Conophylline Protects Cells in Cellular Models of Neurodegenerative Diseases by Inducing Mammalian Target of Rapamycin (mTOR)-independent Autophagy*

Background: Autophagy is essential for prevention of neurodegenerative diseases. Results: Conophylline induces mTOR-independent autophagy and protects against neurotoxicity. Conclusion: Conophylline protects cells by enhancement of autophagy in models of neurodegenerative diseases. Significance: Conophylline would be a therapeutic agent for neurodegenerative diseases. Macroautophagy is a cellular response that leads to the bulk, nonspecific degradation of cytosolic components, including organelles. In recent years, it has been recognized that autophagy is essential for prevention of neurodegenerative diseases, including Parkinson disease (PD) and Huntington disease (HD). Here, we show that conophylline (CNP), a vinca alkaloid, induces autophagy in an mammalian target of rapamycin-independent manner. Using a cellular model of PD, CNP suppressed protein aggregation and protected cells from cell death caused by treatment with 1-methyl-4-phenylpyridinium, a neurotoxin, by inducing autophagy. Moreover, in the HD model, CNP also eliminated mutant huntingtin aggregates. Our findings demonstrate the possible use of CNP as a therapeutic drug for neurodegenerative disorders, including PD and HD.

Macroautophagy (herein referred to as autophagy) is a cellular response that leads to the bulk, nonspecific degradation of cytosolic components, including organelles. Autophagy is initiated by the formation of a small membrane particle, called the autophagosomes, mediated by signaling cascades, including autophagy-related genes and microtubule-association 1 light chain 3 (LC3) 2 (1). The completed autophagosome is targeted to the lysosome, and subsequently, the outer membrane of the autophagosome fuses with the lysosomal membrane. Then the inner membrane of the autophagosome and cytoplasmic components are degraded by lysosomal hydrolases. Autophagy is induced under several conditions, including nutrient starva-tion, and is mainly negatively regulated by the serine/threonine protein kinase mammalian target of rapamycin (mTOR). mTOR inhibits the activity of the ULK1 complex, which is essential for autophagosome biogenesis (2). On the other hand, pathways that regulate autophagy independently of mTOR have been reported (3)(4)(5).
In recent years, it has been recognized that autophagy is essential for the prevention of neurodegenerative diseases, including Parkinson (PD) and Huntington (HD) diseases (6). The hallmarks of these neurodegenerative disorders are the presence of intracellular aggregate-prone proteins in the brain. Although the pathogenic role of these aggregates remains controversial, the amount of aggregated protein is positively correlated with neuronal toxicity (7). In addition, the symptoms of the disease are often alleviated by the elimination of aggregates from neurons (8). Therefore, induction of autophagy should provide an attractive therapeutic strategy for these neurodegenerative disorders.
In this study, we searched for novel small molecule autophagy inducers and successfully found that conophylline (CNP) showed such an activity. CNP is a vinca alkaloid first isolated from the tropical plant Tavertaemontana divaricate (9). Also, it was isolated from the leaves of Ervatamia microphylla as a Ras oncoprotein function inhibitor (10). Recently, it has been reported that CNP induces the differentiation of insulin-producing precursor cells in vitro and in vivo (11,12). Here, we reported that CNP induces autophagy and demonstrated its possible use for neurodegenerative disorders.

EXPERIMENTAL PROCEDURES
Materials-CNP was previously isolated from the leaves of E. microphylla as reported before (10). Rapamycin and bafilomycin A1 were purchased from LC Laboratories (Woburn, MA) and Sigma-Aldrich, respectively. htt72Q-AcGFP expression vector was prepared by insertion of huntingtin exon 1 with 72 CAG repeats synthesized by Life Technologies into the pAcGFP1-N1 vector (Clontech). MG132 (Cell Signaling Technology Inc.) was used as a proteasome inhibitor.
Detection of Autophagosome-HeLa cells stably expressing EGFP-LC3 or transiently transfected with EGFP-LC3/EGFP-LC3-G120A (15) for 18 h were treated with chemicals for 24 h at 37°C and were fixed with 4% paraformaldehyde in PBS (Wako Pure Chemical Industries, Ltd., Osaka, Japan) for 20 min. The cells were then washed with PBS and observed under a fluorescence microscope (EVOS FL Cell Imaging System; Life Technologies).
