Expression of alternatively spliced interleukin-1 receptor accessory protein mRNAs is differentially regulated during inflammation and apoptosis

Cellular Signalling 15 (2003) 793 – 802 www.elsevier.com/locate/cellsig

Expression of alternatively spliced interleukin-1 receptor accessory protein mRNAs is differentially regulated during inflammation and apoptosis Liselotte E. Jensen *, Alexander S. Whitehead Department of Pharmacology and Center for Pharmacogenetics, University of Pennsylvania, 156 Johnson Pavilion, 3620 Hamilton Walk, Philadelphia, PA 19104, USA Received 12 December 2002; accepted 21 January 2003

Abstract Two alternative splice variants of the interleukin-1 receptor accessory protein (IL-1RAcP) mRNA are known. Membrane-bound IL1RAcP (mIL-1RAcP) promotes intracellular interleukin-1 (IL-1) signalling whereas soluble IL-1RAcP (sIL-1RAcP) is probably an inhibitor of IL-1 signalling. Here we establish that sIL-1RAcP mRNA levels increase 16-fold in response to phorbol esters in the human hepatoma cell line HepG2 via a mechanism that depends on de novo protein synthesis. Following exposure of cells to UV light, a potent inducer of apoptosis, mIL-1RAcP mRNA is rapidly down-regulated and a new steady-state level established briefly before a gradual return to pretreatment levels. Following treatment with staurosporine, also an inducer of apoptosis, mIL-1RAcP mRNA levels steadily decrease through 72 h, with little change in sIL-1RAcP mRNA levels. A novel alternative splice variant, sIL-1RAcP-h, was identified. Its sequence indicates that sIL-1RAcP-h is secreted and has a unique second half of the third immunoglobulin (Ig) domain. The dramatic changes in levels of IL-1RAcP mRNAs suggest important functions in regulating sensitivity to IL-1 during stress and may play a role in oncogenic processes that are engaged during chronic inflammation. D 2003 Elsevier Science Inc. All rights reserved. Keywords: Apoptosis; Inflammation; Interleukin-1; UV light; Staurosporine; Cytokine receptors; Alternative splicing; Molecular biology

1. Introduction Interleukin-1 (IL-1) is a potent cytokine that initiates many of the immunological responses to infection, tissue damage and stress (reviewed in Ref. [1]). These responses include, but are not limited to, fever, increased vascular permeability, hypotension, increased levels of circulating nitric oxide, infiltration of inflammatory cells, stimulation of the hypothalamic – pituitary – adrenal axis and corticosterone synthesis and a radical alteration in hepatic acute phase protein synthesis. Intracellular responses to IL-1 are initiated via the transmembrane IL-1 receptor type I (IL-1RI). IL-1RI has 319

Abbreviations: IAPs, inhibitors of apoptosis; IL-1, Interleukin-1; IL1RI, interleukin-1 receptor type I; IL-1RAcP, interleukin-1 receptor accessory protein; mIL-1RAcP, membrane-bound IL-1RAcP; PDD, phorbol 12,13-didecanoate; QRT-PCR, quantitative RT-PCR; sIL-1RAcP, soluble IL-1RAcP; UTR, untranslated region. * Corresponding author. Department of Pharmacology, University of Pennsylvania, 156 Johnson Pavilion, 3620 Hamilton Walk, Philadelphia, PA 19104-6084, USA. Fax: +1-215-573-9135. E-mail address: [email protected] (L.E. Jensen).

extracellular amino acid residues comprising three immunoglobulin (Ig) domains, a transmembrane region of 20 residues and 213 intracellular residues that are involved in signal transduction [2,3]. IL-1RI associates with the interleukin-1 receptor accessory protein (IL-1RAcP), which is structurally very similar to IL-1RI with which it shares approximately 25% sequence identity [4]. Its extracellular region of 340 amino acid residues also comprises three Ig domains and has transmembrane and intracellular regions that are 29 and 181 residues, respectively. Crystallographic analyses have established that the three Ig domains of IL-1RI enfold one-half of the IL-1 molecule [5]. Based on mathematical algorithms, it has been predicted that IL-1RAcP in turn wraps around the IL-1/IL-1RI complex, predominantly via interactions with IL-1RI involving only minimal contacts with the partially enclosed IL-1 [6]. Such a model largely explains the lack of affinity of IL-1RAcP itself for IL-1 [4]. However, controversy remains over whether IL-1RAcP affects the binding kinetics and dissociation constant of IL-1 with IL-1RI [4,7 – 9]. Formation of the trimeric IL-1/IL-1RI/IL-1RAcP complex triggers recruitment, and binding to the intracellular domains of IL-1RAcP and IL-1RI [10 – 15], of several intra-

0898-6568/03/$ - see front matter D 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0898-6568(03)00039-1

794

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

cellular adaptor proteins and kinases, including Toll-interacting protein (Tollip), myeloid differentiation factor 88 (MyD88), and members of the IL-1R associated kinase (IRAK) family [10,11,16– 20]. These in turn activate additional down-stream signalling factors. The signalling pathway end-points include stabilization of mRNAs encoding a range of immune-related products including COX-2, IL-2 and IL-8 [21 – 24] and activation of the transcription factors NF-nB and AP-1, which modulate gene expression and hence cellular phenotype. The intracellular signalling cascades activated by the IL-1/IL-1RI/IL-1RAcP complex are also utilized by the Toll receptors, which as part of the innate immune response recognize foreign agents such as LPS, peptidoglycan, bacterial DNA and double-stranded RNA (reviewed in Refs. [25 –27]). The IL-1RAcP gene contains 12 exons spanning more than 100 kb [28]. Two alternatively spliced IL-1RAcP mRNAs have previously been identified [4,28]. The longer splice variant, an approximately 5.3 kb mRNA derived from all 12 exons, encodes the membrane-bound form of IL1RAcP (hereafter mIL-1RAcP for clarity, Table 1) that is an integral part of the IL-1/IL-1RI/IL-1RAcP trimeric complex. The shorter splice variant, an approximately 2.3 kb mRNA derived from the first 9 exons (Table 1), encodes a smaller protein comprising the three extracellular Ig domains and a short unique C-terminus [4,28]. The latter protein lacks the transmembrane and intracellular domains of mIL-1RAcP, and is secreted in a soluble form ([4], L.E. Jensen, unpublished observation), hence its designation as soluble IL1RAcP (sIL-1RAcP). Since sIL-1RAcP lacks the intracellular domain, which is needed for intracellular signalling, this splice variant cannot have the same function as mIL-1RAcP. Our previous results have suggested that sIL-1RAcP acts as an inhibitor of IL-1 signalling [28], possibly via competitive inhibition of the interaction between mIL-1RAcP and the IL1/IL-1RI complex.

