Inactivation of Integrin-linked Kinase Induces Aberrant Tau Phosphorylation via Sustained Activation of Glycogen Synthase Kinase 3 (cid:1) in N1E-115 Neuroblastoma Cells*

Integrin-linked kinase (ILK) is a focal adhesion serine/ threonine protein kinase with an important role in integrin and growth factor signaling pathways. Recently, we demonstrated that ILK is expressed in N1E-115 neuroblastoma cells and controls integrin-dependent neurite outgrowth in serum-starved cells grown on laminin (Ishii, T., Satoh, E., and Nishimura, M. (2001) J. Biol. Chem. 276, 42994–43003). Here we report that ILK controls tau phosphorylation via regulation of glycogen synthase kinase-3 (cid:1) (GSK-3 (cid:1) ) activity in N1E-115 cells. Stable transfection of a kinase-deficient ILK mutant (DN-ILK) resulted in aberrant tau phosphorylation in N1E-115 cells at sites recognized by the Tau-1 antibody that are identical to some of the phosphorylation sites in paired helical filaments, PHF-tau, in brains of patients with Alzheimer’s disease. The tau phosphorylation levels in the DN-ILK-expressing cells are constant under normal and differentiating conditions. On the other hand, aberrant tau phosphorylation was not observed in the parental control cells. ILK inactivation resulted in an increase in the active form but a decrease in the

Tau is a microtubule-associated protein that stabilizes microtubules within neurites and axons (1). It is hypothesized that tau hyperphosphorylation leads to the destabilization of microtubules and aggregation of tau proteins, which impairs axonal transport and eventually results in neuronal cell death (1)(2)(3). Indeed, tau hyperphosphorylation appears to be an early event preceding the formation of paired helical filaments (PHF) 1 in the brains of Alzheimer's disease patients (4). On the other hand, tau phosphorylation also seems to control microtubule dynamics during neurite outgrowth and neuronal maturation, because embryonic and neonatal tau is much more heavily phosphorylated than adult tau (5)(6)(7)(8). Thus, the study of the regulation of tau phosphorylation in neurons is important for understanding the neurofibrillary degeneration in Alzheimer's disease, as well as the physiologic function of tau in neurite outgrowth and neuronal development.
Integrin-linked kinase (ILK) is a cytoplasmic serine/threonine kinase protein that serves as a mediator in integrin-and growth factor-mediated signal transduction (9,10). ILK is expressed in N1E-115 cells and controls integrin-dependent neurite outgrowth (11). In the present study, stable transfection of a kinase-deficient mutant of ILK (DN-ILK), which behaves as a dominant negative and inactivates endogenous ILK, resulted in aberrant tau phosphorylation. DN-ILK overexpression increased the active form and decreased the inactive form of GSK-3␤. On the other hand, a selective uncompetitive inhibitor of GSK-3␤, lithium, inhibited aberrant tau hyperphosphorylation. These results suggest that inactivation of ILK results in sustained activation of GSK-3␤ and leads to aberrant tau phosphorylation in N1E-115 neuroblastoma cells.

EXPERIMENTAL PROCEDURES
Reagents-LY294002 was obtained from Sigma. Rabbit polyclonal anti-ILK IgG (UB 06 -550 and UB 06 -592) and myelin basic protein were obtained from Upstate Biotechnology, Inc. (Lake Placid, NY). lines were isolated over the next 4 to 5 weeks. Among them, three different cell lines were selected based on the detection of green fluorescent protein fluorescence and confirmation of gene transcription using reverse transcriptase PCR. The cloned cell lines were maintained in Dulbecco's modified Eagle's medium containing 20% FBS and Zeocin (0.5 mg/ml).
