Naturally occurring phytochemicals for the prevention of Alzheimer’s disease


  • Jiyoung Kim,

    1. Major in Biomodulation, Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, Seoul National University, Seoul, Korea
    2. Department of Bioscience and Biotechnology, Bio/Molecular Informatics Center, Konkuk University, Seoul, Korea
    Search for more papers by this author
  • Hyong Joo Lee,

    1. Major in Biomodulation, Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, Seoul National University, Seoul, Korea
    Search for more papers by this author
  • Ki Won Lee

    1. Department of Bioscience and Biotechnology, Bio/Molecular Informatics Center, Konkuk University, Seoul, Korea
    Search for more papers by this author

Address correspondence and reprint requests to Ki Won Lee, Department of Bioscience and Biotechnology, Konkuk University, 1 Hwayang-dong, Gwangjin-gu, Seoul 143-701, Korea. E-mail: and Hyong Joo Lee, Major in Biomodulation, WCU, Department of Agricultural Biotechnology, Research Institute for Agriculture and Life Sciences, Seoul National University, 599 Gwangak-ro, Gwanak-gu, Seoul 151-921, Korea. E-mail:


J. Neurochem. (2010) 112, 1415–1430.


Alzheimer’s disease (AD) is an age-related neurodegenerative disease increasingly recognized as one of the most important medical problems affecting the elderly. Although a number of drugs, including several cholinesterase inhibitors and an NMDA receptor antagonist, have been approved for use, they have been shown to produce diverse side effects and yield relatively modest benefits. To overcome these limitations of current therapeutics for AD, extensive research and development are underway to identify drugs that are effective and free of undesirable side effects. Certain naturally occurring dietary polyphenolic phytochemicals have received considerable recent attention as alternative candidates for AD therapy. In particular, curcumin, resveratrol, and green tea catechins have been suggested to have the potential to prevent AD because of their anti-amyloidogenic, anti-oxidative, and anti-inflammatory properties. These polyphenolic phytochemicals also activate adaptive cellular stress responses, called ‘neurohormesis’, and suppress disease processes. In this commentary, we describe the amyloid-β-induced pathogenesis of AD, and summarize the intracellular and molecular targets of selected dietary phytochemicals that might slow the progression of AD.

Abbreviations used:

Alzheimer’s disease


amyloid precursor protein


Swedish mutant APP







Huntington’s disease


heme oxygenase-1


heat-shock protein






mitogen-activated protein kinase


nuclear factor-κB


nitric oxide


non-steroidal anti-inflammatory drugs


Parkinson’s disease


reactive oxygen species


soluble APP-α


transforming growth factor-β


tumor necrosis factor-α

As individuals increasingly live to an older age, dementia, the best-known risk factor in aging, continues to grow as a serious public health problem. Alzheimer’s disease (AD) is the most prevalent dementia subtype and accounts for about 60% of all cases (van Marum 2008). In 2000, 25 million people were diagnosed with AD worldwide, a number that is expected to increase to 114 million by 2050 (Wimo et al. 2003). The earliest symptoms of AD often appear as subtle and intermittent deficits in the ability to remember minor events of everyday life. At later stages, AD gradually progresses to severe dementia, which affects multiple cognitive and behavioral functions. Fortunately, basic research has identified many of the pathways that contribute to this devastating disease, providing opportunities for the development of new treatments. Among the potential new treatment options are certain naturally occurring phytochemicals. These compounds have been the subject of increasing interest on the part of consumers and manufacturers because numerous epidemiological studies have suggested an association between the consumption of polyphenolic phytochemical-rich foods or beverages and the prevention of certain neurological diseases, including AD (Singh et al. 2008).

Current pharmacotherapy for AD

Post-mortem studies of AD patients in the early 1970s revealed that choline uptake and acetylcholine release are reduced in the brains of AD patients, reductions that are associated with substantial pre-synaptic cholinergic deficits (Francis et al. 1999). These observations led to the establishment of the ‘cholinergic-deficit hypothesis’, which posits that a shortage of acetylcholine caused many symptoms of AD, especially those related to learning difficulties (Francis et al. 1999; van Marum 2008). To slow the progression of AD and improve cognitive and general functioning, researchers tried to restore the cholinergic balance through inhibition of cholinesterase-mediated acetylcholine breakdown (Francis et al. 1999; van Marum 2008). The first drug developed for AD based on the cholinergic-deficit hypothesis was the cholinesterase inhibitor, tacrine. Tacrine was approved for the treatment of cognitive loss in patients with AD by the U.S. Food and Drug Administration in 1993 (van Marum 2008). Since that time, several other cholinesterase inhibitors, including donepezil (1996), rivastigmine (2000) and galantamine (2001), have been approved and used for AD treatment (van Marum 2008). In 2004, the U.S. Food and Drug Administration also approved memantine, a low-to-moderate affinity NMDA antagonist, for patients with moderate-to-severe AD (Lleo et al. 2006). By modulating NMDA receptor activity, memantine reduces abnormal excitatory glutamate signals, which are responsible for neuronal cell dysfunction and the eventual cell death observed in AD (Lleo et al. 2006; van Marum 2008).

Although multiple drugs have now been approved, their expected benefits are modest. For example, cholinesterase inhibitors improve overall memory deficits only temporarily, and the effects on cognitive test scores, behavioral measures, and functional outcomes are modest (Lleo et al. 2006; van Marum 2008). These cholinesterase inhibitors commonly cause gastrointestinal side effects, such as nausea and diarrhea (Table 1) (Lleo et al. 2006; van Marum 2008). Tacrine has also shown considerable hepatotoxicity, and its use has been especially limited by poor oral bioavailability (Lleo et al. 2006; van Marum 2008). Memantine has proved to be even less effective clinically than the cholinesterase inhibitors (van Marum 2008). The disadvantages and advantages of cholinesterase inhibitors and the NMDA receptor antagonist, memantine, in the treatment of cognitive impairment in patients with AD are described in Table 1. Because the benefits of these AD drugs are marginal and evident in only a subset of patients, their effectiveness has been questioned (Maggini et al. 2006; Roberson and Mucke 2006). Ultimately, accumulating evidence has led to the rejection of the cholinergic-deficit hypothesis; the modest beneficial effects of cholinesterase inhibitors might simply have been attributable to a restoration of acetylcholine balance in the brain (van Marum 2008). Fortunately, many new therapies that directly target the root cause of AD are now under investigation.

