Hypertonic stress activates glycogen synthase kinase 3beta-mediated apoptosis of renal medullary interstitial cells, suppressing an NFkappaB-driven cyclooxygenase-2-dependent survival pathway.

The survival of renal medullary interstitial cells (RMICs) requires their adaptation to rapid shifts in ambient tonicity normally occurring in the renal medulla. Previous studies determined that cyclooxygenase-2 (COX 2) activation is critical for this adaptation. The present studies find that these adaptive mechanisms are dampened by the simultaneous activation of an apoptotic pathway linked to a glycogen synthase kinase 3beta (GSK 3beta). Inhibition of GSK 3 by LiCl or specific small molecule GSK inhibitors increased RMIC survival following hypertonic stress, and transduction of RMICs with a constitutively active GSK 3beta (AdGSK 3betaA9) significantly increased apoptosis, consistent with a proapoptotic role of GSK 3beta. Following GSK 3beta inhibition, increased survival was accompanied by increased COX 2 expression and COX 2 reporter activity. In contrast, GSK 3beta overexpression reduced COX 2 reporter activity. Importantly, enhanced RMIC survival produced by GSK 3beta inhibition was completely dependent on COX 2 because it was abolished by a COX 2-specific inhibitor, SC58236. The signaling pathway by which GSK 3beta suppresses COX 2 expression was then explored. GSK 3beta inhibition increased both NFkappaB and beta-catenin activity associated with decreased IkappaB and increased beta-catenin levels. The increase in COX 2 following GSK 3beta inhibition was entirely blocked by NFkappaB inhibition using mutant IkappaB adenovirus. However, adenoviral overexpression of beta-catenin did not increase COX 2 levels. These findings suggest that GSK 3beta negatively regulates COX 2 expression and that GSK 3beta inhibitors protect RMICs from hypertonic stress via induction of NFkappaB-COX 2-dependent pathway.

To survive and function normally, renal medullary cells rely on a unique ability to withstand the rapidly shifting osmotic environment present in the renal medulla (2). Failure to adapt to the harsh environment in the renal medulla contributes to the development of papillary necrosis, observed following nonsteroidal anti-inflammatory drug use, pregnancy, and diabetes mellitus (3). Hypertonic stress activates two opposing cellular signaling cascades that either lead to cell death or promote cell survival. The balance between these two pathways determines cell fate. Several mechanisms have been proposed to contribute to the ability of renal medullary cells to survive hypertonic stress, including accumulation of organic osmolytes (4,5) and induction of heat shock proteins (6,7). In addition, recent studies demonstrate that renal medullary interstitial cells (RMICs) 1 depend on robust COX 2 activity to adapt to hypertonic stress, both in vitro and in vivo (2,8,9). Conversely, the mechanisms contributing to hypertonic stress-induced RMIC cell death are less clearly defined.
GSK 3␤ signaling has been implicated in a variety of biological processes associated with altered cell survival and differentiation, including Wnt-associated developmental patterning and ␤-catenin-driven tumorigenesis (10,11). Recent studies demonstrate that the GSK 3␤ inhibitor lithium protects neurons from stress-induced cell death (12)(13)(14)(15)(16)(17). A link between GSK 3␤ activity and neuronal cell survival following hypertonic stress has also been suggested, though the downstream mechanisms of these effects are uncharacterized (18). Some data suggest that COX 2 is a downstream target of Wnt/GSK 3␤/␤-catenin signaling, promoting proliferation and tumorigenesis (19), and previous studies have shown that COX 2 is critical for renal interstitial cell survival. The aim of the present study was to examine whether GSK 3␤ modulates renal medullary interstitial cell survival through a COX 2-dependent mechanism.

MATERIALS AND METHODS
Cell Culture-Rabbit medullary interstitial cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum and 1% (v/v) penicillin (20). Cultures were incubated at 37°C in 95% air/5% CO 2 . Osmolality of the control medium was 330 mosmol/kg. Hyperosmotic medium was prepared by adding NaCl to achieve the indicated tonicity.