Control mechanisms have evolved to ensure that IL-1 signalling does not continue beyond the time required for the protective physiological processes to be fully engaged. The best described mechanisms involve a decoy receptor (IL-1R type II), which acts as a ligand sink for IL-1 [3,29,30], and a classical receptor antagonist (IL-1R antagonist), which competes with IL-1 for binding to the type I receptor [31,32]. However, these control mechanisms are not always sufficient to prevent IL-1 from playing a pivotal role in the etiology of a range of diseases. The excessive inflammation and destruction of cartilage in the joints of patients with rheumatoid arthritis (reviewed in Refs. [33,34]) is an example of the pathologic consequences that may result from high and sustained local exposure to IL-1 and TNF-a, another cytokine with effects similar to those mediated by IL-1. Activation of NF-nB can protect cells from undergoing apoptosis by up-regulating factors such as inhibitors of apoptosis (IAPs), FLIP and the Bcl-2 homologue A1/Bfl-1 [35 – 41], which inhibit the apoptotic signalling cascade. NFnB may play a role in oncogenesis through such a mechanism by ‘‘rescuing’’ cells, which otherwise have been ‘‘programmed’’ to die (reviewed in Ref. [42]). It has been reported that IL-1 prevents apoptosis in keratinocytes, chondrocytes, osteoclasts, neutrophils, monocytes and lymphocytes via a mechanism that probably involves activation of NF-nB [43 – 49]. IL-1 may similarly promote the growth and metastasis of some cancers, such as leukemia and melanomas (see Ref. [1]). Interestingly, polymorphism in the IL-1 gene cluster has recently been associated with increased risk of chronic hypochlorhydric responses to Helicobacter pylori and gastric cancer [50]. We have previously shown that the proportional expression of mIL-1RAcP and sIL-1RAcP mRNAs changes during an inflammatory response induced by phorbol esters [28]. While screening for additional agents which could

Table 1 Overview of exons and introns in the IL-1RAcP gene Exon

RNA segment/ protein domain

mRNA position (bp)

Exon length (bp)

Length of intron (kb)

Exons in mIL-1RAcP

Exons in sIL-1RAcP

Exons in sIL-1RAcP-h

1 2 3 4 5 6 7 8 9 9 10 11 12

5VUTR 5VUTR Signal peptide Ig 1 Ig 2A Ig 2B

1 – 118 119 – 205 206 – 270 271 – 556 557 – 743 744 – 909 910 – 981 982 – 1108 1109 – 1257 1258 – 2041 1258 – 1407 1408 – 1551 1552 – 4726

118 87 65 286 187 166 72 127 149 784 150 144 3175

42* 8 40* 4.5 10 3 4 1.9 0/15

+ +/
+ + + + + + +
+ + +

+ +/
+ + + + + + + +




+ +/
+ + + + + +

+ + +

Ig 3A Ig 3B Soluble C-terminus + 3VUTR Transmembrane (Ig 3C) Intracellular (3VUTR) Intracellular + 3VUTR (3VUTR)

1.3 2.5

The positions of intron – exon boundaries and the sizes of exons and introns were derived from Ref. [28] and GenBank accession no. AC008249 (labelled with *). Numbers for mRNA position are from GenBank accession no. AB006537. Exons present or absent in any of the three alternatively spliced IL1RAcP isoforms are indicated with + or
, respectively. RNA segments and protein domains indicated in brackets indicate elements specifically encoded by sIL-1RAcP-h.

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

influence the expression of IL-1RAcP mRNA species, we observed that two known inducers of apoptosis, UV light and staurosporine give rise to distinctly different patterns of IL-1RAcP mRNAs expression and identified a novel third alternatively spliced IL-1RAcP mRNA. Here we define the expression of all three IL-1RAcP mRNA species in response to cellular stress and discuss the potential impact that different patterns of IL-1RAcP mRNA expression may have on IL-1-mediated apoptotic rescue and oncogenesis.

795

fied. To reduce the efficiency of amplification from the ribosomal RNA, competimers (primers that have been modified to prevent their participation in amplification) were used according to the manufacturer’s instructions. A 4:1 ratio of competimers to primers was used. Twenty PCR cycles, which in initial optimization experiments had proved to be within the exponential phase of amplification, were performed, and products were quantified using PAGE and image analysis as described for the proportional RT-PCR. 2.3. Identification of exons present in IL-1RAcP mRNAs

2. Materials and methods 2.1. Cell line and treatments Human hepatoma cells (HepG2) were maintained in Dulbecco’s modified Eagle’s medium with 25 mM HEPES and glutamax-1 (L-alanyl-L-glutamine) supplemented with 10% (v/v) foetal calf serum, 1 mM sodium pyruvate, 0.01 mM nonessential amino acids and 50 Ag/ml gentamicin (GibcoBRL, Grand Island, NY). Cells were grown to approximately 90% confluence and treated with 150 ng/ml phorbol 12,13-didecanoate (PDD), 5 AM staurosporine, 50 AM isoquinoline, 1 AM calcium ionophore A23187, 2 AM 1,2 dioctanoyl-rac-glycerol, 10 nM dexamethasone (Sigma, St. Louis, MO), 9 kJ UV-C light, 10 ng/ml IL-1a, 10 ng/ml IL-1h (National Cancer Institute, Frederick, MD), 10 ng/ml IL-6 (AstraZeneca, Wilmington, DE), 50 ng/ml TNF-a (Zeneca Pharmaceuticals, Macclesfield, UK), 10 ng/ml IFN-g, 1 ng/ml IL-2, 10 ng/ml IL-4, 20 ng/ml IL-10, 10 ng/ml IL-13 or 20 ng/ml IL-18 (Peprotech, Rocky Hill, NJ). To inhibit de novo protein synthesis, cells were treated with 10 Ag/ml cycloheximide (Sigma). All experiments were performed at least three times with similar outcome. 2.2. Proportional and quantitative RT-PCR Isolation of total RNA and proportional RT-PCR were performed as described elsewhere [28] with minor modifications. Reverse transcription of 1 Ag total RNA was performed at 42 jC using AMV reverse transcriptase (Promega, Madison, WI) and oligo(dN)6 primer (Amersham Pharmacia Biotech, Piscataway, NJ) and ANTI-RNase inhibitor (Ambion, Austin, Tx). A 357 bp product was amplified from sIL1RAcP mRNA, and a 305 bp product was amplified from mIL-1RAcP. PCR products were analysed on SYBR Green I (Molecular Probes, Eugene, OR) stained 10% polyacrylamide (29:1 acrylamide/bis-acrylamide) gels using STORM and ImageQuant technology (Molecular Dynamics, Sunnyvale, CA). Quantitative RT-PCR (QRT-PCR) for analysis of levels of individual IL-1RAcP mRNA species employed the QuantumRNAk 18S Internal Standards (Ambion). In brief, the RNA was reverse transcribed as described above. In addition to the PCR products derived from IL-1RAcP mRNA, a fragment of the 18S ribosomal RNA was ampli-

