Tumor necrosis factor α–induced proteins: Natural brakes on inflammation


One of the principal tasks of our immune system is to protect our body against pathogens. Of equal importance, however, is the capacity to preserve the integrity of the host by protecting against autoimmunity. To accomplish this, a number of regulatory pathways have been revealed. They include certain regulatory T cell subsets, such as FoxP3+ Treg cells, natural killer T cells, and Tr1 cells, as well as other cell types, such as certain macrophage subsets. While some of these occur naturally as the immune system develops, interestingly, certain subsets may be actively induced in adults. This has sparked interest in their potential use in therapeutic approaches. Also at the cellular level, a number of proteins with antiinflammatory capacity have been discovered, some of which are induced in response to the proinflammatory cytokine tumor necrosis factor (TNF).

TNF operates both upstream and downstream of diverse signaling cascades, and numerous proteins are induced in response to TNF. In recent years, some of these so-called TNFα-induced proteins (TNFAIPs) have gained a lot of interest. One of these proteins is A20, or TNFAIP-3, a deubiquitinating protein that negatively regulates NF-κB–dependent gene expression in response to different immune-activating stimuli, including TNF, interleukin-1 (IL-1), and in response to triggering of Toll-like receptors (TLRs) and nucleotide-binding oligomerization domain–containing protein 2 (NOD-2) receptor 1. A20 is considered to inhibit NF-κB function by deubiquitinating specific NF-κB signaling molecules, such as receptor-interacting protein 1 (RIP-1), RIP-2, and TNF receptor–associated factor 6 (TRAF-6). Several single-nucleotide polymorphisms in the human TNFAIP-3 locus have been shown to be associated with increased susceptibility to rheumatoid arthritis, type 1 diabetes mellitus, systemic lupus erythematosus, celiac disease, Crohn's disease, psoriasis, and multiple sclerosis (for review, see ref.1).

Another TNF-induced protein, TNFAIP-5, or pentraxin 3, is expressed in rheumatoid arthritis synovial tissue and in astrocytes from patients with Alzheimer's disease (2, 3), while the apoptosis regulator TNFAIP-8 contains a death-effector domain and is capable of inhibiting caspase-mediated apoptosis (4). Knock-down of TNFAIP-8 expression in tumor cells decreases their oncogenicity, which suggests that it may be involved in carcinogenesis (5). TNFAIP-6 (TNFα-stimulated gene 6 [TSG-6]), a protein involved in cell–cell and cell–matrix interactions, is expressed within the synovium and cartilage of arthritic joints and is thought to play a protective role in articular inflammation (6). It further inhibits RANKL-induced osteoclast differentiation and activation (7).

Figure 1.

Binding of tumor necrosis factor (TNF) to its receptor (TNF receptor type I [TNFRI]) induces signaling pathways that lead to apoptosis or to the induction of inflammatory genes. TNF-induced protein A20 operates as a negative regulator to dampen NF-κB signaling, whereas TNFα-induced protein 8 (TNFAIP-8) inhibits caspase-mediated apoptosis. The precise biologic function of TNFα-induced adipose-related protein (TIARP) is largely unknown, although its localization at the plasma membrane and its association with caveolae suggest that it could be involved in intercellular communication or trafficking. Encircled P at IKK represents phosphorylation. FADD = Fas-associated death domain; casp 8 = caspase 8; DED = death-effector domain; U = ubiquitin.

The important roles of these proteins in the mammalian immune system have been demonstrated in mice, where a deficiency in the proteins generally leads to systemic inflammation. Mice fully deficient in A20 spontaneously develop multiple organ inflammation and cachexia and die within 2 weeks of birth (8). Mice deficient in TNFAIP-8–like 2 (TIPE-2; a member of the TNFAIP-8 family) develop multiple organ inflammation and splenomegaly, are hypersensitive to septic shock, and die prematurely (9).

TNFα-induced adipose-related protein (TIARP), also called 6-transmembrane protein of prostate 2 (STAMP-2) or the human homolog 6-transmembrane epithelial antigen of prostate 4 (STEAP-4), belongs to a family of 4 mammalian proteins that were all originally characterized in the prostate, including STAMP-1 (STEAP-2), STEAP, and STEAP-3 (pHyde) (10, 11). The expression of TIARP/STAMP-2 dramatically increases during TNFα exposure, and yet also during the course of adipose differentiation, as demonstrated in the 3T3-L1 preadipose cell line (11). Tissue distribution of TIARP messenger RNA is not restricted to white and brown adipose tissues, but is also detectable in liver, kidney, heart, and skeletal muscle.

Since TIARP/STEAP-4/STAMP-2 is strongly expressed during adipose differentiation, a metabolic function could be expected. Adipose tissue STAMP-2 expression was found to be reduced in obese animals deficient in TNFα function, demonstrating that STAMP-2 is regulated by TNFα in vivo (12). STAMP-2–deficient mice accumulated more subcutaneous fat, had higher blood glucose concentrations, hyperinsulinemia, and dyslipidemia. In addition to the involvement of STAMP-2 in glucose and lipid metabolism, visceral white adipose tissue from STAMP-2−/− mice was shown to be infiltrated by F4/80+ macrophages, with significantly elevated systemic levels of IL-6, TNF, and monocyte chemotactic protein 1 (MCP-1). These findings indicate that STAMP-2/TIARP orchestrates inflammatory responses and metabolic balances in adipocytes and may be essential for the maintenance of systemic metabolic homeostasis.