Cell Viability Analysis-Cell viability was assessed using crystal violet (21,22). Following hyperosmotic stress for predetermined time periods, culture medium was removed and plates were washed with phosphate-buffered saline. The remaining viable attached cells were stained with 0.5% crystal violet in 50% methanol for 15 min. The plates were then gently rinsed with water and dried. A solution containing 0.1 M citrate sodium in 20% methanol, pH 5.4, was added; 30 min later, the absorbance at 570 nm was read using a spectrophotometer. The percentage of cell survival was defined as the relative crystal violet absorbance of treated versus untreated cells. Remaining unattached cells in the medium were confirmed to be dead by a trypan blue exclusion assay.
Caspase-3/7 Activity-Fluorometric assay of caspase-3/7 activity was conducted using an Apo-ONE TM homogeneous caspase-3/7 assay kit (Promega). This assay uses rhodamine 110, bis(N-CBZ-L-aspartyl-L-glu-tamyl-L-valyl-L-aspartic acid amide) (Z-DEVD-R110) as substrate. Cells were cultured in 12-well plates with 10% fetal bovine serum, which was substituted with serum-free medium for 2 h prior to experiment. Cells were subjected to hyperosmotic stress of varying tonicity for indicated periods of time and lysed using lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 100 M phenylmethylsulfonyl fluoride, 10 g of aprotinin/ml, 10 g of leupeptin/ml, 5100 M phenylmethylsulfonyl fluoride, 5 g/ml of pepstatin, and 0.2% Nonidet P-40). The protein concentration was measured by the bicinoconinic acid method. All measurements of caspase activity were carried out in triplicate in 96-well clear-bottom plate. Buffer and substrate were mixed at a proportion of 100:1 and incubated with cell lysate containing 20 g of protein. Fluorescence was measured using a plate reader (Bio-Tek, Winooski, VT) set at 485 nm excitation and 530 nm emission. The amount of fluorescent product generated is proportional to the amount of caspase-3/7-cleavage activity present in the sample.
Fluorescence-activated Cell Sorter Analysis of Apoptosis-The level of apoptosis was analyzed in living cultures using a combined annexin V-propidium iodide staining using a Vybrant TM apoptosis assay kit (Molecular Probes, Eugene, OR). Briefly, cells were subjected to hypertonic stress in the presence or absence of GSK 3␤ inhibitors LiCl and SB 216763 for 8 h. Cells, including dead and floating cells, were harvested and washed in cold phosphate-buffered saline. 100 l of 1 ϫ 10 6 /ml cells were taken for each assay and stained with PI and annexin V. Negative and positive controls were maintained individually for annexin V and PI. Stained cells were analyzed by flow cytometry (BD FACScan; BD Biosciences) (23,24).
Immunoblotting-Cells were washed with phosphate-buffered saline and lysed in SDS Lamelli buffer. Protein concentration was determined using bicinchoninc acid protein assay (Sigma). Protein extract (20 g) was loaded in each lane of a 10% SDS-PAGE mini-gel and run at 120 V. Protein was transferred to a nitrocellulose membrane at 22 V overnight at 4°C. The membrane was washed three times with TBST (50 mM Tris, pH 7.5, 150 mM NaCl, 0.05% Tween 20) and then incubated in blocking buffer (150 mM NaCl, 50 mM Tris, 0.05% Tween 20, and 5% Carnation™ nonfat dry milk, pH 7.5) for 1 h at room temperature. The membrane was then incubated with anti-mouse antibodies for GSK-3␤ and ␤-catenin (1:2500 and 1:500, respectively; BD Transduction Laboratories) or anti-rabbit antibodies for pGSK (Ser-9) (1:1000; Cell Signaling), IB-␣ (1:1000; Santa Cruz Biotechnology), and COX 2 (1:1000; Cayman) in blocking buffer overnight at 4°C. After being washed three times, membranes were incubated with horseradish peroxidase-conjugated secondary antibody (1:5000 for anti-mouse and 1:10,000 for anti-rabbit) for 1 h at room temperature, followed by three 15-min washings. Antibody labeling was visualized by addition of chemiluminescence reagent (Renaissance; PerkinElmer Life Sciences), and the membrane was exposed to Kodak XAR-5 film.