The 5V regions of individual IL-1RAcP mRNAs were amplified using the GC-RICH PCR System (Roche, Indianapolis, IN) according to the manufacturer’s instructions with a forward primer (5V-TGCCGGGATCCAGGTCTC-3V, position 1– 18, GenBank accession no. AB006537) located in the 5Vuntranslated region (UTR) of the previously described full-length mIL-1RAcP mRNA and one of the IL-1RAcP mRNA specific reverse primers used for QRT-PCR. The 3V regions were amplified using the IL-1RAcP mRNA specific forward primer used for QRT-PCR and a reverse primer located within the 3V UTR of the previously described fulllength mIL-1RAcP mRNA (5V-GAGGAAGCAGAATAGTCAGC-3V, position 2082– 2063 in GenBank accession no. AB006537). PCR products were ligated into pCR-Blunt IITOPO (Invitrogen, Carlsbad, CA) and sequenced in both directions. Sequences were analysed using programs available from the NCBI blast homepage. 2.4. Caspase-3 activation and DNA fragmentation Total proteins were extracted on ice in 50 mM HEPES, 150 mM NaCl, 20 mM h-glycerophosphate, 1 mM EDTA, 1 mM benzamidine, 50 mM NaF, 1 mM Na3VO4, 2 mM dithiothreitol, 10% (v/v) Protease Inhibitor Mixture (Sigma) and 1% Nonidet P-40. Amounts of total protein were quantified using Coomassie Protein Assay Reagent (Pierce, Rockford, IL), and equal amounts for each sample were separated by electrophoresis in NuPAGE 4 –12% Bis –Tris gels (Invitrogen) and transferred to Hybond ECL nitrocellulose membranes (Amersham Pharmacia Biotech). Excess protein binding sites were blocked with 5% (w/v) nonfat milk in PBS. Immunodetection of caspase-3 was performed using polyclonal rabbit anti-caspase-3 (caspase-3 (H-277), Santa Cruz Biotechnology, Santa Cruz, CA), horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin (Sigma) and enhanced chemiluminescence (ECL Western blotting detection reagents, Amersham Pharmacia Biotech). Cells were detached from tissue culture plates using trypsin, and DNA was extracted using Generation Capture columns (Gentra Systems, Minneapolis, MN), according to the manufacturer’s instructions. DNA was separated by gel electrophoresis and stained with ethidium bromide or SYBR Green I.

796

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

3. Results 3.1. Expression of sIL-1RAcP mRNA is up-regulated in response to phorbol esters Using a proportional RT-PCR method, we have previously shown that in HepG2 cells, the expression of sIL1RAcP mRNA is increased relative to that of mIL-1RAcP mRNA [28] in response to the phorbol esters phorbol 12myristate 13-acetate and PDD, which induce an in vitro inflammatory response. In brief, the above method relies on co-amplification of 357 and 305 bp PCR products from reverse transcribed sIL-1RAcP and mIL-1RAcP mRNA species, respectively, utilizing a common forward primer and mRNA species specific reverse primers. It consequently permits only the proportional expression of the two mRNA species to be determined. Hence, our previous studies could not establish whether the changed mRNA ratios following phorbol ester treatments were due to increased expression of sIL-1RAcP mRNA or decreased expression of mIL-1RAcP mRNA, or both. To define the mechanism(s) leading to the observed changes in sIL-1RAcP to mIL-1RAcP mRNA ratios, we employed the QuantumRNAk 18S Internal Standards available from Ambion. The system utilizes the 18S rRNA as an internal standard in each RNA preparation under test to allow individual mRNA species to be quantified during the exponential phase of PCR amplification. We used the above QRT-PCR method to determine levels of sIL-1RAcP and mIL-1RAcP mRNA in samples derived from cells treated with PDD for up to 72 h (Fig. 1A). At the beginning of the time-course (0 h), mIL-1RAcP and sIL1RAcP mRNAs constitute approximately 70% and 30%, respectively, of the total amount of IL-1RAcP mRNA (Fig. 1B). Six hours after treatment the ratio of the two splice variants are approximately equal. By 12 h, the starting ratio has reversed, and at both the 24 and 36 h time-points, the sIL-1RAcP mRNA is the dominant splice variant comprising approximately 80% of the total. By 72 h, at which point there is an approximately 3:2 ratio of sIL-1RAcP to mIL1RAcP (Fig. 1B), the two alternatively spliced mRNAs are trending back towards their pretreatment ratio. No significant change in the levels of mIL-1RAcP mRNA were observed following PDD treatment (Fig. 1C). In contrast, the levels of sIL-1RAcP mRNA increased dramatically (Fig. 1C); at the 24 h peak, they were 16-fold higher than baseline (Fig. 1C). By 72 h, however, sIL-1RAcP mRNA levels had almost returned to those observed in untreated cells. 3.2. Proportional expression of sIL-1RAcP and mIL-1RAcP mRNAs changes in response to UV light

Fig. 1. Expression of mIL-1RAcP and sIL-1RAcP mRNAs in response to PDD. HepG2 cells were treated with PDD and total RNA extracted at the time-points indicated. (A) The mIL-1RAcP and sIL-1RAcP cDNAs were co-amplified with rRNA using QRT-PCR as described in Materials and methods. PCR products were visualized using SYBR Green I staining after PAGE. Lane marked M represents size markers. Positions of rRNA, sIL1RAcP and mIL-1RAcP PCR products are indicated on the right. (B) Band intensities in (A) were quantified and adjusted for size of PCR product. Levels of mIL-1RAcP (4) and sIL-1RAcP (y) mRNA species are represented as percentage of the total amount of IL-1RAcP mRNA at a given time-point. (C) Band intensities in (A) were quantified. Levels of mIL-1RAcP (4) and sIL-1RAcP (y) mRNA species were standardized against the level of rRNA in each individual sample. The levels of mIL1RAcP and sIL-1RAcP mRNAs are represented as fold increase compared to the levels of each individual mRNA species at the beginning of the experiment (0 h). Two independent samples were analysed per time-point and standard deviations are indicated with error bars.

We used the previously described [28] proportional RTPCR method to screen for agents which may modulate the proportional expression of sIL-1RAcP and mIL-1RAcP mRNAs. The following agents had no effect: isoquinoline, A23187, 1,2 dioctanoyl-rac-glycerol, dexamethasone, IL-

1a, IL-1h, IL-2, IL-4, IL-6, IL-10, IL-13, IL-18, IFN-g and TNF-a. However, UV light and staurosporine both induced changes in the proportional expression of IL-1RAcP mRNAs (not shown).

