Troglitazone, a thiazolidinedione, decreases tau phosphorylation through the inhibition of cyclin-dependent kinase 5 activity in SH-SY5Y neuroblastoma cells and primary neurons

Authors

  • Du-Hyong Cho,

    1. Department of Neurology, Konkuk University Medical Center, Konkuk University, Seoul, South Korea
    2. Department of Pharmacology, Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science School of Medicine, Konkuk University, South Korea
    Search for more papers by this author
  • Eun Joo Lee,

    1. Department of Neurology, Konkuk University Medical Center, Konkuk University, Seoul, South Korea
    2. Department of Pharmacology, Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science School of Medicine, Konkuk University, South Korea
    Search for more papers by this author
  • Kyoung Ja Kwon,

    1. Department of Neurology, Konkuk University Medical Center, Konkuk University, Seoul, South Korea
    2. Department of Pharmacology, Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science School of Medicine, Konkuk University, South Korea
    Search for more papers by this author
  • Chan Young Shin,

    1. Department of Neurology, Konkuk University Medical Center, Konkuk University, Seoul, South Korea
    2. Department of Pharmacology, Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science School of Medicine, Konkuk University, South Korea
    Search for more papers by this author
  • Kee-Ho Song,

    1. Department of Internal Medicine, Konkuk University School of Medicine, Seoul, South Korea
    Search for more papers by this author
  • Jung-Hyun Park,

    1. Department of Molecular Medicine, Ewha Womans University Medical School, Seoul, South Korea
    Search for more papers by this author
  • Inho Jo,

    1. Department of Molecular Medicine, Ewha Womans University Medical School, Seoul, South Korea
    Search for more papers by this author
  • Seol-Heui Han

    Corresponding author
    1. Department of Pharmacology, Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science School of Medicine, Konkuk University, South Korea
    • Department of Neurology, Konkuk University Medical Center, Konkuk University, Seoul, South Korea
    Search for more papers by this author

Address correspondence and reprint requests to Seol-Heui Han, MD, PhD, Department of Neurology, Konkuk University Medical Center and Center for Geriatric Neuroscience Research, SMART Institute of Advanced Biomedical Science, School of Medicine, Konkuk University, 1 Hwayang dong, Gwangjin-gu, Seoul 143-701, Korea. E-mails: alzdoc@kuh.ac.kr; biohyong@hanmail.net

Abstract

The peroxisome proliferator-activated receptor gamma (PPARγ) agonists thiazolidinediones (TZDs) are prescribed for the treatment of type 2 diabetes mellitus. Furthermore, it has been reported that TZDs have a beneficial effect on neurodegenerative disorders, such as Alzheimer's disease. However, the molecular mechanisms underlying this effect are not fully understood. Here, we investigated whether and how troglitazone, a parent TZD drug, inhibits tau phosphorylation. Treatment with troglitazone decreased tau-Thr231 phosphorylation and p35, the specific activator of cyclin-dependent kinase 5 (CDK5), in a dose- and time-dependent manner. Troglitazone also decreased CDK5 enzymatic activity, and ectopic expression of p25, the cleaved and more active form of p35, restored the troglitazone-induced decrease in tau-Thr231 phosphorylation. Treatment with either MG-132, a reversible proteasome inhibitor, or lactacystin, a specific and irreversible 26S proteasome inhibitor, significantly reversed the observed inhibitory effects of troglitazone. However, GW9662, a specific and irreversible PPARγ antagonist, did not alter the observed inhibitory effects. Similar results were also found when other TZD drugs, pioglitazone and rosiglitazone, were used. Treatment with various inhibitors revealed that troglitazone-induced inhibitions of tau-Thr231 phosphorylation and p35 expression were not mediated by glycogen synthase kinase 3β, protein kinase A, and protein phosphatase 2A signaling pathways. Finally, we also found that the same observed inhibitory effects of troglitazone hold true for the use of primary cortical neurons. Taken together, we demonstrated that TZDs repressed tau-Thr231 phosphorylation via the inhibition of CDK5 activity, which was mediated by the proteasomal degradation of p35 and a PPARγ-independent signaling pathway.

image

Thiazolidinediones (TZDs) decrease tau-Thr231 phosphorylation via the inhibition of CDK5 enzymatic activity, which is mediated by the enhanced proteasomal degradation of p35 and a PPARγ-independent signaling pathway. This molecular mechanism implies a role for TZD drugs in the prevention and treatment of Alzheimer's disease.

