<<

iii18 Ann Rheum Dis 2001;60:iii18–iii24 Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from Novel targets for 18 binding

C A Dinarello

Abstract clear that IL18 is a proinflammatory Background— (IL18) is re- and that its mechanism of action can be lated to the IL1 family by structure, independent of its ability to induce IFNã. receptors, signalling molecules, and func- IL18 is related to IL1â more than any other tion. IL18 induces expression and cytokine. The similarities between IL1â and synthesis of tumour necrosis factor IL18 exist at several levels. Firstly, the IL18 (TNF), IL1, Fas , several chemo- precursor form (proIL18), like proIL1â, does kines, and vascular adhesion molecules. not contain a , and proIL18 Similar to IL1â, IL18 is synthesised as a requires cleavage to an active cytokine by the biologically inactive precursor molecule IL1â converting enzyme (ICE, caspase-1). lacking a signal peptide. The IL18 precur- IL1â and IL18 are structurally related because sor requires cleavage into an active, both are primarily all â-pleated sheet 2 mature molecule by the intracellular folded molecules. This structural relationship cysteine protease, IL1â converting enzyme is significant because there are very few all-â (ICE, or caspase-1). Inhibitors of ICE sheet molecules among the diVerent cytokines. activity limit the biological activity of IL18 At the level, the activity of IL18 is in animals and may be useful in reducing through a heterodimeric complex, the IL18 the activity of IL18 in human disease. receptor (IL18R) complex. The IL18R binding However, a constitutively secreted IL18 chain is termed IL18Rá. IL18Rá is a member of the IL1 receptor family,3 previously identi- binding protein (IL18BP) exists which 4 functions as a natural inhibitor of IL18 fied as the IL1R related protein (IL1Rrp). A activity. IL18BP binds IL18 with a high signalling chain (IL18Râ), also termed acces- sory protein-like (AcPL),5 aYnity (Kd of 400 pM) and, at equimolar is related to the IL1R accessory protein.6 Although similar to ratios, inhibits 50–70% of IL18; at twofold the IL1 receptor accessory protein in that the molar excess, IL18BP neutralises nearly IL18Râ does not itself bind its ligand in all IL18 activity. solution, the IL18Râ chain is part of the IL18 Method—IL18 was investigated for its role receptor complex. After binding of IL18 to the in human myocardial function. An IL18Rá, the IL18R AcPL binds to form a high ischaemia/reperfusion(I/R)modelofsupra- aYnity heterodimeric complex with the ligand. fused human atrial myocardium was used The high aYnity IL18R complex recruits the to assess myocardial contractile force. IL1 receptor activating kinase (IRAK), result- Results—The addition of IL18BP to the ing in phosphorylation of nuclear factor êB perfusate during and after I/R resulted in (NF B)-inducing kinase (NIK) with subse- ê http://ard.bmj.com/ improved post-I/R contractile function quent translocation of NFêB to the nucleus.7 from 35% of control to 76% with IL18BP. Initially identified as part of the IL1R signalling Also, IL18BP treatment preserved intra- events,8 IRAK is recruited to the IL1R complex cellular tissue creatine kinase levels (by after exposure to IL1. In cells possessing both 420%). Because active IL18 requires cleav- the IL18Rá and â chains, nuclear translocation age of its precursor form by ICE, inhibi- of NFêB is seen after incubation with IL18, tion of ICE attenuated the depression in and this property helps to explain the pleo-

contractile force after I/R (from 35% of tropic nature of IL18. In IL18 deficient mice, on October 1, 2021 by guest. Protected copyright. control compared with 75.8% in treated production of IFNã and cytotoxic T cells is atrial muscle, p<0.01). markedly diminished despite ample amounts Conclusion—Myocardial ischaemia is a of IL12.9 A similar finding exists in mice target for IL18BP and use of IL18BP may deficient in ICE.10 11 The role of IL12 in IFNã thereby reduce ischaemia-induced myo- production therefore seems to require IL18. cardial dysfunction. (Ann Rheum Dis 2001;60:iii18–iii24)

Constitutive production from Interleukin (IL) 18 was first described as an monocyte/ (IFN) ã-inducing factor1 which The production of IL18 in freshly obtained Department of circulated during endotoxaemia in mice pre- human peripheral blood mononuclear cells Medicine, Division conditioned with an infection of Propionibacte- (PBMC) and in splenic from Infectious Diseases, 12 B168, University rium acnes. Because of its ability to induce non-preconditioned mice has been studied. Colorado Health IFNã, IL18 is by default a member of the T cell Reverse transcriptase polymerase chain reac- Sciences Centre, 4200 helper type I (Th1)-inducing family of cyto- tion and western blot analysis were used to East Ninth Avenue, kines (IFNã, IL2, IL12, IL15). However, compare the production of IL18 with that of Denver, CO 80262, because antibodies to IL18 also reduced the IL1â in the same preparations. USA C A Dinarello hepatotoxicity of endotoxaemia, IL18 was con- There is no constitutive for sidered to possess other biological properties IL1â in freshly obtained human PBMC from Accepted 27 June 2001 beyond that of inducing IFNã. It has become healthy donors using over 40 cycles of PCR.13 14

