Induction of Glia Maturation Factor-β in Proximal Tubular Cells Leads to Vulnerability to Oxidative Injury through the p38 Pathway and Changes in Antioxidant Enzyme Activities*

Proteinuria is an independent risk factor for progression of renal diseases. Glia maturation factor-β (GMF-β), a 17-kDa brain-specific protein originally purified as a neurotrophic factor from brain, was induced in renal proximal tubular (PT) cells by proteinuria. To examine the role of GMF-β in PT cells, we constructed PT cell lines continuously expressing GMF-β. The PT cells overexpressing GMF-β acquired susceptibility to cell death upon stimulation with tumor necrosis factor-α and angiotensin II, both of which are reported to cause oxidative stress. GMF-β overexpression also promoted oxidative insults by H2O2, leading to the reorganization of F-actin as well as apoptosis in non-brain cells (not only PT cells, but also NIH 3T3 cells). The measurement of intracellular reactive oxygen species in the GMF-β-overexpressing cells showed a sustained increase in H2O2 in response to tumor necrosis factor-α, angiotensin II, and H2O2 stimuli. The sustained increase in H2O2 was caused by an increase in the activity of the H2O2-producing enzyme copper/zinc-superoxide dismutase, a decrease in the activities of the H2O2-reducing enzymes catalase and glutathione peroxidase, and a depletion of the content of the cellular glutathione peroxidase substrate GSH. The p38 pathway was significantly involved in the sustained oxidative stress to the cells. Taken together, the alteration of the antioxidant enzyme activities, in particular the peroxide-scavenging deficit, underlies the susceptibility to cell death in GMF-β-overexpressing cells. In conclusion, we suggest that the proteinuria induction of GMF-β in renal PT cells may play a critical role in the progression of renal diseases by enhancing oxidative injuries.

In chronic nephropathies, proteinuria is reportedly one of the best predictors, independent of mean arterial blood pressure, for disease progression toward end-stage renal failure (1,2). Microalbuminuria, which features a small quantity of albumin only (30 -300 mg/24 h), is known as an important early sign of diabetic nephropathy (3,4) and of progressive renal function loss in a non-diabetic population (5). In experimental models, proteinuria caused tubular insults accompanying infiltration of macrophages and T lymphocytes into the kidney (6). Interstitial inflammation can trigger fibroblast proliferation and accumulation of extracellular matrix proteins, which may facilitate tubulointerstitial fibrosis, which is a hallmark of progression of renal disease. In cultured proximal tubular (PT) 1 cells activated by administration of albumin, a number of genes encoding vasoactive and inflammatory molecules, which have potentially toxic effects on the kidney, were transactivated (7). These results strongly suggest that altering the disposition of PT cells by proteinuria must be involved in the process of renal damage. However, the mechanisms by which proteinuria accelerates renal disease progression remain largely unknown.
We recently found that the brain-specific glia maturation factor-␤ (GMF-␤) gene is induced in PT cells by proteinuria by comparison of the gene expression profiles (8,9) 2 of normal and proteinuria disease models (10,11). GMF-␤ is a 17-kDa brainspecific protein that was isolated from bovine brain homogenate as a substance inducing the maturation of normal neurons as well as glial cells (12,13); and at first, it was considered to be a neurotrophic factor. However, later intensive researches provided cumulative evidence that GMF-␤ is involved in cell signal transduction. This evidence comprised the following findings. 1) The GMF-␤ protein lacks a leader sequence and is not secreted by cells (14). 2) It contains consensus phosphorylation sites and is phosphorylated by protein kinase C, protein kinase A, casein kinase II, and ribosomal S6 kinase in in vitro studies (15,16). 3) GMF-␤ inhibits the extracellular signalregulated kinase-1/2 and enhances p38 activity (17). 4) Overexpressed GMF-␤ in primary astrocytes causes secretion of neurotrophic factors such as brain-derived neurotrophic factor and nerve growth factor through activation of the p38 pathway (18). The amino acid sequence of GMF-␤ is highly conserved among many species (19), suggesting that it plays basic roles across many species. The expression of GMF-␤ is largely limited to the brain (19), especially the glial cells and some neu-rons (20). Schwann cells of the distal segment of the transected nerve express GMF-␤, and this induction of GMF-␤ coincides with the temporal expression of nerve growth factor receptors in the cells (21). These results suggest that GMF-␤ may play a protective role in the brain. However, its precise function in neurons and glial cells is still largely a matter of speculation.