ILK Assay-ILK assay was performed as described by Delcommenne et al. (13). Cells were lysed in 50 mM Hepes buffer (pH 7.5) containing 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 10 g/ml leupeptin, 2.5 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, 5 mM sodium fluoride, and 1 mM sodium orthovanadate. The lysates were incubated with anti-ILK antibody (UB 06 -592) at 4°C for 12 h. After incubation, the lysates were precleared, and immune complexes were collected with protein A-Sepharose. After washing twice with lysis buffer and once with 50 mM HEPES (pH 7.0) buffer containing 1 mM EDTA, the immunoprecipitated ILK was incubated for 20 min at 30°C in the presence or absence of 10 g of the exogenous substrate myelin basic protein in a total volume of 50 l kinase reaction buffer (50 mM HEPES (pH 7.0), 10 mM MnCl 2 , 10 mM MgCl 2 , 2 mM NaF, 1 mM Na 3 VO 4 ) containing 6 M [␥-32 P]ATP (10 Ci; NEG-502Z; PerkinElmer Life Sciences). The reaction was stopped by the addition of an equal volume of 2ϫ SDS-PAGE sample buffer. The kinase reaction products were analyzed using SDS-PAGE (5-20% polyacrylamide) and autoradiography. For detection of the immunoprecipitated ILK and DN-ILK proteins, the precipitated proteins were released from the immunobeads by boiling in 80 l of SDS-PAGE sample buffer for 5 min. Equal volumes of the samples were loaded onto SDS-PAGE. Total ILK and DN-ILK proteins were detected by immunoblotting with an anti-ILK antibody (UB 06 -550) that recognized both ILK and DN-ILK proteins.

DN-ILK Inhibits Basal ILK Activity and Prevents Stimulation of ILK Activity after Cell Adhesion on Laminin under
Serum-free Conditions-Mouse N1E-115 neuroblastoma cells grown on a laminin matrix exhibit neurite outgrowth in response to serum deprivation (11,17). We have demonstrated previously (11) that ILK is expressed in N1E-115 neuroblastoma cells and controls integrin-dependent neurite outgrowth in serum-starved cells grown on laminin. To inactivate endogenous ILK, cells were stably transfected with DN-ILK, which behaves as a dominant negative (11,13). Based on the results obtained from immunoblotting ( Fig. 1), the expression level of DN-ILK protein in DN-ILK-transfected cells was estimated to be at least twice that of endogenous ILK protein, and neither the ILK nor DN-ILK expression level changed under differentiating conditions. The ILK activity in the parental cells under serum-free conditions was transiently activated after seeding on the laminin matrix, whereas that in the DN-ILK-transfected cells was not. Also, weak basal ILK activity was detected only in the parental cells under non-differentiating conditions (Fig.  1). These findings are consistent with our previous observations (11). Thus, DN-ILK inactivates endogenous ILK under both differentiating and non-differentiating conditions. For the ILK assay and detection of ILK and DN-ILK proteins, cells were cultured for 60 min, which was required for maximal activation of ILK activity (11), under either differentiating or non-differentiating conditions and then lysed. ILK was immunoprecipitated from cell extracts. ILK activity was determined using myelin basic protein as an exogenous substrate, as described under "Experimental Procedures." ILK and DN-ILK expression levels in cell lysates (20 g) were analyzed using immunoblotting with an affinity-purified polyclonal anti-ILK antibody.
Inactivation of Endogenous ILK Induces Aberrant Tau Phosphorylation-To examine whether tau phosphorylation is involved in neurite outgrowth in N1E-115 cells, we analyzed the tau phosphorylation level in parental and DN-ILK-transfected cells using phosphorylation-and dephosphorylation-dependent anti-Tau antibodies. Western blots of tau from both parental and DN-ILK-transfected cells were probed with three different tau antibodies, anti-Tau-1, which recognizes an epitope of tau only when it is not phosphorylated, antiphospho (Ser 199 ,Ser 202 )-Tau, which recognizes tau phosphorylated at Ser 199 and Ser 202 , and anti-Tau, which recognizes both native and phosphorylated forms of tau. In non-transfected parental cells, tau was recognized by anti-Tau-1 but not antiphospho(Ser 199 ,Ser 202 )-Tau antibody, under both normal and differentiating conditions. Thus, tau in the parental cells was not phosphorylated at sites recognized by those antibodies in either condition. On the other hand, cells stably transfected with a DN-ILK to inactivate the endogenous ILK had dramatically decreased anti-Tau-1 immunoreactivity but an increased anti-phospho(Ser 199 ,Ser 202 )-Tau immunoreactivity (Fig. 1). Thus, tau was phosphorylated in DN-ILK-transfected cells, and the tau phosphorylation level did not change even under differentiating conditions. Total tau detected with anti-Tau antibody migrated as several bands in the 40-to 70-kDa range, but some of the protein bands, which migrate relatively slower on SDS-PAGE, were weaker when tau was phosphorylated (Fig. 1). These results suggest that inactivation of endogenous ILK results in aberrant hyperphosphorylation of tau, at least at Ser 199 and Ser 202 .