www.annrheumdis.com Interleukin 18 binding protein iii19 Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from Surprisingly, the same PBMC expressed con- IL18 induced IL8 gene expression and synthe- stitutive mRNA for IL18. This was also seen sis.20 When IL1 receptor antagonist was used to using western blot analysis for proIL18 in block IL1 activity in these cells, IL18-induced lysates from the same PBMC which contained IL8 production was reduced by 40%. When no proIL1â in the same cells. Constitutive tumour necrosis factor (TNF) activity was IL18 gene expression and the presence of inhibited, IL18-induced IL8 production was proIL18 protein were also observed in freshly reduced by 80%, suggesting that the primary obtained murine splenocytes.12 The promoter action of IL18 was through a TNF dependent regions for IL1â and IL18 gene expression pathway. The source of TNF was the CD3+ have been studied and may provide an insight and the natural killer cells. The induction of into these observations. The promoter for IL18 TNF can be seen at the level of gene expression is TATA-less and IL18 promoter activity within two hours after exposure to IL18. IL18 upstream of exon 2 acts constitutively.15 There- also stimulates the synthesis of macrophage fore, it is not unexpected that IL18 mRNA is chemoattractant protein 1 in these cultures.20 constitutively expressed even in whole blood Using neutralising anti-IL18 antibodies, freshly obtained from healthy donors where no Netea et al compared the myeloperoxidase artefact is introduced. The additional finding activity in lungs and livers of mice after that the 3' untranslated region of human IL18 endotoxaemia.23 ICE deficient mice were com- lacks the AUUUA destabilisation sequence15 is pletely resistant to lethal endotoxaemia in- also consistent with these observations. This duced by (LPS) derived would allow for more sustained levels of the from either Escherichia coli or Salmonella polyadenylated species and translation into typhimurium. IFNã deficient mice were not protein. Daily injections of IL12 into mice resistant to S typhimurium LPS, suggesting an result in high circulating levels of IFNã, which IFNã-independent role for IL18. Anti-IL18 is completely prevented by prior treatment protected mice against a lethal injection of with anti-IL18 antibodies, or absent in mice either LPS and also reduced neutrophil deficient in ICE.11 In this model, IL12 may accumulation in liver and lungs. The increased activate ICE. Osteoclasts also produce IL18,16 survival was accompanied by decreased levels and regulation of bone density may be a prop- of , macrophage inflammatory pro- erty of IL18, as it is for IL1â. tein 2.23 In addition to monocyte/macrophages, the epidermal cells from mouse skin produce IL18 IL18 up regulates expression of constitutively.17 The is the major IL18 enhances Fas ligand expression and source of the IL18. Moreover, contact sensitis- induces apoptosis in Fas-expressing human ers increased IL18 synthesis, whereas skin irri- myelomonocytic KG-1 cells.22 KG-1 cells are tants did not. A murine keratinocyte cell line monocyte-like and also produce the intercellu- (PAM 212) also produced IL18 constitu- lar adhesion molecule-1 (ICAM-1). IL18 tively.17 IL18 has also been found constitutively increases ICAM-1 expression in these macro- expressed in colonic specimens and isolated phagic cells.24 These cells also express constitu- mucosal cell populations from patients with tively Fas antigen (CD95) and, after exposure Crohn’s disease.18 Similar to IL1, IL18 is found to IL18, KG-1 cells became apoptotic. Fas lig- in the rat adrenal cortex and pituitary gland.19 and is up regulated in T cells and NK cells after http://ard.bmj.com/ The adrenal IL18 was increased by cold stress, incubation with IL18.25 In the liver, cytotoxic T whereas the pituitary IL18 was not. Two tran- cells are thought to contribute to the destruc- scripts of adrenal IL18 were sequenced: an tion of hepatic cells, particularly when infected abundant form was clearly the rat homologue by viruses. Cytotoxic T cells and NK cells iso- of mouse IL18 and shared 91% lated from the liver and then stimulated with . The other transcript was less abun- IL18 exhibit enhanced killing of Jurkat T 26

dant. It had a 59 deletion and the cells. The IL18-induced enhanced killing by on October 1, 2021 by guest. Protected copyright. corresponding 19 amino acids at the 3' end of liver derived cytotoxic T lymphocytes was the molecule.19 Possibly, this truncated rat IL18 through a perforin dependent pathway and not molecule may act as a receptor antagonist. CD95.26 Liver derived NK cells also exhibited enhanced killing of Jurkat cells after exposure to IL18, but the mechanism was thought to be IL18 induces cytokines and independent of TNF. However, failure to There is no dearth of evidence for the role of detect TNF in cell culture supernatants is not chemokines in promoting inflammation. For an argument that any cellular change is TNF example, the neutrophil-activating chemokine independent because membrane forms of TNF IL8 is a primary agonist in several pathological are biologically active but often hardly measur- processes of the lung. IL18 stimulates the pro- able in supernatant. In fact, the ability of IL18 duction of IL18 from human NK cells as well to induce HIV-1 in U1 cells21 and IL8 in as other CXC and CC chemokines.20 In the PBMC20 is TNF dependent. In both studies, HIV-1 infected macrophage cell line U1, which the dependence on TNF was shown by reduc- is derived from the histiocytic lymphoma cell ing the IL18 eVect with the soluble TNF p55 U937, IL18 induces IL8.21 In addition, the receptor. Therefore, it is likely that IL18- myelomonocytic leukaemia cell line KG-1 also induced killing by liver NK cells may still be releases IL8 when stimulated with IL18.22 The due to induction of membrane TNF. ability of IL18 to induce CC and CXC chemo- In cockroach allergen-induced airway re- kines places IL18 in a strategic role in sponse, IL18 levels were observed within the inflammation. When human PBMC were used, lungs.27 When IL18 was instilled into the

