Transglutaminase 2 Induces Nuclear Factor-κB Activation via a Novel Pathway in BV-2 Microglia*

Transglutaminase 2 (TGase 2) expression is increased in inflammatory diseases. We demonstrated previously that inhibitors of TGase 2 reduce nitric oxide (NO) generation in a lipopolysaccharide (LPS)-treated microglial cell line. However, the precise mechanism by which TGase 2 promotes inflammation remains unclear. We found that TGase 2 activates the transcriptional activator nuclear factor (NF)-κB and thereby enhances LPS-induced expression of inducible nitric-oxide synthase. TGase 2 activates NF-κB via a novel pathway. Rather than stimulating phosphorylation and degradation of the inhibitory subunit α of NF-κB (I-κBα), TGase2 induces its polymerization. This polymerization results in dissociation of NF-κB and its translocation to the nucleus, where it is capable of up-regulating a host of inflammatory genes, including inducible nitric-oxide synthase and tumor necrosis factor α (TNF-α). Indeed, TGase inhibitors prevent depletion of monomeric I-κBα in the cytosol of cells overexpressing TGase 2. In an LPS-induced rat brain injury model, TGase inhibitors significantly reduced TNF-α synthesis. The findings are consistent with a model in which LPS-induced NF-κB activation is the result of phosphorylation of I-κBα by I-κB kinase as well as I-κBα polymerization by TGase 2. Safe and stable TGase2 inhibitors may be effective agents in diseases associated with inflammation.

Transglutaminase 2 (TGase 2; E.C. 2.3.2.13, protein-glutamine ␥-glutamyltransferase) 1 belongs to a family of Ca 2ϩ -dependent enzymes that catalyzes N ⑀ -(␥-L-glutamyl)-L-lysine isopeptide bond formation between peptide bound lysine and glutamine residues (1). N ⑀ -(␥-L-Glutamyl)-L-lysine cross-linking stabilizes intra-and extracellular proteins as macromolecular assemblies that are used for a variety of essential physiological purposes, such as barrier function in epithelia, apoptosis, and extracellular matrix formation (2). TGase 2 is normally expressed at low levels in many different tissues and inappropriately activated in a variety of pathological conditions (3). However, the role TGase 2 specifically plays in disease etiology is unclear.
Autoimmune diseases are strongly associated with aberrant activation of macrophage and T cells, which cause serious inflammation. Abnormal increase of TGase 2 expression in autoimmune inflammatory myopathies and celiac disease has been reported (4 -6). TGase 2 is increased in autoimmune diseases as a result of macrophage activation (7)(8)(9). The increase of TGase 2 expression seems to be closely associated with autoantibody formation. Autoantibodies against TGase 2 are present in the blood of patients with such autoimmune diseases as celiac disease (10), dermatitis herpetiformis (11), type 1 diabetes (12), lupus (13), and rheumatoid arthritis (14). TGase 2 has been detected in the synovial fluid of patients with arthritis (15) and in the serum and cerebral spinal fluid of patients with amyotrophic lateral sclerosis (16). These reports suggest that the inappropriate expression and/or presentation of TGase 2 to T cells might contribute to these diseases in genetically predisposed individuals.
Activation of glia, a process termed reactive gliosis, has been observed during the pathogenesis of neurodegenerative diseases. A hallmark of brain inflammation is the activation of microglia cells that produce neurotoxic factors such as nitric oxide (NO) (17) and TNF-␣ (18). Synthesis and release of these factors constitute a part of the innate immunity that enables the host to destroy invading pathogens. However, excessive production and accumulation of NO seems to contribute neurodegeneration (19). TGase 2 expression in rat brain astrocytes is induced by glutamate (20) or by the inflammation-associated cytokines such as interleukin-1␤ and TNF-␣ (21). Microarray analysis in a monkey model of neuro-AIDS has recently revealed that TGase 2 expression is specifically increased in the infected brain (22). Immunohistochemical staining in the infected brain shows clear increase of TGase 2 in the cells with astrocytic morphology. Multiple factors must contribute to the activation of TGase 2 in oxidative stress and in the elevation of intracellular calcium. This suggests that induced-TGase 2 in the activated astroglial cells may participate in the pathogenesis of neurodegenerative diseases. Neurodegenerative diseases, such as Parkinson's disease (23,24) and Alzheimer's disease (25,26), and neuro-AIDS brains (22) are closely associated with increased brain TGase 2 expression. Inflammatory markers also occur in Parkinson's disease (27), Alzheimer's disease (28), and AIDS dementia (17).
