Collagen-induced arthritis (CIA) is an animal model for rheumatoid arthritis (RA). Lipopolysaccharide (LPS) is known to accelerate CIA; however, the pathogenetic mechanisms are not yet fully understood. In this study, type II collagen (CII)-immunized mice were found to have marked increases in degree of expression of mRNA of inflammatory mediators such as tumor necrosis factor alpha (TNF-α), interleukin (IL)-1β, and macrophage inflammatory protein-2 (MIP-2) in their arthritic paws and of serum anti-CII antibody concentration before the onset of arthritis induced by LPS injection. The gene expression was rapid and continuous after direct activation of nuclear factor κB. The amounts of mRNA of TNF-α, IL-1β, and MIP-2, as well as of matrix metalloproteinases and the receptor activator of nuclear factor κB ligand, increased with the development of arthritis, correlated positively with clinical severity and corresponded with histopathological changes. Moreover, anti-TNF-α neutralizing antibody inhibited the development of LPS-accelerated CIA and a single injection of recombinant mouse TNF-α induced increases in anti-CII antibody concentrations, suggesting TNF-α may contribute to the development of arthritis by both initiation of inflammation and production of autoantibodies. These data suggest that exacerbation of RA by LPS is associated with rapid and continuous production of inflammatory mediators and autoantibodies.
type II collagen
- E. coli
inhibitor of nuclear factor κB alpha
lipopolysaccharide-accelerated collagen-induced arthritis
macrophage inflammatory protein-2
nuclear factor κB
receptor activator of nuclear factor κB ligand
standard error of the mean
tumor necrosis factor alpha
Rheumatoid arthritis is a systemic autoimmune disease that is characterized by chronic inflammation of joints, leading to synovial hyperplasia, infiltration of leukocytes and progressive destruction of cartilage and bone . Various humoral factors are involved in the development of RA. One of these is autoantibodies, which activate complement systems by formation of immune complexes in combination with self-antigens and induce recruitment of inflammatory cells into synovium [1, 2]. Another important factor is inflammatory mediators such as cytokines, chemokines and proteases that induce inflammation and tissue destruction [1, 3-6]. In particular, proinflammatory cytokines including TNF-α, IL-1β, IL-6 and IL-17 amplify inflammation by inducing production of various additional inflammatory mediators and induction of recruitment, differentiation and proliferation of inflammatory cells. These inflammatory mediators are detectable in high concentrations in synovial fluid or serum of RA patients [7-14]. Although some biologics targeting proinflammatory cytokines have improved the effectiveness of RA therapy , some patients are unresponsive to such therapy. Thus, the pathogenesis of RA is complex and has not yet been fully elucidated.
Collagen-induced arthritis is a common animal model with similarities to RA in both cause and pathology. Synovial hyperplasia, infiltration of inflammatory cells and destruction of cartilage and bone occur in the joint tissue of both CIA mice and humans with RA . In addition, an autoantibody, namely anti-CII antibody, triggers CIA . Injection of a cocktail with monoclonal anti-CII antibody induces arthritis in mice [18, 19]. Furthermore, proinflammatory cytokines play important roles in the pathogenetic process of CIA. Antibodies against TNF-α, IL-1β and IL-6 are effective against CIA [20-22]. In addition, it is well known that CIA is accelerated by LPS, a major component of the outer membrane of gram-negative bacteria [23-26]. LPS-CIA is more useful than CIA alone for screening for anti-rheumatic drugs because it enables not only shortening the duration of experiments but also more accurate evaluation of incidence and severity . On the other hand, the characteristics of some pathogenetic mechanisms are similar in LPS-CIA and RA. It is well known that some patients with RA have acute onset and exacerbations of their disease. Administration of LPS reactivates paw inflammation in CIA mice, suggesting LPS plays a role in the exacerbation of autoimmune diseases, including RA [27, 28]. Likewise, LPS-CIA induces acute onset of inflammation after LPS injection in CII-immunized mice, in contrast with unmodified CIA, which develops slowly . Therefore, LPS-CIA is a valuable disease model for investigating the acute phase of RA. However, few studies have analyzed the pathogenetic mechanisms of LPS-CIA. In particular, there are no detailed analyses focusing on the period before the onset of arthritis.
In this study, we have demonstrated an association between induction of paw inflammation and both gene expression of inflammatory mediators and production of anti-CII antibody prior to the onset of arthritis. In addition, we propose that TNF-α contributes to the development of LPS-CIA by both initiation of inflammation and production of anti-CII antibody. Our findings suggest that rapid and continuous production of inflammatory mediators and autoantibodies is involved in the mechanisms of RA exacerbation.
