synovitis and Mankin scores for cartilage changes in tarsal joints from animals with adjuvant-induced arthritis (AIA) treated as indicated, beginning on day 8 after injection of Freund's complete adjuvant. TNFR = tumor necrosis factor receptor; IL-21R = interleukin-21 receptor.

  • P < 0.05 versus phosphate buffered saline (PBS) vehicle and IgG groups.

  • P < 0.05 versus PBS vehicle group.

  • §

    P < 0.05 versus all other groups.

Synovial structure (0–3)2.92 ± 0.212.83 ± 0.261.08 ± 0.491.33 ± 0.68
Fibroplasia (0–3)2.67 ± 0.412.50 ± 0.450.58 ± 0.490.67 ± 0.61
Inflammatory cells (0–3)2.92 ± 0.212.92 ± 0.210.67 ± 0.611.00 ± 0.63
Pannus (0–2)1.83 ± 0.412.00 ± 00 ± 00 ± 0
Total synovitis score (0–11)10.33 ± 0.7510.25 ± 0.522.33 ± 1.513.00 ± 1.73
Cartilage structure (0–6)3.42 ± 0.383.08 ± 0.201.08 ± 0.380.92 ± 0.38
Cartilage cells (0–6)2.83 ± 0.262.75 ± 0.420.33 ± 0.410.67 ± 0.52
Safranin O–fast green staining (0–4)3.00 ± 0.002.67 ± 0.260.92 ± 0.38§1.58 ± 0.20
Tidemark integrity (0–1)0000
Total Mankin score (0–14)9.25 ± 0.528.50 ± 0.632.33 ± 0.983.00 ± 0.63

The efficacy of IL-21/IL-21R inhibition by IL-21R.Fc was compared in this model with inhibition of the TNF pathway by mouse soluble TNFRII.Fc. Blockade of either cytokine pathway resulted in a dose-dependent amelioration of the clinical signs of arthritis, indicating that both cytokines play a key role in this disease. Similar histologic scores were obtained after treatment in each case (Table 2), indicating that blockade of either cytokine pathway is sufficient to reverse the clinical and histologic disease features in AIA.

Representative photomicrographs of the paws of rats with AIA treated with vehicle, mouse soluble TNFRII.Fc, or IL-21R.Fc in this experiment are shown in Figure 5. When compared with IgG-treated specimens, both IL-21R.Fc and mouse soluble TNFRII.Fc reduced all of the parameters of synovial inflammation and all of the components of the Mankin score for articular cartilage lesions. In the vehicle-treated animals, the articular cartilage was damaged and in some cases was completely destroyed. In these areas, hyperplastic synoviocytes and an evolving pannus reaction invaded the articular cartilage region. Extensive proliferation of the synovial membrane and pannus resulted in villus formation that filled the joint space.

thumbnail image

Figure 5. Photomicrographs of the tarsal joints from rats in the second adjuvant-induced arthritis experiment. A, Tarsal joint from a vehicle-treated rat, showing loss of articular cartilage (arrow), hyperplasia of the synovial membrane (H), and organization of the aerolar connective tissue to produce pannus (P). B, Tarsal joint from a murine tumor necrosis factor receptor type II Fc fusion protein–treated rat, showing normal articular cartilage (C), hyperplasia of the synovial membrane, minimal villus formation, and normal aerolar connective tissue (AC). C, Tarsal joint from an interleukin-21 receptor Fc fusion protein–treated rat, showing slight articular cartilage damage from overlying, hyperplastic synovial membrane and minimal organization of the aerolar connective tissue (O). (Hematoxylin and eosin stained; original magnification × 200.)

