Galectin 3 aggravates joint inflammation and destruction in antigen-induced arthritis

Authors


Abstract

Objective

Galectin 3, an endogenous β-galactoside–binding lectin, plays an important role in the modulation of immune responses. The finding that galectin 3 is present in the inflamed synovium in patients with rheumatoid arthritis suggests that the protein is associated with the pathogenesis of this disease. We undertook this study to investigate the influence of galectin 3 deficiency in a murine model of arthritis.

Methods

Wild-type (WT) and galectin 3–deficient (galectin 3−/−) mice were subjected to antigen-induced arthritis (AIA) through immunization with methylated bovine serum albumin. The concentration of serum cytokines (interleukin-6 [IL-6] and tumor necrosis factor α [TNFα]) and antigen-specific antibodies was evaluated using a cytometric bead array platform and enzyme-linked immunosorbent assay (ELISA). Cellular IL-17 responses were examined by flow cytometry, ELISA, and enzyme-linked immunospot assay.

Results

The joint inflammation and bone erosion of AIA were markedly suppressed in galectin 3−/− mice as compared with WT mice. The reduced arthritis in galectin 3−/− mice was accompanied by decreased levels of antigen-specific IgG and proinflammatory cytokines. The frequency of IL-17–producing cells in the spleen was reduced in galectin 3−/− mice as compared with WT mice. Exogenously added recombinant galectin 3 could partially restore the reduced arthritis and cytokines in galectin 3−/− mice.

Conclusion

Our findings show that galectin 3 plays a pathogenic role in the development and progression of AIA and that the disease severity is accompanied by alterations of antigen-specific IgG levels, systemic levels of TNFα and IL-6, and frequency of IL-17–producing T cells. To our knowledge, this is the first report of in vivo evidence that galectin 3 plays a crucial role in the development of arthritis.

Rheumatoid arthritis (RA) is an autoimmune disease characterized by a massive infiltration of mononuclear and polymorphonuclear cells into the joints and the development of inflammation that results in destruction of articular cartilage and adjacent bone (1). The mechanisms that give rise to RA are only partly understood. Activated T helper cells comprise a large proportion of the inflammatory cells that invade the synovial tissue and may therefore be a cell type of pathogenic importance (2, 3). A number of inflammatory mediators have also been implicated in the establishment and progression of inflammatory joint destruction, including proinflammatory cytokines such as tumor necrosis factor α (TNFα), interleukin-6 (IL-6), and IL-17 and proteins from a novel superfamily of animal lectins, the galectins (4, 5).

The galectins are evolutionarily conserved, β-galactoside–binding lectins that have received much attention in the fields of cancer and autoimmune disease due to their modulating activities on both pro- and antiinflammatory responses (6, 7). The galectins share a high degree of amino acid sequence similarity in the C-terminal, carbohydrate-recognition domain, are synthesized as cytosolic proteins in a number of different cell types (e.g., myeloid cells, fibroblasts, and endothelial and epithelial cells [8]), and are externalized by a nonclassic secretion pathway (6). The galectins are thought to crosslink β-galactoside–containing cell surface glycoconjugates, a process that has been correlated with altered cellular adhesion, differentiation, and survival (9, 10). The involvement of galectins in RA has been demonstrated for galectin 1 and galectin 9 in mice, both of which suppressed the development of collagen-induced arthritis (CIA) (5, 11).

Galectin 3 has been suggested to be a proinflammatory mediator since this lectin induces the production of reactive oxygen species (ROS) in human neutrophils and promotes chemotaxis in monocytes (12–14). The interaction of galectin 3 with T cells induces antiapoptotic activity, a phenomenon often correlated with a prolonged inflammatory response (15). Recently, it was shown that galectin 3 selectively promoted the secretion of mononuclear cell–specific chemokines from synovial, but not skin, fibroblasts in RA, suggesting that galectin 3 can amplify joint inflammation (16). A proinflammatory role for galectin 3 in RA gains further support from the observation that the level of galectin 3 is elevated in synovial fluid from RA patients compared with that from patients with osteoarthritis (4). Nevertheless, the mechanism(s) by which galectin 3 modulates chronic inflammation in RA has yet to be elucidated.