Table 1.   U.S. Food and Drug Administration-approved drugs for treating cognitive loss in patients with AD
  1. Modified from Lleo et al. (2006).

Cholinesterase inhibitors
 Cognex (tacrine)Nausea, diarrhea, urinary incontinence, hepatotoxicity 
 Aricept (donepezil)Nausea, diarrhea, vomiting, insomnia, nightmares, dizziness✓Temporarily improve memory deficits and provide some symptomatic relief
 Exelon (rivastigmine)Nausea, diarrhea, weight loss, vomiting, dizziness, fatigue, headache✓Beneficial, but modest, effects on cognitive test scores, behavioral measures, and functional outcomes
 Razadyne (galantamine)Nausea, vomiting, diarrhea, dizziness, weight loss, headache, abdominal pain, asthenia, somnolence 
NMDA receptor antagonist
 Namenda (memantine)Hallucinations, confusion, fatigue, dizziness, headache✓Beneficial, but modest, effects on cognitive test scores, behavioral measures, and functional outcomes
✓Beneficial in patients taking cholinesterase inhibitors; combined treatment may reduce side effects of either treatment modality

Amyloid hypothesis in the pathogenesis of AD

The successor to the cholinergic-deficit hypothesis – the ‘amyloid hypothesis’– was first introduced in 1992 to explain the pathophysiology of AD (Hardy and Higgins 1992). According to the amyloid hypothesis, abnormal proteolytic cleavage of amyloid precursor protein (APP) by β- and γ-secretase leads to excessive extracellular accumulation of amyloid-β (Aβ) in the AD brain (Selkoe 1996). Increased Aβ production or aggregation, or deficient clearance of Aβ, results in the build-up of a variety of pathogenic Aβ assemblies (Roberson and Mucke 2006). Pathologically accumulated Aβ might directly interact with neuronal membranes or indirectly stimulate diverse intracellular signaling pathways to impair neuronal synapses and dendrites, and cause local oxidative stress reactions and sustained inflammatory responses (Figs 1 and 2) (Roberson and Mucke 2006; Heneka and O’Banion 2007; van Marum 2008). Through these various and multifactorial mechanisms, Aβ accumulation leads to the progressive loss of neurons, disintegration of neural circuits, and neurological decline characteristic of AD (Roberson and Mucke 2006). At its root, the amyloid hypothesis suggests that Aβ protein is the key to the profound neuronal and synaptic degeneration in the brain regions implicated in learning and memory (Hardy and Higgins 1992; Mattson 2004).

Figure 1.

 Aβ-mediated oxidative stress in the pathogenesis of AD, and possible molecular targets of the inhibitory actions of curcumin, resveratrol, and catechin.

Figure 2.

 Aβ-mediated inflammation in the pathogenesis of AD, and possible molecular targets of the inhibitory actions of curcumin, resveratrol, and catechin.

New approaches for preventing and treating AD have emphasized strategies to reduce the pathogenesis of Aβ peptides (Roberson and Mucke 2006). Many new therapies that seek to reduce Aβ levels by directly targeting the mechanisms that regulate Aβ production are currently under investigation (Roberson and Mucke 2006). γ-Secretase inhibitors for AD treatment have reached clinical trials (Roberson and Mucke 2006). β-Secretase, which cleaves APP, is another prime target of efforts to inhibit Aβ production (Roberson and Mucke 2006). Stimulating α-secretase, which cleaves APP within Aβ and yields biologically inactive peptides, also effectively reduces Aβ production (Roberson and Mucke 2006). Small molecules that disrupt Aβ aggregation have also shown positive results in animal models of AD (Roberson and Mucke 2006).

In addition to pharmacological strategies that target key steps in the control of Aβ production, Aβ immunotherapy has received considerable attention as a promising approach for reducing the level of Aβ in the central nervous system of AD patients (Schenk et al. 1999; Bard et al. 2000; Janus et al. 2000; Morgan et al. 2000; DeMattos et al. 2001; Dodart et al. 2002; Kotilinek et al. 2002). The focus of the majority of this research has been on (i) active immunotherapy, exemplified by the AN1792 vaccine (QS-21), a pre-aggregated synthetic Aβ1–42 preparation; and (ii) passive immunization using injections of previously prepared polyclonal anti-Aβ antibodies (intravenous immunoglobulin) (Bard et al. 2000; DeMattos et al. 2001; Dodart et al. 2002; Kotilinek et al. 2002; Hock et al. 2003; Gilman et al. 2005; Masliah et al. 2005). Immunotherapy with active vaccination against Aβ has been shown to reduce cerebral plaque load and improve deficits in learning and memory in several independent transgenic animal models of AD (Schenk et al. 1999; Janus et al. 2000; Morgan et al. 2000). However, initial clinical trials using full-length aggregated Aβ1–42 (AN1792) had to be stopped in phase II because of the development of aseptic meningoencephalitis in 6% of the treated patients (Orgogozo et al. 2003; Ferrer et al. 2004). A new generation of Alzheimer’s vaccines based on AFFITOPE technology, which uses short peptides that mimic parts of the native Aβ sequence as their antigenic component, is currently under development (Schneeberger et al. 2009). Because these vaccines do not rely on full-length Aβ or Aβ fragments, their specificity can be more precisely controlled, preventing cross-reactivity with APP and potentially improving the clinical profile of such AFFITOPE-based therapies.

Aβ-mediated oxidative stress in the pathogenesis of AD

The central nervous system is very rich in polyunsaturated fatty acids and has a high content of transition metals and ascorbate, which together act as potent pro-oxidants (Pratico 2008). Because the brain possesses a relative paucity of antioxidant systems compared with other organs, it is very prone to oxidative imbalance and is thus vulnerable to oxidative damage (Pratico 2008). This susceptibility may come into play in the pathogenesis of AD, as evidenced by the fact that the brains of AD patients consistently exhibit evidence of injury caused by reactive oxygen species (ROS) and reactive nitrogen species during the early development of the disease (Pratico and Sung 2004; Pratico 2008). Importantly, the evidence of oxidative stress has been shown to precede the formation of senile plaques in the brain, the neuropatholgical hallmark of AD (Petersen et al. 2007). Markers of oxidative stress observed in the brains of AD patients are described in Table 2.

