A clinical perspective of IL-1β as the gatekeeper of inflammation

Authors

  • Charles A. Dinarello

    Corresponding author
    1. Department of Medicine, University of Colorado, Aurora, CO, USA
    2. Department of Medicine, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands
    • University of Colorado Denver, Division of Infectious Diseases, B168, 12700 East 19th Ave., Aurora, CO 80045, USA
    Search for more papers by this author

Abstract

An expanding spectrum of acute and chronic non-infectious inflammatory diseases is uniquely responsive to IL-1β neutralization. IL-1β-mediated diseases are often called “auto-inflammatory” and the dominant finding is the release of the active form of IL-1β driven by endogenous molecules acting on the monocyte/macrophage. IL-1β activity is tightly controlled and requires the conversion of the primary transcript, the inactive IL-1β precursor, to the active cytokine by limited proteolysis. Limited proteolysis can take place extracellularly by serine proteases, released in particular by infiltrating neutrophils or intracellularly by the cysteine protease caspase-1. Therefore, blocking IL-1β resolves inflammation regardless of how the cytokine is released from the cell or how the precursor is cleaved. Endogenous stimulants such as oxidized fatty acids and lipoproteins, high glucose concentrations, uric acid crystals, activated complement, contents of necrotic cells, and cytokines, particularly IL-1 itself, induce the synthesis of the inactive IL-1β precursor, which awaits processing to the active form. Although bursts of IL-1β precipitate acute attacks of systemic or local inflammation, IL-1β also contributes to several chronic diseases. For example, ischemic injury, such as myocardial infarction or stroke, causes acute and extensive damage, and slowly progressive inflammatory processes take place in atherosclerosis, type 2 diabetes, osteoarthritis and smoldering myeloma. Evidence for the involvement of IL-1β and the clinical results of reducing IL-1β activity in this broad spectrum of inflammatory diseases are the focus of this review.

Historical perspective

IL-1 has a long history 1; it begins with interest in the most salient manifestation of inflammation, fever. Indeed, the discovery of IL-1 as the quintessential inflammatory cytokine can be traced to the purification of the endogenous fever-producing molecule, leukocytic pyrogen, in 1977 2. Interest in this molecule increased when we reported that leukocytic pyrogen was the same molecule as lymphocyte activating factor 3, thus necessitating invention of the IL nomenclature. The term for IL-1 was assigned to the macrophage product and IL-2 for the T-cell product, even though there was no N-terminal amino acid sequence at that time that these were indeed different molecules. By the early 1980s, it was hypothesized, following on from experiments using purified preparations of IL-1 from human blood monocytes and other preparations by the laboratories of Jean-Michel Dayer and Joost Oppenheim, that IL-1 possessed multiple and diverse properties such that it mediated the acute phase response to infection, injury and immune challenge 4. The concept that IL-1 possessed these seemingly unrelated properties was diagramed in 1984 (4 and Fig. 1), without the benefit of recombinant IL-1 to validate the concept. The scientific community, being skeptical of the concept that a single small protein could have such a spectrum of activities, demanded confirmation with recombinant IL-1.

Figure 1.

This figure was published in 1984 4 and is reprinted with permission. It was published on the eve of the isolation of the first cDNA for IL-1β in 1984 (reviewed in 1).

Following the isolation of the cDNA for IL-1α 5 and IL-1β 6 in 1984, studies using the recombinant forms confirmed the growing list of inflammatory properties of IL-1. Indeed, recombinant IL-1α or IL-1β provided ample evidence for the broad role of IL-1 in health as well as disease (Fig. 2) The availability of recombinant forms also allowed for the development specific assays such as radioimmunoassays and later ELISAs. These assays changed how many viewed cytokines since the immunoassays liberated the investigator from the non-specific bioassays that had dominated and confused the field for 20 years. The specific assays now told another story and that was the ability to follow a disease process or a therapy in terms of changes in cytokine levels. However, the greatest contributions of the recombinant forms of IL-1 were the responses they triggered upon administration to humans. Cancer patients undergoing bone marrow transplantation were injected with either IL-1α or IL-1β to stimulate hematopoiesis Table 1 summarizes the human responses observed, and physiologic responses such as fever following injection of 10 ng/Kg IL-1α or IL-1β match those observed using purified human leukocytic pyrogen injected into rabbits in 1977 2.

Figure 2.

A summary of the key in vitro and in vivo effects of recombinant IL-1α or IL-1β that have been discovered to date; the summary is not exhaustive Abbreviations: REM sleep, rapid eye movement sleep; PAF, platelet activating factor; COX-2, type 2 cyclooxygenase; PLA2, phspholipase-A2; PAI-1. plasminogen activating factor inhibitor-1; ICAM-1-1, intracellular adhesion molecule-1; VEGF, vascular endothelial cell growth factor; G-CSF, granulocyte colony stimulating factor; GM-CSF, granulocyte-macrophage colony stimulating factor.

Table 1. Human responses to intravenous IL-1α or IL-1βa)
ResponseConcentration IL-α/β required
  • a)

    a) These responses were reviewed in 104. Data on human responses to IL-1α or IL-1β can also be found in 105–113.

