Cell signaling pathways in the neuroprotective actions of the green tea polyphenol (-)-epigallocatechin-3-gallate: implications for neurodegenerative diseases

Authors

  • Silvia Mandel,

    1. Eve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases Research and Department of Pharmacology, Technion-Faculty of Medicine, Haifa, Israel
    Search for more papers by this author
  • Orly Weinreb,

    1. Eve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases Research and Department of Pharmacology, Technion-Faculty of Medicine, Haifa, Israel
    Search for more papers by this author
  • Tamar Amit,

    1. Eve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases Research and Department of Pharmacology, Technion-Faculty of Medicine, Haifa, Israel
    Search for more papers by this author
  • Moussa B. H. Youdim

    1. Eve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases Research and Department of Pharmacology, Technion-Faculty of Medicine, Haifa, Israel
    Search for more papers by this author

Address correspondence and reprint requests to M. B. H. Youdim, Eve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases Research and Department of Pharmacology, Technion-Faculty of Medicine, Efron Street, PO Box 9697, Haifa 31096, Israel. E-mail: Youdim@Tx.Technion.ac.il

Abstract

Accumulating evidence supports the hypothesis that brain iron misregulation and oxidative stress (OS), resulting in reactive oxygen species (ROS) generation from H2O2 and inflammatory processes, trigger a cascade of events leading to apoptotic/necrotic cell death in neurodegenerative disorders, such as Parkinson's (PD), Alzheimer's (AD) and Huntington's diseases, and amyotrophic lateral sclerosis (ALS). Thus, novel therapeutic approaches aimed at neutralization of OS-induced neurotoxicity, support the application of ROS scavengers, transition metals (e.g. iron and copper) chelators and non-vitamin natural antioxidant polyphenols, in monotherapy, or as part of antioxidant cocktail formulation for these diseases. Both experimental and epidemiological evidence demonstrate that flavonoid polyphenols, particularly from green tea and blueberries, improve age-related cognitive decline and are neuroprotective in models of PD, AD and cerebral ischemia/reperfusion injuries. However, recent studies indicate that the radical scavenger property of green tea polyphenols is unlikely to be the sole explanation for their neuroprotective capacity and in fact, a wide spectrum of cellular signaling events may well account for their biological actions. In this article, the currently established mechanisms involved in the beneficial health action and emerging studies concerning the putative novel molecular neuroprotective activity of green tea and its major polyphenol (-)-epigallocatechin-3-gallate (EGCG), will be reviewed and discussed.

Abbreviations used

amyloid-beta peptide

AD

Alzheimer's disease

ADNF

activity-dependent neurotrophic factor

ALS

amyotrophic lateral sclerosis

APP

amyloid precursor protein

ARE

antioxidant regulatory element

BDNF

brain-derived neurotrophic factor

DA

dopamine

EC

(-)-epicatechin

EGC

(-)-epigallocatechin

EGCG

(-)-epigallocatechin-3-gallate

ERK1/2

extracellular signal-regulated kinases

GSH

glutathione

JNK

c-jun-N-terminal kinase

LDL

low-density lipoprotein

NSAIDs

non-steroid anti-inflammatory drugs

MAPK

mitogen-activated protein kinases

MPTP

1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine

6-OHDA

6-hydroxydopamine

OS

oxidative stress

PD

Parkinson's disease

PKC

protein kinase C

R-APO

DA receptor agonist R-apomorphine

ROS

reactive oxygen species

TRAIL

tumor necrosis factor-related apoptosis-inducing ligand

VIP

vasoactive intestinal peptide

Green tea consumption is considered a source of dietary constituents endowed with biological and pharmacological activities with potential benefits to human health (Graham 1992). Several of these have been subject, in the last few years, to intensive investigation in diverse medical disciplines, such as cardiology, oncology, inflammatory diseases, and neurology. Most of the experimental and epidemiological studies concerning green tea effects have been targeted at its possible cardiovascular, anti-inflammatory, and anticarcinogenic effects, which have been linked to the antioxidant/pro-oxidant properties of its polyphenol constituents. (Wang et al. 1994; Lin and Lin 1997; Yang et al. 1998; Lin et al. 1999a, 1999b; Pan et al. 2000). Among the various bioactive compounds, tea extract is particularly rich in flavonoids, a family of polyphenols found in fruits and vegetables, as well as in plant beverages such as tea, pomegranate juice, raspberry, blueberries, and red wine. Fresh tea (Camellia sinensis) leaves contain a high amount of catechins, a group of flavonoids or flavanols, known to constitute 30–45% of the solid green tea extract (Yang and Wang 1993; Wang et al. 1994). Among the tea catechins, (-)-epigallocatechin-3-gallate (EGCG) is the major constituent, accounting for more than 10% of the extract dry weight, followed by (-)-epigallocatechin (EGC) > (-)-epicatechin (EC) ≥ (-)-epicatechin-3-gallate (ECG). Ingested tea catechins are absorbed mainly in the small intestine and metabolized by enzymatic reactions of glucuronidation, sulfation and O-methylation. These forms are detected in plasma and excreted in bile and urine (for review see Bravo 1998). However, there are species differences concerning the bioavailability of EGCG relative to other tea catechins, as in mouse and humans it has been found to be conjugated less than EGC and EC (Kim et al. 2000; Chow et al. 2001). There is evidence that polyphenol metabolites and their parent compounds have access to the brain. Studies with radioactively labeled EGCG in mouse or chemiluminescence-based detection of EGCG in rats, demonstrated its incorporation into brain, as well as in various organs including kidney, heart, liver, spleen, and pancreas (Nakagawa and Miyazawa 1997; Suganuma et al. 1998). Furthermore, a recent study has shown that the methylated and glucuronidated derivatives of epicatechin are both detected in rat brain following oral administration (Abd El Mohsen et al. 2002).