Abbreviations used
AD

Alzheimer's disease

amyloid beta

BBB

blood–brain barrier

CAMKII

Ca2+/calmodulin-dependent protein kinase II

CDK5

cyclin-dependent kinase 5

DM

diabetes mellitus

ERK1/2

extracellular signal-regulated kinase 1/2

FTDP

frontotemporal dementia

GSK-3β

glycogen synthase kinase 3β

JNK

c-Jun N-terminal kinase

MAP

microtubule-associated protein

NFT

neurofibrillary tangles

PCN

primary cortical neurons

PHF

paired helical filaments

PKA

protein kinase A

PP2A

protein phosphatase 2A

PPARγ

peroxisome proliferator-activated receptor γ

TZD

thiazolidinedione

Thiazolidinediones (TZDs) are synthetic ligands that activate the nuclear receptor peroxisome proliferator-activated receptor γ (PPARγ). These ligands are known to have effects that are beneficial to insulin sensitization, and are therefore used in the treatment of type 2 diabetes mellitus (DM) (Derosa and Maffioli 2012; Schwartz et al. 1998). Furthermore, TZDs also exhibit anti-cancer activities, both in vitro and in vivo, in various cancer models (Ondrey 2009). PPARγ-dependent or -independent signaling pathways mediate these effects according to the TZD concentration and target cell types (Ondrey 2009; Galli et al. 2006). Moreover, there is growing evidence indicating the ameliorating effects of TZDs on pathologically abnormal protein depositions and deficits in memory in various neurodegenerative disorders, especially Alzheimer's disease (AD) (Pedersen et al. 2006; O'Reilly and Lynch 2012). For example, long-term treatment with a TZD drug pioglitazone exhibits several beneficial effects, such as strikingly decreased hyperphosphorylated tau deposit in hippocampal region of brain, enhanced learning on the active avoidance task, and increased short- and long-term plasticity in a mouse model of AD (Searcy et al. 2012). Recently, multiple signaling pathways have been proposed for the underlying mechanisms (Landreth et al. 2008). These include: (i) decrease in amyloid β (Aβ) production by suppressing β-secretase 1 expression and subsequent enhancing degradation of amyloid precursor protein, (ii) enhancement of cerebral energy metabolism by increasing mitochondrial function, (iii) improvement of insulin sensitivity and suppression of glucocorticoid level, and (iv) increase in anti-inflammation through suppressing NF-κB and consequently inhibiting pro-inflammatory cytokine expression and microglial activation.

Tau is a microtubule-associated protein that stabilizes polymerized microtubules in physiological situations, and is also a principal component of paired helical filaments, which is comprised of neurofibrillary tangles (NFTs) in pathological conditions such as AD (Takashima 2010). In this regard, tau is genetically linked to frontotemporal dementia (FTDP-17) (Lee et al. 2001). Six isoforms of the tau protein are produced in the human brain via alternative splicing from one gene and are expressed differentially according to developmental stage (Hernandez et al. 2008). The phosphorylation state of tau determines its binding affinity to microtubules and hyperphosphorylation of tau inhibits the microtubule-assembling activity of the protein and promotes its self-aggregation (Gong and Iqbal 2008; Iqbal et al. 2010). More than 80 putative serine (Ser) or threonine (Thr) phosphorylation sites have been reported in the longest human brain tau isoform (441 amino acids); among these, at least 30 Ser/Thr residues have been found to be phosphorylated (Lovestone and Reynolds 1997; Hanger et al. 1998). Phosphorylation of tau at different sites has different pathogenic effects. A quantitative in vitro study demonstrated that phosphorylation of tau at Ser262, Thr231, and Ser235 inhibits its binding to microtubules by ~35%, ~25%, and ~10%, respectively (Sengupta et al. 1998). In vitro kinetic studies of hyperphosphorylated tau suggest that Ser199/Ser202/Thr205, Thr212, Thr231/Ser235, Ser262/Ser356, and Ser422 are among the critical phosphorylation sites that convert tau to an inhibitory molecule that sequesters normal microtubule-associated proteins from microtubules (Alonso Adel et al. 2004).

Abnormal hyperphosphorylation of tau is believed to cause loss of biological activity, gain of toxic activity, and, consequently, aggregation into paired helical filaments (Alonso Adel et al. 2004; Gong et al. 2005). To date, families of kinases have been identified as phosphorylating tau. These include proline-directed protein kinases [CMGC family, including cyclin-dependent kinase 5 (CDK5), mitogen-activated protein kinases such as the extracellular signal-regulated kinase 1/2, the c-Jun N-terminal kinase and p38, and glycogen synthase kinase 3β (GSK3β)] (Hanger et al. 2009). Members of the AGC kinase family, such as protein kinase A (PKA), and of the CAMK family, such as the Ca2+/calmodulin-dependent protein kinase II, have also been reported as tau kinases (Hanger et al. 2009; Sengupta et al. 2006). In this regard, different degrees of tau pathology have been mimicked in transgenic mice over-expressing GSK-3β (Lucas et al. 2001) or the regulatory activating subunit, p25, of CDK5 (Noble et al. 2003).