Fever(1–10 ng/Kg)
Myalgias and joint pain(1–10 ng/Kg)
Generalized fatigue(1–10 ng/Kg)
Headache; nausea(1–10 ng/Kg)
Neutrophilia(1–10 ng/Kg)
Thrombocytosis(1–10 ng/Kg)
GM-CSF, G-CSF(1–10 ng/Kg)
IL-6, IL-8, TNFα(1–10 ng/Kg)
IL-1Ra and soluble TNFα receptor(130 ng/Kg)
ACTH, cortisol(1–10 ng/Kg)
Gastrointestinal discomfort(>30 ng/Kg)
Nitric oxide (serum)(3 ng/Kg)
Hypotension(30 ng/Kg)
Hemodynamic shock(300 ng/Kg)

Next in the history of IL-1 was the identification of the naturally occurring and specific inhibitor of IL-1 activity 7–9, later found to be the IL-1 receptor antagonist (IL-1Ra). IL-1Ra was developed into a therapeutic (anakinra) and tested in humans. Anakinra is a pure receptor antagonist binding tightly to the type I IL-1 receptor (IL-1RI) and preventing activation of this receptor by either IL-1β or IL-1α. Approved for treating patients with rheumatoid arthritis, the use of anakinra validated the importance of IL-1 in a broad spectrum of inflammatory diseases. More recently, soluble receptors for IL-1 (rilonacept) and human mAbs to IL-1β (canakinumab and Xoma 052) have been used to neutralize IL-1β specifically. In most reports, summarized in Table 2, there is a dramatic, rapid and sustained improvement in patients following a reduction in IL-1β activity. Thus, from clinical studies using IL-1β neutralization, one concludes that this cytokine should be considered a gatekeeper of inflammation.

Table 2. Blocking IL-1β in treatment of acute and chronic inflammatory diseases
DiseaseReferences
  • a)

    a) Rheumatoid arthritis is also treatable with other anti-cytokines, anti-receptors, immunomodulating agents and B-cell-depleting antibodies.

Classic auto-inflammatory diseases 
FMF14, 114–118
Pyogenic arthritis, pyoderma gangrenosum, acne (PAPA)119–121
CAPS17, 23, 29, 46, 122–126
HIDS127–130
Refractory adult Still's disease60, 131–133
SOJIA22, 134–140
Schnitzler's syndrome141–143
TRAPS11, 12, 144
Blau's syndrome; Sweet's syndrome119
Probable auto-inflammatory diseases 
Recurrent idiopathic pericarditis145, 146
MAS147, 148
Urticarial vasculitis149
Anti-synthetase syndrome (myosistis)150
Relapsing chondritis151
Behçet's disease152, 153
Erdheim-Chester syndrome (histiocytosis)154
Synovitis, acne, pustulosis, hyperostosis and osteitis (SAPHO)155
Common diseases treatable with IL-1β blockade 
Rheumatoid arthritisa) 
Periodic fever, aphthous stomatitis, pharyngitis and adenitidis syndrome (PFAPA)156
Urate crystal arthritis (gout)30, 77, 78, 80, 157, 158
Type 2 diabetes90–92
Smoldering myeloma100
Post myocardial infarction heart failure95
Osteoarthritis69

What is auto-inflammation?

How did this concept start?

The term was first used to describe a rare disease characterized by recurrent bouts of fever and systemic inflammation due to a mutation in the coding region of the p55 TNF-receptor 10. The disease was traditionally called Familial Hibernian Fever but is now called TNF-receptor-associated periodic syndrome or TRAPS. The mutation results in a failure to translocate this receptor to the cell membrane and thus the protein builds up in the endoplasmic reticulum, resulting in a form of “endoplasmic reticulum stress.” Since the inflammation was triggered by an endogenous protein, albeit an abnormal protein due to malfolding, the term “auto-inflammation” was coined. Initially the disease was treated by administration of the soluble TNF-receptor etanercept since, due to the mutation, circulating levels of the soluble receptor are low; however, subsequently the inflammation has been shown to respond to anakinra 11, 12. Thus, TRAPS emerges as an IL-1-mediated disease. In some studies, neutralization of TNF-α with infliximab has worsened the inflammation of TRAPS 13.

The second disease that was considered due to “auto-inflammation” is familial Mediterranean fever (FMF), also characterized by life-long bouts of fever with local and systemic inflammation, is due to a mutation in a protein. The mutation in FMF is found in the intracellular protein called pyrin (reviewed in 14). WT pyrin binds to ASC (apoptosis-associated speck-like protein containing a caspase activation and recruitment domain), an essential component for the activation of caspase-1 and the processing of IL-1β. It is thought that pyrin functions to sequester ASC and prevent its participation in caspase-1 activation; however, mutated pyrin appears to lose part of the ASC binding and, as a result, there is a greater activation of caspase-1 and secretion of IL-1β. Indeed, attacks of FMF are fully prevented by anakinra (see Table 1), although the disease is usually controlled by daily colchicine. However, in patients whose disease is poorly controlled by colchcine, blocking IL-1 rapidly returns the patient to normalcy. The attacks of FMF are seemingly unprovoked, but it is likely that constitutional changes such as stress, viral infections or dietary components trigger the activation of caspase-1 and release of IL-1β.

In 2001, Hal Hoffman described a mutation in a protein in families who experience systemic and local inflammatory responses upon exposure to cold 15. Termed familial cold auto-inflammatory syndrome (FCAS), the mutation was found to be in a protein that Hoffman named cryopyrin (now termed nucleotide-binding domain and leucine-rich repeat containing protein 3 (NLRP3)). Together with ASC, NLRP3 participates in the activation of caspase-1 16. Patients with FCAS are treated with anakinra or the IL-1 soluble receptor rilonacept 17. Two other diseases with mutations in NLRP3 are Muckle–Wells syndrome (MWS), which can also be triggered by exposure to cold, and chronic infantile neurological, cutaneous and articular (CINCA) syndrome (also termed neonatal onset multisystem inflammatory disease, NOMID). Together FCAS, MWS and CINCA are called cryopyrinopathy-associated periodic syndrome (CAPS) and are uniquely IL-1β-mediated diseases. The mAb to IL-1β, canakinumab, is approved for the treatment of CAPS.