www.annrheumdis.com iii20 Dinarello Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from airway at the time of allergen challenge, a is an eVective inhibitor of immune and inflam- significant increase in peribronchial eosinophil matory disease associated with increased pro- accumulation was seen in both allergic and duction of IL18. non-allergic mice. After IL18 instillation, there was a significant increase in eotaxin, but not other eosinophil chemotactic factors, in bron- Cytokines in the heart choalveolar lavage fluid.27 The role of eotaxin It must be emphasised that measuring cyto- was confirmed using eotaxin −/− mice, which kines in human disease and making direct cor- demonstrated significantly less eosinophil ac- relations between any particular cytokine and cumulation than control mice. any particular disease is not proof that the cytokine has a causal role in the disease process. This is also the case with assessing the IL18 binding protein role of cytokines in the pathogenesis of an Methods established for the purification and atherosclerotic lesion of the coronary artery, an characterisation of ligand binding cytokine acute ischaemic event associated with myocar- receptors in human urine28–31 were used to dial infarction, progression of myocardopa- purify a 38 kDa IL18 binding protein to thies, or loss of myocardial function in conges- homogeneity from urine using ligand aYnity tive heart failure. Nevertheless, a highly chromatography, and the N-terminal amino significant correlation of the level of a particu- acids were determined.32 The amino acid lar cytokine or cytokines with severity of sequence of the IL18 binding protein disease contributes to an understanding of the (IL18BP) is not that of a known cytokine role of cytokines in that disease process. Two receptor presently listed in the data banks. experimental approaches do provide absolute Although this binding protein probably repre- evidence of a causal role of any particular sents the extracellular domain of the ligand cytokine in the pathogenesis of a particular binding receptor chain, to date a transmem- disease. Firstly, in animal models of the brane and cytosolic component to the IL18BP disease, administration of highly specific cyto- has not been found. Even genomic clones do kine antagonists reduces the severity of the dis- not contain sequences that would translate into ease. In heart disease, several animal models a transmembrane domain for insertion into the exist which have application to human disease. cell membrane.32 Therefore, IL18BP is a Animal models include gene deletion studies or constitutively produced inhibitor of IL18. transgenic mice overexpressing specific . However, from its structure and function, it has Secondly, specific blockade or neutralisation of probably been at one time in evolution, a cell some cytokines—namely IL1, IL18, or TNF, is associated membrane receptor. now possible in patients. Similar to neutralising antibodies to IL18,10 In studies on the eVects of cytokines on the IL18BP prevents LPS-induced IFNã myocardial function in septic shock, the data production.32 Also, similar to anti-IL18 anti- have been primarily physiological and derived bodies, IL18BP does not aVect IFNã produc- from experimental models of septic shock. In tion after stimulation with mitogens such as those animal models, intravenous infusion of concanavalin A. However, unlike antibodies to IL1 or TNF, but particularly the combination IL18, the IL18BP does not exhibit species spe- of IL1 and TNF, exhibited a profound ability http://ard.bmj.com/ cificity, and hence human IL18BP neutralises to suppress myocardial function.35 The local both human and murine IL18.32 With a single production of cytokines from the myocardium immunoglobulin domain, IL18BP resembles itself (as would take place in ischaemia) is more the extracellular segment of cytokine receptors likely to aVect myocardial function at concen- with immunoglobulin-like structures; however, trations that are clinically relevant. For exam- IL18BP is a novel protein distinct from the IL1 ple, in human atrial muscle strips, suspended in

and IL18 receptor families. Located in chro- an oxygenated balanced salt buVer and electri- on October 1, 2021 by guest. Protected copyright. mosome 11q3 at the inverted position of the cally paced, contractile function as measured nuclear mitotic apparatus protein-1, the by developed force is dramatically reduced by human IL18BP gene encodes for at least four the presence of <10 pg/ml of exogenous TNFá distinct isoforms (IL18BPa, b, c, and d), each in the bath.36 IL1â induces a similar reduction derived from a single mRNA by splice variants. in developed force at similar low concentra- They diVer primarily in their carboxyl termini, tions. Therefore, the human heart appears to whereas at the N-terminal one third to two be exquisitely sensitive to the direct eVects of thirds of the amino acids are identical. IL1 and TNF, and in both cases these A BIAcore sensor chip with immobilised cytokines depress function. human IL18 bound IL18BP. The mean disso- The first evidence that IL1 and TNF directly ciation constant for human IL18BPa was suppress myocardial function came from stud- 0.399±0.034 nM. The binding of increasing ies on human serum of patients with acute sep- amounts of IL18BPa to immobilised human tic shock marked reversible myocardial depres- mature IL18 reveals a rapid on-rate and a sion.37 Using immunoabsorption, removal of markedly slow oV-rate.33 both TNFá and IL1â (but not either alone) IL18BP circulates in healthy humans at con- from these sera resulted in elimination of centrations of 2.15 (0.15) ng/ml (range 0.5– serum myocardial depressant activity. IL2, IL4, 7).34 In general, the molar excess of IL18BP to IL6, IL8, IL10, or IFNã failed to cause signifi- IL18 is in the order of 20–30-fold greater. cant cardiac myocyte depression over a wide Given the high aYnity of IL18BP for neutrali- range of concentrations. Individually, TNFá sation of IL18 (400 pM), it seems that IL18BP and IL1â each depressed peak velocity and

