Immunity, Vol. 22, 131–142, January, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.immuni.2004.12.003 -32: A and Inducer of TNF␣

Soo-Hyun Kim,1,* Sun-Young Han,1 Tania Azam,1 (Aksentijevich et al., 2002; Hawkins et al., 2004) reduces Do-Young Yoon,2 and Charles A. Dinarello1 the signs and symptoms of these diseases as well as 1Department of Medicine the immunological and biochemical markers of inflam- University of Colorado Health Sciences Center mation. Blocking IL-18 in animal models of disease has Denver, Colorado 80262 reduced disease severity, including arthritis (Banda et 2 Korea Research Institute of Bioscience al., 2003; Plater-Zyberk et al., 2001), inflammatory bowel and Biotechnology disease (Siegmund et al., 2004), graft versus host dis- Yuseong Daejon 305-600 ease (Min et al., 2004), ischemia-reperfusion injury (Pom- South Korea erantz et al., 2001), and spontaneous atherosclerosis (Mallat et al., 2001). Therefore, IL-18-inducible might contribute to autoimmune and inflammatory dis- Summary eases. However, there is an obstacle to studying IL-18- inducible genes because IL-18 activity often requires We describe the structure, regulation, signal costimulatory factors, such as IL-2, IL-12, or IL-15, in transduction. and functions of a cytokine, interleukin order to manifest its responsiveness (Ahn et al., 1997; (IL)-32. An IL-18 unresponsive cell was converted to Hoshino et al., 1999; Lauwerys et al., 2000; Ohtsuki et a responsive cell by transfection of the IL-18 receptor al., 1997). This requirement of costimulatory factors pre- ␤ chain, and IL-18-induced microarray revealed high vents an independent assessment of solely IL-18-induc- expression of a cytokine-like gene. Although IL-32 ible genes. does not share with known cyto- IL-18 has two known receptor chains, a ligand binding kine families, IL-32 induces various , human IL-18R␣ chain and a signal-transducing IL-18R␤ chain. TNF␣, and IL-8 in THP-1 monocytic cells as well as Both chains of the IL-18 receptor belong to the IL-1 mouse TNF␣ and MIP-2 in Raw cells. IL- receptor family and consist of three Ig-like domains in 32 activates typical cytokine signal pathways of nu- the extracellular region (Kato et al., 2003; Yamamoto et clear factor-kappa B (NF-␬B) and p38 -acti- al., 2004). An additional component of IL-18 complex is vated kinase. IL-32 mRNA is highly expressed IL-18 binding protein (IL-18BP), which is not a part of in immune tissue rather than other tissues. Human IL- the IL-18 signaling complex but, rather, binds and neu- 32 exists as four splice variants, and IL-32 from other tralizes IL-18 activity (Kim et al., 2000; Novick et al., species were found as expressed sequence tag clones 1999). IL-18BP is a secreted receptor-like molecule, in the databank. Induced in human peripheral lympho- which consists of a single Ig-like domain. IL-18BP cyte cells after mitogen stimulation, in human epithe- shares a significant homology with viral , which lial cells by IFN␥, and in NK cells after exposure to the combination of IL-12 plus IL-18, IL-32 may play a role neutralize the biological activities of human IL-18 (Es- in inflammatory/autoimmune diseases. teban and Buller, 2004; Esteban et al., 2004; Reading and Smith, 2003; Xiang and Moss, 1999). Introduction In the absence of costimulants, nonimmune cells do not respond to IL-18 due to a low level or absence of ␤ IL-18 is a multifunctional cytokine having roles in both IL-18R chain (Chandrasekar et al., 2003). Therefore, it innate and adaptive immune responses. IL-18 has been was necessary to generate a stable clone expressing the ␤ studied for its effects in the broad spectrum of Th1- IL-18R chain, an IL-18 receptor component required for or Th2-related autoimmune diseases (Nakanishi et al., transmitting an IL-18 signal (Azam et al., 2003; Debets 2001; Okamura et al., 1995, 1998). IL-1 and IL-18 belong et al., 2000; Kim et al., 2001b). Stable expression of the to the IL-1 ligand family because they share structural IL-18R␤ chain was accomplished in the human lung similarity and require caspase-1 for processing (Bazan carcinoma A549 cell (A549-R␤), which converted these et al., 1996; Gu et al., 1997). IL-18 also uses a similar cells to IL-18-responsive cells in the absence of a costi- signaling pathway as that of IL-1 family members. They mulant. The A549-R␤ stable clone was used for IL-18- both recruit IL-1 receptor-associated kinases (IRAKs), inducible genes with microarrays independent of costi- form IRAK complexes with the mulants. The IL-18-induced microarray study revealed (TNF) receptor-associated factor-6, and activate the the induction of a cytokine-like molecule, which was cascade of inhibitor of kappa B/nuclear factor-kappa B described 12 years ago as transcript (IkBa/NF-kB). (Kojima et al., 1998; Matsumoto et al., 4 (NK4) (Dahl et al., 1992). Recently, increased gene 1997; Robinson et al., 1997). expression of NK4 has been reported in peripheral blood Members of the IL-1 ligand family play pathological mononuclear cells (PBMC) from patients receiving high- roles in autoimmune and inflammatory diseases (Dina- dose IL-2 therapy for malignant melanoma (Panelli et rello, 2004). Blocking IL-1 in humans with rheumatoid al., 2002), but the function of NK4 has remained unknown arthritis (Bresnihan et al., 1998), ankylosing spondylitis until the present study. We demonstrate here that NK4 (Haibel et al., 2005), and mutations in the NALP3 gene is an (IL-32) and describe its bio- logical function, regulation of production, genomic *Correspondence: [email protected] structure, and signal transduction. Immunity 132

Figure 1. The Reconstitution of Functional IL-18R␤ Chain in A549 Human Lung Carcinoma Cells and Microarray Study (A) Human IL-18R␤ expression in A549 cells. RT-PCR of IL-18R␤ in the A549 cell line (A549-R␤). (B and C) Stable clone (A549-R␤) was tested for the induction of IL-6 and IL-8 after IL-18 (50 ng/ml) stimulation for 16 hr (n ϭ 7). (D) IL-18-inducible cytokine-related genes by microarray. A549-R␤ cells were stimulated with IL-18 (50 ng/ml) for 6 hr then subjected for microarray analysis. The figure shows that only cytokine-related genes were induced over 3-fold (log base 2) and added the new nomenclature of cytokines (Bacon et al., 2002). The data present two independent experiments. (E) The induction of NK4 in the same RNA used for IL-18-inducible gene microarray analysis.

Results al., 1992), and yet no functional data exist on this gene product. The NK4 mRNA was examined by RT-PCR with Identification of the Cytokine the same RNA used for the microarray study. The tran- Although the human lung carcinoma A549 cell line script of NK4 was detected only in IL-18-treated A549- (A549-WT) expresses the endogenous IL-18 receptor ␣ R␤ cells and not in unstimulated A549-R␤ cells (Fig- chain, the IL-18 receptor ␤ chain is not expressed as ure 1E). determined by RT-PCR (Figure 1A). After transfection We next examined the NK4 gene expression in the of IL-18R␤ in A549 (A549-R␤), A549-R␤ responded to human NK cell line (Hoshino et al., 1999) because the IL-18 in the absence of a costimulus and produced IL-6 NK4 gene was described as a NK cell product. The NK4 (Figure 1B) as well as IL-8 (Figure 1C), whereas the A549 gene was constitutively and highly expressed in this parent cell line (A549-WT) remained unresponsive to IL- cell line, which is maintained in conditioned medium 18. The A549-R␤ stable clone was used to study IL-18- containing both IL-2 and IL-15. The presence of IL-2 in inducible genes by microarray, which revealed that IL- the conditioned medium likely contributes to the high 18 induced several cytokine genes including IL-6 and level of NK4 gene expression (data not shown). IL-8 (Figure 1D). Among IL-18-inducible genes, which Because the biological function(s) of NK4 has re- were greater than 3-fold (log base 2), there were several mained unknown, the RT-PCR product of NK4 gene , IL-1␤, and granulocyte colony-stimulating was cloned for the purpose of producing a recombinant factor, each known as an IL-18-inducible gene. How- protein and to examine its biological activities. The re- ever, NK4 was highly induced at 5.3-fold (log base 2). The combinant NK4 protein was purified by using three se- NK4 gene was first described 12 years ago (accession quential steps: His6-tag affinity, size exclusion, and ion numbers MN-004221 and AB007827) as a transcript in exchange chromatography. This process yielded a ho- IL-2-activated NK or mitogen-stimulated T cells (Dahl et mogenous band (Figure 2, top). The Raw 264.7 macro- Interleukin-32: A TNF␣-Inducing Factor 133

databank, and the blast program revealed an expressed sequence tag (EST) clone of equine, bovine, ovine, and swine IL-32. The equine IL-32␤ (AC: CD469554 and BI961524) shares the highest homology with human (31.8% identity) followed by bovine (AC: CK832489), ovine (AC: CO202364), and swine (AC: CB287292); mouse EST clones were not available. The multiple align- ments present three species of IL-32␤ human, equine, and bovine (Figure 3C). The ovine and swine have only isoform A. Equine has isoform A (AC: CD469554) and B, whereas bovine has isoform B and C. The sequence of equine IL-32␤ or bovine IL-32␥ was combined in two EST clones: CD469554 and BI961524 for equine IL-32␤ or CK834399 and CK832489 for bovine IL-32␥. The open reading frame of equine IL-32␤ or bovine IL-32␥ was deduced from the combined sequence.