It has recently been established that CDK5 is a mainstay in the progression of neurodegenerative diseases as well as in the normal neuronal development process (Jessberger et al. 2009; Camins et al. 2006). To exert its enzymatic activity, CDK5 must associate with a binding partner, such as p35 (or the p25 cleaved form of p35) or p39 (or the p29 cleaved form of p39) (Lalioti et al. 2010). It is well known that during the development of AD, various insults cause an increase in intracellular Ca2+ levels, which ultimately leads to the aberrant activation of CDK5 (Su and Tsai 2011). Finally, aberrant CDK5 activity causes tau hyperphosphorylation and NFT deposition, and contributes to the development of AD (Su and Tsai 2011; Lee and Tsai 2003). It has also been reported that inhibition of CDK5 activity by pharmacological inhibitors in AD model mice reversed the pathogenesis observed (Crews et al. 2011).

Several reports showed that TZDs have ameliorating effects on impaired learning and memory and on aggregated Aβ in AD animal and cell models (Pedersen et al. 2006; De Felice et al. 2009; Escribano et al. 2010). However, the nature of the direct role of TZDs in tau pathology remains elusive. Here, we clearly demonstrated that TZDs alleviated tau-Thr231 phosphorylation in a PPARγ-independent manner via the repression of CDK5 activity, which was mediated by the proteasomal degradation of p35.

Materials and methods

Materials

Troglitazone and rosiglitazone were purchased from Cayman Chemical Company (Ann Arbor, MI, USA), and pioglitazone was purchased from Adipogen International, Inc. (San Diego, CA, USA). GW9662, okadaic acid, and LiCl were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). MG-132 and lactacystin were obtained from Calbiochem (Darmstadt, Germany) and Enzo Life Sciences Inc. (Farmingdale, NY, USA), respectively. [γ-32P]ATP was obtained from PerkinElmer Life Sciences (Boston, MA, USA). The purified histone H1 protein and roscovitine were purchased from Calbiochem. H-89 was purchased from Cell Signaling Technology Inc. (Boston, MA, USA). Protein A agarose was obtained from Thermo Scientific Inc. (Rockford, IL, USA). Antibodies against p-tau-Thr212, p-tau-Thr231, p-tau-Ser262, and tau were purchased from Invitrogen (Carlsbad, CA, USA). Antibody against p-tau-Ser396 was purchased from Cell Signaling Technology Inc. Antibodies against CDK5 and p35 were obtained from Santa Cruz Biotechnology (La Jolla, CA, USA). Antibody against β-actin was purchased from Sigma-Aldrich Co. Dulbecco's modified Eagle's medium (DMEM), Eagle's minimal essential medium, Dulbecco's phosphate-buffered saline, fetal bovine serum (FBS), penicillin and streptomycin antibiotics, l-glutamine, trypsin–EDTA solution, and the plastic ware used for cell culture were purchased from Gibco-BRL (Gaithersburg, MD, USA). All other chemicals were of the purest analytical grade.

SH-SY5Y cell culture, drug treatment, and transfection

SH-SY5Y neuroblastoma cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and maintained in DMEM supplemented with 10% FBS at 37°C under 5% CO2 in air. SH-SY5Y cells that grew to 70% confluence were maintained further for the indicated times in DMEM supplemented with 1% FBS. In some experiments, cells were co-treated with various concentrations of the indicated drugs or chemicals. For transfection experiments, pcDNA3.1 mammalian expression vector-containing cDNAs encoding bovine p25 (Cho et al. 2010) were transfected into the cells grown to 70% confluence in 60 mm culture dishes using Lipofectamine 2000 (Invitrogen), according to the manufacturer's instructions. For control experiments, equal copy numbers of pcDNA3.1 mammalian expression vector were used.

Rat primary cortical neuron culture

Primary cortical neurons were obtained from dissociated embryonic day 18 cortices of Sprague–Dawley rats, as described previously (Kwon et al. 2011). All experimental procedures were conducted in accordance with the approved institutional animal care procedures, and the protocols were approved by the Institutional Animal Care and Use Committee of Konkuk University. Pregnant female Sprague–Dawley rats (n = 2) in each experiment at 8 weeks of age (250–300 g) were purchased from Orient Bio Inc. (Seoul, South Korea) and resided at a controlled temperature (22 ± 1°C) and humidity (50 ± 10%) on a 12-h alternate light–dark cycle. Food and water were provided ad libitum throughout the experiment. Briefly, the cortices, freed of meninges, were mechanically dissociated and gently triturated three times using a flame-polished Pasteur pipette in culture Eagle's minimal essential medium supplemented with 20 mM glucose. Cells were then seeded onto 50 μg/mL poly-d-lysine-coated plates in culture medium supplemented with 5% FBS, 5% horse serum, and 2 mM glutamine. The cultures were maintained further at 37°C in a humidified 5% CO2 incubator. For pure neuronal cultures, 2 μM cytosine-β-arabinofuranoside was added after 2 days of culture, as described previously (Kwon et al. 2011), and cultured neuronal cells were used at 8 days in vitro.