It was thus of interest to determine what impact could be served by the induction of the brain-specific protein GMF-␤ in PT cells by proteinuria. In this study, we demonstrate that proteinuria induced GMF-␤ in PT cells in a time-dependent manner. Overexpression of GMF-␤ in a mouse PT cell line and NIH 3T3 cell lines resulted in susceptibility to cell death upon stimulation with tumor necrosis factor-␣ (TNF-␣), angiotensin II, and other oxidative stress-inducing agents (buthionine sulfoximine (BSO) and H 2 O 2 ). Further studies clarified that GMF-␤ overexpression caused enhancement of oxidative injury, leading to F-actin reorganization and apoptosis under oxidative stress. We demonstrate that GMF-␤-overexpressing cells, as a mechanism of the vulnerability to oxidative stress, caused a prolonged increase in H 2 O 2 through changes in several antioxidant enzyme activities and that p38 was significantly involved in this process. These results suggest that the induction of the brain-specific GMF-␤ gene in the kidney may play a key role in renal disease progression caused by proteinuria.

EXPERIMENTAL PROCEDURES
Protein-overloaded Proteinuria Murine Model-Five-week-old C57BL6 male and female mice weighing ϳ20 g were intraperitoneally given bovine serum albumin (10 mg/g of weight; Sigma) dissolved in saline for 5 days during a 1-week period. The final dose of 10 mg/g of weight was reached by incremental increases in the dose over the first week, beginning with 2.5 mg/g of weight. Control mice were treated with saline (7,11). The load of bovine serum albumin was continued to 4 weeks. At 1, 2, 3, and 4 weeks, 24 h after loading of bovine serum albumin, kidneys were removed for investigations.
Tissue Preparation and Laser Capture Microdissection-The proteinoverloaded proteinuria model mice were separated into three random groups. After 1, 2, 3, or 4 weeks of loading of bovine serum albumin, kidneys from the first group (n ϭ three each) were removed after perfusion with phosphate-buffered saline (PBS) for Northern blot analysis, and those from the second group (n ϭ one each) after perfusion first with PBS and then with 4% paraformaldehyde. They were made into specimens with the paraffin sectioning method after paraformaldehyde fixation and used for in situ hybridization. Kidneys from the third group (n ϭ three each) were removed after perfusion first with PBS and then with 99.5% ethanol. Kidney tissue sections were prepared and subjected to laser microdissection using an LM200 Image Archiving workstation (Arcturus Engineering, Mountain View, CA) along with real-time PCR as described (25,26). Total RNA was extracted from samples attached to LCM transfer film using TRIzol (Invitrogen) and reverse-transcribed with SuperScript TM II RNase H reverse transcriptase (Invitrogen). Quantitation of GMF-␤ mRNA/rRNA was performed with this real-time PCR system according to the manufacturer's instructions. The GMF-␤ TaqMan probe was 5Ј-TGGTTTCA-GTCTCTGCTAGTTCATACCGCA-3Ј. The GMF-␤ forward primer sequence was 5Ј-GAGGCTTGAAACATTGGTGGTT-3Ј, and its reverse primer sequence was 5Ј-CAAGCACCATGCTTACCAAAAG-3Ј. The rRNA TaqMan probe and the forward and reverse primers were obtained from TaqMan rRNA control reagents (Applied Biosystems, Foster City, CA).
Northern Blot Analysis-Total RNA from mouse kidney was extracted with TRIzol according to the manufacturer's instructions. Ten micrograms of the RNAs were fractionated on formaldehyde-agarose gels, transferred onto nylon membranes (Hybond-N ϩ , Amersham Biosciences), and subjected to Northern analyses. For the probes, we used the mouse GMF-␤ sequences (bp 3080 -4130) obtained from the Gen-Bank TM /EBI Data Bank (accession number AF297220). Glyceraldehyde-3-phosphate dehydrogenase cDNA was used as an internal control.
In Situ Hybridization-GMF-␤ was subjected to in situ hybridization with the DNA nucleic acid detection kit (Roche Applied Science) according to the manufacturer's instructions. The GMF-␤ sense and antisense cRNA probes were prepared by PCReaction. The primers for PCR were 5Ј-GTCGCAGTAGAGTGGAGTGTGTTG-3Ј and 5Ј-CAGGTCAGGGCC-ATTCACTCTATG-3Ј. The PCR product was then subcloned into pST-Blue-1, and the sequence was confirmed to be identical to that of mouse GMF-␤ (data not shown). The clone was digested by XhoI to produce the template of the antisense probe and by BamHI to produce that of the sense probe. The digoxigenin-labeled antisense and sense cRNA probes were generated using 1 g of template and T7 or SP6 RNA polymerase, respectively, in combination with the digoxigenin RNA labeling mixture (Roche Applied Science).
Cell Culture and Transfection-The mouse renal proximal tubular cell line mProx24 (60) and NIH 3T3 cells were cultured in Dulbecco's modified Eagle's medium (Sigma) containing 10% fetal calf serum (FCS) (Invitrogen) in 5% CO 2 at 37°C. The mProx24 cells were maintained in 0.1% gelatin-coated culture dishes. The mouse GMF-␤ coding region was amplified by PCR using forward primer 5Ј-CGGGATCCC-GCTGACGACCGGAAGGAAAATGAGTGAG-3Ј and reverse primer 5Ј-CGGGATCCCGCCAGTACCCAGGAGTGGTCAGAGGAGG-3Ј, cut with BamHI, and inserted into the BglII site of FLAG-linked pCMV-Taq1 (Stratagene, La Jolla, CA). Permanent transfection of cells with mammalian expression vectors was achieved with LipofectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. The transfectants were maintained in Dulbecco's modified Eagle's medium containing 10% FCS and 200 g/ml Geneticin (G418, Sigma).