The Biological Activity of rIL-32 and Antibody Neutralization Assay IL-32␣ induced significant amounts of TNF␣ and MIP-2 Figure 2. The Biological Activities of Recombinant IL-32␣ in Macro- and increased the production of both cytokines in a phage Raw 264.7 Cells dose-dependent manner (Figure 4A). We next compared ␣ ␤ The three step purified recombinant IL-32␣ was examined for biolog- the biological activity of IL-32 with that of IL-32 , which ical activity measuring the induction of TNF␣. The E. coli recombi- has a full C terminus. IL-32␤ induced similar levels of nant IL-32␣ induced TNF␣ from Raw cells. The data represent TNF␣ and MIP-2, as did IL-32␣ in the mouse macro- mean Ϯ SEM of five separate experiments. Each ion exchange chro- phage Raw cell line (Figure 4B). Although experiments matograph fraction is shown after 10% SDS-PAGE with Coomassie were performed in the presence of polymyxin B (5 blue staining in the upper panel. ␮g/ml), we cannot rule out endotoxin contamination in recombinant proteins from E. coli. Therefore, we pro- duced recombinant IL-32 in mammalian cells. The rIL- phage cell line responded to stimulation of recombinant 32␤ was produced from three different sources of mam- NK4 and produced large amounts of TNF␣ (Figure 2, malian cells: a stable clone of Anjou65 (a derivative of bottom), which coincided with the peak fractions of pro- the human epithelial 293T cell line), Cos7 stable clone tein eluting from ion exchange chromatography. At this (Cos7-S), and Cos7 transient transfection (Cos7-T). The point, we recognized that the recombinant NK4 protein maximum yield of mammalian rIL-32␤ in supernatants possesses a typical cytokine activity and renamed the was as low as 1 ng/ml, and IL-32␣ was below 100 pg/ml. molecule as IL-32. Therefore, each mammalian rIL-32␤ was purified with an immobilized anti-IL-32␤ monoclonal antibody affinity Genomic Structure of IL-32 and IL-32 column in the presence of sodium azide (0.02%) to pre- from Other Species vent microbial growth during purification. Each of the The genomic structure of IL-32 and its localization in three mammalian rIL-32␤ induced both human and human were analyzed. The blast program mouse TNF␣ in PMA-differentiated THP-1 cells and in the NCBI gene databank revealed that IL-32 resides in mouse Raw cells, respectively (Figure 4C). E. coli rIL- the human 16 p13.3. IL-32 has four splice 32␣ and mammalian rIL-32␤ increased human TNF␣ variants due to mRNA splicing, as shown in Figure 3A. (Figure 4D) and human IL-8 production (data not shown) ␣ ␤ ␦ Three splice variants, IL-32 , IL-32 , and IL-32 , were in PMA-differentiated THP-1 cells in a dose-dependent isolated from human NK cells by RT-PCR, whereas the manner. Furthermore, the high concentrations of E. coli ␥ IL-32 variant was previously reported as the NK4 tran- rIL-32␣ and mammalian rIL-32␤ of Anjou65 induced hu- script. IL-32 is organized with eight small exons; the man IL-8 in nondifferentiated THP-1 cells, but superna- ␥ second exon contains an ATG codon. IL-32 has an tant of affinity-purified mock-transfected Anjou65 did additional 46 amino acids in the N terminus because not induce IL-8 (Figure 4E). of the absence splicing between exon 3 and exon 4. The affinity-purified monoclonal antibody (mAb; 32-4) However, in IL-32␦ the second exon is absent, which against rIL-32␤ resulted in a high background level of results in a shift ATG codon in the third exon. Because IL- mTNF␣. Therefore, we generated the FabЈ fragment of 32␣ is the most abundant cDNA clone, the recombinant the mAb. The background level was reduced dramati- protein of this isoform was used for the experiments. cally after removing the Fc fragment. We assessed neu- IL-32␣ has a deletion of 57 amino acid residues in the tralizing rIL-32␣ and rIL-32␤ activities with the FabЈ mAb. C terminus (Figure 3B) due to the splicing of exons 7 Figure 4F depicts FabЈ mAb (50 ng/ml) reduction in the and 8, in contrast to the other splice variants, which biological activity of E. coli rIL-32␣ of more than 70% have a large single exon at the end. The analysis of in a dose-dependent manner. In addition, the same Fab the IL-32 amino acid sequence revealed three potential inhibited the affinity-purified rIL-32␤ Cos7-S activity by N-myristoylation sites and one N-glycosylation site that greater than 65% (Figure 4G). A higher concentration are indicated (Figure 3B). of FabЈ mAb did not reduce further IL-32 activities in We searched for other species of IL-32 in the NCBI these assays. Immunity 134

Figure 3. The Genomic Organization of Four IL-32 Splice Variants and Other Species of IL-32␤ by Amino Acid Sequence Alignment (A) IL-32 resides in human chromosome 16p13.3. The four IL-32 splice variants cDNA sequences were aligned with the 5 kb genomic sequence. (B) The multiple alignments present four human IL-32 splice variants. A potential posttranslational modification site was analyzed (http:// ca.expasy.org), and Myr or Gly indicates three potential N-myristoylation or one N-glycosylation site, respectively. (C) The alignment exhibits three IL-32␤ species: human, equine, and bovine. The consensus amino acid residue was shown at bottom of the each line. The alignment has been done by using the LALIGN program (www.ch.embnet.org/software/LALIGN_form.html) and manually corrected.

Expression of IL-32 mRNA and IL-32 Protein N-glycosylation and three N-myristolation sites as indi- We analyzed the expression of IL-32 in various human cated (Figure 3B). tissues by Northern blot. The 1.2 kb IL-32 transcript was We further examined the induction of IL-32 in human detected in several tissues, but IL-32 expression was epithelial Wish cells treated with IFN␥ because the Wish more prominent in immune cells than in nonimmune cells are commonly used for antiviral assay and other tissues (Figure 5A). The high expression of IL-32 mRNA biological activities of IFN␥. As shown in Figure 5C, IL- in immune tissues is not due to a difference in the 32 was induced in cell lysates of human epithelial cells amount of mRNA in each lane. in a time-dependent manner and diminished after 46 hr, Endogenous expression of IL-32 was also detected whereas IL-32 was not induced in unstimulated cells. at the protein level. A549-R␤ cells were treated with IL-18 and IL-1␤ increased the expression of endogenous either IL-18 or IL-1␤ as indicated (Figure 5B). Endoge- IL-32 in the lysates of A549-R␤ cells in a time-dependent nous IL-32 (27 kDa) was detected by immunoblot by manner (Figure 5D), whereas IL-32 was induced by only using affinity-purified rabbit anti-human IL-32␣ poly- IL-1␤ in A549-WT (Figure 5E), similar to IL-6 and IL-8 clonal antibody. Although IL-32 does not possess a typi- products from this cell line (Figures 1B and 1C). In addi- cal hydrophobic signal peptide in its N terminus, IL-32 tion, we observed IFN␥-induced IL-32 in the lysate of was detected as a secreted molecule. By Western blot A549-WT (data not shown). analysis, IL-32 was present in the culture medium of It was next determined whether the recombinant pro- A549-R␤ cells after 48 hr of stimulation in the presence teins from the cDNA of IL-32␣ and IL-32␤ expression in of IL-18 or IL-1␤, but there was no synthesis in unstimu- mammalian cells were comparable to endogenous IL- lated cells. A difference in molecular weight between 32. Cos7 cells were transiently transfected with IL-32␣ the recombinant IL-32 from E. coli (right lane of Figure and IL-32␤ cDNAs. Both cell culture media and lysates 5B) and endogenous IL-32 may be due to posttransla- were used to detect the recombinant IL-32 by immu- tional modifications. For example, the amino acid se- noblot or quantified by using a specific ECL assay. IL- quence analysis revealed that IL-32 has one potential 32␣ and IL-32␤ were present in the culture media (Figure Interleukin-32: A TNF␣-Inducing Factor 135

