Role of IL-33 in inflammation and disease

Asthma is a chronic inflammatory disease classically characterized by airway hyper-responsiveness, allergic inflammation, elevated serum IgE levels, and increased Th2 cytokine production. Given that IL-33 is a strong inducer of Th2 immune responses its role in asthma has been extensively studied (reviewed in [42]). Initial gene expression studies in a range of tissues using human and mouse cDNA libraries revealed expression of IL-33 in lung tissue, and high expression in bronchial smooth muscle cells [3]. More recently, expression of IL-33 was found in higher levels in endobronchial biopsies from human asthmatic subjects compared to controls. The IL-33 expression was particularly evident in those with severe asthma [43], and the expression was mainly located in bronchial epithelial cells [44]. Studies to investigate which cells were the main IL-33 responsive cells in lung demonstrated that both epithelial and endothelial cells, but not smooth muscle cells or fibroblasts were important [45]. Several animal model studies have highlighted a functionally important role for IL-33/ST2 in asthma and allergic airways inflammation. In a murine ovalbumin-induced airway inflammation model, intranasal administration of IL-33 induces antigen-specific IL-5⁺ T cells and promotes allergic airway disease even in the absence of IL-4 [24]. Furthermore, intranasal IL-33 also promotes airways hyper-responsiveness, goblet cell hyperplasia, eosinophilia, polarization of macrophages towards an M2 phenotype, and accumulation of lung IL-4, IL-5 and IL-13 [34, 46, 47]. More recently, an IL-33 transgenic mouse was generated in which IL-33 expression was controlled under a CMV promoter and released as a cleaved 18 kDa protein in pulmonary tissue [48]. These mice developed massive airway inflammation with infiltration of eosinophils, hyperplasia of goblet cells and accumulation of pro-inflammatory cytokines in bronchoalveolar lavage fluid. In contrast, intraperitoneal anti-IL-33 antibody treatment inhibited allergen-induced lung eosinophilic inflammation and mucus hypersecretion in a murine model [49]. Furthermore, administration of blocking anti-ST2 antibodies or ST2-Ig fusion protein inhibited Th2 cytokine production in vivo, eosinophilic pulmonary inflammation and airways hyper-responsiveness [50]. At present, the role of IL-33/ST2 in studies using ST2-deficient mice is unclear as these mice are not protected in the ovalbumin-induced airway inflammation model but have attenuated inflammation in a short-term priming model of asthma. Furthermore, there is also an exacerbation of disease in wild-type or Rag-1^-/- mice that had undergone adoptive transfer of ST2^-/- DO11.10 Th2 cells [24, 51, 52]. In order to clarify the role of IL-33/ST2 in lung inflammation, several groups have generated mice deficient in IL-33. Oboki and co-workers demonstrated that 2 sensitizations of IL-33^-/- mice with ovalbumin emulsified in alum showed attenuated eosinophil and lymphocyte recruitment to the lung, airway hyper-responsiveness and inflammation [19]. A similar study by Louten and colleagues has also shown that endogenous IL-33 contributes to airway inflammation and peripheral antigen-specific responses in ovalbumin-induced acute allergic lung inflammation using IL-33^-/- mice [53]. Collectively, the data suggest that IL-33 is involved in lung inflammation and supports the concept of ST2 as a therapeutic target in asthma.

Recent evidence suggests a role for IL-33/ST2 in several rheumatological diseases, including rheumatoid arthritis (RA), osteoarthritis (OA), psoriatic arthritis (PsA) and systemic lupus erythematosus (SLE). The first study to link IL-33 expression with arthritis utilized in situ hybridization to show that IL-33 mRNA expression in the RA synovium is primarily in endothelial cells [11]. Subsequently, IL-33 protein has been found in endothelial cells of synovial tissue and in cells morphologically consistent with synovial fibroblasts in a subset of RA, PsA and OA patients [54]. IL-33 is also expressed in cultured synovial fibroblasts from patients with RA and expression was markedly elevated in vitro by inflammatory cytokines [55, 56]. Circulating IL-33 protein has also been detected in 94/223 RA patient serum samples by ELISA, but was completely absent in healthy controls or OA samples [57]. Furthermore, the level of serum IL-33 decreased after anti-TNF treatment and correlated with production of IgM and RA-related autoantibodies including Rheumatoid Factor and anti-citrullinated protein antibodies. Serum and synovial fluid levels of IL-33 have also been shown to decrease in patients who respond to anti-TNF treatment, while they did not change in non-responders [58]. Similarly, Talabot-Ayer and co-workers show that serum and synovial fluid IL-33 levels were higher in RA than in OA patients, and undetectable in PsA serum and synovial fluid [54]. Another study has demonstrated that neutrophils from patients with RA successfully treated with anti-TNF treatment expressed significantly lower levels of ST2 than patients treated with methotrexate alone [59]. In SLE, one study has shown serum IL-33 levels were significantly increased, compared with healthy controls, but to a lower extent than in patients with RA [60]. The other study reported no change in serum IL-33 levels between controls and SLE patients, but did report a significant increase in sST2 that correlated with SLE disease activity [61].