Immunoblotting-Total cell lysates were separated by SDS-PAGE (12% acrylamide) and then electrotransferred to polyvinylidene difluoride membranes (Hybond-P, Amersham Biosciences). FLAG-GMF-␤ was detected using mouse monoclonal anti-FLAG antibody M2 (Sigma) at a 5000:1 dilution. The blots were probed with individual primary antibodies as described above and then incubated with horseradish peroxidase-conjugated antibodies. Rabbit polyclonal antibody was raised against mouse GMF-␤ (the amino acid sequence of the epitope polypeptide was NH 2 -YQHDDGRVSYPLC-COOH) and confirmed by immunoblotting to bind specifically the GMF-␤ protein band. Anti-CuZn-SOD and anti-Mn-SOD antibodies were purchased from Stressgen Biotech Corp. (Victoria, British Columbia, Canada). Proteins were visualized with an enhanced chemiluminescence reagent (Pierce).
Cell Viability Assay-Cell viability for mouse PT and NIH 3T3 cells was determined 84 and 48 h, respectively, after treatments by the addition of the tetrazolium compound 3-(4,5-dimethylthiazol-2-yl)-5-(3carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) (inner salt; Promega) (27). The MTS tetrazolium compound is bioreduced by cells into a colored formation product that is soluble in culture medium. The solution of MTS mixed with the electron-coupling reagent phenazine methosulfate was added directly to the culture wells and incubated for 1 h, after which absorbance at 490 nm was recorded with a 96-well plate reader. Mouse TNF-␣, angiotensin II, H 2 O 2 , and BSO were purchased from Sigma (Sigma).
Lactate Dehydrogenase Release Assay-The results obtained in the MTS assay were confirmed in a lactate dehydrogenase release assay. A cytotoxicity detection kit (Roche Applied Science) was used to quantitate cytotoxicity/cytolysis based on the measurement of lactate dehydrogenase activity released from damaged cells. The lactate dehydrogenase activities in the culture supernatants of mProx24 and NIH 3T3 cells were determined 48 and 24 h, respectively, after treatments according to the manufacturer's instructions.
Detection of Apoptotic Cells by Annexin V Flow Cytometer Analysis-Detection of apoptotic cells was performed with an annexin V-enhanced green fluorescent protein apoptosis detection kit (Medical & Biological Laboratories, Nagoya, Japan). The assay takes advantage of the properties of binding of annexin V to membrane phosphatidylserine, which is translocated from the inner face of the plasma membrane to the apoptotic cell surface to make early detection of apoptosis by flow cytometry possible. After the addition of H 2 O 2 following one overnight FCS starvation, mProx24 cells were gently trypsinized and washed once with 10% FCS-containing Dulbecco's modified Eagle's medium before incubation with annexin V-enhanced green fluorescent protein at room temperature for 5 min in the dark. The analysis of annexin V-enhanced green fluorescent protein binding was performed with a FACSCalibur flow cytometry (excitation at 488 nm and emission at 530 nm; BD Biosciences) using a fluorescein isothiocyanate signal detector.
Activities of Caspase-3-Caspase-3 activities were detected using a colorimetric assay kit (Promega). Ac-DEVD-p-nitroaniline was used as the substrate for caspase-3. p-Nitroaniline was released from the substrate after cleavage with the DEVDase caspase-3. Free p-nitroaniline produced a yellow color that was monitored with a Emax microplate reader (Molecular Devices, Sunnyvale, CA) at 405 nm. After this treatment, the adherent mProx24 cells, both the stable transfectants and wild-type cells, were scraped off and washed with PBS by centrifugation. The cells were then resuspended in hypotonic cell lysis buffer (5 mM EDTA, 5 mM dithiothreitol, 25 mM HEPES (pH 7.5), 5 mM MgCl 2 , 2 mM phenylmethylsulfonyl fluoride, 10 g/ml pepstatin A, and 10 g/ml leupeptin) and lysed by three cycles of freezing and thawing. The cells were finally centrifuged at 16,000 ϫ g for 20 min at 4°C, and the supernatants were used as samples for caspase assay. The protein contents of the same supernatants were assayed with the Bio-Rad protein assay. Benzyloxycarbonyl-VAD-fluoromethyl ketone (50 M) was included in the caspase-3 assay kit.
Detection of Nucleosomal Ladders in Apoptotic Cells-DNA was extracted using the DNAzol reagent (Invitrogen). To detect nucleosomal ladders for apoptotic cells, the PCR-based amplification kit for DNA ladders (Clontech, Palo Alto, CA) was used according to the manufacturer's instructions.