www.annrheumdis.com Interleukin 18 binding protein iii21 Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from myocyte shortening in vitro, but the combina- are suspended in a blood-free, physiologically tion of TNFá and IL1â depressed myocardial oxygenated buVer bath and then exposed to an cell contractility at substantially lower concen- episode of simulated ischaemia. During this trations, consistent with a synergistic eVect.37 time, contractile force decreases dramatically; The role of nitric oxide (NO) and cyclic guanyl when the tissue is re-exposed to oxygen, the monophosphate (cGMP) in depression of contractile force returns but is diminished myocardial contractility induced by either (60–70% reduction), and evidence of myocar- IL1â, TNFá, or the combination of these dial damage is seen by release of creatine cytokines, was investigated.38 Inhibitors of NO kinase.43 44 When TNF bioactivity is specifically synthase such as N-methyl-L-arginine (L- neutralised during ischaemia/reperfusion (I/R), NMA) and methylene blue for inhibition of a greater return of contractile force is seen, guanylate cyclase prevented cytokine induced suggesting that endogenous myocardial TNF myocardial suppression; an excess of L-arginine activity contributes to the contractile dysfunc- 45 with L-NMA restored the eVect. In addition, tion induced by the ischaemic event. TNFá, IL1â, or TNFá plus IL1â induced NO We asked whether the cytokine IL18 con- formation in myocyte cultures; moreover, tributes to human ischaemia-induced myocar- serum samples from patients with acute septic dial dysfunction. As discussed above, IL18 is a shock also induced NO production. proinflammatory cytokine structurally and In considering the direct eVects of IL1, IL18, functionally related to IL1â.24647 Both IL1â or TNF on myocardial function, one should and IL18 are initially synthesised as inactive keep in mind several pathogenic situations— precursors requiring ICE for cleavage to namely, (a) the local production of IL1, IL18, mature, biologically active molecules.48 49 Al- and TNF from resident myocardial macro- though IL1â and IL18 have distinct cell surface phages; (b) the ability of ischaemia to induce receptors, the receptor chains for each cytokine the release of IL1, IL18, or TNF from these are members of the same receptor super- macrophages; (c) the level of IL1, IL18, and family,35and signal transduction is similar.7 To TNF receptor expression on myocytes; and (d) assess a role for endogenous IL18 in the heart, relevant post-receptor events after IL1, IL18, a specific, natural inhibitor of IL18 activity, or TNF in triggering of their respective recep- IL18BP, was added to the suprafusing bath tors on myocytes. during I/R.

IL18 in the heart Neutralisation of endogenous IL18 with During ischaemia, as well as reperfusion, IL18BP prevents the loss in numerous endogenous mediators, such as post-ischaemic developed force small molecule second messengers, are pro- Control human atrial trabeculae were supra- duced which aVect myocardial function. fused under normoxic conditions throughout Within minutes of an ischaemic episode, myo- the experiment. Under these conditions, the cardial contractile force diminishes and the developed force in the control trabeculae is overall recovery of contractile force is largely reduced (10%). Trabeculae subjected to 20 dependent on the duration of the ischaemic minutes of ischaemia exhibit a rapid decline in period.39 For example, during an ischaemic contractile function; upon reperfusion, con- event, Ca2+ homeostasis is perturbed, oxygen tractile force returns to approximately 25% of http://ard.bmj.com/ derived free radicals are generated, and nitric the control developed force. In contrast, oxide (NO) synthesis and release takes place. trabeculae exposed to ischaemia, but in the In addition, there is also local production of presence of IL18BP, returned to 55% of the cytokines, particularly TNFá and IL1â.40 In control developed force.50 the intact heart, these cytokines contribute to These results suggest that I/R leads to release ischaemia-induced myocardial dysfunction by of biologically active IL18 after processing the

inducing gene expression for inducible NO endogenous precursor IL18 by ICE. There- on October 1, 2021 by guest. Protected copyright. synthase (iNOS), cyclo-oxygenase-2 (COX-2), fore, IL18 was measured in freshly obtained

and phospholipase A2, as well as vascular adhe- atrial tissue. Basal IL18 was present in sion molecules and several chemokines. As a trabeculae obtained before the insertion of the result, there is immediate depression of myo- pump-oxygenator cannula into the right cardial contractile force mediated by small atrium. After 90 minutes of equilibration, 30 molecule messengers followed by cytokine minutes of ischaemia, and 45 minutes of mediated neutrophil infiltration, which further reoxygenation, trabeculae were homogenised damages heart muscle. Animal hearts studied and IL18 levels determined. A 4.5-fold in- in the absence of blood or blood products crease in IL18 in the post-I/R tissue was elaborate TNFá41 and IL1â during an ischae- found.50 mic challenge. Cardiomyocytes also lose con- Steady state mRNA levels for IL18 and tractile force owing to the action of these IL18BP were also determined in these tissues. endogenous cytokines.42 We observed basal gene expression for IL18 Most of the experimental data concerning and IL18BP in the freshly obtained, pre- TNFá and IL1â mediated myocardial dysfunc- ischaemic atrial homogenates. Similar to the tion are derived from animal studies. However, increase in IL18 protein, I/R induced a further human myocardial tissues obtained from pa- increase in steady state IL18 mRNA levels tients undergoing elective cardiopulmonary (4.7-fold increase). Gene expression for bypass procedures have been studied under IL18BP was also present in freshly obtained controlled, ex vivo conditions.43 44 In this atrial tissue and increased only modestly (1.3- experimental model, human atrial trabeculae fold) after I/R.