The beneficial effects ascribed to tea drinking are believed to rely on the pharmacological actions of catechins and their derivatives' components. They act as radical scavengers in vitro and exert indirect antioxidant effects through activation of transcription factors and antioxidant enzymes (for a review see Higdon and Frei 2003). In addition, increased consumption of relatively high concentrations of tea polyphenols has been found to correlate with a reduced incidence of certain cancers, especially skin, lung, esophagus, stomach, liver, and prostate in animal models (Yang and Wang 1993; Higdon and Frei 2003), and to induce apoptosis and cell cycle arrest in a wide array of cell lines as a consequence of pro-oxidant, proapoptotic actions (Ahmad et al. 1997; Masuda et al. 2001; Levites et al. 2002a; 2002b). Despite these in vivo and in vitro results, epidemiological studies on tea intake and risk of cancer are still inconclusive. One possible explanation for this poor correlation may result from lack of a priori, well-designed studies, to assess the impact of tea consumption on risk of cancer contraction. Additional factors may include lifestyle, poor correlation between animal models and humans, and differences in metabolic system among individuals. A deeper understanding of the molecular mechanisms involved will contribute to a better knowledge of the causes of cancer and will allow revelation of novel targets for drug intervention.

Significant data have been accumulated during the last few years, on the positive effects of natural antioxidant nutrients on cardiovascular dysfunction, resulting from multiple factors, such as atherosclerotic plaques, low-density lipoprotein (LDL) oxidation, blood pressure, triglycerides, and cholesterol metabolism. Previous studies demonstrated that consumption of flavonoid-rich nutrients such as red wine (Fuhrman et al. 1995), licorice (ethanolic extract) (Fuhrman et al. 1997), and pomegranate juice (Aviram et al. 2002) results in a reduction in the propensity of plasma LDL to undergo lipid peroxidation in response to copper ions. Similarly, ingestion of either black tea or green tea protects plasma LDL oxidation in humans (Serafini et al. 1996) and in rats fed with green tea extract (Anderson et al. 1998). However, epidemiological studies do not provide conclusive evidence for a protective effect of tea in cardiovascular diseases such as hypertension, coronary heart disease, myocardial infarction, and stroke (Dufresne and Farnworth 2001). This emphasizes the importance of well-designed controlled studies to assess risk reduction in consumers of green and black tea.

The intense investigation on the mechanisms by which neurons die has led to the therapeutic use of antioxidants in aging and neurodegenerative diseases. Oxidative damage to neuronal biomolecules, increased accumulation of iron in specific brain areas, and inflammatory processes with proliferation of reactive microglia are considered major pathological aspects of the aging process and neurodegenerative disorders such as Parkinson's (PD) and Alzheimer's diseases (AD; Gerlach et al. 2003; Selkoe and Schenk 2003). Additional complementary processes include release of cytochrome c, increased α-synuclein aggregation, loss of tissue reduced glutathione (GSH; an essential factor for removal of hydrogen peroxide), reduction in mitochondrial complex I activity, and increased lipid peroxidation (Mizuno et al. 1989; Riederer et al. 1989; Sofic et al. 1992; Jenner and Olanow 1996; Olanow and Youdim 1996; Spillantini et al. 1997; Linazasoro 2002; Perry et al. 2002). There is also evidence for increased expression of apoptotic proteins (for review see Blum et al. 2001), as well as mitochondria and ubiquitin–proteasome system dysfunction, which may lead to breakdown of energy metabolism and consecutive intraneuronal calcium overload (Linazasoro 2002; McNaught et al. 2002). Thus, neurodegeneration in PD or other neurodegenerative diseases appears to be multifactorial, where a complex set of reactions lead to the demise of neurons. This receives further support from recent cDNA microarray gene expression studies, showing the existence of a gene cascade of events occurring in the nigrostriatal pathway of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP; Grunblatt et al. 2001a) and also in the 6-hydroxydopamine (6-OHDA) and methamphetamine animal models of PD (Mandel et al. 2003b). Thus, novel therapeutic neuroprotective strategies support the application of reactive oxygen species (ROS) scavengers, transition metals (e.g. iron and copper) chelators, non-steroid anti-inflammatory drugs (NSAIDs), non-vitamin natural antioxidant polyphenols, antiapototic drugs, such as calcium channel and caspases inhibitors and bioenergetic drugs, in monotherapy, or as part of antioxidant cocktail formulation for the treatment of these diseases (Slikker et al. 1999; Mandel et al. 2003a).

Dietary antioxidants, especially tea and tea flavonoids, have attracted increasing interest due to their relatively potent radical scavenging (Salah et al. 1995; Nanjo et al. 1996), iron chelating (Guo et al. 1996), and anti-inflammatory activities (Haqqi et al. 1999; Pan et al. 2000). In addition to these established properties of green tea polyphenols, an increasing number of studies provided novel molecular targets, possibly implicated in their neuroprotective action. These include control of calcium homeostasis (Ishige et al. 2001), activation of extracellular mitogen-activated protein kinases (MAPK; Schroeter et al. 2002; Chung et al. 2003), protein kinase C (PKC; Levites et al. 2002b, 2003), antioxidant enzymes (Chen et al. 2000; Levites et al. 2001), survival genes (Levites et al. 2002b), as well as the processing of the amyloid precursor protein (APP) pathway (Levites et al. 2003). These findings suggest that green tea extract may be a source of neuroprotectants, with particular relevance to PD, AD, and other neurodegenerative diseases where oxidative stress (OS) has also been implicated.

This review will focus on the most recent findings from in vivo and in vitro studies relating the neurochemical effects of the major green tea polyphenol component EGCG, with special emphasis on the cellular signal transduction pathways implicated in its neuroprotective and neurorescue actions.