Figure 4. Biological Activities of rIL-32 and FabЈ Monoclonal (32-4) Antibody Neutralization Assay The rIL-32 induces proinflammatory cytokines in macrophage Raw and human THP-1 cells. (A) Dose-dependent E. coli derived IL-32␣-induced TNF␣ and MIP-2. IL-32 concentrations are indicated in units/ml on the horizontal axis. (B) Both IL-32␣ (10 U/ml) and IL-32␤ (10 U/ml) variants are active in Raw cells. (C) Human and mouse TNF␣ induced by various sources of mammalian IL-32␤. Raw and PMA-differentiated THP-1 cells were treated with 1 U/ml of Anjou65 and Cos7-S, or 2 U/ml of Cos7-T. (D) Dose-dependent E. coli IL-32␣ and mammalian IL-32␤ (Cos7-T) induced human TNF␣ in PMA-differentiated THP-1 cells. The units are indicated at the bottom of the graph. (E) The high concentrations of E. coli rIL-32␣ (20 U/ml) or mammalian rIL-32␤ (Anjou65, 10 U/ml) induced human IL-8 in nondifferentiated THP-1 cells. (F) Dose-dependent FabЈ mAb neutralization of E. coli rIL-32␣ (3 U/ml). The FabЈ concentration is indicated at the bottom of the graph. (G) Mammalian rIL-32␤ (Cos7-S, 2 U/ml)-induced mTNF␣ is shown in the presence of FabЈ mAb (40 ng/ml). The data represent mean Ϯ SEM of three separate experiments.

5F) and lysates (Figure 5G). The molecular size of recom- activity, we stimulated NK cells in the presence of a binant mammalian IL-32 in the immunoblot is identical neutralizing anti-huIFN␥ (Choi et al., 2001), and there in molecular size to the endogenous IL-32 shown in was no effect on IL-32 induction (data not shown). Figure 5B. The ECL assay revealed similar distribution Human PBMC contain mostly T cells with fewer num- of IL-32␣ and IL-32␤ in the same culture media and bers of monocytes and B cells. Freshly prepared PBMC lysates. IL-32␤ was more efficiently secreted compared were stimulated with LPS or Con A, and the supernatants to IL-32␣, which was mostly associated with the cell and lysates were assayed for level of IL-32. After 60 hr, (Figure 5H). there was no detectable IL-32 in the supernatants or lysates of PBMC stimulated with LPS (data not shown). Production of IL-32 in NK Cells and Primary PBMC However, Con A consistently induced IL-32, which was found in the supernatants and lysates primarily after 60 To determine the regulation of IL-32 production, we ex- hr. The lysates contained more IL-32 than the superna- amined different cell lines as well as human PBMC. IL- tants (Figure 6C). 32 was detected in the cell culture medium of A549-R␤ We further examined Con A-induced TNF␣ in human after IL-18 or IL-1␤ stimulation, but not in control me- PBMC through IL-32. Freshly isolated PBMC were stim- dium (Figure 6A). Because the NK4 (IL-32) gene was ulated with Con A in the presence of various concentra- isolated from IL-2-activated NK cells, we stimulated the tions of the same FabЈ anti-IL-32␤ that neutralized the human NK cell line with the combination of IL-12 plus biological activity of rIL-32 (Figures 4F and 4G), but there IL-18. As shown in Figure 6B, there was a significant was no effect on Con A-induced TNF␣ or IFN␥ produc- induction of IL-32 by IL-12 or IL-18. The effect of the tion (data not shown). combination of these two cytokines appeared additive. By comparison, when this cell line is stimulated by either IL-12 or IL-18, there is little or no induction of IFN␥, Signal Transduction of IL-32 whereas IFN␥ production by the combination of IL-12 Signal transduction of the cytokine IL-32 was also stud- plus IL-18 is highly synergistic (Kim et al., 2002). To ied. The degradation I␬B was examined because IL-32 examine the induction of IL-32 as a function of IFN␥ induced TNF␣ and MIP2, two cytokines known to be Immunity 136

Figure 5. The Expression and Regulation of IL-32 Endogenous expression of IL-32 at both mRNA and protein levels. (A) The Northern blot depicts the expression of IL-32 mRNA in various human tissues as listed in each lane. The numbers on the most left lane indicate the size of mRNA and the /Northern blot. (B) The detection of released endogenous IL-32 in the cell culture supernatant. The A549-R␤ stable clone was stimulated for 48 hr in the presence of the following stimuli, IL-18 (50 ng/ml) or IL-1␤ (10 ng/ml) as indicated. The supernatants were harvested then probed with the affinity-purified anti-IL-32␣ antibody. (C–E) The detection of IL-32 in cell lysates at various times as indicated. (C) Human Wish cells were stimulated with IFN␥ (100 U/ml) or (D) A549- R␤, and (E) A549-WT cells were stimulated with IL-18 (50 ng/ml) or IL-1␤ (10 ng/ml). The data represent one of three independent experiments. (F and G) IL-32␣ and ␤ cDNA were transiently transfected into Cos7 cells. After 20 hr of transfection, immunoblot or electrochemiluminescence (ECL) analysis of cell culture supernatants and lysates. (F) Cell culture supernatants and (G) lysates were subjected for immunoblot with anti- IL-32 antibody. (H) ECL measured IL-32 concentration in the same supernatants and lysates. The data represent mean Ϯ SEM of five separate experiments. activated by NF-␬B. As shown in Figure 7A, IL-32␣ in- production, signal transduction, and function of IL-32 duced I␬B degradation in a time-dependent manner, are described. IL-32 was deposited in the databank 12 beginning 15 min after exposure and reaching a maximal years ago as NK4, a transcript of activated human T cells level at 45 min followed by recovery after 90 min. The and NK cells after stimulation by or IL-2, re- phosphorylation of p38 MAPK, another characteristic spectively (Dahl et al., 1992). It should be noted that the of the inflammatory cytokine signal pathway, was also term NK4 is presently used to describe a variant of studied. Phosph-p38 MAPK was dramatically increased hepatocyte (Martin et al., 2002). There is 5 min after stimulation by IL-32␣ and then decreased no sequence homology between the hepatocyte growth from 15 to 30 min thereafter (Figure 7B). Interestingly, factor variant and NK4/IL-32. Although IL-32 lacks se- a second increase in p38 MAPK phosphorylation was quence homology to the presently known cytokine fami- observed at 45 min and then decreased slowly. lies, its ability to induce the proinflammatory cytokines such as TNF␣, IL-8, and MIP-2 in different cells and to Discussion signal through the classical signal pathways of proin- flammatory cytokines places the gene in the class with In the present study, we add a cytokine, which we have other cytokines. termed IL-32, to the growing number of molecules with The discovery of IL-32 as a cytokine was carried out functions for both innate and acquired immune systems. with the specific goal of understanding IL-18-inducible The gene structure, splice variants, tissue distribution, genes in the absence of costimulants. Therefore, we Interleukin-32: A TNF␣-Inducing Factor 137