www.annrheumdis.com iii22 Dinarello Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from Atrial tissue was obtained just before inser- calcium, and myofilaments to induce contrac- tion of the pump-oxygenator cannula and tile depression. In addition to xanthine oxidase, immediately snap frozen. IL18 was localised in TNFá is an inducer of oxygen metabolites. macrophages and endothelial cells before any Recent data indicate that IL18 primes human operation-related ischaemia took place and neutrophils for superanion production. also was present in the absence of contact with Ischaemia is a direct stress signal to the myo- any foreign surfaces. The localisation of IL18 cyte and as a result, gene expression of stress in resident macrophages and endothelial cells related molecules is increased. For example, is consistent with previous studies of constitu- after 15 minutes of ischaemia in rodent hearts tive preformed precursor IL18 in freshly perfused with Kreb’s buVer, TNFá gene obtained human peripheral monocytes from expression is up regulated.40 However, the sud- healthy subjects.12 Therefore, we conclude that den and marked reduction in atrial trabecular preformed precursor IL18 exists in the myo- function in the present study is apparent within cardium of patients scheduled for coronary minutes and it is unlikely that cytokines artery bypass for ischaemic heart disease. account for the early dysfunction. During reperfusion, however, the failure to return ICE inhibition also reduces completely to functionality appears to be post-ischaemic developed force cytokine mediated because specific cytokine Because IL18BP eVectively attenuated blockade or neutralisation restores functional- ischaemia-induced myocardial dysfunction, we ity to a greater degree than ischaemic controls. suggested that inhibition of the conversion of Depressed function during reperfusion may be preformed precursor IL18 to mature IL18 due to oxygen radical-induced loss of myocyte would also attenuate ischaemia-induced myo- integrity, increased production of NO, or cardial dysfunction. Therefore, the specific altered calcium flux. Therefore, do IL1â or ICE inhibitor YVAD was added to the IL18, or both, trigger the above changes? The suprafusion bath before the onset of ischaemia. addition of IL1â to oxygenated human ICE inhibition by the addition of YVAD was trabeculae suppresses function,36 and it is continued throughout the ischaemic period known that IL1 induces NOS in cardiac myo- and during reperfusion. Inhibition of ICE â cytes.52 However, it is not known if IL18 acts resulted in attenuation of ischaemia-induced myocardial dysfunction as shown by the similarly. improvement in contractile function from 35% Nitric oxide is a myocardial depressant. of control in I/R to 60% at 10 µg/ml and 75.8% However, the eVect of of NO after ischaemia is at 20 µg/ml.50 These results confirm that controversial. This controversy stems from the biologically active IL18 in human myocardium diVerent tissue levels of NO present depending is the result of cleavage of preformed precursor on which pathway of NO synthesis is activated. IL18 by ICE. In addition, these results suggest Lower levels of NO resulting from synthesis that myocardial ischaemia may activate latent through the constitutive NOS pathway seem to 53 ICE. protect the myocardium, whereas the NO produced from inducible nitric oxide synthase (iNOS), which is significantly higher, leads to Implications of IL18 blockade in 54 ischaemic heart disease myocardial injury. After a moderate ischaemic http://ard.bmj.com/ The generation of oxygen derived free radicals, insult, induction of iNOS occurs in the rat nitric oxide, calcium overload, or decreased myocardium, followed by increased NO pro- 55 responsiveness of the myofilaments to calcium duction. This NO subsequently leads to myo- may each contribute to post-I/R contractile cardial contractile depression. Using the same dysfunction.39 In addition to these immediate trabeculae model as the present study, Cain et acting mediators, the relationship of cytokines al showed that specific inhibition of nitric oxide synthase attenuated TNFá- and IL1â-induced to post-I/R myocellular dysfunction remains on October 1, 2021 by guest. Protected copyright. unclear. The data in the present studies suggest human myocardial dysfunction.36 As previously that IL18 and IL1â are processed and released discussed, endogenous TNFá accounts for from their endogenous precursor forms in some of the post-ischaemic myocardial dys- human heart tissue during ischaemic injury function. There are numerous hypotheses as to and function to suppress contractile force. how TNFá mediates ischaemia-induced myo- Furthermore, the processing of the precursors cardial dysfunction. Finkel et al, showed that appears to be ICE dependent, and latent ICE is TNFá induced contractile dysfunction in probably activated by ischaemia. Previously, isolated hamster papillary muscle.54 This eVect neutralisation of endogenous TNFá was was abolished with inhibition of NO synthase. shown to protect human trabeculae from Nitric oxide has been shown to have a role in ischaemia-induced dysfunction.44 At present, it TNFá-induced myocardial dysfunction is likely that a combination of IL18, IL1â, and through desensitisation of the myofilaments to TNFá accounts for the ischaemia-induced calcium.52 In addition, TNFá may also lead to dysfunction. phosphorylation of troponin, which further Oxygen metabolites present after ischaemia desensitises the myofilaments to calcium. depress myocardial contractile function in sev- Calcium is a vital mediator of myocardial eral animal models, both in vitro and in vivo.39 contractile function. Changes in intracellular The source of the oxygen radicals is unclear, Ca2+, cellular calcium overload, and modula- though xanthine oxidase may be an important tion of the myofilaments’ response to Ca2+ each mediator of oxy radical production.51 Oxy radi- aVect contractile force. Most investigation has cals may interact with cellular , lipids, focused on the role of calcium as the eVector of