Quantification of the Number of Autophagosomes-The number of EGFP-LC3 vesicles was counted, and the number of vesicles per cell was calculated in HeLa cells stably expressing EGFP-LC3. At least 50 cells were counted from four different fields selected at random.
Immunofluorescence Microscopy-PC12D cells were treated with chemicals for 24 h at 37°C and were fixed with 4% paraformaldehyde in PBS for 20 min. The cells were then washed with PBS and were permeabilized with 0.1% Triton X-100 in PBS for 10 min. After incubation with 3% bovine serum albumin in PBS for 30 min, the cells were immunostained with anti-␣-synuclein antibody (BD Biosciences) for 1 h and incubated with anti-mouse IgG tagged with Alexa Fluor 488 (Life Technologies) for 1 h. The cells were then washed with PBS and were incubated with 2 g/ml Hoechst 33258 (Wako Pure Chemical Industries, Ltd.) for 10 min to stain the nuclei. Then the cells were washed three times with PBS and observed under a fluorescence microscope (EVOS FL Cell Imaging System). To detect aggresomes, the ProteoStat aggresome detection kit (Enzo BioChem. Inc., New York, NY) was used according to the manufacturer's protocol.
Quantification of Aggresome Formation in PC12D Cells-The number of aggresomes was counted and the number of aggresomes per cell was calculated in PC12D cells stained with ProteoStat aggresome detection kit or anti-␣-synuclein antibody. At least 50 cells were counted from four different fields selected at random.
Detection of htt72Q-AcGFP Aggregates and Quantification-Transfection of HeLa cells or Atg7 Ϫ/Ϫ or Atg7 ϩ/ϩ MEFs with htt72Q-AcGFP was performed by using Lipofectamine LTX reagents (Life Technologies) according to the manufacturer's instructions. 6 h after transfection, cells were treated with CNP for 24 h. Then cells were fixed and observed under a fluorescence microscope. For quantification of aggresome formation, we have calculated the percentage of cells that have at least one htt72Q-AcGFP aggregate to AcGFP-positive cells. At least 80 cells were counted from 10 different fields selected at random.
Statistical Analysis-For immunoblotting, densitometry analysis was done by using ImageJ software (National Institutes of Health) from three independent experiments, and the control condition was set to 100%. The values that we obtained were expressed as the means Ϯ S.D. and compared using Student's t test. In the figures, significant p values are shown as * for p Ͻ 0.05 and ** for p Ͻ 0.01.

RESULTS
Conophylline Induces Autophagy-To identify small molecules that could protect neuronal cells, we screened for autophagy inducers from an in-house chemical library, and we found that CNP, a vinca alkaloid, induces autophagy (Fig. 1A). We first examined the levels of LC3-II, a promising autophagosomal marker. LC3-I is conjugated to phosphatidylethanolamine to form LC3-II, which can associate with the autophagosome membrane and is an essential protein for autophagosome formation (15). In HeLa cells, treatment with 100 ng/ml CNP significantly increased the expression levels of LC3-II (Fig. 1B). Autophagosome formation induced with CNP was also confirmed by using HeLa cells stably expressing EGFP-LC3, in which autophagosomes can be observed as EGFP-positive dots (13). In CNP (100 ng/ml)-treated cells, EGFP-positive dots were clearly observed (Fig. 1C). To determine whether the dots formed by CNP treatment is the result of phosphatidylethanolamine conjugation with LC3, we constructed the vector that encodes EGFP-LC3-G120A that is defective in phosphatidylethanolamine conjugation. As shown in Fig. 1D, 100 ng/ml CNP did not form GFP-positive dots in HeLa/EGFP-LC3-G120A, indicating that CNP increased the phosphatidylethanolamine conjugation in HeLa/EGFP-LC3 cells. These data indicate that CNP increases the number of autophagosomes. Moreover, CNP also up-regulated the expression levels of LC3-II in neuronal cell lines: NGF-differentiated PC12D and SH-SY5Y cells (Fig. 1, E and F).
