Tumor necrosis factor receptor I from patients with tumor necrosis factor receptor–associated periodic syndrome interacts with wild-type tumor necrosis factor receptor I and induces ligand-independent NF-κB activation

Authors

  • Nasim Yousaf,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • David J. Gould,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Ebun Aganna,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Linda Hammond,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Rita M. Mirakian,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Mark D. Turner,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Graham A. Hitman,

    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Search for more papers by this author
  • Michael F. McDermott,

    Corresponding author
    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    Current affiliation:
    1. St. James's University Hospital, Leeds, UK
    • Academic Unit of Musculoskeletal Disease, Clinical Science Building, St. James's University Hospital, Beckett Street, Leeds LS9 7TF, UK
    Search for more papers by this author
    • Drs. McDermott and Chernajovsky contributed equally to this work.

  • Yuti Chernajovsky

    Corresponding author
    1. Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, London, UK
    • Bone and Joint Research Unit, William Harvey Research Institute, Barts and The London Queen Mary's School of Medicine and Dentistry, University of London, Charterhouse Square, London EC1M 6BQ, UK
    Search for more papers by this author
    • Drs. McDermott and Chernajovsky contributed equally to this work.


Abstract

Objective

To investigate the molecular consequences of expressing mutated forms of tumor necrosis factor receptor I (TNFRI) as found in patients with TNFR-associated periodic syndrome (TRAPS).

Methods

We cloned and expressed full-length wild-type (WT) and T50K and P46L variants of TNFRI using a new tightly regulated doxycycline-dependent expression system. This system enabled the study of molecular interactions between these receptors at both physiologic and pathophysiologic levels of expression.

Results

We used chemical crosslinking on the cell surface to show that WT and mutant forms of TNFRI, derived from TRAPS patients, interact in the absence of TNF ligand. Doxycycline-controlled up-regulation of one TNFRI allele, either WT or mutant, caused down-regulation of the other allele, indicating dynamic control of cell surface assembly. We also demonstrated that increased expression of mutant TNFRI (T50K) was associated with a parallel increase in NF-κB p65 (RelA) subunit activation, which did not occur with increased expression of WT TNFRI.

Conclusion

The T50K TRAPS-related variant is capable of sustaining inappropriate NF-κB activation, resulting in persistent autoinflammation in target organs such as skin, synovial membrane, and the central nervous system. We conclude that some of the inflammatory processes seen in TRAPS do not involve direct interaction of TNF with its receptors, but that other proinflammatory mechanisms capable of up-regulating TNFRI expression may cause cellular activation through the NF-κB signaling pathway.

Tumor necrosis factor receptor–associated periodic syndrome (TRAPS; MIM no. 142680) is a dominantly inherited chronic inflammatory disorder caused by mutations in the extracellular domains of tumor necrosis factor receptor I (TNFRI), the gene for which is TNFRSF1A (1), and characterized by recurrent fevers and abdominal pain. The most common mutations involve cysteine residues, but several variants involving other residues have also been reported (1, 2). More than 30 different pathogenic TNFRSF1A mutations have now been identified (see http://fmf.igh.cnrs.fr/infevers/ and ref. 2), mostly located in either the first or second cysteine-rich N-terminal extracellular domain (CRD1 or CRD2) of TNFRSF1A (2, 3). Impaired cleavage of the TNFRI (p55) ectodomain upon cellular activation has been demonstrated as a pathogenic mechanism in some but not all TNFRI variants (3–6) and did not relate to the severity of the phenotype (7, 8).

Biologic activities of TNF are mediated through 2 TNFRs, TNFRI and TNFRII (TNFR superfamily 1B [TNFRSF1B] p75) (9, 10). TNFR trimerization is necessary for the functional activities mediated by the receptors, and this self-assembly occurs in the absence of the ligand (11). The self-association is mediated by a pre–ligand binding assembly domain (PLAD) located within the CRD1 (11), whereas the ligand binding pocket is formed by CRD2 and CRD3 (12, 13). Engagement of TNFRI by TNF activates several transcription factors that include NF-κB and c-Jun/activator protein 1 (AP-1), leading to up-regulation of a large number of genes involved in inflammatory and immune responses (9, 10, 14).

TNFRI signaling studies to date have been difficult to perform due to the proapoptotic consequences of overexpression, and some researchers have resorted to the use of cytoplasmic truncations of TNFRI that do not reflect the physiologic events or the natural structure of these molecules. In addition, the data obtained were based on fluorescence-activated cell sorting or enzyme-linked immunosorbent assay analysis using monoclonal antibodies whose specificity for the mutated forms of TNFRI has not been carefully assessed, and biochemical analyses such as immunoblotting were not performed. In an attempt to overcome these difficulties, regulated expression of TNFRI in a tetracycline-regulated system has been tried; however, maximal levels of antibiotic were required to obtain high levels of gene expression (15, 16). Despite these caveats, investigators in those studies have hypothesized (but not demonstrated) a gain of function with proinflammatory consequences for the TNFRI mutations (15, 16).