Download figure to PowerPoint

In animals treated with either mouse soluble IL-21R.Fc or mouse soluble TNFRII.Fc, significant reductions in articular cartilage damage were found, and the development of pannus was greatly attenuated. Moderate synoviocyte proliferation still occurred, but minimal villus formation was seen. The beneficial effects of mouse soluble IL-21R.Fc and mouse soluble TNFRII.Fc on reducing the synovial and articular cartilage lesions were similar in the AIA model.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

We have explored the role of the T cell–derived cytokine IL-21 in pathology development in animal models of RA. In these experiments, we found that IL-21 contributes to disease in both CIA and AIA, each of which is a distinct animal model of RA. Administration of IL-21R.Fc, a potent neutralizing reagent for IL-21, resulted in reversal of clinical disease in a rat model of AIA. Administration of IL-21R.Fc to mice with CIA after the onset of disease symptoms also resulted in amelioration of disease progression. Blockade of IL-21 was correlated with enhanced IFNγ production and suppression of serum IL-6 levels. Thus, the IL-21–IL-21R interaction defines a new pathway contributing to the inflammatory arthritis response in these models.

Both the rat AIA model and the mouse CIA model are driven by pathogenic T cell responses. Pathogenicity has been associated with the production of inflammatory cytokines, including TNFα, IL-6, and IL-17 (29–31). Herein we identified IL-21 as a new T cell cytokine that is contributing to the pathologic processes of both CIA and AIA. In both models, inhibition of the cytokine using a soluble mouse IL-21R.Fc fusion protein resulted in a reduction of the clinical signs of disease in animals in which the disease process had already been initiated. This implicates IL-21 as having a role in the effector phase of both diseases.

IL-21 can enhance the proliferative response of naive mouse T cells to TCR stimulation (5). Similarly, in the present study, the addition of IL-21 to rat T cells stimulated with anti-CD3 resulted in enhanced proliferation. This suggests that IL-21 may contribute to the initial expansion of pathogenic T cell populations. However, inhibition of IL-21 using soluble IL-21R.Fc had no effect on the proliferative responses of previously primed mouse T cells or Con A–activated rat T cells, suggesting that IL-21 is not the predominant cytokine driving the expansion of primed T cells.

The complete reversal of disease and histologic changes observed in the AIA model of arthritis suggest that IL-21 is necessary for the prolonged accumulation of inflammatory cells in the joints. AIA in the Lewis rat is mediated by migration of activated T cells and neutrophils into the joints of immunized animals (15, 16). Activated macrophages are also thought to contribute to the disease. The role of the IL-21 pathway in macrophage-mediated events has not been well described. In mice, IL-21 can contribute to the expansion of the macrophage lineage (32). In RA patients, expression of IL-21R has been described on synovial macrophages and fibroblasts, suggesting that these cell types may potentially participate in IL-21–mediated effects (33). In addition, IL-21, which is made by activated CD4+ T cells, may be contributing to the expansion, activation, or recruitment of inflammatory cells, either within the peripheral compartment or directly in the joints.

B cells are required for disease in CIA, since mice deficient in B cells are resistant to CIA induction (34). Production of type II collagen–reactive antibodies is necessary for disease induction (35), and arthritis can be induced by transfer of mixtures of type II collagen–reactive antibodies (36). However, production of antibodies to type II collagen is not sufficient to induce disease, since disease can be suppressed in the presence of anti–type II collagen antibody responses (25). IL-21 is required for the induction of IgG responses, as shown by the reduced IgG responses of IL-21R–deficient mice (6). We did not detect a significant difference in anti–type II collagen antibody titers in those experiments in which IL-21R.Fc was administered after type II collagen boosting, suggesting that anti–type II collagen antibody production had already proceeded past the IL-21–dependent step.

Pathogenic antibodies in CIA are associated with the IgG2a and IgG2b isotypes, which are less profoundly affected than IgG1 in IL-21R–deficient mice (6). It may be that IgG2a responses in our experiments were being driven by IL-21–independent mechanisms. However, we also did not see any significant changes in the levels of IgG1 or IgG2b anti–type II collagen antibody titers. Thus, the mechanism of disease amelioration in these CIA studies is independent of modulation of anti–type II collagen antibody responses and may share a common pathway with that observed in AIA. The lack of complete disease amelioration in CIA may reflect the contribution of the anti–type II collagen antibodies to disease, which are still present in these treated animals.