LC3-II up-regulation is attributed to the increase in autophagosome formation or impairment of autophagosome maturation. To investigate whether CNP promotes autophagosome formation, we assessed the expression level of p62, a protein that is selectively degraded by autophagy (17). In HeLa cells, treatment with 100 ng/ml CNP significantly decreased the expression levels of p62, which was blocked by treatment with bafilomycin A1 (BMA), an inhibitor of autophagosome-lysosome fusion (18). In addition, cotreatment with CNP and bafilomycin A1 increased the LC3-II expression level compared with their single treatment ( Fig. 2A). Similar results were obtained in PC12D cells (Fig. 2B). These data indicate that CNP increases autophagy flux.
A recent study identified ARL6ip1 as a direct binding protein of CNP (19). To examine whether the interaction of CNP with  ARL6ip1 is necessary for CNP-induced autophagy, we used RNA interference to down-regulate the protein expression of ARL6ip1. When ARL6ip1 was knocked down by its specific siRNAs, LC3-II expression levels were not altered. Moreover, CNP induced autophagy in ARL6ip1 knocked down cells, as well as control cells (Fig. 2C). These data indicate that CNP induces autophagy independently of ARL6ip1.
Conophylline Enhances mTOR-independent Autophagy-To explore the molecular mechanisms underlying CNP-induced autophagy, we examined whether CNP affects a known pathway that is negatively regulated by the serine/threonine protein kinase mTOR. mTOR kinase activity can be evaluated by the detection of phosphorylation states of the mTOR downstream effectors, such as ribosomal S6 protein kinase (p70S6K) and ribosomal S6 protein (S6) at Thr 389 and Ser 235/236 , respectively. Although rapamycin, a specific mTOR inhibitor, suppressed the phosphorylation of p70S6K and S6, CNP had no such effects in HeLa cells (Fig. 3A). Similar results were obtained in PC12D cells (Fig. 3B). These results indicate that CNP does not interfere mTOR signaling. To confirm whether CNP acts in an mTOR-independent manner, we detected autophagosome formation using HeLa cells stably expressing EGFP-LC3, when cells were co-treated with CNP and a sufficient dose (10 nM) of rapamycin to inhibit mTOR activity. Treatment with CNP and rapamycin had an additive effect on the increase of EGFP-LC3 puncta, compared with their single treatment (Fig. 3C), indicating that CNP and rapamycin induce autophagy through an independent mechanism. These data indicate that CNP regulates autophagy through an mTOR-independent signaling pathway.
Conophylline Protects Cells from Cell Death Induced by 1-Methyl-4-phenylpyridinium-1-Methyl-4-phenylpyridinium (MPP ϩ ), a mitochondrial complex I inhibitor, is the toxic metabolic product of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (20). MPTP is converted to MPP ϩ by monoamine oxidase-B in astrocytes, and MPP ϩ is selectively taken up by dopaminergic neurons. As in animals, MPTP treatment produces clinical, biochemical, and neuropathological changes similar to those found in idiopathic PD (21); thus, MPP ϩ is widely used as a cellular model of PD. MPP ϩ treatments induced iron signal, and it is responsible for intracellular oxidant generation, ␣-synuclein expression and aggregation, proteasomal dysfunction, and apoptosis (22,23). Thus, we examined the effect of CNP on protein aggregation caused by MPP ϩ treatment, assessed by using ProteoStat aggresome dye, which is a red fluorescent molecular rotor dye, to specifically detect denatured protein within aggresomes and aggresome-like inclusion bodies. In PC12D cells, treatment with MPP ϩ induced protein aggregation detected by ProteoStat aggresome dye, whereas CNP markedly decreased the number of aggresomes (Fig. 4A). In addition, immunofluorescence analysis showed that CNP also suppressed the aggregation of ␣-synuclein, which is a major component of aggresomes in PD (Fig. 4B). To investigate whether CNP can degrade the aggresomes, CNP was added after aggresomes were synthesized by MPP ϩ treatment for 24 h. As a result, CNP degraded aggresomes remarkably by 12 h of treatment, suggesting that CNP suppressed aggresome formation by promotion of its degradation (Fig. 4C).
Next, we examined the effect of CNP on cell death induced by MPP ϩ . Our results showed that the decrease in cell viability caused by MPP ϩ was significantly restored by treatment with CNP (Fig. 4D). These data indicate that CNP inhibited MPP ϩinduced protein aggregation and cell death.