1 MATERIALS AND METHODS
Female DBA/1JNCrj mice were purchased from Charles River (Tokyo, Japan). All mice were used at the age of 7–10 weeks. All experimental procedures were performed in accordance with the in-house guidelines of the Institutional Animal Care and Use Committee of Daiichi Sankyo (Tokyo, Japan).
1.2 Induction and assessment of arthritis in mice
The mice were immunized by intradermal injections at the base of the tail with an emulsion containing 150 μg of bovine CII (Collagen Gijutsu-Kenshukai, Tokyo, Japan) in Freund's complete adjuvant containing Mycobacterium butyricum (Difco, Detroit, MI, USA). After 17–21 days, the mice were injected s.c. with 0.2 mg/kg of LPS of E. coli O111:B4 (Difco) or with 2, 20 and 200 ng/kg of recombinant mouse TNF-α i.v. (Genzyme, Boston, MA, USA) in place of LPS. Anti-TNF-α neutralizing antibody (25 mg/kg) or control IgG (25 mg/kg) purified from the ascites of mice injected with hybridoma (MP6-XT22 and Y13-259, respectively) were administered i.p. once daily from the day before to 6 days after LPS injection. Clinical severity of paw inflammation was scored periodically on a scale of 0–3 according to the following criteria: 0, normal; 0.5, swelling of one digit; 1, swelling of more than two digits or redness of paw; 2, swelling of part of paw; and 3, swelling of entire paw.
The hind paws of the mice were severed between the knee and ankle, fixed in PBS containing 10% formaldehyde, decalcified in 10% ethylenediamine tetraacetic acid and embedded in paraffin. The paws were sliced horizontally to the footpad, sectioned, and stained with HE. Synovium and bone/cartilage tissues of their tarsal joints were evaluated by light microscopy. Infiltration of leukocytes, proliferation of synovial cells, edema, destruction of bone and cartilage tissues, and osteoclast formation in tarsal joints were scored on a scale of 0–3 according to the following criteria: 0, no abnormality detected; 0.5, very slight; 1, mild; 2, moderate; and 3, severe or marked.
1.4 Quantitative real-time polymerase chain reaction
Total RNA was extracted from hind paws of the mice that had been homogenized in ISOGEN (Nippon Gene, Tokyo, Japan) using Polytron and cDNA-synthesized using SuperScript III First-Strand Synthesis Super Mix (Life Technologies, Carlsbad, CA, USA). PCR was performed using a 7500 real-time PCR system (Life Technologies) with TaqMan Gene Expression Assays (Life Technologies) and Premix Ex-Taq (Takara Bio, Otsu, Japan) according to the manufacturers' instructions. The degree of mRNA expression was normalized with each degree of GAPDH expression.
1.5 Measurement of anti-type II collagen antibody concentration
The concentrations of anti-CII IgG in mouse serum were measured using ELISA. Briefly, a 96-well plate coated with 0.1 μg/well of bovine CII was incubated overnight at 4°C, then washed with PBS containing 0.05% Tween 20 followed by blocking with PBS containing 1% BSA, 5% sucrose, and 0.05% sodium azide for 1 hr at room temperature. After washing, the plates were incubated with the serum for 2 hrs at room temperature. The plates were then washed and incubated with goat anti-mouse IgG conjugated with horseradish peroxidase for 2 hrs at room temperature. After washing, the plates were incubated with tetramethyl benzidine substrate for 20 mins and the optical density of the wells at 450 nm measured. The concentrations of anti-CII IgG were quantitated by using a standard curve derived from anti-human CII IgG (Daiichi Fine Chemical, Takaoka, Japan).
1.6 Statistical analysis
Statistical significance was determined by Dunnett's test or Student's t-test for degree of expression of mRNA and anti-CII IgG concentrations, by nonparametric Dunnett's test for histopathological scores and by Wilcoxon's rank sum test for clinical scores. To elucidate the correlation between the amount of mRNA and clinical score, Spearman's rank order correlation test was used.
2.1 Lipopolysaccharide induces arthritis in type II collagen-immunized mice
Type II collagen-immunized DBA/1JNCrj mice were subcutaneously injected with LPS. Paw swelling was observed from Day 2: it reached a peak on Day 7, then declined slowly until Day 17 (Fig. 1a). The prevalence of swelling remained stable and high until Day 17 (Fig. 1b). These findings show that LPS induces arthritis in CII-immunized mice.