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

797

using QRT-PCR (Fig. 2). Proportional analysis (Fig. 2B) established that a change occurred within 1 h of UV exposure such that the relative proportions of mIL-1RAcP and sIL-1RAcP mRNAs changed from 70% and 30% to 55% and 45% (Fig. 2B). Examination of relative levels of individual IL-1RAcP mRNA species during the time-course revealed that sIL1RAcP mRNA remained essentially unchanged (Fig. 2C), whereas mIL-1RAcP mRNA was rapidly down-regulated after UV exposure. After only 1 h, the mIL-1RAcP mRNA level was 83% that of untreated cells, and by 6 h, the level had declined to approximately 70%. mIL-1RAcP mRNA levels remained depressed through at least 36 h, but by 72 h had recovered to 86% of levels observed in untreated cells (Fig. 2C). UV light is a well-known inducer of apoptosis. To establish whether apoptosis had been initiated in the cell population examined above, we harvested total cellular protein at different times following UV exposure and measured levels of caspase-3, a key intermediate in many of the signalling pathways that lead to apoptosis. Inactive caspase3, a 30 kDa protein, is activated by cleavage to an active 20 kDa fragment, thereby permitting the biological engagement of caspase-3 in response to a given stimulus to be monitored. Levels of inactive full-length caspase-3 decreased at the later time-points, i.e. 24, 36 and especially 72 h (Fig. 3). However, only very low amounts of the activated 20 kDa fragment were observed at 24 and 36 h and none at 72 h. We observed no evidence of apoptosis as assessed by DNA fragmentation (not shown) in the adherent cell population. 3.3. Identification of sIL-1RAcP-b mRNA, a novel IL1RAcP alternative splice variant

Fig. 2. Expression of mIL-1RAcP and sIL-1RAcP mRNAs in response to UV light. HepG2 cells were exposed to UV light and total RNA extracted at the time-points indicated. (A) The two IL-1RAcP cDNAs were co-amplified with rRNA using QRT-PCR as described in Materials and methods. PCR products were visualized using SYBR Green I staining after PAGE. Lane marked M represents size markers. Positions of rRNA, sIL-1RAcP and mIL1RAcP PCR products are indicated on the right. (B) Band intensities in (A) were quantified and adjusted for size of PCR product. Levels of mIL-1RAcP (4) and sIL-1RAcP (y) mRNA species are represented as percentage of the total amount of IL-1RAcP mRNA at a given time-point. (C) Band intensities in (A) were quantified. Levels of mIL-1RAcP (4) and sIL-1RAcP (y) mRNA species were standardized against the level of rRNA in each individual sample. The levels of mIL-1RAcP and sIL-1RAcP mRNAs are represented as percentage of the levels of each individual mRNA species at the beginning of the experiment (0 h). Two independent samples were analysed per time-point and standard deviations are indicated with error bars.

To examine further the UV inducible changes in mIL1RAcP and sIL-1RAcP mRNA expression, HepG2 cells were exposed to UV light and total RNA harvested at different time-points from cells that remained attached to the culture plates. IL-1RAcP mRNA levels were examined

We also examined the consequences of treating cells with staurosporine, another general inducer of apoptosis and inhibitor of protein kinases. RNA was extracted from cells treated with staurosporine for 0– 72 h and analysed by QRTPCR. From the 3 h time-point through the end of the timecourse, an approximately 150 bp product was observed in addition to the 305 and 357 bp products derived from the sIL1RAcP and mIL-1RAcP mRNAs, respectively (Fig. 4A). The

Fig. 3. Activation of caspase-3 in response to UV light. Cells were exposed to UV light and total protein extracted at the time-points indicated. Equal amounts of total protein per sample were analysed using Western blotting with an antibody directed against full-length caspase-3. Positions of the uncleaved 30 kDa caspase-3 and the activated/cleaved 20 kDa caspase-3 fragment are indicated. NS indicates position of protein nonspecifically detected by the assay.

798

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

Fig. 4. Expression of sIL-1RAcP, mIL-1RAcP and sIL-1RAcP-h mRNA in response to staurosporine. HepG2 cells were treated with staurosporine and total RNA extracted at the time-points indicated. (A) The three IL-1RAcP cDNAs were co-amplified with rRNA using QRT-PCR as described in Materials and methods. PCR products were visualized using SYBR Green I staining after PAGE. Lane marked M represents size markers. Positions of rRNA, sIL-1RAcP, mIL-1RAcP and sIL-1RAcP-h PCR products are indicated on the right. (B) Band intensities in (A) were quantified and adjusted for size of PCR product. Levels of sIL-1RAcP (y), mIL-1RAcP (4) and sIL-1RAcP-h (.) mRNA species are represented as percentage of the total amount of IL-1RAcP mRNA at a given time-point. (C) Band intensities in (A) were quantified. Levels of sIL-1RAcP (y), mIL-1RAcP (4) and sIL-1RAcP-h (.) mRNA species were standardized against the level of rRNA in each individual sample. The levels of sIL-1RAcP and mIL-1RAcP mRNAs are represented as percentage (scale to the left) of the levels of each individual mRNA species at the beginning of the experiment (0 h). Levels of sIL-1RAcP-h mRNA are represented as fold increase (scale to the right) compared to the levels at the beginning of the experiment (0 h). Two independent samples were analysed per time-point and standard deviations are indicated with error bars.