Table 2.   Oxidative modifications of biomolecules in AD brains
  1. Modified from Pratico (2008).

DNA and RNA oxidation
 DNA breaks, DNA nicking, DNA fragmentation
Protein oxidation
 Protein carbonyls
 Advanced glycation end products
Lipid peroxidation
 Thiobarbituric acid reactive substrates

The source of oxidant species in the AD brain may include unbound transition metals, damaged mitochondria and Aβ peptides themselves, as illustrated in Fig. 1 (Reddy and Beal 2008). Aggregated Aβ is neurotoxic and spontaneously produces free radical spin adducts in aqueous solutions (Hensley et al. 1995). Increased iron (Fe2+), copper (Cu2+), and zinc (Zn2+) levels are observed in the brains of AD patients, and these high levels of transition metals stimulate free radical generation (Good et al. 1992; Lovell et al. 1998). Aβ produces hydrogen peroxide (H2O2), which, in the presence of Fe2+, leads to hydroxyl radical formation via the Fenton reaction (Schubert and Chevion 1995). In the brains of AD patients and transgenic mouse models of AD, Aβ and APP are present in mitochondrial membranes and disrupt mitochondrial electron transport (Pratico 2008; Reddy and Beal 2008). Dysfunctional mitochondria release excessive amounts of H2O2 and ultimately cause irreversible cellular damage in the brain (Reddy and Beal 2008).

In defective AD neurons, oxidative stresses and free radicals drive modifications in almost every critical cellular macromolecule, including DNA, RNA, protein, and lipid (Pratico 2008). ROS produced endogenously and exogenously cause functional and structural alterations in these biomolecules and ultimately lead to neuronal cell death and tissue damage (Fig. 1). In addition to modifying macromolecules, oxidative stresses can further enhance Aβ formation by increasing APP levels and modulating the activity and expression of β-secretase and γ-secretase (Pratico 2008).

Thus, antioxidant enzymes and antioxidants have been demonstrated to protect against Aβ-induced cytotoxicity. Increases in the activity of superoxide dismutase and catalase in particular have been shown to block Aβ-induced neuronal cell death (Behl et al. 1994; Qin et al. 2002). Pre-clinical and clinical studies have suggested an association between dietary intake of antioxidants and a reduced risk for AD (Gonzalez-Gross et al. 2001; McDaniel et al. 2003).

Aβ-mediated inflammation in the pathogenesis of AD

Recent evidence suggests that inflammatory responses may significantly contribute to the progression and chronicity of AD (Heneka and O’Banion 2007). Inflammatory changes, including the presence of activated microglia, are observed throughout the AD brain, but particularly at amyloid deposits (Heneka and O’Banion 2007). As shown in Fig. 2, fibrillar Aβ can bind the complement factor C1 and potentially activate the classical complement pathway, which plays an important role in local recruitment and activation of microglia (Rogers et al. 1992; Heneka and O’Banion 2007). Aβ also activates microglia directly by binding to receptors, including the receptor for advanced glycation end products (Yan et al. 1998), the lipopolysaccharide (LPS) receptor CD14 (Fassbender et al. 2004), and other scavenger receptors (Paresce et al. 1996). Prolonged simulation of microglia induces the release of large amounts of pro-inflammatory mediators, including complement components, cytokines, various free radicals and nitric oxide (NO), all of which potentially contribute to further neuronal dysfunction and eventual death (Heneka and O’Banion 2007).

Complement expression is up-regulated in the AD brain, especially in the areas of primary pathology, including the entorhinal cortex, hippocampus, and mid-temporal gyrus (Yasojima et al. 1999). In a C1q complement-deficient transgenic mouse model of AD, the number of activated glia is decreased and neuronal integrity is improved, suggesting that C1q exerts a detrimental effect on neuronal integrity in AD, most likely by activating the complement cascade and enhancing inflammatory responses (Fonseca et al. 2004). Aβ also activates the nuclear factor-κB (NF-κB)-dependent cytokine production pathway (Combs et al. 2001) and stimulates the production of chemokines, monocyte chemoattractant protein-1, and macrophage inflammatory protein-1α/β in human macrophages and astrocytes (Smits et al. 2002). A number of cytokines, including interleukin-1α/β (IL-1α/β), IL-6 and tumor necrosis factor-α (TNF-α), are highly elevated in the brains of AD patients (Akiyama et al. 2000; Tuppo and Arias 2005). Chemokines and chemokine receptors are also up-regulated in the AD brain and might play an important role in recruiting microglia and astroglia to the sites of Aβ deposition (Xia and Hyman 1999; Heneka and O’Banion 2007). Recruited microglia and astroglia further produce additional cytokines, chemokines, and other neurotoxic pro-inflammatory molecules, thus perpetuating a deleterious cascade (Heneka and O’Banion 2007). The neuroinflammatory mediators observed in Aβ-challenged cell culture systems or AD brains are summarized in Table 3.

Table 3.   Activated neuroinflammatory responses found in Aβ-challenged cell culture systems or AD brains
  1. Modified from McGeer and McGeer (2001) and Tuppo and Arias (2005).

Complement proteins
 C1 through C9
Acute phase proteins
 Amyloid P component
 C-reactive protein
 Monocyte chemoattractant protein-1
 Macrophage inflammatory protein-1α/β
 Inducible nitric oxide synthase
 Free radicals and NO

Apart from self-propagation and direct cytopathic effects on neurons, cytokines may contribute more directly to the pathogenesis of AD (Heneka and O’Banion 2007). Studies performed in transgenic animals suggest that cerebral amyloid deposition is increased under inflammatory conditions (Guo et al. 2002). Moreover, these animals do not develop amyloid plaques unless inflammation is induced (Guo et al. 2002). Several lines of evidence suggest that inflammatory molecules either directly influence APP processing or increase the susceptibility to Aβ deposition/aggregation (Heneka and O’Banion 2007). For example, the combination of TNF-α and interferon-γ induces the production of Aβ peptide (Blasko et al. 1999), and IL-1 has been implicated in the transformation of diffuse Aβ into Aβ plaques (Akiyama et al. 2000). In addition, cytokines transcriptionally up-regulate the enzymatic activity of β-secretase, a key enzyme for generating Aβ from APP (Sastre et al. 2003). An association of AD with several polymorphisms of pro-inflammatory genes, including those for IL-1 (Nicoll et al. 2000), IL-6 (Papassotiropoulos et al. 1999), TNF-α (McCusker et al. 2001; Perry et al. 2001) and acute phase protein (Kamboh et al. 1995), has also been reported. These observations suggest a new mechanism in which inflammatory components themselves exacerbate the fundamental pathology of AD (Heneka and O’Banion 2007).

