Treatment of inflammatory disorders by blocking IL-1

Interleukin-1β (IL-1β) is one of the most potent pyrogens, inducing fever even at picomolar doses. In hereditary fever syndromes increased local and systemic IL-1β was observed years ago.1 These rare diseases include familial cold autoinflammatory syndrome, Muckle–Wells syndrome and the most severe manifestation, chronic infantile cutaneous neurological articular syndrome (also called neonatal-onset multisystem inflammatory disease (NOMID)). All of them are clinically defined by recurrent episodes of fever with skin rashes. However, most patients also have arthritis and possess increased inflammatory markers such as C-reactive protein or serum amyloid A. Patients of the most severe form (NOMID) often have meningitis, conjunctivitis and polyserositis, affecting the lung or joints. In the past decade, it became apparent that these periodic fever syndromes are associated with gain of function mutations in NLRP3, providing a probable biochemical mechanism for increased IL-1β levels.2 ,3 Over the past decade the clinical management of periodic fever syndromes has been revolutionised by the use of anakinra, the recombinant form of naturally occurring IL-1 receptor antagonist (IL-1Ra). Since then numerous inflammatory conditions have been shown to be responsive to anakinra (see table 1). Although the pathogenesis of many of the diseases remain unclear, anakinra has been used successfully in metabolic diseases such as gout,15 type II diabetes14 or in inflammatory diseases such as pyogenic acne,19 and even autoimmune diseases such as rheumatoid arthritis20 (table 1).

However, anakinra, which needs to be injected several times per week, often induces localised inflammatory responses at the site of injection. To overcome these adverse effects, two new biological agents targeting the IL-1 pathway were introduced into clinics—namely, rilonacept and canakinumab. Rilonacept (IL-1 Trap) is a decoy receptor for IL-1, inhibiting both IL-1α and IL-1β signalling, while canakinumab is a humanised monoclonal antibody selectively binding to IL-1β. These two drugs have been successful in treating the periodic fever syndromes mentioned above.

While Canakinumab is highly efficient in the treatment of periodic fever syndromes,4 clinical studies are currently underway to test its efficacy in other putative IL-1-mediated diseases. Results of the Canakinumab Anti-inflammatory Thrombosis Outcomes Study will help us to decipher the effect of anti-IL-1β treatment in metabolic diseases. In this multinational study, 17 200 patients with postmyocardial infarction will be treated with canakinumab or placebo. After 4 years of cardiovascular events both in the arterial and venous system, metabolic parameters will be evaluated to monitor the efficacy of anti-IL-1β treatment on secondary prevention of cardiovascular diseases and diabetes. Preliminary results of the MRC-ILA-HEART Study, however, could not demonstrate any benefit of anakinra treatment for the reduction of inflammation in acute coronary syndromes.21

Properties and regulation of IL-1α and IL-1β

As described above, the successful treatment of inflammatory diseases by inhibiting IL-1 signalling was in most cases achieved by blocking the IL-1R1 using anakinra. This receptor, however, is activated by two distinct ligands—namely, IL-1β and its close homologue IL-1α. These cytokines share similar functional characteristics as they are potent pyrogens, strong costimulators of T cells and highly proinflammatory.22 In addition, IL-1α and IL-1β share a similar structure. Neither contains a secretion sequence, suggesting a non-canonical secretion mechanism, which remains elusive. In contrast to IL-1β, which requires proteolytic processing to become active, IL-1α is biologically active in its intact, full-length form. The cleavage of pro-IL-1β is achieved through the formation of a large protein complex, termed the inflammasome. The formation of the inflammasome results in the activation of caspase-1, which in turn can process IL-1β and other substrates.23 To date, four distinct inflammasomes have been described either consisting of NLRP3, NLRP1, NLRC4 or AIM2.24 ,25 Most of these inflammasomes, at least partially, require the adaptor protein, ASC (apoptosis-associated speck-like protein containing a caspase recruitment domain), to recruit caspase-1 to the inflammasome complex. Upon binding to the inflammasome, caspase-1 is cleaved and activated, leading to cleavage and secretion of IL-1β.

