A Novel Cytokine Receptor-Ligand Pair IDENTIFICATION, MOLECULAR CHARACTERIZATION, AND IN VIVO IMMUNOMODULATORY ACTIVITY*

As part of a large scale effort to discover novel secreted proteins, a cDNA encoding a novel cytokine was identified. Alignments of the sequence of the new protein, designated IL-17B, suggest it to be a homolog of the recently described T cell-derived cytokine, IL-17. By Northern analysis, EST distribution and real-time quan-titative polymerase chain reaction analysis, mRNA was detected in many cell types. A novel type I transmembrane protein, identified in an EST data base by homology to IL-17R, was found to bind specifically IL-17B, as determined by surface plasmon resonance analysis, flow cytometry, and co-immunoprecipitation experiments. Readily detectable transcription of IL-17BR was restricted to human kidney, pancreas, liver, brain, and intestines and only a few of the many cell lines tested. By using a rodent of IL-17BR as a probe, IL-17BR message was found to be drastically up-regulated during intestinal inflammation

Cytokines are secreted regulatory peptides that mediate a wide range of biological activities by binding to specific cell surface receptors on target cells. Cytokine actions include control of cell proliferation and differentiation, regulation of hemopoiesis, and immune and inflammatory responses (1). Cytokines are also major orchestrators of host defense processes and, as such, are involved in responses to exogenous as well as endogenous insults and in repair or restoration of tissue integrity.
Except for the presence of an N-terminal signal peptide usually required for secretion, the cytokines known thus far are members of many distinct and structurally unrelated families of molecules. In an attempt to identify new cytokine-like peptides, we have searched for novel open reading frames containing signal sequences in a large sequence data base of expressed human genes. One group of candidate clones was found to encode a novel homolog of interleukin-17. 1 IL-17 2 is a cytokine-inducing glycoprotein of 155 amino acids, produced predominantly by activated CD4 ϩ T cells and double negative (CD4 Ϫ CD8 Ϫ ) T cells exhibiting indirect proinflammatory and hematopoietic properties (2)(3)(4)(5)(6). In vivo, its expression has been reported elevated in the rheumatoid synovium, in multiple sclerosis blood and cerebrospinal fluid, and in peripheral blood mononuclear cells following ischemic stroke (7)(8)(9)(10)(11). Is is also produced by tumorinfiltrating lymphocytes and increases tumorigenicity of human cervical tumors in nude mice (12,13).
Here we report on the discovery of a novel IL-17 homolog (IL-17B), together with the identification of a novel cell surface receptor that specifically binds to it. We detail the molecular cloning, tissue distribution, and expression of both IL-17B and IL-17BR, describe the in vivo activity of IL-17B, and demonstrate binding to IL-17BR. Our results illustrate the utility of searching large sequence data bases for the rapid matching of ligand-receptor pairs. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

EXPERIMENTAL PROCEDURES
Sequence and Expression Analysis-Full-length cDNAs for human IL-17B and IL-17BR were identified, sequenced, and submitted to Gen-Bank TM and were assigned the accession numbers AF212311 and AF212365, respectively.
DNA sequencing was performed using ABI 377 automated DNA sequencers and Perkin-Elmer Biosystems Big Dye Terminator sequencing chemistries.
Northern blot analysis of poly(A) RNA samples was performed using CLONTECH (Palo Alto, CA) multiple tissue Northern blots.
For analysis of murine IL-17B transcripts, total RNA was prepared from rodent organs, separated on agarose gels containing formamide, and blotted onto Nylon filters.
Membranes were hybridized overnight in Hybrisol solution (Oncor), preheated to 42°C before use, followed by two subsequent washes in 2ϫ SSC, 0.1% SDS, and 0.2ϫ SSC, 0.1% SDS at the same temperature.
Double-stranded cDNA probes, used at a minimum specific activity of 2 ϫ 10 9 cpm/g, were generated by restriction digestion, 32 P-labeled using the Rediprime random primer labeling system (Amersham Pharmacia Biotech), and purified with NucTrap ion exchange push columns (Stratagene, La Jolla, CA).