approximately 150 bp product was not observed after PDD treatment or UV exposure (Figs. 1A and 2A). As the forward and reverse primers used for RT-PCR are located in exon 8 (forward primer), exon 9 (reverse primer specific for sIL1RAcP) and exon 10 (reverse primer specific for mIL1RAcP), we hypothesized that the approximately 150 bp product arose from a novel splice variant of the IL-1RAcP pre-mRNA in which exon 9 is excluded (Table 1, predicted size, 156 bp). Sequencing of the approximately 150 bp product confirmed this hypothesis (sequence submitted to GenBank under accession no. AF4873351). The lack of exon 9 leads to a frame shift that begins at the 5Vend of exon 10 and results in an alternative open reading frame specifying a unique 45 amino acid sequence followed by a stop codon (Fig. 5). This unique sequence does not encode a transmembrane domain and the resulting protein is therefore likely to be secreted. Consequently, we have named this novel alternative splice variant ‘‘sIL-1RAcP-h’’. To determine which additional exons down-stream of exon 10 are present in the sIL-1RAcP-h mRNA, we used the forward QRT-PCR primer and a reverse primer complementary to sequence from the known 3V UTR (exon 12) of the mIL-1RAcP mRNA to amplify an RT-PCR product from RNA extracted from staurosporine treated cells. A product of approximately 1 kb was obtained and cloned into a sequencing vector. Individual clones lacking exon 9 were identified using the primers from the QRT-PCR method. Sequencing of such clones revealed that sIL-1RAcP-h encodes exon 11 and 12 (intracellular domains, Table 1) in addition to exons 8 and 10. Exons up-stream of exon 8 were identified in a similar manner using a forward primer corresponding to sequence in the known 5V UTR of mIL1RAcP and the mIL-1RAcP mRNA specific reverse primer from the proportional RT-PCR. Sequencing of clones lacking exon 9 established that sIL-1RAcP-h encodes all exons from exon 3 (the signal peptide) to exon 8 (Table 1). Consequently, the full-length sIL-1RAcP-h protein comprises Ig domains 1 and 2, Ig domain 3A (from exon 8) and a unique second half of the third Ig domain, which we have named Ig 3C (Table 1). The sIL-1RAcP-h mRNA encodes a polypeptide of 346 amino acids, which is 10 residues shorter than the previously described sIL-1RAcP. Alignment of the full-length novel Ig 3 domain and surrounding sequence from sIL-1RAcP-h with the Ig domains of mIL-1RAcP and sIL-1RAcP (Fig. 5) revealed that 8 amino acids from the unique 45 residue peptide are strictly conserved (18%). One of these conserved residues is Cys(332), which forms a disulfide bond with Cys(266), thereby contributing to the characteristic ‘‘closing’’ of the Ig domain structure (Fig. 5). This allows for the possibility of maintaining the overall Ig domain structure and size of the novel Ig 3 domain of sIL-1RAcP-h. The previously described sIL-1RAcP contains an additional Cys 1 The sequences published in this paper have been submitted to GenBank under the accession nos. AF487335 and AF538730 – 538734.

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

799

Fig. 5. Alignment of Ig three domains and surrounding sequence of mIL-1RAcP, sIL-1RAcP and sIL-1RAcP-h. Numbers above the sequence indicate position within the full-length proteins (including signal-peptide). Three gaps, indicated with (-), in the sIL-1RAcP-h sequence were introduced for optimal alignment. Stop codons are represented as asterisk (*). Peptide sequence encoded by exon 8 (Ig 3A domain), which is identical in all three isoforms, is shown in lower case letters. The Ig 3B sequence of mIL-1RAcP and sIL-1RAcP and the Ig 3C sequence of sIL-1RAcP-h are shown in upper case letters. Positions of conserved residues between Ig 3B and Ig 3C are indicated with arrowheads (^). Positions of the two Cys residues, which form the characteristic Ig domain disulfide bond, are indicated with triangles (4). Cys residues, which may be involved in dimer formation, are indicated with diamonds ( w ).

residue, Cys(354), within the C-terminal 6 residues, which are unique to this isoform (Fig. 5), suggesting that sIL1RAcP forms dimers. The sIL-1RAcP-h isoform may also form dimers since two Cys residues (Fig. 5), in addition to the ‘‘Ig domain’’ pair, are present in Ig 3C (sIL-1RAcP-h) but not Ig 3B (mIL-1RAcP and sIL-1RAcP). Sequencing of the sIL-1RAcP-h clones additionally revealed alternative splicing between exon 1 and exon 3 (Table 1) in two of three clones. The two full-length alternatively spliced sIL-1RAcP-h mRNA sequences have been submitted to GenBank under accession nos. AF538730 and AF538731. To determine whether the consequent generation of alternative 5V UTRs is specific to sIL-1RAcP-h mRNA, we also sequenced clones, generated as above, containing exon 9 (i.e. clones derived from mIL-1RAcP mRNA); three of four such clones contained contiguous exon 1 – exon 3 without exon 2. We next amplified the 5V region of the sIL1RAcP mRNA using the forward 5VUTR primer from exon 1 and the QRT-PCR reverse primer specific for this isoform. Sequencing identified the exon 1 –exon 3 splice variation in two of four clones. Thus, the exon 1 – exon 3 and exon 1 – exon 2 – exon 3 splice variants are present in all three IL1RAcP mRNA species. The sequences of the mIL-1RAcP mRNA containing the shorter 5V UTR and the two forms of the sIL-1RAcP mRNA have been submitted to GenBank under accession nos. AF538733, AF538732 and AF538734, respectively. 3.4. Proportional expression of three IL-1RAcP mRNAs in response to staurosporine Due to the induction of mIL-1RAcP-h, analysis of the proportional levels of the IL-1RAcP mRNAs following treatment of cells with staurosporine necessarily revealed a pattern (Fig. 4A,B) different from those induced by PDD or UV light. With respect to the mIL-1RAcP and sIL-1RAcP mRNAs, the beginning of the time-course resembles that following PDD treatment in that there were approximately equal amounts of the two splice variants (40% each of the total IL-1RAcP mRNA versus 70% and 30% of mIL-1RAcP and sIL-1RAcP mRNAs, respectively, at 0 h) after 6 h, and the sIL-1RAcP isoform predominated at later time-points. However, in contrast to the PDD response, the proportional bias in favour of sIL-1RAcP became more marked over time. At 72 h, the sIL-1RAcP and mIL-1RAcP mRNAs constituted 74% and 12%, respectively, of total IL-1RAcP mRNA

(Fig. 4B). The novel sIL-1RAcP-h splice variant was only present in trace amounts at the beginning of the time-course (Fig. 4A,B); however, within 6 h of adding staurosporine, sIL-1RAcP-h constituted approximately 20% of the total IL1RAcP mRNA, a proportion that was maintained at least through 36 h. By 72 h, the proportion of sIL-1RAcP-h had declined to 14% of total IL-1RAcP mRNA(Fig. 4B). Analysis of individual levels of the three IL-1RAcP mRNAs following staurosporine treatment (Fig. 4C) revealed that although early changes in the proportional expression of mIL-1RAcP and sIL-1RAcP mRNAs resembled those induced by PDD, the underlying mechanism was different. Whereas sIL-1RAcP mRNA levels were dramatically increased by PDD, the amount of this mRNA species remained largely unchanged through 24 h in response to staurosporine. Subsequently, at the 36 and 72 h time-points, modest decreases in sIL-1RAcP mRNA levels to 90% and 83% of those observed at 0 h were observed (Fig. 4C). In contrast, mIL-1RAcP mRNA levels declined steadily throughout the entire time-course, falling to 50% of baseline after 12 h and 12% after 72 h (Fig. 4C). Most striking was the specific induction of the novel sIL1RAcP-h mRNA from trace amounts in untreated cells to a 17-fold increase only 3 h after staurosporine treatment, a level that was maintained through 12 h (Fig. 4C). Although levels declined steadily thereafter, the 72-h level was still four times higher than the pretreatment level (Fig. 4C). Activation of caspase-3 and DNA fragmentation was examined after treatment with staurosporine. As with the UV treated cells, no signs of DNA fragmentation were observed at any time-point (not shown). However, low levels of activated caspase-3 were observed at 24, 36 and 72 h (Fig. 6),

Fig. 6. Activation of caspase-3 in response to staurosporine. Cells were treated with staurosporine and total protein extracted at the time-points indicated. Equal amounts of total protein per sample were analysed using Western blotting with an antibody directed against full-length caspase-3. Positions of the uncleaved 30 kDa caspase-3 and the activated/cleaved 20 kDa caspase-3 fragment are indicated. NS indicates position of protein nonspecifically detected by the assay.