The potentially significant contribution of inflammatory mechanisms to the progression and chronicity of AD has prompted consideration of anti-inflammatory treatment strategies. In this context, non-steroidal anti-inflammatory drugs (NSAIDs) have been shown to down-regulate the pro-inflammatory signals of microglia and astrocytes and reduce Aβ production (Breitner et al. 1994). NSAIDs appear to exert a protective effect against the development of AD (Breitner et al. 1994; in t’ Veld et al. 2001; Zandi et al. 2002). However, their potential beneficial effects apparently do not extend to cases of pre-existing AD (Lleo et al. 2006): clinical trials in patients with established AD reported little (Rogers et al. 1993; Scharf et al. 1999; Aisen et al. 2002) or no (Aisen 2000; Van Gool et al. 2001; Aisen et al. 2003) benefit of NSAIDs.

Selected polyphenolic phytochemicals for the prevention of AD

Phytochemicals are plant-derived chemical compounds that have potential health-promoting properties. A wide variety of phytochemicals has been shown to prevent certain chronic diseases, such as cancers and cardiovascular diseases, by mitigating or correcting cellular dysfunctions (Manach et al. 2005; Duthie 2007). Polyphenolic phytochemicals are the most abundant dietary antioxidants; however, numerous studies performed in animal models or cell culture demonstrated that the antioxidant activity of these compounds is unlikely to be the sole explanation for their protective cellular effects.

Among the phytochemicals of emerging interest are curcumin, resveratrol, and green tea catechins, based on epidemiological studies that have suggested a positive relationship between consumption of these compounds and the prevention of AD (Singh et al. 2008). Specially, curcumin (diferuloyl methane), resveratrol (trans-3,4′,5-trihydroxystilbene), and (−)-epigallocatechin-3-gallate (EGCG) exhibit antioxidant activity as well as anti-amyloidogenic properties which reduce the formation of neurotoxic Aβ fibrils (see the structure in Fig. 3) (Ono et al. 2004; Marambaud et al. 2005; Rezai-Zadeh et al. 2005). In the following sections, we summarize the intracellular and molecular targets of selected phytochemicals that might be capable of slowing the progression of AD (Table 4).

Figure 3.

 The structures of curcumin, resveratrol, and (−)-epigallocatechin-3-gallate.

Table 4.   Molecular evidence for beneficial effects of selected phytochemicals against AD
 ↓ Formation of Aβ fibrils from fresh Aβ in in vitro (Ono et al. 2004)
 ↓ Preformed Aβ fibrils in in vitro (Ono et al. 2004)
 ↓ Aβ-induced expression of cytokines (TNF-α and IL-1β) and chemokines (macrophage inflammatory protein-1β, monocyte chemoattractant    protein-1 and IL-8) (Giri et al. 2004)
 – Protects cells from Aβ-induced toxicity in multiple in vitro culture systems (Kim et al. 2001)
 – Crosses blood–brain barrier (Yang et al. 2005; Garcia-Alloza et al. 2007)
 ↓ Aβ deposits in the brain of aged Tg2576 mice (Lim et al. 2001; Yang et al. 2005) as well as Aβ-injected rats (Frautschy et al. 2001)
 ↓ Senile plaques in the brains of Tg2576 mice (Lim et al. 2001; Yang et al. 2005; Garcia-Alloza et al. 2007)
 ↓ Oxidized proteins in the brains of Tg2576 mice (Lim et al. 2001)
 ↓ IL-1β and pro-inflammatory cytokines elevated in the brains of Tg2576 mice (Lim et al. 2001)
 ↑ Post-synaptic density-95 in the brains of Aβ-injected rats (Frautschy et al. 2001)
 ↓ Cognitive deficits in animal models of AD (Frautschy et al. 2001; Ishrat et al. 2009)
 ↓ Production of Aβ peptides in vitro (Marambaud et al. 2005)
 – Protects cells from Aβ-induced toxicity in multiple in vitro culture systems (Jang and Surh 2003; Savaskan et al. 2003; Han et al. 2004)
 – Crosses blood–brain barrier (Wang et al. 2002; Mokni et al. 2007)
 ↓ Malondialdehyde in animal models (Sharma and Gupta 2002; Kumar et al. 2007)
 ↓ Cognitive impairment and memory deficit in animal models (Sharma and Gupta 2002; Sharma et al. 2005; Kumar et al. 2007)
Green tea catechins, including EGCG
 ↓ Translation of APP mRNA and membrane-bound holoprotein APP (Levites et al. 2003; Singh et al. 2008)
 ↑α-Secretase (α-desintegrin and metalloprotease or TNF-α converting enzyme) activity, cleavage of the α-C-terminal fragment of APP and  soluble amyloid precursor protein-α (Levites et al. 2003; Rezai-Zadeh et al. 2005; Obregon et al. 2006)
 ↓ Activity of β-secretase (Jeon et al. 2003)
 ↓ Production of Aβ peptides in APP695 over-expressing neurons (Rezai-Zadeh et al. 2005)
 ↓ Formation of Aβ fibrils by directly binding to the native unfolded Aβ (Ono et al. 2003; Bastianetto et al. 2006; Ehrnhoefer et al. 2008)
 ↓ Preformed Aβ fibrils in vitro (Ono et al. 2003)
 ↓ Aβ-induced levels of malondialdehyde in hippocampal neuronal cells (Choi et al. 2001)
 ↓ Aβ/IL-1β-induced IL-6, IL-8, COX-2 and prostaglandin E2 in human astrocytoma U373MG cells (Kim et al. 2007b)
 ↓ Aβ-induced caspase activity in hippocampal neuronal cells (Choi et al. 2001)
 ↓ Aβ-induced neuronal cell death (Choi et al. 2001; Levites et al. 2003; Bastianetto et al. 2006)
 – Brain permeable (Nakagawa and Miyazawa 1997; Suganuma et al. 1998; Mandel et al. 2006)
 ↓ Aβ levels and plaques in Tg2576 AD mice (Rezai-Zadeh et al. 2005, 2008)
 ↓ Hippocampal lipid peroxide and cortico-hippocampal ROS in animal models of AD (Haque et al. 2008)
 ↓ Tau pathology in Tg2576 AD mice (Rezai-Zadeh et al. 2008)
 ↓ Cognitive impairment in animal models of AD (Haque et al. 2008; Rezai-Zadeh et al. 2008)


Curcumin, a yellow pigment present in the rhizome of turmeric (Curcuma longa), is used as a food preservative and a spice to impart a characteristic flavor to Indian curries (Singh et al. 2008). Interestingly, it was reported that the prevalence of AD in people 70–79 years of age in India is 4.4-fold less than that in the United States, suggesting that a diet rich in curcumin might be responsible for the reduced AD risk in aged Indians (Ganguli et al. 2000).