To overcome some of these problems, we have developed a carefully controlled tetracycline-regulated system in which we tagged full-length wild-type (WT) and mutant TNFRI molecules in order to study the pathophysiologic effects of mutant TNFRI expression along with their associated biochemical and cellular consequences. We focused on the functional characterization of 2 full-length TNFRI variants, T50K and P46L. Both of these mutations are located within the CRD1 of TNFRI and are associated with defects of ectodomain shedding but represent opposite ends of the clinical spectrum. We have reported about a German family in which a T50K TNFRI mutation was present in 3 affected members who showed not only typical TRAPS features, but also varying degrees of central nervous system (CNS) involvement (17). Although P46L may be associated with TRAPS, it is not fully penetrant, being intermediate between the T50K phenotype and healthy subjects with WT TNFRI. Using both constitutive and tetracycline-regulated transient expression systems, we determined the functional properties of mutant TNFRI, both individually and in combination with WT TNFRI. Our findings show for the first time that both T50K and P46L variants participate in self-association with the WT TNFRI subunits. Furthermore, in contrast to WT TNFRI, the T50K variant exhibits enhanced NF-κB activation in the absence of TNF.

MATERIALS AND METHODS

Antibodies and reagents.

Anti-human TNFRI/TNFRSF1A monoclonal antibody (MAB 225), biotinylated anti-human TNFRI polyclonal antibody (BAF 225), anti-human TNFRII/TNFRSF1B monoclonal antibody (MAB 726), and biotinylated anti-human TNFRII/TNFRSF1B polyclonal antibody (BAF 726) were purchased from R&D Systems (Abingdon, UK). Horseradish peroxidase (HRP)–conjugated F(ab′)2 goat anti-mouse IgG was obtained from Serotec (Kidlington, UK). Anti-V5 antibody and HRP-conjugated anti-V5 antibody were purchased from Invitrogen (Paisley, UK), and antihemagglutinin (anti-HA) antibody and peroxidase-conjugated anti-HA antibody (clone 12CA5) were from Roche Diagnostics (Lewes, UK). Anti–NF-κB1 (p50) (NLS, sc-114X) and anti–p65 (RelA) (sc-372X) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Recombinant human TNFα was purchased from ImmunoKontact (AMS Biotechnology [Europe], Oxon, UK). All protease inhibitors were purchased from Calbiochem (Merck Biosciences, Nottingham, UK).

TNFRI cloning and expression constructs.

Total RNA was prepared, using the guanidinium thiocyanate method (18), from Epstein-Barr virus–transformed cells that had been established previously from healthy controls and patients with TRAPS (7). Informed consent was approved by the East London and City Health Authority Research Ethics Committee. Following reverse transcription, the full-length coding region of TNFRI was amplified by polymerase chain reaction (PCR) using the sense primer 5′-CGCGGATCCACCATGGGCCTCTCCACCGTGCC-3′ containing a Bam HI restriction site at the 5′ end, and using the antisense primer 5′-TCTGAGAAGACTGGGCGCGGG-3′ (GenBank accession no. M58286). DNA was amplified with Pfu DNA polymerase (Promega, Madison, WI). PCR products were digested with Bam HI, and the resulting fragments were cloned in-frame into the Bam HI and Eco RV cloning sites of the pcDNA6/V5-HisB vector (Invitrogen) to create tags at the C-terminus of the receptors (Figure 1). The resulting plasmids were analyzed by restriction enzyme digestion and by DNA sequencing of both strands.

Figure 1.

Schematic representation of constructs. Both wild-type (WT) and mutant tumor necrosis factor receptor I (TNFRI) cloned into pcDNA6/V5-HisB and pGT vectors expressed V5-His tags. Only WT TNFRI cloned into the pcDNA6B vector carried a hemagglutinin (HA) tag (see Materials and Methods). CMV = immediate-early enhancer/promoter of cytomegalovirus; Ptet = tetracycline-responsive promoter; SV40 = early/late promoter of SV40; rtTA-2SM2 = reverse tetracycline transactivator; IRES = internal ribosome entry site; tetR-KRAB = tetracycline repressor. Upright triangle indicates bovine growth hormone poly(A) signal. Inverted triangle indicates SV40 poly(A) signal.

To generate HA-tagged WT TNFRI, a linker containing the HA coding sequence and a stop codon with a 5′ Bst XI end and a 3′ Xba I end (sense 5′-GTGGCTATCCATATGACGTCCCAGACTATGCTGGCTGAT-3′, antisense 5′-CTAGATCAGCCAGCATAGTCTGGGACGTCATATGGATAGCCACTGTG-3′) was inserted in-frame between the Bst XI and Xba I sites of the WT-TNFRI/pcDNA6/V5-HisB construct (Figure 1). This HA tag insertion resulted in the elimination of V5-His expression. For the tetracycline-regulated gene expression system, a Pme I fragment containing the full-length tagged TNFRI complementary DNA was excised from the TNFRI/pcDNA6/V5-HisB plasmid and inserted into the Eco RV cloning site of the pGT vector that contains the tetracycline-regulated promoter (19), generating the TNFRI/pGT-V5-His construct (Figure 1).

Construction of pMIK/ZeoSV2.

The tetracycline repressor (tetR-KRAB) cassette was removed from pCMV-tetR(B/E)-KRAB (20) by restriction with Eco RI and Bam HI and inserted into the Eco RI and Bam HI cloning sites of pZeoSV2(−) (Invitrogen) to produce pK/ZeoSV2. A linker containing a 5′ Nhe I end, internal Mfe I, Sma I, Eco RV, and Eco RI sites, and an Mfe I end was cloned between the Nhe I and Eco RI sites of pK/ZeoSV2, generating pLK/ZeoSV2. The encephalomyocarditis virus internal ribosome entry site was removed from pCITE (Novagen, Merck Biosciences, Nottingham, UK) by restriction with Eco RI and Bal I and cloned into pLK/ZeoSV2 digested with Xba I, end-blunted with Klenow (Roche Diagnostics, East Sussex, UK), and then cut with Eco RI to create the pLIK/ZeoSV2 construct. Finally, the fragment encoding the reverse tetracycline transactivator (rtTA-2SM2) was removed from pUHDrtTA-2SM2 (21) by restriction with Bam HI (end-blunted) and Eco RI and inserted into the Mfe I and Eco RV sites in pLIK/ZeoSV2 to create the pMIK/ZeoSV2 construct (Figure 1).