Neuroprotection by flavonoids

In vivo and in vitro studies have demonstrated that polyphenolic flavonoids exert a protective role in neurodegeneration. Clinical trials with AD patients have reported potential benefits in cognitive function and memory impairment from treatment with the antioxidant extract of Gingko biloba, known to be enriched with flavonoids (for review see Pratico and Delanty 2000 ). Depressive symptoms of patients with AD and aged non-Alzheimer patients have been shown to respond well to treatment with Ginko biloba extract EGb 761, probably due to its anxiolytic-anti-stress effect (for review see Ward et al. 2002 ). Moreover, the consumption of flavonoid-rich blueberries or strawberries reverses the age-related cognitive and motor behavioral deficits in rats ( Joseph et al. 1999 ) and prevents loss of neurons in the CA1 and CA2 regions, but not CA3 region, of the ischemic rat hippocampus ( Sweeney et al. 2002 ). The antioxidant cocktail drink, β-catechin, has been reported to extend the mean lifespan of senescence-accelerated mice, as compared to their respective controls that received normal drinking water ( Kumari et al. 1997 ). In vitro studies with Ginkgo biloba extract was shown to prevent amyloid-beta peptide (Aβ)-induced neurotoxicity in cultured hippocampal neurons ( Bastianetto et al. 2000 ; Choi et al. 2001 ) and in neuroblastoma cells ( Luo et al. 2002 ). Furthermore, pre-treatment with the flavonoid epicatechin attenuated neurotoxicity induced by oxidized LDL in mouse-derived striatal neurons ( Schroeter et al. 2001 ), as evidenced by apoptotic DNA fragmentation and caspase-3 activation.

Green tea extract and EGCG have also been reported to exert potent neuroprotection in both in vivo and in vitro models of neurodegeneration. EGCG prevented striatal dopamine (DA) depletion and substantia nigra DA-containing neuron loss (Fig. 1), when given chronically to mice treated with the parkinsonism-inducing neurotoxin, N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP; Levites et al. 2001). Additionally, EGCG prevented neuronal cell death in culture caused by the neurotoxins 6-OHDA (Levites et al. 2002b), 1-methyl-4-phenylpyridinium (MPP+Levites et al. 2002b), and Aβ (Choi et al. 2001; Levites et al. 2003). A similar effect of tea polyphenols has also been described in a mouse model of cerebral ischemia, where intravenous (i.v.) injection of tea catechins (Matsuoka et al. 1995) or intraperitoneal (i.p.) injection of EGCG (Lee et al. 2000) improved the memory impairment and reduced hippocampal neuronal damage, respectively.

Figure 1.

Effect of EGCG on striatal dopamine content (a), TH levels (b) and activity (c) in MPTP-treated mice. C57-BL mice received EGCG (2 and 10 mg/kg, orally) for 10 days and the following four days the animals received a combination of EGCG and MPTP (24 mg/kg/day, i.p.). Controls received saline or EGCG only. The results represent the mean ± SEM ( n  = 6 mice/group). * p <  0.05, ** p <  0.005 versus MPTP; † p <  0.05 versus control. Adapted from Levites et al. (2001 ).

Neuroprotective action of EGCG: cell culture studies

The majority of the pharmacological actions of catechins, especially EGCG, relates to their lethal-antitumorigenic action, an effect observed with relatively high doses. These studies have demonstrated that most antioxidants and metal chelating agents, including polyphenols, display biphasic mode of action; they exert pro-oxidant and proapoptotic activity at high doses, but at low micromolar and submicromolar concentrations are neuroprotective against a wide spectrum of neurotoxic agents. This has been demonstrated for typical antioxidants such as ascorbic acid (vitamin C; Halliwell 1996), and for green tea extract (Levites et al. 2002a), EGCG (Levites et al. 2002b), and the DA receptor agonist R-apomorphine (R-APO; Gassen et al. 1998), This concentration range constitutes the pharmacological neuroprotective window of action of a particular drug and may explain why only high doses of catechins are responsible for the anticancer cell death effect. Thus, it seems unlikely that the radical scavenging property of polyphenolic compounds account for their antitumorigenic action, as this antioxidant effect occurs at concentrations that do not induce cell mortality. Until very recently, no explanatory mechanism for this concentration-dependent anti- (neuroprotective) and proapoptotic (neurotoxic) effect of green tea polyphenols has been proposed. Emerging evidence suggests that the antioxidant activity cannot be the sole mechanism responsible for their neuroprotective action, but rather that their ability to alter signaling pathways may significantly contribute to the cell survival effect.

Molecular mechanisms in neuroprotection by EGCG

Modulation of cell survival/death genes

Recently, our group conducted customized cDNA array-based studies to clarify the molecular mechanisms involved in the cell survival and cell death action of EGCG (Levites et al. 2002b). Its gene expression profile was compared to three other neuroprotective drugs, DA, R-APO (both catechol derivatives), and the pineal indoleamine hormone melatonin, at low and high concentrations (Weinreb et al. 2003a, 2003b). EGCG (1 µm), DA (10 µm), and R-APO (1 µm), behaved as potent neuroprotective agents, decreasing the expression of proapoptotic genes bax, bad, gadd45, and fas ligand. However, EGCG did not affect the expression of antiapoptotic bcl-w, bcl-2 and bcl-xL, while the other three compounds increased them. Thus, EGCG neuroprotective effect is thought to be mediated through downregulation of proapoptotic genes as shown for mdm2, caspase-1, cyclin-dependent kinase inhibitor p21, and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL; Levites et al. 2002b), rather than upregulation of antiapoptotic genes (Table 1). Furthermore, when cell viability of neuroblastoma SH-SY5Y cells was challenged with 6-OHDA, low micromolar concentration of EGCG abolished the induction of proapoptotic-related mRNAs and the decrease in Bcl-2, Bcl-w and Bcl-xL.

Table 1.  Comparison of the effects of a protective and toxic conentration of EGCG as determined by cDNA microarray and confirmatory quantitative real-time RT–PCR
Gene nameEGCG (1 µm)EGCG (50 µm)
  1. NB SH-SY5Y cells were treated with EGCG for 6 h. The amount of each product was normalized to the housekeeping genes 18S-rRNA and β-actin. up, upregulated; down, downregulated; NC, not changed. Taken from Levites et al. (2002b) and Weinreb et al. (2003a).