Substantial in vitro and in vivo evidence indicates that curcumin has anti-amyloidogenic, anti-oxidative, and anti-inflammatory properties that carry the potential to prevent AD (Ringman et al. 2005). In vitro studies have demonstrated that curcumin and its analog, rosmarinic acid, have anti-amyloidogenic properties, dose-dependently inhibiting the formation and extension of neurotoxic Aβ fibrils from fresh Aβ, and destabilizing preformed Aβ fibrils to regenerate Aβ monomers (Ono et al. 2004). Inhibiting Aβ fibril formation represents an attractive therapeutic strategy for the treatment of AD; thus, given its primary effect on Aβ fibrilization, curcumin is a promising agent for use in treating and/or preventing AD. Although the precise anti-amyloidogenic mechanism of curcumin is not clear, it has been suggested that the compact and symmetric structure of curcumin, which has two 3,4-methoxyhydroxyphenyl rings bound by a short carbohydrate chain, might be capable of specifically binding free Aβ and subsequently inhibiting polymerization of Aβ into Aβ fibrils (Ono et al. 2004). Alternatively, curcumin may specifically bind to Aβ fibrils and subsequently destabilize the β-sheet-rich conformation of Aβ in Aβ fibrils (Ono et al. 2004).

Curcumin exhibits potent antioxidant activity, a property that led to its early use as a food preservative (Ringman et al. 2005). In fact, curcumin is a much stronger free radical scavenger than vitamin E (Zhao et al. 1989). Curcumin scavenges NO-based radicals and protects the brain from lipid peroxidation (Wei et al. 2006). It also functions as a hydroxyl radical scavenger to prevent 8-hydroxy-2-deoxyguanosine formation within the DNA molecule, preventing oxidative damage of DNA in mouse fibroblasts (Shih and Lin 1993). In in vitro experiments, curcumin (or its metabolites) at physiologically relevant concentrations was shown to bind Cu2 + and Fe2 + ions (Baum and Ng 2004). Because these redox-active metals exacerbate Aβ aggregation or cause subsequent oxidative damage in the AD brain, curcumin may prevent reactive metal-mediated aspects of AD pathogenesis (Baum and Ng 2004). In this context, copper–curcumin complexes act as radical scavengers and exhibit superoxide dismutase-mimetic properties (Baum and Ng 2004). Curcumin also activates glutathione S-transferase (Nishinaka et al. 2007) and partially restores glutathione content in the brain (Ishrat et al. 2009). Curcumin induces the antioxidant enzyme, heme oxygenase-1 (HO-1) (Motterlini et al. 2000), which has been shown to increase tolerance of the brain to stresses and serve an important anti-degenerative function in AD pathogenesis (Schipper 2004). Additional evidence for the antioxidant property of curcumin is provided by studies using an AD transgenic mouse model, in which feeding curcumin reduced brain levels of oxidized proteins containing carbonyl groups (Lim et al. 2001).

Curcumin has been reported to suppress the activity of the transcription factors, NF-κB and activator protein-1, and regulate inflammatory responses (Singh and Aggarwal 1995; Nanji et al. 2003; Bengmark 2006; Sandur et al. 2007; Shishodia et al. 2007). Curcumin is capable of blocking the induction of inducible nitric oxide synthase, possibly by inhibiting NF-κB activation (Nanji et al. 2003; Bengmark 2006; Sandur et al. 2007). Curcumin inhibits lipoxygenase and cyclooxygenase-2 (COX-2), which are responsible for the synthesis of pro-inflammatory leukotrienes, prostaglandins and thromboxanes (Nanji et al. 2003; Bengmark 2006; Rao 2007; Sandur et al. 2007). Importantly, curcumin has been shown to abrogate Aβ-induced expression of cytokines and chemokines in both peripheral blood monocytes and THP-1 monocytic cells (Giri et al. 2004). In Tg2576 AD-like mice, curcumin significantly reduced the elevated level of IL-1β (Lim et al. 2001). Given the demonstrated importance of Aβ-mediated inflammatory processes in the pathogenesis of AD, the anti-inflammatory characteristic of curcumin may prove beneficial for the prevention or treatment of AD.

The anti-AD effects of curcumin have been demonstrated in several animal models of AD. When fed to aged Tg2576 mice, low doses of curcumin reduced the level of Aβ plaques in the brain (Lim et al. 2001). Multi-photon microscopy demonstrated that curcumin can cross the blood–brain barrier to target senile plaques and disrupt existing plaques in these mice (Yang et al. 2005; Garcia-Alloza et al. 2007). The preventive effects of curcumin were demonstrated in a rat model as well, in which infusion of human Aβ1–40 and Aβ1–42 into the cerebral ventricles of aged female rats produces an AD-like phenotype (Frautschy et al. 2001). Curcumin-fed/Aβ-injected rats exhibited better memory function compared to Aβ-injected rats (Frautschy et al. 2001). In rats fed curcumin, there was an increase in post-synaptic density-95, a post-synaptic marker that plays a fundamental role in synaptic transmission, suggesting that curcumin might prevent Aβ-mediated synaptic deficits in these rats (Frautschy et al. 2001). These observations raise the possibility that dietary supplementation with curcumin could be an effective prophylaxis against AD.