Staining of F-actin and Confocal Fluorescence Microscopy-Before the various treatments, dish-cultured cells were transferred to 0.1% gelatin-coated coverslips. After 24 h, the medium was removed, and the cells were starved in serum-free medium for 12 h, after which they were fixed with 3.7% formaldehyde and permeated with 0.5% Triton X-100 (Sigma) in PBS (pH 7.5). F-actin was detected using 0.165 M TRITClabeled phalloidin (Sigma). The cells were examined by confocal microscopy using a Radiance 2100 BLD (Bio-Rad).
Determination of Reactive Oxygen Species (ROS) Production in PT Cells-Intracellular ROS generation was assessed using the oxidantsensitive dye 2Ј,7Ј-dichlorodihydrofluorescein diacetate (DCFDA) (Lambda Fluoreszenztechnologie, Graz, Austria) (28) in wild-type PT cells and those stably overexpressing GMF-␤. One of the ROS, intracellular H 2 O 2 , could be assayed (29). PT cells were treated with 5 g/ml DCFDA for 30 min at 37°C in Krebs-Ringer bicarbonate solution (118.3 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 2.5 mM CaCl 2 , 25.0 mM NaHCO 3 , and 11.1 mM glucose). After DCFDA incubation, the attached PT cells were cooled and harvested by trypsinization at 4°C, followed by flow cytometry analysis using a FACSCalibur. To examine the influence of intracellular esterase and probe efflux on the DCFDA assay, the oxidant-insensitive fluorescein diacetate derivative carboxyfluorescein diacetate (Molecular Probes, Inc., Eugene, OR) was used at 10 M. The assay method using carboxyfluorescein diacetate was the same as that using DCFDA.
SOD Activity Assay-SOD activity was measured as described by Sutherland and Learmonth (30) using an SOD assay kit (Trevigen, Gaithersburg, MD). This method is based on the inhibition of the reduction of nitro blue tetrazolium (NBT) by SOD. Superoxide ions convert NBT into NBT-diformazan. NBT-diformazan absorbs light at 550 nm. SOD reduces the superoxide ion concentration and thereby lowers the rate of NBT-diformazan formation. The extent of reduction in the appearance of NBT-formazan reflects the amount of SOD activity in a sample. CuZn-SOD was extracted with an ethanol/chloroform method, and its activity was measured. Mn-SOD activity was determined by subtracting CuZn-SOD activity from total SOD activity.
Catalase Activity Assay-Catalase activity was measured as described by Zhou et al. (31) using the Amplex® Red catalase assay kit (Molecular Probes, Inc.). Cells were lysed and processed in isotonic buffer (10 mM Tris-Cl (pH 7.4), 200 mM mannitol, 50 mM sucrose, and 1 mM EDTA). In the assay, catalase first reacts with H 2 O 2 to produce water and oxygen. Next, the Amplex Red reacts, as calculated with a 1:1 stoichiometry, with any unreacted H 2 O 2 in the presence of horseradish peroxidase to produce the highly fluorescent oxidized product resorufin. Since resorufin has also strong absorption of 563 nm, detection of this absorbance made it possible to determine the quantity of resorufin and therefore catalase activity.
Glutathione Peroxidase (GPX) Activity Assay-GPX activity was measured as described by Paglia and Valentine (32) using a GPX assay kit (Oxis Research, Portland, OR). This assay is an indirect measure of the activity of cellular GPX. Oxidized glutathione is recycled to its reduced form by glutathione reductase. The oxidation of NADPH resulting in NAD ϩ is accompanied by a decrease in absorbance at 340 nm so that GPX activity can be monitored. To assay cellular GPX, a cell homogenate is added to a solution containing glutathione, glutathione reductase, and NADPH. The enzyme reaction is started by adding t-butyl hydroperoxide as a substrate, and the absorbance at 340 nm is recorded every 30 s for 3 min. The rate of decrease in the absorbance at 340 nm is directly proportionate to the GPX activity in the sample.
Cellular Glutathione Content Assay-Cellular glutathione content was measured using a glutathione content assay kit (GSH-400 TM , Oxis Research). The method used with this kit is based on a chemical reaction that proceeds in two steps. The first step is the formation of substitution products (thioethers) between 4-chloro-1-methyl-7-trifluromethylquinolinium methylsulfate and all mercaptans that are present in the sample. The second step is a ␤-elimination reaction under alkaline conditions, which transforms the substitution product (thioether) obtained with glutathione into a thione with a maximal absorbance at 400 nm. Cell pellets are resuspended in an ice-cold metaphosphoric acid working solution, and the cell lysate is produced by homogenizing the cell suspension.