Hoffman's discovery of the mutation in NLRP3 led to a series of studies from the laboratory of the late Jürg Tschopp. Taking monocytes from patients with MWS, the Tschopp group demonstrated that the processing and secretion of IL-1β was markedly elevated in comparison with monocytes from healthy individuals and further demonstrated that this was due to oligomerization of intracellular proteins with NLRP3 for the conversion of pro-caspase-1 to active caspase-1 and hence the cleavage of the IL-1β precursor. The complex required NLRP3 and ASC and the mutation was a gain of function mutation for the processing and secretion of active IL-1β. Tschopp named the caspase-1-activating complex “the inflammasome” 16. Mice deficient in NLRP3 or ASC often resisted IL-1β-mediated inflammation similar to that observed in mice deficient in caspase-1. In the present issue of this journal, Jürg Tschopp summarizes his views on the importance of the molecular contribution of mitochondria to the activation of the NLRP3 inflammasome and states that “mapping the connections between mitochondria, metabolism and inflammation is of great interest, as malfunctioning of this network is associated with many chronic inflammatory diseases” 18. One cannot overstate the importance of Jürg Tschopp's contributions for understanding the molecular mechanisms of IL-1β-mediated inflammation and its impact on human disease.

The broadening concept of auto-inflammation

From the above three discoveries, the concept of auto-inflammation emerges as due to gain of function mutations that participate in the activation of caspase-1 and the secretion of active IL-1β. Although one can also consider auto-inflammatory diseases as due to poor control of caspase-1, any non-infectious disease brought under rapid and sustained control with neutralization of IL-1β may be due to endogenous molecules that trigger active IL-1β, regardless of caspase-1 processing. For example, patients with identical disease manifestations in FMF, FCAS, MWS and CINCA who are highly responsive to neutralization of IL-1β and have no mutations in pyrin or NLRP3. Second, another chronic inflammatory disease called hyper IgD syndrome (HIDS) is due to a mutation in mevalonic acid synthesis but patients with HIDS are successfully treated with IL-1β blockade (see Table 1). Third, a growing list of systemic and local diseases are treated by blocking IL-1β activity, but there are no mutations in any component through which caspase-1 activation occurs. However, upon in vitro culture of fresh monocytes from these seemingly unrelated diseases, there is increased release of processed IL-1β 16, 19–23.

The rate-limiting step in auto-inflammation

The rate-limiting step in the release of IL-1β appears to be the translation of the mRNA into the IL-1β precursor. In circulating human blood monocytes, caspase-1 is present in an already active state 24; caspase-1 is also constitutively active in highly metastatic human melanoma cells 25. When the monocyte is stimulated to synthesize the IL-1β precursor, cleavage of the precursor takes place and mature IL-1β is secreted over several hours. Thus, the rate-limiting step for the release of active IL-1β is the synthesis of the IL-1β precursor. In general, the release of active IL-1β from blood monocytes is tightly controlled with less than 20% of the total synthetic IL-1β precursor being processed and released. Although the release of active IL-1β from the blood monocytes of healthy subjects takes place over several hours 24, the process can be accelerated by the exogenous addition of ATP 19, which triggers the P2X7 purinergic receptor 26.

In tissue macrophages, caspase-1 is not constitutively active 24. Extracellular ATP is required to activate the P2X7 receptor, which opens the potassium channel. Simultaneously, intracellular potassium levels fall, caspase-1 is activated, the IL-1β precursor is cleaved and secretion takes place 26. Thus, in ischemic diseases where there is cell death, release of ATP contributes to caspase-1 activation. A similar process may take place in the inflammatory process of gouty arthritis. In this disease, the synovial macrophage is induced to synthesize the IL-1β precursor following exposure to uric acid crystals in combination with free fatty acids 27. In the presence of large numbers of neutrophils, crystal-induced cell death causes the release of ATP and triggering of the P2X7 receptor. In addition, there may be a hypoxic component to the production of IL-1β in gout since the disease characteristically occurs in the most distal joints.

IL-1 can be the “auto” in auto-inflammation

Most human disease is sterile and, in many cases, the release of cell contents upon necrotic death releases the IL-1α precursor. The IL-1α precursor is fully active and does not require caspase-1 processing. Here the concept of auto-inflammation may find its fundamental mechanism, as auto-inflammation needs auto-stimulants. One auto-stimulant is IL-1 itself as IL-1 induces itself 28. The clinical evidence behind this concept can be found in treating patients with the classic auto-inflammatory diseases such as CAPS. For example, the elevated levels of caspase-1 mRNA as well as that of IL-1β in the blood monocytes from the CINCA syndrome patients decreases dramatically with anakinra treatment but rapidly returns with cessation of anakinra 23. In addition, a single administration of an anti-IL-1β mAb results in prolonged resolution of disease activity after the antibody is cleared from the circulation 29. Similar observations have been made in patients treated with a single dose of canakinumab for gout 30. In those studies of IL-1-induced IL-1, IL-1α was used to stimulate gene expression and release of active IL-1β since the IL-1α precursor is constitutively present in all mesenchymal cells. Furthermore, the IL-1α precursor, which unlike the IL-1β precursor, binds to the IL-1 receptor and is active. Not unexpectedly, IL-1α is also the cytokine that has been consistently implicated as causing sterile inflammation due to cell death 31, 32. Upon death of cells, the IL-1α precursor is released and induces myeloid cell-mediated inflammation in vivo 31, 32. In this sense auto-inflammatory diseases are likely to have ischemia as part of the induction of IL-1α 32. IL-1α is also expressed as an integral membrane protein, which is highly active in inducing chemokines from mesenchymal cells 33.