We demonstrated previously that the recombinant peptide R2, which had dual inhibitory effects on purified TGase 2 and soluble phospholipase A 2 , reversed the inflammation of allergic conjunctivitis to ragweed in a guinea pig model (29). Recombinant peptides were constructed with the sequences from pro-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Both authors contributed equally to this work. ʈ To whom correspondence should be addressed. Tel.: 914-597-2500, ext. 3041; Fax: 914-597-2757; E-mail: tgase@hanmail.net. 1 The abbreviations used are: TGase 2, transglutaminase 2; TNF-␣, tumor necrosis factor-␣; LPS, lipopolysaccharide; NF-B, nuclear factor-B; PBS, phosphate-buffered saline; iNOS, inducible nitric-oxide synthase; NMMA, N G -monomethyl-L-arginine; RT, reverse transcription; LDH, lactate dehydrogenase; IKK, I-B kinase; I-B␣, inhibitory subunit ␣ of nuclear factor-B; NIK, nuclear factor-B inducing kinase; SEAP, secreted alkaline phosphatase; RA, retinoic acid. elafin (for inhibition of TGase 2), and antiflammins (for inhibition of soluble phospholipase A 2 ). It is interesting that we found that the only pro-elafin sequence, E2 (a part of elafin, a TGase inhibitor), itself also has dramatic anti-inflammatory effects (29). This strongly suggests that TGase activity may play a key role in macrophage activation resulting from inflammatory stress. We observed recently that TGase 2 expression is increased by LPS treatment in BV-2 microglia, and NO release is dramatically reduced by TGase inhibitors (30). During the LPS-induced microglia activation, TGase activity is increased about 5-fold in microglia after 24-h exposure to LPS in a timedependent manner (30). The increase of NO synthesis is correlated with increase of TGase 2 expression. Furthermore, we observed that transient transfection of TGase 2 in BV-2 microglia increases NF-B activity (30). This suggests that TGase 2 may aggravate inflammation by activating the NF-B cascade in microglia. To test this hypothesis, we determined the identity of the major in vivo target of TGase 2 in the NF-B cascade. We also tested whether TGase inhibition might be effective at reducing NF-B activation in the LPS-induced rat brain injury model.
Nitrite Measurements-Accumulated nitrite was measured in the cell supernatant after LPS treatment for 24 h by the Griess reaction (33). In brief, 200-l aliquots of cell supernatant from each well are mixed with 100 l of Griess reagent [1% sulfanilamide (Fluka), 0.1% naphthylethylenediamine dihydrochloride (Fluka), and 2.5% H 3 PO 4 ] in a 96-well microtiter plate, and the absorbance was read at 540 nm using a plate reader.
Semiquantitative RT-PCR of Mouse TGase 2 and iNOS with Competitive Standard Mimics-Semi-quantitative RT-PCR was performed by using competitive mimic templates (internal amount control) (34). To prepare total RNA for RT-PCR, the cells were lysed by adding TRIzol reagent according to the manufacturer's instructions. Samples of total RNA were reverse-transcribed at 42°C using the first strand synthesis kit (Promega) with avian myeloblastosis virus reverse transcriptase, and PCR was performed for the transcripts of iNOS and TGase 2 using specific primer sets. All PCRs contained the following in a volume of 20 l: 1.5 mM MgCl 2 , 200 M dNTP, 0.2 M concentration of each upstream and downstream primer, 0.5 units of Taq polymerase, and variable amounts of templates as required. The mimic templates of TGase 2 and iNOS were constructed by PCR to retain each target PCR primer. The mimics of mouse TGase 2 and mouse iNOS were made from 2014 -2338 and 1451-2043 bp, respectively. The products of RT-PCR were: 526 bp for target TGase 2, 345 bp for mimic TGase 2, 593 bp for target iNOS, and 417 bp for mimic iNOS. The PCR primer sequences were; mouse TGase 2 sense strand (5Ј-CCA AGC AAA ACC GCA AAC TG-3Ј), mouse TGase 2 antisense strand (5Ј-TGA TGG CTC TCC TCT TAC CCT TTC-3Ј); mouse iNOS sense strand (5Ј-ACT ACC AGA TCG AGC CCT GGA AC-3Ј), and mouse iNOS antisense strand (5Ј-GCA AGC TGA GAG GCT CCC AGG-3Ј).
Stable Transfection of TGase 2-The human neuroblastoma cell line SH-SY5Y used for stable transfection studies was obtained from American Type Culture Collection. The SH-SY5Y cells were grown in Dulbecco's modified Eagle's medium/Ham's F12 medium (50:50) supplemented with 10%-heat inactivated fetal bovine serum, glutamine, and penicillin/streptomycin as described previously. To avoid clonal variation, we adopted the Flp-In™ System (35) (Invitrogen, Co). We used SH-SY5Y/FRT cells carrying empty vector as a control and SH-SY5Y/TG cells expressing full-length human TGase 2 in pcDNA5/FRT vector. After selection, there was no increase of cell death signs in SH-SY5Y/TG cells by the criteria of normal cell growth, lactate dehydrogenase (LDH) release, 4Ј,6Ј-diamidino-2-phenylindole, dihydrochloride staining, caspase activity, and annexin V staining (data not shown). This is in accordance with the previous report that TGase 2-transfected neuroblastoma cells do not show increased apoptosis unless they are subjected to oxidative stress (36,37).