Immunofluorescent Staining of Aberrantly Phosphorylated Tau-To examine the intracellular localization of aberrantly phosphorylated tau, cells were stained with antibody against phosphorylated tau. Immunofluorescent staining of cell monolayers with anti-phospho(Ser 199 ,Ser 202 )-Tau antibody is shown in Fig. 2. DN-ILK-transfected cells were strongly stained with the antibody against phosphorylated tau under normal and differentiating conditions. On the other hand, parental cells were not significantly stained under differentiating conditions, but very weak dot-like structures were observed only in the non-differentiated cells, suggesting that a small minority of the tau, which could not be detected by the immunoblotting analysis, might be phosphorylated under non-differentiating conditions. In the DN-ILK-transfected cells, intracellular cytoplasm, except the nucleus, was stained with strong immunofluorescent intensity, and microtubule-like structures were observed. Further analysis of the DN-ILK-transfected cells using confocal laser scanning microscopy indicated that microtubule-like structures spread and cover right under the whole plasma membrane of the cells and form basket-like structures (Fig. 2,  top, left and right).
Active Form of GSK-3␤ Increases in DN-ILK-transfected Cells-To examine signal pathways involved in tau phosphorylation, we analyzed the activation status of GSK-3␤, because GSK-3␤ is one of the candidate kinases that can phosphorylate tau at both Ser 199 and Ser 202 (18) and also has an important role in the ILK-mediated signal pathway (13,19,20). Activation of GSK-3␤ is dependent on Tyr 216 phosphorylation (21). On the other hand, GSK-3␤ activity is inhibited by direct phosphorylation at Ser 9 by ILK (20,22) and by protein kinase B/Akt, which is also activated via ILK (13,23). We therefore examined whether the levels of these phosphorylated GSK-3␤ forms are different between parental and DN-ILK-transfected cells and are changed under different culture conditions by Western blot analysis using anti-phospho(Ser 9 )-GSK-3␤ and anti-phospho-(Tyr 279 / 216 )-GSK-3␣/␤ antibodies. Tyr 216 in GSK-3␤ was highly phosphorylated in DN-ILK-transfected cells but was very weakly phosphorylated in parental cells (Fig. 3A). In contrast, Ser 9 in GSK-3␤ was highly phosphorylated in parental cells but not in DN-ILK-transfected cells. Moreover, these phosphorylation levels were not significantly different between non-differentiating and differentiating conditions. The expression level of GSK-3␤ was not different among all cells tested. These results suggest that ILK inactivation results in Ser 9 dephosphorylation and increased Tyr 216 phosphorylation in GSK-3␤, thereby activating the enzyme.
GSK-3␤ Activity Is Activated by Inactivation of Endogenous ILK in DN-ILK-transfected Cells-To test whether the level of GSK-3␤ activity increases in DN-ILK-transfected cells, we performed a GSK-3␤ kinase assay on proteins immunoprecipitated from cell lysates with a monoclonal GSK-3␤ antibody. The level of GSK-3␤ activity in DN-ILK-transfected cells was approximately three times more than that in parental cells (Fig. 3B). The level of GSK-3␤ activity was not different between differentiating and non-differentiating conditions in either DN-ILKtransfected or parental cells (Fig. 3B). Thus, inactivation of endogenous ILK results in increased GSK-3␤ activity. This result is consistent with the fact that ILK inactivation induced an increase in the levels of active GSK-3␤ formed via Tyr 216 phosphorylation (Fig. 3A).