Two signals in auto-inflammation

In addition to IL-1 auto-induction, other endogenous stimulants have been identified. For example, activated complement, uric acid crystals, high concentrations of glucose, cholesterol, and free fatty acids, particularly oxidized free fatty acids, can participate in the production of IL-1β. The role of each of these is discussed below within the context of specific disease processes. Moreover, these endogenous stimulators of IL-1β production often act together. Uric acid crystals alone do not stimulate IL-1β production and neither does free fatty acids but it requires the combination of both 27. In general, translation of the IL-1β precursor requires two signals; one signal is for IL-1β gene expression and the second is for completion of the synthesis of the protein. Without a second signal, polyadenylated IL-1β mRNA falls off the ribosome 34, 35. C5a is generated in most inflammatory conditions and induces marked gene expression for IL-1β but without significant translation. However, a small amount of IL-1α or IL-1β drives the mRNA to complete translation 36.

What are the endogenous mechanisms for the control of IL-1-induced auto-inflammation? The naturally occurring IL-1Ra is clearly essential for controlling IL-1-induced inflammation as deletion of IL-1Ra in mice results in the spontaneous development of a rheumatoid arthritis-like inflammatory joint disease 37 and lethal arthritis 38. In humans, a deletion of IL-1Ra or a mutation that affects the ability of IL-1Ra to inhibit IL-1 results in severe and lethal systemic inflammation at birth 39, 40. IL-1 activity can also be controlled by its own decoy receptor, IL-1R type II, which shunts IL-1β away from the signaling receptor 41. Type I interferon such as interferon-α (IFN-α) is also an endogenous mechanism by which the activity of IL-1β is suppressed and is particularly relevant for auto-inflammation. IL-1α-induced IL-1β gene expression and secretion of processed IL-1β is reduced by 60–95% in the presence of equimolar concentrations of either IFN-α or IFN-γ 42. A report from the laboratory of the late Jürg Tschopp also observed that type I IFN-β reduced the activation of NLRP3 and the maturation of IL-1β 43. In that study, the authors demonstrated that the ability of IFN-β to suppress the maturation of IL-1β was due to the STAT1 transcription factor, which also repressed the activity of the NLRP1 43. Not unexpectedly, IFN-β induced IL-10 in a STAT1-dependent manner; autocrine IL-10 then signaled via STAT3 to reduce the abundance of the IL-1α as well as the IL-1β, precursors. Moreover, blood monocytes from multiple sclerosis patients undergoing IFN-β treatment are found to produce substantially less IL-1β as compared with monocytes from healthy donors 43.

Blocking IL-1 in inflammatory diseases

Treating CAPS

Together, FCAS, MWS and CINCA syndrome are grouped and called CAPS. These syndromes are characterized by recurrent fevers, leukocytosis, elevated acute phase proteins, myalgias and generalized fatigue. CINCA syndrome is a severe form of CAPS beginning in neonatal life. The term “cryopyrin” was coined by Hoffman during his studies regarding the mutation in FCAS 15. Upon exposure to cold, the affected subjects develop fevers, leukocytosis and generalized flu-like symptoms, hence the use of “cryo” for cold and “pyrin” for fever. Blood monocytes from these patients release more IL-1β upon incubation in the cold as compared with monocytes from persons without the mutation 21. CAPS patients treated with either anakinra 23, 44, 45, a soluble IL-1 receptor (rilonacept) 17 or a monoclonal anti-human IL-1β (canakinumab) 29, experience a rapid, sustained and near complete resolution of the disease. Of particular importance is the amelioration of the central nervous system abnormalities in children with CINCA during sustained treatment with anakinra 23 or canakinumab 46.

Treating FMF in colchicine-intolerant patients

Colchicine is routinely used to prevent attacks of FMF 47. Although the mechanism of action of colchicine in FMF is poorly understood, one effect of colchicine is a reduction in the migration of monocytes into an inflamed area 47. Because oral colchicine is converted in the liver to an active compound by p450 cytochrome C, some patients are resistant to colchicine because they harbor a mutation in p450 cytochrome C. As a result, these patients are treated with anakinra. Other patients are intolerant of the loose stools associated with colchicine use. Anakinra brings about a rapid cessation of the local and systemic inflammation of an attack. However, periodic anakinra is effective in preventing FMF attacks when administered early during the prodrome and in some patients daily anakinra is used. Colchicine-resistant FMF disease severity can present as bilateral pneumonia; initiation of anakinra therapy in such patients has been shown to result in a rapid improvement in clinical symptoms as well as radiographic resolution within 2 days 48.

The surprise in treating TRAPS

Since TRAPS was originally believed to be due to a lack of endogenous soluble TNF-α receptor, disease activity was thought to be best controlled by administration of agents that neutralized TNF-α such as etanercept and infliximab. However, TRAPS turns out to be an IL-1β-mediated auto-inflammatory disease and optimally responsive to IL-1β blockade. Blood monocytes from TRAPS patients release IL-1β in greater amounts than cells from healthy subjects 13, a characteristic of auto-inflammatory diseases. In fact, treating patients with TRAPS with infliximab worsened disease severity 13, 49. Another characteristic of patients with auto-inflammatory diseases is the response to reducing IL-1β activity, which is observed in patients who are refractory to corticosteroids, cyclosporine, azathiaprine or colchicine. Whereas anti-TNF-α is often used to treat such refractory patients, the response to TNF-α blockade is modest at best and often not sustained. It is also possible that even a modest response to anti-TNF-α in these patients is due to a reduction in IL-1 activity since TNF-α induces IL-1 50. Not all patients with TRAPS respond to anakinra, and neutralizing antibodies to IL-6 receptor have been effective in reducing disease activity, as reported in a single patient 51.

