Locally produced proinflammatory cytokines such as tumor necrosis factor-α, interleukin (IL)-1, and IL-6 play a pivotal role in the pathology of rheumatoid arthritis (RA). These cytokines, by upregulation of several genes, are responsible both for the recruitment and continuous activation of the inflammatory cells and for inducing production of the enzymes that destroy bone and cartilage. However, these inflammatory events need to be balanced by the production of endogenous inhibitors, as inflammatory responses are generally localized and the consequent destruction of affected joints is less severe [1].

The ubiquitously expressed heme enzyme indoleamine 2,3-dioxygenase (IDO) catalyzes the non-hepatic oxidative degradation of tryptophan, the initial and rate-limiting step in tryptophan metabolism, resulting in depletion of this least abundant essential amino acid. Tryptophan starvation enables the host to restrict the growth of intracellular pathogens [2]. The expression of IDO in healthy tissues is generally quite low but is markedly upregulated in response to infection and inflammation. Recent studies have established an entirely novel important biological function for IDO. These studies indicate that IDO-mediated tryptophan depletion in vitro results in inhibition of matrix-degrading metalloproteinase enzyme production, suppression of inflammatory responses, and promotion of immune tolerance [3, 4]. Accordingly, an emerging immunological paradigm regards IDO as a key regulatory control enzyme in both innate and adaptive immune responses. Indeed, modulation of cellular IDO expression and/or activity is now increasingly implicated in inflammation and autoimmunity, organ transplantation and fetal rejection, and evasion of immune surveillance by tumor cells. Furthermore, agents that inhibit IDO are under active investigation as potential adjuvants for tumor immunotherapy [5].

IDO-mediated tryptophan catabolism results in depletion of tryptophan with concomitant generation of kynurenine and ultimately NAD (nicotinamide adenine dinucleotide). Expression of the IDO gene is induced in most cell types in response to infection with microbial agents via activation of toll-like receptors. Furthermore, interferon-gamma (IFN-γ), IL-10, and cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) have also been shown to stimulate IDO [4, 6]. In contrast, IL-4, IL-13, and transforming growth factor-β are suppressors of IDO [7]. In previous studies, we have shown that IFN-γ caused time-dependent induction of IDO gene expression and activity in cultured normal fibroblasts [4]. The consequent tryptophan catabolism and ensuing tryptophan starvation in these cultures were associated with profound suppression of collagenase and stromelysin gene expression induced by IL-1β. Abrogation of IL-1β-induced matrix metalloproteinase stimulation upon activation of IDO in these fibroblasts was directly due to the reduction of local tryptophan concentrations rather than the accumulation of kynurenine and other tryptophan metabolites. These observations led us to propose the hypothesis that, through its ability to induce transient tryptophan starvation, IDO represented an important endogenous anti-inflammatory mechanism and that disruption of IDO induction or function would be associated with exaggerated inflammatory responses.

Collagen-induced arthritis (CIA) is generated in genetically susceptible mouse or rat strains by immunization with type II collagen (CII) dissolved in complete Freund's adjuvant (CFA). CIA highly resembles RA according to overlapping immunopathogenic pathways and similar histopathologic features [8]. Among other reasons, the central role of class II major histocompatibility complex [9], the involvement of CD4⁺ T cells in the immunopathogenesis of the disease, the predominantly Th1 type immune response to CII in the induction phase [10, 11], as well as the cytokine profile throughout the evolution of CIA [12] make this experimental animal model one of the most adequate models of RA. To address the question of whether IDO plays a negative regulatory role in an animal model of arthritis, we assessed IDO activity in the sera of mice at different stages of arthritis and the effect of 1-methyl-tryptophan (1-MT), a specific inhibitor of IDO, on arthritis development and antigen-specific immune responses in CIA.

RA and experimental inflammatory joint diseases have a progressive character with involvement of increasing numbers of joints; however, the initial and aggressive acute phase in affected joints slows down over time and the inflammatory processes burn out. Several lines of evidence indicate that the effector mechanism that initially attacks the joints is T cell-driven in response to the effect of proinflammatory cytokines, but the mechanisms responsible for the limitation of acute inflammatory processes are much less understood. The novel finding of our study is that IDO activity is upregulated in the acute phase of CIA reflected by the increased K/T ratio in the serum. Furthermore, we could also demonstrate that inhibition of IDO in this experimental model augments the incidence and severity of the disease and increases the immune responses to the autoantigens and alloantigens. These data suggest that IDO plays a central role in the negative regulatory feedback of immunological mechanisms in inflammatory joint diseases.