Next, the histopathological features of synovium and bone/cartilage tissues in the tarsal joints of LPS-CIA mice were assessed. In the synovium, infiltration of leukocytes, proliferation of synovial cells and edema were observed from Day 1, reaching maximal and sustained levels from Days 3 to 7 (Fig. 1c, d–f). The infiltrating leukocytes were mainly neutrophils. The proliferating synovial cells were both macrophage-like and fibroblast-like. Destructions of bone and cartilage tissues were observed from Day 1, reaching a peak on Day 7 (Fig. 1c, g–i). Along with development of bone destruction, an increase in the number of osteoclasts around the area of destruction was observed (Fig. 1c). On Day 19, most histopathological features had resolved almost completely, although proliferation of synovial cells and tissue destruction remained a mild to moderate (Fig. 1c). These findings indicate that joints of LPS-CIA mice have similar histopathological features to those of humans with RA.
2.2 Gene expression of inflammatory mediators is increased in lipopolysaccharide-accelerated collagen-induced arthritis paws
To evaluate the association of inflammatory mediators such as cytokines, chemokines and proteases with the development of LPS-CIA, degree of gene expression of TNF-α, IL-1β, IL-6, IL-17A, RANKL, MIP-2 (the functional IL-8 homolog in mice), MMP-3 and MMP-9 in the mouse paws were measured. The degree of expression of TNF-α was increased at 6 hrs after LPS injection (Fig. 2a). TNF-α expression returned to baseline levels on Day 3 (after the onset of arthritis) and increased again on Day 6 in parallel with the development of arthritis (Fig. 2a). Additionally, on Days 3 and 6 the degree of expression of the gene for TNF-α was positively correlated with the clinical scores (Fig. 2b). Moreover, the other assessed genes also increased and their expression also correlated with the development of arthritis (Table 1). In contrast, saline-injected mice did not have upregulation of the gene expression at these time points (data not shown). These data suggest that these inflammatory mediators are related to the development of arthritis in this model.
|Fold change over pre-LPS injection†||Correlation‡|
|Day 3||Day 6*||R|
|IL-1β||8.0 ± 2.2||21.5 ± 2.8||0.82|
|IL-6||21.2 ± 8.8||32.9 ± 6.6||0.75|
|IL-17A||7.2 ± 2.6||13.0 ± 2.9||0.62|
|RANKL||5.9 ± 1.2||20.4 ± 2.7||0.84|
|MIP-2||4.9 ± 1.6||16.4 ± 2.8||0.79|
|MMP-3||14.2 ± 3.8||48.3 ± 6.9||0.89|
|MMP-9||2.3 ± 0.5||19.5 ± 2.5||0.78|
Next, the period before the onset of arthritis was focused on and short-term changes in gene expression up to 24 hrs after LPS injection analyzed. Because LPS directly induces expression of many inflammatory mediators including TNF-α, IL-1β, IL-6 and MIP-2 through NF-κB, gene expression of IκBα, which represents the NF-κB target gene, was measured. The amount of IκBα mRNA in the paws had increased by 2 hrs after LPS injection, then had decreased to a small amount at 6 hrs, which was maintained until 24 hrs (Fig. 2c). These data indicate that LPS directly activated NF-κB in the paws around 2 hrs after LPS injection. In addition, expression of TNF-α, IL-1β, IL-6 and MIP-2 had also increased by 2 hrs after LPS injection (Fig. 2d). However, expression of TNF-α continued to increase, at least until 24 hrs, whereas that of IL-1β and MIP-2 increased again 8 hrs after LPS injection. In contrast, although the amount of MMP-3 had increased by 24 hrs, expression of MMP-3 and MMP-9 did not change until 8 hrs. These data demonstrate that some inflammatory mediators are upregulated rapidly and continuously after direct activation of NF-κB and before the onset of LPS-CIA.
2.3 Lipopolysaccharide induces rapid production of anti-type II collagen antibody in type II collagen-immunized mice
To examine whether LPS induces production of specific autoantibodies, anti-CII IgG concentrations in the mice sera were measured. Anti-CII IgG concentrations in CII-immunized mice were increased by Day 6 after LPS injection although the difference in anti-CII IgG concentrations between saline and LPS injection was not statistically significant (Fig. 3a). Furthermore, analysis of short-term changes (up to 24 hrs after LPS injection) showed that anti-CII IgG concentrations increased from 2 hrs after LPS injection (Fig. 3b). Anti-CII antibody concentrations were greater on Day 6 than at 24 hrs; however, these were separate experiments. These data suggest that LPS induces rapid and continuous production of anti-CII antibody in CII-immunized mice.