800

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

Fig. 7. Dependence of PDD mediated changes in sIL-1RAcP mRNA expression on de novo protein synthesis. HepG2 cells were treated with medium only or PDD in the presence or absence of cycloheximide. Total RNA was extracted at 24 h and expression of sIL-1RAcP mRNA was evaluated using QRT-PCR as described in Materials and methods and legend in Fig. 1C.

indicating that apoptotic signalling pathways had been activated by these later time-points. 3.5. Changes in proportional expression of mIL-1RAcP and sIL-1RAcP in response to PDD require de novo protein synthesis To determine whether the observed changes in expression of alternatively spliced IL-1RAcP mRNAs in response to PDD, UV light and staurosporine were mediated by mechanisms that require de novo protein synthesis, we pretreated cells with cycloheximide for 1 h before exposure to the above agents. RNA was extracted after 24 h, and expression of the three alternatively spliced IL-1RAcP mRNAs was analysed by QRT-PCR. Cycloheximide had no effect on the changes in proportional expression induced by UV light or staurosporine (not shown). In contrast, cycloheximide treatment blocked the PDD-induced increase in sIL-1RAcP mRNA levels (Fig. 7), establishing that the capacity of PDD to up-regulate this IL-1RAcP mRNA species is dependent upon de novo protein synthesis.

4. Discussion IL-1RAcP is an essential component of IL-1 signalling. The two previously described membrane-bound and soluble/secreted alternative splice variants of IL-1RAcP are unlikely to have identical functions as it has been shown that the intracellular portion of the membrane-bound form (which is absent in the soluble isoform) is necessary for intracellular signalling [11 –13,15]. Furthermore, our published data have suggested that sIL-1RAcP down-regulates IL-1 signalling [28], probably by competitive inhibition of the association of mIL-1RAcP with the IL-1RI/IL-1 complex. The likelihood that the two major IL-1RAcP isoforms have opposite roles (i.e. supporting and suppressing IL-1 signalling) suggests that the relative levels at which they are expressed would be an important determinant of cellular

responses during inflammation. The existence of sIL1RAcP-h, the novel splice variant of IL-1RAcP described above, potentially adds an additional level of complexity to the regulation of IL-1 signalling. Exon 9, which encodes the second half of the third Ig domain of both sIL-1RAcP and mIL-1RAcP (Table 1), is absent from the sIL-1RAcP-h mRNA. Surprisingly exon 10, which in mIL-1RAcP encodes the transmembrane domain, contains a second open reading frame that appears to encode a unique second half of the third immunoglobulin domain in sIL-1RAcP-h. Of the 45 amino acids in the novel partial Ig domain, 8 residues are conserved. Most significantly, one of these residues is a Cys that probably participates in the formation of the characteristic Ig domain disulfide bond (Fig. 5). Interestingly, the second half of the third Ig domain has minimal interaction with either IL-1RI or IL-1 in the predicted model of the trimeric receptor– ligand complex [6]. We have previously demonstrated that the three extracellular Ig domains are sufficient for association of mIL1RAcP and sIL-1RAcP with the IL-1RI/IL-1 complex [28]. Since sIL-1RAcP-h encodes the first two Ig domains and half of the common third Ig domain, it may be able to interact with IL-1RI and IL-1 to form a trimeric complex similar to that formed by mIL-1RAcP. If IL-1RAcP-h does form a trimeric complex with IL-1 and IL-1RI, it may have a function similar to that proposed for the previously described sIL-1RAcP, i.e. inhibition of IL-1 signalling via competition with the signal-facilitating mIL-1RAcP for association with the IL-1RI/IL-1 complex. The above results have established that under different conditions of stress, the three IL-1RAcP mRNAs are differentially expressed. The patterns of expression appear to be shaped by distinct mechanisms. Whereas the increase in sIL-1RAcP mRNA induced by PDD (Fig. 1) requires de novo protein synthesis (Fig. 7), changes induced by apoptotic signals do not. The PDD mediated increase in sIL1RAcP mRNA levels may be due to a transient up-regulation of transcription from the IL-1RAcP gene. However, given that there were no concomitant changes in mIL1RAcP mRNA levels in response to PDD, we consider it more likely that the increased levels of sIL-1RAcP mRNA arise from enhanced stability mandating its accumulation over time. PDD treatment does not result in changes in mIL-1RAcP mRNA levels, but treatment with either UV light or staurosporine results in significantly decreased levels of this mRNA species (Figs. 1C, 2C and 4C). Interestingly, several AU-rich sequence elements are present within the 3V untranslated region of the mIL-1RAcP mRNA (L.E. Jensen, unpublished observation). Such AU-rich sequence elements are associated with mRNA instability [51,52], and it is possible that in response to pro-apoptotic stress, mIL1RAcP mRNA stability/instability may be modulated through such regulatory elements. However, the precise mechanisms governing the effect of UV light and staurosporine on mIL-1RAcP mRNA stability must differ, at least