Resveratrol, a polyphenolic phytoalexin found in grapes, red wine and berries, has been reported to possess a wide range of biological and pharmacological activities (Soleas et al. 1997; Singh et al. 2008). Several epidemiological studies have indicated an inverse relationship between wine consumption and the incidence of AD (Orgogozo et al. 1997; Lindsay et al. 2002), leading to speculation that resveratrol might contribute to the beneficial effect of wine in AD patients. Resveratrol, like curcumin, readily crosses the intact blood-brain barrier and penetrates into brain tissue (Wang et al. 2002; Mokni et al. 2007). Resveratrol has been reported to possess potent neuroprotective properties in several models, both in vitro and in vivo (Virgili and Contestabile 2000; Zamin et al. 2006; Ates et al. 2007; Della-Morte et al. 2009). Recent evidence indicates that resveratrol can reduce neuronal cell death and mitigate cerebral damage after ischemia/hypoxia, trauma, and excitotoxicity (Virgili and Contestabile 2000; Zamin et al. 2006; Ates et al. 2007; Della-Morte et al. 2009).

Resveratrol is reported to possess anti-amyloidogenic activity, reducing the level of secreted or intracellular Aβ peptides in several cell lines expressing Swedish mutant APP695; however, it does not affect the Aβ-producing enzymes, β- or γ-secretase (Marambaud et al. 2005). Selective proteasome inhibitors and siRNA-directed silencing of the proteasome subunit β5 prevented the resveratrol-induced decrease in Aβ levels, suggesting that resveratrol suppresses the level of Aβ by modulating the proteasome (Marambaud et al. 2005). Resveratrol also protected SH-SY5Y neuroblastoma cells and primary hippocampal neurons from Aβ-induced toxicity, an effect that was associated with resveratrol-induced activation of protein kinase C (Savaskan et al. 2003; Han et al. 2004).

Several lines of experimental evidence suggest that resveratrol may block oxidative stresses involved in the pathogenesis of AD, making resveratrol a promising therapeutic agent to prevent or treat AD. Resveratrol can scavenge free radicals and protect neurons and microglia (Jang and Surh 2003; Savaskan et al. 2003; Zhuang et al. 2003; Candelario-Jalil et al. 2007). Pre-treatment of PC-12 rat pheochromocytoma cells with resveratrol attenuated Aβ-induced intracellular ROS accumulation and subsequently suppressed NF-κB activation and apoptotic features (Jang and Surh 2003). Resveratrol also up-regulates cellular antioxidants, including glutathione, induces the gene expression of phase 2 enzymes, and protects against oxidative and electrophilic injury (Cao and Li 2004). Like curcumin, resveratrol potentiates the neurohormetic HO-1 pathway in primary neuronal cultures (Zhuang et al. 2003). In the brain of rats, resveratrol exerted a neuroprotective action against oxidative stresses induced by streptozotocin or the microtubule-disrupting agent, colchicine (Sharma and Gupta 2002; Kumar et al. 2007). Chronic administration of resveratrol also significantly reduced the elevated levels of malondialdehyde in these rats (Sharma and Gupta 2002; Kumar et al. 2007).

Resveratrol not only reduced neuronal cell death but also suppressed activation of astrocytes and microglia (Wang et al. 2002; Bi et al. 2005; Candelario-Jalil et al. 2007). In an earlier study, resveratrol was found to suppress LPS-induced TNF-α and NO production in the N9 mouse microglial cell line by inhibiting NF-κB activation and p38 mitogen-activated protein kinase (MAPK) phosphorylation (Bi et al. 2005). Resveratrol has been shown to block the expression of COX-2 and inducible nitric oxide synthase, possibly by inhibiting NF-κB activation (Rahman et al. 2006). Resveratrol significantly diminished LPS-induced free radical formation and expression of microsomal prostaglandin E synthase-1, the most important and terminal synthase responsible for prostaglandin E2 synthesis, in activated microglial cells (Candelario-Jalil et al. 2007). Low concentrations of resveratrol dramatically reduced the production of 8-iso-prostaglandin F2α, a reliable indicator of free radical generation (Candelario-Jalil et al. 2007). Because the reduction in 8-isoprostane formation is a direct and potent contributor to the subsequent attenuation of microglial activation, the neuroprotective effects of resveratrol might be due in part to the attenuation of neuroinflammatory responses (Candelario-Jalil et al. 2007).

The effects of resveratrol on memory function have been tested in several animal models. It was reported that trans-resveratrol prevents cognitive impairment and spatial memory deficits (Sharma and Gupta 2002; Sharma et al. 2005; Kumar et al. 2007). For example, central administration of colchicine caused cognitive dysfunction in rats that was associated with excessive free radical generation and a loss of cholinergic neurons (Kumar et al. 2007). Chronic treatment with resveratrol significantly restored acetylcholinesterase activity in the brains of colchicine-injected rats and improved the colchicine-induced cognitive impairment (Kumar et al. 2007). Collectively, these findings suggest that resveratrol has chemopreventive or therapeutic potential in AD.

Green tea catechins

Catechin intake has been associated with a wide variety of beneficial health effects. Green tea catechins are brain permeable (Nakagawa and Miyazawa 1997; Suganuma et al. 1998; Mandel et al. 2006), and have been increasingly ascribed neuroprotective actions in recent studies (Rezai-Zadeh et al. 2005, 2008; Singh et al. 2008). Green tea catechins have been reported to serve several molecular biological roles, including activation of MAPKs, protein kinase C, antioxidant enzymes and survival genes, and control of calcium homeostasis and APP processing (Chen et al. 2000; Ishige et al. 2001; Levites et al. 2001, 2002, 2003). Several in vitro studies have shown that green tea extract may protect neurons from Aβ-induced damages (Choi et al. 2001; Levites et al. 2003; Bastianetto et al. 2006).