Antibody Preparation-For bacterial production of IL-17B, an open reading frame coding for the mature form of IL-17B (residues Gln 21 -Phe 180 ), as predicted by SignalP (34), was amplified by PCR and cloned as an NdeI-Asp718I restriction fragment (495-bp product) downstream of an inducible lacZ promoter. For efficient translation, the first 50 nucleotides of mature IL-17B were codon-optimized for expression in Escherichia coli. The primers used were sense, 5Ј-GACTCATATG-CAGCCGCGTTCCCCGAAATCCAAGCGTAAA-3, and antisense, 5Ј-GACTGGTACCTTATCAGAAGATGCAGGTGCAGC-3Ј. The reading frame and adjacent areas were sequence-confirmed following cloning. After transformation and expression in E. coli, IL-17B was present in the inclusion bodies. Inclusion bodies were solubilized with 4 M guanidine HCl and dialyzed against 50 mM sodium acetate buffer, pH 5, containing 0.1 M NaCl. Antisera were prepared by immunizing rabbits with IL-17B (Q21-F180). The sera were used for immunoblot analysis after 1000-fold dilution.
Transient Transfections-Plasmid DNA was transfected into 293T cells using LipofectAMINE reagent (Life Technologies, Inc.) according to the manufacturer's instructions.
Generation of Stably Transfected CHO Clones-The complete open reading frame of human IL-17B was amplified by PCR (601-bp product). The primers used were sense, 5Ј-GACTGGATCCGCCATCATGGACT-GGCCTCACAACC-3, and antisense, 5Ј-GACTGGTACCGGATG-GTCTCGGGCTGCTG-3Ј. Full-length IL-17B was cloned as a BamHI-Asp718I restriction fragment into a cytomegalovirus-enhancer/Rous sarcoma virus-long terminal repeat promoter-based expression vector. The clones were sequence-confirmed before transfection into CHO cells. IL-17B-positive CHO clones were selected by reverse transcriptase-PCR and amplified to 1 micromolar methotrexate. Conditioned media (CHO-5 serum-free media without methotrexate) from 7 CHO clones were analyzed for IL-17B expression by SDS-PAGE followed by silver staining. Three CHO clones with the highest expression were selected for continued amplification in the presence of 10 M methotrexate.
Purification of IL-17B-Four-day conditioned media from IL-17Bexpressing clones was used for protein purification. The media were adjusted to 25 mM HEPES buffer, pH 7.2, and applied to the strong cation exchange resin (Poros HS-50) using a BioCad 60 (PE, Perkin-Elmer/Perseptive Biosystems). The HS-50 bound material was eluted using a step gradient of NaCl in 25 mM HEPES buffer, pH 7.2, and fractions were analyzed by SDS-PAGE. The 0.8 M NaCl pool was applied to weak anion exchanger (CM HyperD, BioSepra) and eluted with a NaCl gradient. The IL-17B-positive fractions were pooled, subjected to size-exclusion chromatography on a Superdex-75 column (Amersham Pharmacia Biotech) equilibrated with phosphate-buffered saline, and pooled again. Protein concentration was determined using the BCA procedure (Pierce). Endotoxin was measured using the LAL assay (Cape Cod Assoc.). Measured endotoxin levels were between 11 and 20 EU/mg protein.
Purification of Epitope-tagged IL-17B-For synthesis of N-terminal Flag fusion protein, the mature portion (Arg 23 -Phe 184 ) coding region of IL-17B was amplified by PCR and cloned into pFLAG-CMV-1 vector (Sigma) as an EcoRI-BamHI restriction fragment. The primers used were 5Ј-GCCCCGGAATTCAAGGAGCCCCAAAAGCAAGAGG-3Ј (sense) and 5Ј-GCCCGC GGATCC TCAGAAGATGCAGGTGCAGCC-3Ј (antisense, 550-bp product). Conditioned media from 293T cells transiently transfected with pFLAG-CMV-1:IL-17B was prepared and purified using anti-FLAG affinity chromatography according to the manufacturer's instructions. Approximately 300 g of purified protein was recovered from 500 ml of culture supernatant.