BaxDownUp
BadDownUp
Bcl-XlNCDown
Bcl-2NCDown
Bcl-wNCDown
Caspase-6DownUp
Caspase-1DownUp
fasNCUp
fas ligandDownUp
Gadd45DownUp
Gadd45βNCUp
p21 (waf1)DownUp
TRAILDownUp
Mdm2NCUp

A high degree of homology in the expression pattern of DA and R-APO was also observed (Weinreb et al. 2003a), suggesting that these drugs share a similar mechanism in their cell survival/death action. Significant similarities were noticed between EGCG and R-APO. This might be predictable, given that these compounds share similar attributes, being catechol-like derivatives, iron chelators, and protecting against neurotoxicity induced by 6-OHDA or MPTP (Gassen and Youdim 1999; Mandel et al. 2003a). These findings are supported by an earlier cDNA expression array study conducted in midbrain-containing (nigrostriatal) tissue from mice chronically treated with R-APO and EGCG. The results revealed a significant functional group homology between both drugs, in genes coding for signal transducers, transcriptional repressors, and growth factors, which may account for their neuroprotective activity (Mandel et al. 2002; Table 2). Similar to EGGC, the three neuroprotective drugs are also well-established antioxidant agents, further supporting the assumption that complementary mechanisms, in addition to radical scavenging, are involved in their neuronal survival effect.

Table 2.  Comparison of gene expression induced by the neuroprotective drugs R-APO and EGCG in the substantia nigra-striatum axes of mice
Gene/Protein nameR-APO
(10 mg/kg)
14 days
EGCG
(2 mg/kg)
14 days
  1. From Mandel et al. (2002).

Cell surface antigens
 Gap junction α 1 protein (GJA1); connexin 43 (CXN43; CX43)UpUp
Transcription factors and DNA-binding protein
 Zinc finger protein of the cerebellum 3 (ZIC3)NSUp
 Erf; Ets-related transcription factorUpUp
Cell adhesion receptors and proteins
 Semaphorin JNSDown
Oncogenes and tumor suppressors
 Adenomatous polyposis coli protein (APC)UpUp
Ion channels and transport proteins
 CCHB3; calcium channel (voltage-gated; dihydropyridine-sensitive; l-type) β-3 subunit)DownDown
 SYNAPTOTAGMIN III (SYTIII)DownDown
 Insulin-like growth factor I receptor α subunit (IGF-I-Rα)DownUp
Post-translational modification and folding
 ERp72 endoplasmic reticulum stress protein; protein disulfide isomerase-related proteinDownDown
Receptors
 Adenosine A2A receptor (ADORA2A)DownDown
 Granulocyte-macrophage colony-stimulating factor receptor low-affinity subunit precursor (GM-CSF-R)NSDown
 Lymphotoxin receptor (TNFR family)NSDown
Extracellular cell signaling and communication
 β-Protachykinin aUpUp
 Secretogranin II precursor (SGII); chromogranin CUpUp
 Pleotrophin precursor (PTN) (Heparin-binding growth factor 8) (HBGF-8)UpUp
 Neurogenic locus notch homolog 1 precursor (notch1); motch proteinDownDown
Modulators, effectors and intracellular transducers
 TGF-β-activated kinase 1 (TAK1); mitogen-activated protein kinase 7 (MAP3K7)NSUp
 Interleukin-6 receptor β chain; membrane glycoprotein gp130DownDown
 Rab GDI α; Rab GDP-dissociation inhibitor α; GDI-1; XAP4DownDown

In contrast to the antiapoptotic effect observed with low micromolar concentrations (< 10 µm), a proapoptotic pattern of gene expression is observed with high concentrations of EGCG, DA, R-APO, and melatonin (> 20 µm). It includes the expression of bax, gadd45, caspase family members (3, 6 and 10) and tumor necrosis factor receptor family member fas and fas-ligand mRNAs (Table 1). The effect of melatonin was comparably lower than that of EGCG, R-APO, and DA; for example, neither bcl-2 nor bcl-xL mRNA levels were decreased by the high concentration of melatonin in contrast to the other antioxidants, possibly explaining the low index of mortality observed with very high doses. A similar high concentration of EGCG has been recently reported to activate apoptotic pathways in the prostate carcinoma cell line LNCaP, elevating the transcriptional activity of p53 and activating its downstream targets p21/WAF1 and Bax (Hastak et al. 2003).

Given the central role mitochondria play in OS-induced apoptosis, it may be speculated that EGCG-mediated inhibition of apoptosis might implicate mitochondrial targets. This may be a consequence of the blockade of mitochondrial permeability transition pore (mPTP) opening, as flavonoids have been reported to bind to mitochondrial benzodiazepine binding sites of GABA-A and adenosine receptors, which are associated with the mPTP, thus modulating cytochrome c release (for a review see Schroeter et al. 2002). Secondly, flavonoids may affect mitochondrial integrity by modulating calcium homeostasis, as previously reported Ishige et al. 2001). Finally, our findings with cDNA microarrays, show that EGCG represses the expression of bad and bax mRNA, which may favor the increase in the ratio of Bcl-2/Bcl-xL to Bax/Bad proteins, thereby contributing to mitochondrial stability and regulation of mPTP (Merry and Korsmeyer 1997). Taken together, these findings suggest that EGCG might function as a bioenergetic drug such as coenzyme Q, which is currently in clinical trials for PD patients (Beal 2003), modulating mitochondrial responses to oxidative insults.

The modulation of cellular survival and cell death pathways has significantly biological consequences that are important in understanding the various pharmacological and toxicological responses of antioxidant drugs. This discriminatory target action of the different neuroprotectants, may allow for a more specific therapy design, especially when a multipharmacological action drug cocktail, is considered for formulation.