Rheumatoid arthritis

Several trials have shown the benefit of anakinra in treating the signs and symptoms of rheumatoid arthritis. After one full year of treatment, the reduction in disease severity in patients with rheumatoid arthritis treated with anakinra is comparable to other treatments 52, 53. IL-1 is a potent inhibitor of proteoglycan synthesis in cartilage 54, and joint space narrowing and erosions in patients with rheumatoid arthritis treated with anakinra are clearly improved 52, 55, 56. Moreover, unlike TNF-α blocking therapies, there have been no reports of opportunistic infections, particularly reactivation of Mycobacterium tuberculosis, in patients treated with IL-1β blocking agents. In an analysis of anakinra use in rheumatoid arthritis, Mertens and Singh 57 reviewed five trials involving 2065 anakinra-treated patients compared with 781 patients treated with placebo and reported that there was significant improvement in various clinical and biochemical markers of disease activity as well as in the Larsen radiographic scores of the anakinra-treated patients. The authors concluded that anakinra is a relatively safe and modestly efficacious therapy for rheumatoid arthritis.

Given that anakinra is injected each day and because the first weeks of anakinra injections can cause painful injection site reactions, anakinra is not as popular with patients or with rheumatologists as anti-TNF-α. By comparison, there is widespread use of anti-TNF-α agents in treating rheumatoid arthritis, which is due to both the reduction in joint inflammation as well as the rapid (within a day) reduction in the depressive effects of TNF-α on the central nervous system. For example, with the use of functional magnetic resonance imaging, it can be observed that within 24 h of an intravenous infusion of infliximab, not only is nociceptive central nervous system activity both in the thalamus and somatosensoric cortex, but also activation of the limbic system, blocked 58. These results explain the rapid and sustained feeling of well-being reported by patients receiving anti-TNF-α treatment. The efficacy and safety of anakinra was evaluated in patients with active psoriatic arthritis; anakinra led to an improvement in signs and symptoms in nine out of 19 patients; two patients had an American College of Rheumatology (ACR) score of 70 59 (an American College of Rheumatology score of 70 indicates that the patient has experienced an overall improvement of 70% in disease activity).

Adult onset Still's disease (AOSD)

AOSD, a rheumatologic condition found world-wide, is characterized by a variety of clinical features including intermittent fever, arthritis, evanescent rash, sore throat and leukocytosis, as well as hepatic dysfunction, polyserositis, lymphadenopathy and splenomegaly. The etiology of AOSD remains unknown but viral infection has been suspected in its pathogenesis. Death in association with systemic features such as hepatic failure, amyloidosis, infection and disseminated intravascular coagulation has been reported and progression into macrophage activation syndrome (MAS) is known. Several clinical and biochemical markers of inflammation observed in AOSD are similar to those of the systemic inflammatory response syndrome as fever, neutrophilia and hepatic acute phase protein synthesis are prominent in AOSD. Reducing TNF-α is often without effect whereas anakinra results in a rapid resolution of systemic and local manifestations of the disease within hours and days of the initial subcutaneous injection 60. Reducing IL-1β activity in AOSD is now the standard therapy.

Systemic onset juvenile idiopathic arthritis

Systemic onset juvenile idiopathic arthritis (SOJIA) is thought to be an auto-immune disease and treatable with tocilizumab (anti-IL-6 receptor); however, the disease has the characteristics of an auto-inflammatory disease with increased secretion of IL-1β from blood monocytes and dramatic responses to anakinra or canakinumab in patients resistant to glucocorticoids 22. SOJIA patients usually do not respond to anti-TNF-treatment 22, 61. Gattorno et al. 20 reported heterogeneous responses to IL-1 blockade by anakinra, with approximately one-half of the patients treated with anakinra experiencing rapid improvement whereas the other half exhibited either an incomplete or no response. The responders in that study were characterized by higher absolute neutrophil counts but a lower number of disease-active joints before entering the trial. Thus, it is likely that a more systemic disease predicts a positive response to IL-1 blockade. Indeed, clinical experience reveals that in approximately 50% of SOJIA patients, arthritis tends to remit when the systemic features are controlled. In the other half, unremitting chronic arthritis and joint damage occurs. Thus, durable treatment of SOJIA patients depends on the phase of the disease, that is, whether it is systemic or arthritic. Whereas anakinra treatment of SOJIA does not distinguish between a causative role for IL-1α or IL-1β, sustained responses to canakinumab have been consistently observed implying a role for IL-1β.

MAS

MAS is also known as hemophagocytic syndrome and there is an inherited variant of MAS due to a mutation in perforin. Another related disease is termed cytophagic histiocytic panniculitis, which is characterized by daily high spiking fevers and severe panniculitis 62, 63. There is abnormal activation and proliferation of well-differentiated macrophages/histiocytes, together with increased phagocytic activity. The primary clinical and biochemical features of MAS include non-remitting high fever, hepatosplenomegaly, cytopenia, hypertriglyceridemia and hyperferritinemia and the disease is often fatal due to hepatic failure. Although numerous well-differentiated macrophages phagocytosing hematopoietic cells in the bone marrow can be observed, MAS is diagnosed clinically. Despite treatment of MAS with cyclosporine, which improves the outcome, the prognosis remains severe with 50% mortality. The disease is most commonly secondary to infections, usually infection of intracellular organisms and particularly viruses of the herpes family, but it is also secondary to malignancy, notably non-Hodgkin lymphoma, as well as inflammatory/auto-immune diseases such as 64 and AOSD 60.