BIAcore Binding Analyses-High density BIAcore flow cells were prepared for initial ligand screening by covalent immobilization to the sensor chip (CM5 chip) via amine groups using N-ethyl-NЈ-(dimethylaminopropyl) carbodiimide/N-hydroxysuccinimide chemistry. huIL-17R and huIL-17BR proteins were dialyzed against 10 mM sodium acetate buffer, pH 5, and coupled at densities of 7900 and 9600 relative units for IL-17R and IL-17BR, respectively. Various concentrations of purified IL-17B and IL-17 (R & D Systems, Minneapolis, MN) in 50 ml of HEPES-buffered saline (HBS) buffer were examined for receptor binding at a flow rate of 15 l/min. After injection of sample the flow cell was equilibrated with HBS. A low density BIAcore flow cell (ϳ300 relative units) derivatized with IL-17B receptor was used for kinetic analysis. Various concentrations of IL-17B were flown over the huIL-17BR flow cell and a control flow cell immobilized with irrelevant Fc protein, TR2-Fc. Injection of IL-17B in HBS at different flow rates (15-75 l/ml) and at different concentrations (0.04 -5 g/ml) were used to determine the kinetic binding constants. IL-17B binding was followed during the binding (k a , on-rate) and dissociation (k d , off-rate) phases. After analysis the chip was regenerated by dissociation of IL-17B⅐receptor complexes by washing with 20 ml of 10 mM glycine HCl, pH 2.3. All methods were run on a BIAcore 3000 using the BIAcore control software version 1.3 (BIAcore AB). The on and off rates were determined using the kinetic evaluation program in BIAevaluation 3 software using a 1:1 binding model and the global analysis method. The 2 value was 2.54.
Flow Cytometric Evaluation of IL-17BR Transfectants-For detection of IL-17B, cells (10 6 in 100 l) were incubated with either preimmunized rabbit serum (1:100) or IL-17B immunized rabbit serum (1: 100). Cells were washed and then incubated with phycoerythrinconjugated goat anti-rabbit F(ab) 2 . Cell were washed, resuspended in 5 g/ml propidium iodide solution, and acquired on the FACScan (Becton Dickinson Immunocytometry Systems, San Jose, CA). Alternatively, cells were first incubated 10 min at room temperature with 1 g of soluble IL-17B, and then the anti-IL-17B serum was added as described. Ability of soluble IL-17BR-Fc to block IL-17B binding to IL-17BRϩ cells was also tested, using 1 or 10 g of IL-17BR-Fc added in solution with the soluble IL-17B. Analysis was performed using an electronic gate on propidium iodide-negative live cells.
Peritoneal Exudate Cells-BALB/c mice (n ϭ 8 per group) were injected intraperitoneally with 0.2 ml of rhIL-17B at indicated amounts plus 50 g of the human chemokine HCC-1 (35) as a carrier. At 4 h after injection the mice were sacrificed by CO 2 asphyxiation. The peritoneal cavity was then exposed, and the exudate was collected by washing the cavity with 4 ml of phosphate-buffered saline. Cell counts performed in triplicate on each peritoneal exudate sample were quantitated by complete blood chemistry analyzer and hemocytometer. Cytocentrifuge (Shandon, Inc., Pittsburgh, PA) smears of peritoneal exudate cells from each mouse were stained with Wright's stain for differential counts. Total numbers of PMN accumulating in the peritoneal cavity were calculated by multiplying total peritoneal exudate cells by the percent PMN determined from differential counts. Both % and total values of PMN were expressed as means Ϯ S.E. Significance of difference was determined by an analysis of variance t test.
Reagents-Recombinant human IL-17 was from R & D Systems (Minneapolis, MN). Dextran sulfate sodium (DSS, M r 36,000 -44,000) was purchased from American International Chemistry (Natick, MA). LipofectAMINE reagent and geneticin (G418) were from Life Technologies, Inc. Indomethacin and methotrexate were obtained from Sigma.
Animals-Female Swiss Webster mice (20 -25 g) and female Lewis rats (160 -180 g) were obtained from Charles River Laboratories (Raleigh, NC) and kept under standard conditions for 1 week before being used in experiments. The animal protocols used in this study were reviewed and approved by the Human Genome Sciences, Inc., Institutional Animal Care and Use Committee.