Western blot analysis

Cells after treatment with troglitazone in the absence or presence of various chemicals were lysed in lysis buffer [20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 mM β-glycerophosphate, 1 mM NaF, 1 mM Na3VO4, and 1×  Protease Inhibitor Cocktail (Roche Molecular Biochemicals, Indianapolis, IN, USA)]. Equal quantities of protein (~20 μg) in the lysates were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis under reducing conditions, and then electrophoretically transferred onto nitrocellulose membranes. Blots were probed with the appropriate antibody directed against tau (1 : 2000 dilution), p-tau-Thr231 (1 : 1000 dilution), p35/25 (1 : 500 dilution), and actin (1 : 100 000 dilution), followed by the corresponding secondary antibody, and finally developed using enhanced chemiluminescence reagents (Amersham, DE, USA).

Immunoprecipitaion

SH-SY5Y cells were treated with 20 μM troglitazone or vehicle [dimethylsulfoxide (DMSO)] for 24 h and lysed in lysis buffer, and the lysates were then centrifuged at 12 000 g for 10 min. The supernatant (400 μg of protein) was immunoprecipitated using 4 μL of anti-CDK5 or anti-p35 antibodies. As a control experiment, 4 μL of non-immune serum (normal rabbit IgG) was used. The immunoprecipitates were washed three times with lysis buffer and subjected to western blot analysis.

In vitro kinase activity assay

After immunoprecipitation using 4 μL of anti-CDK5 or anti-p35 antibodies, the immunoprecipitates were washed twice with lysis buffer and twice with kinase assay buffer [20 mM HEPES (pH 7.4), 5 mM MgCl2, and 1 mM dithiothreitol]. For a control experiment, the immunopreciptates were obtained from the lysates using 4 μL of non-immune serum. Each immunoprecipitate was re-suspended in 25 μL of kinase assay buffer and the kinase assay was started by adding 2 μCi of [γ-32P]ATP and 2 μg of histone H1, as a substrate. After 60 min incubation at 30°C, the reaction was terminated by adding Laemmli sample buffer. The samples were boiled for 5 min and subjected to 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and then the dried gel was exposed to X-ray films at −80°C. Kinase activity was quantified by measuring the densitometry of bands obtained using an image analysis program (ImageJ; NIH, Bethesda, MD, USA).

Statistical analysis

All results are expressed as the Mean ± SD, with n indicating the number of experiments. Statistical significance between two points in the in vitro kinase assay was evaluated using Student's t-test. Statistical significance among various doses or time points was determined by one-way anova and DUNCAN analysis as post hoc using SPSS package program (PASW Statistics 18, Quarry Bay, Hong Kong). Statistical significance among groups with more than two variables was determined by two-way anova. All differences were considered significant at a p < 0.05.

Results

Troglitazone decreases tau-Thr231 phosphorylation in SH-SY5Y cells

We examined the effect of troglitazone on tau phosphorylation in SH-SY5Y cells and found that troglitazone decreased tau-Thr231 phosphorylation in a dose- and time-dependent manner, with no alteration of total tau expression (Fig. 1a and b). As treatment with 20 μM troglitazone for 24 h significantly decreased (to 30–40% that of the control) tau-Thr231 phosphorylation, all subsequent experiments were performed using this condition, unless otherwise mentioned. At least under our experimental condition, troglitazone did not alter the phosphorylation states of tau at other sites such as Thr212, Ser262, and Ser396 (Figure S1).

Figure 1.

Troglitazone decreases tau-Thr231 phosphorylation in SH-SY5Y cells. SH-SY5Y cells were treated with various doses of troglitazone (0, 5, 10, or 20 μM) for 24 h (a) or with 20 μM troglitazone for varying times (0, 3, 8, or 24 h) (b). Cell lysates were obtained after treatment and equal amounts of proteins were separated on a sodium dodecyl sulfate polyacrylamide gel and electrophoretically transferred onto nitrocellulose membranes. The p-tau-Thr231 and tau proteins were detected via western blot analysis using antibodies specific for each protein. The blot shown is representative of at least three experiments (n = 3). Densitometry was used to quantify p-tau-Thr231 relative to total tau, and the graphs show the mean fold alterations below the controls (± SD). Statistical significance among groups was determined by one-way anova. All differences were considered significant at a < 0.05.