In both CIA and AIA, the initiating pathogenic T cell response is predominantly a Th1 response. However, the role of INFγ, the canonical Th1 cell cytokine, does not correlate with disease susceptibility and appears to have a protective role in CIA (37–39). Recent studies examining the distinct roles of IL-12 and IL-23 have raised questions about the role of Th1 cells in CIA (40). Interestingly, IL-21 has been shown to inhibit the differentiation of naive T cells into Th1 cells that produce IFNγ (28). The production of IFNγ from reactivated type II collagen–primed mouse lymph node T cells or Con A–activated rat T cells was enhanced in the presence of IL-21R.Fc, and mRNA for IFNγ was also reduced in the paws of mice isolated after treatment with IL-21R.Fc. These results suggest that IL-21R.Fc may be affecting disease by modulation of IFNγ. Mice deficient in IL-21R have enhanced IFNγ responses after priming for a delayed-type hypersensitivity response (28). Experiments in IFNγ-deficient mice might help to elucidate whether the protective effect of IL-21R.Fc is mediated through IFNγ.

Serum IL-6 levels were decreased consistently in multiple experiments with IL-21 or TNFα blockade. Expression of IL-6 and TNFα mRNA in TaqMan analyses of the diseased paws was also reduced in IL-21R.Fc–treated mice. A decrease in serum IL-6 levels has also been observed in therapeutic CIA studies in which IL-17 was inhibited (29). IL-21 has been reported to promote production of IL-17 from activated human T cells (41).Thus, blockade of IL-21 in CIA may function, in part, by modulation of IL-6–driven inflammation, either directly or through down-regulation of TNFα or IL-17. IL-6 has a pathogenic role in CIA (26, 31, 42) and is a critical inflammatory cytokine in RA. Blockade of IL-6R has demonstrated efficacy in clinical trials of RA (43).

Our data show a role for IL-21 in 2 models of RA in which therapeutic intervention was started after onset of the disease process. These data identify IL-21 as a new T cell cytokine contributing to the pathogenesis of animal models of RA. The regulation of IFNγ by IL-21 further suggests that IL-21 may have a role in determining the type of Th cell response generated in these autoimmune models. Blockade of IL-21 resulted in down-regulation of IL-6, a cytokine that is implicated in the pathologic processes of both CIA and RA, demonstrating that IL-21 contributes to the production of IL-6 in animal models of arthritis. These data support the need for further evaluation of the role of IL-21 in human RA.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

Dr. Collins had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Young, Hegen, Ma, Albert, Senices, Wu, Nickerson-Nutter, Keith.

Acquisition of data. Hegen, Ma, Whitters, Albert, Lowe, Senices, Wu, Sibley, Leathurby, Brown, Keith.

Analysis and interpretation of data. Young, Hegen, Ma, Whitters, Albert, Lowe, Senices, Wu, Brown, Nickerson-Nutter, Keith, Collins.

Manuscript preparation. Young, Hegen, Ma, Albert, Lowe, Senices, Brown, Keith, Collins.

Statistical analysis. Ma, Senices, Keith.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

This study was designed and carried out by staff members of Wyeth Research; these members are all authors of the manuscript. All authors independently reviewed and contributed to the manuscript during its development, agreed to submit the manuscript, and approved the content of the submitted manuscript. The study was supported by Wyeth Research, and decisions on study design, data collection, data analysis, and interpretation of the data were made by Wyeth Research scientists.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES

We thank Qiuna Bi for assistance with the anti-collagen ELISAs, and Neil Wolfman, Beatriz Carreno, and Kyri Dunussi-Joannopoulos for helpful discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. Acknowledgements
  9. REFERENCES