GSK 3␤ Kinase Activity-Kinase activity of GSK 3␤ immunoprecipitated from cell lysate was analyzed in vitro. After being subjected to hypertonic stress for the indicated periods of time, cells were lysed on ice in freshly prepared buffer (20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 1 mM EDTA, 20 mM NaF, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM sodium vanadate, 5 g of aprotinin/ml, 5 g of leupeptin/ ml, 1 mM phenylmethylsulfonyl fluoride, 1 M microcystin LR). Lysates were clarified by centrifugation for 10 min at 13,000 rpm in a microcentrifuge and precleared with protein G-Sepharose. 100 g of total cellular protein diluted in lysis buffer was immunoprecipitated with 1 g of anti-GSK 3␤ antibody (mouse monoclonal anti-rat GSK 3␤) overnight by rocking incubation at 4°C followed by addition of protein G-Sepharose and incubation for another 1 h. Immunoprecipitates were washed twice with lysis buffer, twice with wash buffer (10 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 0.2 mM sodium vanadate, 1 M microcystin LR), and twice with kinase reaction buffer (20 mM HEPES, pH 7.4, 10 mM MgCl 2 ,1 mM dithiothreitol, 0.2 mM EGTA). Activity of the immunoprecipitated GSK-3␤ was assayed in a total volume of 40 l of kinase buffer containing 3.75 g of phospho-glycogen synthase peptide 2, 15 M cold ATP, and 10 Ci of [ 32 P]ATP. After 20 min of incubation at 30°C, 20 l of SDS lysis buffer was added and incubated at 70°C to stop the reaction. Reaction mixtures were centrifuged, and 15 l of the supernatant was spotted onto Whatman P81 phosphocellulose paper. Filters were washed in three changes of 0.75% phosphoric acid, rinsed in acetone, dried, and 32 p incorporation measured in a liquid scintillation counter. Non-phosphorylated glycogen synthase peptide was used as negative control, and nonspecific 32 p incorporation was subtracted from values obtained using the phospho-glycogen synthase peptide.

FIG. 1. Effect of hypertonicity on GSK 3␤ activity.
A, GSK 3␤ kinase activity. RMICs were subjected to hypertonic stress (550 mosM) for 12 h. Total cell lysates (100 g of protein) were analyzed for GSK 3␤ activity as described. Specific activity of control ϭ 89 pmol/mg protein/ min. *, p Ͻ0.005 versus control. B, pGSK 3␤ and GSK 3␤ expression levels. RMICs were subjected to hypertonic stress (550 mosM) by the addition of NaCl for the indicated periods of time. Total cell lysates (20 g of protein/lane) were analyzed via immunoblot for GSK 3␤ and pGSK 3␤ Ser-9.

FIG. 2.
A, effect of GSK 3␤ inhibitor LiCl on RMIC survival following hypertonic stress. RMICs were cultured to confluence and shifted to media with the indicated tonicity. Experiments were carried out in the presence (LiCLϩ) or absence (LiClϪ) of LiCl. Following 12 h of hypertonic stress, cell survival was assessed using a crystal violet assay as described under ''Materials and Methods.'' *, p Ͻ0.00; **, p Ͻ0.0001 versus individual control. B, LiCl treatment increases pGSK 3␤ levels. Cultured RMICs were treated with LiCl (30 mM) for 6 and 12 h. Total cell lysates (20 g of protein) were immunoblotted for GSK 3␤ and pGSK 3␤ Ser-9.
Reporter Gene Assay: ␤-Catenin-TCF/LEF, NFB, and COX 2-␤-Catenin-TCF/LEF reporter plasmid activity (generously provided by Drs. Kinzler and Volgelstein) (25) was determined using 12-well culture plates seeded at a density of 2 ϫ 10 5 RMICs/well. Following 24 h of growth, cells were transfected with ␤-catenin-TCF/LEF reporter and a plasmid containing Renilla luciferase driven by TK promoter using SuperFect (Qiagen). Forty-eight h post-transfection, cells were subjected to hypertonic stress in the presence or absence of GSK-3 inhibitors. After the indicated incubation time, cells were lysed and luciferase activity measured using the dual luciferase assay system (Promega Corp., Madison, WI).