www.annrheumdis.com Interleukin 18 binding protein iii23 Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from myocardial contractile dysfunction. The rela- within the immediate post-I/R injury. ICE tionship between myocardial calcium changes inhibition also preserves cell viability because and myocellular contractile dysfunction has creatine kinase levels remained high in post- been well described.39 After an I/R injury, the ischaemic tissues treated with an ICE inhibitor. myofilaments’ responsiveness to calcium de- IL1â and TNFá have also been implicated in creases and is thought to account for most of the pathogenesis of human myocardial sup- the decrease in contractile function after pression in sepsis.37 59 The mechanism(s) by ischaemia. In addition to calcium overload, an which IL1â and TNFá induce contractile dys- ischaemic insult leads to the production and function has also been linked to NO and activation of intracellular calcium dependent changes in cellular calcium handling.59 In addi- proteases. Upon activation, these proteases tion, inhibition of the sphingomyelin signalling begin intracellular myofilament proteolysis, pathway abrogated TNFá/IL1â-induced myo- leading to post-ischaemic contractile dysfunc- cardial contractile dysfunction.36 Although the tion. Given the protection aVorded by the anti- present study does not deal with the role of NO cytokine interventions in the present study, it is in IL18 mediated, ischaemia-induced dysfunc- likely that IL1â or IL18, or both, alter intracel- tion, TNFá depresses the myocardium in a NO lular calcium homeostasis during and after dependent pathway.44 Blockade of IL1 recep- ischaemia. tors revealed a role for endogenous IL1â in the Although mature IL1â has been shown to I/R injury, a finding that was not unexpected suppress function directly when added to given the large amount of animal data. That human atrial trabeculae,36 it has not been endogenous IL18 also plays a part in the injury shown whether endogenous IL1â in the heart was unexpected but based on the fact that participates in ischaemia-induced dysfunction. IL18BP only neutralises mature IL18.32 33 In the present study, inhibition of IL1â activity Because ICE inhibition prevents the cleavage by IL1 receptor blockade suggests that biologi- of both precursor IL1â and IL18, it would not cally active endogenous IL1â is present in the be surprising if IL18 and IL1â acted synergis- heart after ischaemia. Furthermore, the forma- tically in suppressing myocardial function. In tion of active IL1â in the ischaemic heart is fact, specific TNFá, IL1â, and IL18 blockade ICE dependent. The data are consistent with each attenuate, but do not entirely reverse, I/R- the concept that synthesis of the precursor for induced myocardial dysfunction, suggesting IL1â, and activation of ICE, take place during that the three cytokines act together to I/R. suppress myocardial function. The studies support the concept that human The ability to modulate or interrupt cytokine atrial myocardium is highly sensitive to IL18 signalling has been demonstrated in numerous and IL1â and that the combination of these disease states in both animal models and clini- two cytokines seems synergistically to depress cally in humans—for example, rheumatoid myocardial function. We have previously arthritis and Crohn’s disease. IL1â and TNFá shown that the presence of exogenous IL1â or have both been associated with acute myocar- TNFá decreases contractile force in human dial dysfunction and hence are logical targets in trabeculae in the absence of ischaemia.36 In patients. However, IL18, a new member of the addition, the presence of these two cytokines IL1 super family, has to date not been together has a synergistic eVect on the depres- associated with myocardial dysfunction but http://ard.bmj.com/ sion of myocardial contractility. Furthermore, should be considered a target for ischaemic we have preliminary data to suggest that exog- treatment based on the present studies. IL1â enous IL18, under normoxic conditions, also and IL18 share numerous properties, including depresses myocardial contractile function. cleavage of the inactive precursor forms to the The ability of ICE inhibition to reduce post- active forms by ICE (caspase-1). Given the ischaemic dysfunction suggests that the myocardium’s response to proinflammatory

processing of precursor IL1â and IL18 is nec- cytokines and the ability to interrupt the on October 1, 2021 by guest. Protected copyright. essary for cytokine mediated myocardial sup- inflammatory process we suggest that the inhi- pression. The immunohistochemical studies bition of caspase-1 protects ischaemia-induced showed that IL18 is preformed in the resident human myocardial dysfunction through the macrophages and endothelial cells of atrial tis- inhibition of IL1â and IL18 processing. sues from patients with ischaemic heart These studies are supported by NIH Grant AI15614. disease, but it is not clear if precursor IL1â is also preformed. However, IL1â mRNA is rap- 1 Nakamura K, Okamura H, Wada M, Nagata K, Tamura T. Endotoxin-induced serum factor that stimulates gamma idly increased in rat hearts within 15 minutes inteferon production. Infect Immun 1989;57:590–5. after an ischaemic insult,40 and therefore it is 2 Bazan JF, Timans JC, Kaselein RA. A newly defined interleukin-1? Nature 1996;379:591. likely that there is also increased precursor 3 Torigoe K, Ushio S, Okura T, Kobayashi S, Taniai M, IL1â synthesis in atrial trabeculae during Kunikate T, et al. Purification and characterization of the human interleukin-18 receptor. J Biol Chem 1997;272: ischaemia. Ischaemia itself may be an activator 25737–42. of latent ICE activity in heart tissue. Several 4 Parnet P, Garka KE, Bonnert TP, Dower SK, Sims JE. IL-1Rrp is a novel receptor-like molecule similar to the investigators have reported that ICE inhibition type I interleukin-1 receptor and its homologues T1/ST2 during myocardial I/R injury in animals and IL-1R AcP. J Biol Chem 1996;271:3967–70. 5 Born TL, Thomassen E, Bird TA, Sims JE. Cloning of a reduces apoptotic cell death. The criteria used novel receptor subunit, AcPL, required for interleukin-18 for determining cell death were DNA fragmen- signaling. J Biol Chem 1998;273:29445–50. 6 Greenfeder SA, Nunes P, Kwee L, Labow M, Chizzonite tation and cleavage of poly ADP-ribose RA, Ju G. Molecular cloning and characterization of a sec- polymerase.56–58 Importantly, the present stud- ond subunit of the interleukin-1 receptor complex. J Biol Chem 1995;270:13757–65. ies expand these observations by demonstra- 7 Robinson D, Shibuya K, Mui A, Zonin F, Murphy E, Sana tion that ICE inhibition preserves functionality T, et al. IGIF does not drive Th1 development but