In the present study, a murine, antigen-induced arthritis (AIA) model resembling chronic arthritis was used to examine the influence of galectin 3 on the development and progression of disease. AIA was induced by methylated bovine serum albumin (mBSA) in wild-type (WT) and galectin 3−/− mice. Joint inflammation and destruction were assessed histologically. Galectin 3−/− mice displayed significantly reduced synovitis and joint erosion compared with WT mice, an effect that was partially restored by the addition of exogenous galectin 3 to galectin 3−/− mice. Reduced arthritis in galectin 3−/− mice was accompanied by a significant decrease in circulating mBSA-specific antibodies, the proinflammatory cytokines TNFα and IL-6, and IL-17–producing T cells. Taken together, we provide evidence that galectin 3 plays a pathogenic role in development and progression of AIA and that disease severity was accompanied by alterations in the levels of antigen-specific IgG, the systemic levels of TNFα and IL-6, and the frequency of IL-17–producing cells.

MATERIALS AND METHODS

Mice.

Galectin 3−/− mice generated on a 129/Sv background (17) were backcrossed for 4 generations with C57BL/6 mice, resulting in a 93.75% C57BL/6 background (18). Galectin 3 heterozygote mice were bred to obtain littermate WT and galectin 3−/− mice and genotyped as previously described (18). Mice ages 16–20 weeks were maintained in the animal facility at the Department of Rheumatology and Inflammation Research, University of Gothenburg. The experiments were approved by the Ethical Committee for Animal Experimentation, Gothenburg.

Galectin 3 preparation.

Recombinant human galectin 3 was produced in Escherichia coli and stored at 4°C in phosphate buffered saline (PBS), pH 7.2, containing 150 mM lactose until further purification by gel filtration (PD-10; Pharmacia) to remove lactose. Endotoxin was reduced to <10 pg per μg of galectin 3 (as determined by Limulus amebocyte lysate assay) using an AffinityPak Detoxi-Gel column (Pierce). Purified galectin 3 was stored at −80°C in Krebs-Ringer phosphate buffer (1 mM Ca2+, 1.5 mM Mg2+, 10 mM glucose, pH 7.3).

AIA model.

The antigen, mBSA (Sigma-Aldrich), was emulsified in Freund's complete adjuvant (CFA; Sigma-Aldrich). Three groups of 10–11 mice were immunized by subcutaneous injection of 200 μg of mBSA in 200 μl CFA in one flank on day 0 and 100 μg of mBSA in 100 ml CFA in the base of the tail on day 7. On day 21, arthritis was induced in 1 knee joint by intraarticular injection of 30 μg of mBSA in 20 μl PBS. On day 28, the mice were anesthetized, bled, and killed by cervical dislocation. On days 0, 3, 21, and 24, WT and galectin 3−/− mice were given Krebs-Ringer phosphate buffer (vehicle control) or recombinant galectin 3 (1.3 μg per gm body weight) intraperitoneally (IP). Blood was collected on day 0, day 10, and immediately following termination (day 28). Blood was stored at room temperature for a minimum of 4 hours, and then centrifuged at 4,000 revolutions per minute for 10 minutes to collect sera. Sera were transferred to clean tubes and stored at −80°C for further use. The scheme for injection and sampling is shown in Figure 1.

Figure 1.

Scheme for antigen-induced arthritis (AIA) and administration of exogenous galectin 3 (gal-3). Wild-type and galectin 3−/− mice were immunized by subcutaneous (SC) injection of methylated bovine serum albumin (mBSA) emulsified in Freund's complete adjuvant (CFA) on day 0 (200 μg mBSA in 200 μl CFA) and day 7 (100 μg mBSA in 100 μl CFA). AIA was induced on day 21 in 1 knee joint by intraarticular (IA) injection of mBSA. On day 28, mice were killed for joint histologic examination and analysis of cellular as well as humoral responses. Blood was drawn on days 0, 10, and 28 for quantification of cytokines and mBSA-specific antibodies. Galectin 3−/− mice received either vehicle control or recombinant galectin 3 intraperitoneally (IP) on days 0, 3, 21, and 24.