NFB reporter plasmid containing firefly luciferase linked to 3 NFB consensus sequences was a gift from Dr. Timothy Blackwell (26). COX 2 reporter plasmid was obtained from Dr. Hiroyasu Inoue (27). RMICs were seeded and transiently transfected as mentioned above. In some studies, cells were co-transfected with mouse GSK 3␤ cDNA. The mouse GSK 3␤ cDNA expressed sequence tag clone was obtained from Invitrogen. GSK 3␤ sequence was confirmed by complete nucleotide sequencing and subcloned into expression vector pCDNA 3.
Transduction of RMICs with Recombinant GSK 3␤A9, Enhanced Green Fluorescent Protein, and IB(mut) Adenovirus-AdGSK 3␤A9, carrying a Ser-to-Ala substitution at Ser-9 in the NH 2 -terminal region of GSK 3␤ and a HA epitope tag at the COOH terminus was a gift from Dr. Thomas Force (28). Ad␤-catenin (Myc-tagged) was a gift from Dr. Howard C. Crawford (19,29). The recombinant viruses were propagated in HEK 293 cells, and high titer stocks (ϳ2 ϫ 10 10 particles/ml) were purified by CsCl density gradient centrifugation. For infection of RMICs, virus of 30 -100 multiplicity of infection was added to each culture dish.
AdGFP was prepared as previously described (4). A full-length green fluorescent protein cDNA (Green Lantern or GL; Invitrogen) was subcloned into pACCMV. The pACCMV shuttle plasmid contains the cytomegalovirus immediate early enhancer and promoter and the SV40 polyadenylation sequence. The resulting shuttle plasmids were cotransfected into HEK 293 cells along with the pJM17 vector by Super-Fect (Qiagen). AdGFP was generated by a homologous recombination event, resulting in plaque formation in the HEK 293 cells. The resulting infectious adenovirus was plaque-purified. 100 multiplicity of infection was added to each RMIC culture dish. After a 2-h incubation, the virus was removed and fresh Dulbecco's modified Eagle's medium with 10% fetal bovine serum added. Experiments were carried out 48 -72 h after infection AdIB(mut) was constructed, amplified, and purified as described previously (20,30). Briefly, the transdominant inhibitor of NFB, avian IB-␣S36/40A, was constructed by oligonucleotide site-directed mutagenesis and subcloned into a shuttle plasmid pACCMV (20,31). The infectious AdIB was generated by co-transfecting with pJM17 as described above.
Flow cytometric analysis of relative levels of annexin V-PI staining of cells showed that within 8 h of hypertonic stress, 49% of cells become apoptotic compared with control (18%), with 32% cells in the early apoptotic stage and 10 and 7%, respectively, in the late apoptotic and necrotic stages. However, the apoptotic effects could be considerably decreased by simultaneously treating the cells with GSK 3␤ inhibitors (LiCl, 24%, and SB 216763, 19%) (Fig. 3B). These studies are consistent with a role for GSK 3␤ activity in promoting cell death following hypertonic stress.
To examine whether protective effects of GSK 3␤ inhibitors on RMIC survival depend on COX 2 activity, the effect of a COX 2-selective inhibitor (SC58236, 10 M) (20) was examined. Treatment of cells with SC58236 completely abolished the protective effects of GSK 3␤ inhibitors Li or the SB compounds (Fig. 5) following hypertonic stress. This concentration of COX 2 inhibitor did not have any effect on RMIC survival in nonstressed cells.
GSK 3␤ Activity Modulates COX 2 Expression via an NFBdependent Pathway-Activation of NFB contributes to increased COX 2 expression following stress (2,33). Recent studies indicate that NFB activity may be regulated by GSK 3␤ (34,35). To test whether COX 2 induction by GSK 3␤ inhibitors required NFB, RMICs were transduced with an adenovirus encoding mutant dominant negative inhibitor of NFB (IB). The IB(mut) reduced COX 2 expression in cells treated with the GSK 3␤ inhibitor SB216763 (Fig. 6A). These findings support a role for NFB in the COX 2 induction by GSK 3 inhibitors.