Figure 6. The Measurement of Endogenous IL-32 by ECL (A) IL-32 was detected by ECL assay with the same samples that were used for the immunoblot of Figure 5B. (B) IL-12, IL-18, or IL-12 plus IL-18 induced IL-32 in the supernatant of human NK cell line. The data represent mean Ϯ SEM of four separate experiments. (C) Con A-induced IL-32 was detected in both supernatants and lysates of human PBMC (n ϭ 7). designed an IL-18-responsive cell line (A549-R␤) and monocytic THP-1 cells, which may implicate a specific subjected the IL-18-stimulated mRNA for gene chip IL-32 receptor expression only in these differentiated analysis. Several of the transcripts increased by IL-18 cells. The high concentrations of E. coli rIL-32␣ and were not unexpected, but NK4 was induced to a rela- mammalian rIL-32␤ provoked a weak response in non- tively high level and was without known function. There- differentiated THP-1 cells (Figure 4E). High concentra- fore, the first approach was to express the molecule tions of both E. coli and mammalian rIL-32 did not induce and examine the function(s) of the recombinant protein. cytokines in freshly obtained human PBMC, further sup- The homogenous protein peak eluting from the purifica- porting the observation that biological activity of rIL-32 tion coincided with the induction of TNF␣ by macro- is not due to endotoxin contamination. phages (Figure 2). IL-32 triggered the typical cytokine We tested two mAbs (clones 32-4 and 32-9) against signal pathways of NF␬B and p38 MAPK and also in- E. coli rIL-32␤, and both antibodies inhibited IL-32 activ- duced TNF␣ and chemokines. This is a demonstration ity about 30% (data not shown). However, the high con- of a biological function and signaling transduction of IL- centration of both mAbs induced mTNF␣, which is un- 32 as that of a typical proinflammatory cytokine. usual because the same species of mAb that stimulate We produced mammalian rIL-32␤ from different mouse Raw cells do not often also result in TNFa induc- sources. The mammalian rIL-32␤ induced human and tion. Therefore the Fab fragment of mAb (32-4) was pre- Ј mouse TNF␣ in PMA-differentiated human monocytic pared and used for neutralization assay. The Fab mono- THP-1 cells and mouse Raw cells. Affinity-purified su- clonal antibody neutralized 75% of IL-32 biological pernatant from mock-transfected cells did not induce activity in mouse Raw cells (Figure 4F), and a similar TNF␣ (Figure 4C). IL-32 is active on highly differentiated result was observed with MIP-2 in the same assays (data cells such as mouse macrophage Raw and human not shown). The original description of NK4 was reported in IL-2 and mitogen-activated NK or T cells. In the present study, we found that IL-32 is expressed constitutively in various organs (Figure 5A), which is similar to IL-15. IL-15 is produced by a wide variety of tissues, whereas IL-2 is exclusively produced by activated T cells (Bam- ford et al., 1996; Grabstein et al., 1994). The induction of IL-32 in either immune or nonimmune cells may be an important property for the cytokine (Figure 5). We have detected elevated levels of IL-32 in circulation of some, but not all, patients with sepsis compared to healthy individuals (data not shown), but this finding is not unexpected and is rather consistent with several cy- tokines. The secretion of IL-32 in the absence of a clear signal peptide is characteristic of several cytokine families (Cerretti et al., 1992; Ghayur et al., 1997; Kuida et al., Figure 7. IL-32␣-Induced I␬B Degradation and Phosphorylation of 1995). Similar to IL-1␤ and IL-18, IL-32 is secreted as a p38 MAPK soluble protein in cell culture media of stimulated pri- ␣ Raw cells were stimulated with IL-32 (20 U/ml) at the indicated mary cells as well as cell lines (Figures 5 and 6). IL-32 times. does not possess potential caspase-1 cleavage sites (A) The membrane was probed with rabbit anti-IkB and then normal- ized with goat anti-actin. by primary sequence analysis. New splice variants have (B) The membrane was probed with rabbit anti-phospho-p38 MAPK been found, and none of the IL-32 splice variants pos- and then normalized with rabbit anti-p38 MAPK. sess a typical hydrophobic signal peptide at the N termi- Immunity 138

nus. IL-32␤ isoform was sufficiently secreted to cell cul- nent source of the cytokine. Human peripheral T cells ture supernatant by transient transfection into Cos7 cell, stimulated with Con A are a source of TNF␣ (Brod et al., though no signal peptide exists at the N terminus. It 1996; Chang and Shaio, 1995), but the link of mitogen- remains unclear how IL-32 is secreted from stimu- activated CD4ϩ T cells by Con A is best found in an lated cells. animal model of Con A-induced hepatitis (Kusters et al., The existence of IL-32 in other species is evidence 1997; Tiegs et al., 1992). In that model, either IL-18 or of an evolutionarily conserved molecule that may have TNF␣ neutralization prevents the development of the important functions in immune responses (Figure 3C). disease (Fantuzzi et al., 2003; Tiegs et al., 1992). It is Not as many IL-32 isoforms are found in other species possible that Con A-induced TNF␣ involves IL-32 and as in the human because in other species, the EST data- is based on the ability of Con A-stimulated CD4ϩ T cells bank is not sequenced as extensively as in the human. to mediate the hepatitis (Tiegs et al., 1992) and to pro- Alternatively, they do not exist as in the case of human duce IL-32. IL-32, in turn, would activate macrophagic IL-18BP, which exists as four isoforms, whereas mouse Kupffer cells to release TNF␣ and cause hepatocyte cell has only two isoforms (Kim et al., 2000; Novick et al., death. Such a mechanism is supported by the findings 1999). Intensive searching for a mouse IL-32 homolog in the present paper that recombinant IL-32 induced with various IL-32 species did not reveal a mouse homol- TNF␣ from differentiated . ogous gene. It is possible that a mouse IL-32 gene has Although several immune regulatory molecules have very low homology with human because equine or bo- been discovered during the past decade, many phenom- vine shares only 31.8%–28.1% identities, respectively, ena remain unexplained in understanding immune regu- with human IL-32, which is contrary to other cases of lation. The functional portfolio of the cytokine IL-32 may evolutionary well-conserved genes. help unravel the complexity of cytokine biology. The After transient transfection into Cos7 cells, IL-32␣ iso- identification of IL-32 receptor components, functional form remained inside the cell unlike IL-32␤ (Figure 5H). characterization of the different splice variants, and Also, IL-32 was detected in cell lysates after stimulation analysis of the variation of IL-32 expression in diseases (Figures 5C–5E). Therefore, we cannot exclude the pos- will contribute to understanding a precise role of IL-32 sibility of an intracellular function for IL-32. There are in the network of cytokine biology. examples of such functions; cytokines such as IL-1␣ (Stevenson et al., 1997) and high mobility group-1 (Wang Experimental Procedures et al., 1999) are active as intracellular as well as extracel- lular proteins (Werman et al., 2004). Stable Expression of IL-18R␤ Chain in A549 Splice variants are unusual in cytokine genes; never- Human lung carcinoma A549 cells (3 ϫ 105 per well) were seeded theless, splice variants exist in cytokines such as IL-15, in a 6-well plate a day before transfection. The cells were washed with 2 ml of Opti-MEM (Invitrogen, Carlsbad, CA) and incubated for IL-1F7, and VEGF (Bufler et al., 2002; Kumar et al., 2002; 25 min in 1 ml of the same medium prior to transfection. In the Meazza et al., 1996; Nishimura et al., 1998; Pufe et al., meantime transfection mixture solution for each well was prepared 2001). IL-15 exists as two isoforms, a long signal peptide as follows. We mixed 5 ␮l of Lipofectamine 2000 in 100 ␮l of Opti- (48 amino acids) and a short signal peptide (21 amino MEM and then incubated the mixture for 5 min in a hood. Next, 2 ␮g acids) due to an in exon 5 (Tagaya of plasmid DNA, pTARGET/huIL-18R␤ (Kim et al., 2001b) was added et al., 1997). Although both IL-15 isoforms possess iden- and incubated for 20 min. The transfection mixture solution was added to a well, and the plate was incubated for an additional 4 tical biological properties, sharing the same receptors hr in a cell culture incubator. Adding 2 ml of cell culture medium and biological functions, each IL-15 isoform has a dis- containing 10% FCS in each well terminated the transfection. The tinct regulation and expression pattern. The long signal next day, cells were trypsinized and transferred in a plate (15 cm). peptide IL-15 isoform is more efficiently secreted in su- After cells adhered to the plate, the culture medium was replaced pernatant (Tagaya et al., 1997), which is similar to the with fresh medium containing 800 ␮g/ml of Neomycin (G-418, In- IL-32␤ isoform. vitrogen). The selection medium was exchanged every 3 days with Inflammatory cytokines, such as IL-1␤, IL-18, and fresh selection medium until individual colonies appeared, which took place between 10 and 14 days. Small pieces of circular 3 MM IFN␥, induced IL-32 from two different epithelial cell papers were sterilized and wet with trypsin and then used for picking types (Figures 5B–E), a common target in several inflam- individual colonies. Colonies were transferred into 24-well plates matory diseases. Moreover, most members of the vari- containing 1 ml of selection medium in each well. An individual clone ous cytokine families induce each other, and this also was grown until the cells reached a population of 106, and each served as part of the rationale for renaming the NK4 clone was tested for the RT-PCR of transfected gene IL-18R␤ as transcript as IL-32. In addition, IFN␥ induces IL-32 in a well as for bioassays measuring IL-6 and IL-8 in cell culture medium time-dependent manner (Figure 5C), suggesting a role after IL-18 stimulation for 24 hr. Three clones were positive in both the RT-PCR of IL-18R␤ and bioassays. Limiting dilutions in order for IL-32 in innate and adaptive immune responses. The to obtain a single clone cultured these clones. A single clone used ␣ induction of TNF by IL-32 suggests that IL-32 has an for this study is one of the subclones from the limiting dilution. important role in autoimmune/inflammatory diseases because blockade of TNF␣ activity is a highly effective Cells and Bioassays therapeutic approach in rheumatoid arthritis (Firestein, Human NK cell line was obtained from Dr. Hans Klingerman (Rush 2003), Crohn’s disease (van Dullemen et al., 1995), and Medical Center, Chicago, IL) and cultured in RPMI-1640 medium psoriasis (Leonardi et al., 2003). containing 10% FCS, IL-2 (50 pg/ml), and IL-15 (200 pg/ml) (Pepro- Similar to the original description of NK4, we observed tech, Rocky Hill, NJ). Mouse macrophage Raw 264.7, human A549 lung carcinoma, monkey kidney Cos7, Anjou65 (a derivative of the that Con A induced IL-32 from freshly isolated human human epithelial 293T cell line), human epithelial Wish, and human PBMC. However, in these same PBMC cultures, LPS monocytes THP-1 cell line were obtained from American Type Cul- was without effect. It is unlikely that peripheral blood ture Collection (ATCC) and maintained according to ATCC instruc- monocytes triggered by Toll-like receptors is a promi- tions. Bioassays were performed in 96-well plates. Briefly, Raw cells Interleukin-32: A TNF␣-Inducing Factor 139

