Tumor necrosis factor (TNF)-receptor-associated periodic fever syndrome (TRAPS) is a rare monogenic autoinflammatory disorder characterized by recurrent attacks of fever and inflammation, predominantly affecting fibroblast-rich tissues such as serosal surfaces, synovial joints, and the deep dermis of the skin (1). The disease is both clinically and genetically heterogeneous, with over 140 mutations reported to date (2). Likewise, a number of molecular mechanisms have been implicated in the pathogenesis of TRAPS, including abnormal TNF-receptor 1 (TNFR1) receptor cleavage (3), ligand-independent activation of mutant TNFR1 (4–6), activation of nuclear factor kappa B (NF-κB)/mitogen-activated protein kinase (6–10), generation of mitochondrial reactive oxygen species (7, 11), and TNFR1 misfolding and retention within the endoplasmic reticulum (ER), leading to activation of the unfolded protein response (UPR) (7). This latter cellular stress response appears to be limited to selective activation of the ER-associated endonuclease, inositol-requiring enzyme 1 (IRE1) (8, 12); however, this activation can have significant proinflammatory effects as shown by Martinon et al. They demonstrated that lipopolysaccharide (LPS) can activate IRE1, via the TLR4 signaling pathway. The spliced X-box binding protein 1 (XBP1) transcription factor, generated via activation of IRE1, can subsequently bind to the promoter regions of TNF and interleukin (IL)-6 (13). In turn, we have shown that inflammatory cytokines, such as TNF, can also activate the IRE1 arm of the UPR resulting in synthesis of XBP1s (14). Therefore, in TRAPS patients, the coexistence of low-level ER stress, with resultant local production of proinflammatory cytokines, can promote chronic activation of IRE1, and subsequent heightened responsiveness of TRAPS cells. These findings corroborate the observations that TRAPS cells are hyperresponsiveness to low-dose LPS, with increased production of proinflammatory cytokines, the release of which leads to clinical manifestations persisting for periods of weeks to months (1, 15, 16).

The traditional role for IRE1, as part of the UPR, relates to the endonuclease function of this enzyme and its ability to target a variety of mRNA and microRNA (miRs) species, and, in this way, limit protein production and help to resolve ER stress (17). IRE1 regulates the expression of mRNA, and miRs via regulated IRE1-dependent decay (RIDD) (18). In this way, IRE1 controls protein exit from the ER, including the levels of proteins that go on to be involved in regulation of ER processes, at both the genetic and epigenetic level. Control of miRs can lead to significant modulation in activity of cellular pathways by determining either cell survival or death (19). miRs, which are small non-coding RNAs that regulate mRNA expression by translational inhibition (20), usually have multiple targets, which may be found on the same and/or different molecular pathways (20).

The co-expression of miR-155 and miR-146 in human monocytes, in response to LPS, was first shown in 2006 (21). Despite evidence suggesting both pro- and anti-inflammatory actions for miR-155, in different contexts, numerous publications have demonstrated that both miR-155 and -146a target a number of downstream signaling pathways involved in toll-like receptor 4 (TLR4)-mediated LPS responses (22, 23), suggesting that, collectively, these miRs regulate a negative-feedback loop to prevent excessive TLR4 activation. In 2011, Schulte et al. suggested a more refined role for these two miRs (24); they used a graded LPS challenge to show that miR-146a was necessary for prevention of TLR4 responses, at sub-inflammatory doses of LPS, which might be relevant to maintaining tolerance to the host’s own microbiome. On the other hand, miR-155 was found to limit TLR4 responses following exposure to higher, proinflammatory doses of LPS. Thus, failure to upregulate these miRs may lead to chronic hyperresponsiveness of the TLR4 pathway.

We therefore hypothesized that the intracellular levels of miR-155 and miR-146a may be reduced in TRAPS cells, possibly due to targeted destruction by IRE1. Furthermore, the proinflammatory milieu of TRAPS cells, particularly due to paracrine effects of TNF and IL-1β, would facilitate this process. We decided to focus on the effects of IL-1β, since this cytokine appears to be critical to the disease pathogenesis and clinical manifestations of TRAPS, and also because anti-IL-1 therapy is now the treatment of choice for this condition (25). Furthermore, we elected to study these effects in dermal fibroblast (DF) from three TRAPS patients, who harbored three different TNFRSF1A mutations, since many of the clinical manifestations of TRAPS are localized to fibroblast-rich tissues, and we wanted to show that this is not a mutation-specific phenomenon.

In this study, we show, for the first time, that TRAPS DF with T50M, C472, and C88R mutations in TNFRSF1A are hyperresponsive to LPS. We were also able to induce a similar phenotype in healthy DF following stimulation with IL-1β, which was partially reversed using the IRE1 inhibitor 4μ8C. In both cases, IRE1 activation was associated with significantly reduced levels of miR-146a and miR-155 in the cells under study. Finally, we have demonstrated the presence of global miR dysregulation in TRAPS patients, which, alongside the downregulation of miR-146a and miR-155, contributes to the proinflammatory phenotype observed in TRAPS T50M, C472, and C88R genotypes.