GSK 3␤ is a highly conserved protein kinase thought to be constitutively active in differentiated cells (1). It is an important component of the Wnt signaling pathway, and its inhibition by Wnt signaling is believed to play a critical role in cell proliferation during embryogenesis. Constitutive GSK 3␤ activity can be suppressed by a variety of stimuli, including Wnt ligands, insulin, epidermal growth factors, and fibroblast growth factor (10), mainly via phosphorylation of the NH 2terminal serine 9 residue (38). Conversely, GSK 3␤ may be activated by stress, including hypoxia in A7r5 cells (39) or potassium deprivation of sensory neurons (40). Altered GSK 3␤ activity has been implicated in several human diseases, including cancer, diabetes mellitus, and Alzheimer's (1).
The present studies provide evidence suggesting a role for GSK 3␤ activation regulating renal medullary interstitial cell survival following hypertonic stress in renal medullary cells. Hypertonic stress increased GSK 3␤ kinase activity and decreased levels of the inactive phospho-GSK 3␤ (Ser-9) in a time-dependent manner. Whether reduced activity of inhibitory kinases (e.g. protein kinase B), which normally phosphorylate and inactivate GSK 3␤, or higher phosphatase activity, or both contribute to this activation of GSK 3␤ remains unexplored.
Three distinct GSK 3 inhibitors, LiCl, SB216763, and SB415286, reduced RMIC apoptosis following hypertonic stress as indicated by caspase 3/7 activity and annexin V-PI staining. The concordance of results obtained with LiCl and selective small molecule ATP-competitive inhibitors of GSK 3 activity, including SB415286 and SB216763, is important because the latter compounds are structurally distinct and highly selective for GSK 3 (32). Furthermore, the potency of these GSK 3␤selective inhibitors in preventing RMIC death correlated well with their relative in vitro potency of inhibiting GSK 3␤, with SB216763 being more potent than SB415286 (40). Pro-apoptotic effects of GSK 3␤ in RMICs agree with previous observations, showing GSK 3␤ activation to be involved in neuronal cell death following hypoxia and serum deprivation (17,39).
The mechanism by which activated GSK 3␤ promotes cell death is poorly understood. Previous studies have established that COX 2-derived prostanoids play a critical role in RMIC survival (2,43). Of relevance to the present study are previous observations that LiCl induces COX 2 expression in mouse mammary epithelial cells (44). The present studies demonstrate that GSK 3 inhibition not only induces COX 2 expression but also protects RMICs from death following hypertonic stress. COX 2 activity is crucial for RMIC survival because the COX 2-selective inhibitor SC58236 completely abolished the protective effect of GSK 3␤ inhibitors. These findings provide a functional correlate to the induction of COX 2 by GSK 3␤ inhibitors and link COX 2 function to cytoprotection afforded by GSK 3␤ inhibitors. COX 2 expression can be induced via multiple mechanisms, including increased mRNA transcriptional rates, mRNA stabi- lization, and altered COX 2 protein stability (45). The fact that GSK 3␤ inhibitors increased COX 2 reporter activity in the present study suggests an important role for increased COX 2 transcription. GSK 3␤ can phosphorylate and modulate activity of several transcriptional systems, including ␤-catenin, c-Jun, c-Myc, IRS-1, C/EBP, and cAMP-response element-binding protein (46). Although, GSK 3␤ inactivation by Wnt/␤catenin signaling has been implicated as a means of inducing COX 2 expression (44,47), this pathway does not seem critical for COX 2 induction in RMICs. Though GSK inhibition increased ␤-catenin expression and TCF/LEF reporter activity in RMICs, ␤-catenin overexpression failed to increase COX 2 lev-els. Thus, the GSK 3␤/␤-catenin pathway is unlikely to play a significant role in modulating RMIC COX 2 expression following osmotic stress. In contrast, the present evidence supports a major role for NFB in mediating COX 2 up-regulation by GSK.