Statistical Analyses-Values are expressed as means Ϯ S.D. Except for the fluorescence-activated cell sorter analyses, all values were derived from measurements done in triplicate. Statistical analyses for multiple comparisons were performed with analysis of variance (ANOVA) and post hoc Bonferroni's correction. Unpaired Student's t test was used for the comparison of two groups in the studies with NIH 3T3 cells (Figs. 3, 4, and 8). A p value Ͻ0.05 was considered to indicate statistical significance. ANOVA and unpaired Student's t test were performed using a StatView software package (Abacus Concepts Inc., Berkeley, CA).

RESULTS
Northern Blot Analysis-The gene expression profile of the proximal tubules of the proteinuria mouse model (11) confirmed that the expression of GMF-␤ was up-regulated by proteinuria in 1 week. 2,3 To examine the gene expression of GMF-␤ in the diseased mouse model kidney after Ͼ1 week of proteinuria, Northern blot analysis was performed using RNA from kidney tissues of control and diseased mouse models. The gene expression of GMF-␤ was found increase in a time-dependent manner (Fig. 1A), resulting in a significant increase in GMF-␤ gene expression after 2 weeks of proteinuria.
In Situ Hybridization of GMF-␤ in the Kidney-To examine the localization of GMF-␤ mRNA expression in the diseased mouse model kidney, in situ hybridization was performed. An increase in GMF-␤ gene expression was seen mainly in the proximal tubules after 3 weeks of proteinuria (Fig. 1B), whereas no staining was detected in the control and 1-week kidney specimens (data not shown).
Laser Microdissection and Real-time PCR Analysis-Laser microdissection and real-time PCR analysis (25,26) were used to quantify the proximal tubule-specific gene expression of GMF-␤ induced by proteinuria. The laser microdissection enabled us to obtain a sufficient quantity of mRNA for the PT cells only (Fig. 1C, left). The increased expression of GMF-␤ mRNA in disease model PT cells was quantitatively confirmed in real-time PCR analysis (Fig. 1C, right). As shown in Fig. 1C, GMF-␤ mRNA in the PT cells from the proteinuria mouse model increased significantly (16.5 Ϯ 3.7 times) compared with that in control PT cells and in parallel with the progression of proteinuria. These findings matched the results of Northern analyses using whole kidney mRNA (Fig. 1A).
Proximal Tubular Cell Lines (mProx24) That Overexpress GMF-␤-To clarify the pathophysiological roles of GMF-␤ induction by proteinuria in PT cells, we constructed mouse proximal tubular cell lines (mProx24) that overexpress FLAGtagged GMF-␤ as described under "Experimental Procedures." Immunoblotting with anti-FLAG and anti-GMF-␤ antibodies was employed to confirm expression of the GMF-␤ protein in stable transformants and mock-transfected wild-type PT cells. The results indicated that a high level of the 17-kDa GMF-␤ transgene product was present in the GMF-␤ stable transformant clones GMF0 and GMF1 ( Fig. 2A). A basal level of endogenous GMF-␤ expression was detected in mock-transfected wild-type PT cells. Immunoblotting of the three types of cell lysates with anti-FLAG antibody detected FLAG-tagged GMF-␤ only in the GMF-␤ stable transformant ( Fig. 2A). We also transfected the FLAG-GMF-␤ construct into mouse fibro-blast NIH 3T3 cells to establish stable transformant cell lines. Immunoblotting confirmed the increased expression of GMF-␤ gene products (Fig. 2B).
Vulnerability of GMF-␤-overexpressing Cells to Cell Deathinducing Stimuli-In the process of establishing GMF-␤ stable transformants, sudden and extensive cell death was observed as cells grew to confluence, with most cells becoming detached from the culture dish. The apoptosis-inducing stimuli TNF-␣ (100 ng/ml) and angiotensin II (1 M) were added to the nonconfluent PT cells to confirm the susceptibility to cell death of the GMF-␤-overexpressing cells. The MTS cell viability assay showed that, with both stimuli, GMF-␤-overexpressing PT cells (mProx24) were more vulnerable to cell death stimuli than were mock-transfected wild-type cells (Fig. 3, A and B). Similar results were obtained with experiments using NIH 3T3 cells stably transfected with GMF-␤. The results for NIH 3T3 cells demonstrated that the vulnerability to cell death stimuli due to GMF-␤ overexpression was not restricted to the mouse PT cell line (mProx24). The results obtained with the MTS assay were confirmed by lactate dehydrogenase release assay (data not shown).
Oxidative Stress Enhanced by the Expression of GMF-␤ Induced Cell Death in PT Cells-Since both TNF-␣ and angiotensin II stimuli were reported to induce oxidative stress in cells and eventually cell apoptosis (33)(34)(35), we hypothesized that the cell lines could be susceptible to oxidative stress from other stimuli. The effects of oxidative stress on cells stably transfected with GMF-␤ were observed upon the addition of 250 M H 2 O 2 and 500 M BSO, an inhibitor of glutathione synthesis. It was found that GMF-␤ stable transformants were more susceptible to oxidative stress caused by both agents than were mock-transfected wild-type cells (Fig. 4, A and B). These results were confirmed in the case of GMF-␤-overexpressing NIH 3T3 cells (Fig. 4) and also using lactate dehydrogenase release assay in the two types of cells (data not shown), suggesting that GMF-␤ overexpression leads to an increase in cell death caused by oxidative stress.