Lucherini et al. reported a 6-miR signature to be diagnostic of TRAPS (30); however, to our knowledge, no studies to date have suggested a functional role of miRs in TRAPS pathogenesis. Evidence to date unequivocally points toward important role for miR-146a in numerous inflammatory diseases (22). In the case of miR-155, both pro- and anti-inflammatory functions have been reported (31, 32). The expression of miR-155 does increase in a gradual fashion, similar to most other NF-κB-dependent genes in response to environmental stimuli. This may imply a role for miR-155 in propagating inflammatory responses, particularly since it has been shown to stimulate TNF production by stabilizing its mRNA (33). Furthermore, miR-155 inhibits the translation of the anti-inflammatory molecule, SHIP-1 (31) [reviewed in Ref. (22, 34, 35)]. However, at the same time, miR-155 was shown to suppress some proinflammatory components of the TLR4 pathway, such as TAB 2 and IKKε (33, 36), which suggests a more complex role for this miR in regulation of the inflammatory response. Schulte et al. proposed a scenario whereby, on the one hand, miR-155 may induce negative regulation of autocrine signaling whilst, at the same time, promoting paracrine signaling via TNF (24). The end result could activate bystander macrophages, which is essential to deal with infection, but, at the same time, miR-155 may limit the magnitude of the inflammatory response by preventing hyper-activation of the TNF-producing cells. Our data would support such a model. However, to what extent these differential roles of miR-155, or the relationship between miR-146a and miR-155, contribute to specific disease pathogenesis remains to be determined.

Our data also highlight the importance of LPS-hyperresponsiveness in the pathogenesis of TRAPS T50M, C472, and C88R mutations. Our group and others have previously reported this phenomenon in TRAPS, as well as in other inflammatory diseases (37). We have found equivalent levels of surface expression of TLR4 between HC and TRAPS, suggesting that the LPS-hyperresponsiveness observed was not mediated by differences in TLR4 surface expression. Interestingly, TNFR1 knockout mice are resistant to LPS-induced shock, and human monocytes that are deficient in TLR4 are also resistant to LPS (6, 38). Thus, mutant TNFR1 may act in synergy with the TLR4 signaling pathway, triggering increased cellular stress, and consequent lowering of the activation threshold for TLR4-LPS signaling which, in turn, further augments intracellular stress levels in DF with T50M, C472, and C88R mutations in TNFRSF1A. In addition, LPS is expressed at the cell surface of gram-negative bacteria, therefore in the LPS-hyperresponsive state, it may be possible for commensal, or subclinical bacterial loads, to trigger an inflammatory response.

Although TRAPS is associated with mutations in TNFRSF1A it seems that IL-1β is also critical to the disease pathogenesis. Previous data published by our group, as well as others, have reported high levels of IL-1β in the serum of a number of TRAPS patients (9, 39). Furthermore, anti-IL-1 targeted therapy is now emerging as the treatment of choice for this condition (1, 40–44). Our findings suggest that the role exercised by IL-1β in the pathogenesis of TRAPS is, in part, due to its ability to downregulate miR-146a and miR-155. Furthermore, we propose that the effect could be mediated by the IRE1 branch of the UPR, since the IRE1 inhibitor 4u8C partially reversed inhibition of these miRs. We, and others, have shown that proinflammatory cytokines can induce IRE1 activation; for example, it has been shown that both IL-1β and IL-6 are able to induce ER stress in pancreatic islet cells (45). It is also possible that the endoribonuclease activity of IRE1 may only be required for the initiation of IL-1β production, after which inflammatory cytokine production is driven by the autocrine action of proinflammatory cytokines themselves, particularly given that activation of IRE1 is associated with the induction of apoptosis and cell death (18). For example, IRE1 cleaves TRAF2, activating the JNK signaling pathway, which is important in NF-κB activation and proinflammatory cytokine production (46), and, also, TRAF2 is a recognized target of miR-146a (20). Finally, it has been shown that IRE1 and TNFR1 can combine to form an intracellular molecular complex during cellular stress, leading to UPR activation (47). This plethora of supporting evidence shows that IRE1 plays a significant and complex role in cellular stress responses, in both health and disease.

Thus, in summary, overall our data suggest a model whereby TNFR1 protein misfolding within T50M, C472, and C88R mutations, which previously has been shown to occur in TRAPS cells (7, 8) activates the IRE1 branch of the UPR. IRE1 activation leads to cleavage of miR-146a and miR-155, and consequently results in hyperresponsiveness to LPS and increased proinflammatory cytokine production, such as IL-1β. Subsequently, the paracrine effects of these cytokines as well as the proinflammatory local environment observed in TRAPS maintain the UPR activation state and degradation of miR-146a and miR-155.

All patients provided written informed consent for this Leeds (East) Research Ethics Committee-approved study (REC: 04/Q1206/107).

SH, TS, LO, and CW carried out experimental work. SS, MD, and MW provided the patients, designed the study. All contributed toward writing of the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.