Cell signal transduction pathways

A number of intracellular signaling pathways have been described to play central functions in neuronal protection against a variety of extracellular insults. These include, among others, the MAPK (Xia et al. 1995; Singer et al. 1999; Levites et al. 2002b), phosphatidylinositol-3-kinase-AKT (Kaplan and Miller 2000), and PKC (Maher 2001; Cordey et al. 2003) pathways. MAPKs are important transducers of extracellular stimuli into intracellular responses, through a series of phosphorylation cascades. Among the MAPKs the extracellular signal-regulated kinases (ERK1/2) are mainly activated by mitogen and growth factors (Vaudry et al. 2002), while p38 and c-jun-N-terminal kinase (JNK) respond to stress stimuli such as inflammation, UV irradiation and several types of OS, some of them may result in cell death by apoptosis (Harris et al. 2002; Johnson and Lapadat 2002). The involvement of ERK1/2 signaling in neuronal survival, has been demonstrated by the observed decrease in its activity in primary cortical neurons intoxicated with glutamate (Singer et al. 1999), in PC12 cells in which apoptosis was induced by nerve growth factor withdrawal (Xia et al. 1995) and in SH-SY5Y neuroblastoma cells treated with 6-OHDA (Levites et al. 2002b). However, the fate of a cell to survive or die upon exposure to different agents will depend on the cell type, growing conditions and kind of stimuli. Different molecules may use the same pathway to initiate distinct biological effects. For instance, two growth factors, epidermal growth factor (EGF) and nerve growth factor (NGF), use the same pathway to cause PC12 proliferation and differentiation, respectively (Vaudry et al. 2002). Interestingly, there have been reports where activation of ERK1/2 seems to mediate neuronal injury in focal ischemia (Alessandrini et al. 1999), in glutamate and oxidized-LDL-induced toxicity (Yun et al. 1998; Schroeter et al. 2001) and in cytotoxicity and activation of caspase-3 in the extraneuronal hepatoma HepG2 and HeLa cell lines, respectively (Chen et al. 2000).

Flavonoids can activate MAPK signaling cascades in both neuronal and extraneuronal tissues and neutralize the decline in ERK activity caused by exogenous OS-inducing agents. EGCG counteracted the decline in ERK1/2 induced by 6-OHDA in neuroblastoma cells (Levites et al. 2002b). However, neither EGCG nor catechin (Schroeter et al. 2001), at their neuroprotective concentrations (1–10 µm), affected the levels of ERK1/2 phosphorylation by themselves, in the absence of any exogenous damage, in neuronal cell line and primary neuron cultures. Conversely, resveratrol, a natural flavonoid occurring in grapes and wine, was shown to directly induce phosphorylation of ERK1/2 in SH-SY5Y neuroblastoma cells (Miloso et al. 1999). This discrepancy may result from structural differences between the catechins and flavonoids, although no biochemical explanation for this has been presented so far. Alternatively, the cell type (e.g. neuronal vs. peripheral tissue) may also determine the cellular response to a particular stimulus. Recently, it has been shown that topical application of EGCG induced proliferation of human normal epidermal keratinocytes through stimulation of ERK1/2 and AKT (Chung et al. 2003). This finding is not surprising, given that ERK1/2 is strongly activated by mitogens and growth factors (Vaudry et al. 2002). Thus, the differential signal pathway activation by EGCG in neurons compared to peripheral cells, may reflect cell–function diversity.

EGCG neuroprotective activity also involves the intracellular signaling mediator PKC (Levites et al. 2002b, 2003), a family of serine/threonine kinases consisting of 11 isoforms, which are divided into three subclasses: conventional (α, βΙ, βΙΙ, γ), novel (δ, εθ, η, µ), and atypical (ι/λ, ζ). Conventional PKCs require Ca2+ and diacylglycerol for activation, whereas novel PKCs need only diacylglycerol (Battaini 2001). However, neither Ca2+ nor diacylglycerol are involved in the activation of atypical PKCs. PKC is thought to have a fundamental role in the regulation of cell survival, and programmed cell death and long-term potentiation (LTP; Berra et al. 1997; Dempsey et al. 2000; Maher 2001). The rapid loss of neuronal PKC activity is a common consequence of several brain damages, such as ischemia (Cardell and Wieloch 1993; Busto et al. 1994) and glucose deprivation (Small et al. 1996). In effect, the binding of phorbol ester, a direct activator of PKC, is significantly reduced in the substantia nigra, caudate putamen, and pallidum of PD patients, when compared with the relevant controls (Nishino et al. 1989). Thus, it seems surprising that only a small number of compounds have been reported to prevent the decline in PKC activity, observed in response to a variety of insults. These include brain-derived neurotrophic factor (BDNF), activity-dependent neurotrophic factor (ADNF), and vasoactive intestinal peptide (VIP), which were shown to protect primary rat cortical neurons (Gressens et al. 1998) and mouse white matter tissue (Gressens et al. 1999) against excitotoxic insults, through activation of PKC.

The induction of PKC activity in neurons by EGCG is thought to be a prerequisite for neuroprotection against several neurotoxins. This catechin prevented Aβ (Levites et al. 2003), serum withdrawal (Mandel et al. 2003c), and 6-OHDA (Levites et al. 2002b)-induced cell death and these effects were paralleled by a rapid phosphorylative activation of PKC. This was further validated by the finding that a general PKC inhibitor completely abolished the protection induced by EGCG and that phorbol 12-myristate 13-acetate (PMA), a direct activator of PKC, mimicked the protective effect, suggesting a crucial role of PKC in cell survival. Similar results have been obtained with estrogen, which protected against Aβ toxicity through activation of PKC in rat cortical neurons (Cordey et al. 2003). These in vitro findings are supported by the results of Levites et al. (2003), who showed that repeated administration of EGCG for 7 or 14 days caused significant increase in the protein expression levels of PKC isoenzymes α and ε in the membrane and cytosolic fractions of mice hippocampus (Levites et al. 2003). It is well established that these isoforms play a central role in cell survival and differentiation pathways (Gubina et al. 1998; Whelan and Parker 1998; Maher 2001).