Transfections and Western blotting analysis.

We maintained 293T cells in Dulbecco's modified Eagle's medium (Cambrex, Wokingham, UK) containing 10% fetal bovine serum, 100 units/ml penicillin, 100 μg streptomycin, and 2 mML-glutamine. Transient transfections were performed using the calcium phosphate precipitation method described previously (19). Transfected 293T cells were subjected to an osmotic shock, after which fresh medium was added and cells were allowed to recover for 20–24 hours before stimulating with TNF (10 ng/ml) for 5–15 minutes. For regulated gene expression, the cells were cultured in medium with or without doxycycline (Sigma-Aldrich, Poole, UK). Unstimulated or ligand-stimulated cells were scraped, washed in ice-cold phosphate buffered saline (PBS), and lysed in appropriate buffer for 15 minutes on ice. Cellular debris was removed by centrifugation for 15 minutes at 4°C, and lysates were stored at −70°C until analyzed.

To assess the effects of induced expression of V5-His–tagged TNFRI on the constitutive expression of HA-tagged WT TNFRI, 5 × 105 cells were cotransfected with 2 μg each of WT-TNFRI/pcDNA6B-HA, WT- or mutant-TNFRI/pGT-V5-His, and pMIK/ZeoSV2 plasmid DNA. Following the addition of doxycycline (0.0001–1 μg/ml), the cells were lysed in radioimmunoprecipitation assay buffer (150 mM NaCl, 50 mM Tris HCl [pH 7.5], 2 mM EDTA, 1% Nonidet P40 [NP40], 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], and 2 mM Na3VO4 supplemented with the protease inhibitors aprotinin [10 μg/ml], leupeptin [10 μg/ml], and 4-[2-aminoethyl]benzenesulfonylfluoride, HCl [100 μM]). Equal amounts of lysates were subjected to 10% SDS–polyacrylamide gel electrophoresis (PAGE) under reducing conditions and analyzed by Western blotting.

Immunoprecipitation and immunoblotting.

After cell lysates were precleared with protein G–Sepharose (Amersham Pharmacia Biotech, Buckinghamshire, UK), they were mixed either with 2–3 μg of specific antibodies and 50 μl of protein G–Sepharose slurry (50% in PBS) or with 50 μl of Ni-NTA agarose (Qiagen, Crawley, UK) in the presence of 10 mM imidazole (Sigma), and samples were rotated for 16 hours at 4°C. Bead-bound complexes were washed extensively with appropriate lysis buffer and then boiled in sample loading buffer containing SDS. For Ni bead–bound complexes, the SDS–sample buffer contained 500 mM imidazole. Immune complexes were separated by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech), and then blotted with appropriate primary and secondary antibodies. When biotinylated primary antibodies were used, the membranes were incubated with streptavidin–biotinylated horseradish complex (Amersham Pharmacia Biotech) before development with an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).

Cell surface TNFRI crosslinking.

We transfected 293T cells (2 × 106) with 4 μg each of WT-TNFRI/pcDNA6B-HA, WT- or mutant-TNFRI/pcDNA6/V5-HisB, WT- or mutant-TNFRI/pGT-V5-His, and pMIK/ZeoSV2 plasmid DNA, as appropriate, using pcDNA6/V5-HisB vector to compensate for the total amount of DNA transfected. The procedure used for crosslinking cell surface proteins was essentially as described by Chan et al (11). When necessary, the cells were first treated with TNF (100 ng/ml) for 1 hour before the addition of the chemical crosslinker 3,3′-dithiobis(sulfosuccinimidylpropionate) (DTSSP; Pierce, Cheshire, UK) to a final concentration of 2 mM. Following lysis (11), cellular debris was removed by centrifugation. For immunoprecipitation, precleared cell lysates were mixed either with 2 μg of anti-human TNFRI monoclonal antibody (MAB 225) and 50 μl protein G–Sepharose or with 50 μl of Ni-NTA agarose in the presence of 10 mM imidazole and 8M urea. Immune complexes were resolved by 8% or 10% SDS-PAGE under nonreducing (without β-mercaptoethanol) or reducing (with 500 mM β-mercaptoethanol) conditions and analyzed using anti–tag epitope antibodies.

Preparation of cytosolic and nuclear extracts.

Cells (2–4 × 106) were scraped in PBS, washed once, and then resuspended in 200 μl of low-salt buffer (10 mM HEPES [pH 7.9], 10 mM KCl, 2 mM MgCl2, 0.1 mM EDTA, 10 mM NaF, 10 mM Na4P2O7, 2 mM Na3VO4, and protease inhibitors). After 15 minutes of incubation on ice, NP40 was added to a final concentration of 0.6%, and the suspension was vortexed briefly and incubated on ice for an additional 10 minutes. The whole cell lysates were centrifuged at 14,000 revolutions per minute for 10 minutes at 4°C, and the supernatants were saved as cytosolic extracts. The pelleted nuclei were resuspended in 50–100 μl of high-salt buffer (50 mM HEPES [pH 7.9], 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 20% glycerol, 1 mM dithiothreitol [DTT], and protease inhibitors), and the samples were rotated for 40 minutes at 4°C. Following centrifugation at 14,000 rpm for 10 minutes at 4°C, the resulting supernatants were used as nuclear extracts. Protein concentrations were determined using a bicinchoninic acid protein assay reagent kit (Pierce).