To test whether the effect of TGase 2 on cellular targets can be reversed, we employed the tetracycline induced expression system using the EcR 293 cell line (Flp-In T-Rex-293; Invitrogen). After introduction of full-length human TGase 2 using pcDNA5/FRT into EcR 293 cells and selection with hygromycin, TGase 2 was induced by 1 g/ml of tetracycline treatment for 24 h in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum.
IKK Inhibitor Treatment-To test whether TGase 2-induced NF-B activation is IKK-dependent, the IKK-2 inhibitor SC-514 (Calbiochem) was employed (74). As a positive control, BV-2 was activated with LPS with or without SC-514. BV-2 cells were pre-treated with or without 10 M SC-514 for 1 h before 30-min LPS induction. SH-SY5Y and SH-SY5Y/TG cells were also treated with or without 10 M SC-514 for 1 h. Cells were harvested and the cytosolic fraction was collected for Western blotting analysis.
Transglutminase Activity Assay-A modified TGase assay method was used to determine the enzymatic activity by measurement of the incorporation of [1,4-14 C]putrescine into succinylated casein (4,38).
Western Blotting-The cytosolic fraction of samples was prepared using the nuclear extract kit (Sigma). Samples were applied to the wells of 10 -20% gradient SDS gels using Tricine buffer (Invitrogen) and then transferred to polyvinylidene difluoride membranes (Invitrogen). Western blotting was performed as established previously (4). Antibodies to NF-B p65, I-B␣, phospho-IB-␣ (Ser32), I-B kinase ␤ (IKK-␤), phospho-IKK␣ (ser180)/IKK␤(Ser181), and NF-B activating kinase were obtained from Cell Signaling Technologies (Beverly, MA). Antibodies to NIK, IKK␣, and ␣-topoisomerase I were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies to LDH (Research Diagnostics, Inc., Flanders, NJ), ubiquitin (Sigma) and TGase 2 (clone CUB 7402; NeoMarkers, Union City, CA) were obtained as indicated. The concentration of primary and secondary antibodies were 5 and 0.1 g/ml, respectively. The blot was then developed by enhanced chemiluminescence (Pierce, Milwaukee, WI). To determine the purity of extracted cytosol and nuclear fractions, we employed anti-LDH for the cytosolic fraction and anti-␣-topoisomerase for the nuclear fraction.
In Vitro Cross-linking Experiments-To obtain I-B␣, the full-length human I-B␣ was cloned into pET-30 Ek/LIC vector (Novagen) by PCR using full-length I-B␣ cDNA (pCMV-IB␣; BD Biosciences), expressed and purified using a HisTrap column (Amersham Biosciences) according to the supplier's manual. Purified human recombinant NF-B (p52) protein was obtained from Santa Cruz Biotechnology. Purified I-B␣ (2 M) or NF-B (p52) (2 M) was incubated with or without 0.001 units of guinea pig liver TGase 2 for 30 min at 37°C in 20 l of Tris-HCl, pH 7.5, containing 10 mM CaCl 2 . After incubation, the sample was analyzed by Western blotting for I-B␣ and by Coomassie protein staining for NF-B (p52).
Binding Efficiency of Free and Polymerized I-B␣ to NF-B-The full-length I-B␣ was prepared as described above. The full-length human NF-B(p65) was obtained from Active Motif Co. Incubation of 2 M I-B␣ with TGase 2 (0.001 units) for 30 min at 37°C shows complete polymerization of I-B␣ (Fig. 3C). To test the binding efficiency of free or polymerized I-B␣ to NF-B(p65), various concentrations of I-B␣ (0.25-2.0 M) were incubated with or without TGase 2 (0.001 units) for 30 min at 37°C in 50 mM Tris-Cl buffer, pH 7.5, containing 10 mM CaCl 2 , and the reaction was terminated by addition of 20 mM EDTA. NF-B (2 M) was added to the I-B␣ mixture for 1 h at RT. For immunoprecipitation, the mixture was gently mixed with 5 g of NF-B(p65) antibody for 1 h at RT, and a protein A/G-agarose-conjugated slurry (Pierce) was added for 1 h at RT. The precipitate was collected after centrifugation at 2,000 ϫ g for 5 min, boiled in the loading buffer, and loaded onto 10 -20% gradient Tricine-polyacrylamide gels. After electrophoresis, proteins were transferred onto the polyvinylidene difluoride membrane and Western blotting was performed against I-B␣ Transient Transfection of TGase 2-Experiments involving transient expression of TGase 2 were conducted using cDNAs encoding for fulllength human TGase 2 cloned in the pSG5 vector (Stratagene) (23). The transient transfection was performed by the calcium phosphate method. When mouse BV-2 cells reached 80% confluence in 6-well tissue culture dishes, the medium was replaced with 2 ml of fresh culture medium. The plasmids (1 g) were prepared in the presence of 25 mol of calcium in 100 l of medium. The same amount of 2ϫ HEPES-buffered saline was prepared. The plasmid and calcium mix were added slowly to the 2ϫ HEPES-buffered saline buffer, and the resulting mixture was incubated for 20 min at room temperature. The mixture was gently vortexed and added drop-wise to the culture medium.