The role of PKC in neurodegenerative diseases and neuroprotection is acquiring major relevance, in light of the recent findings with the novel anti-PD drug rasagiline, developed in our laboratory (Parkinson Study Group 2002) and more recently chosen for neuroprotective studies in PD, by the National Institute of Health (Ravina et al. 2003). This drug displays potent neuroprotective and antiapoptotic features against a variety of insults in cell cultures (Youdim et al. 2001; Yogev-Falach et al. 2003) and in vivo (Sagi et al. 2003), related to its PKC-dependent activation of ERK1/2 and induction of Bcl-2 (Ban Am et al. 2004a,b). The mechanism by which PKC activation leads to neuroprotection has to be further defined. Studies with extra-neuronal tissues support a role for PKCα as a functional Bcl-2 kinase that can suppress cell apoptosis, probably through direct or indirect phosphorylation of Bcl-2 (Ruvolo et al. 1998). Moreover, overexpression of PKCε in a hematopoietic cell line, resulted in increased expression of Bcl-2 (Gubina et al. 1998). Whether EGCG activation of PKC leads to a direct phosphorylation and subsequent activation of Bcl-2, thereby suppressing apoptosis, remains to be determined.

The signaling pathways participating in OS-induced cell damage and suggested major targets for EGCG actions are depicted in Fig. 2.

Figure 2.

A hypothetical model diagramming cell signaling pathways and cell survival/death genes affected by neurotoxin-induced oxidative stress. Potential targets of green tea polyphenols (GT, 0.06–6 µ m ) or EGCG (0.01–10 µ m ) are suggested: (a) direct scavenging of ROS induced by neurotoxins; (b) counteracting the negative effect of ROS on both PKC and ERK1/2 pathways. Alternatively, ERK1/2 might be subjected to direct regulation by PKC, thus, activation of this kinase by EGCG may lead to restoration of the reduced phosphorylated ERK1/2 levels; (c) direct phosphorylative activation of PKC; (d) inhibiting the translocation of NF-kB to the nucleus; (e) preventing the expression of cell-death genes. Sharp arrows indicate positive inputs, whereas blunt arrows are for inhibitory inputs. PKC, protein kinase C; ERK, extracellular signal-regulated kinase; SAPK/JNK, stress-responsive mitogen-activated protein kinase/c-jun N-terminal kinase; EGCG (-)-epigallocatechin-3-gallate; ROS, reactive oxygen species; 6-OHDA, 6-hydroxydopamine; MPP + , 1-methyl-4-phenylpyridinium; Aβ, amyloid-beta peptide.

Antioxidant function: scavenging of ROS and induction of endogenous antioxidants

Tea catechins are powerful hydrogen-donating antioxidants and free radical scavengers of ROS and nitrogen species in in vitro systems (Salah et al. 1995; Nanjo et al. 1996; Morel et al. 1999). In brain tissue, green tea and black tea extracts were shown to strongly inhibit lipid peroxidation promoted by iron-ascorbate in homogenates of brain mitochondrial membranes (IC50: 2.44 and 1.40 µmol/L polyphenols, respectively; Levites et al. 2002a). A similar effect was also reported using brain synaptosomes, in which the four major polyphenol catechines of green tea were shown to inhibit iron-induced lipid peroxidation (Guo et al. 1996). In the majority of these studies, EGCG was shown to be more efficient, as a radical scavenger, than its counterparts ECG, EC and EGC, which might be attributed to the presence of the trihydroxyl group on the B ring and the gallate moiety at the 3′ position in the C ring (Nanjo et al. 1996). In this regard, it has been shown that the major metabolite recovered in bile after catechin administration to rats, 3′-O-methyl-(-)epicatechin-5-O-β-glucuronide, methylated at the 3′-position, has very low superoxide scavenging activity (Harada et al. 1999). However, 3′-O-methyl-(-)epicatechin and its parent compound epicatechin, were equally effective against oxidative stress-induced in striatal neurons, despite the differences in their antioxidant potency (Schroeter et al. 2001). These findings support the notion that green tea polyphenols might exert neuroprotection independently of their classical antioxidant activity.

The neuroprotective actions of green tea polyphenols and flavonoids may also result from their ability to induce of endogenous antioxidant defenses. The various flavonoids and phenolic antioxidants activate the expression of some stress-response genes, such as phase II drug metabolizing enzymes, glutathione-S-transferase and heme-oxygenase 1 (Chen et al. 2000), probably via their binding to the antioxidant regulatory element (ARE) present in the promoter of those genes. Additionally, during OS, the transcriptional activation of these stress-response genes, correlated with an increase in the activity and nuclear binding of the transcription factors Nrf1 and Nrf2 to the ARE sequences contained in their promoters and with activation of the MAPK pathway (Owuor and Kong 2002). These authors suggested that low, protective concentrations of antioxidants, may activate these transcription factors by increasing their phosphorylation and activity, through modulation of MAPKs function.

Iron chelation, inhibition of NF-κB nuclear translocation, and iNOS activation

Green tea polyphenols have been shown to possess relatively potent metal chelating properties (Guo et al. 1996; Grinberg et al. 1997), which have been attributed to the gallate moiety present in the C-ring of both EGCG and ECG (Kumamoto et al. 2001). Thus, the ability of polyphenols to act as radical scavengers and chelate transitional metals such as iron and copper, may be of major significance for treatment of neurodegenerative diseases (PD, AD, ALS, Huntington's disease, multiple sclerosis, etc.), where accumulation of iron at brain areas where neurodegeneration occurs, has been shown (Youdim and Riederer 2003). The localization of iron and ferritin in PD patients is restricted to definite brain areas (Riederer et al. 1989; Jellinger et al. 1990; Sofic et al. 1991), specifically in the substantia nigra pars compacta (SNPC) and not the reticulata, even though the latter region has a higher iron content than the SNPC (Jellinger et al. 1990). Similarly, AD pathogenesis is also associated with iron accumulation and is linked to the characteristic neocortical Aβ deposition, phosphorylation of tau and tangle formation, which may be mediated by abnormal interaction with excess of free-chelatable iron. Ionic iron can, in turn, participate in Fenton reaction with subsequent generation of ROS, thought to initiate the processes of OS and inflammatory cascade resulting in production of cytotoxic cytokines (TNFα, IL-1, and IL-6) in the microglia and in the surrounding neurons (Mogi et al. 1994; Blum-Degen et al. 1995; McGeer and McGeer 1995; Sakaguchi et al. 1996).