Histologic evaluation of arthritis.

For histologic examination, knee joints that received mBSA were removed, fixed in 4% formaldehyde, decalcified, embedded in paraffin, and stained with hematoxylin and eosin. Tissue sections were coded and examined in a blinded manner for synovitis and joint destruction (histopathologic index). The extent of synovitis was graded on an arbitrary scale from 0 to 3, where 0 = no signs of inflammation; 1 = mild synovial hypertrophy in up to 5 cell layers; 2 = moderate inflammation characterized by hyperplasia of synovial membrane in up to 10 cell layers and influx of inflammatory cells throughout the synovial tissue; and 3 = marked synovial hypertrophy in >10 cell layers, with synovial tissue infiltrated by inflammatory cells. Bone and cartilage erosions were scored from 0 to 3, where 0 = no erosion of cartilage/bone; 1 = mild erosion of cartilage and no erosion of bone; 2 = moderate erosion of cartilage and mild erosion of bone; and 3 = severe destruction of cartilage and moderate/severe erosion of bone.

Cell proliferation.

Spleen and lymph nodes (popliteal, inguinal, axillary) were pressed through 70-μm cell strainers, incubated with ACK lysis buffer (0.15M NH4Cl, 10 mM KHCO3, 0.1 mM Na2EDTA) for 5 minutes to lyse red blood cells, washed, and suspended in Iscove's medium supplemented with 10% fetal calf serum, 4 mML-glutamine, 10 mM HEPES, 1 mM sodium pyruvate, 50 μM 2-mercaptoethanol, and 50 μg/ml gentamicin. Cells were seeded onto duplicate 96-well tissue culture plates at 2 × 106 cells/ml and incubated in the presence and absence of 50 μg/ml mBSA or 1.25 μg/ml concanavalin A (Con A) for 48 hours. From 1 plate, cell-free supernatants were collected after 48 hours and stored at −20°C for cytokine determinations. On the duplicate plate, cells were pulsed with 0.5 μCi 3H-thymidine (Amersham-Buchler) for the last 6 hours of culture. Thereafter, cells were harvested onto 96-well glass fiber filters (Packard Bioscience), and 3H-thymidine incorporation was measured with a scintillation counter (Top-Count; Packard Bioscience). All reagents were from Sigma-Aldrich unless indicated otherwise.

Flow cytometry.

Surface expression of CD3, CD4, and CD25 on splenocytes was analyzed by flow cytometry using allophycocyanin-conjugated anti-mouse CD3, peridinin chlorophyll A protein–conjugated anti-mouse CD4, and fluorescein isothiocyanate–conjugated anti-mouse CD25 antibodies (all from BD PharMingen). Intracellular staining of FoxP3 was performed using phycoerythrin-conjugated anti-mouse/rat FoxP3 (eBioscience) according to the manufacturer's instructions. Treg cells were defined as cells positive for CD4, CD25, and FoxP3. Flow cytometry was performed using a FACSCanto II (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).

Enzyme-linked immunospot (ELISpot) assay.

The frequency of IL-17–producing cells in the spleen was analyzed with a mouse IL-17 ELISpot kit (eBioscience) according to the manufacturer's instructions. Splenocytes were stimulated for 42 hours at 37°C and 5% CO2 with 50 μg/ml of mBSA or 50 ng/ml of phorbol myristate acetate (PMA; Sigma-Aldrich) plus 1 μg/ml of ionomycin (Sigma-Aldrich). In order to relate the number of IL-17 spots to the frequency of CD3+ cells in the spleen, the ELISpot data were combined with the flow cytometry data from the CD3+ spleen cells, and the data are presented as the number of IL-17 spots per 105 CD3+ cells.

Antigen-specific IgG in sera.

Microplates (96-well; Fisher Scientific) were precoated with 0.1 μg/ml mBSA at 4°C overnight. After coating, the plates were washed, blocked with 2% casein, and incubated with serum. Titration experiments were performed in order to find the optimal dilution of the sera for enzyme-linked immunosorbent assay (ELISA) analysis. Bound IgG was detected with a peroxidase-conjugated anti-mouse IgG (ICN Biochemicals) and tetramethylbenzidine (Sigma-Aldrich) substrate. Absorbance was measured at 450 nm on an ELISA plate reader (Tecan).