Whereas inflammasome-mediated release of IL-1β has been repeatedly shown to represent an important effector mechanism in local and systemic inflammatory responses, a role for IL-1α in innate immune responses remains largely uncharacterised. The release of IL-1β is thought to occur through an active secretion mechanism, but most current publications describe IL-1α as a passively released danger signal resulting from non-apoptotic cell death.26 ,27 As anakinra blocks the function of both of these cytokines, it is impossible to dissect their distinct roles in inflammatory conditions based on the trials mentioned above. Yet, as caspase-1 processes and thereby activates IL-1β, but not IL-1α,28 pathogenesis is often attributed solely to IL-1β.22 The possibility of a direct, inflammasome-mediated contribution of IL-1α has been largely ignored. Moreover, the few studies that provide data on IL-1α and inflammasomes are inconclusive.29–31

Evidence for a role of IL-1α in inflammatory conditions

While data on the role of IL-1α in inflammatory conditions is scarce, current publications do provide some evidence implicating it in driving these conditions. Murine in vivo models often possess milder phenotypes in IL-1β knockout mice than IL-1R1-deficient or even caspase-1-deficient animals.32–37 These observations already point to additional effector(s) acting through IL-1R1. Indeed, studies that investigated the phenotype of IL-1α-deficient or IL-1α/IL-1β double knockout mice do indicate that this additional effector is IL-1α. IL-1R1 signalling has been shown to be required for the clearance of Candida albicans infections. IL-1β and IL-1α knockout mice showed a milder phenotype, as measured by fungi clearance and survival, in comparison with IL-1R1 knockout mice, suggesting that both cytokines are important in Candida albicans immunity.36 Similarly, inflammation induced by nanoparticles, as measured by neutrophil infiltration into the peritoneum, has been shown to depend on IL-1R1 and caspase-1. In contrast, deficiency in IL-1β or IL-1α only resulted in partial reduction of inflammation.31 Finally, in the same model of peritoneal inflammation, cholesterol crystal-induced neutrophil infiltration was dependent on IL-1R1 and caspase-1 and also on IL-1β and IL-1α to a similar extent.38 Intriguingly, all the inducers of inflammation used in these models have been shown to be stimuli of the NLRP3 inflammasome. Given the clear role of IL-1α in these models, it is intriguing to speculate whether the NLRP3 inflammasome might have a direct role in the regulation of its activity. Indeed, two recent studies provide evidence to support the notion that IL-1α secretion is regulated by the NLRP3 inflammasome.39 ,40

Regulation of IL-1α expression and secretion

In myeloid cells, the best-characterised cells for IL-1 secretion, pro-IL-1α and pro-IL-1β, need to be transcriptionally induced. Human keratinocytes, on the other hand, constitutively express IL-1α and IL-1β.39 To induce IL-1α and IL-1β transcription in myeloid cells, an initial signal is required. Typically, Toll-like receptor (TLR) activation by lipopolysaccharide treatment to activate TLR4, or CpG treatment to activate TLR9 is used to induce IL-1 expression. However, the TLR signal alone is not sufficient to evoke the secretion of any of the two IL-1 family members. Until last year, IL-1α was considered to be released passively by dying cells,26 while only IL-1β was considered to be actively secreted in a caspase-1-dependent manner. Recently, we and others39 ,40 demonstrated a controlled release of both IL-1α and IL-1β from myeloid and non-myeloid cells, including keratinocytes.

Fettelschoss et al reported that, upon TLR activation, IL-1α is expressed and localised to the cell surface as an active membrane-bound ligand.40 After inflammasome activation IL-1α is then actively secreted. The authors hypothesised that the secretion of IL-1α depends on the presence of IL-1β. According to their model, IL-1α directly binds to IL-1β and uses it as a shuttle for cosecretion.

We, on the other hand, identified two distinct pathways for IL-1α cleavage and secretion. Depending on the NLR activator, cleavage and secretion of IL-1α can be either caspase-1 dependent or independent. In general, particles that are phagocytosed by cells induce caspase-1 independent IL-1α secretion, while soluble NLRP3 agonists such as nigericin or ATP require caspase-1, ASC and NLRP3 for IL-1α release. In both cases IL-1α processing is mediated by Ca²⁺-dependent, calpain-like proteases. The activation of calpain-like proteases can be downstream of the inflammasome (in the case of soluble NLRP3 stimuli) or independent of inflammasome activity (in the case of particles). Surprisingly, although the presence of caspase-1 is critical for IL-1α secretion upon treatment of soluble NLRP3 stimuli, the catalytic activity of caspase-1 is not required. IL-1α lacks a caspase-1 cleavage site and its processing is not influenced by pan-caspase inhibitors. Nonetheless, caspase-1 knockout cells are unable to secrete IL-1α in response to soluble NLRP3 stimuli while caspase inhibitors have no effect. This controversy might be explained by a protease-independent function of caspase-1, possibly regulating the active release of IL-1α. In our conditions, IL-1α secretion remains unaltered in IL-1β knockout mice, contrary to the proposed function of IL-1β as a shuttle for IL-1α secretion. Both publications do, however, independently describe controlled release of cleaved IL-1α. Moreover, in the different in vivo settings used, IL-1α seems to be more important for inflammatory responses than the inflammasome proteins, NLRP3 and the adaptor ASC, implicating additional pathways that mediate IL-1α secretion.31 ,39 ,40