Troglitazone decreases CDK5 activity and down-regulates p35

As it is well known that CDK5 mediates the hyperphosphorylation of tau, especially in pathogenic states (including AD) and that tau is phosphorylated by CDK5 at the Thr231 residue (Hanger et al. 1998, 2009), we examined whether troglitazone inhibits CDK5 kinase activity. As shown in Fig. 2a, an in vitro kinase assay using CDK5 immunoprecipitates (IP) revealed that troglitazone (20 μM treatment for 24 h) significantly inhibited CDK5 kinase activity. Furthermore, we also found a significant decrease in the activity of p35-associated kinase(s) in troglitazone-treated cells, suggesting that p35 is likely to play an important role in troglitazone-induced decrease in CDK5 kinase activity under our experimental condition (Fig. 2a). Compared with control cells, a significant decreased level of p35, but not CDK5, was also found in p35 IP in trogiltazone-treated cells, indicating that the inhibitory effects of troglitazone on CDK5 activity may be because of decreased p35 level. In fact, troglitazone dramatically decreased p35 expression in a dose- and time-dependent manner (Fig. 2b and c). At present, it is noted that troglitazone significantly, but not completely, decreased tau-Thr231 phosphorylation in cells (Fig. 1), suggesting that not all CDK5 activity was inhibited under our condition (Fig. 2a). To assess a residual CDK5 activity in troglitazone-treated cells, we used a specific CDK5 inhibitor, roscovitine, to inhibit a residual CDK5 activity, and found that roscovitine further decreased the troglitazone-inhibited tau-Thr231 phosphorylation (Figure S2).

Figure 2.

Troglitazone decreases CDK5 activity and down-regulates p35. The CDK5 enzymatic activity or p35-associated kinase(s) activity was measured, and existence of CDK5 or p35 in immunoprecipitated complex was assessed in SH-SY5Y cells treated with or without 20 μM troglitazone for 24 h (a). Cell lysates were obtained and immunoprecipitated using antibodies against CDK5 or p35, and an assay of the kinase activity of CDK5 or the p35-associated protein, or western blot analysis was performed as described in the 'Materials and methods' (a). The blot shown is representative of at least three experiments (n = 3). Densitometry was used to quantify the optical density of 32P incorporated into histone H1; the graph shows the mean optical density with standard deviation (± SD). For the measurement of p35 expression, SH-SY5Y cells were treated as described in the legend to Fig. 1 (b, c). p35 and β-actin were detected via western blot analysis using antibodies specific for each protein. The blot shown is representative of at least three experiments (n = 3). p35 expression was quantified and is shown as a bar graph, as described in the legend to Fig. 1. Statistical significance was determined using Student's t-test for two points and differences were statistically significant at *p < 0.05 (a). Statistical significance among groups was determined by one-way anova and all differences were considered significant at a p < 0.05 (b, c).

Ectopic expression of p25 reverses the troglitazone-mediated decrease in tau-Thr231 phosphorylation

In an attempt to confirm directly whether the troglitazone-induced decrease in p35 expression causes a decrease in tau-Thr231 phosphorylation, we transfected a cDNA encoding bovine p25 into the cells. p25 is an active, cleaved form of p35 that associates with CDK5, resulting in longer activation times. Our results showed that ectopic expression of p25 significantly reversed the troglitazone-induced decrease in tau-Thr231 phosphorylation (Fig. 3).

Figure 3.

Ectopic expression of p25 reverses the troglitazone-mediated decrease in tau-Thr231 phosphorylation. SH-SY5Y cells were prepared, transfected with a cDNA encoding bovine p25, and treated with 20 μM troglitazone for 24 h. Cellular proteins were obtained and subjected to western blot analysis. The blot shown is representative of at least three experiments (n = 3). The level of p-tau-Thr231 was quantified and is shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by two-way anova. All differences were considered significant at a p < 0.05.

Troglitazone reduces p35 expression via the enhancement of proteasomal degradation

It has been reported that p35 is regulated by proteasomal degradation (Patrick et al. 1998) and that troglitazone mediates the ubiquitin-dependent proteasomal degradation of cyclin D1 (Wei et al. 2008). On the basis of these findings, we used the well-known proteasomal inhibitors MG-132 or lactacystin to explore the mechanism via which troglitazone decreases the cellular levels of p35. As shown in Fig. 4, we found that these inhibitors significantly reversed both the decrease in tau-Thr231 phosphorylation and the down-regulation of the p35 protein in SH-SY5Y cells.

Figure 4.

Troglitazone reduces p35 expression via the enhancement of proteasomal degradation. SH-SY5Y cells were prepared and cotreated with either 1 μM MG-132 (a) or 10 μM lactacystin (b) in the presence of 20 μM troglitazone for 24 h. Cell lysates were prepared and tau-Thr231 phosphorylation and p35 expression were detected via western blot analysis, as described in the legend to Fig. 1. The blot shown is representative of at least three experiments (n = 3). The levels of p-tau-Thr231 and p35 were quantified and are shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by two-way anova. All differences were considered significant at a < 0.05.