Determination of proinflammatory cytokines.

Serum cytokines were measured using a BD cytometric bead array mouse inflammation kit (BD Biosciences) according to the manufacturer's instructions. The lower detection limits for IL-6, TNFα, IL-10, monocyte chemotactic protein 1 (MCP-1), interferon-γ (IFNγ), and IL-12p70 were 5 pg/ml, 7.3 pg/ml, 17.5 pg/ml, 52.7 pg/ml, 2.5 pg/ml, and 10.7 pg/ml, respectively. The serum levels of IFNγ, IL-10, MCP-1, and IL-12p70 were below the detection limits at all time points (days 0, 10, and 28) (data not shown). IL-17 and IFNγ in supernatants from cultured lymph node cells were measured by standard ELISA (R&D Systems) according to the manufacturer's instructions. The detection limits for IL-17 and IFNγ were 12 pg/ml and 10 pg/ml, respectively.

Statistical analysis.

For statistical evaluation, a nonparametric one-way analysis of variance (Kruskal-Wallis) with post hoc analysis was used. P values less than 0.05 were considered significant.

RESULTS

Decreased AIA severity in galectin 3−/− mice.

The murine AIA model of chronic joint inflammation was used to study the role of galectin 3 in arthritis development. This model, resembling chronic arthritis (19), is effective in the C57BL/6 mouse strain, which constitutes the genetic background of the galectin 3–deficient mice. Galectin 3−/− and WT mice were challenged with mBSA according to the regimen illustrated in Figure 1. As expected, by day 28, all WT mice had developed severe arthritis in the immunized knee joint, as revealed by massive cellular infiltration of the synovial membrane and joint erosion (Figures 2A and B). In contrast, histologic examination of joints from galectin 3−/− mice showed significantly less synovial inflammation and joint damage, indicating that the absence of galectin 3 lessens disease severity (Figures 2A and B).

Figure 2.

Deficiency of galectin 3 (gal-3) reduces the severity of antigen-induced arthritis. Knee joints obtained on day 28 (7 days after arthritis induction) from wild-type (WT) mice treated with vehicle (veh) (n = 8), galectin 3−/− mice treated with vehicle (n = 10), and galectin 3−/− mice treated with exogenous galectin 3 (n = 10) were examined histologically for inflammation. A, Representative images of joint histology are shown. Boxed areas showing sites of erosion in the upper panels are shown at higher magnification in the lower panels. B, The severity of arthritis is shown as synovitis (synovial membrane inflammation and hyperplasia) and joint erosion. Each symbol represents 1 individual mouse. Horizontal lines indicate the median. ∗ = P < 0.05; ∗∗ = P < 0.01.

The galectin 3 deficiency in mice was partially rescued by the addition of recombinant galectin 3 at 4 time points (days 0, 3, 21, and 24). We chose these time points in order to intervene with all possible biologic mechanisms leading to arthritis. Thus, interventions were scheduled both prior to (day 0) and during (day 3) the development of an initial immune response as well as before (day 21) and during (day 24) the enhancement (booster) of the immune reaction initiating the establishment of pathologic signs. The joints from galectin 3−/− mice receiving galectin 3 treatment had increased synovitis compared with those from galectin 3−/− mice that received vehicle control, and the level of synovitis was comparable with that in the WT group (severe synovitis in 8 of 8 WT mice and in 7 of 10 galectin 3−/− mice receiving galectin 3 treatment in contrast to 3 of 10 galectin 3−/− mice). While 3 galectin 3−/− mice receiving recombinant galectin 3 displayed severe bone erosion, the mode bone erosion in these mice was comparable with that in galectin 3−/− mice that did not receive exogenous galectin 3 (Figures 2A and B). These data clearly indicate that galectin 3 plays a pathogenic role in murine AIA.