GMF-␤ Overexpression Enhances Apoptosis Caused by H 2 O 2 -To further investigate the property of the cells with
GMF-␤ overexpression, we studied the effect of GMF-␤ overexpression on apoptosis of PT cells (clone GMF0) under oxidative stress. At 250 M H 2 O 2 , it took 84 h to detach the GMF0 cells. H 2 O 2 at 8 mM was needed to cause apoptosis in this PT cell line in 24 h (data not shown). Binding of annexin V to phosphatidylserine to indicate early apoptotic events was also examined. There was no difference in apoptosis in the two cell types under control conditions and 1 mM H 2 O 2 could not generate apoptotic change in either of the PT cell lines (Fig. 5A), whereas a clear increase in the numbers of apoptotic cells was observed only in GMF-␤-overexpressing PT cells at 8 mM H 2 O 2 (Fig. 5B). This increase in apoptotic cells was completely inhibited by the addition of the p38 inhibitor SB203580. Caspase-3 activity was significantly increased in GMF-␤ stable transformants after H 2 O 2 treatment and was inhibited by SB203580 and the caspase-3 inhibitor benzyloxycarbonyl-VAD-fluoromethyl ke-  cont and 1w, 2w, 3w, and 4w, control mice and mice at 1, 2, 3, and 4 weeks of proteinuria, respectively. B, in situ hybridization of GMF-␤. The data obtained with the antisense cRNA probe are shown (left, magnification ϫ100; right, magnification ϫ400) after a 3-week protein overload. Positive signals (dark purple grains) were observed mainly in the epithelial PT cells, but not in the specimen with the sense cRNA probe. C, expression of GMF-␤ mRNA in PT cells quantified by laser microdissection in combination with real-time PCR as described under "Experimental Procedures." As indicated by the arrowheads (left), only PT cells were isolated and collected by laser microdissection from 5-m thick kidney section specimens. The GMF-␤ mRNA/rRNA ratio in the renal sections of the disease model PT cells resulting from 3-week proteinuria increased by a factor of 16.5 Ϯ 3.7 (mean Ϯ S.D., n ϭ 3) compared with that in control PTs (right). *, p Ͻ 0.05 versus control mice (ANOVA).

FIG. 2. Overexpression of GMF-␤ in mouse PT cells and NIH 3T3 cells.
Shown are the results from immunoblotting of GMF-␤ using cell lysates of two clones of GMF-␤ (GMF0 and GMF1) and of mocktransfected wild-type (WT) PT cells (mProx24) (A) and NIH 3T3 cells (B). After overnight FCS starvation, total cell lysates of the five types of cell lines were immunoblotted with mouse monoclonal anti-FLAG antibody M2 and rabbit polyclonal anti-mouse GMF-␤ antibody. tone (Fig. 5C). As shown in Fig. 5D, the GMF-␤-overexpressing cell line showed ladder formation of DNA under H 2 O 2 stress, whereas much less DNA ladder formation was seen in control cells and in the p38 inhibitor-treated GMF-␤-overexpressing cells. These findings indicate that the expression of GMF-␤ in non-brain cells enhances the effect of oxidative stress and make the cells vulnerable to apoptotic changes governed by the p38 pathway.

F-actin Reorganization by Oxidative Stress Is Enhanced in GMF-␤-overexpressing Cells-A change in F-actin organization
is a well known hallmark of early oxidative cell injury in many cell types (36,37), and this change is also found in renal diseases (22). To further examine the modification of oxidative stress by expression of GMF-␤, we observed cytoplasmic Factin reorganization under oxidative stress in GMF-␤-overexpressing cells. As shown in Fig. 6A, it appears that cytoplasmic F-actin was disrupted and concentrated around the nuclei of wild-type PT cells 15 min after the addition of H 2 O 2 . In cells stably transfected with GMF-␤ (clone GMF0), this change in the organization of actin was more clearly observed, and the size of the cells was decreased by H 2 O 2 treatment. Of interest, F-actin reorganization in GMF-␤ stable transformants was attenuated by the 1-h pretreatment with 10 M SB203580 (Fig.  6B). In the case of the NIH 3T3 cell line, virtually no F-actin reorganization was observed in mock-transfected wild-type cells under oxidative stress for 15 min (Fig. 6C). However, in NIH 3T3 cells stably transfected with GMF-␤, F-actin reorga-nization was clear, and cells contracted significantly, as seen in the case of the PT cell line (Fig. 6D). These results suggest that GMF-␤ overexpression aggravates early oxidative injury represented by the F-actin reorganization under stress in both PT and NIH 3T3 cells.