Oligonucleotides and electrophoretic mobility shift assay (EMSA).

DNA oligonucleotides (Invitrogen) containing an NF-κB binding site from the murine κ–light chain enhancer were used as a probe (sense 5′-GGAGTTGAGGGGACTTTCCCAGGC-3′, antisense 5′-GCCTGGGAAAGTCCCCTCAACT-3′). The oligonucleotides for AP-1 were 5′-GGCGCTTGATGAGTCAGCCGGAA-3′ (sense) and 5′-TTCCGG- CTGACTCATCAAGCG-3′ (antisense). Double-stranded annealed oligonucleotides were radiolabeled with α32P-dCTP (Amersham Pharmacia Biotech) using Klenow enzyme. Nuclear extracts (1 μg total protein) were mixed with 2 μl of dilution buffer (20 mM HEPES [pH 7.9], 60 mM KCl, 0.25 mM EDTA, 0.125 mM EGTA, 20% glycerol, 1 mM DTT, and protease inhibitors) in a final volume of 7 μl.

For competition assays, unlabeled double-stranded oligonucleotides were added to appropriate reaction mixtures and incubated for 20 minutes at room temperature prior to the addition of 7 μl of the binding reaction mixture. The binding reaction mixture was prepared using 10× binding buffer (100 mM Tris HCl [pH 7.9], 500 mM NaCl, 25 mM MgCl2, 10 mM EDTA, 10 mM DTT, 50 mM spermidine, and 40% glycerol), poly(dI-dC) (0.5 μg/reaction), and radiolabeled probe (0.5 ng/reaction). For supershift analyses, antibodies (1–2 μg) were added 20 minutes after the addition of binding reaction mixture containing the labeled DNA probe, and mixtures were incubated for an additional 25 minutes at 4°C. The reaction mixtures were then subjected to electrophoresis in 6% polyacrylamide gels (Accugel 29:1 sequencing grade; National Diagnostics, Hessle, UK) prepared in 0.25× TBE buffer (22.5 mM Tris–borate, 0.5 mM EDTA [pH 8.0]). Following electrophoresis in 0.25× TBE buffer, the gels were dried and exposed to x-ray films (Hyperfilm; Amersham Pharmacia Biotech) at −70°C for 8–48 hours.

RESULTS

Participation of WT TNFRI and mutant TNFRI subunits in PLAD-mediated self-association.

To examine PLAD-dependent self-assembly of WT or mutant TNFRI, we used a thiol-cleavable, membrane-nonpermeable chemical crosslinker, DTSSP, for crosslinking cell surface proteins constitutively expressed from 293T cells transfected with V5 epitope–tagged TNFRI. Following immunoprecipitation of cell lysates with an anti-TNFRI monoclonal antibody and blotting with an anti-V5 antibody, complexes with discrete molecular size bands migrating in the ranges of 100–150 kd and 150–250 kd were observed for WT and P46L TNFRI and, to a lesser extent, for T50K TNFRI (Figure 2A). Since the monomers were clearly seen, the complexes migrating between 100 kd and 150 kd may have reflected receptor dimers (Figure 2A). The molecular sizes of larger complexes (>150 kd) correlated well with glycosylated and nonglycosylated TNFRI trimers (Figure 2A). Furthermore, the presence of TNF did not affect the profile of either monomers or large complexes for the WT or mutant TNFRI receptors.

Figure 2.

Pre–ligand binding assembly domain (PLAD)–mediated self-association of mutant TNFRI with WT TNFRI subunits. We transiently transfected 293T cells with WT-TNFRI/pcDNA6/V5-HisB (WT + Vect), T50K-TNFRI/pcDNA6/V5-HisB (T50K + Vect), and P46L-TNFRI/pcDNA6/V5-HisB (P46L + Vect). Empty pcDNA6/V5-HisB vector was used as a control (Vect) and also to compensate for the total amount of DNA in each transfection. Nontransfected (Unt) or transfected cells were harvested and incubated with or without TNF, followed by treatment with 3,3′-dithiobis(sulfosuccinimidylpropionate) (DTSSP), after which cell lysates were immunoprecipitated (IP) either with an anti-human TNFRI monoclonal antibody or with Ni-agarose beads. Positions of molecular weight markers are shown. A, Self-assembly of trimeric TNFRI by PLAD. Anti-human TNFRI antibody–precipitated complexes were subjected to 8% nonreducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and blotting with horseradish peroxidase (HRP)–conjugated anti-V5 antibody. Positions of TNFRI complexes are indicated by arrows. B, Heterotrimer formation by T50K TNFRI with WT TNFRI subunits. We cotransfected 293T cells with WT-TNFRI/pcDNA6/V5-HisB + WT-TNFRI/pcDNA6B-HA (WT + WT), T50K-TNFRI/pcDNA6/V5-HisB + WT-TNFRI/pcDNA6B-HA (T50K + WT), or pcDNA6/V5-HisB vector alone (Vect) for constitutive expression. Harvested cells were incubated with or without TNF followed by treatment with DTSSP, and then cell lysates were immunoprecipitated with Ni-agarose beads. Bound complexes were resolved by 10% SDS-PAGE under reducing conditions and then blotted either with anti-V5 (upper panel) or with anti-HA (lower panel) primary antibodies followed by an HRP-conjugated anti-mouse secondary antibody. Positions of monomers are indicated by arrows. See Figure 1 for other definitions.

