plant polyphenols as a double-edged sword in health

AGri-Bioscience Monographs, Vol. 6, No. 1, pp. 1–57 (2016) www.terrapub.co.jp/onlinemonographs/agbm/

Received on March 3, 2014Accepted on July 9, 2014Online published on

October 17, 2016

Keywords• polyphenol• antioxidant• electrophile• phase II drug metabolizing

enzyme• health promotion

© 2016 TERRAPUB, Tokyo. All rights reserved.doi:10.5047/agbm.2016.00601.0001

erized molecules such as tannins are reportedly identi-fied. Phenolic acids include caffeic acid, vanillic acid,and coumaric acid, that account for approximately one-third of the total dietary intake of polyphenols.Flavonoids, account for the remaining two-thirds ofpolyphenol intake, can be further subdivided into 13classes, including flavones, flavanones, flavonols,isoflavones, catechins (flavanols), anthocyanins andproanthocyanidins, with more than 5,000 identifiedcompounds. Epidemiological studies have supportedthe hypothesis that high dietary intake of polyphenolsis associated with decreased risk of a range of diseasesincluding cardiovascular disease, cancer andneurodegenerative diseases (Fraga 2007; Martin andAppel 2010). Since the French paradox was describedtwo decades ago (Renaud and Lorgeril 1992), several

AbstractPolyphenols in foods possess potential to exhibit various beneficial effects in health pro-motion. Although polyphenols are well known to have antioxidant properties, a wealth ofdata suggests that most of the relevant mechanisms of disease prevention by polyphenolsare not mainly related to their antioxidant properties. The protective function of phenoliccompounds, which can be metabolically converted into electrophiles, might be ascribedto the induction of the cytoprotective responses against oxidative stress or toxic chemi-cals. Based on compelling evidence regarding the beneficial potential of polyphenols,various polyphenol-rich dietary supplements are being developed for public use. It isrecently becoming apparent that commonly used dietary compounds can exert deleteri-ous effects at pharmacological (supraphysiological) doses. Moreover, some human inter-vention trials showed not only failure to protect by polyphenols but accelerated develop-ment of chronic diseases in susceptible subjects. Thus, the present monograph tries toprovide experimental evidence supporting the idea that optimal polyphenol supplemen-tation can confer beneficial effects but high doses elicit adverse effects in an electrophilicreaction-dependent manner, thereby establishing a double-edge sword in health promo-tion. Current knowledge of polyphenols as phase II inducers is also discussed, with con-sideration of the metabolic activation, binding targets and factors influencing the bio-logical activities of polyphenols. In addition, this monograph attempts to provide a briefperspective on the beneficial and harmful effects of food polyphenols.

Plant Polyphenols as a Double-Edged Sword inHealth Promotion: Lessons from the ExperimentalModels Using Simple Phenolic Acids

Yoshimasa Nakamura

Division of Agricultural and Life SciencesGraduate School of Environmental and Life ScienceOkayama University1-1-1 Tsushima-naka, Kita-ku, Okayama 700-8530, Japane-mail: [email protected]

1. Introduction

Increased consumption of fruit and vegetables hasbeen consistently recommended to globally reduce therisk of chronic diseases including cardiovascular dis-ease, cancer, hypertension, type-2 diabetes, and so on(Fraga 2007). The protection against chronic diseasesis regarded to be mainly due to the excessive digestionof bioactive compounds including polyphenols as wellas the nutrients (Bravo 1998). Polyphenols, contain-ing at least one aromatic ring with one or more hy-droxyl groups in addition to other substituents, areknown as secondary metabolites involved in the chemi-cal defense of plants against biotic and abiotic stresses.More than 8,000 distinct structures of polyphenols fromsimple phenolic acids and flavonoids to highly polym-

2 Y. Nakamura / AGri-Biosci. Monogr. 6: 1–57, 2016

doi:10.5047/agbm.2016.00601.0001 © 2016 TERRAPUB, Tokyo. All rights reserved.

studies have focused their attention on the polyphenoliccomponents of red wine in order to explain the inverseassociation observed between wine consumption andthe incidence of cardiovascular disease, thereby open-ing the debate of which type of polyphenols is moreprotective than others. Polyphenols have also beenshown to be effective against underlying mechanismsof disease processes. With respect to cardiovasculardisease, polyphenols may alter lipid metabolism, in-hibit low-density lipoprotein (LDL) oxidation, inhibitplatelet aggregation, improve endothelial function andreduce blood pressure (Arranz et al. 2012). Polyphenolshave also been shown to exert a variety of anti-carci-nogenic effects, including an ability to induce apoptosisin tumor cells, inhibit cancer cell proliferation and pre-vent angiogenesis and tumor cells invasion (Arranz etal. 2012).

Oxidative stress is one of the causative factors thathave long been identified as being involved in thepathogenesis of cardiovascular disease as well as manyother degenerative diseases such as cancer, and immunedysfunction (Ames et al. 1993). Although reactive oxy-gen species (ROS) are essential participants in thephysiology of aerobic organisms, excessive and uncon-trolled ROS production leads to extensive oxidationof cell components and irreversible damage underly-ing the acute pathology of many diseases. Therefore,antioxidants, molecules that inhibit the oxidation ofother molecules, have been investigated for the pre-vention of oxidative stress-related diseases.Polyphenols are well known to have antioxidant prop-erties, which is often claimed to be responsible for their

disease-protective effects (Scalbert et al. 2005). Alter-natively, a wealth of data suggests that most of the rel-evant mechanisms of disease prevention bypolyphenolic flavonoids are not related to their anti-oxidant properties, but are rather due to the pro-oxi-dant action and direct interaction of flavonoids withtarget molecules. Figure 1 summarizes the major re-actions of a catechol type-polyphenol; antioxidant, pro-oxidant, and electrophilic conjugation reactions(Nakamura and Miyoshi 2010). Polyphenols induce anantioxidant reaction by being oxidized themselves, inwhich case an electron and hydrogen are transferredto free radicals. Paradoxically, polyphenols can act asconditional pro-oxidants that produce ROS via auto-catalytic oxidation or metal-catalyzed reaction to semi-quinone radicals, even when no free radicals existaround polyphenols. Finally, two-electron oxidation ofa catechol type polyphenol leads to the formation of aquinone, an electrophile agent.

An electrophile (Fig. 2) is a chemical that partici-pates in a chemical reaction by accepting an electronpair of a nucleophile. Interest in reactive electrophiles,which are naturally occurring or endogenously gener-ated, stems from the fact that they show a series ofcharacteristics and wide ranging biological responses(Nakamura and Miyoshi 2010). Unlike many classicalsignals and hormones, exposure of animals or their cellsto electrophilic chemicals potentially induces power-ful pathophysiological effects by covalently attachingto nucleic acids, proteins, and small molecules, as wellas by indirectly lowering pools of cellular reductants.For example, the discovery of carcinogen-DNA adductshas established a DNA damage-mutation-carcinogen-esis paradigm. The electrophile-protein covalent bind-ing contributes to the toxicity of drugs and other chemi-cals possibly through irreversible regulatory modifi-cations to proteins, including phosphorylation, acyla-tion, methylation, ubiquitination, and sumoylation. Inaddition to the toxic potential , food-derivedelectrophiles have been explored as promising protec-tors against neoplasia, toxicity, and many chronicpathological conditions. Talalay and Benson reported

Fig. 1. The major reactions of a catechol type-polyphenol;antioxidant, pro-oxidant, and electrophilic conjugation re-actions. Reprinted from Bioscience, Biotechnology and Bio-science, 74(2), Nakamura and Miyoshi, Electrophiles infoods: the current status of isothiocyanates and their chemi-cal biology, 242–255, Copyright (2010), with permissionfrom Japan Society for Bioscience, Biotechnology andAgrochemistry.

Fig. 2. Conceptual i l lustration of electrophile andnucleophile. Reprinted from Bioscience, Biotechnology andBioscience, 74(2), Nakamura and Miyoshi, Electrophiles infoods: the current status of isothiocyanates and their chemi-cal biology, 242–255, Copyright (2010), with permissionfrom Japan Society for Bioscience, Biotechnology andAgrochemistry.

Y. Nakamura / AGri-Biosci. Monogr. 6: 1–57, 2016 3


that administration of phenolic antioxidants to micereduced the formation of mutagenic metabolites froma carcinogen and raised the activity levels of classicaldetoxification enzymes in many tissues (1982). Theseobservations led to the suggestion that the protectivefunction of phenolic antioxidants, which can bemetabolically converted into electrophiles, might beascribed to the induction of the cytoprotective re-sponses, that is electrophile counterattack responses(Prestera et al. 1993).

Based on compelling evidence regarding the benefi-cial potential of polyphenols, various polyphenol-richdietary supplements are being developed for public use.Polyphenols are consumed daily, with an estimatedtotal consumption of 1 g/day in western countries,which could be much higher if dietary supplements arealso consumed. Bioavailability can be modified to fur-ther increase absorption and ultimately plasma concen-trations of polyphenols (Martin and Appel 2010). How-ever, the upper limit for plasma concentrations is stillunknown for many polyphenols. It is recently becom-ing apparent that commonly used dietary compoundscan exert deleterious effects at pharmacological(supraphysiological) concentrations. Moreover, somehuman intervention trials showed not only failure toprotect by polyphenols but accelerated developmentof cancers or cardiovascular disease in susceptible sub-jects (Halliwell 2007). Thus, the present monographtries to provide experimental evidence supporting theidea that optimal polyphenol supplementation can con-fer beneficial effects but high-doses elicit adverse ef-fects in an electrophilic reaction-dependent manner,thereby establishing a double-edged sword in healthpromotion.

The outline of this monograph is as follows: Sec-tions 2 to 5 present the author’s research papers inharmful aspects of high doses of polyphenols in mouseskin and other organs such as liver and kidney.

Section 2 (Nakamura et al. 2000d) demonstrates thedose- and timing-dependent dual effects of topical ap-plication of the phenolic antioxidant protocatechuicacid (PCA) on 12-O-tetradecanoylphorbol-13-acetate(TPA)-induced mouse skin tumor promotion. Pretreat-ment with PCA at a low dose significantly inhibitedthe number of papillomas, whereas PCA pretreatmentat a high dose significantly enhanced tumor numbers.This opposing effect was also pretreatment duration-dependent. These results suggested that topically ap-plied PCA can be converted to compound(s) lackingantioxidative properties and/or rather possessing thepotential to enhance tumor development. A similar ten-dency was also demonstrated in the short-term experi-ment of TPA-induced inflammation and oxidativestress. For example, mice pretreated with PCA at a quitehigh dose for 3 h did not show suppression or evensignificantly enhanced TPA-induced leukocyte infil-tration, hydrogen peroxide (H2O2) generation, and

thiobarbituric acid-reacting substances level, whilePCA treatment together with TPA significantly sup-pressed these parameters. Oxidative stress responsesinduced by a high dose of PCA such as consumptionof glutathione reduced form (GSH) and H2O2 genera-tion were counteracted by the tyrosinase inhibitor ar-butin (ArB). This study suggests that tyrosinase-de-pendent metabolism is involved in the tumor-enhanc-ing effect of PCA on mouse skin.

Section 3 (Nakamura et al. 2001a) provides furtherevidence showing that a catechol antioxidant PCAmodifies inflammatory responses in mouse skinthrough the tyrosinase-dependent bioactivation. Thisstudy showed that the responsibility of PCA-inducedleukocyte infiltration in a B6C3F1 mouse, having muchhigher tyrosinase activity, is more severe than that ofan ICR albino mouse. Furthermore, the possibility thatPCA alone can induce tyrosinase-dependent contacthypersensitivity in ICR mouse skin was suggested.Based on the present data, it was concluded that thetyrosinase-derived electrophilic quinoneintermediate(s) of PCA, which binds nucleophilicresidues of proteins including sulfhydryl group andconjugates of which are recognized as antigen, mightbe involved in alteration of the cellular immune func-tions including oxygen radical-generating leukocytesmigration to inflamed regions.

Section 4 (Nakamura et al. 2014) tries to encompassthe scope of toxic phenomena from PCA to otherpolyphenolic compounds. This study evaluated the ty-rosinase-dependent modifying effect of several sim-ple phenolic compounds on GSH and discussed thestructure-activity relationship. The in vitro chemicalmodification in GSH and a model protein by thebioactivated PCA was characterized. This study alsoconfirmed the enhancing effects of the phenolic com-pounds as well as PCA on TPA-induced inflammationin mice, suggesting that higher doses of the polyphenolscommonly have the potential to enhance skin inflam-mation.

Section 5 (Nakamura et al. 2001b) provides the evi-dence showing that overdoses of PCA can disturb thedetoxification system in organs other than the skin suchas the liver and kidney. This study showed that the i.p.administration of PCA causes significant hepatic andnephrotic GSH depletion and thus slight hepatotoxic-ity and nephrotoxicity. The protective role of GSH foracute hepatotoxicity was demonstrated using GSH-depleted mice administered with a GSH synthesis in-hibitor buthionine sulfoximine. The present resultssuggested that the possible toxic effects of high dosesof phenolic antioxidants in the liver and kidney aredependent on their metabolic conversion intoelectrophilic reactive species.

Sections 6 and 7 differ from the previous four sec-tions because they present the role of polyphenolelectrophilicity in the beneficial responses including



induction of the cytoprotective gene expression. Sec-tion 6 (Nakamura et al. 2003a) tries to clarify the mo-lecular mechanism of glutathione S-transferase (GST)induction by tert-butylhydroquinone (tBHQ) as com-pared with its analogue, 2,5-di-tert-butylhydroquinone(DtBHQ). This study indicated that tBHQ and DtBHQshows similar one-electron oxidization potentials,whereas only tBHQ acts as an electrophile but DtBHQdoes not, which is possibly due to the difference insteric hindrance. It should be noteworthy that tBHQmore potently induces the GSTP1 gene expression inRL34 cells than DtBHQ does. The results led us to ahypothesis that an electrophilic quinone product thatcan react with intracellular nucleophiles is the ultimateinducer of the GSTP1 gene expression. Section 7 (Ishiiet al. 2009) demonstrates the catechol-modification ofprotein sulfhydryls by 3,4-dihydroxyphenyl acetic acid(DPA), a model of the catechol-type polyphenol. A newprobe was also developed to directly detect proteinmodification by catechol-type polyphenols using abiotinylated DPA (Bio-DPA). The oxidation-depend-ent electrophilic reactivity of DPA with peptidesulfhydryls was confirmed by both mass spectrometryand nuclear magnetic resonance spectroscopy. Thepresent results indicated that b-actin is one of the ma-jor targets of protein modification by catechol-typepolyphenols, and that Bio-DPA is a useful probe forunderstanding the redox regulation by dietarypolyphenols. This study provides an alternative ap-proach to understand that catechol-type polyphenol isa potential modifier of redox-dependent cellular eventsthrough sulfhydryl modification.

Finally, the conclusions and perspective derived fromthe sections are summarized in Section 8. Table 1 sum-marizes the doses and treatment periods of PCA thatare required for the beneficial or harmful effects de-termined in the present studies.

2. A simple phenolic antioxidant enhances tumorpromotion in mouse skin: dose- and timing-dependent enhancement and involvement ofbioactivation by tyrosinase

2-1. Introduction

Various kinds of antioxidants have been investigatedfor their potential usefulness as cancerchemopreventive agents. Antioxidants have beenthought to mainly suppress carcinogenesis during theinitiation phase, since most act as radical scavengers,or inducers or inhibitors of xenobiotics metabolizingenzymes including phase I and II enzymes. On the otherhand, several antioxidants that can inhibit the initia-tion events have been found to be second stage tumorpromoters (Ito and Hirose 1989). Ogawa et al. (1998)clearly demonstrated stage- and organ-dependent pro-motional effects of antioxidants in a rat multiorgancarcinogenesis model (1998). Thus, chemopreventiveagents that act in the initiation stage do not necessar-ily exert beneficial effects in the post-initiation phase.Moreover, radical scavengers are known to have pro-oxidative potential because of their conversion to morereactive or stable radicals after they react directly withROS, which may contribute to the induction of sec-ondary oxidative damage to the target organs.

The simple phenolic PCA (Fig. 3) is one of the ma-jor benzoic acid derivatives from edible plants andfruits and shows a strong antioxidative effect, 10-foldhigher than that of a-tocopherol (Ueda et al. 1996).PCA even at a 100 ppm shows potent chemopreventiveeffects on colon and oral carcinogenesis in rats (Tanakaet al. 1993, 1994). The present study was initially per-formed to estimate the effectiveness of PCA againstTPA-induced tumor promotion in mouse skin. Inter-estingly, differences in modifying effects were depend-

Table 1. Summary of the doses and treatment periods of PCA that are required for the beneficial or harmful effect in thismonograph.

aPretreatment period before stimulus or treatment period.bMyeloperoxidase.cThiobarbituric acid-reacting substances.dInterferon-g.eAspartate aminotransaminase.

Assay model or biomaker Beneficial effect (dose, perioda) Harmful effect (dose, perioda) Chapter

Skin carcinogenesis by DMBA/TPA (papillomas) Decreased (16 nmol, 0.5 h) Increased (1,600 nmol, 0.5 h) 2Acute skin inflammation by TPA (edema/MPOb) Decreased (81 nmol, 0.5 h) No effect (>8,100 nmol, 0.5 h) 2Acute skin inflammation by TPA (edema/MPOb) Decreased (20,000 nmol, 0 h) Increased (20,000 nmol, 3 h) 2, 3Glutathione peroxidase activity in the skin æ Decreased (20,000 nmol, 3 h) 2

TPA-induced oxidative stress (TBARSc) Decreased (20,000 nmol, 0 h) Increased (20,000 nmol, 3 h) 2, 3Subchronic skin inflammation by TPA (edema/MPOb) Decreased (16 nmol, 0.5 h) Increased (1,600 nmol, 0.5 h) 3Contact hypersensitivity (edema/IFN-gd) æ Increased (30,000 nmol, 48 h) 3

Acute hepato- and nephro-toxicity (ASTe/plasmatic urea) æ Increased (500 mg/kg, 6 h) 5

Glutathione level in the liver and kidney æ Decreased (500 mg/kg, 6 h) 6



ent on the dose (8.1~20,000 nmol) and timing (5 min~ 3 h before TPA treatment) of PCA. We demonstratednot only the lack of an inhibitory effect but also sig-nificant enhancement of mouse skin tumor promotionby pretreatment with a high dose of PCA 3 h beforeTPA application. The possibility of metabolism by ty-rosinase activity of PCA to certain compound(s) with-out antioxidative properties and/or with tumor promo-tional potency was also suggested.

2-2. Materials and methods

2-2A. Chemicals and animalsPCA was purchased from Nacalai Tesque, Inc.,

Kyoto, Japan. TPA was obtained from ResearchBiochemicals International, Natick, MA, USA. RPMI1640 medium was purchased from Gibco RBL,Rockville, MD, USA. All other chemicals were pur-chased from Wako Pure Chemical Industries, Osaka,Japan. Female ICR mice (7 weeks old) were obtainedfrom Japan SLC, Shizuoka, Japan. Mice used in eachexperiment were supplied with fresh tap water ad libi-tum and rodent pellets (MF, Oriental Yeast Co., Kyoto,Japan) freshly changed twice a week. Animals weretreated in accordance with the Guidelines for AnimalExperimentation of Kyoto University. Animals weremaintained in a room controlled at 24 ± 2∞C with arelative humidity of 60 ± 5% and a 12-h light/dark cycle(06:00 to 18:00). The back of each mouse was shavedwith surgical clippers two days before each experiment.

2-2B. Tumor promotion experimentThe modifying effect of PCA on TPA-induced tumor

promotion was examined by a standard initiation-pro-motion protocol with dimethylbenz[a]anthracene(DMBA) and TPA as previously reported (Nakamuraet al. 1999). One group was composed of 15 femaleICR mice housed 5 per cage. The back of each mousewas shaved with surgical clippers two days before ini-tiation. The mice at 7 weeks old were initiated with

DMBA (190 or 95 nmol/0.1 ml acetone). One weekafter initiation, the mice were promoted with TPA (1.6nmol/0.1 ml acetone) twice a week for 20 weeks. Infour other groups, the mice were treated with PCA (16,160, 1600 or 20,000 nmol/0.1 ml acetone) 0 min, 40min, or 3 h before each TPA treatment. The modifyingeffect on TPA-induced tumor promotion was evaluatedby both the ratio of tumor-bearing mice and the numberof tumors, more than 1 mm in diameter, per mouse.The data were statistically analyzed using the Student’st-test (two sided), that assumed unequal variance, forthe average number of tumors per mouse and by thec2-test for the incidence of skin tumors.

2-2C. Inflammatory biomarker determinationThe modifying effect on single TPA application-in-

duced inflammation was determined by twobiomarkers, edema formation and myeloperoxidase(MPO) activity, as previously reported (Nakamura etal. 1998a). Mice were sacrificed by cervical disloca-tion 18 h after a single application of TPA. Mouse skinpunches were obtained with an 8-mm-diameter corkborer and weighed in an analytical balance. The IE wereexpressed by the relative increasing ratio of the weightof a treated punch to that of a control punch; inhibi-tory effect (IE) (%) = [(TPA alone) – (test compoundplus TPA)]/[(TPA) – (vehicle)] ¥ 100. The MPO activ-ity was calculated from the linear portion of the curveand expressed as units of MPO per skin punch. Oneunit of MPO activity was defined as the activity thatdegraded 1 mmol of H2O2 per min at 25∞C.

2-2D. Double TPA treatment protocolOne group was composed of 5 female ICR mice

housed at 5 per cage. PCA was topically applied to theshaved area of dorsal skin at various times before ap-plication of TPA solution (8.1 nmol/0.1 ml in acetone).This TPA dose (8.1 nmol) was used for the potentiationof oxidative responses compared with the dose fortumor promotion (1.6 nmol). The same doses of PCAand TPA or acetone were applied twice at an intervalof 24 h for H2O2 determination. Although the timing(24 h apart) of double TPA application was differentfrom tumor promotion protocol, the level of oxidativestress was nearly the same when the time between thetwo TPA treatments was 24–72 h (data not shown).