NF-B Activity Assay-We measured NF-B activity using the secreted alkaline phosphatase (SEAP) reporter system 3 (pNFB-SEAP; BD Biosciences Clontech). The culture medium was replaced with fresh medium at 12 h after transient transfection. The medium was collected for SEAP assay after 24 h and the cells were harvested for ␤-galactosidase assay. The vehicle vector pSG5 (Stratagene) was used as a control. Cells treated with pGAL plasmid (1 g) were co-transfected with expression vectors to normalize expression to ␤-galactosidase activity (39). The SEAP assay was performed according to the supplier's instructiosn (BD Biosciences Clontech). Values are the means of three determinations (S.D. Ͻ 10%).
Effect of TGase Inhibitors on Reduced I-kB␣ in SH-SY5Y/TG Cells-Cystamine is known to inhibit TGase activity by blocking access of the quences (29,46). The effectiveness of R2 and E2 as TGase 2 inhibitors was previously demonstrated in vitro and in vivo (29). To determine the effects of TGase inhibitors on I-B␣ reduction, different inhibitors were added to the SH-SY5Y/TG culture for 30 min. Thereafter, the cytosolic fraction was obtained by using the nuclear extract kit (Sigma) for Western blotting analysis of I-B␣.
Effect of TGase Inhibitors in Vivo on LPS-induced Rat Brain Injury-Male Sprague-Dawley rats (Samtako, Osan, Korea) weighing 190 -220 g were used for the experimental model of intraperitoneal LPS injection described previously (47). All experimental procedures were approved by the Seoul National University Care of Experimental Animals Committee. Rats were injected intraperitoneally with LPS (2.5 mg/kg) dissolved in 0.9% saline or 0.9% sterile saline. To determine the effect of TGase inhibitors, animals were intraperitoneally injected with R2 peptide (25 M), E2 peptide (25 M), and dexamethasone (1 mg/kg) once at 30 min before and once at the time of the LPS injection. Rats were killed 1 h after the LPS injection. Dexamethasone injection was employed as a positive control.
Immunohistochemistry-At 1 h after intraperitoneal injection of LPS or saline, rats were deeply anesthetized with 1% ketamine (30 mg/kg) and xylazine hydrochloride (4 mg/kg). Brains were perfused through the heart with saline containing 0.5% sodium nitrite and 10 units/ml heparin followed by perfusion with 4% paraformaldehyde in PBS (0.1 M, pH 7.2). Brains were removed and postfixed, rinsed in PBS, and cryoprotected in sucrose. Sections were prepared on a sliding microtome (40 m) at the level of subfornical organ (47). Immunohistochemical stain-ing of TGase 2 was performed with a monoclonal antibody to TGase 2 (TG-100; NeoMarkers). Brain sections were blocked with 1% bovine serum albumin in PBS and incubated overnight with primary antibody solution (1:200 dilution). After washing for 30 min with PBS, the sections were incubated with biotinylated goat anti-mouse IgG for 1 h followed by incubation with peroxidase-avidin for 1 h, and visualized with the Vector Elite kit (Vector Laboratories, Burlingame, CA). Floating sections were mounted on slides, dehydrated through graded alcohols, and coverslipped (48). Controls for staining specificity were: preabsorption of the primary TGase 2 antibody with purified guinea pig liver TGase 2 (Sigma); omission of the primary antibody; or its replacement with nonimmune serum. None of the controls showed positive staining (data not shown).
Comparative RT-PCR-Samples of total RNA from rat brain tissues including subfornical organ (2 mm from the optic chiasm) were reversetranscribed using the first strand synthesis kit (Roche Molecular Biochemicals), and PCR was performed on the transcripts of TNF-␣ and ␤-actin using specific primer sets. RT-PCR primers for the targets were made from 923-1242 bp of rat TNF-␣ and 91-760 bp of rat ␤-actin. Therefore, the products of RT-PCR were 320 and 670 bp, respectively. The PCR primer sequences were as follows: rat TNF-␣ sense, CCC CAT TAC TCT GAC CCC TT; rat TNF-␣ antisense, AGG CCT GAG ACA TCT TCA GC; rat ␤-actin sense, GGC ATT GTA ACC AAC TGG GAC; rat ␤-actin antisense, TGT TGG CAT AGA GGT CTT T. To ensure a linear relationship between the amount of PCR product and amount of total RNA, we used variable cycles for rat TNF-␣ and ␤-actin.

RESULTS
The induction of TGase 2 in LPS-induced BV-2 microglia is shown in Fig. 1. TGase activity was increased about 5-fold concomitant with a 10-fold increase of NO release after 24-h LPS treatment (Fig. 1A). RT-PCR analysis of iNOS and TGase 2 was performed in BV-2 cells after LPS treatment (Fig. 1B). TGase 2 is increased about 3-fold concomitant with a 10-fold increase of iNOS. Furthermore, we observed that transient transfection of TGase 2 in the BV-2 microglia increases NF-B activity (30). iNOS is triggered by NF-B activation. Therefore, the data suggest that TGase 2 is probably involved in the regulation of the NF-B cascade. To test whether NO induces TGase 2 expression, BV-2 was treated with LPS and NMMA (iNOS inhibitor). NMMA did not affect TGase activity, but it did reduce NO secretion in a dose-dependent manner (Fig. 1C).