In murine models of RA, IL-33 mRNA has also been detected in the joints of mice undergoing collagen-induced arthritis (CIA) [56], and in mouse knee joints injected with methylated bovine serum albumin [59]. Furthermore, ST2^-/- mice developed attenuated CIA and reduced ex vivo collagen-specific induction of pro-inflammatory cytokines (IL-17, TNFα, and IFNγ), and antibody production [55]. Conversely, treatment with IL-33 exacerbated CIA and elevated production of both pro-inflammatory cytokines and anti-collagen antibodies through a mast cell-dependent pathway. Administration of blocking anti-ST2 antibodies at the onset of CIA also attenuated the severity of disease and reduced joint destruction [56]. This was also associated with reduced IFNγ and IL-17 production. In a model of anti-glucose-6-phosphate isomerase autoantibody-induced arthritis, IL-33 treatment exacerbated disease. Conversely, ST2^-/- mice were protected against disease and had reduced expression of articular pro-inflammatory cytokines [62]. The IL-33 effects in this model also appear to be mast cell-dependent as IL-33 failed to increase the severity of the disease in mast cell-deficient mice, and mast cells from wild-type, but not ST2^-/- mice restored the ability of ST2^-/- recipients to respond. IL-33 has also been shown to chemoattract neutrophils to a knee joint injected with methylated bovine serum albumin [59].

Various rheumatological diseases can have effects on bone including erosion (e.g. RA) and ossification and the formation of new bone (e.g., ankylosing spondylitis and OA). Recently, the role of IL-33 in bone metabolism and remodeling has been studied with conflicting results. Bone structure and metabolism are determined by the formation and activity of osteoclasts and osteoblasts. Mun and co-workers showed that IL-33 can stimulate the formation of multi-nuclear osteoclasts from monocytes, and enhanced expression of osteoclast differentiation factors including TRAF6, nuclear factor of activated T cells cytoplasmic 1, c-Fos, c-Src, cathepsin K, and calcitonin receptor [63]. However, in contrast two other studies have shown that IL-33 completely abolished the generation of multinucleated osteoclasts [64] or had no direct effect [65, 66].

IL-33 also appears to have direct effects on osteoblast cells. IL-33 expression increases during osteoblast differentiation, and that while ST2^-/- mice displayed normal bone formation they had increased bone resorption, thereby resulting in low trabecular bone mass [64]. Furthermore, IL-33 mRNA levels are increased in osteoblasts following treatment with the bone anabolic factors parathyroid hormone or oncostatin M. In addition, IL-33 treatment promoted matrix mineral deposition by osteoblasts in vitro [65]. However, a recent study reports conflicting data that while IL-33 mRNA is present in human osteoblasts, ST2L is not constitutively expressed and IL-33 treatment has no effect on these cells [66]. The reasons for these differences in the biology of IL-33 in osteoclasts and osteoblasts are unclear at present but may reflect different cell culture conditions and differentiation protocols used. In summary, IL-33 appears to have pro-inflammatory effects in various rheumatological diseases activating synovial fibroblasts and mast cells within joints.

Skin and activated dermal fibroblasts contain a high level of IL-33 mRNA expression compared to other tissues and cell types [3]. IL-33 mRNA and protein is also substantially higher in the skin lesions of patients with atopic dermatitis compared with non-inflamed skin samples [67], and in affected psoriatic skin compared to healthy skin [68, 69]. Elevated serum IL-33 levels have also been detected in patients with systemic sclerosis, and levels correlated positively with the extent of skin sclerosis [70]. Furthermore, subcutaneous administration of IL-33 can induce IL-13-dependent fibrosis of skin in murine models [71]. Recently, it was shown that ST2^-/- mice exhibited reduced cutaneous inflammatory responses compared to WT mice in a phorbol ester-induced model of skin inflammation [69]. Furthermore, intradermal injections of IL-33 into the ears of mice induced a psoriasis-like inflammatory lesion that was partially dependent on mast cells.