2-2E. Determination of oxidative stress parametersMice treated by the double-treatment protocol were

sacrificed 1 h after the last TPA treatment. The H2O2content was determined by the phenol red-horseradishperoxidase (HRPO) method (Nakamura et al. 1996,1998a, 1998c). The final results are expressed asequivalents of nmol of H2O2 per skin punch, on thebasis of a standard curve of HRPO-mediated oxida-tion of phenol red by H2O2. Thiobarbituric acid-react-

Fig. 3. Chemical structure of PCA. Reprinted from Carcino-genesis, 21(10), Nakamura et al., A simple phenolic anti-oxidant protocatechuic acid enhances tumor promotion andoxidative stress in female ICR mouse skin: dose-and tim-ing-dependent enhancement and involvement ofbioactivation by tyrosinase, 1899–1907, Copyright (2000),with permission from Oxford University Press.



ing substances (TBARS) level of mouse epidermis wasdetermined by our previously reported method (10, 12).The final results are expressed as equivalents of nmolsof malondialdehyde per cm2, on the basis of a stand-ard line of TBARS formation using the authenticmalondialdehyde.

2-2F. Histological examinationMice treated with the double-treatment protocol were

sacrificed 1 h after the second TPA treatment. Excisedskin was fixed in 10% buffered formalin, and thenembedded in paraffin. Skin samples were cut at 3 mm,mounted on silanized slides, dewaxed in xylene, de-hydrated through an ethanol series, and stained withhematoxylin and eosin. For each section of skin, thethickness of the epidermis from the basal layer to thestratum corneum was measured at five equidistantinterfollicular sites utilizing an image analysis systemLeica Q500IW-EX (Leica Co., Ltd., Tokyo, Japan) witha microscope Leica DMRE HC (Leica Co., Ltd.). Thenumbers of infiltrating leukocytes were counted at fivedifferent areas of each section using this image analy-sis system. Proliferating cell nuclear antigen (PCNA)immunohistochemistry was performed as previouslyreported (Nakamura et al. 1998a). The PCNA labelingindex was counted at six different areas of each sec-tion using the image analysis system, and expressedas the number of positive squamous cell nuclei dividedby the total number of squamous cell nuclei ¥100.

2-2G. GSH assayMeasurement of GSH level was performed spectro-

photometrically using a commercial kit(BIOXYTECH® GSH-400™ Assay; OXIS Interna-

tional, Inc., Portland, OR, USA). Mice were treatedwith PCA at different time intervals and sacrificed bycervical dislocation. The skin samples were homog-enized and extracted with 5% metaphosphoric acidsolution containing 5 mM EDTA. Then centrifugation(10,000 g, 20 min), 50 ml of 12 mM chromogenic rea-gent in 0.2 M HCl was added to the resultingsupernatant (300 ml) and mixed thoroughly. After 50ml of 7.5 M NaOH was added and mixed, the mixturewas incubated at 25∞C for 10 min, and then absorb-ance was determined spectrophotometrically at 400 nm.

2-2H. GPx and GST AssaysMeasurement of glutathione peroxidase (GPx) activ-

ity was performed spectrophotometrically using a com-mercial kit (BIOXYTECH® GPx-340™ Assay; OXIS,International, Inc.). The skin samples were minced in3 ml of 50 mM Tris-HCl buffer (pH 8.0) containing 5mM sodium azide and 1 mM 2-mercaptethanol and thenhomogenized at 4∞C for 30 s twice. After centrifuga-tion (10,000 g, 20 min), 700 ml of NADPH reagentcontaining GSH and glutathione reductase was addedto the resulting supernatant (70 ml) and mixed thor-oughly. After 350 ml of 0.007% (w/w) tert-butylhydroperoxide solution was added and mixed, increasesin absorption at 340 nm for 3 min at 0.2-min intervalswere recorded. The GPx activity was calculated fromthe linear portion of the curve and expressed as unitsof GPx per skin punch. One unit of GPx activity wasdefined as the activity that degraded 1 mmol of NADPHper min at 25∞C.

Total GST activity was measured using 1-chloro-2,4-dinitrobenzene as a substrate according to the methodof Habig et al. (1974).

Treatment Tumors/mouse (% tumor-bearing mouse)

6 wk 10 wk 15 wk 20 wk

Acetone 0 (0) 0 (0) 0 (0) 0 (0)PCA (1600 nmol) 0 (0) 0 (0) 0 (0) 0 (0)TPA (1.6 nmol) 0.90 ± 1.66 (30) 6.80 ± 6.53 (60) 12.60 ± 7.28 (80) 14.00 ± 6.73 (100)

PCA (16 nmol)/TPA 0.58 ± 1.24 (25) 2.67 ± 4.27 (50) 6.33 ± 6.41a (58) 6.75 ± 7.37a (58b)

PCA (160 nmol)/TPA 1.92 ± 3.00 (50) 8.25 ± 7.86 (83) 12.67 ± 7.45 (83) 12.83 ± 7.26 (92)

PCA (1600 nmol)/TPA 4.17 ± 4.71a (58b) 10.83 ± 7.11 (92b) 17.58 ± 4.72 (100) 19.33 ± 4.03a (100)

Table 2. Modifying effects of pretreatment (40 min) with different doses of PCA on TPA-induced papilloma formation in ICRmouse skin. ICR mice were treated topically with 190 nmol of DMBA. One week later, the mice were treated with acetone,1.6 nmol of TPA alone or pretreated with PA twice a week for 20 weeks. The number of tumors per mouse are expressed asmeans ± SD (n = 10 or 12). Reprinted from Carcinogenesis, 21(10), Nakamura et al., A simple phenolic antioxidant protocate-chuic acid enhances tumor promotion and oxidative stress in female ICR mouse skin: dose-and timing-dependent enhance-ment and involvement of bioactivation by tyrosinase, 1899–1907, Copyright (2000), with permission from Oxford UniversityPress.

aStatistically different from TPA alone (Student’s t-test); P < 0.05.bStatistically different from TPA alone (c2-test); P < 0.05.



2-3. Results

2-3A. Modifying effect of the dose or timing of PCApretreatment on TPA-induced tumor promo-tion

The modifying effects of different doses of topicallyapplied PCA on TPA-induced tumor promotion wereexamined in a two-stage mouse skin carcinogenesismodel and the results are shown in Table 2. Tumorsbegan to develop 6 weeks after tumor promotion byTPA. The incidence of tumor-bearing mice and theaverage number of tumors per mouse in the group givenDMBA and TPA were 100% and 14.0, respectively, atthe end (20 weeks) of the experiment. In the grouptreated with 16 nmol of PCA 40 min prior to each TPAtreatment, the average number of tumors per mousewas reduced by 52% (P < 0.05) and tumor incidenceby 42% (P < 0.05). On the other hand, pretreatmentwith 160 nmol PCA showed slight inhibitory effectson tumor numbers (IE = 8%) and the tumor incidence(IE = 8%). Moreover, PCA 1600 nmol significantlyenhanced the number of tumors per mouse by 38% (P< 0.05). Mice initiated with DMBA and then treatedwith acetone or 1,600 nmol PCA twice weekly for 20weeks did not develop any tumors.

Subsequently, we examined whether the time inter-val between PCA and TPA treatments attenuates itsantioxidative properties and thus decreases antitumorpromotional activity (Fig. 4). In this experiment, miceinitiated with 95 nmol DMBA and promoted with 1.6nmol of TPA twice weekly for 20 weeks developed anaverage of 8.7 ± 6.1 tumors/mouse. In the group treatedwith 20,000 nmol PCA 5 min prior to 1.6 nmol TPAapplication, the average number of tumors per mousewas reduced by 38%, but this was not statistically sig-nificant. On the other hand, pretreatment with 20,000nmol PCA 3 h before TPA treatment significantly in-creased the number of tumors per mouse (16.0 ± 7.4)by 84% (P < 0.01).

2-3B. Modifying effect of PCA on inflammatory re-sponses and oxidative stress in ICR mouse skin

To assess the possibility that PCA could modify TPA-induced inflammation and oxidative stress, we inves-tigated the effects of different doses and different tim-ing of topical application of PCA on single TPA appli-cation-induced edema formation and MPO activityenhancement as well as double TPA application-in-duced H2O2 generation in ICR mouse skin. As shownin Table 3, the application of a low dose (81 nmol) ofPCA 30 min before 8.1 nmol TPA treatment inhibitededema formation, MPO activity, and H2O2 generationby 20%, 58% (P < 0.05) and 22%, respectively.Whereas, the application of 8,100 or 20,000 nmol PCA30 min before TPA treatment showed slight inhibitionof expression of these three biomarkers (IEs = –4~8%).

To determine whether the time interval between PCAand TPA treatments modifies its antioxidative proper-ties, the effects of pretreatment with 20,000 nmol PCA0, 0.5, 1, 3 or 6 h before TPA treatment were evaluated(Table 3). The simultaneous application of 20,000 nmolPCA together with TPA treatment inhibited expressionof all biomarkers by 27%, 47% (P < 0.05) and 51% (P< 0.001), respectively. However, as the time intervalsbetween PCA and TPA treatment were increased, fewerinhibitory effects on inflammation and oxidative stresselevation were observed. An application 3 h prior toTPA treatment showed the strongest enhancement ofMPO activity and H2O2 generation (166% of control,P < 0.05 and 125% of control P < 0.05, respectively).

As topically applied PCA significantly modified dou-ble TPA application-induced H2O2 generation, we as-sessed whether such PCA treatment influences doubleTPA application-induced TBARS formation, a well-known biomarker of overall oxidative damage to cel-lular constituents such as membrane lipids. The quan-titative data for the levels of TBARS formation inmouse epidermis homogenate are shown in Fig. 5. Theincreased level of TBARS caused by the double TPAapplication was significantly higher than that of thecontrol (0.43 ± 0.09 versus 0.15 ± 0.06 nmol/cm2, P <0.01). The simultaneous application of PCA (20,000

Fig. 4. Modifying effect on TPA-induced tumor promotionof PCA in DMBA-initiated ICR mouse skin. One group wascomposed of 15 female ICR mice. The mice at 7 weeks oldwere initiated with DMBA (95 nmol). One week after initia-tion, the mice in group 1 (closed circles) were promoted withTPA (1.6 nmol/0.1 ml in acetone) twice a week for 20 weeks.In the PCA-treated experiments, the mice in group 2 (opentriangles) and 3 (open squares) were treated with PCA(20,000 nmol/0.1 ml in acetone, respectively) 5 min and 3h, respectively, prior to each TPA treatment. The tumor pro-motion-modifying activity was evaluated by both the ratioof tumor-bearing mice (left) and the number of tumors permouse (right). Statistical analysis was performed by the c2-test on tumor-bearing mouse and the Student’s t-test on thenumber of tumors per mouse. Significance is expressed as:a, P < 0.005; b, P < 0.05. Reprinted from Carcinogenesis,21(10), Nakamura et al., A simple phenolic antioxidant pro-tocatechuic acid enhances tumor promotion and oxidativestress in female ICR mouse skin: dose-and timing-depend-ent enhancement and involvement of bioactivation by tyro-sinase, 1899–1907, Copyright (2000), with permission fromOxford University Press.



nmol) with TPA treatment significantly inhibited theincrease in TBARS level (0.24 ± 0.05 nmol/cm2, P <0.01 versus double TPA). On the other hand, the appli-cation of PCA 3 h before TPA treatment significantlyenhanced double TPA application-induced TBARS for-mation (0.63 ± 0.13 nmol/cm2, P < 0.05 versus doubleTPA).

2-3C. Effects of PCA on morphological alterationsin mouse skin treated with double TPA appli-cation

The results that PCA treatment caused the inverseeffects on MPO activity, H2O2 generation and TBARSformation led us to select double TPA application modelof mouse inflammation for histological observation todetermine whether PCA application enhances leukocyteinfiltration and hyperplasia in the cutis. Double appli-cation of TPA caused morphological alteration of in-flammatory response (Table 4, Fig. 6B) as comparedwith the control group (Fig. 6A), which was well cor-related to the results of skin edema formation and MPOactivity (Table 3). Mouse skin treated with TPA twicewith a 24-h interval displayed severe epidermal hy-

perplasia (Fig. 6B) and resulted in a marked increasein leukocyte infiltration as compared with that treatedwith acetone as shown in Table 3 (8 ± 1 versus 312 ±41/mm2). Also, double TPA treatment increased thenumber of mitosis in epidermal squamous cells (Fig.6B). On the other hand, pretreatment with PCA (20,000nmol) together with TPA application diminished TPA-induced hyperplasia, mitosis (Fig. 6C) and leukocyteinfiltration (Table 4). PCNA immunohistochemistry re-vealed PCNA-positive nuclei in many epidermalkeratinocytes, only at the basal and first suprabasallayer in mice treated with acetone alone (PCNA-labeling index: 36.4 ± 6.7, Table 4). Double applica-tion of TPA significantly enhanced the PCNA index(73.4 ± 16.1, P < 0.05) as compared with acetone alone.The PCNA index (36.1 ± 3.2) in mice given doubleapplication of TPA and PCA at the same time was al-most equivalent to that of control mice, and the valuewas significantly lower than that of mice treated twicewith TPA (P < 0.02). While a high proportion of epi-thelial cell nuclei in hair follicles and glandular ap-pendages were also stained, no significant differencesin this pattern were observed between different treat-

Group Pretreatment (timea) Stimuli Oxidative stress parameter

Edema formationb MPO activityb H2O2 productionb

(mg/punch, %c) (unit/punch, %c) (nmol/punch, %c)

1 acetone (0.5) acetone 32.9 ± 3.7 æ 0.10 ± 0.03 æ 0.31 ± 0.14 æ2 PCA 8100 nmol (0.5) acetone 31.6 ± 4.5 æ 0.47 ± 0.28 æ 0.67 ± 0.35 æ3 acetone (0.5) TPA 8.1 nmol 47.0 ± 7.3d 100 1.69 ± 0.30e 100 2.60 ± 0.36e 100

4 PCA 81 nmol (0.5) TPA 8.1 nmol 37.7 ± 11.4 80 0.70 ± 0.22e,h 42 2.03 ± 0.44e 78

5 PCA 8,100 nmol (0.5) TPA 8.1 nmol 48.1 ± 10.6d 102 1.62 ± 0.18e 96 2.70 ± 0.84e 104

6 PCA 20,000 nmol (0) TPA 8.1 nmol 34.3 ± 8.5 73 0.90 ± 0.24e,g 53 1.28 ± 0.13e,h 49

7 PCA 20,000 nmol (0.5) TPA 8.1 nmol 45.9 ± 11.0d 98 1.59 ± 0.58e 92 2.70 ± 0.59e 104

8 PCA 20,000 nmol (1) TPA 8.1 nmol 50.8 ± 11.4d 108 2.35 ± 0.88e 139 2.81 ± 0.48e 108

9 PCA 20,000 nmol (3) TPA 8.1 nmol 53.8 ± 8.0d,f 114 4.50 ± 1.24e,f 266 3.24 ± 0.07e,g 125

10 PCA 20,000 nmol (6) TPA 8.1 nmol 52.7 ± 8.2d 111 3.58 ± 1.14e,f 211 2.89 ± 0.45e 111

Table 3. Modifying effects of pretreatment at different times or different doses of PCA on TPA-induced enhancement ofoxidative stress parameters in ICR mouse skin. ICR mice (5 mice in each group) were treated as described in “Materials andmethods”. The mice for edema and MPO activity determination were sacrificed 18 h after TPA treatment (8.1 nmol), and skinpunches were obtained. The mice for H

2O

2 production were sacrificed 1 h after the second TPA treatment (8.1 nmol), and skin

punches were obtained. Reprinted from Carcinogenesis, 21(10), Nakamura et al., A simple phenolic antioxidant protocate-chuic acid enhances tumor promotion and oxidative stress in female ICR mouse skin: dose-and timing-dependent enhance-ment and involvement of bioactivation by tyrosinase, 1899–1907, Copyright (2000), with permission from Oxford UniversityPress.

aThe time interval between PCA pretreatment and TPA (8.1 nmol) stimulation.bValues are means ± SD of 5 mice. Significance was determined by the Student t-test.cValues are percentages of those in group 3 (TPA stimulation alone).dSignificant versus group 1 (Student’s t-test); P < 0.05.eSignificant versus group 1 (Student’s t-test); P < 0.001.fSignificant versus group 3 (Student’s t-test); P < 0.05.gSignificant versus group 3 (Student’s t-test); P < 0.01.hSignificant versus group 3 (Student’s t-test); P < 0.001.



Group Pretreatment (time) Stimuli No. of leukocytes in the cutisa PCNA-stained cell nuclei indexb

1 acetone (0.5) acetone 8 ± 1 36.4 ± 6.72 acetone (0.5) TPA 8.1 nmol 312 ± 41c 73.4 ± 16.1c

3 PCA 20,000 nmol (0) TPA 8.1 nmol 12 ± 5d 36.1 ± 3.2c,d

4 PCA 20,000 nmol (0.5) TPA 8.1 nmol 115 ± 19c,d 46.7 ± 12.8c,d

5 PCA 20,000 nmol (3) TPA 8.1 nmol 226 ± 66c 65.6 ± 6.0c

Table 4. Modifying effects of PCA pretreatment at different times on TPA-induced morphological changes in mouse skin. Themice were sacrificed 1 h after the second TPA treatment (8.1 nmol), and skin punches were obtained for determination of themorphological changes. Reprinted from Carcinogenesis, 21(10), Nakamura et al., A simple phenolic antioxidant protocate-chuic acid enhances tumor promotion and oxidative stress in female ICR mouse skin: dose-and timing-dependent enhance-ment and involvement of bioactivation by tyrosinase, 1899–1907, Copyright (2000), with permission from Oxford UniversityPress.

aMean ± SD of 15 values taken from three individual mice.bMean ± SD of 15 values, counted in various regions of the cutis (dermis and subcutis) in an area of 1 mm2 under themicroscope.cSignificant versus group 1 (Student’s t-test); P < 0.05.dSignificant versus group 2 (Student’s t-test); P < 0.05.

Fig. 5. Modifying effects of PCA on TBARS formation inthe mouse epidermis. ICR mice (5 mice in each group) weretreated by the double treatment protocol as described in“Materials and methods”. Mouse skin was treated with PCA(20,000 nmol) or acetone 0 or 3 h prior to each TPA treat-ment. The mice were sacrificed 1 h after the second TPAapplication, and their epidermis was removed for TBARSassays. Significance determined by the Student’s t-test isexpressed as: a, versus acetone, P < 0.05; b, versus TPA, P <0.01; c, versus TPA, P < 0.05. Reprinted from Carcinogen-esis, 21(10), Nakamura et al., A simple phenolic antioxi-dant protocatechuic acid enhances tumor promotion andoxidative stress in female ICR mouse skin: dose-and tim-ing-dependent enhancement and involvement ofbioactivation by tyrosinase, 1899–1907, Copyright (2000),with permission from Oxford University Press.

ment groups (data not shown). On the other hand, asthe interval between PCA and TPA treatment was in-creased, inhibitory effects on TPA-induced increasesin numbers of leukocytes and PCNA index were notsignificant as compared with group 2 treated only withTPA (Table 4).

2-3D. PCA causes tyrosinase-dependent reductionof GSH level and antioxidative enzyme activi-ties

We assessed whether redox alteration is involved inthe inverse effects of PCA on inflammatory responsesin mouse skin. Since the treatment of mouse skin withTPA alone showed little influence on total GSH level(Reiners et al. 1991), the modifying effect of PCA onGSH level in mouse skin was examined. As shown inFig. 7, GSH level in mouse skin was reduced by treat-ment with 20,000 nmol PCA but not by 81 nmol PCA.This reduction was completely inhibited bycoadministration of the tyrosinase inhibitor ArB (10mmol), which showed no modulatory effect on GSHlevel by itself. The possibility that oxidative metabo-lism of PCA to benzoquinone form, which can readilyconjugate to nucleophiles such as GSH, prompted usto determine whether treatment with a high dose of PCAinhibits enzyme activity due to protein modificationsuch as conjugation with sulfhydryl residue at the ac-tive site. GPx and GST are abundant antioxidative en-zymes that can diminish cellular H2O2 level. As shownin Table 5, treatment with 20,000 nmol PCA signifi-cantly inhibited both GPx and GST activities (P < 0.05,respectively), and this effect was also counteracted byArB.

We also examined the inhibitory effects of ArB onPCA-enhanced inflammatory responses and oxidativestress. As shown in Fig. 8, the tyrosinase inhibitor ArB

counteracted PCA-induced enhancement of edema for-mation (127% of TPA group versus 107%), MPO ac-tivity (256% versus 145%, P < 0.05), and H2O2 level(136% versus 120%).



Fig. 6. Effects of PCA on TPA-induced morphological changes in mouse skin. The protocol for animal treatment was asdescribed in “Materials and methods”. Treatment with: A, acetone; B, double doses of TPA; C, double dose of TPA and PCA(20,000 nmol, 0 h prior to TPA); D, double dose of TPA and PCA (20,000 nmol, 3h prior to TPA). X20. Reprinted fromCarcinogenesis, 21(10), Nakamura et al., A simple phenolic antioxidant protocatechuic acid enhances tumor promotion andoxidative stress in female ICR mouse skin: dose-and timing-dependent enhancement and involvement of bioactivation bytyrosinase, 1899–1907, Copyright (2000), with permission from Oxford University Press.

2-4. Discussion

The results of this study demonstrated that topicalapplication of PCA exerts contrasting effects on TPA-induced tumor promotion in mouse skin in a mannerthat is both dose- and timing-dependent. A low dose ofPCA (16 nmol; 10-fold dose of TPA) 40 min prior toTPA application significantly decreased both the inci-dence and the multiplicity of tumors (Table 2). Thedose-dependent inhibitory effect of PCA was confirmedat doses ranging from 1/10-fold (0.16 nmol) to 10-fold(16 nmol) that of TPA (data not shown). This inhibi-tory effect of PCA was prominent when compared withknown food phytochemicals (Nishino et al. 1984;Huang et al. 1988; Murakami et al. 1995; Murakamiet al. 1996; Nakamura et al. 1999). On the other hand,the antitumor promotional effect of PCA was inverselycorrelated with the dose ranging from 10-fold to 1000-fold that of TPA. As the dose of PCA was increased,less antitumor promoting activity was observed withinthis range (Table 2). Tseng et al. reported that fairlyhigh doses (5, 10, and 20 mmol) of PCA 5 min prior to

TPA (15 nmol) treatment twice weekly to mice initi-ated with benzo[a]pyrene inhibited the incidence oftumors by 19~44% (1998). The effective dose of PCAin the present study was ~250-fold lower than thosereported previously. Interestingly, experiments usingtwo groups treated with 20 mmol (=20,000 nmol) ofPCA 5 min and 3 h prior to TPA treatment clearly dem-onstrated completely opposite effects on tumor devel-opment: the application of PCA 5 min before TPA treat-ment inhibited tumor formation as described previouslyby Tseng et al. (1998), while the application of PCA 3h before TPA treatment lacked a tumor inhibitory ef-fect and even enhanced tumor formation (Fig. 4). Thus,it is very likely that the excessive amount of PCA isconverted to compound(s) without antioxidative activ-ity or antitumor promotional properties.