Tissue Collection-Tissues from several models of inflammation were tested for expression of the IL-17B receptor. Models included murine colitis, rat jejunitis, mouse graft versus host disease, and Listeria-induced bacteremia in mice.
In the DSS-induced murine colitis model, female Swiss Webster mice were given a 4% solution of DSS ad libitum for 7 days. Animals were euthanized on day 7, and the distal third of the colon flushed with saline and snap-frozen with liquid nitrogen in preparation for RNA extraction.
In indomethacin-induced rat jejunitis, female Lewis rats were injected subcutaneously on day 0 and 1 with indomethacin. Indomethacin was prepared by solubilizing in absolute ethanol, sonicating for 30 s, and then diluting 1:4 v/v with 5% sodium bicarbonate to create a stock solution of 10 mg/ml. The stock solution was diluted further with 5% sodium bicarbonate, and rats were injected subcutaneously (sc) with a final dose of 8 mg/kg in a volume of 0.2 ml. On day 4, 3 days after the final indomethacin injection, rats were euthanized, and 10 cm of the small intestine was removed, starting 20 cm up from the cecum. The intestinal tissue was flushed with saline and snap-frozen with liquid nitrogen prior to RNA analysis.

RESULTS
Primary Structure of IL-17B-An EST coding for a putative signal peptide was initially discovered in a human thymus cDNA library. Three additional clones were subsequently identified in libraries from thymus tumor and from 9-and 12-weekold human embryo tissue. All four clones were fully sequenced and found to be identical over the entire open reading frame. They are predicted to code for a protein of 184 amino acids with an N-terminal leader sequence of 20 amino acids. The predicted molecular mass for this protein is 20.4 kDa, with an estimated isoelectric point of 9.24 (Fig. 1A). There is one potential Nlinked glycosylation site and eight cysteine residues. The short 3Ј-untranslated region contains a single near-consensus polyadenylation site and is devoid of the characteristic AU repeats found in several other cytokines, growth factors, and protooncogenes (30).
A comparison of both nucleotide and amino acid sequences with the GenBank TM or EMBL data bases revealed significant homology of the translation product with the amino acid sequence of the recently described T cell-derived cytokine, IL-17 ( Fig. 1B). At the amino acid level, human IL-17B shared 21.3, 19, and 20.7% identity with human, mouse, and rat IL-17, respectively, and 21.9% identity with the product of the 13th ORF of Herpesvirus saimiri (HVS13). The degree of conserva- tion is higher in the C-terminal portion of the protein, and six of the eight cysteines present in IL-17B are conserved and identically spaced between IL-17B and IL-17 (2,3).
A putative murine ortholog of IL-17B was identified in a mouse EST data base and found to be 87.8% similar to the human IL-17B and 21.3, 19.6, 22, and 21.9% similar to the human, mouse, rat, and viral IL-17 sequences (Fig. 1B).
The map position of the human IL-17B gene was determined by somatic cell hybrid and radiation hybrid mapping. Amplification of the standard G3 radiation hybrid panel using genespecific oligonucleotide primers showed linkage to the SHGC-33930, SHGC-4655, and SHGC-11215 markers on chromosome 6 at distances of 13, 14, and 18 centirads, with LOD scores of 10.13, 9.25, and 8.94, respectively, corresponding to a cytogenetic location at 6p21.2.
Cellular and Tissue Distribution of the hIL-17B mRNA-By Northern blot analysis of human tissues, a very strong signal at ϳ1.0 kb was seen in spinal cord, testis, and small intestines, and less pronounced in prostate, colon mucosal lining, ovary, and in the K-562 chronic myelogenous leukemia cell line (Fig.  2). Furthermore, a weak transcript of similar length was routinely observed in trachea, uterus, adrenal gland, substantia nigra, and fetal kidney. Even though IL-17B cDNA was initially isolated from thymus, the signal observed on all blots with spleen or thymus poly(A) RNA was either faint or not visible. The tissue distribution of murine IL-17B was also determined. An ϳ1.0-kb band was observed on poly(A) mouse RNA blots probed with a murine reading frame-specific cDNA probe (Fig. 1B). The signal was strongest in brain, heart, and testis and weaker in lung, liver, and skeletal muscle (mouse spinal cord was not tested).