(5 ϫ 105/ml, 0.1 ml/per well), A549-R␤ cells (2 ϫ 105/ml, 0.1 ml/per were subjected to ion-exchange chromatography (HiTrapQFF, well), and human NK cells (5 ϫ 105/ml, 0.1 ml/per well) were seeded A¨ KTAFPLC). The three step-purified recombinant proteins were in 96-well plates and cultured until cells adhered to the plate. For tested for biological activities. Raw cell assays, the medium was removed and then stimulated with fresh medium (0.2 ml) containing various concentrations of IL- Expression of Mammalian Recombinant Proteins 32 in the presence 5 ␮g/ml of polymyxin B (Bedford Laboratories, pTARGET/IL-32␤ containing His6-tag at the C terminus was tran- Bedford, OH). For A549-R␤ cells assays, the medium was removed siently transfected into Cos7 cells (8 ϫ 106) with DEAE-Dextran as and then stimulated with fresh medium (0.2 ml) containing IL-18, described (Kim et al., 2000; Novick et al., 1999). For stable clones, which was produced as described (Kim et al., 2001a). For NK cell pTARGET/Mock and pTARGET/IL-32␤-his6 were transfected into assays, cells were stimulated with IL-12 (Peprotech), IL-18, or both Cos7 cells (1 ϫ 106) or Anjou65 cells (1 ϫ 106) with Opti-MEM as cytokines. THP-1 cells (5 ϫ 105/ml, 0.1 ml/per well) were seeded in described above for the stable transfection of IL-18R␤. The transient 96-well plates in the absence of FCS and then stimulated with rIL- and stable clone cells were cultured in media (0.5% FCS) for 4 32s. For PMA-differentiated THP-1 assays, THP-1 cells (0.25 ϫ days then harvested for the next step of purification. The transient 105/ml, 0.1 ml/per well) were seeded in 24-well plates and stimulated transfection cell supernatants were purified as described for the with 100 ng/ml of Phorbol 12-myristate 13-acetate (PMA, Sigma, St three-step purification of E. coli rIL-32. The stable clone supernatant Louis, MO) for 48 hr. The plate was washed and left overnight in was purified with the affinity column of mAb anti-IL-32␤ immobilized medium without FCS and then treated with rIL-32 in fresh medium. to agarose beads (Affi-Gel Hz, Bio-Rad) according to manufacturer’s The plates were placed in a cell culture incubator for 16–20 hr, and instruction. The purified IL-32 was dialyzed against RPMI containing then the culture supernatants were collected for cytokine measure- penicillin/streptomycin at 4ЊC prior to bioassay. ments. Human PBMC were isolated from residual leukocytes after plate- Immunization and Fusion letpheresis of healthy donors with Histopaque (Sigma) and was A 5-week-old female Balb/c mouse was immunized with 20 ␮gof approved by Combined Colorado Investigational Review Board. rIL-32␤ antigen emulsified in Freund’s complete adjuvant (Sigma). ϫ 7 PBMC (1.5 10 ) were seeded in 6-well plates in 3 ml of RPMI On days 14 and 21, the mouse was given an intravenous and intra- ␮ containing 10% FCS then stimulated with 20 g/ml of Con A (Sigma). peritoneal injection with the antigen emulsified in Freund’s incom- The plates were placed in a cell culture incubator for 60 hr. The plete adjuvant (Sigma). After three injections, the mouse was sacri- culture supernatant and cell lysate were collected for IL-32 mea- ficed, the spleen was aseptically removed, and splenocytes were surement. prepared for fusion. Briefly, 1 ϫ 107 splenocytes and 1 ϫ 106 NS-1 IL-32 units were established for the biological activity of IL-32 mouse myeloma cell line (ATCC) were fused by using polyethylene because of differences between E. coli rIL-32 and mammalian rIL- glycol 1500 (Roche Applied Science, Indianapolis, IN). Fused cells 32. In addition, E. coli rIL-32 showed variations among different were resuspended at 1 ϫ 106 cells/ml in hybridoma growth media ␣ batches. One unit/ml of IL-32 is a 2-fold increase of mouse TNF with 10% FCS and HAT. Cells were then plated in 96-well plates. ␣ in mouse Raw cells or of human TNF in PMA-differentiated THP-1 After 2 weeks, the culture supernatants of hybridoma were screened cells. The conditions of assays are described above. Approximate with rIL-32␣ coated plates for direct ELISA. mAb classes and sub- concentration of 1 unit for E. coli rIL-32 is 20 ng/ml (mammalian rIL- classes were determined by using an Immuno-TypeTM mouse mAb 32 is 2 ng/ml) as calculated by using ECL and Western blot. isotyping kit (BD Bioscience, San Diego, CA) according to the manu- facturer’s instructions. Microarray A549-R␤ cells were seeded at 2 ϫ 106 in plates (9 cm) one day Production of Ascites and Purification of Antibodies before the experiment. Cells were stimulated with IL-18 (50 ng/ml) 6 Two hybridoma 32-4 (IgG1) and 32-9 (IgG1) clones (5 ϫ 10 cells) week-old female Balb/c-9 ف for 6 hr then total RNA was isolated with RNeasy kit (QIAGEN, were intraperitoneally injected into an 8 Valencia, CA) according to the manufacture’s instruction. Total RNA mouse. After a week, the ascites fluid was removed with a sterile (10 mg/ml) was converted to first strand cDNA by using Superscript needle. Supernatants were collected by centrifugation, purified by II RT (Invitrogen). The second-strand cDNA was synthesized with using protein A Sepharose (Bio-Rad), and eluted with 0.1 M glycine- the use of T4 DNA polymerase I. After second-strand synthesis, the HCl, (pH 2.7). reaction mixture was cleaned with a kit supplied by GeneChip. cRNA Immobilized IL-32 was also used to purify the FabЈ fragment of synthesis was performed to generate biotin-labeled cRNA by using a mouse monoclonal antibody (32-4). Clarified ascites fluid was RNA transcript-labeling kit. The biotinylated-cRNA was fragmented passed over the IL-32 ligand affinity column and eluted with 0.1 M prior to hybridization. The samples were hybridized to the Affymetrix glycine-HCl, (pH 2.7). The affinity-purified antibody was incubated GeneChip HG-U133, and data were analyzed in the microarray core with immobilized pepsin (Pierce, Rockford, IL) for 4 hr at 37ЊC. The laboratory of the University of Colorado Health Sciences Center. cleaved and noncleaved Fc products of digestion were then passed over protein G Sepharose (Amersham, Bioscience, Sweden). The RT-PCR and Cloning for E. coli or Mammalian Vectors FabЈ preparation was subjected to 7.5% SDS-PAGE to confirm ho- Total RNA was isolated with Tri-Reagent (Sigma, St Louis, MO) from mogeneity, and this preparation was used for neutralization. the human NK cells. A pair of sense primer, 5Ј-CTGTCCCGAGTCTG GACTTT-3Ј, and reverse primer, 5Ј-GCAAAGGTGGTGTCAGTATC- Northern Blot 3Ј, were used for the RT-PCR. Superscript II RT (Invitrogen) con- The cDNA probe of both IL-32␣ and actin (Azam et al., 2003) were verted 2 ␮g of total RNA to first strand cDNA, and then PCR reaction excised from plasmid vectors by using restriction enzymes, then was performed at 94ЊC for 45 s, 70ЊC for 2 min, and 59ЊC for 1 min the inserts of cDNA were purified with GeneClean II kit (Q-BIO gene, for 30 cycles. The PCR products were ligated into pGEMT-Easy Carlsbad, CA). The cDNA probes were 32P-dCTP-labeled (NEN Life (Promega, Madison, WI) for DNA sequencing and then IL-32 inserts Science, Boston MA) by random priming and by using Klenow (New were transferred into pPROEX/HTa (Invitrogen) for E. coli or pTAR- England Biolabs, Beverly, MA). A membrane containing human poly GET (Promega) for mammalian expression. (A)ϩ RNA from different tissues (human MTN Blot II, BD) was probed with the cDNA probes and autoradiographed. Expression of E. coli Recombinant Proteins Recombinant IL-32␣ and IL-32␤ proteins were expressed in E. coli Immunoblots and purified by a TALON affinity column (Invitrogen) by using His6- For detection of IkB degradation and phosphorylation of p38 MAPK, tag at the N terminus of recombinant proteins. The TALON affinity- Raw 264.7 macrophage cells were stimulated with IL-32␣ (20 U/ml) purified proteins were subjected to size exclusion chromatography in the presence of 5 ␮g/ml polymyxin B (Bedford Lab) at various (Superdex 75, A¨ KTAFPLC) and then digested with TEV (tobacco times. Cells were lysed with kinase lysis buffer (Han et al., 2002) etch virus, Invitrogen) for 16 hr at 4ЊC in order to remove the His6- then subjected to 10% SDS-PAGE. The samples were transferred tag fusion protein. The TEV-cleaved recombinant proteins were to nitrocellulose membranes. The membranes were blocked in 3% dialyzed in phosphate buffer (20 mM, [pH 9]). These materials BSA (Santa Cruz Biotechnology, Santa Cruz, CA). The first mem- Immunity 140