www.annrheumdis.com iii24 Dinarello Ann Rheum Dis: first published as 10.1136/ard.60.90003.iii18 on 1 November 2001. Downloaded from synergizes with IL-12 for interferon-ã production and acti- 34 Novick D, Schwartsburd B, Pinkus R, Suissa D, Belzer I, vates IRAK and NFêB. Immunity 1997;7:571–81. Sthoeger Z, et al. A novel IL-18BP ELISA shows elevated 8 Cao Z, Xiong J, Takeuchi M, Kurama T, Goeddel DV. serum IL-18BP in sepsis and extensive decrease of free Interleukin-1 receptor activating kinase. Nature 1996;383: IL-18. Cytokine (in press). 443–6. 35 Okusawa S, Gelfand JA, Ikejima T, Connolly RJ, Dinarello 9 Takeda K, Tsutsui H, Yoshimoto T, Adachi O, Yoshida N, CA. Interleukin 1 induces a shock-like state in rabbits. Syn- Kishimoto K, et al. Defective NK cell activity and Th1 ergism with and the eVect of response in IL-18-deficient mice. Immunity 1998;8:383– cyclooxygenase inhibition. J Clin Invest 1988;81:1162–72. 90. 36 Cain BS, Meldrum DR, Dinarello CA, Meng X, Joo KS, 10 Fantuzzi G, Puren AJ, Harding MW, Livingston DJ, Banerjee A, et al. Tumor necrosis factor-alpha and Dinarello CA. IL-18 regulation of IFN-ã production and interleukin-1beta synergistically depress human myocardial cell proliferation as revealed in interleukin-1â converting function. Crit Care Med 1999;27:1309–18. enzyme-deficient mice. Blood 1998;91:2118–25. 37 Kumar A, Thota V, Dee L, Olson J, Uretz E, Parrillo JE. 11 Fantuzzi G, Reed DA, Dinarello CA. IL-12-induced IFNã Tumor necrosis factor alpha and interleukin 1beta are is dependent on caspase-1 processing of the IL-18 precur- responsible for in vitro myocardial cell depression induced sor. J Clin Invest 1999;104:761–7. by human septic shock serum. J Exp Med 1996;183:949– 12 Puren AJ, Fantuzzi G, Dinarello CA. Gene expression, syn- 58. thesis and secretion of IL-1â and IL-18 are diVerentially 38 Kumar A, Brar R, Wang P, Dee L, Skorupa G, Khadour F, regulated in human blood mononuclear cells and mouse et al. Role of nitric oxide and cGMP in human septic spleen cells. Proc Natl Acad Sci USA 1999;96:2256–61. serum-induced depression of cardiac myocyte contractility. 13 Mileno MD, Margolis NH, Clark BD, Dinarello CA, Burke Am J Physiol 1999;276:R265–76. JF, Gelfand JA. Coagulation of whole blood stimulates 39 Bolli R. Mechanism of myocardial “stunning”. Circulation interleukin-1â gene expression: absence of gene transcripts 1990;82:723–38. in anticoagulated blood. J Infect Dis 1995;172:308–11. 40 Herskowitz A, Choi S, Ansari AA, Wesselingh S. Cytokine 14 Shapiro L, Dinarello CA. Cytokine expression during mRNA expression in postischemic/reperfused myocar- osmotic stress. Exp Cell Res 1997;231:354–62. dium. Am J Pathol 1995;146:419–28. 15 Tone M, Thompson SAJ, Tone Y, Fairchild PJ, Waldmann 41 Meldrum DR, Cleveland JC Jr, Cain BS, Meng X, Harken H. Regulation of IL-18 (IFN-ã-inducing factor) gene AH. Increased myocardial tumor necrosis factor-alpha in a expression. J Immunol 1997;159:6156–63. crystalloid-perfused model of cardiac ischemia-reperfusion 16 Martin TJ, Romas E, Gillespie MT. in the con- injury. Ann Thorac Surg 1998;65:439–43. trol of osteoclast diVerentiation. Crit Rev Eukaryot Gene 42 Gurevitch J, Frolkis I, Yuhas Y, Paz Y, Matsa M, Mohr R, et Expr 1998;8:107–23. al.Tumor necrosis factor-alpha is released from the isolated 17 Stoll S, Mueller G, Kurimoto M, Saloga J, Tanimoto T, heart undergoing ischemia and reperfusion. J Am Coll Yamauchi H, et al. Production of IL-18 (IFN-ã-inducing Cardiol 1996;28:247–52. factor) messenger RNA and functional protein by murine 43 Cleveland JCJ, Meldrum DR, Cain BS, Banerjee A, Harken . J Immunol 1997;159:298–302. AH. Oral sulfonylurea hypoglycemic agents prevent 18 Pizarro TT, Michie MH, Bentz M, Woraratanadharm J, ischemic preconditioning in human myocardium. Two Smith MF Jr, Foley E, et al. IL-18, a novel immunoregula- paradoxes revisited. Circulation 1997;96:29–32. tory cytokine, is up-regulated in Crohn’s disease: expres- 44 Cain BS, Meldrum DR, Dinarello CA, Meng X, Banerjee A, sion and localization in intestinal mucosal cells. J Immunol Harken AH. Adenosine reduces cardiac TNF-alpha 1999;162:6829–35. production and human myocardial injury following 19 Conti B, Jeong JW, Tinti C, Son JH, Joh TH. Induction of ischemia-reperfusion. J Surg Res 1998;76:117–23. IFN-ã-inducing factor in the adrenal cortex. J Biol Chem 45 Cain BS, Meldrum DR, Meng X, Dinarello CA, Shames 1997;272:2035–7. BD, Banerjee A, et al. p38 MAPK inhibition decreases 20 Puren AJ, Fantuzzi G, Gu Y, Su MS-S, Dinarello CA. TNF-alpha production and enhances postischemic human Interleukin-18 (IFN-ã-inducing factor) induces IL-1â and myocardial function. J Surg Res 1999;83:7–12. IL-8 via TNFá production from non-CD14+ human 46 Okamura H, Nagata K, Komatsu T, Tanimoto T, Nukata Y, blood mononuclear cells. J Clin Invest 1998;101:711–24. Tanabe F, et al. A novel costimulatory factor for gamma 21 Shapiro L, Puren AJ, Barton HA, Novick D, Peskind RL, Su interferon induction found in the livers of mice causes MS-S, et al. Interleukin-18 stimulates HIV type 1 in mono- endotoxic shock. Infect Immun 1995;63:3966–3972. cytic cells. Proc Natl Acad Sci USA 1998;95:12550–5. 47 Dinarello CA. IL-18: a Th1-inducing, proinflammatory 22 Ohtsuki T, Micallef MJ, Kohno K, Tanimoto T, Ikeda M, cytokine and new member of the IL-1 family. J Allergy Clin Kurimoto M. Interleukin-18 enhances Fas ligand expres- Immunol 1999;103:11–24. sion and induces apoptosis in Fas-expressing human 48 Gu Y, Kuida K, Tsutsui H, Ku G, Hsiao K, Fleming MA, et myelomonocytic KG-1 cells. Anticancer Res 1997;17: al. Activation of interferon-ã inducing factor mediated by 3253–8. interleukin-1â converting enzyme. Science 1997;275: 23 Netea MG, Fantuzzi G, Kullberg BJ, Stuyt RJ, Pulido EJ, 206–9. McIntyre RC Jr, et al. Neutralization of IL-18 reduces neu- 49 Ghayur T, Banerjee S, Hugunin M, Butler D, Herzog L, trophil tissue accumulation and protects mice against lethal Carter A, et al. Caspase-1 processes IFN-gamma-inducing Escherichia coli and Salmonella typhimurium endotox- factor and regulates LPS-induced IFN-gamma production. emia. J Immunol 2000;164:2644–9. Nature 1997;386:619–23.