Lithium Reduces Tau Phosphorylation in a Dose-dependent Manner-Lithium is an uncompetitive GSK-3␤ inhibitor (24). We therefore examined the effect of lithium on tau phosphorylation. Cells were treated with varying concentrations of LiCl for 16 h under non-differentiating conditions. Treatment of DN-ILK-transfected cells with LiCl reduced antiphospho(Ser 199 ,Ser 202 )-Tau immunoreactivity but increased anti-Tau-1 immunoreactivity in a dose-dependent manner (Fig.  4). Moreover, reaction of slowly migrated tau bands with anti-Tau, which were supposed to be non-phosphorylated tau, was recovered after treatment with LiCl in a dose-dependent manner. The same results were obtained in DN-ILK-transfected cells under differentiating conditions (data not shown). These results suggest that LiCl inhibited tau phosphorylation at Ser 199 and Ser 202 and also at the sites recognized by anti-Tau-1. On the other hand, the same treatment of parental cells with 25 mM LiCl did not affect either anti-Tau-1 or antiphospho(Ser 199 ,Ser 202 )-Tau immunoreactivity. These results suggest that GSK-3␤ activation induced by ILK inactivation is directly involved in tau phosphorylation at Ser 199 and Ser 202 and also at the sites recognized by anti-Tau-1.

Aberrant Tau Phosphorylation Is Partly Involved in Inhibition of Neurite Outgrowth in DN-ILK-transfected Cells-We
demonstrated previously (11) that activation of p38 mitogenactivated protein (MAP) kinase is involved in ILK-mediated signal transduction leading to integrin-dependent neurite outgrowth in N1E-115 cells. In our previous report (11), we suggested that signaling pathways other than p38 MAP kinase, which is also activated via ILK activation, might be involved in integrin-dependent neurite outgrowth, because a p38 MAP kinase inhibitor, SB203580, blocked ϳ80% of the ILK-dependent neurite outgrowth (see Fig. 5). To examine whether aberrant tau phosphorylation is involved in the inhibition of neurite outgrowth in DN-ILK-transfected cells, we treated the cells with 10 mM LiCl, a dose that prevents aberrant tau phosphorylation, under serum-free conditions. Treatment of the DN-ILK-transfected cells with LiCl partially recovered neurite outgrowth, in the presence or absence of SB203580, to the levels of that in the SB20358-treated parental cells (Fig. 5). On the other hand, treatment of the parental cells with LiCl did not affect neurite outgrowth in either the presence or absence of SB203580 (Fig. 5). These results suggest that aberrant tau phosphorylation is partly involved in the inhibition of neurite outgrowth in DN-ILK-transfected cells. DISCUSSION The present study demonstrates that ILK inactivation induces aberrant tau phosphorylation via GSK-3␤ activation in N1E-115 cells, which is partly involved in the inhibition of neurite outgrowth, based on the following observations: 1) tau was phosphorylated at Ser 199 and Ser 202 and at the sites recognized by anti-Tau-1 in DN-ILK transfected cells but not in parental cells, 2) ILK inactivation increased the active form but decreased the inactive form of GSK-3␤, leading to increased GSK-3␤ activity, and 3) treatment of the DN-ILK-transfected cells with LiCl, an uncompetitive inhibitor of GSK-3␤, blocked aberrant tau phosphorylation and partly recovered the levels of neurite outgrowth in DN-ILK-transfected cells.
Integrin signaling is required for neurogenesis in serumstarved N1E-115 cells (25). N1E-115 cells grown on a laminin matrix exhibit neurite outgrowth in response to serum deprivation (11,17). This neurite outgrowth depends on an integrindependent signal pathway, because pre-treatment of cells with anti-␤1 integrin antibody inhibited neurite outgrowth (11,17). We demonstrated recently (11) that ILK is expressed in N1E-115 cells and controls integrin-dependent neurite outgrowth in serum-starved cells grown on laminin. In the present study, stable transfection of DN-ILK, which behaves as a dominant negative and inactivates endogenous ILK, resulted in aberrant tau phosphorylation. On the other hand, the aberrant tau phosphorylation was not observed in the parental cells under either normal or differentiating conditions. ILK activity in the parental cells was increased after seeding on the laminin matrix under serum-free conditions, whereas that in the DN-ILK- transfected cells was not. Thus, both cell adhesion to laminin and serum deprivation were necessary for full activation of ILK. On the other hand, weak basal ILK activity was detected in the parental cells under non-differentiating normal conditions (Fig. 1). It remains unknown how basal ILK activity is maintained under non-differentiation. These results suggest that endogenous ILK prevents aberrant tau phosphorylation. Moreover, aberrant tau phosphorylation did not significantly occur in the parental cells under normal conditions, suggesting that weak basal ILK activity is sufficient to protect tau from aberrant phosphorylation.