Molecular Characterization of an IL-17B Receptor-In order to identify target cell types that respond to IL-17B, we searched for candidate receptors. Since IL-17B is distantly related to IL-17, we screened the EST data bases for novel homologs of the recently described murine and human IL-17R amino acid sequences (31,32). A cDNA clone containing an open reading frame predicted to code for a type I transmembrane protein was identified in a library from human adult lung tissue. Overlapping clones were subsequently discovered in libraries from various other tissues, predominantly eosinophils, brain, pancreas, kidney, thyroid, and osteoclastomas.
A large open reading frame is predicted to encode a receptor of 426 amino acids. Computer-assisted analysis suggests that this protein has an N-terminal signal peptide with a cleavage site after proline 17. The signal peptide is followed by a 273amino acid extracellular domain (Arg 18 -Gly 289 ), a 22-amino acid transmembrane stretch (Trp 290 -Leu 311 ), and a 115-amino acid cytoplasmic tail (Met 312 -Leu 426 , Fig. 3A). There are six potential N-linked glycosylation sites in the extracellular domain, at positions 67, 103, 156, 183, 197, and 283. The predicted molecular mass for this protein is 47.9 kDa, with an isoelectric point of 8.16. Overall, the IL-17BR protein sequence is 19.2 and 18.2% identical to the human and murine IL-17R sequences, respectively (Fig. 3B). There is no WSXWS motif in the extracellular domain (33). The cytoplasmic portion of this new receptor is much shorter than the unusually long tail described for IL-17R (32). Furthermore, a segment (TPPPLR-PRKVW) located proximal to the IL-17R transmembrane domain, which is highly conserved among cytokine receptors (33), is absent.
By Northern blot analysis of human tissues using an open reading frame-specific hybridization probe, two specific transcripts of ϳ3.5 and 1.4 kb can be detected in several endocrine tissues, most pronounced in fetal and adult liver, kidney, pancreas, testis, colon, and small intestines but are absent in peripheral blood leukocytes and lymphoid organs (Fig. 4). Only a few of a large series of transformed human cell lines grown in culture expressed IL-17BR transcript detectable by Northern and real-time PCR analysis (not shown). These were predominantly derived from organs found to be positive for IL-17BR message above and included the WRL-68 human embryonic liver, Colo587 pancreas adenocarcinoma-mesothelioma, PANC-1 pancreatic epithelioid carcinoma, HeLa S3 cervical carcinoma, K562 leukemia, Raji Burkitts lymphoma and SW480, colorectal adenocarcinoma lines.
The map location of IL-17BR was determined at 3p21.1 by radiation hybrid mapping, with an LOD score of 12. It is noted that several chemokine receptors and trypsin inhibitors have been mapped in the 3p21.1, 3p21.2, and 3p21 regions.
Tissue Distribution of Rodent IL-17BR-The constitutive mRNA and protein levels of many cytokine receptors are extremely low, and detection of their transcription and surface expression to levels that can be detected by Northern blot or cell sorting is very often dependent on specific activation mechanisms. To gain insight into possible roles of this novel cytokine-receptor pair in disease processes, a partial cDNA clone for the putative murine IL-17BR ortholog was identified and hybridized with total RNA prepared from a series of rodent disease model organs. Because of the proinflammatory roles of IL-17, RNAs from several inflammatory models were used. These included kidney and liver RNAs from a murine model of graft versus host disease, liver following Listeria infection, as well as colon and intestinal tissues from DSS-induced colitis and from indomethacin-induced intestinal inflammation in rats. Among the models tested, IL-17BR message was found to be significantly up-regulated only in the intestines after indomethacin treatment (Fig. 5). However, the up-regulation was drastic, from weak or undetectable in most untreated samples to a readily detectable or intense signal in total RNAs from several differently treated animals. As seen above with the human probe and human tissues (Fig. 4) clones expressed a novel protein of ϳ20 kDa not present in control media transfected with expression vector plasmid only (not shown). One clone was chosen for scale-up, and conditioned media were obtained after 4 days. Immunoblot analysis of conditioned media using a polyclonal antibody revealed the presence of several species of IL-17B, which suggested the presence of proteolytic processing and/or differential glycosylation of the protein. IL-17B was purified to apparent homogeneity as described under "Experimental Procedures." The cation exchange-bound IL-17B was then subjected to sizeexclusion chromatography and N-terminal sequencing. SDS-PAGE of purified IL-17B revealed the presence of three major species of 23, 22, and 18 kDa under reducing conditions (Fig. 6).