Troglitazone-mediated decreases in tau-Thr231 phosphorylation and p35 expression are PPARγ independent

As TZDs exert their effects in either a PPARγ-dependent or -independent manner, depending on the cell type or concentration used (Ondrey 2009; Blanquicett et al. 2008), we investigated whether the observed effects induced by troglitazone are PPARγ dependent. As shown in Fig. 5, GW9662, a specific and irreversible PPARγ antagonist, did not alter the observed effects, suggesting that the inhibitory effects of troglitazone on tau-Thr231 phosphorylation and p35 degradation are mediated by a PPARγ-independent signaling pathway.

Figure 5.

Troglitazone-mediated decreases in tau-Thr231 phosphorylation and p35 expression are peroxisome proliferator-activated receptor γ independent. SH-SY5Y cells were prepared and co-treated with 5 μM GW9662 in the presence of 20 μM troglitazone for 24 h. Cell lysates were prepared and tau-Thr231 phosphorylation and p35 expression were detected via western blot analysis, as described in the legend to Fig. 1. The blot shown is representative of at least three experiments (n = 3). The levels of p-tau-Thr231 and p35 were quantified and are shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by two-way anova. All differences were considered significant at a < 0.05.

Other TZDs, pioglitazone and rosiglitazone, inhibit tau-Thr231 phosphorylation and p35 expression

We investigated whether other TZDs, namely pioglitazone and rosiglitazone, also inhibit tau-Thr231 phosphorylation and p35 expression. As shown in Fig. 6, pioglitazone and rosiglitazone repressed tau-Thr231 phosphorylation and p35 expression in a dose-dependent manner in SH-SY5Y cells. However, the efficacy of these drugs on tau-Thr231 phosphorylation and p35 expression was different from that of troglitazone, as higher doses of pioglitazone and rosiglitazone were required to yield equal inhibitory effects to those of troglitazone.

Figure 6.

Other thiazolidinediones, pioglitazone and rosiglitazone, inhibit tau-Thr231 phosphorylation and p35 expression. SH-SY5Y cells were treated with various doses (0, 50, 100, or 200 μM) of pioglitazone (a) or rosiglitazone (b) for 24 h. After treatment, tau-Thr231 phosphorylation and p35 expression were detected via western blot analysis, as described in the legend to Fig. 1. The blot shown is representative of at least three experiments (n = 3). The levels of p-tau-Thr231 and p35 were quantified and are shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by one-way anova. All differences were considered significant at a < 0.05.

Troglitazone-mediated decreases in tau-Thr231 phosphorylation and p35 expression are independent to GSK3β, PKA, and protein phosphatase 2A signaling pathways

As it is well established that several kinases such as GSK3β and PKA, and protein phosphatase such as protein phosphatase 2A (PP2A) play also an important role in regulation of tau phosphorylation, we investigated whether these signaling pathways are also involved in troglitazone-mediated decrease in tau-Thr231 phosphorylation. As shown in Fig. 7a and b, LiCl, a GSK3β inhibitor, and H-89, a PKA inhibitor, decreased basal tau-Thr231 phosphorylation, with no alterations of p35 level. Furthermore, treatment with each of these two chemicals further repressed the troglitazone-inhibited tau-Thr231 phosphorylation in an additive manner. These results suggest that either GSK3β or PKA may not be involved in troglitazone-mediated inhibition of tau-Thr231 phosphorylation. Unlike to two chemicals mentioned, okadaic acid, a PP2A inhibitor, increased basal tau-Thr231 phosphorylation and p35 level. However, okadaic acid did not reverse the observed effects of troglitazone (Fig. 7c), suggesting that PP2A signaling pathways does not mediate the observed inhibitory effects of troglitazone on tau-Thr231 phosphorylation and p35 expression.

Figure 7.

Troglitazone-mediated decreases in tau-Thr231 phosphorylation and p35 expression are independent to glycogen synthase kinase 3β, protein kinase A, and protein phosphatase 2A signaling pathways. SH-SY5Y cells were prepared and co-treated with 10 mM LiCl (a) or 10 μM H89 (b) or 5 nM okadaic acid (c) in the presence of 20 μM troglitazone for 24 h. After treatment, tau-Thr231 phosphorylation and p35 expression were detected via western blot analysis, as described in the legend to Fig. 1. The blot shown is representative of at least three experiments (n = 3). The levels of p-tau-Thr231 and p35 were quantified and are shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by two-way anova. All differences were considered significant at a < 0.05.