The implication of the pivotal role for iron in DA-containing neuron degeneration, has been strengthened by the observations that both MPTP and 6-OHDA significantly increased iron in SNPC of mice, rats, and monkeys treated with these neurotoxins (Monteiro and Winterbourn 1989; Mochizuki et al. 1994; Oestreicher et al. 1994; Temlett et al. 1994). Accordingly, the first studies on pre-treatment with the iron chelator desferrioxamine (Ben-Shachar et al. 1991; Lan and Jiang 1997; Matarredona et al. 1997), demonstrated a significant protection against the neurotoxic actions of MPTP and 6-OHDA. More recent studies examining the effect of iron chelation by either transgenic expression of the iron binding protein ferritin, or oral administration of the metal chelator clioquinol on the susceptibility to MPTP, have shown to be protective against the toxin (Kaur et al. 2003). Furthermore, we have recently developed a brain-permeable iron chelator VK-28, that induces neuroprotection against 6-OHDA lesion of nigrostriatal DA neurons when administered peripherally systematically (Ben Shachar et al. 2004). Similarly, chelation of zinc and copper by the metal chelator clioquinol was shown to markedly inhibit beta-amyloid accumulation in AD transgenic mice (Cherny et al. 2001). Confirmatory gene expression cDNA microarray studies have shown that MPTP, 6-OHDA and methamphetamine increased a number of inflammatory and proapopototic genes while antioxidant defense genes were downregulated (for review see Mandel et al. 2003b). The MPTP-induced events were completely reversed by pre-treatment of mice with the dopamine D1/D2 receptor agonist R-APO (Grunblatt et al. 2001a).

OS-induced ROS generation is associated with the activation of the inflammation and iron-responsive transcription factor, NF-κB (Schreck et al. 1991; Lin and Lin 1997). Indeed, Hunot et al. (1997) found a 70-fold increase in NF-κB immunoreactivity in the nucleus of melanized dopaminergic neurons of the Parkinsonian SNPC, compared to normal brains. EGCG was found to inhibit the nuclear translocation of NF-κB in many cellular systems: immunofluorescence and electromobility shift assays showed that introduction of green tea extract before 6-OHDA-induced OS, inhibited both NF-κB nuclear translocation and binding activity in NB SH-SY5Y cells (Levites et al. 2002a). Furthermore, the reduced activity of NF-κB by EGCG and the theaflavin-3,3′-digallate polyphenol from black tea, was associated with inhibition of LPS-induced TNFα (Yang et al. 1998) and iNOS (Lin and Lin 1997; Lin et al. 1999b; Pan et al. 2000) inflammatory mRNA and protein levels in macrophages. In light of this accumulated information, our group strongly advocate the use of brain-permeable iron chelators as neuroprotective drugs to ‘iron out iron’ from those brain areas where it preferentially accumulates in neurodegenerative diseases. Unfortunately, most iron chelators do not cross the blood–brain barrier and thus special emphasis should be put in development of novel, centrally acting, non-toxic iron chelators, such as VK-28 (Ben Shachar et al. 2004), clioquinol (Kaur et al. 2003), and EGCG (Suganuma et al. 1998).

Neuroprotective action of EGCG: in vivo studies

Parkinson's disease

In spite of the lack of well-controlled clinical trials with tea polyphenos in neurodegenerative diseases, one recent epidemiological study has shown reduced risk of PD associated with consumption of two cups/day or more of tea and two or more cola drinks/day (Checkoway et al. 2002). Neuroprotective studies employing MPTP have shown that both green tea extract and EGCG possess highly potent activities in preventing mice striatal dopamine depletion and substantia nigra dopaminergic neuron loss, induced by MPTP (Levites et al. 2001). These compounds are poor inhibitors of MAO-B, as compared to the selective irreversible MAO-B inhibitors l-deprenyl or rasagiline. Therefore, they are unlikely to act by interfering with the conversion of MPTP to its active metabolite MPP+, a reaction mediated by striatal MAO-B. One possible mechanism underlying the effectiveness of green tea and EGCG against MPTP neurotoxicity, may involve its catechol-like structure, as it is known that catechol-containing compounds are potent radical antioxidants and chelators of ferric ion (Guo et al. 1996; Gassen and Youdim 1997, 1999; Grinberg et al. 1997). In agreement, the catechol derivatives, R-APO and its S-isomer, S-APO (not a dopamine agonist), both shown to induce neuroprotection in cell cultures (Gassen et al. 1998) and in animal models of PD (Grunblatt et al. 1999; Grunblatt et al. 2001b), are potent iron chelators and radical scavengers in brain mitochondrial fraction (Gassen et al. 1996). The structural cathecol-resemblance may account for a recently reported inhibitory effect of green tea polyphenols on the DA pre-synaptic transporters, thereby blocking MPP+ uptake (because of competition for the vesicular transporter) and protecting dopamine-containing neurons against MPP(+)-induced injury (Pan et al. 2003). In addition, in vitro studies conducted in rat liver cytosol homogenates, show that EGCG greatly inhibit catechol-O-methyltransferase (COMT) activity at a low IC50 concentration (0.2 µm; Lu et al. 2003). This action may be of particular significance for PD patients given that DA and related catecholamines are physiological substrates of COMT. The COMT inhibitors entacapone and tolcapone, clinically prescribed to PD-affected individuals, dose-dependently inhibit the formation of the major metabolite of levodopa, 3-O-methyldopa, thereby improving its bioavailability in the brain (Deleu et al. 2002).

The neuroprotective effect of green tea polyphenols in vivo may also involve the regulation of antioxidant protective enzymes (as discussed earlier). EGCG was found to elevate the activity of two major oxygen-radical species metabolizing enzymes, superoxide dismutase and catalase in mice striatum (Levites et al. 2001). This is supported by a previous finding where 1-month administration of a catechin-containing antioxidant preparation increased superoxide dismutase (SOD) activity in the mitochondria fraction of striatum and midbrain and decreased thiobarbiturate reactive substance formation in the cortex and cerebellum of aged rats (Komatsu and Hiramatsu 2000). The significance of these findings and the previously discussed in vitro mechanisms implicated in EGCG neuroprotection, on the incidence and progression of PD, needs to be established.