There is increasing evidence indicating importantroles of oxidative stress in tumor promotion. Tumorpromoters such as TPA enhance the generation of ROSand decrease the ROS detoxification enzymes in bothepidermal and inflammatory cells. Kensler et al. (1987)proposed the requirement of two applications of TPA



for massive ROS generation, and the hypothesis thatthe first treatment of mouse skin with TPA applicationcauses a chemotactic action, i.e., recruitment ofneutrophils responsible for ROS generation with thesecond TPA treatment; i.e., priming and activation,respectively (Ji and Marnett 1992). ROS productionby double or multiple TPA treatments is closely asso-ciated with the metabolic activation of proximate car-cinogens (Kensler et al. 1987, 1989; Ji and Marnett

1992) and the increased levels of oxidized DNA bases(Wei and Frenkel 1991, 1992, 1993). We recently dem-onstrated that a potent inhibitor of leukocyte-derivedROS generation (Nakamura et al. 1998a) effectivelyprevented inflammation-related carcinogenesis(Ohnishi et al. 1996; Tanaka et al. 1997a, b). The re-sults of the present study clearly demonstrated thatapplication of PCA exerted application dose- and tim-ing-dependent inverse effects on TPA-induced inflam-

Fig. 7. Effects of PCA on total GSH level in ICR mouse skin. A, Time-dependent reduction of total GSH level induced by PCA(20,000 nmol). The mice were treated with PCA for different time intervals as indicated in the figure and sacrificed for GSHassay. Significance determined by the Student’s t-test is expressed as: *, versus acetone, P < 0.01. B, Counteracting effects ofthe tyrosinase inhibitor arbutin on PCA-induced reduction of GSH level. The mice (5 mice in each group) were treated witharbutin (10 mM) or acetone 1 h prior to PCA (81 or 20,000 nmol) treatment. The mice were sacrificed 3 h after PCA applica-tion, and skin samples were removed for GSH assays. Significance determined by the Student’s t-test is expressed as: a,versus acetone, P < 0.01; b, versus PCA 20,000 nmol, P < 0.01. Reprinted from Carcinogenesis, 21(10), Nakamura et al., Asimple phenolic antioxidant protocatechuic acid enhances tumor promotion and oxidative stress in female ICR mouse skin:dose-and timing-dependent enhancement and involvement of bioactivation by tyrosinase, 1899–1907, Copyright (2000), withpermission from Oxford University Press.

Group Treatment (time) Inhibitor GPxa GSTa

(mU/mg protein) (mU/mg protein)

1 acetone (3) acetone 50.8 ± 9.7 5.80 ± 0.212 acetone (3) arbutin 10 mmol 51.0 ± 7.2 5.66 ± 0.44b

3 PCA 20,000 nmol (3) acetone 35.6 ± 5.3b 5.08 ± 0.37b

4 PCA 20,000 nmol (3) arbutin 10 mmol 54.2 ± 10.5c 5.68 ± 0.32c

Table 5. Effects of PCA treatment on GPx and GST activities in mouse skin. Reprinted from Carcinogenesis, 21(10), Nakamuraet al., A simple phenolic antioxidant protocatechuic acid enhances tumor promotion and oxidative stress in female ICR mouseskin: dose-and timing-dependent enhancement and involvement of bioactivation by tyrosinase, 1899–1907, Copyright (2000),with permission from Oxford University Press.

aValues are means ± SD of 5 mice.bSignificant versus group 1 (Student’s t-test); P < 0.05.cSignificant versus group 3 (Student’s t-test); P < 0.05.



matory oxidative stress, which trends were very closelycorrelated with modifying effects on tumor develop-ment. Even a high dose of PCA treatment concurrentlywith TPA significantly inhibited edema formation,MPO activity and H2O2 generation, whereas PCA treat-ment 3 h prior to TPA significantly enhanced these in-flammatory parameters. The present result also indi-cated that although quite high doses of PCA showedstrong antioxidative effects in lipid peroxidation invitro (Ueda et al. 1996), PCA when applied 3 h priorto TPA treatment significantly enhanced TBARS for-mation (Fig. 5). TBARS is known as an overalloxidative damage biomarker formed downstream ofH2O2 generation in the presence of a metal ion as acatalyst. TBARS formation in vivo is considered notto reflect a single particular phenomenon but to indi-cate widespread oxidative damage including lipidperoxidation, mitochondrial deenergization and deg-radation of protein or sugar rather than DNA (Nakaeet al. 1994). Thus, the opposite effects of PCA onoxidative stress consequent on inflammatory responsesare likely to, at least in part, be determinants for TPA-induced tumor promotion. Histological studies alsodemonstrated that PCA treatment 3 h prior to TPA di-minished the inhibitory effects, which were observedwith simultaneous application of PCA, on TPA-inducedinflammatory responses such as leukocyte infiltration

in the cutis and PCNA-labeling index, also supportingour assumption that topically applied PCA was con-verted to the plausible toxic compound(s).

Compounds with a catechol moiety such as caffeicacid derivatives are known to be strong radical scav-engers in vitro and to show more potent antioxidativeactivity than monophenolic compounds (Tsuda et al.1996; Nakamura et al. 1998b). On the other hand, en-dogenous and exogenous catechols are oxidants as wellas electrophiles. The catechols are enzymatically orspontaneously oxidized to ortho-quinones that undergoredox cycling mediated by cytochrome P450/P450oxidoreductase or transition metals. In addition, ortho-quinones are Michael reaction acceptors that can read-ily react with nucleophiles such as thiol group. Thus,ortho-quinone or semiquinone formation fromcatechols has been suggested to explain their cytotoxicand/or genotoxic effects at high doses, since their mo-lecular mechanisms of action are attributed to ROSgeneration by redox cycling and covalent binding witha variety of cellular macromolecules (Monks et al.1992, 1997). The former events by high doses of PCAmay occur on the basis of the results that the pretreat-ment of a high dose PCA 3 h prior to TPA inducedTBARS formation in larger quantities than did treat-ment with TPA alone. It is well known that the reac-tion of catechol semiquinone with oxygen leads to gen-eration of superoxide anions with subsequent Fe(II)/Fe(III)-catalyzed production of hydroxyl radicals(Monks et al. 1992). Cu(II) also strongly mediates theoxidation of hydroquinone producing benzoquinoneand H2O2 through a Cu(I)/Cu(II) redox mechanism (Liet al. 1994). The in vitro abilities of PCA to induceFe(II)/Fe(III) redox-cycle-dependent lipid peroxidationand Cu(I)/Cu(II) redox-cycle-dependent oxidativeDNA damage have recently been observed (data notshown), suggesting that the effects of PCA are medi-ated by ROS generation. The oxidative phenomenonof nucleophilic addition to sulfur groups was supportedby the findings that a high dose of PCA significantlyreduced the total level of GSH, the most abundant thiolin cells, which can protect against the attack ofelectrophilic toxicants (Fig. 7). GSH is also well knownto be an endogenous antioxidant decomposing H2O2spontaneously or as a substrate for GPx or GST. Thepresent study clearly demonstrated that the in vitroenzyme activities of GPx and GST were reduced bytreatment with PCA, although excessive GSH wasadded in the enzyme activity determination systems.This inhibition may be due to nucleophilic attachmentof enzyme proteins, since the metabolism of PCA toquinone by tyrosinase was required as discussed be-low. Thus, the disturbance of ROS detoxification sys-tems such as not only GSH but also GPx and GST bythe possible electrophilic metabolite(s) of PCA maybe partially involved in enhanced oxidative stress in-

Fig. 8. Counteracting effects of arbutin on PCA-inducedenhancement of inflammatory biomarkers. ICR mice (5 micein each group) were treated as described in “Materials andmethods”. The mice (5 mice in each group) were treated witharbutin (10 mM) or acetone 1 h prior to PCA (81 or 20,000nmol) treatment. Mice were treated with PCA (20,000 nmol)3 h prior to each TPA treatment. Significance determined bythe Student’s t-test is expressed as: *, versus acetone + PCA+ TPA, P < 0.05. Reprinted from Carcinogenesis, 21(10),Nakamura et al., A simple phenolic antioxidant protocate-chuic acid enhances tumor promotion and oxidative stressin female ICR mouse skin: dose-and timing-dependent en-hancement and involvement of bioactivation by tyrosinase,1899-1907, Copyright (2000), with permission from OxfordUniversity Press.



duced by PCA.On the other hand, it is within the range of possibil-

ity that PCA-derived electrophilic quinones also playsome important roles in anti-carcinogenic activity ofPCA. An appropriate amount of electrophilic agentssuch as isothiocyanates is well known to stimulatephase II enzyme induction to eliminate chemical car-cinogens. Moreover, suppressive effects of PCA ontumor development, when its dosage is applied only ashort time before the TPA application, could be ex-plained by the cytotoxic properties of ortho-quinonesattenuating TPA-induced mitogenic effect, whereas theGSH depletion and consequent oxidative stresspotentiate the oxidative and thus tumor-promoting ac-tion of TPA when applied 3 h before TPA. Althoughmore extensive studies on action mechanisms shouldbe examined, the present results lead us to a hypoth-esis that PCA, in both the protective and enhancingdosage regimes, exerts i ts action through theelectrophilic quinone oxidation products.

Compounds with a carboxylic acid (COOH) moietyincluding benzoic acid are also known to be immedi-ate- and non-immunological-type irritants in skin(Lahti et al. 1995). In addition, prior treatment ofmouse skin with acetic acid, whose pKa (acidity) valueis quite similar to benzoic acid derivatives, increasedthe yield of tumors initiated by urethane (Pound 1966).However, the action mechanism remains unclear. Asfor disturbance of GSH detoxification systems, the in-volvement of free COOH group could be ruled out,since some catechol compounds having no COOHgroup exhibited significant GSH consumption (data notshown). The permissive role of the COOH moiety of

PCA should be elucidated.In the present study, the competitive tyrosinase in-

hibitor ArB (Maeda and Fukuda 1996) counteractednot only GSH consumption and enzyme activity inhi-bition by a high dose of PCA but also PCA-inducedenhancement of inflammatory oxidative stress (Table6 and Figs. 7, 8). Tyrosinase, which catalyzes thehydroxylation of monophenols and the oxidation ofcatechols to their quinone compartments (Mayer 1987;Cooksey et al. 1996; Riley et al. 1997) as well as theconversions of L-tyrosine to L-dopa and L-dopa to L-dopaquinone, is regarded as the rate-limiting steps ofmammalian melanin synthesis (Hearing and Jiménez1987). Actually, we have detected significant tyrosi-nase activity in ICR mouse skin (Nakamura et al.2001a, Section 3). Evidence for the generation ofsuperoxide anion through enzymatic action of tyrosi-nase has been reported (Koga et al. 1992). As men-tioned above, PCA induced redox-cycle-dependent li-pid peroxidation and oxidative DNA damage, suggest-ing that PCA itself chemically generates ROS. In ad-dition, several phenolic toxicants such as urushiols(poison ivy) cause immunomodulation such as type IVhypersensitivity through hapten-protein complex for-mation in skin keratinocytes and Langerhans cells (Parket al. 1998). We have reported that treatment with highdoses of PCA induced hypersensitivity in mouse skin(Nakamura 2001a, section 3). Tyrosinase was a rate-limiting factor of inhibition of GST and GPx by PCAin vitro. PCA alone did not inhibit enzyme activitieswhereas addition of tyrosinase resulted in inhibition(data not shown). These results suggested that the for-mation of PCA-quinone intermediate(s) might occur

Group Pretreatment (timea) Stimuli Oxidative stress parameter


(mg/punch, ratioc) (unit/punch, ratioc) (nmol/punch, ratioc)

1 acetone (3) acetone 26.2 ± 2.8 æ 0.20 ± 0.04 æ 0.83 ± 0.06 æ2 acetone (3) TPA 8.1 nmol 57.9 ± 9.2d 100 0.85 ± 0.10e 100 1.12 ± 0.41 100

3 PCA 20,000 nmol (0) TPA 8.1 nmol 34.6 ± 7.1d,g 60 0.25 ± 0.04e,h 29 1.04 ± 0.23d 93

4 PCA 20,000 nmol (3) TPA 8.1 nmol 79.5 ± 9.8d,f 137 3.67 ± 0.39e,h 431 1.64 ± 0.21e 146

Table 6. Modifying effects of PCA pretreatment on short term TPA application-induced enhanced oxidative stress parametersin B6C3F1 mouse skin. Reprinted from Free Radical Biology & Medicine, 30(9), Nakamura et al., A catechol antioxidantprotocatechuic acid potentiates inflammatory leukocyte-derived oxidative stress in mouse skin via a tyrosinase bioactivationpathway, 967–978, Copyright (2001), with permission from Elsevier.

ICR mice (5 mice in each group) were treated as described in Fig. 1.aThe time interval (h) between PA pretreatment and TPA (8.1 nmol) stimulation.bValues are means ± SD of 5 mice.cValues are percentages of those in group 2 (TPA stimulation alone).dSignificant versus group 1 (Student’s t-test); P < 0.05.eSignificant versus group 1 (Student’s t-test); P < 0.001.fSignificant versus group 2 (Student’s t-test); P < 0.05.gSignificant versus group 2 (Student’s t-test); P < 0.01.hSignificant versus group 2 (Student’s t-test); P < 0.001



by means of tyrosinase. In any case oxidative conver-sion of PCA to PCA-quinonoid(s) by tyrosinase islikely to be the critical step in enhanced oxidative stressin mouse skin.

Antioxidants have been considered as a double-edgedsword in cancer control. Recently, an interesting studyon the evaluation of the sensitivity of transgenic micewith overexpression of GPx or both GPx andsuperoxide dismutase to skin tumor promotion wasreported by Lu et al. (1997). Surprisingly, thesetransgenic mice showed an enhanced tumorigenic re-sponse to application of DMBA/TPA. The mechanismresponsible for this phenomenon is not yet well under-stood. They, however, have presumed that the alteredROS-detoxification enzyme levels might influence theprocess of carcinogenesis by modulating cell growthphenotype, increasing resistance of cells with oxidativedamage, or by altering immune function. These resultsindicated the difficulty in regulation of ROS level andin prevention of cancer by radical scavenging-typeantioxidants.

3. A catechol antioxidant potentiates inflamma-tory leukoc yte-derived oxidative stress inmouse skin via a tyrosinase bioactivation path-way

3-1. Introduction

Acute and chronic inflammatory states have beenimplicated as mediators of a number of pathologicaldisorders including cancer. Chronic inflammation ap-pears to be linked to tumorigenesis in the lung, thebowel, the bladder, the colon and the skin. Althoughthe mechanisms by which inflammatory cells showtheir carcinogenic effects remain unclear, some poten-tial pathways have been proposed (Cerutti and Trump1991). Inflammatory cells produce a highly compli-cated mixture of growth and differentiation cytokinesas well as biologically active arachidonate metabolites.In addition, they possess the ability to generate andrelease a spectrum of ROS and free radicals duringoxidative burst. Among inflammatory cells, polymor-phonuclear leukocytes (PMNs) are particularly adeptat generating and releasing ROS, including superoxide(O2

–), H2O2, hypochlorous acid (HOCl), singlet oxy-gen (1O2) and hydroxyl radical (•OH) (Hurst andBarette 1989; Ramos et al. 1992; Steineck et al. 1992).The generation of O2

– by PMNs is attributed to theactivation of a plasma membrane NADPH oxidase.Utilization of O2

–-derived H2O2 by MPO results in theformation of HOCl. Further reaction of HOCl withH2O2 generates 1O2. In addition, •OH is generated fromthe interaction of HOCl with O2

– (Ramos et al. 1992).Current evidence suggests that these ROS-derived in-flammatory cells may be important in tumorigenesis.Recently we have reported that O2

– from leukocytes

plays an important role for continuous and excessiveproduction of chemotactic factors, leading to chronicinflammation and hyperplasia in mouse skin(Nakamura et al. 1998a, 1999). As mentioned above,ROS production by double or multiple TPA treatmentsis closely associated with the metabolic activation ofproximate carcinogens (Kensler et al. 1987, 1989; Jiand Marnett 1992) and the increased levels of oxidizedDNA bases (Wei and Frenkel 1991, 1992, 1993). Takentogether, the regulation of ROS from activatedleukocytes is proposed to be one of the most promis-ing strategies for cancer control (Nakamura et al.2000b).

Section 2 (Nakamura et al. 2000d) demonstrates notonly the lack of an inhibitory effect but also signifi-cant enhancement of mouse skin tumor promotion bypretreatment with a high dose (>1,600 nmol) of a sim-ple phenolic compounds, PCA, at an appropriate in-terval (3 h before TPA application). The possibility thatmetabolism by tyrosinase activity of PCA to certaincompound(s) without antioxidative properties and/orwith tumor promotional potency has also been specu-lated. Thus the present study addresses whether or notPCA modifies TPA-induced chronic inflammation inmouse skin using the multiple-application model. Togain further evidence that the tyrosinase-derived reac-tive quinone intermediate(s) of PCA is indeed involvedin skin inflammatory responses, we examined the en-hancing effects of PCA on TPA-induced inflammationin both a TPA-resistant and tyrosinase-dominant mousestrain. In addition, contact hypersensitivity by PCA isdocumented.


3-2A. Chemicals and animalsFemale B6C3F1 mice (7 weeks old) were obtained

from Charles River Japan Inc., Yokohama City, Japan.Other chemicals are described in Section 2.

3-2B. Tyrosinase activity determinationTyrosinase L-dopa oxidation activity was determined

as previously reported (Dawley and Flurkey 1993).Mice were sacrificed by cervical dislocation. The skinsamples were minced in 2 ml of PBS (pH 7.4) contain-ing 1% Triton X-100, and then homogenized twice at4∞C for 30 s. The homogenate was centrifuged at10,000 g for 20 min at 4∞C. To each 2-ml cuvette, 1 mlof 0.02% (w/w) L-dopa solution and 100 ml of homoge-nate supernatant were added and quickly mixed. In-crease in absorption at 475 nm for 2 min at 0.1 minintervals were recorded. The oxidation activity wascalculated from the linear portion of the curve and ex-pressed as units of tyrosinase per mg protein. One unitof tyrosinase activity was defined as the activity thatoxidized 1 mmol of L-dopa per min at 25∞C.



3-2C. Double and multiple TPA treatment protocol(short and medium term assay)

One group was composed of 5 female ICR or B6C3F1mice housed 5 in a cage. PCA (81, 8100 nmol, and 20mmol/0.1 ml in acetone) was topically applied to theshaved area of dorsal skin at various times before ap-plication of a TPA solution (8.1 nmol/0.1 ml in acetone).This TPA dose (8.1 nmol) was used for the potentiationof oxidative responses compared with the dose for skincarcinogenesis experiment (1.6 nmol). In the short termassay, the same doses of TPA and test compounds oracetone were applied twice at an interval of 24 h forH2O2 determination, and these agents were appliedonce for edema and MPO activity measurement. In themedium term assay, PCA (16 or 1600 nmol/0.1 ml inacetone) was topically applied to the shaved area ofdorsal skin 30 min before application of a TPA solu-tion (1.6 nmol/100 ml in acetone). The same doses ofTPA and test compounds or acetone were applied twicea week for 5 weeks, as shown in Fig. 9.

3-2D. Determination of inflammatory biomarker,H2O2 level, TBARS and GSH level

See Section 2.

3-2E. Testing for contact hypersensitivityThe contact hypersensitivity test was done as previ-

ously reported (Kalergis et al. 1997) with some modi-fication. One group was composed of 5 female ICRmice housed 5 in a cage. The mice were treated withArB (10 mmol) or acetone 1 h prior to PCA. Contactsensitization was performed by the topical applicationof 100 ml of a solution of 300 mM PCA or 280 mMisoeugenol (5% w/w) in acetone to the shaved area of

dorsal skin. Five days after sensitization, mice werechallenged with 20 ml of PCA or isoeugenol solutionapplied to both sides of one ear; the other ear was usedas control, being treated only with acetone. After oneweek, mice were re-challenged by the same treatment.Mice were sacrificed by cervical dislocation 48 h afterthe last challenge. The response was determined by theweight of an ear punch, obtained with an 8-mm-diam-eter cork borer, and MPO activity as described above.The ear punch was extracted in 300 ml cold PBS, witha tissue homogenizer on ice, followed by centrifuga-tion at 10,000 g for 20 min to obtain the supernatantextract.

The content of interferon-g (IFN-g) in a mouse earextract 24 h after the PCA-challenge was determinedusing the protocol outlined by the manufacturers in anIFN-g ELISA kit (Endogen, Woburn, MA).

3-2F. Cytotoxicity testHL-60 cells were obtained from the Health Science

Research Resources Bank, Osaka, Japan (Collins et al.1977). The cells were grown in RPMI1640 mediumsupplemented with 10% heat-inactivated fetal bovineserum, penicillin (100 U/ml), streptomycin (100 mg/ml), L-glutamine (0.3 mg/ml), pyruvic acid (0.11 mg/ml) and 0.37% NaHCO3 at 37∞C in an atmosphere of95% air and 5% CO2. The cells, inoculated at 4 ¥ 105

cells/ml in a medium with FBS, were preincubated with1.25% (v/v) DMSO at 37∞C for 6 days, differentiatingthem into granulocyte-like cells. After washing withPBS, the differentiated cells were suspended in a me-dium with FBS were plated at 2 ¥ 106/ml in 12 wellplates and exposed to various concentrations (0.1–1mM) of acetaminophen (AA), eugenol (EU), and PCA

Fig. 9. Experimental protocol of single, double (short-term) and multiple (medium-term) TPA application experiments inmouse skin. Reprinted from Free Radical Biology & Medicine, 30(9), Nakamura et al., A catechol antioxidant protocatechuicacid potentiates inflammatory leukocyte-derived oxidative stress in mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright (2001), with permission from Elsevier.



with TPA (100 nM) or mushroom tyrosinase (25 unit)for 90 min. After incubation, cytotoxicity was assessedby the modified MTT assay using a commercial kit(Cell Counting Kit, Dojindo Laboratories, Kumamoto,Japan). Measurement of GSH was performed spectro-photometrically using a commercial kit(BIOXYTECH® GSH-400™ Assay) as mentionedabove.