We also noted that the levels of monomers and larger complexes for V5-His–tagged T50K TNFRI were reduced compared with those for V5-His–tagged WT TNFRI or V5-His–tagged P46L mutant TNFRI (Figure 2A). This difference might have been due to inefficient binding of the T50K mutant by the anti-TNFRI monoclonal antibody used for immunoprecipitation. This possibility was examined further by immunoprecipitating the cell lysates containing V5-His–tagged proteins with Ni-agarose beads and blotting them with anti-V5 antibody. Similar levels of monomers and larger complexes were observed for both WT and T50K TNFRI (results not shown), suggesting that the T50K mutant does form complexes in the absence of TNF.

We next investigated whether mutant TNFRI subunits associated with WT TNFRI subunits. In coexpression experiments, cell lysates from 293T cells constitutively overexpressing HA-tagged WT TNFRI together with either WT-TNFRI-V5-His or T50K-TNFRI-V5-His were analyzed by Ni-agarose bead immunoprecipitation and SDS-PAGE under reducing conditions. Western blotting analysis with anti-tag antibodies showed that V5-His–tagged T50K TNFRI and V5-His–tagged WT TNFRI associated equally well with HA-tagged WT TNFRI (Figure 2B). In order to exclude the possibility that PLAD-dependent self-associated TNFRI (Figure 2B) crosslinked with TNFRII, we performed further coexpression experiments with both types of TNF receptors. The 293T cells constitutively overexpressing untagged full-length human TNFRII together with either WT-TNFRI-V5-His or T50K-TNFRI-V5-His were either left untreated or treated with TNF, and then their surface proteins were crosslinked with DTSSP, as described in Figure 2B. The results showed that V5-His–tagged WT TNFRI or T50K TNFRI did not crosslink with constitutively expressed TNFRII, which self-associated independently of TNFRI in these coexpression experiments (results not shown).

Induced expression of WT and mutant receptors exerts differential down-regulatory effects on constitutively expressed WT TNFRI.

We consistently observed in whole cytosolic extracts that constitutively expressed HA-tagged WT TNFRI was down-regulated when the expression of V5-His epitope–tagged WT TNFRI or mutant TNFRI was maximally induced by 1 μg/ml doxycycline (results not shown). This was investigated further in coexpression experiments by gradually increasing the expression of V5-His–tagged WT, T50K, or P46L TNFRI and assessing the effects on constitutively expressed HA-tagged WT TNFRI (Figure 3).

Figure 3.

Induced expression of WT and mutant TNFRI down-regulates constitutively expressed WT TNFRI. We cotransfected 293T cells with WT-TNFRI/pGT-V5-His + WT-TNFRI/pcDNA6B-HA (WT + WT) (A), T50K-TNFRI/pGT-V5-His + WT-TNFRI/pcDNA6B-HA (T50K + WT) (B), P46L-TNFRI/pGT-V5-His + WT-TNFRI/pcDNA6B-HA (P46L + WT) (C), and WT-TNFRI/pcDNA6B-HA + pGT empty vector (WT), with pMIK/ZeoSV2 plasmid DNA included in all transfections. Receptor expression was induced by the addition of doxycycline (Dox) at 0.0001 μg/ml (lane 5), 0.001 μg/ml (lane 6), 0.01 μg/ml (lane 7), 0.1 μg/ml (lane 8), and 1 μg/ml (lanes 3 and 9). Cell lysates from nontransfected (Unt) and transfected cells were analyzed by Western blotting using anti-HA (upper panels) and anti-V5 (lower panels) antibodies. Lanes 2 and 3 show the constitutive expression of WT-TNFRI/pcDNA6B-HA together with pGT vector in the absence (lane 2) or presence (lane 3) of doxycycline (1 μg/ml). The band intensities of TNFRI in lanes 4–9 (AC) for each combination shown at the right (DF), respectively, were determined with Adobe Photoshop 6.0, using a relative position in lane 1 as background. The levels of induced expression of V5-His–tagged WT, T50K, and P46L TNFRI with 1 μg/ml doxycycline (lane 9) and of constitutive expression of HA-tagged WT TNFRI without doxycycline (lane 4) were taken as 100%. The higher expression of T50K (B and E) in lane 8 (0.1 μg/ml doxycycline) than in lane 9 (1 μg/ml doxycycline) was not consistent between different experiments. Values in DF are the mean of 1 experiment and are representative of 3 independent experiments. WT (V5) = V5-His–tagged WT TNFRI; WT (HA) = HA-tagged WT TNFRI; T50K (V5) = V5-His–tagged T50K TNFRI; P46L (V5) = V5-His–tagged P46L TNFRI (see Figure 1 for other definitions).

The expression of V5-His–tagged WT TNFRI (Figure 3A), T50K TNFRI (Figure 3B), or P46L TNFRI (Figure 3C) was detectable at a doxycycline concentration of 0.01 μg/ml (lane 7) and was elevated further by increasing doses of doxycycline (Figures 3A–C, lanes 8 and 9). The constitutive expression of HA-tagged WT TNFRI in the presence of the pGT empty vector together with pMIK/ZeoSV2 (Figures 3A–C, lane 2) remained unaffected by the maximum concentration of doxycycline (Figures 3A–C, lane 3), suggesting that the observed inhibitory effects on HA-tagged WT TNFRI were not due to the presence of vector-related elements or doxycycline alone. Interestingly, the inhibitory effects on constitutively expressed HA-tagged WT TNFRI were highest with induced V5-His–tagged WT TNFRI: a 50% inhibition occurred at a doxycycline concentration of <0.1 μg/ml (Figure 3D). In contrast, induced V5-His–tagged T50K mutant caused the least inhibition of HA-tagged WT TNFRI expression, since a 50% inhibition was seen with a doxycycline concentration of >0.1 μg/ml (Figure 3E), while the inhibitory effects of P46L were intermediate (Figure 3F).