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

in part, as levels are immediately down-regulated and subsequently stabilize in response to the former, whereas levels decline gradually over time in response to the latter (Figs. 2C and 4C). These differences could be explained as follows. In response to UV light, mIL-1RAcP mRNA may be more rapidly degraded through activation of a presynthesized destabilizing factor and/or inactivation of a stabilizing factor. In the absence of transcriptional modulation, a lower steady-state concentration of mIL-1RAcP mRNA would eventually be reached. In contrast, the decreased level of mIL-1RAcP mRNA following treatment with staurosporine may be due to partial or complete inhibition of transcription or redirection of pre-mRNA splicing factors (see below) leading to a gradual decrease in mRNA levels that do not reach a new steady-state within the time-frame of the experiment. sIL-1RAcP-h is exclusively expressed in staurosporine treated cells, in which the dramatically increased levels of this mRNA species (Fig. 4C) suggest up-regulated, or at least continued, transcription. However, given the concurrent decrease in mIL-1RAcP mRNA, an attractive hypothesis is that in response to staurosporine, the splicing machinery is redirected such that sIL-1RAcP-h mRNA rather than mIL-1RAcP mRNA is generated. At the later time-points (36 and 72 h), transcription of pre-mRNA may additionally be down-regulated such that levels of all three IL-1RAcP splice variants decrease (Fig. 4D). Staurosporine, a microbial cell-permeable alkaloid from Streptomyces staurosporeus, is a well-known inducer of apoptosis and inhibitor of protein kinases. Whether the effects on IL1RAcP mRNA expression are due to the former and/or latter activity of staurosporine cannot be determined from our data. However, it should be noted that two kinases that are inhibited by staurosporine are IKK (inhibitor of NF-nB kinase) 1 and 2 [53,54], which are responsible for the phosphorylation of InB, an event that targets InB for degradation, which in turn allows NF-nB to translocate to the nucleus. A tantalising possibility is that staurosporine is part of an immune evasion mechanism that first targets the local Toll-mediated activation of NF-nB by inhibiting the IKKs, and then targets the systemic responses to IL-1 by not only inhibiting the IKKs but also by modulating IL-1RAcP pre-mRNA splicing to down-regulate mIL-1RAcP mRNA and up-regulate sIL-1RAcP-h mRNA, thereby allowing targeting of all intracellular signalling pathways activated by IL-1, e.g. activation of both NF-nB and AP-1. In mice, mIL-1RAcP mRNA is up-regulated in spleen, lung and thymus in response to intraperitoneal IL-1; however, both mIL-1RAcP and sIL-1RAcP are constitutively expressed at high levels in brain and liver, respectively [4]. In rats, the sIL-1RAcP mRNA has been shown to be upregulated in several parts of the brain after intracerebroventricular administration of IL-1 [55,56]. Conflicting observations (up- or down-regulation or constitutive expression) have been reported regarding the expression of IL-1RAcP mRNAs in brain sections and liver following intraperitoneal

801

administration of LPS [57 –59]. The different patterns of IL1RAcP mRNA expression observed during stress will probably mandate distinct cellular responses to IL-1 with consequently distinct adaptive phenotypes. The overall trend is consistent with a phenotypic modulation towards reduced sensitivity to IL-1 via a proportional increase in the expression of sIL-1RAcP (and sIL-1RAcP-h), the inhibitor(s) of IL-1 signalling, at the expense of the expression of mIL-1RAcP, the promoter of signalling. Such modulation would accord with an intuitive expectation that cells would engage mechanisms to prevent the adoption of an inappropriately ‘‘overstimulated’’ phenotype. IL-1 can prevent certain cell types from undergoing apoptosis [43 – 48], possibly through inducing the expression of specific inhibitors that block the apoptotic pathways. The changes in proportional expression of the three alternative splice variants of IL-1RAcP mRNA that are induced by UV light and staurosporine occur earlier than caspase-3 activation. This suggests that there is a window of time during which IL-1-mediated rescue may be affected prior to a cell becoming irreversibly committed to apoptosis. Published studies in keratinocytes have shown that IL-1 can protect cells from apoptosis induced by CD95, but not from apoptosis induced by UV light, observations that were, in part, explained by differential regulation of expression of TNF-a and IAPs [44]. Differential expression of the alternatively spliced isoforms of IL-1RAcP offers an additional means of achieving such context-dependent outcomes. Thus, if sIL1RAcP-h inhibits IL-1 signalling, cells co-stimulated with IL-1 and staurosporine (or other inducers of apoptosis with similar effects on IL-1RAcP expression) may become insensitive to IL-1 and undergo apoptosis, in part, because of reduced IL-1 signalling from the IL-1 receptor complex due to decreased levels of mIL-1RAcP and proportionally higher levels of sIL-1RAcP and sIL-1RAcP-h. In contrast, other apoptotic signals, including UV light, may lead to only modestly reduced sensitivity to IL-1 (via slightly decreased levels of mIL-1RAcP and a minor proportional up-regulation of sIL-1RAcP), leaving IL-1-dependent apoptotic rescue mechanisms intact. Our results using UV light to treat HepG2 cells differ from those previously reported using keratinocytes [44]; modulation of phenotypes in response to both inducers of apoptosis and cytokines are likely to be tissue-specific, and evaluation of IL-1RAcP isoforms (and other factors involved in IL-1 signalling) expression may prove valuable in determining mechanisms of contextdependent signalling and/or apoptosis. Chronic inflammatory conditions, such as hepatitis and H. pylori infections, may cause cancer. Although the etiology of such cancers is largely unknown, it is possible, under certain circumstances, that IL-1 may play a role in preventing cells which receive pro-apoptotic signals (e.g. following exposure to a carcinogen or integration of a viral genome) from undergoing programmed cell death, hence facilitating the survival of phenotypically modified pre-cancerous or malignant cells. Our results suggest that different apoptotic

802

L.E. Jensen, A.S. Whitehead / Cellular Signalling 15 (2003) 793–802

signals may have context-dependent consequences, i.e. some pro-apoptotic signals may still function to eliminate damaged/transformed cells under inflammatory conditions whereas other signals are overridden due to the presence of IL-1. Further in vitro and in vivo studies may reveal whether modulated sensitivities to IL-1 following ‘‘natural’’ activators of apoptosis will impact cell survival and whether IL-1 plays a direct role in malignant transformation associated with certain inflammatory conditions. If this proves to be the case, IL-1 signalling and its constituent sub-pathways would be ideal targets for the development of therapeutic strategies to reduce the risk of cancer in patients with inflammatory conditions such as hepatitis and H. pylori infections.