To identify major targets of TGase 2 in the NF-B cascade, a series of Western blotting experiments was done using SH-SY5Y cells stably transfected with TGase 2 (Fig. 2). TGase activity is increased about 8-fold in the cytosolic fraction of the SH-SY5Y/TG cells (Fig. 2A). Western blotting analyses found no change of NF-B activating kinase, NIK, IKK␣, and p-IKK [p-IKK␣ (Ser-180) and p-IKK␣ (Ser-181)] (data not shown).
Western blotting of I-B␣ and NF-B followed by densitometry showed a ϳ50% decrease in the cytosol and a ϳ30% increase in the nucleus, respectively (Fig. 2B). No change of p-I-B␣ was observed between SH-SY5Y and SH-SY5Y/TG cells (Fig. 2B). TGase 2 transfection in the BV-2 microglia, which resulted in a ϳ5-fold increase of TGase activity (30), was associated with a decrease of I-B␣ level as assessed by Western blotting (Fig.  5A). To test whether the decrease of free I-B␣ by TGase 2 transfection is IKK-dependent or not, SC-514 treatment (IKK-2 inhibitor) was employed (Fig. 2C). We found no change of p-I-B␣ between SH-SY5Y and SH-SY5Y/TG cells with or without SC-514, whereas LPS-treated BV-2 showed a decrease of p-I-B␣ with SC-514 (Fig. 2C).
To test whether TGase 2 reduces I-B␣ level via the ubiquitin-proteasome system, SH-SY5Y/TG cells were incubated for 6 h with proteasome inhibitors including MG 132 (0, 10, 100 M), lactacystin (0, 1, 10 M), or carbobenzoxy-L-isoleucyl-␥-tbutyl-L-glutamyl-L-alanyl-L-leucinal (0, 1, 10 M) (Fig. 3A). The cytosol was extracted from cells and analyzed by Western blotting of I-B␣ and ubiquitin. LDH activity in the medium and caspase-9 expression by Western blotting in the treated cells were not detected during course of the experiment (data not  (Fig. 3C). The mixture of I-B␣ and NF-B(p65) was immunoprecipitated (IP) against NF-B(p65) and subjected to Western blotting against I-B␣. The free I-B␣ is very efficiently bound to NF-B(p65) in a dose-dependent manner (arrow). However, the polymerized I-B␣ (asterisk) lost almost all binding capacity to NF-B(p65). Densitometry analysis revealed that the binding efficiency of polymerized I-B␣ to NF-B is reduced to less than 10% the level of free I-B␣. Data are presented as mean Ϯ S.D. from three repeats. shown). If TGase 2-induced NF-B activation depends on the IKK/ubiquitin/proteasome pathway, the I-B␣ and ubiquitinated I-B␣ should be increased. The level of I-B␣ (arrowhead) in SH-SY5Y/TG cells was decreased by proteasome inhibition, whereas ubiquitinated high molecular weight proteins were increased in a dose-dependent manner (Fig. 3A). Increased ubiquitinated I-B␣ (asterisk) was not detected by Western blotting (Fig. 3A). In SH-SY5Y/TG cells, Western blotting of I-B␣ showed a reduced level of I-B␣ (Fig. 3B, arrowhead) accompanied by highly polymerized I-B␣ (Fig. 3B, asterisk). In vitro incubation of purified I-B␣ (Fig. 3C,  arrowhead) with 0.001 units of purified guinea pig liver TGase 2 for 30 min resulted in completely polymerized I-B␣ (Fig. 3C,  asterisk) by Western blotting. In vitro incubation of NF-B (p52) with TGase 2 did not result in polymerization of this protein (Fig. 3D).
We determined whether the polymerized I-B␣ no longer interacts with NF-B in vitro (Fig. 4). Under the condition of TGase treatment shown in Fig. 3C, free I-B␣ (0.25-2 M) is completely cross-linked to a high molecular weight polymer. After free I-B␣ was treated with or without TGase 2, the mixture was incubated with NF-B. After the incubation, the mixture was immunoprecipitated using NF-B antibody, and the precipitates were subjected to Western blotting analysis against I-B␣. The free form of I-B␣ is able to bind to NF-B very efficiently in a dose-dependent manner (Fig. 4B, arrow). However, the polymerized I-B␣ (Fig. 4B, asterisk) lost binding capacity. Densitometry analysis revealed that the binding efficiency of polymerized I-B␣ to NF-B is reduced to less than 10% of the level of free I-B␣ (Fig. 4B).