A somewhat higher expression level of T50K TNFRI (Figures 3B and E) at a doxycycline concentration of 0.1 μg/ml (lane 8) compared with that seen at a higher doxycycline concentration (1 μg/ml, lane 9) was probably due to small variations in transient transfections of 293T cells, since this effect was not consistent between different experiments or different clones of T50K constructs. We also observed that both V5-His–tagged and HA-tagged TNFRI, whether expressed constitutively or induced by doxycycline, exhibited 2 bands, a feature which may reflect posttranslational differential glycosylation events. Furthermore, slight variations seen in the migration patterns of the doublet bands for V5-His–tagged TNFRI and HA-tagged TNFRI could have been due to structural differences between the 2 tags.

To investigate whether the down-regulation of HA-tagged WT TNFRI expression observed in whole cytosolic lysates (Figure 3) also occurred at the cell surface, we performed further coexpression experiments with constitutively expressed HA-tagged WT TNFRI together with induced expression of either V5-His–tagged WT TNFRI or V5-His–tagged T50K TNFRI, and then we analyzed DTSSP-crosslinked complexes by SDS-PAGE and Western blotting (Figure 4). Little expression of HA-tagged WT TNFRI was detectable when the level of V5-His–tagged WT TNFRI expression was increased with doxycycline concentrations of 0.1–1 μg/ml (Figure 4A, lanes 8 and 9). Furthermore, immunoprecipitation with anti-TNFRI monoclonal antibody and blotting with anti-HA antibody revealed that HA-tagged WT TNFRI was expressed before the induction of V5-His–tagged WT TNFRI (Figure 4B, upper panel, lanes 5–7). Western blotting with a polyclonal anti-TNFRI antibody revealed total cellular TNF receptor expression (Figure 4B, lower panel, lanes 5–9). These results strongly suggest that the expression of HA-tagged WT TNFRI on the cell surface was reduced following up-regulation of V5-His–tagged WT TNFRI, a finding that correlates well with the inhibitory effects of induced WT TNFRI on constitutively expressed HA-tagged WT TNFRI in whole cytosolic extracts (Figure 3A). However, as shown in Figure 4C, T50K expression induced by doxycycline concentrations of 0.1–1 μg/ml (upper panel, lanes 8 and 9) also reduced the surface expression of HA-tagged WT TNFRI (lower panel, lanes 8 and 9).

Figure 4.

Competition of receptor expression on the cell surface. We cotransfected 293T cells with WT-TNFRI/pGT-V5-His + WT-TNFRI/pcDNA6B-HA (WT + WT) (A and B), T50K-TNFRI/pGT-V5-His + WT-TNFRI/pcDNA6B-HA (T50K + WT) (C), or empty vectors alone (Vect), with pMIK/ZeoSV2 plasmid DNA included in all transfections. Receptor expression was induced by the addition of doxycycline (Dox) at 0.001 μg/ml (lane 6), 0.01 μg/ml (lane 7), 0.1 μg/ml (lane 8), and 1 μg/ml (lanes 2, 4, and 9), and 22 hours later nontransfected (Unt) and transfected cells were harvested and treated with 3,3′-dithiobis(sulfosuccinimidylpropionate). Controls (lanes 1, 3, and 5) were without doxycycline. Cell lysates were immunoprecipitated (IP) either with Ni-agarose beads (A and C) or with anti-TNFRI monoclonal antibody (B). Bound immune complexes were analyzed by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis under reducing conditions, followed by Western blotting using anti-V5 (upper panels) and anti-HA (lower panels) antibodies (A and C) or using peroxidase-conjugated anti-HA (upper panel) and biotinylated anti-TNFRI (lower panel) antibodies (B). Specific bands in each panel are indicated by arrows. The earliest induction of V5-His–tagged WT TNFRI expression (A, upper panel, lane 7) increased further with higher concentrations of doxycycline (lanes 8 and 9), while the expression of HA-tagged WT TNFRI decreased (A, lower panel, lanes 8 and 9). Total cellular TNFRI expression is shown in B (lower panel, lanes 5–9), and the absence of HA-tagged WT TNFRI is clearly seen in lanes 8 and 9 (B, upper panel). See Figure 1 for other definitions.

T50K TNFRI expression induces NF-κB activation independently of TNF.

We next used EMSA to examine the activation of NF-κB after overexpressing WT TNFRI and T50K mutant in 293T cells. The profile seen for the empty vector control (Figure 5A, lane 2) did not differ from that for untreated 293T cells (results not shown). Interestingly, constitutively expressed T50K exhibited a strong extra NF-κB DNA binding complex band (Figure 5A, lane 10), whereas only a faint band was visible at the same position with the WT (Figure 5A, lane 6). This NF-κB DNA binding complex was TNF receptor dependent, since untreated 293T cells or vector-transfected cells produced similar NF-κB DNA binding complexes when stimulated by TNF (10 ng/ml) (results not shown).

Figure 5.