3-2G. Protein quantificationConcentrations of proteins were determined with the

BCA protein assay reagent (Pierce) with bovine serumalbumin as the standard.

3-2H. Statistical analysisThe statistical significance of differences between

groups in each assay was assessed by the Student’s t-test (two-sided) that assumed unequal variance.

3-3. Results

3-3A. Tyrosinase activity in the ICR and B6C3F1mouse

The dopa oxidation activity of tyrosinase in the skinfrom albino ICR and black B6C3F1 mice was assayedusing a typical colorimetric method (Dawley andFlurkey 1993). A weak but significant tyrosinase ac-tivity in albino ICR mouse skin was detected. In theextract from B6C3F1 mouse skin, dopa quinone wasproduced in more than ten times the amount as thatfrom ICR mouse skin (Fig. 10). TPA application tothese mouse skins did not affect the tyrosinase activ-ity (data not shown).

3-3B. The modifying effect of PCA on acute inflam-mation in B6C3F1 mouse skin

Because the enhancing effects of high doses of PCAon TPA-induced acute inflammation in TPA-sensitiveand tyrosinase-negative strain albino ICR mouse skinwere observed in the previous study (Nakamura2000d), we determined whether PCA modifies inflam-matory responses in a TPA-resistant and tyrosinase-dominant strain B6C3F1 mouse skin. Inflammatory re-sponses were measured by the weight of a skin punch(edema formation magnitude) and MPO activity, whichis well correlated with the number of infiltratedleukocytes in inflamed regions as confirmed by histo-logical studies (Nakamura et al. 1998a, 2000d;Murakami et al. 2000a, b). Since the maximum re-sponses of acute inflammation were observed in theICR mice used a 3 h-pretreatment protocol (Nakamuraet al. 2000d), we utilized this protocol for the short-term assay in the present study (Fig. 9). As shown inTable 6, the application of 8.1 nmol TPA to B6C3F1mice causes significant edema formation and leukocyteinfiltration by 1.4-fold and 4.3-fold, respectively, com-

Fig. 10. Tyrosinase dopa-oxidation activity in ICR andB6C3F

1 mouse skin. Values are means ± SD of 5 mice. Re-

printed from Free Radical Biology & Medicine, 30(9),Nakamura et al., A catechol antioxidant protocatechuic acidpotentiates inflammatory leukocyte-derived oxidative stressin mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright (2001), with permission from Elsevier.

Fig. 11. Effects of TPA and PCA on MPO activity in ICRand B6C3F

1 mouse skin. ICR and B6C3F

1 mice (5 mice in

each group) were treated as described in “Materials andmethods”. The mice (5 mice in each group) were treated withPCA (20,000 nmol) 3 h before TPA treatment (8.1 nmol).The mice were sacrificed 18 h after TPA treatment and skinpunches were obtained for MPO activity determination. Therelative TPA response value (left column) was expressed bya relative increasing ratio of MPO activity of mice treatedwith TPA (treated) to that of mice with acetone alone (con-trol). The relative PCA response value (right column) wasexpressed by a relative increasing ratio of activity of micetreated with TPA and PCA (treated) to that of mice with TPAalone (control). Reprinted from Free Radical Biology &Medicine, 30(9), Nakamura et al., A catechol antioxidantprotocatechuic acid potentiates inflammatory leukocyte-de-rived oxidative stress in mouse skin via a tyrosinasebioactivation pathway, 967–978, Copyright (2001), with per-mission from Elsevier.



pared with control. On the other hand, double TPAapplication did not significantly increase the residualH2O2 level, which was different only from the case ofICR mice (Nakamura et al. 1998a, 2000d). Althoughthe simultaneous application of PCA (20,000 nmol)with TPA treatment significantly inhibited both edemaformation and MPO activity, the application of the samedose of PCA 3 h before TPA application extremelyenhanced MPO activity by 4.3-fold in B6C3F1 mice.Although we have already reported the modifying ef-fect of a high dose of PCA on inflammation in ICRmouse skin (Nakamura et al. 2000d), we re-examinedthe enhancing effect on MPO activity in ICR mouseskin to compare with that of B6C3F1 mouse skin (Fig.11). We found that the sensitivity of ICR mouse skinto TPA-stimulated inflammation was much higher thanthat of B6C3F1 mouse skin (14-fold vs. 4-fold). Onthe other hand, the sensitivity of tyrosinase-dominantB6C3F1 mice to PCA-enhanced inflammatory responsewas higher than that of albino ICR mice (4-fold vs 2-fold).

We next assessed whether redox alteration is involvedin the inverse effects of PCA on inflammatory re-sponses in mouse skin. Since the treatment of mouseskin with TPA alone showed little influence on totalGSH level (Reiners et al. 1991), the modifying effect

of PCA on GSH level in B6C3F1 mouse skin was ex-amined. As shown in Fig. 12, the GSH level in mouseskin was reduced by more than 50% with 20,000 nmolof PCA for 3 h. This reduction was completely inhib-ited by the co-administration of the tyrosinase inhibi-tor ArB (10 mmol), which alone showed no modula-tory effect on GSH level (data not shown).

3-3C. Modifying effect of PCA on multiple TPA ap-plication-induced inflammatory responses andoxidative stress in ICR mouse skin

Although the dose-dependency of modification ofskin tumor development by PCA was observed in theresults of a short-term experiment of TPA-inducedacute inflammation, the dose of PCA required to ex-hibit an enhancing effect on a single TPA application-induced acute inflammation (20,000 nmol) was morethan ten times higher than that for tumor developmentenhancement (1,600 nmol) (Nakamura et al. 2000d).Therefore, the possibility that multiple applications ofan appropriate dose of PCA could modify chronic in-flammation was assessed by multiple TPA application-induced edema formation and MPO activity enhance-ment as well as ROS generation in ICR mouse skinusing the same protocol (Fig. 9) as skin carcinogen-esis experiment (Nakamura et al. 2000d). As shown inTable 7, the application of a low dose (16 nmol) ofPCA, which strongly inhibited tumor promotion(Nakamura et al. 2000d), 30 min before each 1.6 nmolTPA treatment inhibited edema formation, MPO ac-tivity, and the increase in H2O2 level by 27% (P < 0.01),72% (P < 0.01) and 25%, respectively. Whereas, theapplication of 1,600 nmol PCA 30 min before each TPAtreatment significantly enhanced edema formation andMPO activity by 29% (P < 0.01) and 136% (P < 0.001),respectively, and showed little effect on H2O2 level.On the other hand, the application of 1,600 nmol PCA30 without TPA stimulation exhibited no significantchanges in these parameters.

As topically applied PCA significantly modifiedmultiple TPA application-induced inflammation, weinvestigated whether such PCA treatment influencesmultiple TPA application-induced oxidative damageevaluated by the TBARS level. The quantitative datafor the levels of TBARS formation in mouse epider-mis homogenate are shown in Fig. 13. The increasedlevel of TBARS caused by the multiple TPA applica-tion was significantly higher than that of the control(1.01 ± 0.23 versus 0.40 ± 0.08 nmol/cm2, P < 0.01).The multiple application of PCA (16 nmol) 30 minbefore each TPA treatment inhibited the increase inTBARS level (0.73 ± 0.30 nmol/cm2). On the otherhand, the application of PCA (1,600 nmol) 30 min be-fore TPA treatment significantly enhanced TBARS for-mation (1.60 ± 0.39 nmol/cm2, P < 0.05 versus TPA).

Fig. 12. Effects of PCA on total GSH level in B6C3F1 mouse

skin. The mice (5 mice in each group) were treated with ar-butin (10 mM) or acetone 1 h prior to PCA (20,000 nmol)treatment. The mice were sacrificed 3 h after PCA applica-tion, and skin samples were removed for GSH assays. Sig-nificance determined by the Student’s t-test is expressed.Reprinted from Free Radical Biology & Medicine, 30(9),Nakamura et al., A catechol antioxidant protocatechuic acidpotentiates inflammatory leukocyte-derived oxidative stressin mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright (2001), with permission from Elsevier.



3-3D. PCA-induced contact hypersensitivity in ICRmouse skin

The possibility that oxidative metabolism of PCA toa benzoquinone form, which can readily conjugate tonucleophiles such as GSH, prompted us to determinewhether treatment with a high dose of PCA exhibitedcontact hypersensitivity due to protein modification byhapten such as conjugation with a sulfhydryl residue.Protocols that lead to the induction of contact hyper-sensitivity to haptens such as dinitrofluorobenzene arewell established. As shown in Table 8, isoeugenol, apotent sensitizer from clove bud oil, induced signifi-cant ear swelling after 48 h post-challenge (113% ofcontrol, P < 0.05). PCA also significantly increasedear weight (119% of control, P < 0.05), redness (datanot shown), MPO activity (145% of control, P < 0.05)and IFN-g production (242% of control, P < 0.05) com-parable to isoeugenol. On the other hand, pretreatmentof the tyrosinase inhibitor ArB 1 h before sensitizationsignificantly inhibited inflammatory responses (inhibi-tory effect = 58%, 73%, and 63%, respectively).

3-3E. Cytotoxic effect of PCA in cultured cells de-pendent on tyrosinase and independent onMPO

Some phenolic compounds, including EU, have beenreported to be oxidized by MPO to a reactive interme-diate and also to be cytotoxic (Thompson et al. 1989).The in vitro data regarding the oxidation of catecholsby tyrosinase, the reactivity of the corresponding qui-none to thiol groups, and the cytotoxicity of tyrosi-nase-derived ortho-quinones have been described(Cooksey et al. 1996; Riley et al. 1997). These obser-

vations led us to examine whether MPO and/or tyrosi-nase can activate PCA. The cytotoxicity of AA, EUand PCA to differentiated HL-60 cells was observedin both the presence and absence of TPA, being able tostimulate the generation of a large amount of O2

– inthese cells (Thompson et al. 1988: Nakamura et al.1998a). In the absence of both TPA and tyrosinase, allcompounds dose-dependently and significantly inducedcytotoxicity (data not shown). While TPA stimulationsignificantly enhanced the cytotoxicity of EU and AA,no significant change was observed in the PCA-treatedcells (Table 9). A concentration of 0.1 mM EU caused34% cell death in TPA-stimulated cells as opposed to18% cell death in unstimulated cells. TPA stimulationto differentiated HL-60 cells treated with EU also sig-nificantly reduced intracellular total GSH level to 73%of the unstimulated control as previously reported(Thompson et al. 1989). The same concentration ofPCA caused 36% cell death in both cells. On the otherhand, addition of tyrosinase to the culture medium sig-nificantly enhanced PCA-induced cell death (43% sur-vival, P < 0.001), but did not alter cytotoxicity of EUor AA. Addition of tyrosinase to PCA-treated cells alsoreduced GSH level to 72% of control (P < 0.001). Inaddition, PCA alone did not directly react with GSHwhereas addition of tyrosinase resulted in formation aPCA-GSH adduct (Nakamura et al. 2014).

3-4. Discussion

PCA is well known to be one of the major strongantioxidants from vegetables and fruit, and a promis-ing cancer chemopreventor as mentioned above. The

Group Oxidative stress parameter


(mg/punch, ratioc) (unit/punch, ratioc) (nmol/punch, ratioc)

1 acetone acetone 27.8 ± 4.6 æ 0.17 ± 0.05 æ 0.83 ± 0.05 æ2 PCA 1,600 nmol acetone 27.9 ± 6.6 æ 0.20 ± 0.09 æ 0.90 ± 0.18 æ3 acetone TPA 1.6 nmol 44.8 ± 5.0e 100 1.83 ± 0.45e 100 2.27 ± 0.32e 100

4 PCA 1,600 nmol TPA 1.6 nmol 57.8 ± 7.7e,f 129 4.32 ± 0.39e,g 236 1.94 ± 0.50d 85

5 PCA 16 nmol TPA 1.6 nmol 32.5 ± 4.3f 73 0.52 ± 0.07e,f 28 1.71 ± 0.34d 75

ICR mice (5 mice in each group) were treated as described in Fig. 2.aThe time interval between pretreatment and stimuli application was 30 min.bValues are means ± SD of 5 mice.cValues are percentages of those in group 3 (TPA stimulation alone).dSignificant versus group 1 (Student’s t-test); P < 0.01.eSignificant versus group 1 (Student’s t-test); P < 0.001.fSignificant versus group 3 (Student’s t-test); P < 0.01.gignificant versus group 3 (Student’s t-test); P < 0.001.

Table 7. Modifying effects of PCA pretreatment on multiple TPA application-induced enhanced oxidative stress parameters inICR mouse skin. Reprinted from Free Radical Biology & Medicine, 30(9), Nakamura et al., A catechol antioxidant protocate-chuic acid potentiates inflammatory leukocyte-derived oxidative stress in mouse skin via a tyrosinase bioactivation pathway,967–978, Copyright (2001), with permission from Elsevier.



recent study, however, demonstrated the significantenhancement of mouse skin tumor promotion by pre-treatment with a high dose of PCA at an appropriateinterval, and the possibility of metabolism of PCA tocertain compound(s) without antioxidative propertiesand/or with tumor promotional potency (Nakamura etal. 2000d). The present study utilizing acute andchronic inflammation systems in mouse skin providedfurther evidence that the tyrosinase-derived reactivequinone intermediate of PCA may be involved in skininflammatory responses such as recruitment ofleukocytes generating tumor-promoting oxygen radi-cals. The puzzling result that the treatment of the typi-cal phenolic antioxidant unexpectedly enhancedoxidative stress, even as the application dose is in-creased, is the topic of this study.

Previous studies had revealed the significant differ-ences in sensitivity to multistage skin carcinogenesisbetween mice that are sensitive (e.g., SENCAR, CD-1, etc.) and resistant (C57BL/6J, B6C3F1, etc.) to TPA(Slaga and Fischer 1983). The plausible primary de-terminants in the strain-dependent differences relatedto tumor promotion including TPA-induced epidermalhyperplasia, induction of dark cells, 8-lipoxygenase,serum granulocyte-macrophage colony-stimulatingfactor levels, leukocyte infiltration but neither orni-

thine decarboxylase nor protein kinase C activities havebeen demonstrated (Sisskin et al. 1982; Fischer et al.1988, 1989; Mills and Smart 1989). In particular, lo-cal production of ROS by inflammatory cells has beensuggested to significantly contribute to skin tumor de-velopment (Yoon et al. 1993). In addition, Wei et al.(1993) clearly indicated that the TPA-mediatedoxidative events and oxidative DNA modification bydifferent doses of TPA closely correlated with the dose-dependent promoting potencies of TPA in both mousestrains. Our findings further confirmed that TPA-re-sistant strain mice exhibited fewer inflammatory re-sponses to topical TPA application than the sensitivestrain ICR does. This confirmation is based on the factthat TPA application with B6C3F1 mice resulted inslight MPO activity enhancement and little H2O2 gen-eration less than those in ICR mice (Fig. 11 and seeWei et al. 1993).

It is certain that tyrosinase is not involved in TPA-induced tumor promotion because tyrosinase activityis expressed in the TPA-resistant B6C3F1 mouse tentimes greater than in TPA-sensitive albino ICR mouseskin (Fig. 10). On the other hand, pretreatment of ahigher dose of PCA dramatically enhanced infiltrationof inflammatory leukocytes but slightly increased theH2O2 level in B6C3F1 mouse skin (Table 6), since TPA-stimulated peritoneal macrophages of B6C3F1 micehave far less ability to generate O2

– than those of theTPA-sensitive SENCAR mouse (Yoon et al. 1993). Itshould, however, be noted that the tyrosinase-domi-nant B6C3F1 mouse was more sensitive to inflamma-tory responses by PCA application than the albino ICRmouse (Fig. 11). In addition, GSH level in skin wasalso reduced by treatment of 20,000 nmol PCA (Fig.12). As profoundly discussed below, PCA itself, de-pendent on tyrosinase and independently of TPA, canreact with a variety of cellular components, includingGSH, and may have an immunoinflammatory effectthrough recognition of PCA-protein adducts as anti-gen by macrophages (Park et al. 1998).

Our recent study demonstrated that topical applica-tion of PCA exerts contrasting effects on TPA-inducedtumor promotion in mouse skin in both dose- and tim-ing-dependent manners. A high dose of PCA (1,600nmol; 1,000-fold dose of TPA) 40 min prior to TPAapplication significantly enhanced the multiplicity ofskin tumors (Nakamura 2000d). A similar tendency ofdose-dependent effects of PCA application on skin car-cinogenesis was also observed in the short-term ex-periment of TPA-induced acute inflammation. How-ever, the dose of PCA required to exhibit an enhancingeffect on a single TPA application-induced acute in-flammation (20,000 nmol) was much higher than thatfor enhancement of tumor development (1,600 nmol).The chronic inflammation experiment of multiple TPAtreatments for 5 weeks clearly demonstrated that PCA1,600 nmol, enough to enhance skin tumor promotion,

Fig. 13. Modifying effects of PCA on TBARS formation inthe mouse epidermis. ICR mice (5 mice in each group) weretreated by the multiple treatment protocol as described inFig. 1 and “Materials and methods”. Mouse skin was treatedwith PCA (1,600 nmol) or acetone 30 min prior to each TPAtreatment. The mice were sacrificed 1 h after the last TPAapplication, and their epidermis was removed for TBARSassays. Entry 1, untreated control treated only with acetone;2, PCA (1,600 nmol) + acetone instead of TPA stimulation;3, acetone + TPA (1.6 nmol); 4, PCA (1,600 nmol) + ac-etone instead of TPA stimulation; 5, PCA (16 nmol) + TPA.Significance determined by the Student’s t-test is expressedas: a, versus group 1, P < 0.05; b, versus group 3, P < 0.05.Reprinted from Radical Biology & Medicine , 30(9),Nakamura et al., A catechol antioxidant protocatechuic acidpotentiates inflammatory leukocyte-derived oxidative stressin mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright (2001), with permission from Elsevier.



significantly elevated leukocyte infiltration comparedto TPA-treated mice (Table 7). As mentioned above,MPO, found in monocyte and in the primary granulesof PMNs (Klebanoff 1991), catalyzes the formation ofHOCl using H2O2 as a substrate. HOCl is high toxic,mutagenic (Bernofsky 1991), and also activates car-cinogens (Petruska et al. 1992). The results in this re-port also obviously demonstrated that multiple treat-ment of high doses of PCA significantly enhanced

TBARS formation (Fig. 13), known as an overalloxidative damage biomarker formed downstream ofH2O2 generation in the presence of a metal ion as acatalyst. Thus, trends of oxidative stress consequenton inflammatory responses of PCA were very closelycorrelated with modifying effects on tumor develop-ment (Nakamura et al. 2000d). These results stronglysuggested that inflammatory leukocytes as oxygen radi-cal generators and activators in inflammatory regions

Table 8. Contact hypersensitivity to PCA and isoeugenol in ICR mouse skin. Reprinted from Free Radical Biology & Medi-cine, 30(9), Nakamura et al., A catechol antioxidant protocatechuic acid potentiates inflammatory leukocyte-derived oxidativestress in mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright (2001), with permission from Elsevier.

ICR mice (5 mice in each group) were treated as described in Fig. 1 (medium term).aThe time interval between pretreatment and stimuli application was 30 min.bValues are means ± SD of 5 mice.cValues are percentages of those in group 1 (TPA stimulation alone).dSignificant versus group 1 (Student’s t-test); P < 0.05.eSignificant versus group 1 (Student’s t-test); P < 0.001.fSignificant versus group 3 (Student’s t-test); P < 0.05.gSignificant versus group 3 (Student’s t-test); P < 0.001.

Table 9. Cytotoxicity and GSH consumption of AA, EU, and PCA in the presense and absense of TPA or tyrosinase. Reprintedfrom Free Radical Biology & Medicine, 30(9), Nakamura et al., A catechol antioxidant protocatechuic acid potentiates in-flammatory leukocyte-derived oxidative stress in mouse skin via a tyrosinase bioactivation pathway, 967–978, Copyright(2001), with permission from Elsevier.

aConcentrations; phenolic compounds, 0.1 mM; TPA, 100 nM; tyrosinase, 25 unit (20 mg protein).bValues are means ± SD (n = 3).cValues are percentages of those in control (DMSO alone).dSignificant versus control group (Student’s t-test); P < 0.05.eSignificant versus control group (Student’s t-test); P < 0.001.fSignificant versus group treated only with each phenolic compound (Student’s t-test); P < 0.01.gSignificant versus group treated only with each phenolic compound (Student’s t-test); P < 0.001.

Group Challengea Ear swellingb MPO activityb IFN-g levelb

(mg/punch, ratioc) (unit/punch, ratioc) (ng/punch, ratioc)

1 acetone 14.6 ± 0.7 100 0.095 ± 0.012 100 215 ± 21 100

2 isoeugenol 16.5 ± 0.9d 113 0.118 ± 0.015 129 467 ± 36d 217

3 PCA 17.3 ± 1.5d 119 0.138 ± 0.016d 145 521 ± 42d 242

4 PCA/arubutin 15.7 ± 0.7f 108 0.106 ± 0.013f 112 330 ± 32d,g 153

Phenolic compounda Activatora Cell viabilityb GSH levelb

(% to controlc) (nmol/106 cells)

Control (DMSO) æ 100.0 ± 3.2 2.12 ± 0.06AA æ 89.2 ± 1.2d 2.03 ± 0.12AA TPA 74.7 ± 3.6e,f 1.98 ± 0.18AA tyrosinase 92.2 ± 7.3 2.05 ± 0.13EU æ 82.5 ± 3.7e 2.00 ± 0.04EU TPA 66.3 ± 3.5e,g 1.46 ± 0.10e,f

EU tyrosinase 85.1 ± 6.2d 2.11 ± 0.14PCA æ 64.3 ± 5.2e 1.97 ± 0.16PCA TPA 63.7 ± 2.3e 2.08 ± 0.19PCA tyrosinase 43.1 ± 6.8e,g 1.53 ± 0.07e,f



play an important role in enhancement of skin tumorpromotion by PCA.