Reduced serum levels of mBSA-specific IgG antibodies in galectin 3−/− mice.

The antigen-specific antibody response in AIA was investigated by measuring serum levels of mBSA-specific IgG. Sera from galectin 3−/− mice contained significantly lower concentrations of mBSA-specific IgG on day 10 than those in sera from WT mice (Figure 3A). Methylated BSA–specific IgG levels in sera from galectin 3−/− mice receiving vehicle treatment also differed significantly from those in sera from galectin 3−/− mice that received exogenous galectin 3 (Figure 3A). The levels of mBSA-specific IgG in all murine sera were increased by day 28, and at this time point, there were no significant differences in mBSA-specific IgG levels between the groups (Figure 3A). This suggests that galectin 3 deficiency results in a delayed antibody response in AIA and that addition of exogenous galectin 3 can reverse this effect.

Figure 3.

Altered levels of circulating methylated bovine serum albumin (mBSA)–specific IgG antibodies, interleukin-6 (IL-6), and tumor necrosis factor α (TNFα) in sera from galectin 3−/− (gal-3−/−) mice. Sera obtained on day 10 (D10) and day 28 (D28) from wild-type (WT) mice treated with vehicle (veh), galectin 3−/− mice treated with vehicle, and galectin 3−/− mice treated with exogenous galectin 3 were analyzed for mBSA-specific IgG antibodies (A), IL-6 (B), and TNFα (C). Dashed lines indicate detection limits. Each symbol represents 1 individual mouse. Horizontal lines indicate the median. ∗ = P < 0.05. AU = arbitrary units.

Reduction of circulating proinflammatory cytokine levels in galectin 3−/− mice.

The proinflammatory cytokines TNFα and IL-6 are known to play important roles in the pathogenesis of RA (1, 20). In order to investigate whether the reduced arthritis seen in galectin 3−/− mice was accompanied by altered proinflammatory cytokine responses, the levels of TNFα and IL-6 in mouse serum were monitored during the course of the experiment. On day 0, the (basal) levels of IL-6 and TNFα in serum were below the detection limit in all groups of mice (data not shown). Levels of IL-6 and TNFα were detectable in sera on day 10 and day 28, with the highest concentrations found on day 10, that is, following immunization but prior to the induction of joint arthritis. On day 10, sera from WT mice contained significantly higher concentrations of both IL-6 and TNFα than did sera from galectin 3−/− mice. Administration of galectin 3 to galectin 3−/− mice caused the levels of IL-6 and TNFα to increase toward those found in WT mouse sera, but these increased cytokine levels did not differ significantly from the levels in galectin 3−/− mice (Figures 3B and C). In contrast to serum concentrations on day 10, the levels of IL-6 and TNFα in sera obtained on day 28 (at the termination of the experiments) were substantially diminished, and no significant differences were found in sera obtained from all groups of mice at this time point. These data suggest that galectin 3 modulates TNFα and IL-6 only at early stages of AIA.

Levels of IFNγ and IL-17 in lymph node culture supernatants.

The inflammatory environment locally in the inflamed joint is closely reflected in the draining lymph nodes, and the proinflammatory cytokines IFNγ and IL-17 are both suggested to be important for the pathologenesis of arthritis (21, 22). To determine whether galectin 3 deficiency modulates the production of IFNγ and IL-17 locally, lymph node cells obtained from WT or galectin 3–deficient mice were stimulated with mBSA, and the proliferative response and culture supernatants were analyzed. A comparable proliferation index was found in WT and galectin 3−/− mice after 48 hours of stimulation with either mBSA or Con A (data not shown). We found no differences between galectin 3−/− and WT mice in the levels of IFNγ from lymph node cell culture supernatants (Figure 4A), indicating that the Th1 response is of minor importance in the galectin 3–mediated proinflammatory effects in AIA. However, although no statistically significant differences in the levels of secreted IL-17 from cultured lymph node cells could be detected, there was a clear tendency toward lower levels of IL-17 in the cultured lymph node cell supernatants from vehicle-treated galectin 3−/− mice than in those from WT mice (Figure 4B).

Figure 4.