To examine whether aberrantly phosphorylated tau participates in microtubule formation, we stained cells with antibody against phosphorylated tau (Fig. 2). Intracellular cytoplasm, except for the nucleus, was stained with strong immunofluorescent intensity, and microtubule-like structures were observed only in DN-ILK-transfected cells. Moreover, analysis of the DN-ILK-transfected cells using confocal laser scanning microscopy indicated that microtubule-like structures spread and cover right under the whole plasma membrane of the cells and form basket-like structures (Fig. 2, top, left and right). Thus, aberrantly phosphorylated tau participates in microtubule-like structures but is located only immediately under the plasma membrane in the cytosol without being able to form neurites in DN-ILK-transfected cells. On the other hand, parental control cells under non-differentiating conditions were very weakly stained with antibodies against phosphorylated tau, which could not be detected by the immunoblotting analysis, but neurite-bearing control cells under differentiating conditions were not (Fig. 2, a and c). These results suggest that tau phosphorylation at Ser 199 and Ser 202 might negatively control microtubule rearrangement necessary for neurite outgrowth via microtubule instability.
Tau is phosphorylated at Ser 199 and Ser 202 and at the sites recognized by anti-Tau-1, which correspond to Ser 195 , Ser 198 , Ser 199 , Ser 202 , and Thr 205 . Ser 199 and Ser 202 are phosphorylated by GSK-3␤ (18), which is a candidate kinase involved in PHF-tau formation (26,27). Moreover, GSK-3␤ has an important role in the ILK-mediated signal pathway (10,19). We therefore examined the involvement of GSK-3␤ in aberrant tau phosphorylation in DN-ILK-transfected cells. Activation of GSK-3␤ is dependent upon the Tyr 216 phosphorylation (21). On the other hand, GSK-3␤ activity is inhibited by direct Ser 9 phosphorylation by ILK (19,22) and by protein kinase B/Akt whose activity is also activated via ILK (13,23). Tyr 216 in GSK-3␤ was highly phosphorylated in DN-ILK-transfected cells but was very weakly phosphorylated in parental cells. In contrast, Ser 9 in GSK-3␤ was highly phosphorylated in parental cells but not in DN-ILK-transfected cells. These phosphorylation levels were not significantly different between nondifferentiating and differentiating conditions (Fig. 3A). Thus, GSK-3␤ was phosphorylated at both Ser 9 and Tyr 216 in parental cells, whereas Tyr 216 phosphorylation was considerably lower (Fig. 3A). Recently, Bhat et al. (28) suggested that Ser 9 phosphorylation is sufficient to override the Tyr 216 phosphorylation-induced activation of GSK-3␤. Therefore, the level of GSK-3␤ activity seems to maintain lower via Ser 9 phosphorylation in parental cells. These results suggest that ILK inactivates GSK-3␤ via phosphorylation at Ser 9 and prevents activation. In contrast, ILK inactivation results in Ser 9 dephosphorylation and increased Tyr 216 phosphorylation in GSK-3␤, thereby activating the enzyme. Indeed, the level of GSK-3␤ activity in DN-ILK-transfected cells was significantly higher than that in parental cells (Fig. 3B).