CIA, like RA in humans, is characterized by the accumulation of T cells, plasma cells, macrophages, B cells, mast cells, natural killer (NK) cells, and dendritic cells in the synovial sublining [17, 18]. Furthermore, inflammatory cells infiltrating the synovial tissue in RA and in the acute phase of CIA exhibit a predominantly Th1 pattern of cytokine expression [10, 11]. By priming the Th1-type inflammatory cell responses, IFN-γ is one of the most important proinflammatory factors in the induction of T cell-driven autoimmune arthritis, such as CIA. However, IFN-γ plays an ambiguous role in autoimmunity. After the activation of self-reactive lymphocyte clones and of bystander and accessory cells in the acute phase, IFN-γ downregulates the autoimmune processes [19]. Indeed, CIA and CFA developed more readily in IFN-γ receptor-deficient mice than in wild-type littermates [20]. As a possible explanation, it has emerged that CFA elicits strong myelopoiesis and expansion of Mac-1⁺ cells, which play a crucial role in disease pathogenesis, and this process is downregulated by IFN-γ [20]. However, the exact mechanism responsible for the effect of IFN-γ in CIA has not been fully elucidated.

One of the most likely mechanisms for the downregulation of CIA by IFN-γ is the increased expression of IDO by non-T cells [21]. In fibroblasts [22], macrophages [23], and dendritic cells [3], IFN-γ stimulates the enzyme IDO, which degrades the amino acid tryptophan to form kynurenine, resulting in the inhibition of autoimmune processes. According to this assumption, we demonstrated an elevated K/T ratio, indicating high IDO activity during the acute phase of CIA. Moreover, the K/T ratio was even higher in mice that received CII and CFA but developed no clinical and histological signs of arthritis. These data suggest that IDO acts as a negative feedback in this model, and the onset and severity of experimental arthritis are inversely proportional to IDO activity.

To confirm the regulatory role of IDO in CIA, we used 1-MT, a known competitive inhibitor of IDO. 1-MT did not influence IFN-γ, but it significantly suppressed IL-4 production by spleen cells, resulting in an increased Th1/Th2 response. In 1-MT-treated mice, we could demonstrate a significant decrease of K/T ratio in the immediate pre-arthritic and acute phase of arthritis compared to vehicle-treated animals, suggesting the high activity of IDO only in these stages of inflammatory processes. In other words, the low blocking activity of 1-MT either in the pre-arthritic phase or in the chronic phase of arthritis denotes the anti-inflammatory effect of IDO only in the case of upregulation of inflammatory processes, especially Th1 responses.

As a tryptophan-catabolizing enzyme, IDO can induce the peripheral tolerance and reduce the persistent immune activation. On one hand, IDO decreases the tryptophan concentration in the microenvironment of inflammatory cells. Although tryptophan is an essential amino acid indispensable for the biosynthesis of proteins, the low tryptophan concentration results in the arrest of cell proliferation in the mid-G₁ arrest point. T cells are specifically sensitive to tryptophan deprivation [2, 5] and thus IDO activity can block the potential harmful autoimmune response. On the other hand, Zhu and colleagues [24] proposed that synovial T cells derived from RA synovial fluids are resistant to IDO-mediated tryptophan deprivation. This may be one mechanism by which autoreactive T cells are sustained in vivo in patients with arthritis [24]. In addition, selected metabolites on the tryptophan-kynurenine pathway are able to suppress proliferation of allogeneic T cells and, to a lesser extent, B and NK cells [25]. Moreover, some of the kynurenine derivates can induce in vitro the selective apoptosis of Th1 cells, but not Th2 cells [7]. In accordance with these results, 1-MT treatment in our study resulted in an increased, mainly Th1 cell-mediated immune response to the CII and consequently in the significant worsening of severity and increase of onset of CIA mediated by Th1 response.

In this regard, recent studies with the CIA model of arthritis using an antibody to the costimulatory molecule CD137, a member of the tumor necrosis factor receptor superfamily, are of great interest. These results showed that treatment of mice with the anti-CD137 antibody resulted in induction of IDO in vivo, which was associated with significant amelioration of the severity of CIA [26]. Furthermore, pharmacological inhibition of IDO reversed the effects of the anti-CD137 antibody and aggravated the arthritis in this model.

The tryptophan-IDO pathway may have important relevance for the biological therapy of RA. In a study of Boasso and colleagues [27], CTLA-4-Fc treatment of human peripheral blood CD4⁺ T cells resulted in increased IDO expression by these cells. This effect was not observed in CD8⁺ T cells. Thus, abatacept (CTLA4-Ig) therapy may act, at least in part, by the stimulation of IDO production.

The importance of IDO activity in the regulation of CIA supports the hypothesis that IDO expression by antigen-presenting cells is responsible for suppression of undesirable Th1 cells in human inflammatory joint diseases, as tryptophan degradation could be demonstrated in patients with RA [28]. These findings theoretically raise the possibility that the local or systematic induction of IDO activity could be tested in inflammatory joint diseases.