24 Kohka H, Yoshino T, Iwagaki H, Sakuma I, Tanimoto T, 50 Pomerantz BJ, Reznikov LL, Harken AH, Dinarello CA. http://ard.bmj.com/ Matsuo Y, et al. Interleukin-18/interferon-gamma-inducing Inhibition of caspase-1 reduces human myocardial factor, a novel cytokine, up- regulates ICAM-1 (CD54) ischemic dysfunction via inhibition of IL-18 and IL-1â. expression in KG-1 cells. J Leukoc Biol 1998;64:519–27. Proc Natl Acad Sci USA 2001;98:2871–6. 25 Dao T, Ohashi K, Kayano T, Kurimoto M, Okamura H. 51 Charlat MI, O’Neill PG, Egan JM, Abernethy DR, Michael Interferon-ã-inducing factor, a novel cytokine, enhances LH, Myers ML, et al. Evidence for a pathogenetic role of Fas ligand-mediated cytotoxicity of murine T helper cells. xanthine oxidase in the “stunned” myocardium. Am J Cell Immunol 1997;173:230–5. Physiol 1987;252:H566–577. 26 Dao T, Mehal WZ, Crispe IN. IL-18 augments perforin- 52 Kelly RA, Smith TW. Cytokines and cardiac contractile dependent cytotoxicity of liver NK-T cells. J Immunol function. Circulation 1997;95:778–81. 1998;161:2217–22. 53 Schluter KD, Weber M, Schraven E, Piper HM. NO donor 27 Campbell E, Kunkel SL, Strieter RM, Lukacs NW. SIN-1 protects against reoxygenation-induced cardiomyo- DiVerential roles of IL-18 in allergic airway disease: induc- cyte injury by a dual action. Am J Physiol 1994;267: on October 1, 2021 by guest. Protected copyright. tion of eotaxin by resident cell populations exacerbates H1461–6. eosinophil accumulation. J Immunol 2000;164:1096–102. 54 Finkel MS, Oddis CV, Jacob TD, Watkins SC, Hattler BG, 28 Novick D, Engelmann H, Wallach D, Rubinstein M. Soluble Simmons RL. Negative inotropic eVects of cytokines on the cytokine receptors are present in normal human urine. J heart mediated by nitric oxide. Science 1992;257:387–9. Exp Med 1989;170:1409–14. 55 Chandrasekar B, Streitman JE, Colston JT, Freeman GL. 29 Novick D, Engelmann H, Wallach D, Leitner O, Revel M, Inhibition of nuclear factor kappa B attenuates proinflam- Rubinstein M. Purification of soluble cytokine receptors matory cytokine and inducible nitric-oxide synthase from normal human urine by ligand-aYnity and immu- expression in postischemic myocardium. Biochim Biophys noaYnity chromatography. J Chromatogr 1990;510: Acta 1998;1406:91–106. 331–7. 56 Piot CA, Martini JF, Bui SK, Wolfe CL. Ischemic precondi- 30 Engelmann H, Novick D, Wallach D. Two tumor necrosis tioning attenuates ischemia/reperfusion-induced activation factor-binding proteins purified from human urine. Evi- of caspases and subsequent cleavage of poly(ADP-ribose) dence for immunological cross-reactivity with cell surface polymerase in rat hearts in vivo. Cardiovasc Res 1999;44: tumor necrosis factor receptors. J Biol Chem 1990;265: 536–42. 1531–6. 57 Holly TA, Drincic A, Byun Y, Nakamura S, Harris K, 31 Novick D, Cohen B, Rubinstein M. The human interferon Klocke FJ, et al. Caspase inhibition reduces myocyte cell á/â receptor: characterization and molecular cloning. Cell death induced by myocardial ischemia and reperfusion in 1994;77:391–400. vivo. J Mol Cell Cardiol 1999;31:1709–15. 32 Novick D, Kim S-H, Fantuzzi G, Reznikov L, Dinarello CA, 58 Okamura T, Miura T, Takemura G, Fujiwara H, Iwamoto Rubinstein M. Interleukin-18 binding protein: a novel H, Kawamura S, et al. EVect of caspase inhibitors on myo- modulator of the Th1 cytokine response. Immunity cardial infarct size and myocyte DNA fragmentation in the 1999;10:127–36. ischemia-reperfused rat heart. Cardiovasc Res 2000;45: 33 Kim S-H, Eisenstein M, Reznikov L, Fantuzzi G, Novick D, 642–50. Rubinstein M, et al.Structural requirements of six naturally 59 Schulz R, Panas DL, Catena R, Moncada S, Olley PM, occurring isoforms of the interleukin-18 binding protein to Lopaschuk GD. The role of nitric oxide in cardiac depres- inhibit interleukin-18. Proc Natl Acad Sci USA 2000;97: sion induced by interleukin-1 beta and tumour necrosis 1190–5. factor-alpha. Br J Pharmacol 1995;114:27–34.

www.annrheumdis.com