EGCG, a major catechin isolated from the polyphenolic fraction of green tea, reduced Aβ generation in APP695-over-expressing neurons and decreased Aβ levels and plaques in Tg2576 AD mice (Rezai-Zadeh et al. 2005). These effects were associated with elevated α-secretase cleavage activity and increased generation of the α-C-terminal fragment of APP and soluble amyloid precursor protein-α, indicating that EGCG stimulates the non-amyloidogenic α-secretase proteolytic pathway (Levites et al. 2003; Rezai-Zadeh et al. 2005). EGCG was also found to significantly increase the expression of TNF-α converting enzyme, an APP α-secretase candidate (Rezai-Zadeh et al. 2005). In addition, EGCG markedly increased the protein level of active α-desintegrin as well as metalloprotease 10, another APP α-secretase candidate, ultimately leading to non-amyloidogenic APP processing (Obregon et al. 2006). Furthermore, EGCG acted as a β-secretase inhibitor in a cell-free system, raising the possibility that EGCG-mediated inhibition of Aβ generation is accomplished via blockade of β-secretase activity (Jeon et al. 2003). Prolonged administration of EGCG to mice has been shown to down-regulate holo-APP in the hippocampus, suggesting that EGCG might reduce Aβ levels by suppressing the expression of holo-APP (Levites et al. 2003). Green tea catechins are also able to inhibit the formation, extension, and stabilization of Aβ fibrils (Ono et al. 2003). EGCG specifically has been shown to efficiently inhibit the fibrillogenesis of Aβ by directly binding to the native unfolded polypeptides, possibly through formation of stable hydrogen bonds, and preventing their conversion into toxic, on-pathway aggregation intermediates (Ehrnhoefer et al. 2008).

Overall, the antioxidant properties of catechins are more potent than those of α-tocopherol or vitamin C and E (Zhao et al. 1989). Catechins exert their antioxidative activity by chelating metal ions, such as iron (Fe2+) and copper (Cu2+), and preventing the generation of potentially damaging free radicals (Singh et al. 2008). After the oxidation of catechins by free radicals, a dimerized product with an increased iron-chelating potential and ability to scavenge superoxide anions is formed (Singh et al. 2008). It has been suggested that using EGCG chelation to reduce the free iron pool might suppress the translation of APP mRNA and influence the level of APP (Singh et al. 2008). Catechins may also transfer an electron to ROS-induced radical sites on DNA and thereby prevent oxidative DNA modifications (Singh et al. 2008). In addition, EGCG scavenges ROS and inhibits lipid peroxidation (Choi et al. 2001). In cells exposed concurrently to Aβ, EGCG decreases malondialdehyde levels and caspase activity, protecting against Aβ-induced apoptosis and enhancing hippocampal neuronal survival (Choi et al. 2001). EGCG also mitigates Aβ-induced oxidative stress in vivo, reducing hippocampal lipid peroxide in the brains of rats (Haque et al. 2008).

The anti-inflammatory properties of catechins are well documented. EGCG inhibits the activation of MAPK and NF-κB, and subsequently attenuates the production of IL-6, IL-8, and vascular endothelial growth factor in human astrocytoma U373MG cells (Kim et al. 2007b). EGCG also suppresses the inflammatory activities of various cytokines (Ahmed et al. 2002; Han 2003) and attenuates IL-1- and Aβ-induced COX-2 expression and prostaglandin E2 production (Kim et al. 2007b). It has also been reported that EGCG inhibits LPS-induced microglial activation and protects against inflammation-mediated neuronal injury (Li et al. 2004).

EGCG modulates Aβ-mediated tau pathology in Tg2576 mice, reducing potentially toxic sarkosyl-soluble phospho-tau isoforms (Rezai-Zadeh et al. 2008). In a recent study, heat-shock protein (HSP) 90 inhibitors were found to reduce the levels of soluble phospho-tau isoforms (Dickey et al. 2007). EGCG has been found to directly bind HSP90 and inhibit its activity (Palermo et al. 2005); thus, EGCG might modulate the level of phospho-tau through the inhibition of HSP90.

The effects of green tea catechins on memory function have been tested in several animal models. For example, EGCG reduces cognitive impairment in Tg2576 mice (Rezai-Zadeh et al. 2008). Rats provided drinking water with high levels of green tea catechins (mostly EGCG) for 5 months showed less memory impairment following intracerebroventricular injection of Aβ1–40 than those supplied normal drinking water (Haque et al. 2008). In a mouse model of cerebral ischemia, injection of tea catechins ameliorated hippocampal neuronal damage and memory impairment (Matsuoka et al. 1995; Lee et al. 2003). Thus, treatment with green tea catechins, especially EGCG, might be a viable therapeutic approach for treating AD-like brain neuropathologies and associated cognitive impairment.

Neuroprotective hormetic responses induced by polyphenolic phytochemicals

The term hormesis has long been used to describe the phenomenon where a specific chemical is able to induce biologically opposite effects at different doses; most commonly is a stimulatory or beneficial effect at low doses and an inhibitory or toxic effect at high doses (Calabrese et al. 2007a). Phytochemicals have presumably evolved, in part, to allow plants to cope with environmental challenges including exposure to radiation and toxins, and defense against pests and infectious agents (Tuteja et al. 2001; Huffman 2003). Many of these phytochemicals exhibit hormetic properties, being harmful at high doses yet beneficial at relevantly low doses (Son et al. 2008).

Since ingested in low concentrations by human, dietary polyphenolic phytochemicals are reported to activate adaptive cellular stress responses, by which cells have the ability to counteract stressful conditions (Calabrese et al. 2008, 2009b). Adaptive cellular stress responses require the activation of pro-survival pathways and the production of molecules with anti-oxidative and anti-apoptotic activities (Calabrese et al. 2007b, 2008, 2009a). In the field of neuroscience, the adaptive process by which neurons (and hence the nervous systems and organisms) respond to a moderate level of stress by enhancing their ability to resist a more severe stress that might otherwise be lethal or cause dysfunction or disease, refer to ‘neurohormesis’ (Mattson and Cheng 2006). As one of the main neuroprotective intracellular systems, the vitagene system is emerging as a neurohormetic potential target for novel cytoprotective interventions of polyphenolic phytochemicals (Calabrese et al. 2008). Vitagenes encode survival proteins such as HSP70 and HO-1 as well as thioredoxin/thioredoxin reductase (Calabrese et al. 2008, 2009b). In several cases, neurohormetic phytochemicals have been shown to suppress disease processes in animal models relevant to neurodegenerative disorders such as Alzheimer’s disease (Mattson et al. 2007).