MAS is unusual as an IL-1β-mediated disease because of the lack of neutrophilia. Nevertheless, anakinra is used to treat MAS and also the variant of MAS (secondary hemophagocytic syndrome). MAS is probably the best example of an acute, and often lethal, disease due to “hyper-caspase-1 activity” processing and release of IL-18 65. IL-18 is a proinflammatory cytokine belonging to the IL-1 family; IL-18 is present constitutively in monocytes/macrophages, antigen presenting cells and epithelial cells of healthy humans and mice as an inactive precursor and requires caspase-1 for processing to an active cytokine. Indeed, IL-18 appears to be the agonistic cytokine in MAS as IL-18 drives IFN-γ and IFN-γ is known as an activator of macrophages. IL-18-driven IFN-γ also explains the pancytopenia that characterizes MAS, as IFN-γ therapy is known to suppress hematopoiesis. However, IL-18 directly accounts for the hepatic failure in MAS as IL-18 induces FAS ligand leading to the death of hepatocytes. In MAS, the balance between free IL-18 and its naturally occurring antagonist, the IL-18-binding protein, is shifted toward high levels of free IL-18, as there is insufficient IL-18-binding protein to oppose the activity of IL-18 66.

IL-1β in treating osteoarthritis

In the joints, IL-1β is the mediator of reduced chondrocyte proteoglycan synthesis, increased synthesis of matrix metallo-proteinases and the release of nitric oxide 67. Mice deficient in IL-1β are protected from inflammation-induced loss of cartilage 54 whereas mice deficient in TNF-α are not. The role of IL-1β in the destructive processes of osteoarthritis has also been studied in rabbits, pigs, dogs and horses 68 and there has been a placebo-controlled trial of intraarticular anakinra treatment. Although there was a clear dose-dependent (50 versus 150 mg) reduction in pain and stiffness scores, the benefit did not extend beyond one month 69. The modest reduction may be due not only to the heterogeneity of the osteoarthritis population in general but also to the short duration of IL-1RI blockade by anakinra. To address the latter, there is an ongoing study of anti-IL-1β mAbs in osteoarthritis using direct intraarticular injection.

Chronic arthrofibrosis with a limited range of motion of the knee joint can fail to respond to intensive physical therapy, corticosteroid injections, and non-steroidal anti-inflammatory agents. In a preliminary study, eight patients with refractory arthrofibrosis received intraarticular anakinra and the joints of 75% of patients (i.e. six patients) returned to activity levels seen prior to disease onset 70.

Recurrent bouts of gouty arthritis

In 1983, using a specific immunoadsorbant chromatography of anti-IL-1, IL-1 activity was isolated from human joint fluids of patients with gouty arthritis 71. In that same year, monosodium urate (MSU) crystals incubated with PBMC in vitro were reported to induce the release of IL-1 activity into the supernatants 72. Therefore, the concept that IL-1 activity is related to gouty arthritis and that MSU induces IL-1β goes back over 20 years and is hardly a new concept 73, 74; however, MSU crystals can be present in joints without triggering a gouty attack. Indeed, pure MSU crystals do not induce IL-1β release from PBMC alone 75 but rather require a second signal such as priming by low levels of endotoxin 73 or free fatty acids 27, 75; the co-stimulant free fatty acid triggers TLR2 27. Not unexpectedly, mice deficient in caspase-1 or ASC exhibited markedly reduced synovial inflammation in response to the MSU-free fatty acid combination, and in mice deficient in ASC, histological examination of the joints revealed near complete protection; however, mice deficient in NLRP3 responded with same inflammatory response as did wild-type mice 27.

Since neutrophils dominate the inflammation of gouty arthritis in humans, the role of the neutrophil needs to be considered. Cell death of neutrophils provides a wealth of possibilities for inflammation. For the synovial macrophage, dead neutrophils provide a source of ATP and other small molecules for activating caspase-1. Neutrophils also provide a source of proteinase-3, which can process the IL-1β precursor into an active cytokine 76. The gouty attack is likely triggered by over nutrition with free fatty acids providing the second signal in MSU-primed cells, followed by the secretion of active IL-1β, which in turn, induces IL-8 and the infiltration of neutrophils. Large numbers of neutrophils augment the inflammation by providing enzymes and ATP, which induces more active IL-1β.

Clinical trials with IL-1β blockade have revealed an impressive and sustained reduction in patients with recurrent attacks of gouty arthritis 77–80. Even with the use of allopurinol to reduce the systemic levels of uric acid and the anti-inflammatory properties of colchicine, there is no dearth of patients with recurrent episodes of painful gouty arthritis poorly controlled with these regimens. These patients often require intermittent courses of glucocorticoids. Thus, the success of IL-1β-blocking therapies is a welcome addition for treating refractory gouty arthritis in these patients. A single dose of canakinumab has been used successfully in patients with acute gout refractory to standards of therapy in a blinded comparison with a injection of triamcinolone acetonide 30. Within 3 days, canakinumab-treated patients experienced less pain than those treated with triamcinolone acetonide. Moreover, canakinumab significantly reduced the risk of recurrent flares as compared with triamcinolone acetonide. Thus, neutralization of IL-1β provides rapid and sustained pain relief and reduced the number of recurrent flares compared with steroid use.