To investigate whether the neuroprotective effect of CNP is dependent on autophagy, we used RNA interference to downregulate the expression levels of Atg7, an essential protein for autophagosome formation (24). Transfection with Atg7 siRNA resulted in a significant decrease in Atg7 protein levels and subsequent LC3-II down-regulation in PC12D compared with control experiments with luciferase siRNA sequence (Fig. 5A). Cell viability was then assessed by trypan blue dye exclusion assay after treatment with CNP and/or MPP ϩ in siRNAtransfected PC12D cells. As shown in Fig. 5B, decreased expression of Atg7 with its specific siRNA canceled the cytoprotective effect of CNP, whereas control siRNA did not. Moreover, fluorescence microscopy analysis showed that CNP failed to eliminate the aggregate proteins formed by MPP ϩ in Atg7 siRNA-transfected cells (Fig. 5C). Rapamycin used as a positive control showed similar effect with CNP. These data indicate that CNP enhances autophagy upstream of Atg5/7-dependent autophagosome formation, leading to elimination of aggregates, which allows for escape from cell death induced by MPP ϩ .
Conophylline Reduces Mutant Huntingtin Aggregates-Next, we examined the effect of CNP on a cellular model of HD. HD is caused by a CAG trinucleotide repeat expansion encoding an abnormally long polyglutamine (polyQ) in the N terminus of huntingtin protein. Mutant huntingtin is cleaved to form N-terminal fragments comprising the first 100 -150 residues with expanded polyQ repeats, and the fragments will aggregate and cause toxicity (25). Thus, HD pathogenesis is frequently modeled with exon 1 fragment containing expanded polyQ repeats that cause aggregation and toxicity in cell models and in vivo models (25,26). As expected, transfection with the vector encoding AcGFP-tagged huntingtin exon 1 with 72 polyQ repeats (htt72Q-AcGFP) resulted in the aggregation of htt72Q-AcGFP. CNP significantly reduced htt72Q-AcGFP aggregates (Fig. 6A). Because CNP regulates autophagy through an mTOR-independent signaling pathway, we examined whether CNP and rapamycin have additive effects on reducing htt72Q-AcGFP aggregates by autophagy. As shown in Fig. 6B, single treatment with rapamycin at a sufficient dose to inhibit mTOR activity also reduced the number of htt72Q-AcGFP aggregates. Moreover, co-treatment with CNP and rapamycin significantly reduced the number of htt72Q-AcGFP aggregates compared with their single treatment.
Next, we decided to examine whether this effect of CNP is also dependent on autophagy by using autophagy-deficient MEFs lacking the Atg7 gene (Atg7 Ϫ/Ϫ MEFs) and matched wild-type MEFs (Atg7 ϩ/ϩ MEFs), in which we confirmed that 100 ng/ml CNP induced autophagy (Fig. 6C). As shown in Fig.  6D, there are a larger number of htt72Q-AcGFP aggregates in untreated Atg7 Ϫ/Ϫ MEFs compared with in untreated Atg7 ϩ/ϩ MEFs, because mutant htt is a substrate of autophagy. Moreover, the aggregation of htt72Q-AcGFP was eliminated by treatment with CNP in Atg7 ϩ/ϩ MEFs, whereas in Atg7 Ϫ/Ϫ MEFs, CNP failed to clear the htt72Q-AcGFP aggregates. These data strongly indicate that CNP also eliminates mutant huntingtin aggregates by enhancement of mTOR-independent autophagy.
Previous reports showed that mutant htt protein is degraded through both autophagy and ubiquitin-proteasome system (27,28). We examined whether CNP removed htt72Q-AcGFP aggregates through ubiquitin-proteasome system using MG132, a well known proteasome inhibitor. Fig. 6E shows that the addition of MG132 further increased the number of cells with htt72Q-AcGFP aggregates; however, CNP reduced these by ϳ20% in both the absence and the presence of MG132. This suggests that CNP reduces htt72Q-AcGFP aggregates mainly through autophagy.