Curcumin has been reported to activate the expression of several intracellular defense systems both in vitro and in vivo. A hormetic mechanism of action of curcumin is suggested from studies showing that levels of expression of the stress response protein HO-1 were increased in cultured hippocampal neurons treated with curcumin (Scapagnini et al. 2006). At non-toxic concentrations, curcumin induces HO-1 expression by activating the NF-E2-related factor 2/antioxidant response element pathway both in vitro (Pae et al. 2007) and in vivo (Farombi et al. 2008). Resveratrol also activates neurohormetic transcription factor forkhead box O, resulting in the up-regulation of genes involved in antioxidant pathways (Frescas et al. 2005; Robb et al. 2008; Son et al. 2008). Green tea contains chemicals capable of activating adaptive cellular stress responses so protects neurons (Son et al. 2008). Catechins and epicatechins in green tea can induce the production of cytoprotective proteins (Mandel et al. 2005). EGCG up-regulates HO-1 expression by activation of the NF-E2-related factor 2/antioxidant response element pathway and confers resistance against cell death, suggesting a hormetic mechanism of action (Wu et al. 2006; Son et al. 2008).

Challenges for research on polyphenolic phytochemicals in AD

The anti-AD activities of these polyphenolic phytochemicals in vitro and in vivo have been documented, but it would be unwise to extrapolate these results to the human situation without proper clinical trials (Singh et al. 2008). Most reports on the beneficial effects of curcumin, resveratrol and green tea catechins are based on studies either in cell cultures or in animal models, where there is not extensive neuronal damage. On the contrary, in human clinical trials, the patients already suffer from extensive neuronal loss and damage. Therefore, these polyphenolic phytochemicals should be tested for their therapeutic efficacy in human. On the other hand, most cell culture or animal studies have been conducted on a short-term basis. More long-term studies should be undertaken to determine their beneficial effects in slowly developing AD.

No clinical trials of curcumin, resveratrol and green tea catechins in AD have been completed. Curcumin is generally accepted as safe, although some animal studies have suggested risks of gastric ulceration, thyroid follicular cell hyperplasia and hepatotoxicity at very high doses (Kelley and Knopman 2008). Although it is clear that curcumin has a wide variety of beneficial activities, not all studies are consistent with this positive picture. In one animal study, for example, curcumin showed to promote lung tumor (Dance-Barnes et al. 2009); however, the other study demonstrated that curcumin inhibits airway inflammation and lung cancer progression (Moghaddam et al. 2009). In case of green tea, it was touted to reduce cancer, but a recent comprehensive article was unable to arrive at this conclusion (Boehm et al. 2009). These are useful to remind that the many desirable medicinal effects of polyphenolic phytochemicals should not obscure the need for caution until the data have been fully assessed. Unexpected drug–drug interactions also need to be observed and included in the list of consideration for their use.

In human studies to date, doses of 1200 mg/day curcumin were well tolerated in general, although one study noted gastric irritation in two of 19 subjects receiving this dose (Kelley and Knopman 2008). Another study found that doses up to 8 g/day curcumin were well tolerated, and that higher doses were not tolerated simply due to the bulk of the agent (Kelley and Knopman 2008). A NIA-funded phase II trial of curcumin will enroll 33 subjects to determine the safety and tolerability of two doses (2 and 4 g/day) of curcumin (Kelley and Knopman 2008). A separate Chinese study will examine the safety and tolerability of curcumin in combination with ginkgo (Kelley and Knopman 2008). On the other hand, the Personnes Agees QUID study showed that people drinking three to four glasses of wine per day have 80% decreased incidence of dementia and AD compared with those who drank less or did not drink at all (Larrieu et al. 2004). Given the number of ongoing clinical trials, data on effectiveness, dose-efficacy and side effects of these polyphenolic phytochemicals should be available in the near future.


The studies highlighted here indicate that phenolic phytochemicals, such as curcumin from turmeric, resveratrol from grape and wine, and EGCG from green tea, not only exhibit potent anti-oxidative and anti-inflammatory properties, acting to scavenge radicals and regulate inflammatory responses, but also readily cross the blood–brain barrier to act on specific targets that have been implicated in the pathogenesis of AD. An important consideration in this context is their ability to attenuate the generation of neurotoxic Aβ. These compounds also activate endogenous pro-survival pathways including vitagenes to enhance cellular defense mechanisms, which may protect brain against Aβ proteins and deposition of plaques. These phytochemicals might also exert a protective effect against other neurodegenerative diseases including Parkinson’s disease (PD), Huntington’s disease (HD), and amyotrophic lateral sclerosis. For example, curcumin, resveratrol, and green tea catechins have been shown to protect against MPTP-, MPP+- and 6-hydroxydopamin-induced PD-like pathogenesis in vitro and in vivo (Choi et al. 2002; Gelinas and Martinoli 2002; Nie et al. 2002a,b; Zbarsky et al. 2005; Chen et al. 2006; Li et al. 2006; Jin et al. 2008; Lu et al. 2008; Rajeswari and Sabesan 2008; Bournival et al. 2009; Wang et al. 2009). Curcumin inhibits the aggregation of α-synuclein, the protein involved in the pathogenesis of PD (Pandey et al. 2008). On the other hand, resveratrol may be useful in HD, as evidenced by its ability to protect cells against the toxicity of mutant huntingtin protein in HD experimental models (Parker et al. 2005). Similarly, EGCG has been shown to modulate the early events in huntingtin misfolding and reduce toxicity in HD models (Ehrnhoefer et al. 2006). Both resveratrol and EGCG protect motor neurons against an amyotrophic lateral sclerosis-associated mutation of superoxide dismutase 1 (Koh et al. 2006; Kim et al. 2007a; Barber et al. 2009).

Because they are largely innocuous, neuroprotective phytochemicals derived from fruits and vegetables are attractive alternatives to pharmaceuticals such as NSAIDs and anti-degenerative molecules, which lack conclusively demonstrated clinical efficacy and are associated with significant safety concerns. Although there are clear limits for their immediate widespread use, dietary polyphenolic phytochemicals hold great promise as safe, inexpensive and readily available prophylactic agents for AD and other neurodegenerative diseases. Continued effort to extrapolate in vitro and in vivo results to the human situation through properly designed clinical trials should help to realize the potential of this class of compounds.


This work was supported by the Basic Research Program (KRF-2008-313-C00757), the World Class University Program (R31-2008-00-10056-0), and the Priority Research Centers Program (2009-0093824), National Research Foundation, Ministry of Education, Science and Technology, and the Korea Healthcare Technology R&D Project (A090964), Ministry for Health, Welfare and Family Affairs, Republic of Korea.