NF-B activation analyses were done by NF-B/SEAP reporter assay normalized to ␤-galactosidase activity (Fig. 5, A and B, lower bar graph) and electrophoretic mobility shift assay with nuclear fractions after transfection with TGase 2 (Fig. 5C). Western blotting of I-B␣ and TGase 2 were performed (Fig. 5, A and B, top inset). Transient transfection of TGase 2 in BV-2 cells, using cDNAs encoding for full-length human TGase 2 cloned in the pSG5 vector (23), reduced the level of I-B␣ in the cytosol. This was accompanied by a 2-fold increase of NF-B activity (Fig. 5A). Stable transfection of TGase 2 in SH-SY5Y cells reduced I-B␣ in the cytosol, which resulted in a 3-fold increase of NF-B activity (Fig. 5B). Binding reactions were performed with nuclear extractions from BV-2 and SH-SY5Y cells with or without transfection of TGase 2 using a double-stranded consensus oligonucleotide for NF-B end-labeled with [ 32 P]ATP (Fig. 5C). Gel shift showed 3-and 2-fold increases of NF-B in BV-2 and SH-SY5Y cells, respectively, after TGase 2 transfection.
The effect of regulation of TGase expression on the level of I-B␣ was tested in EcR 293 and SH-SY5Y/TG cells (Fig. 6). To control TGase 2 expression, a tetracycline-induced expression system in the EcR 293 cell line was employed (Fig. 6A). EcR 293 cells were collected before incubation (Fig. 6A, left, Ϫ), after 24-h incubation with medium containing 1 g/ml tetracycline FIG. 5. NF-B activation by TGase 2 transfection. Analyses were done by Western blotting of I-B␣ and TGase 2 (top) and by NF-B/SEAP reporter assay normalized to ␤-galactosidase (␤-gal) activity (lower bar graph) (A and B). Electrophoretic mobility shift assay was also performed with nuclear fractions of BV-2 and SH-SY5Y after transfection with TGase 2 (C). A, transiently transfected TGase 2 in BV-2 cells reduced I-B␣ in the cytosol, which resulted in a 2-fold increase of NF-B activity. Data are presented as mean Ϯ S.D. from three samples. (W, non-transfected; Ct, transfected with empty vector, pSG5; TG, transfected with TGase 2, pSG5/TG). B, stable transfection of TGase 2 in SH-SY5Y cells reduced I-B␣ in the cytosol, which resulted in a 3-fold increase of NF-B activity. Data are presented as mean Ϯ S.D. from three samples (W, SH-SY5Y; Ct, SH-SY5Y/FRT; TG, SH-SY5Y/TG cells). C, binding reactions were performed with nuclear extractions from BV-2 and SH-SY5Y cells with or without TGase 2 transfection using a double-stranded consensus oligonucleotide for NF-B end-labeled with [ 32 P]ATP. A gel-shift assay showed 3-and 2-fold increases of NF-B in BV-2 and SH-SY5Y after TGase 2 transfection, respectively. Data are presented as mean Ϯ S.D. from three repeats. (Fig. 6A, center, ϩ), and after another 24-h incubation with fresh medium without tetracycline followed by a 24-h incubation with tetracycline (Fig. 6A, right, Ϫ). The level of free I-B␣ was regulated reciprocally by the level of TGase 2 expression without phosphorylation of I-B␣ (Fig. 6A). To test whether we can obtain the same effect with TGase inhibitors, SH-SY5Y/TG cells were incubated for 30 min with TGase inhibitors, including cystamine (50 M), iodoacetamide (50 nM), E2 peptide (100 M), and R2 peptide (100 M). The I-B␣ level in the cytosolic fraction was measured by Western blotting. TGase inhibitors reduced the level of I-B␣ almost to the control level (Fig. 6B). Data are presented as mean Ϯ S.D. from three samples.