References [1] Dinarello CA. Blood 1996;87:2095 – 147. [2] Sims JE, Acres RB, Grubin CE, McMahan CJ, Wignall JM, March CJ, et al. Proc Natl Acad Sci 1989;86:8946 – 50. [3] Stylianou E, O’Neill LA, Rawlinson L, Edbrooke MR, Woo P, Saklatvala J. J Biol Chem 1992;267:15836 – 41. [4] Greenfeder SA, Nunes P, Kwee L, Labow M, Chizzonite RA, Ju G. J Biol Chem 1995;270:13757 – 65. [5] Vigers GP, Anderson LJ, Caffes P, Brandhuber BJ. Nature 1997;386: 190 – 4. [6] Casadio R, Frigimelica E, Bossu P, Neumann D, Martin MU, Tagliabue A, et al. FEBS Lett 2001;499:65 – 8. [7] Korherr C, Hofmeister R, Wesche H, Falk W. Eur J Immunol 1997; 27:262 – 7. [8] Wesche H, Resch K, Martin MU. FEBS Lett 1998;429:303 – 6. [9] Cullinan EB, Kwee L, Nunes P, Shuster DJ, Ju G, McIntyre KW, et al. J Immunol 1998;161:5614 – 20. [10] Burns K, Clatworthy J, Martin L, Martinon F, Plumpton C, Maschera B, et al. Nat Cell Biol 2000;2:346 – 51. [11] Wesche H, Henzel WJ, Shillinglaw W, Li S, Cao Z. Immunity 1997; 7:837 – 47. [12] Volpe F, Clatworthy J, Kaptein A, Maschera B, Griffin AM, Ray K. FEBS Lett 1997;419:41 – 4. [13] Huang J, Gao X, Li S, Cao Z. Proc Natl Acad Sci 1997;94:12829 – 32. [14] Wesche H, Korherr C, Kracht M, Falk W, Resch K, Martin MU. J Biol Chem 1997;272:7727 – 31. [15] Radons J, Gabler S, Wesche H, Korherr C, Hofmeister R, Falk W. J Biol Chem 2002;277:16456 – 63. [16] Muzio M, Ni J, Feng P, Dixit VM. Science 1997;278:1612 – 5. [17] Burns K, Martinon F, Esslinger C, Pahl H, Schneider P, Bodmer JL, et al. J Biol Chem 1997;273:12203 – 9. [18] Wesche H, Gao X, Li X, Kirschning CJ, Stark GR, Cao Z. J Biol Chem 1999;274:19403 – 10. [19] Li S, Strelow A, Fontana EJ, Wesche H. Proc Natl Acad Sci 2002;99: 5567 – 72. [20] Cao Z, Henzel WJ, Gao X. Science 1996;271:1128 – 31. [21] Gou Q, Liu CH, Ben-Av P, Hla T. Biochem Biophys Res Commun 1998;242:508 – 12. [22] Faour WH, He Y, He QW, de Ladurantaye M, Quintero M, Mancini A, et al. J Biol Chem 2001;276:31720 – 31. [23] Greene C, O’Neill L. Biochim Biophys Acta 1999;1:109 – 21.

[24] Holtmann H, Enninga J, Kalble S, Thiefes A, Dorrie A, Broemer M, et al. J Biol Chem 2001;276:3508 – 16. [25] O’Neill LA. Trends Immunol 2002;23:296 – 300. [26] Modlin RL. Ann. Allergy, Asthma, Immunol 2002;88:543 – 7. [27] Triantafilou M, Triantafilou K. Trends Immunol 2002;23:301 – 4. [28] Jensen LE, Muzio M, Mantovani A, Whitehead AS. J Immunol 2000; 164:5277 – 86. [29] Colotta F, Dower SK, Sims JE, Mantovani A. Immunol Today 1994; 15:562 – 6. [30] Sims JE, Gayle MA, Slack JL, Alderson MR, Bird TA, Giri JG, et al. Proc Natl Acad Sci 1993;90:6155 – 9. [31] Eisenberg SP, Evans RJ, Arend WP, Verderber E, Brewer MT, Hannum CH, et al. Nature 1990;343:341 – 6. [32] Carter DB, Deibel Jr MR, Dunn CJ, Tomich CS, Laborde AL, Slightom JL, et al. Nature 1990;344:633 – 8. [33] Arend WP, Dayer JM. Arthritis Rheum 1990;33:305 – 15. [34] Krane SM, Conca W, Stephenson ML, Amento EP, Goldring MB. Ann NY Acad Sci 1990;580:340 – 54. [35] Stehlik C, de Martin R, Binder BR, Lipp J. Biochem Biophys Res Commun 1998;243:827 – 32. [36] Stehlik C, de Martin R, Kumabashiri I, Schmid JA, Binder BR, Lipp J. J Exp Med 1998;188:211 – 6. [37] Wang CY, Mayo MW, Korneluk RG, Goeddel DV, Baldwin ASJ. Science 1998;281:1680 – 3. [38] Chu ZL, McKinsey TA, Liu L, Gentry JJ, Malim MH, Ballard DW. Proc Natl Acad Sci 1997;94:10057 – 62. [39] Wang CY, Guttridge DC, Mayo MW, Baldwin ASJ. Mol Cell Biol 1999;19:5923 – 9. [40] Kreuz S, Siegmund D, Scheurich P, Wajant H. Mol Cell Biol 2001;21:3964 – 73. [41] Micheau O, Lens S, Gaide O, Alevizopoulos K, Tschopp J. Mol Cell Biol 2001;21:5299 – 305. [42] Mayo MW, Baldwin AS. Biochim Biophys Acta 2000;1470:M55 – 62. [43] Kothny-Wilkes G, Kulms D, Poppelmann B, Luger TA, Kubin M, Schwarz T. J Biol Chem 1998;273:29247 – 53. [44] Kothny-Wilkes G, Kulms D, Luger TA, Kubin M, Schwarz T. J Biol Chem 1999;274:28916 – 21. [45] Schmidt M, Pauels HG, Lugering N, Lugering A, Domschke W, Kucharzik T. J Immunol 1999;163:3484 – 90. [46] Watson RWG, Rotstein OD, Parodo J, Bitar R, Marshall JC. J Immunol 1998;161:957 – 62. [47] Lee ZH, Lee SE, Kim CW, Lee SH, Kim SW, Kwack K, et al. J Biochem 2002;131:161 – 6. [48] Kuhn K, Hashimoto S, Lotz M. J Immunol 2000;164:2233 – 9. [49] Tatsuta T, Cheng J, Mountz JD. J Immunol 1996;157:3949 – 57. [50] El-Omar EM, Carrington M, Chow WH, McColl KE, Bream JH, Young HA, et al. Nature 2000;404:398 – 402. [51] Chen CY, Shyu AB. Trends Biochem Sci 1995;20:465 – 70. [52] Ross J. Microbiol Rev 1995;59:423 – 50. [53] Peet GW, Li J. J Biol Chem 1999;274:32655 – 61. [54] Wisniewski D, LoGrasso P, Calaycay J, Marcy A. Anal Biochem 1999;274:220 – 8. [55] Plata-Salaman CR, Ilyin SE. J Neurosci Res 1997;49:541 – 50. [56] Gayle D, Ilyin SE, Plata-Salaman CR. Brain Res Bull 1997;44:311 – 7. [57] Liu C, Chalmers D, Maki R, De Souza EB. J Neuroimmunol 1996; 66:41 – 8. [58] Ilyin SE, Gayle D, Flynn MC, Plata-Salaman CR. Brain Res Bull 1998;45:507 – 15. [59] Turrin NP, Gayle D, Ilyin SE, Flynn MC, Langhans W, Schwartz GJ, et al. Brain Res Bull 2001;54:443 – 53.