T50K TNFRI expression activates NF-κB independently of TNF. A, Constitutive expression of TNFRI. We transiently transfected 293T cells with WT-TNFRI/pcDNA6/V5-HisB (WT + Vect) and T50K-TNFRI/pcDNA6/V5-HisB (T50K + Vect). Empty pcDNA6/V5-HisB vector was used as a control (Vect) and also to compensate for the total amount of DNA in each transfection. Nuclear extracts were prepared 40–42 hours after transfections and analyzed for DNA binding activity by electrophoretic mobility shift assay (EMSA). Specificity of NF-κB complexes was determined by incubating the reactions with 10-fold and 50-fold excess unlabeled NF-κB and activator protein 1 (AP-1) oligonucleotides (cold oligo) prior to the addition of radiolabeled NF-κB probe. Lane 1 (probe) is a negative control without nuclear extract. Positions of NF-κB DNA complexes are indicated by arrows. T50K-induced NF-κB DNA complex is indicated by an arrowhead. B, Induced expression of TNFRI. We cotransfected 293T cells with WT-TNFRI/pGT-V5-His + pMIK/ZeoSV2 (WT), T50K-TNFRI/pGT-V5-His + pMIK/ZeoSV2 (T50K), or pGT empty vector + pMIK/ZeoSV2 (Vect) plasmid DNA. Receptor expression was induced by the addition of doxycycline (Dox) at 0.001 μg/ml (lanes 5 and 10), 0.01 μg/ml (lanes 6 and 11), 0.1 μg/ml (lanes 7 and 12), and 1 μg/ml (lanes 3, 8, and 13). NF-κB DNA binding activity in nuclear extracts was determined by EMSA. The first lane (probe), arrows, and arrowhead are as described for A. C and D, T50K-induced NF-κB complexes contain p65 (RelA). C, Constitutive expression. Nuclear extracts shown in A for WT + Vect (here termed WT) and for T50K + Vect (here termed T50K) were analyzed after performing the reactions in the absence or presence of anti–NF-κB1 (p50) or anti–p65 (RelA) antibodies. The results shown were obtained from the same EMSA gel. D, Induced expression. Nuclear extracts for induced T50K expression (at 0.1 μg/ml doxycycline) were analyzed by electrophoretic mobility supershift assay as described for C. In C and D, the positions of NF-κB DNA binding complexes are indicated by arrows, and the position of the T50K-induced NF-κB complex is indicated by an arrowhead. See Figure 1 for other definitions.

The specificity of either basal (Figure 5A, lane 2) or T50K-induced (Figure 5A, lane 10) NF-κB DNA binding complexes was confirmed by displacement of NF-κB DNA binding complexes in the presence of 10-fold and 50-fold excess of unlabeled NF-κB DNA oligonucleotides (Figure 5A, lanes 3, 4, 11, and 12). T50K-induced NF-κB DNA complex was completely displaced by a 10-fold excess of cold NF-κB DNA oligonucleotides (Figure 5A, lane 11), whereas a 50-fold excess of unlabeled AP-1 DNA oligonucleotides had no effect on NF-κB DNA binding activities (Figure 5A, lane 13). Furthermore, the TNF-stimulated NF-κB DNA complex observed in either vector-transfected cells or nontransfected 293T cells was similarly displaced by a 10-fold excess of unlabeled NF-κB DNA oligonucleotides but not by a 50-fold excess of cold AP-1 DNA oligonucleotides in co-competition assays (results not shown).

In some experiments we observed a faint NF-κB DNA complex in constitutively expressed WT TNFRI (Figure 5A, lane 6), and in order to exclude the possibility that T50K-induced NF-κB activation was not simply due to receptor overexpression, we performed further experiments in which the expression of WT TNFRI or T50K TNFRI was increased by increasing concentrations of doxycycline. In contrast to WT TNFRI, which did not induce any detectable NF-κB DNA binding complexes (Figure 5B, lanes 4–8), T50K-induced NF-κB DNA binding activity (Figure 5B, lanes 9–13) was clearly seen at a doxycycline concentration of 0.01 μg/ml (Figure 5B, lane 11). Since the expression of both WT TNFRI and the T50K mutant first appeared when the doxycycline concentration was 0.01 μg/ml (Figures 3A and B), these results demonstrate that the T50K mutant induces intrinsic NF-κB DNA binding activity at lower expression levels.

The composition of NF-κB DNA binding complexes induced by T50K expression was determined using antibodies reactive with human p50 and p65 (RelA) subunits of NF-κB proteins. Addition of anti–p65 (RelA) to the binding reaction mixture resulted in abolition of the constitutively expressed T50K-associated NF-κB DNA binding complex (Figure 5C). Similar results were seen when the expression of T50K was induced by doxycycline (0.1 μg/ml) (Figure 5D). Anti–NF-κB1 (p50) was consistently less effective in these assays (Figures 5C and D). The NF-κB DNA binding complex observed following TNF stimulation of nontransfected 293T cells was also abolished by anti–p65 (RelA), while anti–NF-κB1 (p50) was less effective (results not shown). These findings suggest that, in contrast to WT TNFRI expression, T50K TNFRI expression activates NF-κB containing the p65 (RelA) subunit in the absence of TNF.

DISCUSSION

The results presented here show that TNFRI variants with mutations within the CRD1 are capable of participating in PLAD-dependent self-association independently of ligand. The individual variant receptor subunits not only form self-assembled complexes, but can also form structures containing WT TNFRI subunits. Our data suggest that the T50K variant is functionally active and initiates the NF-κB activation pathway in the absence of TNF. To our knowledge, this is the first study demonstrating activation of NF-κB by a TNFRI variant associated with an inflammatory disorder such as TRAPS. It remains to be established whether this feature is common to all TRAPS mutations.