Troglitazone has the same effects on tau-Thr231 phosphorylation and p35 degradation in primary cortical neurons

Finally, we tested whether the effects of troglitazone on tau-Thr231 phosphorylation and p35 expression shown in SH-SY5Y cell lines are also exhibited in primary neurons. We treated primary neurons with various doses of troglitazone (0, 10, 20, 40, and 80 μM) for 3 h and found the same inhibitory effects (Fig. 8a). We also treated primary neurons with 60 μM troglitazone in the absence or presence of MG-132 (5 μM), lactacystin (50 μM), or GW9662 (5 μM). As shown in Fig. 8b and c, both MG-132 and lactacystin significantly reversed the troglitazone-inhibited tau-Thr231 phosphorylation and p35 expression in primary cortical neurons. However, GW9662 did not alter the tau-Thr231 phosphorylation and p35 expression inhibited by troglitazone at all (Fig. 8d).

Figure 8.

Troglitazone has the same effects on tau-Thr231 phosphorylation and p35 degradation in primary cortical neurons (PCN). PCN were prepared and treated with various doses of troglitazone (0, 10, 20, 40, or 80 μM) for 3 h (a). For inhibitor studies, PCN were prepared and co-treated with either 5 μM MG-132 (b) or 50 μM lactacystin (c) or 5 μM GW9662 (d) in the presence of 60 μM troglitazone for 3 h. Cell lysates were obtained and tau-Thr231 phosphorylation and p35 expression were detected via western blot analysis, as described in the legend to Fig. 1. The blot shown is representative of at least three experiments (n = 3). The levels of p-tau-Thr231 and p35 were quantified and are shown as a bar graph, as described in the legend to Fig. 1. Statistical significance among groups was determined by one-way anova (a) or two-way anova (b, c, d). All differences were considered significant at a < 0.05.

Discussion

Many studies have shown that TZDs have an inhibitory effect on AD pathologies through multiple modes of action of TZDs in AD animal and cell models (De Felice et al. 2009; Landreth et al. 2008). Hyperphosphorylation of tau at multiple sites has been shown frequently in NFTs, suggesting that tau phosphorylation is an important biomarker for NFT formation and subsequent AD progression. Recently, it was reported that rosiglitazone inhibited tau phosphorylation via the c-Jun N-terminal kinase pathway (Yoon et al. 2010). However, the nature of the mechanism that is involved in the TZD-mediated amelioration of tau phosphorylation remains obscure. As CDK5 is known to be a tau kinase, it is worthwhile investigating whether CDK5 mediates the TZD-mediated decrease in tau phosphorylation. Here, we found that TZDs, including troglitazone, pioglitazone, and rosiglitazone, inhibit tau-Thr231 phosphorylation, which is mediated by decreased CDK5 kinase activity via an increase in p35 degradation. These findings provide a molecular mechanism via which TZDs ameliorate the memory deficits and pathogeneses, such as NFT deposition, shown in AD mouse and cell models.

As the aberrant CDK5 activity caused by the cleavage of p35 into p25 has been well established as a major mechanism underlying various neurodegenerative diseases, our results of troglitazone-induced down-regulation of p35 via enhanced proteasomal degradation have great scientific significance. Because p35 is cleaved by calpains in a neuronal damaged state, producing p25 (which in turn results in prolonged CDK5 activation and abnormal CDK5 subcellular localization), the troglitazone-induced down-regulation of p35 might forestall the overproduction of p25. Furthermore, as CDK5 inhibitors, such as roscovitine, are currently under development for the treatment of cancer and neurodegenerative diseases (Goodyear and Sharma 2007; Krystof and Uldrijan 2010; Crews et al. 2011), down-regulation of p35 by TZDs might equally be of therapeutic significance for treating cancer, in addition to neurodegenerative diseases. However, further investigations are needed to evaluate this issue.

It was noted that over-expression of p25 did not completely restore the troglitazone-induced inhibition of tau-Thr231 phosphorylation (Fig. 3), which implies that more than one signaling pathway may be involved in this process. In this regard, it is known that GSK3β is a tau kinase belonging to the CMGC family (Hanger et al. 2009); thus, it is also likely that troglitazone alters GSK3β activity. Furthermore, it has been also reported that rosiglitazone enhances expression of PP2A catalytic subunit leading to decrease in tau-Thr231 phosphorylation (Flaquer et al. 2010). In this context, we also found previously that troglitazone alters the subcellular activity of PP2A (Cho et al. 2006). However, we failed to find that GSK3β and PP2A signaling pathways are likely to mediate troglitazone-induced inhibition of tau-Thr231 phosphorylation under our experimental condition. In this regard, we found that troglitazone inhibited the phosphorylation of CDK5-Tyr15 (data not shown), which is known to exhibit full activity of CDK5 (Zukerberg et al. 2000), indicating that troglitazone may affect CDK5 activity by regulating its phosphorylation as well as by degrading its activator p35. Further studies should be accomplished to understand troglitazone-mediated dephosphorylation of CDK5 and its physiological function. In addition, our results revealed that PKA was also involved in tau-Thr231 phosphorylation (Fig. 7b). As PKA belongs to AGC kinase family, it is likely that PKA does not phosphorylate tau-Thr231 directly (Zhu et al. 2005), but plays a role as an upstream regulator of CMGC kinase family. However, further investigations are required to unravel this issue. Nonetheless, we clearly found no involvement of PKA in trogiltazone-mediated inhibitions of tau-Thr231 phosphorylation and p35 expression in this study.