Contact hypersensitivity induced by several toxicantsincluding urushiols (from poison ivy) is generally ac-cepted to involve cell-mediated immune reactions andto be associated with T cell protective immunity. Atthe first sensitization, this hapten penetrates into theepidermis to form hapten-protein conjugates that actas sensitizers. These sensitizers are then internalizedby epidermal Langerhans’ cells, which migrate to thelymph nodes to express memory CD4

+ cells (Th-1cells). A secondary challenge then elicits a series ofcytokine responses to promote T cell activation andproliferation. We demonstrated for the first time thatPCA alone induced contact hypersensitivity responsessuch as ear swelling and leukocyte infiltration, whichare inhibited by a tyrosinase inhibitor, comparable toan already-known sensitizer, isoeugenol (Table 8).IFN-g is a crucial Th-1 cytokine in contact hypersen-sitivity because its production by activated Th-1 cellsis believed to lead to local recruitment of inflamma-tory leukocytes (Fong and Mosmann 1989; Tsuji et al.1997). We measured IFN-g in PCA-challenged mouseear and found that the tyrosinase-dependent increasedlocal concentration levels at 24 h as expected. Thesedata also strongly indicated the occurrence of T cell-mediated cellular immunoresponse, followed byleukocyte infiltration, due to a tyrosinase bioactivationpathway-dependent hapten-protein formation (Fig. 14).

Bioactivation to chemically reactive species is usu-ally dependent on oxidative phase II biotransforma-tion which is performed principally by cytochrome P-450 (CYP) enzymes. However, a non-specific CYPinhibitor SKF525A, even up to a dose of 20 mmol, didnot counteract a PCA-induced modifying effect on in-flammatory responses in mouse skin (Nakamura et al.2014). Drug bioactivation can also be catalyzed by non-CYP enzymes such as those found in neutrophils. Thereis significant evidence that MPO dependent activationof a broad spectrum of chemicals, including proximatecarcinogens, has been described (Kensler et al. 1987;Thompson et al. 1989; Petruska et al. 1992). For ex-ample, EU, naturally occurring phenolic compoundassociated with severe and acute pulmonary illness inhumans, is oxidized into a more reactive and toxicquinone intermediate requiring MPO in the presenceof H2O2. This bioactivation of EU by MPO has beenconfirmed directly using the isolated MPO and by im-plication using PMNs containing both MPO and H2O2generation system (Thompson et al. 1989). Since HL-60 cells can be differentiated into granulocyte-like cellsexpressing both MPO activity and a H2O2 generatingNADPH oxidase system (Thompson et al. 1988), weutilized this cell line as a model of PMNs in the presentstudy. The treatment of EU with TPA, stimulating H2O2generation, significantly enhanced toxicity and GSHdepletion (Table 9). On the other hand, toxicity of PCA

was not potentiated by stimulation with TPA. PCA-induced cytotoxicity and GSH consumption were alsounchanged by the exogenous addition of MPO (datanot shown). These results indicated that MPO-depend-ent activation of PCA is ruled out in the toxic mecha-nism. The oxidative phenomenon of electrophilic ad-dition to sulfhydryl groups was supported by the invitro findings that PCA significantly reduced the GSHlevel in both a cultured cell system (Table 5) and acell-free system (Nakamura et al. 2014) only when ty-rosinase was co-existing. It is, thus, very likely thatthe formation of PCA-quinone intermediate(s) and fur-ther reaction with proteins may occur by means of ty-rosinase. Although differences in the substratespecificity between mammalian and mushroom tyro-sinase should be taken into consideration, we havemore recently observed that mushroom tyrosinase spe-cifically oxidized some catechol compounds includ-ing caffeic acid but not hydroquinones or monophenols

Fig. 14. A plausible molecular mechanism of PCA for in-flammatory oxidative stress in mouse skin. Reprinted fromFree Radical Biology & Medicine, 30(9), Nakamura et al.,A catechol antioxidant protocatechuic acid potentiates in-flammatory leukocyte-derived oxidative stress in mouse skinvia a tyrosinase bioactivation pathway, 967–978, Copyright(2001), with permission from Elsevier.



(Nakamura et al. 2014). A typical phenolic antioxidantquercetin, having a catechol moiety, can also be oxi-dized by tyrosinase and conjugated with GSH (Boersmaet al. 2000). Further study on structure-activity rela-tionship of phenolic antioxidants to mammalian tyro-sinase-dependent oxidation using isolated mouse skinkeratinocytes should provide valuable information onthe toxicology of antioxidants in skin.

4. Thiol modification by bioactivated polyphenolsand its potential role in skin inflammation

4-1. Introduction

Tyrosinase (EC 1.14.18.1) is one of the most popu-lar phenol oxidases widely distributed through the ani-mal, vegetal, bacterial, and fungal kingdoms, and ca-talyses two types of reactions on phenolic substratesat a binuclear copper center; (a) the ortho-hydroxylation of the monophenols into o-diphenols(monophenolase activity), and (b) the oxidation of o-diphenols to o-quinones (diphenolase activity) (García-Molina et al. 2007). The product of both reactions isthe o-quinone of the corresponding monophenol/diphenol (García-Molina et al. 2007). Besides oxidiz-ing their intact substrates, such as L-tyrosine and L-dopa, tyrosinase has a wide substrate specificity onmonophenols and o-diphenols. Tyrosinase participatesin different physiological processes, such as fruit andvegetable browning and pigmentation in animals(García-Molina et al. 2007). Since tyrosinase is be-lieved to be one of the potential pharmacological tar-gets against cutaneous hyperpigmentation and malig-nant melanoma, tyrosinase substrates as well as tyro-sinase inhibitors have been used as chemotherapeuticagents to treat these diseases (Muñoz-Muñoz et al.2010). For example, 4-hydroxyanisole is a substratefor tyrosinase and exhibits depigmentation and tumorshrinkage by its topical application (Riley 1969) andintra-arterial infusions into the legs (Morgan 1984).However, 4-hydroxyanisole clinical trails were termi-nated because serious kidney and liver damage oc-curred as side effects (Rustin et al. 1992). We also dem-onstrated that topical pretreatment with a high dose(>10 mmol) of PCA, a benzoic acid derivative havinga catechol moiety, significantly enhanced mouse skininflammation, oxidative stress, and tumor promotionin a tyrosinase-dependent manner (Nakamura et al.2000d, 2001a). Treatment with a high dose of PCAalone for 3 h also decreased GSH level and detoxifica-tion enzyme activities in mouse skin, which was coun-teracted by the tyrosinase inhibitor ArB (Nakamura etal. 2000d, 2001a). Therefore, the tyrosinase-depend-ent bioactivation of phenolic substances is a prerequi-site not only for depigmentation and anti-melanoma,but also for dermatotoxicity in normal tissue under

specific conditions.Dietary polyphenols, such as flavonoids and

phenylpropanoids, are the most common and widelydistributed phytochemicals in plants (Bravo 1998). Awealth of data suggests that most of the relevant mecha-nisms of disease prevention by polyphenols, such asin cancer, are not related to their antioxidant proper-ties, but are rather due to their pro-oxidant action anddirect interaction with the target molecules (Galati andO’Brien 2003). For example, in slightly alkaline solu-tions, the green tea catechins undergo autoxidation toform ROS, including superoxide and hydrogen perox-ide, resulting in polymerization and decomposition(Nakayama et al. 2002). We recently found that 3,4-dihydroxyphenyl acetic acid and (–)-epigallocatechin-3-gallate form covalent adducts with GSH or proteinthiols in vitro (Ishii et al. 2008, 2009). ROS andelectrophiles can cause oxidative modifications to sen-sitive proteins that can lead to changes in protein func-tion (Nakamura and Miyoshi 2010). However, the re-lationship between the tyrosinase-dependentbioactivation of polyphenols and their modifying abil-ity of thiols has not yet been well characterized.

In the present study, we evaluated the tyrosinase-dependent modifying effect of several simple phenoliccompounds on GSH, the most abundant intracellularsulfhydryl. We observed the in vitro chemical modifi-cations in GSH and a model protein by the bioactivatedcatechol-type phenolic compound. We also examinedthe enhancing effects of the phenolic compounds onTPA-induced inflammation in mice to suggest the pos-sible involvement of their bioactivation in the inflam-mation enhancement.


4-2A. Chemicals and animalsButylated hydroxytoluene (BHT) was purchased

from Nacalai Tesque (Kyoto, Japan). Laccase was ob-tained from Sigma-Aldrich (St. Louis, MO, USA).Glyceraldehydes-3-phosphate dehydrogenase(GAPDH) was purchased from GE Healthcare UK Ltd.(Buckinghamshire, UK). Iodoacetyl-LC-biotin (IAB)was purchased from Thermo Fisher Scientific(Rockford, IL, USA). All other chemicals were pur-chased as mentioned before.

4-2B. Determination of GSH level, GSH-PCA adductand free protein sulfhydryls

Measurement of the GSH level was performed asmentioned above. Detection of the GSH-PCA adductwas performed by nuclear magnetic resonance (NMR)spectrometry using a Bruker AMX400 (400 MHz;Bruker Daltonics, Ltd., Karlsruhe, Germany) as previ-ously reported (Ishii et al. 2009). For the detection ofthe free protein sulfhydryls, GAPDH (0.2 mg/mL) was



incubated with PCA at the indicated concentrations for1 h at 37∞C in a buffer containing 50 mM phosphatebuffer (pH 7.5) with or without laccase (30 units). Thereaction was terminated by centrifugal filtration(Microcon 30, molecular weight cutoff of 30,000;Millipore Co., Billerica, MA, USA) to remove the lowmolecular weight reactants. The samples were treatedwith IAB according to the manufacturer’s protocol andthen separated by 10% SDS-PAGE. The western blot-ting and band visualization were performed as previ-ously reported (Ishii et al. 2009).

4-2C. Inflammatory biomarker determinationThe modifying effect of the phenolic compounds on

a single TPA application-induced inflammation was de-termined by two biomarkers, i.e., edema formation andmyeloperoxidease (MPO) activity, as mentioned above(Sections 2 and 3). Female ICR mice (7 weeks old)were obtained from Japan SLC, Shizuoka, Japan.

4-2D. Statistical analysisAll values were expressed as means ± SD. Statisti-

cal significance was assessed by Student’s paired two-tailed t-test. A P value of 0.05 was regarded to be sta-tistically significant.

4-3. Results and discussion

We first examined the modification behavior of sim-ple phenolic compounds with GSH in the presence orabsence of tyrosinase. We used 12 simple phenoliccompounds including naturally-occurring PCA,luteolin, (+)-catechin, caffeic acid, (3,4-diOH-type),ferulic acid (3-OMe-4-OH type), p-coumaric acid (4-OH type) and gallic acid (3,4,5-triOH-type) as well asartificial pyrocatechol (o-diphenol), resorcinol (m-diphenol), hydroquinone (p-diphenol), its oxidizedderivative (benzoquinone) and BHT (1-OH-4-Me). Asshown in Fig. 15, the co-incubation of each o-diphenol(catechol)-type phenolic compound (50 mM) with GSH(100 mM) for 30 min resulted in the significant loss ofGSH in a tyrosinase-dependent manner, whereas nei-ther the monophenols nor O-methylated catechol (p-coumaric acid, ferulic acid and BHT) showed any ef-fect. A triphenol gallic acid also showed the tyrosinase-dependent effect. An unsubstituted o-diphenol, pyro-catechol, but not resorcinol (m-diphenol) andhydroquinone (p-diphenol), decreased the GSH con-centration in the presence of tyrosinase. Hydroquinoneand benzoquinone consumed GSH independently oftyrosinase. Hydroquinone might be converted into ben-zoquinone possibly by auto-oxidation, which is sup-ported by the previous report showing theelectrophilicity of tert-butylhydroquinone (Nakamuraet al. 2003a). Among the o-diphenols, caffeic acid andpyrocatechol showed both tyrosinase-dependent and -independent effects on the GSH consumption. Theseresults suggested that substrates for the diphenolaseactivity, but not for the monophenolase activity, mightparticipate in the tyrosinase-dependent GSH consump-tion under the tested conditions.

We used PCA as a model catechol compound to fur-ther investigate the modification mechanism, becausePCA has been reported to decrease the GSH level invivo in a tyrosinase-dependent manner (Nakamura etal. 2000a). The in vitro tyrosinase-dependent GSHconsumption by PCA was concentration- and time-de-pendent (data not shown). For instance, the incubationof 100 mM PCA with GSH in the presence of tyrosi-nase for 1 h resulted in an 80% loss of GSH (Fig. 16A).The a-tocopherol cotreatment significantly inhibitedthe tyrosinase-dependent modification of GSH, sug-gesting that, in addition to the electrophilic reaction ofoxidized catechol, the oxidation of thiols might con-tribute to the reactivity of the bioactivated PCA. Theproducts of PCA and GSH with tyrosinase are too com-plicated to identify by their NMR spectral data (datanot shown). Therefore, we used another phenol oxi-dase, laccase (E.C. 1.10.3.2), since it has overlappingsubstrate ranges to tyrosinase and no monohydroxylaseactivity (Thurston 1994). As shown in Fig. 16A, incu-bation of 100 mM PCA with GSH in the presence of

Fig. 15. GSH consumption by simple phenolic compounds.GSH (100 mM) was incubated with the test compounds (50mM) for 130 min in the presence (filled bars) or absence(open bars) of tyrosinase. The amount of residual GSH wasspectrophotometrically estimated using the commercial kitGSH-400. Reprinted from Bioscience, Biotechnology andBioscience , Nakamura et al . , Thiol modification bybioactivated polyphenols and its potential role in skin in-flammation, in press, Copyright (2014), with permissionfrom Japan Society for Bioscience, Biotechnology andAgrochemistry.



Fig. 16. Modification behavior of PCA to GSH and GAPDH.(A) Phenol oxidase-dependent GSH consumption by PCA.GSH (100 mM) was incubated with or without PCA (100mM) and a-tocopherol (1 mM) for 1 h in the presence tyro-sinase or laccase. The amount of residual GSH was spectro-photometrically estimated using the commercial kit GSH-400. (B) 1H NMR analysis of GSH-PCA adduct. A mixtureof PCA and PCA-GSH produced by laccase was dissolvedin D

2O and then analyzed by NMR spectrometry. (C) Detec-

tion of free sulfhydryls of GAPDH treated with PCA andlaccase. GAPDH (0.2 mg/mL) was incubated with PCA andlaccase for 1 h. The free sulfhydryl groups of GAPDH weredetected by SDS-PAGE/blotting with IAB staining. Re-printed from Bioscience, Biotechnology and Bioscience,Nakamura et al., Thiol modification by bioactivatedpolyphenols and its potential role in skin inflammation, inpress, Copyright (2014), with permission from Japan Soci-ety for Bioscience, Biotechnology and Agrochemistry.

laccase (30 unit) resulted in a 30% loss of GSH, andthe co-treatment with a-tocopherol showed little ef-fect. In addition, the GSH-PCA adduct formed by thediphenolase reaction was identified as 2-glutathionyl-PCA by its NMR spectral data (Fig. 16B). Theelectrophilic potential and position of S-glutathionylation in PCA is also supported by a previ-ous report (Saito and Kawabata 2004).

We next examined the potential reactivity of thebioactivated PCA toward proteins using GAPDH con-taining four thiol groups per subunit. The GAPDHtreated with both PCA and laccase decreased a posi-tive band of approximately 36 kDa in a PCA concen-tration-dependent manner when assayed by SDS-PAGE/blotting with free sulfhydryl-positive IAB stain-ing (Fig. 16C), suggesting that the bioactivated PCAcan modify the free thiol groups in GAPDH. Highermolecular weight products than the original GAPDHwere also observed (data not shown), implying theintermolecular crosslinking of GAPDH. We initiallyobserved that PCA inhibited the GAPDH activity in alaccase-dependent manner (data not shown). Takentogether, these results suggested that the bioactivatedcatechol compounds have a potential to modify pro-tein sulfhydryl groups as well as low molecular weightthiols such as GSH.

Because PCA is capable of reducing the GSH leveland thus enhancing TPA-induced inflammation inmouse skin (Nakamura et al. 2000d), we next exam-ined the effects of (+)-catechin and BHT, the formerof which tyrosinase-dependently consumed GSH andthe latter was inactive, on two TPA-induced inflam-matory biomarkers, including edema formation andMPO activity. As shown in Table 10, the applicationof 20 mmol PCA 3 h before the TPA treatment showeda significant enhancement of edema formation andMPO activity compared to the TPA alone-treated con-trol (163% of control, P < 0.05 and 238% of control P< 0.05, respectively). Similarly, the application of notonly (+)-catechin but also tyrosinase reaction-negativeBHT showed enhancing effects on two inflammatorybiomarkers (edema, 141%, P < 0.05 and 136%, P =0.07, respectively; MPO, 179%, P = 0.07 and 160%, P< 0.05, respectively). We next assessed whether theenzyme-dependent bioactivation is involved in the in-verse effects using a tyrosinase inhibitor ArB and cy-tochrome P450 (CYP) inhibitor SKF252a. The tyrosi-nase inhibitor significantly counteracted the PCA-en-hanced inflammation as previously reported (Nakamuraet al. 2000d, 2001a). Although ArB showed the ten-dency to decrease the (+)-catechin-induced enhance-ment of edema formation (124% versus 141%, P =0.11), it showed no effect on the BHT-enhanced in-flammation. On the other hand, SKF252a did not showany effect on the PCA- and (+)-catechin-induced en-hancement, whereas it showed a tendency to inhibitthe BHT-enhanced edema formation (111% versus



Group Pretreatment Stimuli Inflammatory parameter

Edema formationa MPO activitya

(mg/punch, ratiob) (unit/punch, ratiob)

1 acetone acetone 25.6 ± 1.6 æ 0.6 ± 0.3 æ2 acetone TPA 58.6 ± 10.4 100 3.9 ± 0.9 100

3 PCAc TPA 95.5 ± 15.4d 163 9.2 ± 2.6d 238

4 PCA + ArBc TPA 71.3 ± 14.6e 122 4.7 ± 1.6e 122

5 PCA + SKFe TPA 81.4 ± 14.4d 139 8.0 ± 2.2d 207

6 (+)-catechin TPA 82.7 ± 12.6d 141 7.0 ± 2.8 179

7 (+)-catechin + ArB TPA 72.7 ± 14.2 124 5.4 ± 1.7 138

8 (+)-catechin + SKF TPA 79.7 ± 15.0d 136 7.5 ± 2.3d 194

9 BHT TPA 79.4 ± 20.2 136 6.2 ± 1.4d 160

10 BHT + ArB TPA 81.6 ± 13.1d 139 5.8 ± 1.2d 149

11 BHT + SKF TPA 64.7 ± 14.7 111 4.8 ± 1.0 124

Table 10. Modifying effects of PCA, (+)-catechin and BHT on TPA-induced inflammatory parameters in ICR mouse skin.Reprinted from Bioscience, Biotechnology and Bioscience, Nakamura et al., Thiol modification by bioactivated polyphenolsand its potential role in skin inflammation, in press, Copyright (2014), with permission from Japan Society for Bioscience,Biotechnology and Agrochemistry.

ICR mice were pretreated with the test compounds (20 mmol) with or without enzyme inhibitors (10 mmol) 3 h before TPA(8.1 nmol) treatment. After 24 h, edema formation and MPO activity were detemined as previously reported.7)

aValues are the mean ± SD of 5 mice.bValues are percentages of those in group 2 (TPA stimulation alone).cPCA, protocatechuic; ArB, arbutin; SKF, SKF525A.dSignificant versus group 2 (Student’s t-test); P < 0.05.eSignificant versus group 3 (Student’s t-test); P < 0.05.

136%, P = 0.08). These results suggested that higheramounts of the catechol-type polyphenols, including(+)-catechin, have the potential to enhance skin inflam-mation through a tyrosinase-dependent bioactivation,even though the catechol-type catechins have a lowerantioxidant activity and higher stability compared tothe gallate-type catechins, such as (–)-epigallocatechin-3-gallate (Mori et al. 2010). In addition, the toxic ef-fects of BHT, possibly through CYP, led us to the ideathat more attention should be paid to the administereddose of monophenols not only in the skin, but also inthe liver and kidneys, which have significant CYP ac-tivities and the potential target tissues for toxic effectsof PCA and some tyrosinase substrates (Nakamura etal. 2001b; Moridani et al. 2002).

In this section, we indicate that the catechol-typepolyphenols can modify sulfhydryl groups in a phenoloxidase-dependent manner. The possible involvementof polyphenol bioactivation in the enhancement ofTPA-induced skin inflammation was also suggested.Several phenolic toxicants, such as urushiols (poisonivy), cause type IV hypersensitivity through the hap-ten-protein adduct formation (Park et al. 1998). Thismechanism might be involved in the toxic effect ofcatechol-type polyphenols, because PCA has been re-ported to induce contact hypersensitivity in mouse skin(Nakamura 2001a). The occurrence of hapten-protein

adduct formation in mouse skin treated with (+)-cat-echin or BHT and the relationship to theimmunomodulatory effects remains to be elucidated.In addition, tyrosinase was a rate-limiting factor of theinhibition of the GSH-related enzyme activities by PCA(Nakamura et al. 2000d). In any case, the bioactivatedconversion of polyphenol to quinonoid(s) and/or con-comitantly formed ROS are likely to be the critical stepin the enhanced inflammation in mouse skin. Since thepro-oxidant or electrophilic action of polyphenols isalso important for their health-promoting property(Nakamura and Miyoshi 2010), future studies will beneeded to determine the optimal doses of polyphenolsto control their beneficial/harmful balance.

5. Toxic dose of a simple phenolic antioxidantattenuates the glutathione level in mouse liverand kidney

5-1. Introduction

Food phytochemicals showing antioxidant activitieshave mainly been noticed as most promising candidatesfor chemopreventors against oxidative stress-relateddiseases including cancer, because they have beenfound to strongly inhibit oxidative reactions in vitroand in vivo. However, some antioxidants exert not only



Fig. 17. Toxic biotransformation of protocatechuic acid(PCA). Reprinted from Journal of Agricultural and FoodChemistry, 30(9), Nakamura et al., Toxic dose of a simplephenolic antioxidant, protocatechuic acid, attenuates the glu-tathione level in ICR mouse liver and kidney, 5674–5678,Copyright (2001), with permission from American Chemi-cal Society.

weak anti-tumor promoting activity but also carcino-genic activity in rodents when given at a high dose(Ito and Hirose 1989). For example, both artificial anti-oxidants, including 3-tert-butyl-4-hydroxyanisole, andnaturally occurring compounds including caffeic acidhave shown not only tumor promoting activity in ratforestomach carcinogenesis but also induction offorestomach squamous cell carcinoma of rats (Ito andHirose 1989). Although dramatic pharmacological andbiological activities of naturally occurring antioxidantshave so far been much focused on, documentation oftheir safety and toxicology has been limited except forinitial studies on antioxidative vitamins (Meyers et al.1996).