Lithium, a selective uncompetitive inhibitor of GSK-3␤ (24), inhibited tau phosphorylation at Ser 199 and Ser 202 and also at the sites recognized by anti-Tau-1 in a dose-dependent manner (Fig. 4). These results suggest that GSK-3␤ activation induced by ILK inactivation is directly involved in tau phosphorylation at Ser 199 and Ser 202 and also at the sites recognized by anti-Tau-1. As both Ser 199 and Ser 202 are phosphorylated by GSK-3␤ (18), GSK-3␤ at least directly phosphorylates these Ser residues in the DN-ILK-transfected cells. Although a specific tyrosine kinase, which should be activated by ILK inactivation, is probably involved in the Tyr 216 phosphorylation in GSK-3␤, we have not yet determined the kinase involved in this study. To understand the ILK-mediated regulatory mechanisms of GSK-3␤, the specific tyrosine kinase that is activated by ILK inactivation and phosphorylates Tyr 216 in GSK-3␤ must be determined. We demonstrated previously (11) that p38 MAP kinase in the ILK-mediated signal pathway has an important role in integrindependent neurite outgrowth in N1E-115 cells. On the other hand, a p38 MAP kinase inhibitor, SB203580, blocked ϳ80% of the ILK-dependent neurite outgrowth but not to the levels of the DN-ILK-transfected cells (see Fig. 5). In the previous study (11), we suggested that signaling pathways other than p38 MAP kinase, which are also activated via ILK activation, are involved in integrin-dependent neurite outgrowth. We therefore examined whether aberrant tau phosphorylation is involved in the inhibition of neurite outgrowth in DN-ILK-transfected cells. Treatment of the DN-ILK-transfected cells with 10 mM LiCl completely prevented aberrant tau phosphorylation but only partially recovered neurite outgrowth to the levels of that in the SB20358-treated parental cells. These results suggest that aberrant tau phosphorylation is partly involved in the inhibition of neurite outgrowth in DN-ILK-transfected cells. Furthermore, the results also suggest that a p38 MAP kinase pathway is the sole downstream pathway in ILK-dependent neurite outgrowth.
Tau hyperphosphorylation decreases the association of tau with microtubules (29) and inhibits total neurite number (30 -32). The inhibitory effect of aberrant tau phosphorylation on neurite outgrowth was as expected. The effect of aberrant tau phosphorylation, however, is considered to depend on the tau phosphorylation level and also on the number of phosphorylated tau molecules. Although we could not estimate how many molecules in total tau are phosphorylated in the DN-ILK-transfected cells in this study, it is possible that the aberrant tau phosphorylation induced by ILK inactivation at least affects microtubule stability or dynamics and leads to inhibition of neurite outgrowth. On the other hand, endogenous ILK protects tau from aberrant phosphorylation and probably maintains a kind of equilibrium status responsible for microtubule reorganization. Thus, ILK is not only involved in p38 MAP kinase activation but might also control microtubule dynamics via regulation of GSK-3␤ activity during neurite outgrowth in N1E-115 cells (Fig. 6). A recent study (33) of nerve growth factor-induced neurite outgrowth using pheochromocytoma (PC12) cells demonstrated that ILK is involved in nerve growth factor-induced neurite outgrowth via inhibition of tau hyperphosphorylation. This recent report using PC12 cells and our results obtained using N1E-115 cells suggest that ILK is an important regulator of both integrin-and growth factor-mediated signaling in neurons and controls neurite outgrowth. Moreover, ILK might be critical for the regulation of microtubule stability and rearrangement necessary for integrin-and growth factor-mediated neurite outgrowth. ILK is activated after cell adhesiononlaminininserum-freeconditionsinaphosphatidylinositol3-kinasedependent manner. Stimulation of ILK activity results in activation of p38 MAP kinase activity, which is important for ILK-dependent neurite outgrowth in N1E-115 cells (11). ILK phosphorylates GSK-3␤ at Ser 9 , leading to GSK-3␤ inactivation. GSK-3␤ inactivation results in inhibition of aberrant tau phosphorylation, an increase in microtubule stability, and induces neurite outgrowth in combination with p38 MAP kinase action. On the other hand, ILK inactivation by DN-ILK induces GSK-3␤ phosphorylation at Tyr 216 via activation of an unidentified tyrosine kinase, leading to GSK-3␤ activation. GSK-3␤ activation results in aberrant tau phosphorylation, microtubule instability, and decreased neurite outgrowth.