In addition, IL-33 expression was induced in pericytes in an experimental model of wound healing in rat skin [72]. Surprisingly, IL-33 has also been shown to induce cutaneous hypernociception in mice, a phenomenon traditionally associated with Th1 responses [73]. Collectively, these results demonstrate that IL-33 may play a role in various inflammatory skin disorders (Figure 2).

IBD is a group of chronic inflammatory conditions of the colon and small intestine, including ulcerative colitis (UC) and Crohn's disease, resulting from dysregulated immune responses. Several studies report an upregulation of IL-33 mRNA in human biopsy specimens from untreated or active UC patients compared to controls [72, 74‐77]. The main sites of UC IL-33 expression were myofibroblasts and epithelial cells. Similarly, ST2 transcripts have been detected in mucosa samples from patients with active UC [74, 75]. However, although Carriere and co-workers demonstrated expression of IL-33 in endothelial cells of Crohn's disease intenstine [11], subsequent studies have failed to demonstrate a significant role for IL-33 in Crohn's disease [72, 74, 76]. Serum IL-33 and sST2 levels were elevated in UC patients compared with controls, while anti-TNF treatment decreased circulating IL-33 and increased sST2, thus favorably altering the ratio of the cytokine with its decoy receptor [74]. However, in other studies serum concentrations of IL-33 were low or did not differ between UC patients and healthy controls [75, 78].

Several murine studies highlight a role for IL-33 in innate-type immunity in the gut. Mice treated with IL-33 displayed epithelial hyperplasia and eosinophil/neutrophil infiltration in the colonic mucosa [3]. Furthermore, in a murine model of T-cell independent dextran sodium sulphate (DSS)-induced colitis IL-33^-/- mice had enhanced viability, compared to wild-type controls [19]. In a related study macrophage-specific transgenic mice that express a truncated TGF-β receptor II under control of the CD68 promoter (CD68TGF-βDNRII) and subjected to the DSS model of colitis display an impaired ability to resolve colitic inflammation but also an increase in IL-33⁺ macrophages compared to controls [79]. In addition, IL-33 mRNA is upregulated in the ilea and correlates with disease severity in a murine model of Th1/Th2-mediated enteritis, and induced IL-17 production from mesenteric lymph node cells stimulated ex vivo [74]. In summary, the IL-33/ST2 pathway may be an important regulator of UC, but be of less importance in Crohn's disease.

Basal IL-33 mRNA levels are extremely high in the brain and spinal cord [3], and are elevated under conditions such as experimental subarachnoid hemorrhage [1]. Furthermore, expression of IL-33 in glial and astrocyte cultures is increased by Toll-like receptor ligands [80]. Treatment with IL-33 induces proliferation of microglia and enhances production of pro-inflammatory cytokines, such as IL-1β and TNFα, as well as the anti-inflammatory cytokine IL-10 [81]. It also enhances chemokines and nitric oxide production and phagocytosis by microglia. In mice, IL-33 levels and activity were increased in brains infected with the neurotropic virus Theiler's murine encephalomyelitis virus [80]. Finally, a transcriptional analysis of brain tissue from patients with Alzheimer's disease revealed that IL-33 expression was decreased compared to control tissues [82]. This study also demonstrated that 3 polymorphisms within the IL-33 gene resulting in a protective haplotype were associated with risk of Alzheimer's disease [82]. This data is supported by a study in Chinese population with evidence that genetic variants of IL-33 affect susceptibility to Alzheimer's disease [83]. Furthermore, cell-based assays demonstrate that IL-33 can decrease secretion of β-amyloid peptides [82]. Thus, IL-33 may have a role in regulating pathophysiology and inflammatory responses in the CNS.

IL-33 was initially found in the nucleus of the high endothelial venules (HEV) of secondary lymphoid tissues [2]. More recently, IL-33 expression has been reported in coronary artery smooth muscle cells [3], coronary artery endothelium [89], non-HEV endothelial cells [88, 90], adipocytes [66, 91], and in cardiac fibroblasts suggesting that IL-33 may play a role in various CV disorders [92].