The effect of TGase inhibitors on LPS-induced rat brain injury was also determined (Fig. 7). LPS (2.5 mg/kg) or saline alone was injected intraperitoneally into rats. Immunohistochemical staining of TGase 2 showed increased expression in rats killed 1 h after the LPS injection (LPS/Brain) compared with the rats killed after saline injection (Ct/Brain) (Fig. 7A). Induction of TGase 2 expression is apparent in the subfornical organ and choroid plexus concomitant with an increase of I-B␣ expression as assessed by in situ hybridization histochemistry (47). To test the effect of TGase inhibitors on neuroinflammation, rats were injected twice with TGase inhibitors. The inflammatory cytokine TNF-␣ level measured by RT-PCR, using a ␤-actin control, was significantly reduced by the inhibitors (Fig. 7B) (*, p ϭ 0.017; **, p ϭ 0.04, n ϭ 6). DISCUSSION We and others have observed that TGase 2 is induced after LPS treatment in various tissues and cells (9,30,49). In this study, we showed that an increase of TGase activity is associated with NF-B activation via an IKK-independent pathway in microglia and SH-SY5Y cells. TGase 2 is regulated both at the transcriptional and post-transcriptional levels. In addition to TGase 2 induction by LPS treatment, TGase 2 can be induced by various stresses, including oxidative stress (58), UV (50,51), calcium influx brought about by treatment with a calcium ionophore or maitotoxin (52,53), retinoic acid (RA) (54,55), inflammatory cytokines (56,57), glutamate (20), and virus infection (22). Oxidative stress and reactive oxygen intermediates have been shown to increase TGase 2 expression (58). UV radiation generates singlet oxygen and superoxide in the eye. Superoxide rapidly dismutates to hydrogen peroxide, which elevates Ca 2ϩ (59). Therefore, it is possible that changes in Ca 2ϩ activate TGase 2. TGase 2 is significantly increased and translocated to the nucleus by maitotoxin treatment in neuroblastoma cells (53). Maitotoxin activates both voltage-sensitive and ligand-gated calcium channels, which increases intracellular calcium concentrations. RA increases TGase activity biphasically in Chinese hamster ovary cells (54). A recent report showed that phosphoinositide 3-kinase activity was required for RA to increase TGase 2 protein levels in NIH3T3 cells (55). Activation of the Ras-ERK pathway by epidermal growth factor was sufficient to elicit this effect, because continuous Ras signaling mimicked the actions and inhibited RA-induced TGase 2 expression, whereas blocking ERK activity in these same cells restored the ability of RA to up-regulate TGase 2 expression (60). This is a good example of signal-regulated TGase 2 induction. Exposure of astrocytes in primary culture to glutamate also increases TGase 2 expression and translocation of TGase 2 into the nucleus (20). Glutamate exposure of astroglial cells causes ligand-gated channel receptor activation, associated with an excitotoxic cell response. In the glutamate-exposed astroglial cells, an ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate receptor inhibitor reduced TGase 2 expression (20). TGase 2 is induced by inflammatory cytokines, such as TNF-␣ in liver cells (56) and interferon-␥ in intestinal cells (57). TGase 2 also can be induced directly by NF-B activation in liver cells because the TGase 2 promoter has an NF-B binding motif (61). It is interesting that TGase 2 seems to control NF-B activation. This is an intriguing finding because many stimulations triggering NF-B activation overlap with stimulations of TGase 2 such as nontypeable Haemophilus influenza infection (62), UV (63), reactive oxygen intermediates (pervanadate) (64), interleukin-1␣, and TNF-␣ signaling (65) in addition to LPS. This implies that TGase 2 and NF-B may induce each other (Fig. 8). Several pathways for NF-B activation do not rely on IKK activation. These pathways upregulate inflammatory mediators. An IKK-independent NF-B activation pathway via activation of the MKK3/6-p38 mitogenactivated protein kinase pathway occurs in epithelial cells challenged by nontypeable Haemophilus influenza (62). Therefore, nontypeable Haemophilus influenza seems to activate NF-B in human epithelial cells via two distinct signaling pathways, including TLR2-TAK1-dependent NIK-IKK␣/␤-IkB␣ and MKK3/6-p38 mitogen-activated protein kinase pathways. UVinduced I-B␣ degradation is IKK-independent (63). Even an I-B␣ mutant containing alanines at positions 32 and 36 is also susceptible to UV-induced degradation (75). UV-induced NF-B activation depends on phosphorylation of I-B␣ at Cterminal sites (PEST domain, explained below) that are recognized by CK2 (casein kinase II) (66). Furthermore, CK2 activity toward I-B␣ is UV-inducible through a mechanism that depends on activation of p38 mitogen-activated protein kinase. Inhibition of this pathway prevents UV-induced I-B␣ degradation and increases UV-induced cell death (66). However, this mechanism of I-B␣ degradation remains an enigma. Ionizing radiation induces tyrosine phosphorylation in human B-lymphocyte precursors by stimulation of tyrosine kinases (64). The tyrosine kinase inhibitor herbimycin A and the free radical scavenger N-acetyl-cysteine inhibit both radiation-and H 2 O 2induced NF-B activation, indicating that activation triggered by reactive oxygen intermediates is dependent on tyrosine kinase activity (64). Pervanadate-induced tyrosine phosphorylation leads to degradation of I-B␣, and this degradation is required for NF-B activation in human myeloid U937 and HeLa cells (67). A proteosome inhibitor blocked pervanadateinduced degradation without blocking tyrosine phosphorylation of I-B␣, indicating that phosphorylation alone is insufficient to induce degradation (67). The accumulated evidence suggests that NF-B activation may occur through different pathways depending on the stress. However, a major NF-B regulatory mechanism depends on signal-dependent I-B␣ degradation. In addition to the ubiquitination-mediated degrada-tion, I-B␣ can be degraded also by a lysosomal system or by a proline (P), glutamic acid (E), serine (S), and threonine (T) (PEST) sequence pathway. The lysosomal degradation of I-B␣ is activated under conditions of nutrient deprivation (68). In the Chinese hamster ovary cells, the half-life of I-B␣ is 4.4 days in normal medium but it is reduced to 0.9 day in serumdeprived medium (68). This report also showed that increase of I-B␣ degradation is completely blocked by lysosomal inhibitors (68). Rogers et al. found that the amino acid sequences of proteins with intracellular half-lives of less than 2 h contain one or more regions rich in PEST (69). A PEST region is found in the I-B␣ carboxyl terminus. Deletion of this region in I-B␣ results in increased half-life, suggesting that the PEST sequence may play a direct role in the degradation of I-B␣ (70). I-B␣ is readily converted to homopolymers catalyzed cross- FIG. 7. Effect of TGase inhibitors on LPS-induced rat brain injury. LPS or saline alone was injected intraperitoneally into rats. A, immunohistochemical staining of TGase 2 showed increased expression in rats killed at 1 h after the LPS injection (LPS/Brain) compared with the rats killed after saline injection (Ct/Brain). Induction of TGase 2 expression is apparent in the subfornical organ and choroid plexus (dark purple). Scale bar, 2.0 mm. B, rats were injected intraperitoneally with R2 peptide (200 l of 1 M) and E2 peptide (200 l of 1 M), or with dexamethasone (1 mg/kg) once at 30 min before and once at the time of LPS injection. Rats were killed 1 h after the LPS injection. Brain sections, including subfornical organ (2 mm from optic chiasm), were collected for RNA extraction. The level of the inflammatory cytokine TNF-␣ was measured by RT-PCR with ␤-actin as a control. TGase inhibitors significantly reduced TNF-␣ expression with p value (*, p ϭ 0.017; **, p ϭ 0.04, n ϭ 6).