An analysis of crosslinked surface proteins under nonreducing conditions (Figure 2A) suggested that both WT and variant TNFRI formed dimers and trimers in the absence of exogenous TNF. Although trimeric forms of both TNFRI and TNFRII on the cell surface are more appropriate structural features that are required for biologic functions, which may also lead to greater stability of the trimeric form of the receptor on the cell surface (11), dimers of the extracellular region of TNFRI have been reported in crystallography studies (13). In addition, the concomitant expression of TNFRII did not influence the PLAD-dependent self-association of either WT TNFRI or the T50K variant in coexpression experiments (results not shown). Up-regulation of V5-His–tagged WT or variant TNFRI expression by increasing doxycycline concentration resulted in concomitant down-regulation of constitutively expressed HA-tagged WT TNFRI when whole cytosolic extracts were analyzed (Figure 3). In addition, the induced expression of WT TNFRI exhibited slightly greater inhibitory effects on constitutively expressed HA-tagged WT TNFRI than did induced expression of the T50K variant (Figures 3A, B, D, and E), while the effect of P46L was intermediate (Figures 3C and F).

Interestingly, these observations correlate well with the TRAPS phenotype, in which P46L is less penetrant than the T50K variant. It is noteworthy that both WT and T50K TNFRI also down-regulated the constitutive expression of HA-tagged WT TNFRI on the cell surface (Figure 4). The mechanisms by which cell surface expression of TNFRI may be controlled still remain unclear. Although TNFRI has been shown to accumulate predominantly in the trans-Golgi apparatus within the cell, this localization is dependent upon motifs within the cytoplasmic region of the receptor (22–25). The TNFRI receptors used in this study were full length with tags at their carboxyl termini, since the presence of His tags at the amino terminus of TNFR may interfere with its PLAD-mediated self-assembly (26) and ligand binding. Taken together, our results suggest that TNFRI expression on the cell surface may be regulated independently of its total cellular expression, and this regulation may be mediated by an additional, as-yet-undefined cellular component.

In recent studies, Huggins et al (16) and Todd et al (15) analyzed HEK 293 cells stably transfected with truncated WT TNFRI and TRAPS-associated TNFRI variants lacking the cytoplasmic regions and showed that the surface expression of the T50M variant was similar to that observed for the WT receptor. Moreover, the full-length T50M variant was also expressed on the cell surface, although at a lower level than that of the WT TNFRI. However, T50M expressed on the cell surface showed much lower TNF binding capacity compared with WT TNFRI. In our cell surface crosslinking experiments, T50K was poorly detected compared with either WT TNFRI or the P46L variant when anti-TNFRI monoclonal antibody was used for immunoprecipitation (Figure 2A), and this was probably due to inefficient antibody binding. Peripheral blood leukocytes from TRAPS patients with T50M and T50K variants showed defective TNFRI shedding (3, 17). However, it has recently been reported that variations in TNFRI shedding are not only related to the structural mutations in TNFRI, but may also depend on other cellular factors (16).

Our study reveals the intrinsic ability of the T50K variant to activate NF-κB in the absence of TNF. This acquired capacity of T50K is not related to high overexpression levels (Figures 5A and B), since the WT TNFRI expression similarly induced by doxycycline did not lead to NF-κB activation (Figure 5B). The expression of WT TNFRI or of the variants was first detected at a doxycycline concentration of 0.01 μg/ml (Figures 3A and B). The T50K-induced NF-κB DNA binding complex appearing below the endogenous slowest migrating band (Figure 5) consisted primarily of the p65 (RelA) subunit that was completely abolished in the presence of anti–p65 (RelA), and little p50 was detectable in the same complex (Figures 5C and D). The ability of T50K TNFRI to activate NF-κB in the absence of ligand suggests that this variant is functionally active, and this finding may have important clinical consequences. The fact that this mutant protein also has a shedding defect implies that its concentration on the cell surface will probably be higher than that of the WT receptor that can be shed normally.

TRAPS presents clinically with self-limited inflammatory episodes (27). Therefore, in TRAPS patients, a proinflammatory event triggered by an environmental factor (such as infection [28,29], cancer [30,31], and drug toxicity [32]) that can up-regulate TNFRI expression will have an overall proinflammatory and antiapoptotic effect, because in the case of the T50K variant, which is capable of sustaining inappropriate NF-κB activation, this will result in persistent production of proinflammatory mediators including TNF itself. A similar proinflammatory outcome has been shown in a mouse model of TRAPS (33). Importantly, up-regulation of TNFRI in the CNS is well documented and has been linked with ischemic preconditioning (34), viral infection (35, 36), and peripheral nerve injury (37). This mechanism could also operate in TRAPS patients with neurologic involvement, including our patients with the T50K variant (17). Given the complexities of the range of TRAPS mutations associated with different structural elements of TNFRI, changes in signaling events could alter the subsequent activation profiles of functional mediators important in modulating not only TNFRI expression but also signaling pathways mediated by this receptor.

Acknowledgements

We thank the TRAPS patients, their referring physicians (Dr. K. Minden [Berlin, Germany] and Dr. P. Arkwright [Manchester, UK]) for providing us with material to carry out this research, and Dr. F. Flores-Borjas for help with fluorescence-activated cell sorter analysis.

Ancillary