Neuroprotection by flavonoids in AD

Clinical trials with AD patients have also demonstrated potential benefits in cognitive function and memory impairments from treatment with the antioxidant extract of Gingko biloba, known to be enriched with flavonoids (for review see Pratico and Delanty 2000). Depressive symptoms of patients with AD and aged non-Alzheimer patients may also respond to treatment with Ginko biloba extract as this extract has an ‘antistress’ effect (Ward et al. 2002).

A significant body of evidence point to an ‘amyloid cascade’ event in the pathogenesis of AD, where APP is processed to Aβ, which spontaneously self-aggregates, in the presence of divalent metals (Fe2+, Cu2+), into neurotoxic amyloid fibriles in the neocortex (Bush 2003). APP can be processed via alternative pathways; a non-amyloidogenic secretory pathway, includes cleavage of APP to soluble APP (sAPP) by a putative α-secretase within the sequence of Aβ peptide, thus precluding the formation of Aβ, whereas the formation of the amyloidogenic Aβ peptides is regulated by the sequential action of β- and γ-secretases (Checler 1995; Nunan and Small 2000; Bush 2003). As the proportion of APP processed by β-secretase vsα-secretase may affect the amount of Aβ produced, the regulation of these two pathways may be critically important to the pathogenesis of AD. However, given that Aβ is ubiquitously produced in the brain, there is no direct evidence that its overproduction and aggregation underlies sporadic AD. A recent study (Levites et al. 2003) has shown that EGCG promotes the non-amyloidogenic α-secretase pathway of APP neuronal cell cultures. The increase was dose-dependent and the stimulatory effect of EGCG on sAPPα secretion was inhibited by the hydroxamic acid-based metalloprotease inhibitor Ro31-9790, indicating that this effect was mediated via α-secretase processing. Also, long-term treatment of mice with EGCG resulted in decreases in cell-associated, full-length APP levels, as well as increases in sAPP levels in the hippocampus. Thus, the proteolytic processing of APP can also be regulated by treatment with EGCG under in vivo conditions. Inhibition of PKC activity, whose involvement in sAPP release is well established (Checler 1995; Mills and Reiner 1999), prevented EGCG-induced sAPPα release, indicating the key role of PKC in mediating EGCG effect. Although it is not known which isoenzyme of PKC plays a major role in modulating APP processing, several lines of evidence suggest the involvement of PKCα and PKCε in APP processing (Slack et al. 1993; Benussi et al. 1998). In a rat fibroblast cell line, sAPP was increased after stable overexpression of PKCα and PKCε isoenzymes (Kinouchi et al. 1995), while blockade of PKCε activation attenuated phorbol ester-induced increase of α-secretase-derived sAPP (Yeon et al. 2001), further supporting previous studies in brains of AD patients where PKCε activity in the membrane fraction was reduced (Matsushima et al. 1996). In agreement with these findings, repeated administration of EGCG for 7 or 14 days caused significant increases in the protein expression of PKC isoenzymes α and ε mice hippocampus (Levites et al. 2003). A proposed schematic model for the neuroprotective effect and regulation of APP processing by EGCG is illustrated in Fig. 3.

Figure 3.

Schematic representation of the regulation of APP processing by EGCG. The model suggests that EGCG activates PKCα and PKCε, leading to increased production of potentially neuroprotective, nonamyloidogenic sAPPα. Because sAPPα and Aβ are formed by two mutually exclusive mechanisms, stimulation of the secretory processing of sAPPα might prevent the formation of the amyloidogenic Aβ. ↑ Increased levels/activity, ↓ decreased levels/activity. Sharp arrows indicate positive inputs, whereas blunt arrows are for inhibitory inputs. Taken from Levites et al. (2003 ).

Non-amyloidogenic sAPPα has been demonstrated to posses potent neurotrophic and neuroprotective activities against excitotoxic and oxidative insults in various cellular models (Mattson et al. 1993, 1997; Schubert and Behl 1993), and it was shown to protect against p53-mediated apoptosis (Xu et al. 1999). Moreover, sAPPα promotes neurite outgrowth (Small et al. 1994), regulates synaptogenesis (Morimoto et al. 1998), and exerts trophic effects on cerebral neurons in culture. Thus, it may be suggested that EGCG could not only influence the basic pathogenic mechanism underlying AD but may have a significant benefits for slowing the disease progression.

Conclusion

It is apparent that naturally occurring polyphenols such as EGCG, exert profound pharmacological activities, beyond their presently established concentration-dependent antioxidant/pro-oxidant activities. These include inhibition of DA uptake; inhibition of COMT activity, striatal activation of SOD and catalase; inhibition of the OS-induced translocation of NF-κB to the nucleus; activation of PKC; counteraction of the negative effect of OS on both PKC and ERK1/2 pathways; prevention of the expression of cell death and cell cycle regulator genes; promotion of the secretion of soluble, non-toxic, non-amyloidogenic form of APPα, reputed to have neurotrophic and neuroprotective properties against excitotoxic and oxidative insults (Fig. 4).

Figure 4.

Summary of the suggested potential targets involved in EGCG neuroprotective action.

Tea consumption is changing its status from a mere ancient beverage and a lifestyle habit, to a nutrient, endowed with potential biological–pharmacological actions beneficial to human health. This perception supports the current view that polyphenolic dietary supplementation may have an impact on cognitive deficits in old age. As a consequence, green tea polyphenols are now being considered as therapeutic agents in well-controlled epidemiological studies, aimed to alter brain aging processes and as possible neuroprotective agents in progressive neurodegenerative diseases.

Acknowledgements

The authors gratefully acknowledge the support of the Israel Psychobiology Center (Israel); National Parkinson Foundation (Miami, USA); Technion-Research and Development (Haifa, Israel); Friedman Research Fund and Stein Foundation (Philadelphia, USA).

Ancillary