Levels of interferon-γ (IFNγ) and interleukin-17 (IL-17) in lymph node culture supernatants. Supernatants of lymph node cells from wild-type (WT) mice treated with vehicle (veh), galectin 3−/− (gal-3−/−) mice treated with vehicle, and galectin 3−/− mice treated with exogenous galectin 3 were collected after in vitro stimulation with methylated bovine serum albumin for 48 hours. The levels of IFNγ (A) and IL-17 (B) in supernatants were quantified by enzyme-linked immunosorbent assay. In B, lymph node cells from 5 WT mice were used due to the limited number of lymph node cells obtained from other mice in that group. Each symbol represents 1 individual mouse. Horizontal lines indicate the median.

Reduced arthritis in galectin 3−/− mice is accompanied by a decreased IL-17 response systemically.

We further investigated whether galectin 3 modulates IL-17–producing cells systemically in AIA by measuring the functionally active IL-17–producing splenocytes using the ELISpot assay. Cells were stimulated antigen specifically (with mBSA) or polyclonally (with PMA and ionomycin) for 42 hours, and the number of IL-17–producing cells was determined by counting the spots. In order to relate the number of IL-17 spots to the frequency of CD3+ cells in the spleen, the ELISpot data were combined with the flow cytometry data for the CD3+ spleen cells, and the data are presented as the number of IL-17 spots per 105 CD3+ cells.

On the day of termination, flow cytometry data for the CD3+ spleen cells from WT mice, vehicle-treated galectin 3−/− mice, and galectin 3−/− mice treated with exogenous galectin 3 showed no significant differences in either frequency or absolute number (data not shown). Antigen-specific stimulation resulted in a significantly lower frequency of IL-17–producing T cells in galectin 3−/− mice than in WT mice (Figure 5A). Addition of exogenous galectin 3 resulted in a tendency toward an increased frequency of IL-17–producing cells; however, the difference was not statistically significant (Figure 5A). The difference between galectin 3−/− and WT mice could not be explained by differences in proliferation, since the splenocyte proliferation in response to Con A or mBSA was comparable between the groups (data not shown). Polyclonal stimulation of splenocytes with PMA and ionomycin also resulted in significantly lower frequencies of IL-17–producing T cells in vehicle-treated galectin 3−/− mice than in both WT mice and galectin 3−/− mice treated with exogenous galectin 3 (Figure 5B). In contrast to IL-17–producing cells, the total number of Treg cells in WT mouse spleen did not differ significantly from that in galectin 3−/− mouse spleen (Figure 6). In conclusion, these data show that the reduced inflammatory response in galectin 3−/− mice is accompanied by lower frequencies of functionally active IL-17–producing T cells in the spleen.

Figure 5.

Reduced frequency of interleukin-17 (IL-17)–producing cells in galectin 3−/− (gal-3−/−) mice. Enzyme-linked immunospot assay was used to quantify the number of IL-17–producing cells among the population of CD3+ cells in the spleen from wild-type (WT) mice treated with vehicle (veh), galectin 3−/− mice treated with vehicle, and galectin 3−/− mice treated with exogenous galectin 3. Splenocytes were stimulated with methylated bovine serum albumin (mBSA) (A) or phorbol myristate acetate (PMA)/ionomycin (B) for 42 hours. Each symbol represents 1 individual mouse. Horizontal lines indicate the median. ∗ = P < 0.05; ∗∗ = P < 0.01.

Figure 6.

Frequency of Treg cells in galectin 3−/− (gal-3−/−) mice. Spleen cells derived from wild-type (WT) mice treated with vehicle (veh), galectin 3−/− mice treated with vehicle, and galectin 3−/− mice treated with exogenous galectin 3 were stained with peridinin chlorophyll A protein–conjugated anti-CD4 and fluorescein isothiocyanate–conjugated anti-CD25, followed by fixation, permeabilization, and intracellular staining with phycoerythrin-conjugated anti-mouse/rat FoxP3. Data are presented as the number of Treg cells positive for CD4, CD25, and FoxP3. Each symbol represents 1 individual mouse. Horizontal lines indicate the median.