brane was probed with rabbit anti-IkB (Santa Cruz Biotechnology) lymphotrophic virus type I region/IL-15 fusion message that lacks and then normalized with goat anti-actin (Santa Cruz Biotechnol- many upstream AUGs that normally attenuates IL-15 mRNA transla- ogy). The second membrane was probed with rabbit anti-phospho- tion. Proc. Natl. Acad. Sci. USA 93, 2897–2902. p38 MAPK (, Beverly, MA) and then normalized with Banda, N.K., Vondracek, A., Kraus, D., Dinarello, C.A., Kim, S.H., rabbit anti-p38 MAPK (Cell Signaling). Bendele, A., Senaldi, G., and Arend, W.P. (2003). Mechanisms of ␤ For the detection of an endogenous IL-32, A549-R cells were inhibition of collagen-induced arthritis by murine IL-18 binding pro- 6 seeded at 10 cells per well in 6-well plates, and cell culture media tein. J. Immunol. 170, 2100–2105. were exchanged after the cells had adhered. The cells were stimu- Bazan, J.F., Timans, J.C., and Kastelein, R.A. (1996). A newly defined lated for 48 hr in the presence of the following stimuli, IL-18 interleukin-1? Nature 379, 591. (50 ng/ml) or IL-1␤ (10 ng/ml) in the absence of FCS. The superna- tants were harvested then concentrated ten times with Centricon Bresnihan, B., Alvaro-Gracia, J.M., Cobby, M., Doherty, M., Domljan, (YM-10, Life Sciences, Ann Arbor, MI) prior to detection of IL-32. Z., Emery, P., Nuki, G., Pavelka, K., Rau, R., Rozman, B., et al. For a time course of IL-32, Wish, A549-WT, or A549-R␤ cells were (1998). Treatment of rheumatoid arthritis with recombinant human seeded at 5 ϫ 104 cells per well in 6-well plates for overnight and interleukin-1 receptor antagonist. Arthritis Rheum. 41, 2196–2204. then stimulated with IFN␥ (100 U/ml, Peprotech), IL-18 (50 ng/ml), Brod, S.A., Marshall, G.D., Jr., Henninger, E.M., Sriram, S., Khan, M., or IL-1␤ (10 ng/ml). Affinity-purified rabbit anti-IL-32␣ was used for and Wolinsky, J.S. (1996). -beta 1b treatment decreases detecting endogenous IL-32 in cell culture medium or lysate by tumor necrosis factor-alpha and increases interleukin-6 production immunoblot. Peroxidase-conjugated secondary antibodies (Jack- in multiple sclerosis. Neurology 46, 1633–1638. son ImmunoResearch Laboratories, West Grove, PA) were used Bufler, P., Azam, T., Gamboni-Robertson, F., Reznikov, L.L., Kumar, to develop the blots by using enhanced chemiluminescence (NEN S., Dinarello, C.A., and Kim, S.H. (2002). A complex of the IL-1 Life Science). homologue IL-1F7b and IL-18-binding protein reduces IL-18 activity. Proc. Natl. Acad. Sci. USA 99, 13723–13728. Electrochemiluminescence A polyclonal antibody produced in rabbits against recombinant IL- Cerretti, D.P., Kozlosky, C.J., Mosley, B., Nelson, N., Van Ness, K., 32␣ was used for electrochemiluminescence (ECL). IgG was pre- Greenstreet, T.A., March, C.J., Kronheim, S.R., Druck, T., Canniz- pared with Econo-Pac and then affinity purified through IL-32␣- zaro, L.A., et al. (1992). Molecular cloning of the interleukin-1 beta immobilized agarose bead column of Affi-gel 15. An aliquot of the converting enzyme. Science 256, 97–100. affinity purified anti-IL-32␣ antibody was labeled with biotin and Chandrasekar, B., Colston, J.T., de la Rosa, S.D., Rao, P.P., and another aliquot with ruthenium according to the manufacturer’s in- Freeman, G.L. (2003). TNF-alpha and H2O2 induce IL-18 and IL- structions (BioVeris, Gaithersburg, MD). The biotin, and ruthenium, 18R beta expression in cardiomyocytes via NF-kappa B activation. labeled antibodies were used to construct a standard curve using Biochem. Biophys. Res. Commun. 303, 1152–1158. recombinant IL-32␣ for ECL assay. The liquid-phase ECL method Chang, F.Y., and Shaio, M.F. (1995). Decreased cell-mediated immu- was used to measure TNF␣, IL-6, IL-8, IL-32, and MIP-2 in cell culture nity in patients with non-insulin-dependent diabetes mellitus. Diabe- media and serum samples. The amount of ECL was determined by tes Res. Clin. Pract. 28, 137–146. using an Origen Analyzer (BioVeris). Choi, H.J., Dinarello, C.A., and Shapiro, L. (2001). Interleukin-18 inhibits human immunodeficiency virus type 1 production in periph- Acknowledgments eral blood mononuclear cells. J. Infect. Dis. 184, 560–568.