linking by TGase 2 in vitro (Fig. 3C). Furthermore, polymerized I-B␣ lost binding capacity to NF-B (Fig. 4B). Densitometry analysis shows that polymerized I-B␣ lost more than 90% binding capacity of free I-B␣ to NF-B. This implies that I-B␣ is both an acyl donor and an acyl acceptor for TGase cross-linking. We found that putative acyl donors (glutamine residues) are clustered at the end of the carboxyl terminus before PEST sequences, and putative acyl-acceptors (lysine residues) are clustered at the amino terminus. This resembles many other TGase substrate structures as building blocks found in barrier tissues. We suggest that antiparallel homopolymerization of I-B␣ may lead to inhibition of binding to NF-B, or concentrate PEST sequences on one side, which could accelerate degradation of I-B␣. Therefore, TGase 2-induced I-B␣ polymerization may be a major pathway for rapid depletion of free I-B␣ during stress induced by UV, radiation, and nontypeable Haemophilus influenza (Fig. 8).
We demonstrated recently that TGase inhibition using the peptide inhibitor R2, cystamine, and iodoacetamide in the LPSinduced microglia resulted in decreased NO synthesis (30). Herein, we showed that TGase inhibitors reverse the depletion of I-B␣ in SH-SY5Y/TG cells (Fig. 6). Therefore the effect of TGase inhibition on NO synthesis in the microglia is probably caused by blocking the depletion of I-B␣. In the LPS-induced brain injury model, we found that increased TGase 2 expression is restricted to the subfornical organ and choroid plexus (Fig. 7A) concomitantly with I-B␣ expression (47). The increase of I-B␣ mRNA parallels both I-B␣ protein degradation and NF-B activation, because transcription of I-B␣ is regulated by NF-B. Therefore, the TGase 2 expression pattern shows an early phase activation of inflammatory process (Fig.  7A). In this model, we failed to detect a change of I-B␣ protein in the brain after TGase inhibitor treatment, because I-B␣ is ubiquitously abundant in the brain. Therefore, we analyzed TNF-␣ expression, which is specifically regulated by NF-B.
Our results are similar to those in another report (71), which showed that elafin, including the pro-elafin sequence (secretory leukoprotease inhibitor), prevents inflammation in an LPSinduced acute lung injury model. Elafin treatment before the lung injury greatly reduces NF-B activation by inhibiting I-B␣ degradation (71). Elafin prevents LPS-induced NF-B activation by inhibiting degradation of I-B without affecting the LPS-induced phosphorylation and ubiquitination of I-B in U937 cells (72). In that study, a full-length elafin construct containing pre-elafin sequences was employed (72). It is interesting that the pre-elafin domain of SLPI, which is also known as a 'cementoin,' is an excellent TGase 2 substrate. We used the cementoin sequence to synthesize the strong competitive TGase inhibitors E2 and R2. This cementoin peptide contains four repeats of E2 and R2 sequences. It is possible that the cementoins in elafin may serve as in vivo TGase inhibitors (73), which could interfere with the prevention of NF-B activation (Fig. 8). This hypothesis is supported by the finding that elafin does not affect 20S proteasome peptidase-related activity in the cytoplasmic extract of U937 cells (72).
In conclusion, TGase 2 induces NF-B activation via two different pathways, an IKK-independent pathway and an IKKdependent pathway. We propose here a model for the role of TGase 2 during immune cell activation (Fig. 8). TGase 2-induced NF-B activation may be an important defense against infection, or conversely, a disease-associated mechanism in the inflammation process. This lends greater versatility to the innate immune system responding to various stimuli because TGase 2 is induced by various stresses. We demonstrated in this study that TGase inhibition is effective as a therapeutic approach to LPS-induced brain injury. This suggests that TGase inhibition may be beneficial in diseases associated with inflammation.