DISCUSSION

Galectins have been shown to affect the function of immune cells and to regulate the pathogenesis of various inflammatory diseases (7, 10). With respect to arthritis, numerous studies indicate that various galectins may influence disease development in both proinflammatory and antiinflammatory directions. There is evidence that galectin 1 and galectin 9 have antiinflammatory properties, suppressing the inflammatory process and reducing the severity of CIA (11, 23). In contrast, we show here that mice deficient in galectin 3 had significantly less joint inflammation and cartilage damage compared with WT mice after induction of AIA, demonstrating that galectin 3 contributes to the pathology of AIA. Further, galectin 3−/− mice displayed decreased production of the proinflammatory cytokines IL-6, TNFα, and IL-17, supporting the notion of a proinflammatory role for galectin 3 in AIA. It is intriguing how subtle carbohydrate differences between different galectins can lead to a completely opposite effect in immune regulation, and these aspects of galectins in arthritis need further investigation. To our knowledge, this is the first report providing in vivo evidence that galectin 3 plays a crucial role in the development of arthritis. The consequences of our findings include the possibility of alleviating arthritis processes by modulating galectin 3 expression.

We show that the protection against arthritis development in galectin 3−/− mice was partially reversed when exogenous galectin 3 was administered IP to these mice, but the disease severity was still less pronounced than that in WT mice. Since addition of exogenous galectin 3 will restore galectin 3 only temporarily to local environments and particularly to the extracellular, but not the intracellular, environment, the observation that galectin 3 could enhance disease severity suggests that functions exerted by galectin 3 in the extracellular environment promote disease. The inability of galectin 3 to fully restore disease to levels in WT mice may be due to a number of factors, such as the involvement of intracellular galectin 3 in addition to extracellular galectin 3 or insufficient concentrations in the joint or other sites of importance for driving the inflammation. It remains to be further investigated whether galectin 3 added at other time points could restore disease to a greater or lesser extent. Interestingly, galectin 3 may play a different, protective role in arthritis if administered locally in the joint. This has been shown in adjuvant-induced arthritis, in which bone destruction and osteoclast recruitment were suppressed by injection of galectin 3 into the ankle joints of rats (24).

Consistent with our results, a proinflammatory activity of galectin 3 has also been observed in other murine models of inflammatory disease, including Dextran sulfate sodium–induced colitis, atherosclerosis, and experimental autoimmune encephalomyelitis (25–27). Interestingly, antiinflammatory functions suppressing disease development have also been ascribed to galectin 3. For example, using galectin 3−/− mice in a pneumonia model induced by Streptococcus pneumoniae, Farnworth et al showed that galectin 3 reduces the severity of pneumonia (28). Consistent with this finding, we have preliminary data showing that exogenous galectin 3 exerts an antiarthritic effect when administered to mice with septic arthritis (Forsman H: unpublished observations). It is intriguing to ask whether the effects of galectin 3 can be categorized as proinflammatory or antiinflammatory based on the involvement of pathogens in the induction of inflammation; however, we can only speculate about this at present.

Our study clearly demonstrates that galectin 3 plays a crucial role in arthritis development by aggravating the severity of AIA, and the data on proinflammatory cytokines, the frequency of IL-17–producing cells, and the antigen-specific antibody response support the notion of a proinflammatory effect of galectin 3 in this context. However, it is currently not possible to do more than speculate about the underlying mechanisms leading to the effects of galectin 3; multiple pathways are probably involved, since galectin 3 has the ability to interact with and modulate the functions of a number of targets, immune and nonimmune cells, both extracellularly and intracellularly. Galectin 3 may thus directly induce one or several of the humoral and cellular responses we measured that may be the direct driving forces in the inflammatory process, but the altered proinflammatory responses we observed upon galectin 3 deficiency could also be mere consequences of a reduced arthritis initiated and regulated by other mechanisms.