This study was supported by National Institutes of Health grants Dahl, C.A., Schall, R.P., He, H.L., and Cairns, J.S. (1992). Identifica- AI-15614 and HL-68743 (to C.A.D.). D.-Y.Y. was supported by a tion of a novel gene expressed in activated natural killer cells and KRIBB Research Initiative Program grant. T cells. J. Immunol. 148, 597–603. Debets, R., Timans, J.C., Churakowa, T., Zurawski, S., de Waal Received: June 24, 2004 Malefyt, R., Moore, K.W., Abrams, J.S., O’Garra, A., Bazan, J.F., and Revised: October 25, 2004 Kastelein, R.A. (2000). IL-18 receptors, their role in ligand binding Accepted: December 1, 2004 and function: anti-IL-1RAcPL antibody, a potent antagonist of IL- Published: January 25, 2005 18. J. Immunol. 165, 4950–4956. Dinarello, C.A. (2004). Therapeutic strategies to reduce IL-1 activity References in treating local and systemic . Curr. Opin. Pharmacol. 4, 378–385. Ahn, H.J., Maruo, S., Tomura, M., Mu, J., Hamaoka, T., Nakanishi, Esteban, D.J., and Buller, R.M. (2004). Identification of residues in K., Clark, S., Kurimoto, M., Okamura, H., and Fujiwara, H. (1997). an orthopoxvirus interleukin-18 binding protein involved in ligand A mechanism underlying synergy between IL-12 and IFN-gamma- binding and species specificity. Virology 323, 197–207. inducing factor in enhanced production of IFN-gamma. J. Immunol. Esteban, D.J., Nuara, A.A., and Buller, R.M. (2004). Interleukin-18 159, 2125–2131. and glycosaminoglycan binding by a protein encoded by Variola Aksentijevich, I., Nowak, M., Mallah, M., Chae, J.J., Watford, W.T., virus. J. Gen. Virol. 85, 1291–1299. Hofmann, S.R., Stein, L., Russo, R., Goldsmith, D., Dent, P., et al. Fantuzzi, G., Banda, N.K., Guthridge, C., Vondracek, A., Kim, S.H., (2002). De novo CIAS1 mutations, cytokine activation, and evidence Siegmund, B., Azam, T., Sennello, J.A., Dinarello, C.A., and Arend, for genetic heterogeneity in patients with neonatal-onset multisys- W.P. (2003). Generation and characterization of mice transgenic for tem inflammatory disease (NOMID): a new member of the expanding human IL-18-binding a. J. Leukoc. Biol. 74, 889–896. family of pyrin-associated autoinflammatory diseases. Arthritis Rheum. 46, 3340–3348. Firestein, G.S. (2003). Evolving concepts of rheumatoid arthritis. Azam, T., Novick, D., Bufler, P., Yoon, D.Y., Rubinstein, M., Dinarello, Nature 423, 356–361. C.A., and Kim, S.H. (2003). Identification of a critical Ig-like domain Ghayur, T., Banerjee, S., Hugunin, M., Butler, D., Herzog, L., Carter, in IL-18 receptor alpha and characterization of a functional IL-18 A., Quintal, L., Sekut, L., Talanian, R., Paskind, M., et al. (1997). receptor complex. J. Immunol. 171, 6574–6580. Caspase-1 processes IFN-gamma-inducing factor and regulates Bacon, K., Baggiolini, M., Broxmeyer, H., Horuk, R., Lindley, I., Man- LPS-induced IFN-gamma production. Nature 386, 619–623. tovani, A., Maysushima, K., Murphy, P., Nomiyama, H., Oppenheim, Grabstein, K.H., Eisenman, J., Shanebeck, K., Rauch, C., Srinivasan, J., et al. (2002). /chemokine receptor nomenclature. J. S., Fung, V., Beers, C., Richardson, J., Schoenborn, M.A., Ahdieh, Interferon Cytokine Res. 22, 1067–1068. M., et al. (1994). Cloning of a growth factor that interacts with Bamford, R.N., Battiata, A.P., Burton, J.D., Sharma, H., and Wald- the beta chain of the interleukin-2 receptor. Science 264, 965–968. mann, T.A. (1996). Interleukin (IL) 15/IL-T production by the adult Gu, Y., Kuida, K., Tsutsui, H., Ku, G., Hsiao, K., Fleming, M.A., Hay- T-cell leukemia cell line HuT-102 is associated with a human T-cell ashi, N., Higashino, K., Okamura, H., Nakanishi, K., et al. (1997). Interleukin-32: A TNF␣-Inducing Factor 141