PAGE analysis of purified IL-17B under non-reducing conditions (data not shown) showed that, unlike IL-17, IL-17B migrates as a monomer and thus is not a disulfide-linked dimer under these conditions. However, when eluted from a Superdex-75 size-exclusion column, IL-17B behaves as a dimer. Thus, native IL-17B appears to be a non-disulfide-linked dimer.
The major bands were subjected to N-terminal sequence analysis. The 23/22-kDa species had four closely spaced N termini starting at Arg 23 , Ser 27 , Arg 29 , and Lys 30 (in roughly equal proportion), whereas the 18-kDa band had two N termini starting at residues Leu 49 and Ser 51 . The presence of truncated forms of the protein is suggestive of post-translational proteolytic processing. This appears to be due to the action of a proprotein convertase-like activity as three of the N-terminal residues, Arg 29 , Ser 30 , and Met 52 , are preceded by basic residues. However, Ser 51 is preceded by Val and may not be processed by the same enzyme. When expressed in baculovirus, only one species was detected. The N terminus of baculovirusexpressed IL-17B was Arg 23 , which is two residues downstream of the cleavage site predicted by SignalP (34), Gln 21 . Thus, IL-17B isolated from CHO-conditioned media appears to occur in several forms due to post-translational proteolysis. The effects of processing on biological activity are not yet known.
Binding Experiments-Specific interaction of the extracellular domain of the novel receptor with soluble IL-17B purified as described above was demonstrated independently by three different methods. The predicted extracellular domains of human IL-17 receptor and of IL-17B receptor were expressed as chi-meric proteins, fused to the heavy chain constant region of IgG1. When used as immobilized component in the BIAcore surface plasmon resonance analysis system, purified soluble IL-17B bound to IL-17BR, in a concentration-dependent manner (Fig. 7A). Very poor interaction of this receptor was observed with soluble recombinant human IL-17. In contrast, IL-17 bound well to IL-17 receptor under the same experimental conditions (data not shown).
The kinetics of binding of IL-17B to IL-17B receptor was examined. The calculated on-rate, k a , was 8.35e5 1/ms and the off-rate, k d , was 6.34e-3 1/s for a K d of 7.6e-9 M. Thus, IL-17B binds to IL-17B receptor with a relatively high affinity of 7.6 nM.
293T cells transiently transfected with human IL-17BR expression plasmid were used to measure cell surface binding of IL-17B by flow cytometry (Fig. 7B) as detected by an IL-17B antibody. Significant binding of IL-17B was observed in the FIG. 5. IL-17BR message up-regulation in models of inflammatory bowel disease. Tissue was collected from mouse colitis or rat jejunitis induced by treatment as described under "Experimental Procedures," along with normal control colons and intestines. Total RNA was isolated, blotted onto nylon filter, and hybridized with a murine IL-17BR cDNA-specific probe 88% similar and 83% identical to huIL-17BR at the amino acid level over the area corresponding to Tyr 123 -Phe 278 . IL-17BR transfectants but was undetectable in untransfected cells. IL-17B antibody alone did not bind to untransfected or transfected cells. Furthermore, cell surface binding was inhibited by the addition of soluble IL-17BR-Fc fusion protein. This inhibition of binding was dose-dependent, as the mean fluorescence peak was shifted back by 15 and 90% by the addition of 1 and 10 g of receptor protein, respectively.
Specific cell surface binding of epitope-tagged IL-17B was also demonstrated. The SW480 colorectal adenocarcinoma cell line, shown above to express IL-17BR transcript, was used in this experiment. Binding of N-terminal Flag-IL-17B fusion protein to these untransfected cells was detectable as a quantitative shift after staining with Flag-or IL-17B-specific antibody (not shown), in contrast to the only partial shift observed with the transfected cell population above (Fig. 7B).