With this in mind, we speculate that one possible mechanism in the galectin 3–mediated arthritis-promoting effect is the ability of galectin 3 to up-regulate the frequency of IL-17–producing cells during inflammation. Recent discovery of the novel pathogenic IL-17–producing T cell population (Th17 cells) strongly suggests that these cells are far more efficient at inducing autoimmune inflammation in mice than are classic IFNγ-producing Th1 cells (29, 30). Activation of these cells can drive an inflammatory process that is distinct from the classic Th1 and Th2 responses (29). In addition to Th17 cells, IL-17 is also produced by γ/δ T cells (31). High levels of IL-17 have been detected in synovial fluid from RA patients and joints in arthritic mice (21, 32, 33). However, the contribution of IL-17 to the pathogenesis of experimental (murine) arthritis is a topic of controversy. In CIA, addition of IL-17 enhances disease development, and IL-17 deficiency is protective (34–36). However, in proteoglycan-induced arthritis, IL-17 is dispensable, since the onset and severity of disease are equivalent in WT and IL-17−/− mice (37).

Our data clearly demonstrate that galectin 3−/− mice, which are protected from disease, display a significantly lower frequency of IL-17–producing CD3+ splenocytes compared with WT mice. We also found a tendency toward decreased secretion of IL-17, but not IFNγ, in supernatants from cultured lymph node cells of galectin 3−/− mice compared with WT mice. An elevated level of IL-17–producing cells (and subsequently of IL-17), as seen in the WT mice, may lead to increased production of proinflammatory mediators from cells present in the joint, including synovial fibroblasts, monocytes, macrophages, and chondrocytes, which results in cartilage damage and bone erosion, as described earlier (21, 38).

Another hypothetical mechanism for mediating the proinflammatory activity of galectin 3 is the direct targeting by galectin 3 of synovial fibroblasts and neutrophils locally in the joints; these are important cells involved in amplifying the local inflammation and joint damage. It has been shown that upon galectin 3 stimulation, isolated synovial fibroblasts from RA patients could secrete proinflammatory and mononuclear cell–recruiting chemokines, suggesting an important role for galectin 3 in promoting mononuclear cell infiltration in the joints (16). This might partly explain the significantly lower cell infiltration in joints that was observed in galectin 3−/− mice in our study. Decreased infiltration of cells in the joints may also be a consequence of reduced IL-17 production, since IL-17 plays an essential role in mediating neutrophil migration during inflammation through chemokine release, in addition to a direct role in joint destruction (39). Once accumulated at the site of inflammation, neutrophils are potentially responsible for damage to surrounding tissues by producing ROS and hydrolytic enzymes. Interestingly, our group and others have previously reported that galectin 3 can induce massive production of ROS from exudated human neutrophils (14, 40). Supporting the possibility of a role of ROS in joint destruction, we have indeed found that compared with WT mouse splenocytes, splenocytes derived from galectin 3−/− mice produce much less ROS in vitro in response to the potent ROS inducer PMA (Önnheim K: unpublished observations).

In conclusion, this study shows that targeted deletion of endogenous galectin 3 reduces the severity of AIA and that exogenously added galectin 3 partly restores disease severity. Hence, we conclude that galectin 3 plays a pathogenic role in the development and progression of AIA and that disease severity is accompanied by alterations in the levels of antigen-specific IgG, the systemic levels of TNFα and IL-6, and the frequency of IL-17–producing T cells. We speculate that elevated levels of galectin 3 in the joint can stimulate fibroblasts and locally infiltrated neutrophils to produce chemokines/cytokines and ROS, effector molecules involved in amplifying the immune responses and causing bone destruction. Our data thus support recent research proposing galectin 3 to be a novel biomarker for diagnosing diseases such as heart failure, thyroid cancer, and RA (4, 41). Further, targeting endogenous inflammatory mediators such as galectins to dampen or resolve the inflammatory response should serve as a platform for designing novel drugs for treating autoimmune diseases such as RA.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Forsman 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 conception and design. Forsman, Islander, Magnusson, Brown, Karlsson.

Acquisition of data. Forsman, Islander, Andréasson, Andersson, Önnheim, Karlström, Sävman, Brown, Karlsson.

Analysis and interpretation of data. Forsman, Islander, Magnusson, Brown, Karlsson.

Acknowledgements

We thank Fei Ying for excellent technical assistance.

Ancillary