2.4 Tumor necrosis factor alpha is involved in development of lipopolysaccharide-accelerated collagen-induced arthritis
Because TNF-α was expressed rapidly and continuously after LPS injection, the involvement of TNF-α in the development of LPS-CIA was examined. When TNF-α was injected into CII-immunized mice in place of LPS, severe arthritis did not occur (data not shown). However, in TNF-α-injected mice, serum anti-CII IgG concentrations increased in a dose-dependent manner (Fig. 4a). These data indicate that TNF-α can induce production of anti-CII antibody, but is not sufficient on its own to induce arthritis. Finally, the contribution of TNF-α to LPS-CIA development was investigated by using anti-TNF-α neutralizing antibody. Administration of anti-TNF-α antibody once daily from the day before to 6 days after LPS injection significantly inhibited development of LPS-CIA (Fig. 4b). These data indicate that TNF-α plays some part in the development of LPS-CIA.
Autoimmune diseases sometimes have acute onsets and periods of exacerbation, the causes for which have not yet been clarified. Because a number of studies have shown that LPS may play a role in the exacerbation of autoimmune diseases including RA, LPS-CIA is a valuable disease model for investigation of the acute phase of RA [27, 28]. In this study, we elucidated details of the pathogenesis of LPS-CIA from the perspective of the sequence and timing of production of inflammatory mediators and anti-CII antibody and related this to histopathological findings.
Inflammatory mediators such as cytokines, chemokines and proteases are involved in the development of RA through induction of inflammation and destruction of tissues. Upregulation of inflammatory mediators has been observed in both RA and CIA [7-14, 29]. Previous studies have reported increased expression of mRNA of IL-1β and MMPs in arthritic paws of animals with LPS-CIA [30, 31]. In the present study, we clearly demonstrated that LPS-CIA mice have marked increases in the degree of gene expression of inflammatory mediators, including IL-1β, MMP-3 and MMP-9 in their arthritic paws. The amount of mRNA in them increased in parallel with the development of arthritis and correlated positively with clinical severity. Importantly, TNF-α, IL-1β and MIP-2 were rapidly and continuously expressed after direct activation of NF-κB and before the onset of arthritis. In particular, the degree of expression of TNF-α was greater before the onset of arthritis than after its onset. However, we did not observe these responses for MMP-3 and MMP-9, which are overexpressed as a result of aspects of joint inflammation such as proliferation of synovial cells and infiltration of inflammatory cells . These findings suggest that upregulation of TNF-α, IL-1β and MIP-2 is involved in the mechanisms of induction of LPS-CIA. Furthermore, IL-17A was upregulated in parallel with the development of arthritis in the paws of LPS-CIA mice. It is well known that Th17 cells produce IL-17A, which is involved in human RA and arthritis in animal models . This finding suggests that Th17 cells play important roles in the pathogenesis of LPS-CIA.
We confirmed that arthritic paws of LPS-CIA mice are characterized by similar histopathological features as human RA; these include bone and cartilage destruction, synovial change and infiltration of inflammatory cells. Moreover, expression of inflammatory mediators is likely to correlate with histopathological findings. There is consensus that synovial hyperplasia caused by a marked increase in macrophage-like and fibroblast-like synoviocytes is a hallmark of RA . We observed proliferation of both macrophage-like and fibroblast-like synovial cells in parallel with the development of LPS-CIA, which demonstrates that LPS-CIA has this pathological hallmark in common with RA. In addition, we observed the increase in synovial cells from the early stages of arthritis. The increase in MMP-3 we observed 24 hrs after LPS injection may reflect the proliferation of synovial cells, which generally takes 1 day. Notably, infiltration of neutrophils accompanied by edema occurred first; this coincided with the onset of arthritis, suggesting that neutrophil infiltration is the main cause of the inflammation in LPS-CIA. Recent studies have highlighted the possible contribution of neutrophils in the early phases of RA physiopathology. Although it is well known that mononuclear cells such as T cells, B cells and macrophages infiltrate the synovium in RA, neutrophils migrate to the synovial fluid before mononuclear cells do [33-35]. Previous studies have reported accumulation or participation of neutrophils in commonly-used arthritis models, including CIA [29, 36, 37]. In addition, the histopathology of LPS-CIA is similar to that of anti-CII antibody and LPS-induced arthritis, in both of which neutrophils, but not T and B cells, play crucial roles in the development of arthritis . Thus, neutrophil infiltration may cause the acute inflammation in LPS-CIA; this suggests that LPS-CIA has pathogenetic mechanisms in common with the early phase of RA. Upregulation of MIP-2 may play a part in recruitment of neutrophils to inflammatory sites. Likewise, TNF-α and IL-1β may result in edema formation by initiating acute inflammation with enhanced vascular permeability. In addition, the increase in osteoclasts associated with bone and cartilage destruction is consistent with increased expression of RANKL. Similarly, overexpression of MMP-3 and MMP-9 may cause cartilage destruction.