Finally, binding of huIL-17B to huIL-17BR was confirmed by co-immunoprecipitation. Purified IL-17BR-Fc fusion protein was incubated with soluble CHO-derived recombinant human IL-17B (Fig. 7C). Binding of IL-17B to IL-17BR was demonstrated by detection of IL-17B in the protein A-agarose coprecipitate by Western immunostaining.
Neutrophil Migration Elicited by IL-17B in Vivo--Treatment with IL-17 has been shown to activate the transcription factor NF-B and to induce cytokine secretion in fibroblasts (2,3). In our hands, treatment with CHO-expressed and purified IL-17B did not activate NF-B in NIH3T3 cells. Furthermore, no reproducible induction of cytokine message or protein (IL-6, IL-8, TNF-␣, IFN␥, IL-3, granulocyte-colony-stimulating factor) was observed in HeLa, CHO, or 293T cells after treatment with rhIL-17B (not shown) in vitro.
To examine its possible physiological roles in vivo, recombinant human IL-17B was injected into BALB/c mice. As the abundance of IL-17B transcripts in RNA from colon mucosal lining and small intestines may suggest functions of the cytokine on the gastrointestinal tract walls, intraperitoneal (intraperitoneal) injection was chosen as the route of administration. The results shown in Fig. 8A demonstrate that intraperitoneal injection of rhIL-17B consistently caused a dose-dependent influx of PMN into the peritoneal cavity within 4 h. This influx of PMN was not a result of nonspecific vascular leakage because very few red blood cells were observed in most cytopreparations (Fig. 8B). Red blood cells or clotting visible in some animals was attributed to vasculoepithelial injury during injection, and these preparations were excluded from analysis. Another cytokine, rhHCC-1 (35), showed no effect on PMN infiltration, even over a wide range of protein concentrations, and therefore was chosen to serve as protein carrier for the low dose study of IL-17B. Peritoneal PMN infiltration was still marked in response to 100 ng of IL-17B per mouse but became statistically insignificant at 10 ng (Fig. 8A). The results are not attributable to lipopolysaccharide contaminants since (a) the amount of lipopolysaccharide in rhIL-17B (no more than 20.2 EU/mg, i.e. 0.2 EU/g/mouse), is 10-fold lower than that of rhHCC-1 (53 EU/mg, i.e. 2.65 EU/50 g/mouse), and (b) heating of rhIL-17B at 80°C for 45 min completely abrogated its ability to cause PMN influx. DISCUSSION Here we describe the isolation and molecular characterization of a novel mammalian cytokine, denoted IL-17B. Several observations in this report suggest the physiological role of IL-17B to be distinct from IL-17 or other previously characterized secreted factors. First, whereas IL-17 is found to be expressed almost exclusively by CD4 ϩ and CD4 Ϫ CD8 Ϫ double negative activated T cells (2), IL-17B is highly transcribed in human and murine spinal cord, and low levels of expression can be found in many other organs. Second, the AU-rich repeats indicative of transient expression found in IL-17 and other cytokines are absent from the 3Ј-untranslated region of the IL-17B transcript (30). IL-17B may thus be the translation product of a more stable message that in fact could give rise to a constitutive serum presence of the protein. Third, whereas a specific cell surface receptor for IL-17 is described to be expressed in virtually all cell types (31), the receptor for IL-17B discovered here shows a highly specific message expression pattern, largely restricted to kidney, liver, pancreas, and intestines. Furthermore, the drastically shorter cytoplasmic tail of IL-17BR as compared with human and mouse IL-17R indicates that there may be principal differences in the corresponding downstream signaling processes.