Despite the availability of several widely used TNF-α-blocking therapies for rheumatoid arthritis and other auto-immune diseases, there is a paucity of reports that blocking TNF-α provides an effective reduction in gout severity. One explanation for the lack of clinical trials of TNF-α blockade in gout attacks is that the efficacy of TNF-α blockade in refractory gout is less than expected. One study reports a weak response with rather high doses of infliximab 81. There are also few publications on MSU crystals inducing TNF-α from human and mouse cells unless co-stimulated with endotoxins. Therefore, IL-1β blockade may be used for inducing long-term remissions in refractory patients and replace glucocorticoids. If IL-1β blockade becomes the standard of care in refractory gout, it would be consistent with the unique role of IL-1β in the pathogenesis of auto-inflammatory diseases.

Blocking IL-1β in diabetes

IL-1β and type 2 diabetes

The evidence that IL-1β was toxic for the insulin-producing β-cell begins in 1985 using anti-human IL-1β immunoaffinity chromatography 82. This was a milestone report that advanced the field of “soluble factors” from mononuclear phagocytes playing a pivotal role in the pathogenesis of diabetes. Soon thereafter, recombinant human IL-1β was shown to account for the death of the β-cell while sparing the α-cell 83. The topic has been reviewed by Mandrup-Poulsen and co-workers, Mandrup-Poulsen being responsible for the original studies 84. Initially, IL-1 was considered to play a pathogenic role primarily in type 1 diabetes, but a role for IL-1β in type 2 diabetes was not appreciated at that time.

However, from the studies of Donath et al., IL-1β was implicated in type 2 diabetes, which supported the concept that type 2 diabetes is a chronic inflammatory disease (reviewed in 84). In fact, it was shown that high concentrations of glucose stimulated IL-1β production from the β-cell itself 85 resulting in β-cell death and progressive loss in β cell mass. Relevant to the pathogenesis of type 2 diabetes, glucose-induced IL-1β from the β-cell is enhanced by the presence of free fatty acids. Fundamental to IL-1β-mediated loss of β cell mass is the metabolic upheaval of over-nutrition and obesity and there studies show that the adipocyte in the distant fat stores contributes to the loss of the β-cells 86.

The loss of the β cell by IL-1β can also be mediated by oligomers of islet amyloid polypeptide, a protein that forms amyloid deposits in the pancreas during type 2 diabetes, triggering NLRP3 and generating mature IL-1β 87. As reviewed by Masters and O'Neill 88, in the pathogenesis of type 2 diabetes, Alzheimer's disease and in amyotrophic lateral sclerosis, protein aggregates caused by inappropriate misfolding are sensed by the NLRP3, providing a unifying mechanism for IL-1β maturation in these diseases, as IL-1β has been implicated in each of these diseases.

The differentiation of the adipocyte and insulin sensitivity itself is affected by a caspase-1-dependent IL-1β-mediated mechanism. Mice fed a high fat diet have increased caspase-1 and elevated levels of IL-1β. In contrast, caspase-1-deficient mice have decreased body fat and improved insulin sensitivity 86. In vivo, treatment of obese mice with a caspase-1 inhibitor significantly increases their insulin sensitivity 86. Calorimetry analysis revealed higher fat oxidation rates in caspase-1-deficient animals, and adipocytes from caspase-1-deficient mice or mice deficient in NLRP3 are more metabolically active ex vivo with higher insulin sensitivity and increased production of adiponectin as compared with adipocytes from wild-type mice. Gene expression for PPARγ and GLUT4 was also increased in fat from caspase-1- or NLRP3-deficient mice. In the ob/ob obese mouse, fat tissue reveals higher caspase-1 activity with elevated production of active IL-1β. Thus, in addition to blocking IL-1β in type 2 diabetes, targeting IL-1β in pre-diabetic persons with metabolic syndrome should correct some of the abnormalities. These studies are consistent with those reported by Vandanmagsar et al. 89. In those studies, a reduction in adipose tissue expression of NLRP3 was observed in obese persons WT 2 diabetes following calorie restriction and exercise-mediated weight loss. Not unexpectedly, there was improved insulin sensitivity. Similar to the studies by Stienstra et al. 86, NLRP3-deficient mice did not show obesity-induced inflammasome activation in fat depots 89. Collectively, both studies 86, 89 establish that caspase-1-dependent cytokines play an important and possibly causative role in obesity-induced inflammation and insulin resistance.

Clinical trials blocking IL-1β in type 2 diabetes

The first clinical proof of a role for IL-1 in the pathogenesis of type 2 diabetes was a randomized, placebo-controlled study of anakinra for 13 wk. In that study, improved insulin production and glycemic control was observed in anakinra-treated patients 90. The fall in glycated hemoglobin was nearly 0.5% lower than that in placebo-treated patients. In addition to improved glycemic control, C-peptide levels increased and the ratio of proinsulin to insulin decreased, both indicators of improved β-cell function. Not unexpectedly, serum IL-6 and CRP levels decreased significantly. In the 39 wk following the 13- wk course of anakinra, patients who responded to anakinra used 66% less insulin to obtain the same glycemic control as compared with baseline requirements 91. The proinsulin to insulin ratio also improved. Patients with low circulating levels of endogenous IL-1Ra before the trial responded to anakinra and in the 39 wk following the 13-wk treatment period, these responders maintained the improvement in stimulated C-peptide 91. These observations suggest that blocking IL-1β, even for a short period of time, restores the function of the β cells or possibly allows for partial regeneration of β cells.