Increased and Prolonged Intracellular H 2 O 2 Generation in Response to Various Stimuli as a Result of GMF-␤ Overexpres-
sion-Increase in intracellular H 2 O 2 , when converted into hydroxyl radicals, is thought to cause actin modification and cell killing. We therefore used the intracellular H 2 O 2 marker DCFDA to examine the changes in intracellular H 2 O 2 after the addition of H 2 O 2 from outside the cells. Whereas the intracellular H 2 O 2 level in wild-type cells returned to control levels ϳ30 min after reaching its peak at 15 min, the intracellular H 2 O 2 level in GMF-␤-overexpressing PT cells remained higher even 120 min following the addition of 8 mM H 2 O 2 (Fig. 7A). Moreover, the increase in the intracellular H 2 O 2 level in GMF-␤-overexpressing cells was much higher (about twice as high 15 min after H 2 O 2 treatment) than in wild-type cells (Fig. 7B). On the other hand, interestingly, pretreatment with the p38 inhibitor SB203580 was found to abolish the increase in the intracellular H 2 O 2 level following the addition of 8 mM H 2 O 2 (Fig. 7,  A and B). Next, we examined the effects of GMF-␤ expression on intracellular H 2 O 2 generation following treatment with 100 ng/ml TNF-␣ and 1 M angiotensin II, both of which are reported to induce intracellular H 2 O 2 generation and to use it as a second messenger (29,38). As shown in Fig. 7 (C and D), the generation of intracellular H 2 O 2 in GMF-␤ stable transfor- mants was higher and retained longer than in wild-type PT cells. The prolongation of the increase in H 2 O 2 was also observed in NIH 3T3 cells stably expressing GMF-␤ (data not shown). The assays using acetate esters of fluorescent probes, such as DCFDA, are influenced by intracellular esterase activities and cellular efflux. These two factors were checked in the two cell types using the oxidant-insensitive fluorescein diacetate derivative carboxyfluorescein diacetate. There was virtually no difference in fluorescence between GMF0 and control cells at 0 min as well as 120 min after the addition of 8 mM H 2 O 2 (Fig. 7E). These findings indicate that DCFDA assay could detect the intracellular ROS levels equally in wild-type and GMF-␤-expressing cells and that the intracellular esterase activity and cellular efflux were not changed. Collectively, the data obtained here suggest that GMF-␤ overexpression prolongs the generation of intracellular H 2 O 2 under various stimuli, resulting in aggravated oxidative stress. ( Fig. 8A), whereas the protein content and activity of another type of SOD, Mn-SOD, did not change (Fig. 8B). These findings were also confirmed in the case of the NIH 3T3 cell line (Fig. 8,  A and B). The activities of the peroxide-reducing enzymes catalase and GPX and the cellular GSH content were lower in cells with GMF-␤ stable transformants than in mock-transfected wild-type cells in the mouse mProx24 PT cell line and the NIH 3T3 cell line (Fig. 8, C-E). In particular, the GSH content and GPX activities were greatly reduced in the cells overexpressing GMF-␤. These results suggest that GMF-␤-overexpressing cells have a tendency toward aggravated and prolonged increases in intracellular H 2 O 2 .

DISCUSSION
It has been thought that the expression of GMF-␤ in normal tissues is largely limited to the brain (19), especially in the glial cells and some neurons (20). GMF-␤ is reportedly also induced in the developing retina (39). In this study, we showed for the first time that GMF-␤ is also induced in non-neuronal tissues (renal PT cells) by proteinuria (40,41). From the results obtained in Schwann cells (21) and primary astrocytes (18), GMF-␤ was considered to play a protective role in the cells at least in the brain since the overexpression of GMF-␤ results in secretion of neurotrophic factors and induction of its receptor (18,21). In contrast, our study showed that GMF-␤ overexpression made PT and NIH 3T3 cells vulnerable to oxidative stress by the sustained increase in intracellular H 2 O 2 . This unique but critical phenomenon may be pathologi-cally relevant to the progression of various diseases by exaggerating oxidative stress since the overexpression of GMF-␤ has a similar role in at least two different cell lines.
It has recently been well established that ROS play a key role as second messengers in a number of cellular events, for example, cell proliferation, differentiation, and death (29,42,43). ROS (particularly intracellular H 2 O 2 ) generation is associated with TNF-␣-, Fas-, and angiotensin II-mediated apoptosis and various other stimuli (44 -46). In many cell types, including hepatocytes and fibroblasts, oxidative stress produces a severe F-actin reorganization (47)(48)(49)(50), whereas in human intestinal cell monolayers (Caco-2), H 2 O 2 enhances the chemical modification of actin and reorganizes F-actin (51,52). These findings suggest that a sustained increase in intracellular H 2 O 2 in GMF-␤-overexpressing cells may result in their vulnerability to apoptosis and F-actin reorganization. The reorganization of F-actin is reportedly involved in several kidney disease models, whereas ischemic acute renal failure was found to induce F-actin reorganization, leading to loss of surface membrane polarity of tubules (22). Several lines of evidence have shown that apoptosis of epithelial PT cells is relevant to the progression of renal disease in hypertensive nephrosclerosis (24) and focal and segmental glomerulosclerosis (23). These findings suggest that induction of GMF-␤ may have an implication in the progression of renal diseases.