Recombinant human IL-17B protein did not exert any detectable chemotactic activity upon peripheral blood neutrophils or eosinophils from several human donors. Moreover, IL-17BR message was undetectable in human neutrophils by either Northern or real-time PCR analysis (not shown), and neutrophils failed to bind epitope-tagged recombinant IL-17B by fluorescence-activated cell sorter analysis. Therefore, the dose-dependent neutrophil influx observed after intraperitoneal injection is unlikely due to a direct activity on neutrophils. Rather, IL-17B binding to cell surface receptors on stromal or other connective tissue elements may trigger expression and secretion of chemoattractive factors from these cell types, leading to a guided local accumulation of polymorphonuclear leukocyte populations. Accordingly, our inability to observe transcription factor activation or mRNA and protein expression of several known chemokines in transformed cell lines in culture is most likely due to a requirement for a specific activation process to render cells responsive to IL-17B. In addition, IL-17B could allow or enhance migration of polymorphonuclear cells into the gastrointestinal tract, or other epithelial structures by acting not on these invading cells directly but via some effects on the local microvasculature of these tissues. However, even though recombinant expression of IL-17BR cDNA alone is sufficient to yield specific cell surface binding sites, it cannot be ruled out that IL-17B acts on these cells by additional receptors or receptor components.
Several proinflammatory cytokines are known to elicit neutrophil accumulation in the peritoneal cavity. For example, among a number of cytokines tested by Sayers et al. (40), only IL-1 (rhIL-1␣ and -1␤) and TNF (rhTNF-␣ and -␤) induced significant PMN infiltration within 2 h after injection. In the same study, neither the colony-stimulating factors (granulocyte-colony-stimulating factor and granulocyte macrophagecolony-stimulating factor) nor IL-2 or IFN (IFN-␣, -␤ and -␥) consistently induced PMN influx, even when tested over a wide range of protein doses (0.05-50 ng/mouse). When compared on a protein basis, rhIL-1␣ was significantly more potent than any other cytokine (rhIL-1␤, rhTNF-␣, and rhTNF-␤), with maximal PMN accumulation occurring at about 0.5 ng/mouse. How- FIG. 8. Neutrophil attraction following rhIL-17B injection. A, dose response of PMN infiltration. Peritoneal cells obtained 4 h after injection of various concentrations of rhIL-17B in 50 g of HCC-1 carrier. Total and percentage of PMN were calculated for each mouse. * or **** denotes significant difference from control group with p Ͻ 0.05 or p Ͻ 0.0001, respectively. B, effect of rhIL-17B on PMN infiltration into the peritoneal cavity. Cytospin preparations of cells obtained from the peritoneal cavity of mice 4 h after injection of phosphate-buffered saline (A), rhIL-17B (10 g/mouse) (B), rhIL-17B (10 g/mouse after heating at 80°C for 45 min (C), and HCC-1 carrier alone (D). ever, the dose-response curve for rhIL-1␣ is bell-shaped such that high doses of rhIL-1␣ (Ͼ5 ng/mouse) do not induce significant PMN infiltration. In our hands, IL-17B could induce a maximal PMN infiltration very similar to that elicited by rhIL-1 and rhTNF, but at much higher protein dose (1000 -10000-fold of IL-1 and TNF). When compared with its family member IL-17, IL-17B is about 10 time less potent in inducing PMN infiltration (Ref. 16 and data not shown).
Because of the similarity of IL-17B to the pro-inflammatory cytokine, IL-17, and its association with neutrophil chemotaxis, IL-17B receptor (IL-17BR) message distribution studies were conducted in target tissue from various models of inflammation. Among those models were DSS-induced colitis and indomethacin-induced jejunitis, both models of inflammatory bowel disease. Although Northern blot analysis showed no IL-17BR in the colons of DSS-treated mice, IL-17BR was dramatically up-regulated on day 4 in the mid bowel of rats receiving consecutive indomethacin injections on day 0 and 1. Indomethacin-induced jejunitis is characterized by transmural lesions and an influx of neutrophils. Although there is little evidence for an immunologically driven mechanism of action, indomethacin-induced inflammatory bowel disease bears some resemblance to Crohn's disease, its clinical counterpart, as follows: (a) it induces transmural lesions, (b) causes non-bloody diarrhea, (c) has a genetic component, (d) is dependent on the presence of bacteria, (e) causes granuloma formation, and (f) is accompanied by inflammation (36,37).
In summary, we have described a novel cytokine ligandreceptor system that may be involved in specific local inflammatory processes and in the indirect recruitment of neutrophils to tissue repair and immune reactions at specific target organs. Detailed studies of the mechanisms of receptor expression and activation, made possible using the molecular probes first described here, should lead to an understanding of the physiological sites of action for these new molecules.