PCA is one of the major benzoic acid derivatives fromedible plants and fruits. The study on absorption andmetabolism of cyanidin glucoside (CyG) demonstratedthat PCA, possibly derived from degradation of CyG(Tsuda et al. 1996), is actually present in the plasmaof the CyG-fed rat and might contribute to the in vivoantioxidative activity of CyG (Tsuda et al. 1999). Sec-tions 2 and 3 demonstrate the significant enhancementof mouse skin tumor promotion, inflammation andoxidative stress by topical pretreatment with a highdose of PCA, while a relatively lower dose attenuatedthese responses. The application of PCA alone in a largeamount also showed inflammatory responses includ-ing contact hypersensitivity. The possibility that me-tabolism by dermal tyrosinase activity of PCA tocompound(s) lacking antioxidative properties and/orrather possessing the potential to enhance tumor de-velopment has also been suggested.

GSH, the major cellular antioxidant, is well knownto have diverse biological functions including protec-tion of cells from damage by substances such as reac-tive oxygen species, free radicals, and reactiveelectrophiles including a,b-unsaturated carbonyl com-pounds. Sections 2 and 3 demonstrate that treatmentwith a high dose of PCA alone for 3 h enhancedoxidative stress by reducing glutathione levels in mouseskin, which was counteracted by a tyrosinase inhibi-

tor, ArB. These results strongly suggested that the ty-rosinase-derived reactive quinone intermediates ofPCA, which react with nucleophilic residues of pro-teins including sulfhydryl groups or GSH, were in-volved in dermatotoxicity (Fig. 17). GSH is, therefore,regarded as a protective regulator of PCA-inducedoxidative damage.

In this section, we examined the modifying effect oftoxic doses of PCA on GSH levels in ICR mouse liverand kidney at acute and subchronic phases. We alsoshowed using GSH-depleted mice that GSH plays anegatively regulating role for acute hepatotoxicity.


5-2A. ChemicalsButhionine sulfoximine (BSO) was obtained from

Aldrich Chemical, Co., Ltd. All other chemicals werepurchased as mentioned above.

5-2B. Treatment of animalsFemale ICR mice used in each experiment were sup-

plied with fresh tap water ad libitum and rodent pel-lets (MF, Oriental Yeast Co., Kyoto, Japan) freshlychanged twice a week. One group was composed of 5female ICR mice. In the acute toxicity experiments,the mice were dosed with PCA or acetaminophen(AAP) 50 mg or 500 mg/kg, i.p., in 33% polyethyleneglycol in saline, 37∞C. BSO was given as an i.p. dose

Fig. 18. Effects of AAP and PCA on total GSH level in ICRmouse liver (A) and kidney (B). The mice (5 mice in eachgroup) were treated with AAP (500 mg/kg) or PCA (50 or500 mg/kg). They were sacrificed 6 h after PCA applica-tion, and tissue samples were removed for GSH assays. Sig-nificance is expressed as: *, versus control group, P < 0.05.Reprinted from Journal of Agricultural and Food Chemis-try, 30(9), Nakamura et al., Toxic dose of a simple phenolicantioxidant, protocatechuic acid, attenuates the glutathionelevel in ICR mouse liver and kidney, 5674–5678, Copyright(2001), with permission from American Chemical Society.



Treatment Hepatotoxicity Nephrotoxicity

Plasmatic ALT Plasmatic AST Plasmatic urea Urinary protein(IU/l) (IU/l) (mg/ml) (mg/ml)

Control 22.1 ± 7.9 41.0 ± 9.8 0.20 ± 0.02 1.96 ± 1.02AAP (500 mg/kg) 42.0 ± 9.5* 129.8 ± 36.3* 0.26 ± 0.10 1.68 ± 0.61PCA (50 mg/kg) 33.3 ± 9.0 66.3 ± 6.3* 0.21 ± 0.03 2.07 ± 1.29PCA (500 mg/kg) 36.0 ± 12.4 76.8 ± 29.0* 0.35 ± 0.10* 3.09 ± 0.87

Table 11. Comparative acute hepato- and nephrotoxicity of AAP and PCA in ICR mouse liver and kidney. Reprinted fromJournal of Agricultural and Food Chemistry, 30(9), Nakamura et al., Toxic dose of a simple phenolic antioxidant, protocate-chuic acid, attenuates the glutathione level in ICR mouse liver and kidney, 5674–5678, Copyright (2001), with permissionfrom American Chemical Society.

ICR mice were treated with AAP or PA (i.p. administration). The mice were sacrificed 6 h after AAP or PA application. Thetoxic parameters were determined as mentioned in Section “Materials and Methods”.*Statistically different from control; P < 0.05.

of 800 mg/kg in 0.9% NaCl 2 h before PCA adminis-tration. The mice were sacrificed 6 h after AAP or PCAapplication. In the subchronic toxicity experiment, themice were given drinking water containing PCA orAAP (0.01% or 0.1%) for 60 days. Immediately aftercollecting urine and blood by the methods reportedpreviously (Peters et al. 1996), the mice were sacri-ficed and the livers and kidneys were removed. Bloodwas allowed to coagulate at room temperature, and thesamples were centrifuged to obtain the serum. The liv-ers and kidneys were weighed and homogenized in ice-cold PBS(–) (pH 7.4).

5-2C. Determination of toxic parametersThe GSH content in each tissue was measured spec-

trophotometrically using a commercial kit(BIOXYTECH® GSH-400™ Assay) as mentionedabove. The activities of alanine aminotransaminase(ALT) and aspartate aminotransaminase (AST) weremeasured by a GPT-test kit and by a GOT-test kit (WakoPure Chemical Industries), respectively. The plasmaticurea level was quantified by a uric nitrogen B-test kit(Wako Pure Chemical Industries). These experimentswere performed using the protocol outlined by themanufacturers. The urinary protein level was deter-mined with the BCA protein assay reagent (Pierce) withbovine serum albumin as the standard.

5-2D. Statistical analysisThe data were analyzed with an analysis of variance

(ANOVA) when necessary, followed by Fisher’s test.Specific differences among treatments were examinedusing the Student’s t-test (two sided), that assumedunequal variance.

5-3. Results and discussion

AAP, a well-known analgesic and antipyretic drug,

is regarded as an appropriate positive control for acutetoxicity in liver and kidney since a great number ofstudies of AAP toxicity have been documented. Asshown in Fig. 18, the i.p. administration of AAP andPCA (500 mg/kg) led to a significant and dose-depend-ent decline of the hepatic GSH level by approximately30% 6 h after administration. In contrast, the treatmentof PCA dramatically decreased the GSH level in kid-ney while AAP had little effect. The ALT and AST ac-tivities, representative hepatotoxic markers, alsotrended to increase by treatment with both agents (Ta-ble 11). AAP and PCA application at 500 mg/kg re-sulted in the enhancement of AST activity by 3.1-fold(P < 0.05) and by 1.9-fold (P < 0.05), respectively, ascompared with the control. Bilirubin, a degradationproduct of heme normally excreted in the bile fromliver, was detected in the urine of all PCA and AAP-treated mice but never in the urine of the control mice(data not shown) by a commercial detection kit, Pre-test 8a test paper (Wako Pure Chemical Industries). Theplasmatic urea level and urinary protein level alsoshowed a tendency to increase in mice given PCA muchmore than those treated with AAP. Especially the in-crease in the urea level induced by PCA administra-tion was significant (P < 0.05). Urinary glucose, a pu-tative marker of nephrotoxicity, was detected in PCA-treated mice but never in the urine of the control mice(data not shown). These results obviously indicated thepossibility of an excessive amount of PCA to show atoxic effect towards mouse liver and kidney.

In the subchronic administration experiment, all ani-mals remained healthy throughout the experimentalperiod. The body weight and drinking water consump-tion of mice did not significantly differ among thegroups (data not shown). The relative weights of liverand kidney (weight/100 g body weight) of all groupswere almost comparable (Table 12). The continuousadministration of PCA (0.1% in drinking water) for 60



ICR mice were treated with AAP or PCA (in drinking water). The mice were sacrificed 60 days after starting of AAP or PCAapplication. The toxic parameters were determined as mentioned in Section “Materials and Methods”.*Statistically different from control; P < 0.05.

Treatment Hepatotoxicity Nephrotoxicity

Relative liver weight Hepatic GSH Plasmatic ALT Relative kidey weight Nephrotic GSH(g/100 g body weight) (mmol/g tissue) (IU/l) (g/100 g body weight) (mmol/g tissue)

Control 4.6 ± 0.3 63.1 ± 4.5 5.6 ± 1.2 1.1 ± 0.1 13.0 ± 1.0AAP (0.01%) 4.3 ± 0.1 55.0 ± 4.5* 8.0 ± 1.3* 1.1 ± 0.1 12.0 ± 1.2AAP (0.1%) 4.4 ± 0.5 66.5 ± 3.6 7.0 ± 1.1 1.2 ± 0.1 14.2 ± 1.5PCA (0.01%) 4.2 ± 0.2 62.5 ± 4.6 7.0 ± 1.3 1.1 ± 0.1 11.6 ± 1.5PCA (0.1%) 4.6 ± 0.4 58.3 ± 3.3 8.0 ± 1.4* 1.2 ± 0.1 10.5 ± 0.8*

Table 12. Comparative subchronic hepato- and nephrotoxicity of AAP and PCA in ICR mouse liver and kidney. Reprintedfrom Journal of Agricultural and Food Chemistry, 30(9), Nakamura et al., Toxic dose of a simple phenolic antioxidant,protocatechuic acid, attenuates the glutathione level in ICR mouse liver and kidney, 5674–5678, Copyright (2001), withpermission from American Chemical Society.

days resulted in slight but significant enhancement ofALT activity in plasma and decrease in GSH level inkidney (Table 12) . Although the nephroticthiobarbituric acid-reacting substance level, an over-all oxidative stress marker, was slightly enhanced byrelatively high doses of PCA, no significant changesin other parameters concerning hepatotoxicity and ne-phrotoxicity were observed (data not shown). GSHdepletion is known to be a marker of the presence of athiol reactive chemical species. It is, therefore, withinthe range of possibility that significant attenuation ofthe nephrotic level of GSH by a subchronic adminis-tration of high doses of PCA may disturb the detoxifi-cation of other electophilic toxicants including ultimatecarcinogens.

Because PCA-induced hepatotoxicity is correlatedwith the depletion of hepatic GSH levels, we exam-ined the negatively regulating role of GSH againsthepatotoxicity using GSH-depleted mice exposed toan inhibitor of GSH synthesis, BSO. The hepatic con-centration of GSH was significantly decreased to ap-proximately 50% of the control values 2 h after BSO(800 mg/kg) administration (data not shown). PCA (500mg/kg) administered 2 h after BSO induced the en-hancement of ALT activity more severely than that inmice administered PCA alone (Fig. 19). Although sta-tistical analysis by two-way ANOVA revealed that thetreatment of PCA, but not BSO, significantly affectALT activity (P < 0.05, Fisher’s test), interaction be-tween the effects of PCA and BSO was not observed(P = 0.23). Nephrotoxic parameters, including the con-centration of urea in plasma, were not significantlychanged (data not shown).

AAP, a widely used analgesic and antipyretic, is asafe and effective drug at therapeutic dose. AAP as aprotective agent in cancer has also been described(Yamamoto et al. 1973; Williams and Iatropoulos

1997). However, overdoses of AAP can cause liver andkidney damage or even death (Mitchell et al. 1973;Maruyama and Williams 1988). This study clearly dem-onstrates that an overdose of a naturally occurring anti-oxidant, PCA, could cause temporary damage to liverand kidneys. Early studies on AAP hepatotoxicity in-dicated that liver cell injury is caused by its metabolite,N-acetyl-p-benzoquinoneimine (ABI) formed in a cy-tochrome P450-dependent reaction (Dahlin et al. 1984).At normal therapeutic doses, the major metabolic path-ways of AAP are glucuronidation and sulfation. Withan overdose of AAP, however, these pathways are satu-rated and the production of toxic ABI is increased, lead-ing to rapid depletion of the GSH level. Subsequently,ABI reacts with cellular macromolecules, causinghepatotoxicity (Roberts et al. 1991). Although the ex-act mechanism of ABI-induced hepatotoxicity has notbeen fully understood, the production of ROS by re-dox cycling, leading to protein oxidation, and the cova-lent attachment with protein are suggested to be po-tential mechanisms (Gibson et al. 1996). These indi-cations led us to propose a hypothesis that a plausibletoxic metabolite of PCA, covalently binding nucle-ophilic residues of proteins including sulfhydryl groupsor GSH, might be involved in the molecular mecha-nism of PCA-induced hepatotoxicity. The data that theGSH-deficient mice were more sensitive to PCA-in-duced acute damage in the liver than the control mice(Fig. 19) may also support this hypothesis.

In the kidney, the administration of PCA, but notAAP, significantly reduced the GSH level (Fig. 18).This suggested that a different mechanism might oc-cur for PCA and AAP-induced nephrotoxicity; PCA orits metabolite(s) could additionally reduce the level ofnephrotic GSH while AAP or its counterpart(s) showedlittle effect. The GSH conjugates with redox activecompounds such as tert-butylhydroquinone are known



Fig. 19. Effect of a inhibitor of GSH synthesis, BSO, onplasmatic ALT activity. BSO was received an i.p. dose of800 mg/kg in 0.9% NaCl 2 h before PA administration. Themice were sacrificed 6 h after PCA application. The ALTactivity was measured by using GPT-test kit (Wako PureChemical Industries). Significance is expressed as: *, ver-sus PA treatment group, P < 0.05. Reprinted from Journal ofAgricultural and Food Chemistry, 30(9), Nakamura et al.,Toxic dose of a simple phenolic antioxidant, protocatechuicacid, attenuates the glutathione level in ICR mouse liver andkidney, 5674–5678, Copyright (2001), with permission fromAmerican Chemical Society.

to be potent and selective nephrotoxicants (Peters etal. 1996). Therefore, GSH can be regarded as a carrierof redox-active compounds transferring to the kidney,an organ rich in g-glutamyl transpeptidase (g-GT). Thenephrotoxicity of GSH conjugates of these compoundsis dependent on the relatively high activity of g-GTwithin the brush border membrane of renal proximaltubular epithelial cells. Further metabolism of the con-jugates by g-GT has been found to be a prerequisitefor toxicity (Monks et al. 1988). The products of thereaction by g-GT, the conjugates with cysteinylglycinedipeptide, are substrates for dipeptidases, which simi-larly exist in the brush border membrane of renal proxi-mal tubular epithelial cells. The corresponding cysteineconjugates are then transported across the brush bor-der membrane via an amino acid transport system.Since metabolism of the phenolic compounds-GSHconjugates by g-GT is coupled to cellular uptake, theactivity of g-GT is thus considered to be necessary forthe accumulation of the conjugates into renal cells andalso perhaps for the activation of the conjugates byfacilitating oxidation (Monks et al. 1994). It is thuswithin the range of possibility that GSH-PCAconjugate(s) may be one of the active forms that causesnephrotoxicity. The neurotoxic cysteine conjugate ofdopamine, having a catechol moiety, is further metabo-lized to a benzothiazine derivative, which is also re-dox-active and can react with GSH (Shen and Dryhurst1996). We speculate that further metabolism of PCA-GSH conjugates in the kidney may occur and its

metabolites can thus be potential nephrotoxicants. Fur-ther studies on the identification and toxic effects ofnot only GSH conjugates but also their metabolites arecurrently in progress.

In conclusion, possible toxic effects of phenolic anti-oxidant administration on liver and kidney were ob-served. Previous study demonstrated that administra-tion of PCA at less than 1/1000 of toxic dose showedthe most effective cancer chemopreventive activity inmouse skin (Nakamura et al. 2000d). Although quan-tification of PCA amount in daily-consumed vegeta-bles has not been fully evaluated, available papers re-ported that some kind of vegetables or wine are sig-nificant sources of PCA at concentrations of 2–10 mg/g (Li et al. 1993; Satué-Gracia et al. 1999). It is, there-fore, quite difficult to achieve PCA ingestion by dailyfood intake even at a dose of 1 mg/kg body weight. Onthe other hand, much focus should be paid on the safetyof antioxidants administered excessively, because tab-lets or pills containing antioxidative vitamins or plantpolyphenols, extracted and condensed from vegetablesand fruits, are so far commonly available. Therefore,further extensive studies at the molecular level on thebioactivation/detoxification metabolizing mechanismof antioxidants are essential to provide supporting in-formation.

6. Pivotal role of electrophilicity in glutathioneS - transferase induc tion by ter t -butylhydroquinone

6-1. Introduction

Phase II enzymes such as NAD(P)H:(quinone-accep-tor) oxidoreductase (NQO) and GST play a role in thecellular detoxification of genotoxic and carcinogenicchemicals. The GSTs are a family of enzymes thatcatalyze the nucleophilic addition of the thiol group ofGSH to a variety of electrophiles (for review, see Heyesand Pulford 1995). It is generally accepted that theGSTs are encoded by at least five different gene fami-lies. Four of the gene families (Class a, m, p, and q)encode the cytosolic GSTs, whereas the fifth encodesa microsomal form of the enzyme. Recently, twotransgenic rodent studies clearly demonstrated that theClass pi GST (GSTP1) can profoundly alter the sus-ceptibility to chemical carcinogenesis in mouse skin(Henderson et al. 1998) and rat liver (Nakae et al.1998). The Class pi rat and human GST isozymes havebeen shown to be highly efficient in the GSH conjuga-tion of carcinogenic benz[a]pyrene derivatives(Robertson et al. 1986a, b), widespread environmen-tal pollutants in cigarette smoke and automobile ex-haust. In addition, GSTP1-1 is more effective in thedetoxification of electrophilic a,b-unsaturated carbo-nyl compounds produced by radical reactions, lipidperoxidation, ionizing radiation, and the metabolism



of drugs than other GSTs (Berhane et al. 1994). Thus,the induction of GSTP1-1 is regarded as one of theimportant determinants in the cancer chemoprotectionpotential of food materials, phytochemicals and syn-thetic chemicals.

An antioxidant/electrophile response element (ARE/EpRE; consensus sequence TGACNNNGC) or the re-lated element, regulating both its basal and inducibleexpressions, was mostly found in the 5¢-flanking re-gion of the genes of phase II enzymes and may be rec-ognized by a similar series of transcriptional factors(Primiano et al. 1998). Recently, we developed thecultured hepatocyte cell line RL34 and determined theGST induction potencies of edible plants, structuresand their molecular mechanism of several series ofcompounds (Kawamoto et al. 2000; Nakamura et al.2000a, 2000c, 2002a; Morimitsu et al. 2002). Usingthis cell line, we demonstrated for the first time thatGSTP1 enhancer I (GPEI; ARE/EpRE of GSTP1), con-taining a palindromic dyad of the TPA responsible el-ement (TRE)-like sequence (Sakai et al. 1988), is anessential cis-element required for the activation of theGSTP1 gene through the redox alteration byelectrophiles such as benzyl isothiocyanate or diethylmaleate (DEM) (Nakamura et al. 2000c). The findingsthat chemical agents having diverse structures indicatedthat various chemicals may produce a common signaltransduction responsible for AP-1 or its relatedfactor(s). It has been observed that most chemical in-ducing agents of the phase II enzymes have the poten-tial to induce oxidative stress and depletion of GSH(Daniel 1993; Nakamura et al. 2000c).

Phenolic compounds, termed phenolic antioxidantsdue to their chain-breaking reaction duringautooxidation of lipids, are utilized for food preserva-tion and the suppression of lipid peroxidation in bio-logical materials. Previous studies, based on the as-sumption that these antioxidative properties are impor-tant for the inducible effect on the AP-1-mediated geneexpression including phase II enzymes, have concludedthat the AP-1 binding site is an antioxidant responseelement. Concerning the molecular basis of cancer pro-tection by tert-butylhydroquinone (tBHQ), De Longet al. also concluded in their studies of the inductionof phase II enzymes that oxidative lability was essen-tial for inducer activity since catechol (1,2-diphenols)and hydroquinone (1,4-diphenols) derivatives undergofacile oxidation to quinones, whereas 1,3-diphenols,inactive for phase II induction, cannot participate insuch an oxidation (1986). Although these experimentsdid not establish whether the oxidation products or oxi-dation processes (potentially involving radical scav-enging reaction, multiple one- and two-electron oxidereduction and redox-dependent reactive oxygen spe-cies (ROS) generation) were an inductive signal,electrophilic quinone oxidation products were pre-sumed to be the ultimate inducers since electrophiles

including Michael reaction acceptors (e.g., olefins con-jugated to electron-withdrawing groups) andisothiocyanates potentially induce the phase II enzymeexpression (Prestera et al. 1993). Conversely, Pinkuset al. demonstrated that the auto-oxidation of tBHQ tothe semiquinone radical or 1,4-benzoquinone (BQ) andthe generation of hydroxyl radical were detected us-ing the electron spin resonance spectroscopy technique(1996). They also showed that the induction of an en-dogenous GST a class gene (rGSTA1) in hepatoma cellsby tBHQ was inhibited by antioxidants N-acetylcysteine, GSH, and exogenous catalase. It wasthus expected that the intermediate formation of H2O2during the metabolism of tBHQ might be a critical stepfor the phase II enzyme induction. However, the dose-dependency of the hydroxyl radical generation waspoorly correlated with the GST gene expression; as theconcentration of tBHQ increased, a lower amount ofhydroxyl radical was generated (Pinkus et al. 1996).More recently, Lee et al. pointed out that the ROS-generating property could be ruled out in the mecha-nism for the ARE-dependent gene expression by tBHQ(2001). Anyway, phenolic antioxidants express antago-nistic signals, oxidative stress and antioxidative reac-tions in the cells.

tBHQ is a major metabolite in vivo in dogs, rats, andman of 3-tert-butyl-hydroxyanisole (BHA) (Astill etal. 1962; El-Rashidy and Niazi 1983; Verhagen et al.1989), a synthetic phenolic antioxidant frequently usedas a food additive. BHA protects animals against vari-ous carcinogens, presumably through the induction ofmany phase II detoxifying enzymes as well as the in-hibition of cytochrome P450 monooxygenase. On theother hand, mounting evidence has indicated that BHAcan be a carcinogen or tumor promoter in some tissuein animals. Although opposing biological effects ofBHA on carcinogenesis have been well documented,the precise mechanisms of these effects remain obscure.tBHQ has also been reported to be a double-edgedsword in cancer control, possibly through phase II en-zyme induction including GST activity and through thegeneration of ROS by cytochrome P450/P450 oxidore-ductase- or transition metal-mediated redox cycling,respectively. Thus, the metabolic formation of tBHQis thought to, at least in part, contribute to the modify-ing effects of BHA on carcinogenesis.