Such a high and prolonged increase in H 2 O 2 resulting from GMF-␤ overexpression led us to hypothesize that GMF-␤ expression causes this change in antioxidant enzyme activity. Lim et al. (53) reported that GMF-␤-overexpressing C6 glioma cells show a 3.5-fold increase in CuZn-SOD protein and dismutase activity. Our study showed that GMF-␤ expression caused not only an increase in the activity of the H 2 O 2 -producing enzyme CuZn-SOD, but also a decrease in the activity of the H 2 O 2 -reducing enzymes catalase and GPX as well as a reduction in cellular GSH content. Many studies have found that the balance between the activities of H 2 O 2 -producing enzymes (CuZn-SOD and Mn-SOD) and those of H 2 O 2 -reducing enzymes (catalase and GPX) governs the sensitivity of cells to oxidative stress (54 -56). Therefore, PT and fibroblast cells stably transfected with GMF-␤ may become sensitive to oxidative stress through an imbalance between CuZn-SOD and catalase/GPX, FIG. 8. Antioxidant enzyme activity and GSH content in mouse GMF-␤-overexpressing PT and NIH 3T3 cell lines. A, CuZn-SOD immunoblotting (upper) and activity (lower) in GMF-␤-overexpressing (GMF) and mock-transfected wild-type (WT) PT (left) and NIH 3T3 (right) cell lines. GMF-␤ stable transformants showed an increase in CuZn-SOD protein expression and activity. B, Mn-SOD immunoblotting (upper) and activity (lower) in GMF-␤-overexpressing and mocktransfected wild-type PT (left) and NIH 3T3 (right) cell lines. Expression of the Mn-SOD protein and its activity were not different in the wild-type cells and GMF-␤ stable transformants. C, catalase activity in GMF-␤-overexpressing and mock-transfected wild-type PT (left) and NIH 3T3 (right) cell lines. Catalase activity decreased in GMF-␤-overexpressing cell lines. D, cellular glutathione content in GMF-␤-overexpressing and mock-transfected wild-type PT (left) and NIH 3T3 (right) cell lines. Glutathione content was reduced in GMF-␤-overexpressing cell lines. E, GPX activity in GMF-␤-overexpressing and mock-transfected wild-type PT (left) and NIH 3T3 (right) cell lines. GPX activity in GMF-␤-overexpressing cell lines was highly reduced compared with that in wild-type cells. Each graph shows the mean Ϯ S.D. (n ϭ 3). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 versus wild-type cells (ANOVA (mProx24 cells) and Student's t test (NIH 3T3 cells)). which causes accumulation of intracellular H 2 O 2 . Back et al. (57) reported that oligodendrocytes show a maturation-dependent vulnerability to oxidative stress because of low activity of the H 2 O 2 -reducing enzyme GPX and low GSH content (58,59).
In GMF-␤-overexpressing cells, DNA ladder formation, annexin V binding, and caspase-3 activation caused by H 2 O 2 treatment were inhibited by the p38 inhibitor SB203580, suggesting that this apoptotic signal is related to the p38 pathway. Moreover F-actin reorganization in GMF-␤-overexpressing cells was attenuated by pretreatment with SB203580. This observation shows that the p38 pathway is also involved in F-actin reorganization under oxidative stress in GMF-␤-overexpressing cells. Finally, DCFDA assay showed that the prolonged accumulation of H 2 O 2 was totally blocked by SB203580 (Fig. 7). These results suggest that p38, which significantly participated in the accumulation of intracellular H 2 O 2 , probably regulates the generation and degradation of ROS upstream of the antioxidant enzymes. It has reported that GMF-␤ is involved in cell signal transduction, particularly activation of the p38 pathway (17,18). However, how p38 enhances the accumulation of H 2 O 2 in GMF-␤-overexpressing cells is largely unknown. The mechanisms of p38 action on ROS accumulation in GMF-␤-overexpressing cells need to be clarified in future studies.
On the basis of our findings reported here, we propose a hypothetical model regarding the intracellular accumulation of H 2 O 2 by induced GMF-␤ (Fig. 9). The brain-specific protein GMF-␤ induced by chronic stress, such as proteinuria, causes a change in the activities of antioxidant enzymes and promotes the accumulation of intracellular H 2 O 2 possibly through the p38 pathway, rendering non-brain cells susceptible to oxidative injuries, leading to F-actin reorganization and apoptosis. Our discovery may direct attention to GMF-␤, which is induced by proteinuria, as a new key factor in the progression of kidney diseases by aggravated oxidative stress.