NFB activation occurs in response to a variety of cellular stresses, and inhibition of NFB can lead to apoptosis (47,48). Importantly, previous studies demonstrated that hypertonic stress activates an NFB-COX 2-linked survival mechanism in RMICs (2). In its inactive form, NFB exists as a cytosolic complex bound to the inhibitory protein IB (49). Phosphorylation of IB by a specific kinase (IKK) leads to ubiquitin/ATPdependent proteasomal degradation of IB, releasing p65/rel NFB subunits for translocation to the nucleus and transactivation of target genes (50). The functional relationship between NFB and GSK 3␤ remains controversial, with some reports suggesting that GSK 3␤ activity is necessary for NFB activation (35,51) and others showing that GSK 3␤ activity inhibits NFB (34,52). Genetic disruption of GSK 3␤ did not affect the early steps of NFB activation (degradation of IB and nuclear translocation of NFB) in embryonic fibroblasts (35). Nevertheless, GSK 3␤ was recently found to phosphorylate the COOH terminus of p65 subunit of NFB in vitro, which could increase NFB-mediated gene transactivation (51). Conversely, an inhibitory effect of GSK 3␤ on NFB was shown in PC12 cells where Wnt-1-induced activation of NFB could be mimicked by inhibition of GSK 3␤ (34). Similar observations were made in astrocytes in which overexpression of constitutively active GSK 3␤ stabilized IB and reduced IKK activity (52). The present studies demonstrate that inhibition of GSK 3␤ activates NFB FIG. 7. A, ␤-catenin protein levels. Cultured RMICs were subjected to hypertonic stress (550 mosM) for 12 h. Total cell lysates (20 g of protein/lane) were immunoblotted for ␤-catenin. B, ␤-catenin protein levels in the presence of GSK 3␤ inhibitors. RMICs were treated with SB216763 or SB415286 for 8 h. Total cell lysates (20 g of protein/lane) were immunoblotted for ␤-catenin. C, TCF/LEF activity. RMICs were co-transfected with TCF/LEF-driven luciferase vector and TK driving Renilla luciferase plasmid. 24 h later, cells were subjected to hypertonic stress (500 and 550 mosM). 12 h later, luciferase activities were tested as described. ␤-Catenin reporter activity was normalized by Renilla luciferase activity. *, p Ͻ0.005 versus control (300 mosM). D, TCF/LEF activity in the presence of LiCl. Cultured RMICs were co-transfected with ␤-catenin-TCF/LEF-driven luciferase vector and TK driving Renilla luciferase plasmid. 24 h later, cells were treated with LiCl (30 MM). 12 h later, luciferase activities were tested as described. ␤-Catenin reporter activity was normalized to Renilla luciferase activity. *, p Ͻ0.01 versus control. E, COX 2 protein levels in cells overexpressing ␤-catenin. Cultured RMICs were transduced with Ad␤-catenin (75 multiplicity of infection) as described. After 72 h, cells were lysed and immunoblotted for COX 2. F, TCF/ LEF activity in cells overexpressing ␤-catenin. Cultured RMICs were transfected with TCF/LEF-driven luciferase vector and TK driving Renilla luciferase plasmid. 24 h later, they were transduced with Ad␤-catenin. 32 h later, luciferase activities were tested as described. TCF/ LEF reporter activity was normalized to Renilla luciferase activity. *, p Ͻ0.001 versus control. and COX 2 expression, consistent with findings that GSK 3␤ decreases NFB activity.
The use of GSK 3 inhibitors, including lithium, has been proposed as a therapeutic intervention in a variety of diseases, including diabetes mellitus, neurodegeneration, cancer, and inflammation (53). Lithium has been widely used for the treatment of bipolar disorders and Alzheimer's disease (54). Its use in this setting is associated with reproducible renal side effects, including diabetes insipidus and chronic renal tubulointerstitial nephropathy (41). Though the cellular basis of these renal effects are uncertain, it is of interest that urinary prostaglandin E2 excretion is increased in rats receiving Li (42), consistent with up-regulation of renal cyclooxygenase by lithium. Whether increased renal PGE2 production occurs as a result of increased COX 2 expression following Li treatment remains to be determined. The present findings are consistent with this possibility. Furthermore, increased renal interstitial cell number because of enhanced cell survival could contribute to renal tubulo-interstitial cell infiltrates observed in the setting of chronic lithium exposure (41). The concordance between the effects of lithium and other GSK 3␤ inhibitors on renal interstitial cell function suggests that newer pharmacologic inhibitors of GSK 3␤ could have similar renal effects as lithium.