In the present study, we further investigated the roleof the redox cycling reaction or electrophilic propertyof tBHQ in the induction of GST. GST activity, intrac-ellular oxidative stress and GSH consumption weremonitored in intact RL34 cells or in the cell-free sys-tem applied to tBHQ and 2,5-di-tert-butylhydroquinone(DtBHQ), derived from one of the concomitants ofcommercial BHA, 2,5-di-tert-butylhydroxyanisole(Fig. 20, Lam and Garg 1991), with similar potentialsto construct redox cycling and different capacities toreact with nucleophiles such as GSH. This study dem-



Fig. 20. Chemical and stereoscopic structures of tBHQ andDtBHQ. Reprinted from Biochemistry, 42(14), Nakamura etal., Pivotal role of electrophilicity in glutathione S-trans-ferase induction by tert-butylhydroquinone, 4300–4309,Copyright (2003), with permission from American Chemi-cal Society.

onstrated for the first time that tBHQ is a potential in-ducer of pi class GSTP1-1 isozyme, an important mol-ecule that protects against carcinogens as describedabove, and that GPEI is the responsible element in theGSTP1 induction by tBHQ. We report here that DtBHQ,hardly reacting with GSH probably due to steric hin-drance of the bulky tert-butyl moieties, has much lessability for GST induction. Thus, we conclude that thereaction with intracellular nucleophile including pro-tein thiol or GSH is virtually responsible for the in-duction of GSTP1-1 and/or other phase II enzymes bytBHQ.


6-2A. ChemicalsAll chemicals were purchased from Wako Pure

Chemical Industries, Osaka, Japan. Anti-rat GSTP1-1anti-serum was obtained from Biotrin International,Dublin, Ireland. Horseradish peroxidase-linked anti-rabbit IgG immunoglobulin was purchased from Dako,Glostrup, Denmark. 2¢,7¢-Dichlorofluorescin diacetate(H2DCF-DA) was obtained from Molecular Probes,Inc., Leiden, The Netherlands. Authentic tBHQ-GSHconjugates were synthesized as previously reported(van Ommen et al. 1992).

6-2B. Cell culturesRL34 cells were obtained from the Health Science

Research Resources Bank, Osaka, Japan (Yamada et

al. 1987). The cells were grown as monolayer culturesin Dulbecco’s modified Eagle’s medium supplementedwith 5% heat-inactivated fetal bovine serum, penicil-lin (100 U/ml), streptomycin (100 mg/ml), L-glutamine(0.3 mg/ml), pyruvic acid (0.11 mg/ml) and 0.37%NaHCO3 at 37∞C in an atmosphere of 95% air and 5%CO2. Cells post-confluency were exposed to the testcompounds in a medium containing 5% fetal bovineserum.

6-2C. Radical scavenging activity assayThe 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical

scavenging activity was evaluated as previously re-ported (Nakamura et al. 2003b). A test compoundmixed with a 100 mM Tris-HCl buffer (pH 7.4, 1 ml)was added to 0.5 mM DPPH in ethanol (1 ml), and themixture was shaken vigorously and left to stand for 20min at room temperature in the dark. The DPPH radi-cal scavenging activity is expressed as the ratio of therelative decrease in the absorbance of the test samplemixture at 517 nm to that of the 1 mM Trolox solution:DPPH radical scavenging activity (%) = {(vehicle) –(test compound)}/{(vehicle) – (Trolox)} ¥ 100.

6-2D. Intracellular peroxide determinationIntracellular peroxides were detected by H2DCF-DA

as an intracellular fluorescence probe (Nakamura et al.2000c, 2002b). Briefly, the cells under confluency werepreincubated with tBHQ and DtBHQ (50 mM) for 1 h.After stimulation by DEM (0.5 mM) for 15 min andwashing by PBS, the cells were treated with H2DCF-DA (50 mM) for 30 min at 37∞C. A flowcytometer(CytoACE 150, JASCO, Tokyo, Japan) was used todetect dichlorofluorescein (DCF) formed by the reac-tion of H2DCF with intracellular oxidative products.Experiments were repeated four times with similar re-sults. The data are expressed as one representative his-togram.

6-2E. GSH assay and NMR experimentSee Section 4.

6-2F. Measurement of O2– generation

The quantity of O2– generated by the reaction of

tBHQ or DtBHQ with Cu2+ was determined by cyto-chrome c reduction as previously reported (Hirakawaet al. 2002). The reaction mixture, containing 50 mMferricytochrome c, 20 mM tBHQ or DtBHQ, 20 mMCu2+, and 5 mM DTPA in 1.2 ml of 100 mM sodiumphosphate buffer (pH 7.4) with or without superoxidedismutase (SOD, 150 units/ml), was incubated at 37∞C.We recorded the absorption at 550 nm spectrophoto-metrically.

6-2G. HPLC-ECD analysisA HPLC system equipped with a PU-980 HPLC

pump, 807-IT Integrator, and both UV-970 UV and 840-



EC electrochemical detectors were used. Elution pro-files were monitored at 210 nm on the UV detectorand at 0.5 V of the applied oxidation potential on theelectrochemical detector. A reverse-phase column(NOMURA CHEMICAL, Aichi, Japan; DevelosilODS-HG-5; 25 ¥ 0.8 cm) was used through this study.The tBHQ-GSH conjugates formed intracellularly weremeasured by this system. Confluent monolayer cellswere exposed to 10 mM tBHQ and, after the variousincubation periods, cell monolayers were washed twicePBS (pH 7.0) and extracted with the 5% trichloroace-tic acid solution containing 5 mM EDTA followed bycentrifugation (10,000 g, 20 min). The column wasequilibrated and the cell extracts were eluted with 55%MeOH/water containing 0.1% formic acid at flow rateof 0.8 ml/min. The identification of the conjugates wasmade on the basis of the retention time of authenticsamples as well as on the co-elution test performed byadding standard conjugates to cell samples.

6-2H. Enzyme assayGST activity was measured using 1-chloro-2,4-

dinitrobenzene (CDNB) and ethacrynic acid (EA) assubstrates according to the method of Habig and Jakoby(1981).

6-2I. Western blot analysisFor GST, the tBHQ derivatives-treated and untreated

cells were rinsed twice with PBS (pH 7.0) and lysedby incubation at 37∞C for 10 min with a solution con-taining 0.8% digitonin and 2 mM EDTA (pH 7.8). Eachwhole cell lysate was then treated with Laemmli sam-ple buffer for 3 min at 100∞C (Laemmuli 1970). Thesamples (20 mg) were run on 12.5% SDS-PAGE slabgel. One gel was used for staining with Coomassie bril-liant blue and the other was transblotted on a nitrocel-lulose membrane with a semi-dry blotting cell (Trans-Blot SD, Bio-Rad), incubated with Block Ace (40 mg/ml) for blocking, washed, and treated with the anti-body.

6-2J. Plasmid constructionChloramphenicol acetyltransferase (CAT) fusion

plasmids were constructed by the method of Sakai etal. (1988) and kindly given by Prof. M. Imagawa ofNagoya City University. A 3.0 kb fragment between –2.9 kb and +59 bp of the gene for GSTP1-1 (Sugiokaet al. 1985; Okuda et al. 1987) was inserted into theHindIII site of pSV0CAT (Ohno et al. 1988) and des-ignated ECAT. A series of 5¢ deletion mutants was con-structed from the ECAT using appropriate restrictionenzymes (Sakai et al. 1988).

6-2K. DNA transfection for analysis of CAT activityRL34 cells were transfected with 5 mg of plasmid

construct by a calcium phosphate co-precipitation pro-cedure described by Chen and Okayama (1987). The

test compound was added to the culture medium 48 hafter transfection. Cell lysates were obtained afterfreeze-thawing three times in 0.25 M TrisHCl (pH 7.4)and used for CAT assay. For CAT assay (Gorman et al.1982), the extracts were heated at 65∞C for 10 min andthe precipitates were removed by centrifugation at15000 g for 10 min at 4∞C. The reaction mixtures (fi-nal volume 125 ml) containing the cell extract, 210 mMTrisHCl (pH 7.8), 11 kbp of 1-deoxy[14C]chloramphenicol and 80 nmol of acetyl-CoA were incubated at 37∞C for 90 min. Reactions wereterminated by the addition of 1 ml of ethyl acetate. Theproduct (3-acetyl-1-deoxychloramphenicol) and theunreacted substrate were extracted with ethyl acetate.After ethyl acetate was evaporated, the residue wasdisolved in 20 ml of ethyl acetate and chromatographedon a thin-layer plate with chloroform/methanol (95:5,v/v). The radioactivity of the product and substrate wasanalyzed using the Fuji BAS 2000 system (Fuji Photo,Tokyo, Japan).

6-3. Results

6-3A. Redox cycling potentials of tBHQ and DtBHQQuinones shuffle electrons enzymatically or nonen-

zymatically among their reduced form (hydroquinone),oxidized form (BQ), and/or their semiquinone radicalsto construct redox cycles. To investigate the redox proc-ess, we first examined the free radical-scavenging ac-tivity of tBHQ and DtBHQ because the scavenging ofa free radical is completed by the one-electron oxida-tion of the parent compound. As shown in Fig. 21A,the dose-dependent effect to scavenge a well-knownfree radical, DPPH, was observed when the DPPH radi-cal and DtBHQ was mixed in a solution. This activitywas relatively lesser than that of tBHQ but significant.Thus, it appears that DtBHQ as well as tBHQ can beoxidatively converted to the corresponding semiqui-none radical (one-electron oxidation product) or BQ(fully oxidized by two electrons).

Next, to determine whether tBHQ and DtBHQ canact as an antioxidant within cells, we investigated theinhibitory effect during the short exposure ofelectrophile-induced intracellular ROS generation us-ing a flowcytometry technique. As shown in Fig. 21B,the cells treated for 30 min only with the strong thiolblocker DEM exhibited a significant accumulation ofDCF due to the DEM-induced ROS production. Pre-treatment of tBHQ significantly blocked the DEM-in-duced enhancement of the intracellular ROS accumu-lation. DtBHQ also reduced the intracellular ROS pro-duction. Treatment only with tBHQ or DtBHQ did notaffect the basal level of accumulated DCF (data notshown). The intracellular antioxidant activity ofDtBHQ is weaker than that of tBHQ as expected bythe result in the DPPH assay. These results suggestedthat tBHQ and DtBHQ could be oxidized to the corre-



Fig. 21. tBHQ and DtBHQ act as antioxidants in the cell-free and cultured cell systems. A, Free radical scavengingactivity against DPPH radical of tBHQ (closed triangle), andDtBHQ (closed circle). A test compound mixed with a 100mM Tris-HCl buffer (pH 7.4, 1 ml) was added to 0.5 mMDPPH in ethanol (1 ml), and the mixture was shaken vigor-ously and left to stand for 20 min at room temperature in thedark. The DPPH radical scavenging activity is evaluated bythe absorbance at 517 nm. B, Inhibitory activity of tBHQand DtBHQ against DEM-induced intracellular ROS gen-eration in RL34 cells. The cells were pretreated with tBHQand DtBHQ (50 mM) for 1 h and then stimulated with 0.5mM DEM for 30 min. The cells were incubated with H

2DCF-

DA (50 mM) for 30 min to detect intracellular peroxide for-mation. The DCF fluorescence of more than 10,000 cellswas monitored on a flowcytometer. Reprinted from Biochem-istry, 42(14), Nakamura et al., Pivotal role of electrophilicityin glutathione S-transferase induction by tert-butylhydroquinone, 4300–4309, Copyright (2003), with per-mission from American Chemical Society.

sponding semiquinone radical or benzoquinone in bio-logical systems. Because the exact mechanism respon-sible for their antioxidant activity against intracellularROS accumulation is still not thoroughly understood,further studies of ROS molecule or free radical spe-cies contributed to DEM-induced oxidative stress arerequired.

To gain further evidence for the oxidative conver-

sion of tBHQ and DtBHQ to BQs, a 1H NMR analysisof tBHQ and DtBHQ upon incubation with Cu2+ wasperformed without purification of BQ. The oxidationof tBHQ and DtBHQ by Cu2+ was performed in thepresence of bathcuproine in order to remove Cu+, whichcatalyzes the reversed reductive reaction. The oxidizedproducts of tBHQ and DtBHQ by Cu2+ were extractedin chloroform-d to allow measurement of the 1H NMRspectra (Fig. 22A). When tBHQ and DtBHQ weretreated with Cu2+, the spectra assigned to tert-butyl-BQ and 2,5-di-tert-butyl-BQ, respectively, were ob-served. Although hydroquinone (1,4-dihydroxybenzene) is completely oxidized into BQ byequimolar quantities of Cu2+ within 1 min (Hirakawaet al. 2002), significant amounts of tBHQ and DtBHQremained in the same condition.

On the other hand, during the oxidation process, Cu+

is generated by the reduction of Cu2+ (Monks et al.1992). It is generally accepted that the Cu+ ion canproduce O2

– through its reaction with oxygen. Hence,to investigate whether O2

– is produced in the redoxcycle of tBHQ and DtBHQ, we measured the O2

– gen-eration using a cytochrome c reduction method. Weestimated the amount of O2

– generation from the dif-ference in the cytochrome c reduction with or withoutSOD. As shown in Fig. 22B, tBHQ and DtBHQ pro-duced a significant amount of O2

–. These results sug-gested that DtBHQ has a propensity for redox cyclingsimilar to tBHQ.

6-3B. Electrophilic properties of oxidized tBHQ andDtBHQ

It is suggested that the subsequent oxidation of tBHQand DtBHQ generates BQs, which are capable of re-acting with nucleophiles such as GSH. Hence, we in-cubated tBHQ and DtBHQ with GSH and evaluatedtheir electrophilicity by measuring the reactivity withGSH. As shown in Fig. 23A, the reduced GSH levelwas observed after the addition of tBHQ but notDtBHQ. This effect of tBHQ is dose-dependent andwas counteracted by the co-existence of 2-mercaptethanol (Fig. 23B). As the number of tert-butylmoieties increased, a lower amount of reduced GSHwas consumed. Similar results were obtained by analternative procedure using a reversed-phase HPLCwith an electrochemical detector (HPLC-ECD tech-nique), which can detect a free thiol group (data notshown).

We then examined the formation of the tBHQ-GSHadduct within the cells. To this end, RL34 cells in themonolayer culture were exposed to 10 mM tBHQ andthe tBHQ-GSH conjugate was analyzed by the HPLC-ECD technique. As shown in Fig. 23C, tBHQ was in-corporated very rapidly into the cells, reaching 120 nMat 30 min. The cellular concentration began to gradu-ally decrease thereafter. Accompanied by the incorpo-ration of tBHQ, two tBHQ-GSH adducts (Fig. 23D),



Fig. 23. Differences in electrophilic properties of tBHQ andDtBHQ. A, GSH consumption by tBHQ and DtBHQ. Reac-tions were performed with 100 mM tBHQ (closed triangle)or DtBHQ (closed circle) with 100 mM GSH in 1 ml of 100mM phosphate buffer (pH 7.4). The amount of residual GSHwas estimated spectrophotometrically using commercial kitGSH-400. B, Counteracting effect of 2-ME (1 mM) on GSHconjugates formation. C, Detection of intracellularly accu-mulated tBHQ and tBHQ-GSH conjugates by HPLC-ECD.D, Chemical structures of tBHQ-GSH conjugates. E, Quan-tification of intracellularly accumulated tBHQ. F, Quantifi-cation of intracellularly accumulated GSH-tBHQ conjugates.Reprinted from Biochemistry, 42(14), Nakamura et al., Piv-otal role of electrophilicity in glutathione S-transferase in-duction by tert-butylhydroquinone, 4300–4309, Copyright(2003), with permission from American Chemical Society.

Fig. 22. Redox reactions of tBHQ and DtBHQ. A, 1H NMRspectra of tBHQ (left) or DtBHQ (right) treated with or with-out Cu2+. An aqueous solution (0.75 ml) containing 27 mMDtBHQ (upper) or DtBHQ, 27 mM CuCl

2, and 27 mM

bathocuproine disulfonic acid (lower) was shaken vigorouslyfor 1 min. The organic compounds were extracted in chloro-form-d to measure the spectrum. The arrow points to a sig-nal assigned for oxidized DtBHQ (2,5-di- tert-butylbenzoquinone). B, Cytochrome c reduction by tBHQand DtBHQ with Cu2+. Reactions were performed with 10mM tBHQ or DtBHQ plus 10 mM CuCl

2 and 50 mM cyto-

chrome c with or without 150 units/ml SOD in 1.2 ml of 100mM phosphate buffer (pH 7.4) with DTPA. The amount ofO

2– genaration was estimated by subtracting the amount of

reduced cytochrome c with SOD from that without SOD.Reprinted from Biochemistry, 42(14), Nakamura et al., Piv-otal role of electrophilicity in glutathione S-transferase in-duction by tert-butylhydroquinone, 4300–4309, Copyright(2003), with permission from American Chemical Society.

tration), with a lower amount of 6-GS-tBHQ (0.5% ofintracellular tBHQ concentration), both of which werequickly formed within the cells (Fig. 23F). 3,6-Bis-glutathion-S-yl-tBHQ, detected as a metabolite oftBHQ in rat bile (Peters et al. 1996), was not detectedin RL34 cells. In total, the conjugated GSH accountedfor 3% of the tBHQ concentration. These data takentogether indicated that DtBHQ possesses a redox-ac-tive character but no detectable electrophilic abilitywhereas tBHQ has both properties.

5-glutathion-S-yl-tBHQ (5-GS-tBHQ) and 6-glutathion-S-yl-tBHQ (6-GS-tBHQ), were detected inthe tBHQ-treated RL 34 cells (Fig. 23E). Quantifica-tion of the GSH-conjugates showed that the major onewas 5-GS-tBHQ (2.5% of intracellular tBHQ concen-



Fig. 24. Induction of GST activity by tBHQ and DtBHQ inrat liver epithelial RL34 cells. A, Dose-dependent effect oftBHQ and DtBHQ on cellular GST activity. Cells post-confluency were exposed to the test compounds in the me-dium containing 5% fetal bovine serum. Cells were exposedto the test compounds for 24 h. The cellular GST activitywas evaluated according to the method of Habig et al. (1981).B, The effect of antioxidants on tBHQ-induced GST induc-tion. RL34 cells pretreated with vehicle (PBS), NAC (2.5mM) and catalase (100 unit/ml) for 2 h then treated withtBHQ (10 mM) for 24 h. Reprinted from Biochemistry,42(14), Nakamura et al., Pivotal role of electrophilicity inglutathione S-transferase induction by tert-butylhydroquinone, 4300–4309, Copyright (2003), with per-mission from American Chemical Society.

diphenols including tBHQ among the phenolic anti-oxidant derivatives (Kawamoto et al. 2000). These datawere consistent with that reported previously byProchaska et al. (1985) that the induction of NQO inmurine Hepa1c1c7 hepatoma cells depends upon theoxidation-reduction lability. To further investigatewhich properties, either redox cycling-active orelectrophilic, are involved in the GST induction byphenolic antioxidants, we examined the GST-inducingability of DtBHQ compared with tBHQ. It should benoted that the modifying effect of DtBHQ on the phaseII enzyme expression has not yet been determined, al-though DtBHQ, among the hydroquinone derivatives,is the only exception to protect against benzo[a]pyrene-induced neoplasia of the forestomach in ICR mice(Prochaska et al. 1985). Figure 24A shows the modi-fying effects of 10, 20, 50, and 100 mM concentrationsof tBHQ and DtBHQ in RL34 cells on the GST activi-ties toward EA, a specific substrate of Class pi GSTsas well as CDNB. The mean basal specific activitiesof the cytosolic GSTs in RL34 cells were 16.8 ± 2.1and 8.3 ± 0.7 (¥ 10–3) units per mg protein, respec-tively. tBHQ potently enhanced the activities towardCDNB in a dose-dependent manner. These results sug-gested that half of the GST induction by tBHQ wasaccounted for the induction of Class pi GST isozymes.On the other hand, the inducibility of DtBHQ, signifi-cantly generating ROS, was not efficient even at a con-centration of 100 mM. In addition, the tBHQ-inducedenhancement of Class p GST isozymes activity, mea-sured by EA conversion, was not counteracted by theantioxidant treatment (Fig. 24B). These results led usto the hypothesis that ROS produced by redox cyclingcan be ruled out in the major mechanism of the GSTP1induction by tBHQ.

6-3D. GPEI as the tBHQ response element in GSTP1expression

To further confirm whether tBHQ or DtBHQ influ-ences the expression of the pi Class GST protein, themain GST isozyme in RL34 cells, an immunoblotanalysis was carried out using the GSTP1-1-specificantibody. Figure 25A demonstrates a significant in-crease in the level of GSTP1-1 by the treatment withtBHQ, but not with DtBHQ at a concentration of 25mM. No inducing potency of DtBHQ to induce theGSTP1-1 protein was also observed up to a concentra-tion of 100 mM (data not shown). Next, to determinewhether tBHQ or DtBHQ stimulates the promoter ac-tivity of the GSTP1 gene, we examined the effect oftBHQ or DtBHQ on the transient expression of thebacterial CAT reporter gene harboring the 5¢-flankingregion (–2.9 kb to +59 bp) of the GSTP1 gene in RL34cells. As shown in Fig. 25B, ECAT containing the 5¢-flanking region showed a low but detectable CAT ac-tivity, and tBHQ strongly stimulated it by about 6-fold.On the other hand, the stimulating effect of DtBHQ on

6-3C. Effect of tBHQ and DtBHQ on GST induc-tion in RL34 cells

In a previous study, we found that the activity to-ward CDNB, a general GST substrate, in inducer-sen-sitive RL34 cells was most potently induced by 1,4-



Fig. 26. Determination of the DNA element required for theinduction GSTP1 expression by tBHQ. A, 5¢-deletion con-structs are schematically indicated. B, 5 mg of CAT constructswas transfected into RL34 cells and treated with 10 mMtBHQ. Reprinted from Biochemistry, 42(14), Nakamura etal., Pivotal role of electrophilicity in glutathione S-trans-ferase induction by tert-butylhydroquinone, 4300–4309,Copyright (2003), with permission from American Chemi-cal Society.