It is well established that tau has various phosphorylation sites that affect its biological function collectively in neuronal cells (Gong and Iqbal 2008). Therefore, we examined whether troglitazone alters known phosphorylation residues other than Thr231, such as Thr212, Ser262 and Ser396, and found that phosphorylation states of these residues on tau were not altered by troglitazone (Figure S1). Previously, d'Abramo et al. have reported that troglitazone decreases tau-Ser202 and tau-Ser396/404 phosphorylation in CHOtau4R cells, Chinese hamster ovary cells (CHO) stably transfected with the longest isoform of human tau (4Rtau), through a PPARγ-dependent and/or -independent mechanism (d'Abramo et al. 2006). Furthermore, treatment with pioglitazone in high fat diet mice leads to a significant decrease in tau phosphorylation at the only Ser202/Thr205 residue, but not Thr231 and Ser396 in APOEε3 animals (To et al. 2011). At present time, we are unable to identify the exact reason why tau phosphorylation profiles by TZDs are not fully consistent. Different types of cells and/or experimental conditions may explain these apparently incompatible observations. We used SH-SY5Y neuroblastoma cells and rat primary cortical neuronal cells, while in the previous studies CHOtau4R cells and high fat diet mice were used, respectively. Furthermore, we and d'Abramo et al. used troglitazone while animal study was done with another type of TZD pioglitazone. Nonetheless, it is accepted that hyperphosphorylation of tau at Thr231 plays an important role in its loss of function (Sengupta et al. 1998). Furthermore, the Thr231 residue is shown to be a priming phosphorylation site for the occurrence of further phosphorylation at other sites (Augustinack et al. 2002; Luna-Munoz et al. 2005). Therefore, our observation that troglitazone decreases tau-Thr231 phosphorylation is likely to be deemed especially significant.

TZDs have various effects on cellular processes, such as glucose and lipid metabolism, channel activity, and anti-cancer activity, in a PPARγ-dependent or -independent manner depending on their concentration and the cell types used (Blanquicett et al. 2008; Ondrey 2009). Here, we clearly determined that the troglitazone-mediated inhibition of tau-Thr231 phosphorylation and p35 expression occurred in a PPARγ-independent manner (Fig. 5). Our results were in accordance with those of a previous report showing that ablation of cyclin D1 in breast cancer cells was mediated by increased ubiquitination through PPARγ-independent mechanism (Wei et al. 2008). Furthermore, TZDs such as pioglitazone have been prescribed for the treatment of type 2 DM; in that case, however, the mechanism of action was shown to be attributable to PPARγ-dependent and, to a lesser extent, PPARγ-independent pathways (Hauner 2002; Feinstein et al. 2005). An array of clinical studies have shown that diabetes increases the risk of AD by two- to threefold (Zhao and Townsend 2009). About 80% of AD patients have diabetes or abnormal blood glucose levels (Janson et al. 2004). Therefore, using both PPARγ-dependent and -independent pathways, TZDs such as pioglitazone would be suitable drugs to treat patients with concomitant type 2 DM and neurodegenerative diseases.

We also found that higher doses of pioglitazone and rosiglitazone were required to yield equal inhibitory effects on tau-Thr231 phosphorylation and p35 degradation to those observed for troglitazone (Fig. 6). These findings were fairly consistent with our previous report showing that troglitazone was more effective than other TZDs in inhibiting protein synthesis (Cho et al. 2006). It is speculated that this difference in efficacy may arise from their structural differences, which might affect their plasma membrane-penetrating rate and subcellular distribution. In this regard, it has been reported that pioglitazone passes the blood–brain barrier very poorly and rosiglitazone was initially characterized as being blood–brain barrier impermeant (Maeshiba et al. 1997; Pedersen et al. 2006). However, further investigations are required to elucidate this issue.

In summary, our results demonstrate that TZDs ameliorated tau-Thr231 phosphorylation via the repression of CDK5 activity, which was mediated by proteasomal degradation of p35 in a PPARγ-independent manner. Our findings also suggest that TZDs may be used to treat neurodegenerative diseases, such as AD, as well as type 2 DM.

Acknowledgements

This study was supported by the World Class University (WCU) program through the National Research Foundation of Korea, which is funded by the Ministry of Education, Science and Technology (R33-2008-000-10090-0). The authors have no conflicts of interest to declare.

Ancillary