In support of the role of TIARP in inflammation, Inoue et al (13) describe in this issue of Arthritis & Rheumatism the development of spontaneous enthesitis and synovitis in TIARP-deficient mice. In addition to an erosive joint inflammation, which affects ∼76% of animals by 12 months of age, and consistent with previous reports (12), the white adipose tissue in these mice also showed slight cell infiltration, whereas other organs remained unaffected.

Those authors previously showed that TIARP expression in glucose-6-phosphate isomerase–induced arthritis was significantly higher and was mainly produced in the hyperplastic synovium and in CD11b+ splenocytes (14). Furthermore, in patients with rheumatoid arthritis, the human TIARP homolog STEAP-4 was found to be up-regulated in CD68+ macrophages within joints as well as on peripheral blood monocytes and neutrophils (14, 15).

Inoue and colleagues further report that the spleens of TIARP−/− mice were enriched in CD11b+Gr1low/intermediate macrophages and that circulating levels of IL-6 were higher than those in control mice, whereas no TNF or granulocyte–macrophage colony-stimulating factor was detected. Macrophages from TIARP−/− mice were highly proliferative in response to TNF due to a weak negative regulation of NF-κB signaling and dysregulated apoptosis. As such, macrophages seem to play a pivotal role in the development of enthesitis/synovitis in this model.

The dominant role of macrophages in animal models of arthritis was recently also shown by Matmati et al (16). As mentioned above, mice completely deficient in A20 develop multiple organ inflammation and die prematurely. Interestingly, myeloid cell–specific A20 deficiency provoked an erosive polyarthritis in mice, while other organs remained largely unaffected. Similar to macrophages from TIARP−/− mice, macrophages from myeloid A20–deficient mice displayed sustained degradation of IκBα. Arthritis in these animals developed independently of TNF, B cells, and T cells, but was crucially dependent on TLR-4 signals and was driven by IL-6 (16).

Inoue and colleagues also triggered collagen-induced arthritis (CIA) in TIARP−/− mice and found an exacerbation of the arthritis, accompanied by macrophage and neutrophil infiltration in the joints. Macrophages (thioglycolate-induced peritoneal macrophages) showed increased IL-6–induced STAT-3 phosphorylation. They also describe higher Th1 cell and Th17 cell responses in TIARP−/− mice (by interferon-γ and IL-17 production). Finally, they report that treatment with anti–IL-6 receptor monoclonal antibodies during the induction phase prevented the development of CIA in TIARP−/− mice, while this treatment was less effective in wild-type mice. In contrast, TNF blockade was not able to suppress the progression of arthritis.

A particularly interesting feature of arthritis in the absence of either TIARP or A20 is the lack of dependency on TNF, whereas IL-6 blockade ameliorates both diseases. Thus, while TNF blockade is capable of disease attenuation in CIA (17, 18), this treatment was rather ineffective in TIARP-deficient mice. Previous studies indicated that IL-6 neutralization during the induction phase of CIA was shown to be overall ineffective for the prevention of arthritis (19). However, in TIARP−/− mice, treatment with IL-6–blocking antibodies suppressed both the incidence and the progression of CIA. These findings suggest slightly different kinetics of CIA in the absence of TIARP, although the genetic background of the mouse strains may also play a role (susceptible DBA/1 or B6 mice).

The results reported by Inoue et al (13) add to the growing evidence that TNF-induced proteins, such as TIARP, operate as essential feedback loops to dampen TNF-associated inflammation. However, further research is needed to explore the underlying mechanisms that lead to spontaneous joint inflammation in TIARP−/− mice. It is tantalizing to speculate that macrophages and TNF-induced proteins within myeloid cells are the key regulators in this process. This would be consistent with previous observations that other TNF-induced proteins, such as A20 in myeloid cells, play a crucial role in the prevention of arthritides through the suppression of IL-6 and the negative regulation of NF-κB. Cell type–dependent roles for A20 have been demonstrated to prevent a myriad of inflammatory diseases, including psoriasis (epidermis), lupus (dendritic cells and B cells), spondylarthritis (dendritic cells), and inflammatory bowel disease (intestinal epithelial cells).

The relationship between various TNF-induced proteins and A20 and TIARP in particular, remains to be determined, however (Figure 1). An especially intriguing question is whether the observed effects in the absence of TIARP are A20 dependent or, alternatively, whether they represent distinct regulatory feedback mechanisms. Overall, modulation of TNF-induced proteins such as A20 and TIARP may provide novel treatment options for chronic inflammatory arthritis. However, with respect to TIARP, further research regarding its role in lipid and glucose metabolism is warranted.


Drs. Jacques and Elewaut drafted the article, revised it critically for important intellectual content, and approved the final version to be published.