Activation of interferon-gamma inducing factor mediated by in- Matsumoto, S., Tsuji-Takayama, K., Aizawa, Y., Koide, K., Takeuchi, terleukin-1beta converting enzyme. Science 275, 206–209. M., Ohta, T., and Kurimoto, M. (1997). Interleukin-18 activates NF- Haibel, H., Rudwaleit, M., Listing, J., and Sieper, J. (2005). Open kappaB in murine T helper type 1 cells. Biochem. Biophys. Res. label trial of in active ankylosing spondylitis over 24 weeks. Commun. 234, 454–457. Ann. Rheum. Dis. 64, 124–126. Meazza, R., Verdiani, S., Biassoni, R., Coppolecchia, M., Gaggero, Han, S.Y., Kim, S.H., and Heasley, L.E. (2002). Differential gene regu- A., Orengo, A.M., Colombo, M.P., Azzarone, B., and Ferrini, S. (1996). lation by specific gain-of-function JNK1 proteins expressed in Swiss Identification of a novel interleukin-15 (IL-15) transcript isoform gen- 3T3 fibroblasts. J. Biol. Chem. 277, 47167–47174. erated by alternative splicing in human small cell lung cell lines. Oncogene 12, 2187–2192. Hawkins, P.N., Lachmann, H.J., Aganna, E., and McDermott, M.F. (2004). Spectrum of clinical features in Muckle-Wells syndrome and Min, C.K., Maeda, Y., Lowler, K., Liu, C., Clouthier, S., Lofthus, D., response to anakinra. Arthritis Rheum. 50, 607–612. Weisiger, E., Ferrara, J.L., and Reddy, P. (2004). Paradoxical effects of interleukin-18 on the severity of acute graft-versus-host disease Hoshino, T., Winkler-Pickett, R.T., Mason, A.T., Ortaldo, J.R., and mediated by CD4ϩ and CD8ϩ T cell subsets after experimental Young, H.A. (1999). IL-13 production by NK cells: IL-13-producing allogeneic bone marrow transplantation. Blood 104, 3393–3399. NK and T cells are present in vivo in the absence of IFN-gamma. J. Nakanishi, K., Yoshimoto, T., Tsutsui, H., and Okamura, H. (2001). Immunol. 162, 51–59. Interleukin-18 is a unique cytokine that stimulates both Th1 and Th2 Kato, Z., Jee, J., Shikano, H., Mishima, M., Ohki, I., Ohnishi, H., Li, responses depending on its cytokine milieu. Cytokine Growth Factor A., Hashimoto, K., Matsukuma, E., Omoya, K., et al. (2003). The Rev. 12, 53–72. structure and binding mode of interleukin-18. Nat. Struct. Biol. 10, Nishimura, H., Washizu, J., Nakamura, N., Enomoto, A., and Yoshi- 966–971. kai, Y. (1998). Translational efficiency is up-regulated by alternative Kim, S.H., Eisenstein, M., Reznikov, L., Fantuzzi, G., Novick, D., exon in murine IL-15 mRNA. J. Immunol. 160, 936–942. Rubinstein, M., and Dinarello, C.A. (2000). Structural requirements Novick, D., Kim, S.H., Fantuzzi, G., Reznikov, L.L., Dinarello, C.A., of six naturally occurring isoforms of the IL-18 binding protein to and Rubinstein, M. (1999). Interleukin-18 binding protein: a novel inhibit IL-18. Proc. Natl. Acad. Sci. USA 97, 1190–1195. modulator of the Th1 cytokine response. Immunity 10, 127–136. Kim, S.H., Azam, T., Yoon, D.Y., Reznikov, L.L., Novick, D., Rubin- Ohtsuki, T., Micallef, M.J., Kohno, K., Tanimoto, T., Ikeda, M., and stein, M., and Dinarello, C.A. (2001a). Site-specific mutations in the Kurimoto, M. (1997). enhances expression mature form of human IL-18 with enhanced biological activity and and induces in Fas-expressing human myelomonocytic decreased neutralization by IL-18 binding protein. Proc. Natl. Acad. KG-1 cells. Anticancer Res. 17, 3253–3258. Sci. USA 98, 3304–3309. Okamura, H., Tsutsui, H., Komatsu, T., Yutsudo, M., Hakura, A., Kim, S.H., Reznikov, L.L., Stuyt, R.J., Selzman, C.H., Fantuzzi, G., Tanimoto, T., Torigoe, K., Okura, T., Nukada, Y., Hattori, K., et al. Hoshino, T., Young, H.A., and Dinarello, C.A. (2001b). Functional (1995). Cloning of a new cytokine that induces interferon-␥. Nature reconstitution and regulation of IL-18 activity by the IL-18R␤ chain. 378, 88–91. J. Immunol. 166, 148–154. Okamura, H., Tsutsui, H., Kashiwamura, S.-I., Yoshimoto, T., and Kim, S.H., Azam, T., Novick, D., Yoon, D.Y., Reznikov, L.L., Bufler, Nakanishi, K. (1998). IL-18: a novel cytokine that augments both P., Rubinstein, M., and Dinarello, C.A. (2002). Identification of amino innate and acquired immunity. Adv. Immunol. 70, 281–312. acid residues critical for biological activity in human interleukin-18. Panelli, M.C., Wang, E., Phan, G., Puhlmann, M., Miller, L., Ohn- J. Biol. Chem. 277, 10998–11003. macht, G.A., Klein, H.G., and Marincola, F.M. (2002). Gene-expres- Kojima, H., Takeuchi, M., Ohta, T., Nishida, Y., Arai, N., Ikeda, M., sion profiling of the response of peripheral blood mononuclear cells Ikegami, H., and Kurimoto, M. (1998). Interleukin-18 activates the and melanoma metastases to systemic IL-2 administration. Genome IRAK-TRAF6 pathway in mouse EL-4 cells. Biochem. Biophys. Res. Biol. 3, RESEARCH 0035. Commun. 244, 183–186. Plater-Zyberk, C., Joosten, L.A., Helsen, M.M., Sattonnet-Roche, P., Kuida, K., Lippke, J.A., Ku, G., Harding, M.W., Livingston, D.J., Su, Siegfried, C., Alouani, S., van De Loo, F.A., Graber, P., Aloni, S., M.S., and Flavell, R.A. (1995). Altered cytokine export and apoptosis Cirillo, R., et al. (2001). Therapeutic effect of neutralizing endogenous in mice deficient in interleukin-1 beta converting enzyme. Science IL-18 activity in the collagen-induced model of arthritis. J. Clin. 267, 2000–2003. Invest. 108, 1825–1832. Kumar, S., Hanning, C.R., Brigham-Burke, M.R., Rieman, D.J., Lehr, Pomerantz, B.J., Reznikov, L.L., Harken, A.H., and Dinarello, C.A. R., Khandekar, S., Kirkpatrick, R.B., Scott, G.F., Lee, J.C., Lynch, (2001). Inhibition of caspase 1 reduces human myocardial ischemic F.J., et al. (2002). Interleukin-1F7B (IL-1H4/IL-1F7) is processed by dysfunction via inhibition of IL-18 and IL-1beta. Proc. Natl. Acad. caspase-1 and mature IL-1F7B binds to the IL-18 receptor but does Sci. USA 98, 2871–2876. not induce IFN-gamma production. Cytokine 18, 61–71. Pufe, T., Petersen, W., Tillmann, B., and Mentlein, R. (2001). The Kusters, S., Tiegs, G., Alexopoulou, L., Pasparakis, M., Douni, E., splice variants VEGF121 and VEGF189 of the angiogenic peptide Kunstle, G., Bluethmann, H., Wendel, A., Pfizenmaier, K., Kollias, vascular endothelial growth factor are expressed in osteoarthritic G., and Grell, M. (1997). In vivo evidence for a functional role of both cartilage. Arthritis Rheum. 44, 1082–1088. tumor necrosis factor (TNF) receptors and transmembrane TNF in Reading, P.C., and Smith, G.L. (2003). Vaccinia virus interleukin-18- experimental hepatitis. Eur. J. Immunol. 27, 2870–2875. binding protein promotes virulence by reducing gamma interferon Lauwerys, B.R., Garot, N., Renauld, J.C., and Houssiau, F.A. (2000). production and natural killer and T-cell activity. J. Virol. 77, 9960– Cytokine production and killer activity of NK/T-NK cells derived 9968. with IL-2, IL-15, or the combination of IL-12 and IL-18. J. Immunol. Robinson, D., Shibuya, K., Mui, A., Zonin, F., Murphy, E., Sana, T., 165, 1847–1853. Hartley, S.B., Menon, S., Kastelein, R., Bazan, F., and O’Garra, A. Leonardi, C.L., Powers, J.L., Matheson, R.T., Goffe, B.S., Zitnik, R., (1997). IGIF does not drive Th1 development but synergizes with Wang, A., and Gottlieb, A.B. (2003). Etanercept as monotherapy in IL-12 for interferon-gamma production and activates IRAK and patients with psoriasis. N. Engl. J. Med. 349, 2014–2022. NFkappaB. Immunity 7, 571–581. Mallat, Z., Corbaz, A., Scoazec, A., Graber, P., Alouani, S., Esposito, Siegmund, B., Sennello, J.A., Lehr, H.A., Senaldi, G., Dinarello, C.A., B., Humbert, Y., Chvatchko, Y., and Tedgui, A. (2001). Interleukin-18/ and Fantuzzi, G. (2004). Frontline: interferon regulatory factor-1 as interleukin-18 binding protein signaling modulates atherosclerotic a protective gene in intestinal inflammation: role of TCR gamma lesion development and stability. Circ. Res. 89, E41–E45. delta T cells and interleukin-18-binding protein. Eur. J. Immunol. Martin, T.A., Mansel, R.E., and Jiang, W.G. (2002). Antagonistic ef- 34, 2356–2364. fect of NK4 on HGF/SF induced changes in the transendothelial Stevenson, F.T., Turck, J., Locksley, R.M., and Lovett, D.H. (1997). resistance (TER) and paracellular permeability of human vascular The N-terminal propiece of interleukin 1 alpha is a transforming endothelial cells. J. Cell. Physiol. 192, 268–275. nuclear oncoprotein. Proc. Natl. Acad. Sci. USA 94, 508–513. Immunity 142

Tagaya, Y., Kurys, G., Thies, T.A., Losi, J.M., Azimi, N., Hanover, J.A., Bamford, R.N., and Waldmann, T.A. (1997). Generation of secretable and nonsecretable isoforms through alternate usage of signal peptides. Proc. Natl. Acad. Sci. USA 94, 14444–14449. Tiegs, G., Hentschel, J., and Wendel, A. (1992). A T cell-dependent experimental liver injury in mice inducible by concanavalin A. J. Clin. Invest. 90, 196–203. van Dullemen, H.M., van Deventer, S.J., Hommes, D.W., Bijl, H.A., Jansen, J., Tytgat, G.N., and Woody, J. (1995). Treatment of Crohn’s disease with anti-tumor necrosis factor chimeric monoclonal anti- body (cA2). Gastroenterology 109, 129–135. Wang, H., Bloom, O., Zhang, M., Vishnubhakat, J.M., Ombrellino, M., Che, J., Frazier, A., Yang, H., Ivanova, S., Borovikova, L., et al. (1999). HMG-1 as a late mediator of endotoxin lethality in mice. Science 285, 248–251. Werman, A., Werman-Venkert, R., White, R., Lee, J.K., Werman, B., Krelin, Y., Voronov, E., Dinarello, C.A., and Apte, R.N. (2004). The precursor form of IL-1alpha is an intracrine proinflammatory activa- tor of transcription. Proc. Natl. Acad. Sci. USA 101, 2434–2439. Xiang, Y., and Moss, B. (1999). Identification of human and mouse homologs of the MC51L-53L-54L family of secreted glycoproteins encoded by the Molluscum contagiosum poxvirus. Virology 257, 297–302. Yamamoto, Y., Kato, Z., Matsukuma, E., Li, A., Omoya, K., Hashi- moto, K., Ohnishi, H., and Kondo, N. (2004). Generation of highly stable IL-18 based on a ligand-receptor complex structure. Bio- chem. Biophys. Res. Commun. 317, 181–186.

Accession Numbers

Sequences are available at GenBank under accession numbers AY495331 (IL-32␣), AY495332 (IL-32␤), MN_00421 (IL-32␥), AY495333 (IL-32␦), and AY495334 (5 kb genomic sequence).