The rapid and continuous increase in serum anti-CII antibody concentrations may, in part, contribute to the development of LPS-CIA. CII is the major constituent protein of the cartilage of joints and autoimmunity to CII occurs in RA patients . The rapidity of increase in production of serum anti-CII antibody induced by LPS may be mediated via a direct effect on B cells. LPS injection results in production of polyclonal antibodies via stimulation of B cells [38, 39]. Anti-CII antibody has the potential to induce arthritis, as evidenced by induction of articular inflammation by serum transfer or collagen antibody cocktail injection [18, 19]. Complex formation between anti-CII antibody and CII in cartilage activates the complement systems and initiates recruitment of neutrophils and macrophages, which are activated by FcγR ligation and secrete inflammatory mediators . In the present study, we detected a second increase in gene expression of some inflammatory mediators 2–8 hrs after LPS injection, this interval of time being just after the increase in serum anti-CII IgG. Additionally, a previous study has shown that LPS injection of mice without CII immunization does not induce arthritis [23, 26]. Thus, production of anti-CII antibody and the consequent induction of continuous expression of inflammatory mediators may be necessary for acute onset of LPS-CIA. Furthermore, T cells might be involved in LPS-CIA via modulation of pathogenic B cells, which produce anti-CII antibody.
The expression pattern of its mRNA did not decrease after LPS injection and before the onset of arthritis, implying an important role for TNF-α in the development of LPS-CIA. Previous studies have shown that antibodies and soluble TNF-α receptor fusion protein significantly ameliorate CIA [20, 40, 41]. Additionally, anti-TNF-α antibody was effective in LPS-CIA mice which had been given a booster injection of CII 7 days before LPS injection , this procedure differing slightly from ours. In the present study, we confirmed the effectiveness of anti-TNF-α antibody. Our findings indicate that TNF-α is an important molecule in the pathogenesis of LPS-CIA. A single intra-articular injection of TNF-α accelerates the onset of arthritis in rats immunized with CII . In addition, comparable to LPS, administration of recombinant IL-1β can induce arthritis in CII-immunized mice . However, TNF-α did not induce severe arthritis when it was injected i.v. in place of LPS after CII immunization. Interestingly, this experiment showed TNF-α has the potential to induce production of anti-CII antibody. This finding suggests that TNF-α contributes to the development of LPS-CIA by production of anti-CII antibody. On the other hand, because TNF-α alone is not sufficient to induce arthritis, our findings raise the possibility that involvement of other proinflammatory cytokines, such as IL-1β and IL-6, could be necessary for the development of LPS-CIA.
The development and severity of RA are linked to genetic and environmental factors. Microbial infections, which may be a potent source of LPS, may be one of the triggers for autoimmune diseases, including RA [43-45]. It is well known that systemic infections cause acute exacerbations of RA. One of the multiple possible mechanisms by which infections may trigger autoimmunity is the adjuvant effects of pathogens mediated by TLRs and other pattern-recognition receptors on antigen-presenting cells; this would lead to production of inflammatory mediators which in turn would lead to tissue damage . A number of studies suggest a possible role for LPS or its cellular receptor, TLR4, in RA. The serum and synovial fluids of RA patients contain higher concentrations of antibodies against E. coli or LPS-binding protein than do those of healthy subjects [46, 47]. In animal models, the spontaneous arthritis of IL-1Ra−/− mice is markedly suppressed by crossing with TLR4-deficient mice . Moreover, blocking of TLR4 by receptor antagonists suppresses experimental arthritis in mice [48, 49]. We found that, after LPS injection, LPS-CIA mice had increased inflammatory mediators in their paws, similar to that associated with infection. Taken together with these findings, LPS-CIA may have pathogenetic mechanisms in common with infection-associated RA.
In summary, we have demonstrated that LPS induces arthritis in CII-immunized mice and that the onset of that arthritis is preceded by rapid and continuous production of inflammatory mediators and anti-CII antibody. This suggests that production of these humoral factors induces the development of LPS-CIA. In particular, TNF-α may contribute to the development of LPS-CIA via both initiation of inflammation and production of anti-CII antibody. It is possible that the rapid and continuous production of inflammatory mediators and autoantibodies is the mechanism by which infections exacerbate RA.
We thank Dr. Eiichi Imai for his histologic assessment and Dr. Eriko Michishita-Kioi and Dr. Hiroaki Maeda for valuable discussions.
There are no conflicts of interest, including relating to financial support, for this study.