DISCUSSION
The induction of the autophagy has been thought to be an attractive therapeutic strategy for neurodegenerative disorders, including PD and HD. In this study, we found CNP to be an autophagy inducer. Treatment with CNP increased the expression level of LC3-II and decreased the levels of p62, an autophagic substrate (Figs. 1B and 2A). Furthermore, CNP had no effect on mTOR signaling, detected by the phosphorylation state of its substrates (Fig. 3, A and B). Also, we showed that co-treatment with CNP and rapamycin had an additive effect on the induction of autophagy (Fig. 3C), indicating that CNP acts in an mTOR-independent manner. Moreover, siRNA experiments showed that ARL6ip1, a direct binding protein of CNP (19), is not involved in the autophagy induction of CNP (Fig. 2C). Some pathways that are involved in mTOR-independent autophagy have been identified, including the cAMP-Epac-PLC-⑀-inositol 1,4,5-trisphosphate pathway and the Ca 2ϩ -calpain-G-stimulatory protein ␣ pathway (29). Additional studies are needed to determine whether CNP induces autophagy through known mTOR-independent pathways or through a new pathway.
To further know how CNP increased the levels of LC3-II, we have examined the effect of CNP on LC3 mRNA level in HeLa cells. 12 h of treatment with CNP significantly increased the levels of LC3 mRNA (data not shown). Also, the up-regulation of both LC3-I and LC3-II protein level was suppressed by the co-treatment with cycloheximide, a protein synthesis inhibitor in HeLa cells (data not shown), indicating that CNP increases LC3 at the mRNA level. Some reports showed that the mRNA expression of LC3 increases during amino acid starvation, a well known autophagic condition, indicating that the induction of autophagy is accompanied by increases in LC3 synthesis (30).
Some compounds that induce mTOR-independent autophagy have been reported to show beneficial effects in neurodegenerative disorders, such as PD and HD. Sarkar et al. (31) showed that three compounds induce mTOR-independent autophagy and contribute to protect neuronal cells in vivo and in cellular models of PD and HD. The disaccharide trehalose is another potent mTOR-independent autophagy enhancer. Trehalose has been shown to induce autophagy and enhance clearance of aggregate-prone proteins in a cellular model of HD (4). In addition, Ca 2ϩ channel antagonists, K ATP ϩ channel openers and  G i signaling activators also induce mTOR-independent autophagy (5). These compounds are also reported to eliminate mutant huntingtin aggregates in cultured cells and in vivo (5).
CNP also eliminates mutant huntingtin aggregates in a cellular model of HD (Fig. 6) and MPP ϩ -induced protein aggregation and cell death in a cellular model of PD (Fig. 5) by inducing autophagy. In a cellular model of PD, transfection with Atg7 siRNA for 96 h did not induce cell death, even though autophagy acts cytoprotectively. We assumed that although activation of autophagy is cytoprotective when cells were damaged by aggresomes, the defect of the basal level of autophagy for a short time has no effect on the cell viability in cultured cell consistent with a previous report (32). Moreover, time course analysis showed that 12 h of treatment with CNP increased LC3-II level, and this effect did not continue until 48 h (data not shown) in HeLa cells. As shown Fig. 4C, CNP could degrade aggresomes by 12 h of treatment, suggesting that CNP suppressed aggresome formation and subsequent cell death by the clearance of aggresomes at an early time point.
CNP acts at a much lower concentration than compounds that are reported to enhance mTOR-independent autophagy. Moreover, a compound that induces autophagy in an mTOR-independent mechanism is thought to be attractive, because the combination of an mTOR-dependent autophagy inducer and mTOR-independent inducer could potentially result in additive benefits compared with their single treatment and could reduce the required dose of each treatment, which might reduce the possibility of any side effect (6). CNP was shown to be orally active and comparatively stable in the blood (30). The neuroprotective effect of CNP by oral administration in animal models of PD and HD will be investigated in future studies.
CNP has been reported to induce the differentiation of insulin-producing precursor cells lowering blood glucose (12,30) and to prevent pancreatic fibrosis in vivo (33). However, its molecular mechanism of action has not yet been well investigated. Our findings that show CNP to be an mTOR-independent autophagy inducer also help us to elucidate the molecular mechanisms by which CNP exhibits various pharmacological activities that have been reported previously.
In conclusion, we found a new mTOR-independent autophagy inducer, CNP, that protects cells in cellular models of neurodegenerative diseases. Our results suggest the potential use of CNP as a new agent for neurodegenerative diseases, including PD and HD.