The observations made in the anakinra trial in type 2 diabetes have been confirmed using a specific neutralizing mAb to IL-1β 92 and the mAb has also provided more evidence that short-term blockade of IL-1β restores the function of the β cells and possibly regeneration. Similar to the anakinra trial, the effect of a single administration of the mAb to IL-1β resulted in decreased glycated hemoglobin A1C, increased C-peptide levels, greater insulin production following a glucose challenge and decreased IL-6 and CRP levels 93. The reduction in IL-1β-mediated inflammation is not limited to the islet but is rather systemic. Therefore, it is likely that improved glycemic control reflects not only less toxicity on the β-cell in the islet but also reduced inflammation in the adipose tissue.

Blocking IL-1β in acute ischemic events

Similar to the ability of IL-1β to induce cell death in the β-cell, IL-1β is also toxic for the cardiac myocyte 94, 95. In a placebo-controlled trial of patients with ST elevation myocardial infarction (STEMI), daily anakinra was added to the standard therapy the day after angioplasty for 14 days. Serial imaging and echocardiographic studies after 14 wk revealed that left ventricular remodeling was significantly reduced in patients receiving anakinra as compared with patients receiving 14 days of placebo 95. These findings are consistent with myocardial infarction models in mice, in that blocking IL-1 results in a similar reduction in remodeling 96. Therefore, reducing IL-1β-mediated inflammation in the islet may also benefit IL-1β-induced inflammation in coronary arteries, peripheral arteries and the myocardium itself.

Is smoldering myeloma an auto-inflammatory disease?

Targeting IL-1β in smoldering/indolent myeloma

Smoldering myeloma presents a challenge to medicine as the population ages 97. Decades of research have focused on the role of IL-1β and IL-6 in the pathogenesis of multiple myeloma 98, 99. Similar to mature B cells, the myeloma plasma cell produces IL-1β. In the microenvironment of the bone marrow, stromal cells respond to low concentrations of IL-1β and release large amounts of IL-6, which in turn promotes the survival and expansion of the myeloma cells. Lust, Donovan and co-workers reasoned that in the indolent stages of multiple myeloma, blocking IL-1β would provide better control of IL-6 activity. Bone marrow cells from patients with smoldering myeloma were co-cultured with a myeloma cell line actively secreting IL-1β. Anakinra added to these co-cultures significantly reduced IL-6 by nearly 90% and the combination of anakinra plus dexamethasone induced myeloma cell death 100.

The clinical trial

Based on in vitro data, 47 patients with smoldering/indolent myeloma at high risk for progression to full-blown multiple myeloma were treated with daily anakinra for six months. During the 6 months, there was a decrease in CRP in most but not all patients. After 6 months of anakinra, a low dose of dexamethasone (20 mg/wk) was added. Of the 47 patients who received anakinra (25 anakinra with dexamethasone), progression-free survival ensued for more than three years and in 8 patients for more than 4 years 100. Patients with a decrease in serum CRP of 15% or greater after 6 months of anakinra monotherapy resulted in progression-free survival times greater than 3 years as compared with 6 months in patients with less than a 15% fall during anakinra therapy (p<0.002). Thus, an effective reduction in IL-1β activity using CRP as the marker for IL-1β-induced IL-6 halts progression to active myeloma.

Schnitzler's syndrome

Anakinra results in resolution of all signs and symptoms within hours after the first injection. However, approximately 20% of patients with Schnitzler's syndrome develop a lymphoproliferative disorder, mostly lymphoma or Waldenstrom disease, which is similar to patients with IgM MGUS. This latter point and its consequences have been already been addressed in the literature 101. Blocking IL-1β may reduce the progression to a lymphoproliferative disorder in patients with Schnitzler's syndrome. Similar to smoldering myeloma, the concept that IL-1β drives IL-6 production was tested in a patient with another lymphoproliferative disorder, Castleman's disease, which is usually treated with anti-IL-6 receptor antibodies 102. The patient failed to respond to cladribine, rituximab, steroids, etanercept and anti-IL-6 antibody but within 1 wk of anakinra treatment, the constitutional symptoms markedly improved, and anemia, thrombocytosis, leukocytosis, and elevated markers of systemic inflammation reverted to normality 103.

Concluding comments

In cytokine biology as applied to the treatment of disease, associations of elevated circulating levels of a particular cytokine with a disease do not allow for a conclusion of causation by that cytokine for the pathological process. Rather, only specific blockade or neutralization provides the evidence. This is especially the case with IL-1β, as circulating levels, even in severe systemic inflammatory diseases, are undetectable and yet the disease manifestations are dramatically reduced upon blockade of IL-1 activity. This commonly observed therapeutic response is due to the high specific activity of IL-1β, which can be in the picomolar range in humans. Therefore, establishing a role for IL-1β in inflammatory diseases has succeeded by using short-term IL-1β-blockade and its role and usefulness will likely increase with clinical testing, facilitated by the safety of short-acting anakinra and the availability of neutralizing anti-IL-1β antibodies.

Acknowledgements

Supported by NIH Grants AI-15614, CA-04 6934 and JDRF 26-2008-893. The author thanks Antonio Abbate, Mihai Netea, Leo Joosten, Anna Simon and Jos van der Meer for many helpful suggestions in the preparation of this MS.

Conflict of interest: The author declares no financial or commercial conflict of interest. This article is editorially independent of Novartis.

Ancillary