Fig. 25. tBHQ but not DtBHQ Induces of GSTP1-1 proteinexpression and transcription of an GSTP1 gene 5¢-flankingregion-CAT reporter gene. A, Effect of tBHQ and DtBHQon GSTP1 protein. Cells post-confluency were exposed totBHQ (10 mM) and DtBHQ (100 mM) in the medium con-taining 5% fetal bovine serum for 24 h. GSTP1-1 level wasexamined by an immunoblot analysis. B, Stimulation of pro-moter activity of GSTP1 by tBHQ and DtBHQ. The bacte-rial CAT reporter gene harboring the 5¢-flanking region (–2.9 kb to + 59 bp) of the GSTP1 gene was transfected intoRL34 cells and treated with 10 mM t-BHQ or 100 mMDtBHQ. Reprinted from Biochemistry, 42(14), Nakamura etal., Pivotal role of electrophilicity in glutathione S-trans-ferase induction by tert-butylhydroquinone, 4300–4309,Copyright (2003), with permission from American Chemi-cal Society.

the promoter activity was also detectable but muchweaker (2-fold) than that of tBHQ. This indicates thatthe region contains an element responsible for tBHQbut not DtBHQ. Because DtBHQ does not show anyreactivity with thiols but has the ability to generate ROSsimilar to tBHQ, DtBHQ slightly induces the GSTP1expression possibly through a redox cycling-depend-ent mechanism. On the contrary, an electrophilic prop-erty may play a major role in tBHQ-stimulated GSTP1induction.

We have previously reported that the 5¢-flanking re-gion of the GSTP1 gene contains an element responsi-ble for benzyl isothiocyanate as previously reported(Nakamura et al. 2000c). To identify the element re-sponsible for the tBHQ-stimulated promoter activity,we utilized a series of constructs including the 5¢-flank-ing region of the GSTP1 gene, namely ECAT and itsdeletion mutants (Fig. 26A). As shown in Fig. 26B,the 1CAT had completely lost the basal and tBHQ re-



sponsiveness, indicating that the tBHQ element of theGSTP1 gene is located between –2.9 to –2.2 kb up-stream from the transcription start site. Deletion to –1.4 kb (2CAT) and –0.8 kb (3CAT) showed only slightchanges in the CAT activity, while further deletion to–0.14 kb (4CAT) resulted in a significant recovery ofboth the basal and the tBHQ-induced CAT activities,suggesting that some element suppressing and/or en-hancing the transcription activity is also located be-tween –0.4 kb and –0.14 kb and/or between –0.14 kband +59 kb, respectively. It should be noted that a si-lencer element has previously been identified in theregion between –0.4 kb and –0.14 kb (Imagawa et al.1991). There is a strong enhancer element, GPEI, inthe restricted region between –2.9 and –2.2 kb (Okudaet al. 1989). A further experiment using the constructthat has the minimum GSTP1 promoter connected tothe CAT coding region indicated that tBHQ stimulatedthe CAT activity by about 2~2.5-fold in the cellstransfected with the core GPEI-containing constructwhile mutation of either of the TRE-like elementsstrongly reduced the tBHQ-induced promoter activity(data not shown). These data strongly indicated thatthe GPEI represents the tBHQ responsible element andthat both TRE-like elements in this element are essen-tial for this stimulation of the promoter activity.

6-4. Discussion

Diverse chemical compounds, including Michaelreaction acceptors, quinones, catechols, peroxides,isothiocyanates, mercaptans, heavy metals and planarpolycyclic aromatic hydrocarbons, induce the expres-sion of the phase II detoxifying enzymes. The induc-tion of several genes in the phase II enzyme battery ismediated through ARE or its functional equivalent, anelectrophile response element, EpRE, found in themouse GST Ya gene (Rushmore and Pickett 1990). Ithas been hypothesized that these inducing agents havein common the ability to generate reactive oxygen in-termediates directly or via redox cycling which resultin the activation of transcription factors, possiblythrough their alterations of the redox status. While theROS-dependent mechanism is favored by some re-searchers, Talalay and his colleagues pointed out thatmost of the phase II inducers are capable of generat-ing electrophilic intermediates, and have the potentialto react with nucleophilic residues of proteins includ-ing sulfhydryl groups (Prestera et al. 1993). Therefore,they have speculated that a better designation for theseresponsive elements would be EpREs to reflect thenature of the active intermediates. Recently, the corre-lation between the phase II drug metabolizing abilityand electroplilic feature has clearly been demonstratedby structure-activity relationship studies of the natu-rally occurring and synthetic chemicals (Spencer et al.1991).

Although the antioxidant activity of tBHQ has beenwell studied by many researchers, the fate of tBHQafter exhibiting a radical scavenging ability is not fullyunderstood. We demonstrated here that DtBHQ showeda significant scavenging activity towards a stable wa-ter-soluble radical, DPPH (Fig. 21A), and its activityis similar to that of tBHQ. It should be noted that, notonly during the auto-oxidation process but also duringthe radical scavenging reaction, tBHQ is mainly con-verted to tert-butylbenzoquinone (Kim and Pratt 1992),which is easy to be substituted by nucleophiles byMichael addition type reaction. It is possible thatDtBHQ can be oxidized to convert to the electrophilicBQ form. No correlation between the potential to in-duce the GSTP1 expression and the antioxidant effi-cacy of the HQ derivatives indicated that theantioxidative reaction of the hydroquinone derivativescould be ruled out in the molecular mechanism. There-fore, it is suggested that the metabolic product(s) oftBHQ should be taken into account in the mechanismof the GST induction. In addition, DtBHQ exhibit notonly radical scavenging activity in vitro (Fig. 21A) butalso inhibitory effects against intracellular oxidativestress induced by a GSH-depleting agent, DEM (Fig.21B). It is thus likely that the oxidative conversion ofDtBHQ to the BQ form may occur within cells, asshown in Fig. 27.

The ROS generating ability of tBHQ has been dem-onstrated to play important roles in some biological orchemical systems, i.e., cytotoxicity, oxidative DNAdamage, and activation of transcriptional factors in-cluding AP-1 and NF-kB. As described above, hydroxylradical has been suggested to be a critical intermedi-ate of the tBHQ-triggered ARE-dependent gene expres-sion including rGSTA1 (Pinkus et al. 1996). The redoxcycling between tBHQ (fully reduced), the semiqui-none radical (one-electron oxidation product) and tert-butylbenzoquinone (fully oxidized by two electrons)is concomitant with the formation of O2

–, a one-elec-tron reduction product of molecular oxygen (Fig. 27).Indeed, DtBHQ generated O2

– in an amount compara-ble to tBHQ (Fig. 22B) and thus induces oxidativecleavage of DNA as previously reported (Okubo et al.1997). We thus concluded that DtBHQ possesses a re-dox-active character in a manner similar to tBHQ.However, there is no correlation between the ROS gen-eration ability and GST induction potential of tBHQand DtBHQ.

We also demonstrated in the present study that thesignificant yield of BQs of tBHQ and DtBHQ was ob-tained by the Cu2+-mediated oxidation in vitro (Fig.22A). On the other hand, the in vitro experimentsclearly showed that DtBHQ has a poor reactivity withreduced GSH, while tBHQ was quickly and stoichio-metrically consumed upon incubation with GSH (Fig.23A). The structural difference between DtBHQ andtBHQ is only a bulky t-butyl moiety and DtBHQ has



Fig. 27. Summary of redox cycling and electrophilic potentials of tBHQ and DtBHQ. Reprinted from Biochemistry, 42(14),Nakamura et al., Pivotal role of electrophilicity in glutathione S-transferase induction by tert-butylhydroquinone, 4300–4309, Copyright (2003), with permission from American Chemical Society.

two t-butyl groups on both sides against the axis link-ing the two hydroxyl groups of the hydroquinone moi-ety (Fig. 20). The lower ability of DtBHQ to conju-gate with GSH was supported by the in vivo metabolicstudy of tBHQ which indicated that the Micheal addi-tion of GSH thiol group to tBHQ occurs much easieron the opposite side of the t-butyl moiety than on thesame side (Peters et al. 1996). These results clearlyindicated that interference of the GSH conjugation bythe introduction of the t-butyl moiety might be due toits steric hindrance rather than to its electronic charac-teristics, although the t-butyl moiety is an electron-donating group and reduces the reaction rate of theMichael addition to the m- or p-position. This assump-tion was also supported by the quantum mechanicalcomputer calculation study that the energy of the low-est unoccupied molecular orbital (LUMO) of DtBHQis similar to that of tBHQ (data not shown). The topicof the present study is that DtBHQ, which has a pro-pensity for redox cycling similar to that mentionedabove but is hardly reacting with GSH probably due tosteric hindrance of the bulky tert-butyl moieties, hasless ability for GST induction. Thus, we concluded thatan electrophilic quinone oxidation product that reactswith intracellular nucleophiles including protein thiolor GSH is the pivotal inducer of the GSTP1 gene ex-pression.

It is believed that activation of the phase II enzymegenes is regulated by signal transducing kinase cas-cades (Ainbinder et al. 1995). Yu et al. (1999) have

recently identified the extracellular signal-regulatedprotein kinase (ERK) pathway to be involved in theARE-mediated induction of phase II enzymes by tBHQand an isothiocyanate compound. They have alsoshown that the induction involves the direct activationof Raf-1 (a MAPK kinase kinase), whichphosphorylates and activates MEK (a MAPK kinase).It is therefore likely that activation of Raf-1 followedby stimulation of the ERK kinase pathway is involvedin the induction of GSTP1 by tBHQ. More recently,Yu et al. have also shown that the p38 MAP kinasenegatively regulates the induction of the phase II drug-metabolizing enzymes (2000). Interfering with the p38kinase pathway by overexpression of a dominant-nega-tive mutant of p38 or mitogen activated protein kinasekinase 3 (MKK3), an immediate upstream regulator ofp38, potentiated the activation of the ARE reporter geneby tBHQ, whereas the wild types of p38 and MKK3diminished such activation. These results suggestedthat the coordinate modulation of MAP kinase cascadesmay be critical to the regulation of the phase II en-zyme genes through the ARE induced by variouschemoreventive agents (Kong et al. 2000).

We showed here that tBHQ stimulated the promoteractivity of the 5¢-flanking region of the GSTP1 gene(Fig. 25B) and then induced GST mRNA and proteinin RL34 cells (Fig. 25A). It appeared that this stimula-tion required a specific region containing GPEI whichspecifically responded to tBHQ activation (Figs. 25B,26). It has been shown that the enhancer of the GSTP1



expression is regulated by multiple factors, includingAP-1 which is known to be a heterodimer composedof the products of the jun and fos oncogenes (Diccianniet al. 1992). It may be likely that c-Jun functions as animportant component that activates GPEI followed bythe expression of GSTP1. On the other hand, anothercandidate of the trans-acting factor(s) for the induc-tion of GST and other Phase II enzymes has been re-cently identified. Venugopal and Jaiswal (1996) havereported that the transcription factor NF-E2-relatedfactor 2 (Nrf2) positively regulates the ARE-mediatedexpression of the Phase II detoxification enzyme genes.Itoh et al. (1997) have also shown by gene-targeteddisruption in mice that Nrf2 is a general regulator ofthe Phase II enzyme genes in response to electrophilesand ROS. The general regulatory mechanism underly-ing the electrophile counterattack response has beendemonstrated, in which electrophilic agents alter theinteraction of Nrf2 with its repressor protein, kelch-like ECH associated protein 1 (Keap1), thereby liber-ating Nrf2 activity from repression by Keap1, culmi-nating in the induction of the Phase II enzyme genesand antioxidative stress protein genes via AREs (Itohet al. 1999a). Keap 1 contains 25 cysteine residues, 9of which are expected to have highly reactivesulfhydryl groups (Itoh et al. 1999b). More recently,Dinkova-Kostove et al. reported direct evidence thatsulfydryl groups of Keap1, covalently modified byelectrophilic agents, act as the sensors regulating theinduction of the phase II enzymes (2002). Therefore,the Keap1-Nrf2 complex is the most plausible candi-date for the cytoplasmic sensor system that recognizesinducers including tBHQ. DtBHQ can be used as a non-electrophile control for the identification of sensor pro-tein. Although the participation of Nrf2 in the mecha-nism of the GSTP1 induction by tBHQ is not fully elu-cidated, further studies of the signaling pathways be-tween electrophilic attachment to a plausible targetmolecule and GST gene expression are essential toadvance our understanding of the efficacy and safetyof phase II enzyme inducers as therapeutic and pre-ventive agents.

7. Catechol-type polyphenol is a potential modi-fier of protein sulfhydryls: Development andapplication of a new probe for understandingthe dietary polyphenol actions

7-1. Introduction

Dietary polyphenols, such as flavonoids, are the mostcommon and widely distributed phytochemicals in fruitand vegetables, including green tea, wine, cocoa andberries (Bravo 1998). The antioxidant properties ofthese compounds are often claimed to be responsiblefor the protective effects of these components againstcardiovascular disease, cancer, and/or many other dis-

eases (Scalbert et al. 2005). Indeed, some of these ben-eficial effects have been demonstrated in animal mod-els and in some epidemiological studies (Scalbert etal. 2005). Alternatively, a wealth of data exists thatsuggests that most of the relevant mechanisms of dis-ease prevention by polyphenolic flavonoids, such asin cancer, are not related to their antioxidant proper-ties, but rather are due to the pro-oxidant action andthe direct interaction of flavonoids and target molecules(Galati and O’Brien 2003).

Searching for the dietary polyphenol “sensor” or highaffinity proteins that bind to polyphenolic flavonoidsis the first step in understanding the molecular and bio-chemical mechanisms of the functional effects of di-etary polyphenols. A few proteins that can directly bindwith flavonoids have been identified by predicting thetarget molecule, including fibronectin, fibrinogen, his-tidine-rich glycoprotein, and serum albumin, (Sazukaet al. 1996; Bae et al. 2009). Recently, laminin receptorwas reported as a catechin receptor that mediates theanticancer activity of (–)-epigallocatechin gallate(EGCg) (Tachibana et al. 2004). In addition, vimentinwas identified as a novel molecular target of the EGCg-protein interaction by affinity chromatography usingan EGCg-column (Ermakova et al. 2005). However,the biologic and physiologic significance of the func-tional effects of polyphenolic flavonoids is not cleardue to the existence of other targets and the bindingstructure remains unidentified.

Several lines of evidence indicate that the catechol-type polyphenols have a poor stability in neutral oralkaline solution (Nakayama et al. 2002). Moreover,previous reports showed that the catechol-typepolyphenols were easily oxidized in a cell culture me-dium with slightly alkaline pH (Long et al. 2000).Oxidation of flavonoids with a catechol structural motifin their B ring leads to the formation of a flavonoidquinone, which rapidly reacts with sulfhydryls in GSHor protein cysteine residues to form cysteinyl flavo-noid adducts (Awad et al. 2003; van Zanden et al.2003). More recently, we found that EGCg forms cova-lent adducts with protein sulfhydryls through autoxi-dation (Ishii et al. 2008). Many studies have implicatedcysteine sulfhydryls present in various transcriptionfactors, such as Keap1/Nrf2, nuclear factor-kB (NF-kB), and p53 as redox sensors for the transcriptionalregulation of many genes essential for maintainingcellular homeostasis. Some chemopreventive andcytoprotective agents have been found to targetcysteine sulfhydryls present in key transcription fac-tors or their regulators, thereby suppressing the aber-rant over-activation of carcinogenic signal transduc-tion or restoring/normalizing or even potentiating cel-lular defense signaling (Wakabayashi et al. 2004; Naand Surh 2006).

In the present study, to investigate the proteinsulfhydryl modification by the catechol-type



polyphenols, we used 3,4-dihydroxyphenyl acetic acid(DPA) as a model catechol compound, and developeda new probe to directly detect the protein modificationby polyphenol compounds using a biotin-tagged DPA(Bio-DPA). Thus, a model protein, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), was exposed toDPA or Bio-DPA and the modification behavior wascharacterized using Western blotting and MS. In addi-tion, RL34 cells were treated with Bio-DPA, and thetarget proteins were identified using a proteomic ap-proach.


7-2A. ChemicalsDPA, laccase, RNase from bovine, angiotensin I, EC,

ECg, and quercetin were obtained from Sigma-Aldrich(St. Louis, MO). Cy5-labeled avidin, Protein G-Sepharose, normal goat serum, and enhanced chemi-luminescence (ECL) Western blotting detection rea-gents were purchased from GE Healthcare UK Ltd.(Buckinghamshire, UK). EZ-link 5-(biotinamido)pentylamine, sulfo-N-hydroxysuccinimide, and 1-ethyl-3-(dimethylaminopropyl) carbodiimide wereobtained from Thermo Fisher Scientific, Inc. (San Jose,CA, USA). The anti-b-actin polyclonal antibody wasobtained from Santa Cruz Biotechnology, Inc. (SantaCruz, CA, USA). Sequence grade modified trypsin wasobtained from Promega Co. (Madison, WI, USA). b-Lactoglobulin from bovine milk (b-LG), EDTA,dithiothreitol (DTT), 5,5¢-dithiobis (2-nitrobenzoicacid) (DTNB), GSH, iodoacetoamide, trifluoroaceticacid (TFA), (–)-epicatechin (EC), and (–)-epicatechingallate (ECg) were purchased from Wako Pure Chemi-cal Industries, Ltd. (Osaka, Japan). The carboxyl groupof DPA was modified by amidation with EZ-link 5-(biotinamido) pentylamine by a modification of a pre-viously described procedure (Shibata et al. 2003).

7-2B. GSH assayMeasurement of the GSH level was performed as

mentioned before. Detection and characterization ofGSH-DPA adducts were performed by bothelectrospray ionization-liquid chromatography-massspectrometry (ESI-LC MS; see below) and nuclearmagnetic resonance (NMR) spectrometry using aBruker AMX400 (400 MHz; Bruker Daltonics, Ltd.,Karlsruhe, Germany).

7-2C. Electrospray ionization-liquid chromatogra-phy-tandem mass spectrometry (ESI-LC-MS/MS)

The peptide samples were analyzed by reversed-phase HPLC, a system that consisted of a nanospaceSI-1 HPLC system (Shiseido Co., Ltd., Tokyo, Japan),using a Capcell Pak C18 UG120 column (2.0 ¥ 250mm i.d.; Shiseido). These samples were eluted with a

linear gradient of water containing 0.1% formic acid(solvent A) and acetonitrile containing 0.1% formicacid (solvent B), (time = 0–3 min, 10% B; 3–45 min,10–40% B; 45–50 min, 40–50% B; 50–52 min, 50–80% B). The flow rate was 0.2 mL/min and columntemperature was controlled at 40∞C. The MS (MS/MS)analyses were performed on an LCQ ion trap mass sys-tem (Thermo Fisher Scientific, Inc.) equipped with anelectrospray ion source. The electrospray system em-ployed a 5-kV spray voltage and a capillary tempera-ture of 260∞C. Collision induced dissociation experi-ments in the positive ion mode were performed by set-ting the relative collision energy at 30% using heliumas the collision gas.

7-2D. DTNB assayLoss of sulfhydryl groups in GAPDH was assayed

as previously reported (Ishii et al. 2003).

7-2E. Dot Blot and Western blot analysesFor detection of DPA modified proteins, GAPDH and

whole cell lysates from RL34 cells were treated withSDS sample buffer with reducing agent for 5 min at100∞C. The samples were then transblotted or sepa-rated by 10 or 12.5% SDS-PAGE. The gel wastransblotted onto a nitrocellulose or PVDF membrane.The membranes were incubated with 5% skimmed milkfor blocking, washed, and incubated with the HRP-avidin. This procedure was followed by the additionof ECL reagents. The bands and spots were visualizedby Cool Saver AE-6955 (ATTO, Tokyo, Japan).

7-2F. Comparison of modification behavior of Bio-DPA to proteins

Thirty micro molar protein samples (GAPDH, b-LG,and RNase) were incubated with 0–100 mM Bio-DPAin 100 mM sodium phosphate buffer (pH 7.4) at 37∞Cfor 30 min with laccase. The reaction was terminatedby centrifugal filtration (Microcon 10, molecularweight cutoff of 10,000) to remove the low-molecularweight reactants. Bio-DPA-modified proteins weredetected by Western blot analysis with HRP-avidin,followed by the addition of ECL reagents. The bandswere visualized by Cool Saver AE-6955 in same ex-posure condition.

7-2G. ImmunocytochemistryThe DPA-treated RL34 cells were fixed overnight in

PBS containing 2% paraformaldehyde and 0.2% pic-ric acid at 4∞C. The membranes were permeabilizedby exposing the fixed cells to PBS containing 0.3%Triton X-100. The cells were then sequentially incu-bated in PBS solutions containing blocking serum (4%normal goat serum) and immunostained with Cy5-labeled avidin, rinsed with PBS containing 0.3% TritonX-100, and covered with anti-fade solution. Images ofcellular immunofluorescence were acquired using a



Fig. 28. Modification behavior of DPA to GSH. (A) Chemical structure of DPA. (B) GSH consumption by DPA. GSH (100mM) was incubated with or without DPA (100 mM) in 100 mM sodium phosphate buffer (pH 7.4) at 25∞C for 1 h in thepresence or absence of laccase (30 units). The amount of residual GSH was spectrophotometrically estimated using the com-mercial kit GSH-400. (C) Dose-dependent GSH consumption by DPA. GSH (100 mM) was incubated with 0–100 mM DPA in100 mM sodium phosphate buffer (pH 7.4) at 25∞C for 1 h with or without laccase. The amount of residual GSH was spectro-photometrically estimated using the commercial kit GSH-400. (D) Detection of GSH-DPA adducts by LC-MS analysis. GSH(100 mM) was incubated with DPA in 100 mM sodium phosphate buffer (pH 7.4) at 25∞C for 1 h in the presence of laccase.The LC-MS measurements were performed by monitoring ions at abs. 215 nm (lower) and m/z 478 (upper). (E) 1H-NMRanalysis of GSH-DPA adduct. Mixture of DPA and DPA-GSH was dissolved in D

2O, and then analyzed by NMR spectrometry.

Reprinted from Chemical Research in Toxicology, 22(10), Ishii et al., Catechol type polyphenol is a potential modifier ofprotein sulfhydryls: development and application of a new probe for understanding the dietary polyphenol actions, 1689–1698, Copyright (2009), with permission from American Chemical Society.



confocal laser microscope (Bio-Rad) with a 40 ¥ ob-jective (488 nm excitation and 518 nm emission).

7-2H. Peptide mass fingerprints (PMF)PMF was performed as previously reported (Bae et

al. 2009). Briefly, peptide extracts from gel pieces wereanalyzed with an UltraFLEX MALDI-TOF MS (BrukerDaltonics, Ltd., Bremen, Germany). Proteins wereidentified with MASCOT (Matrix Science, London,UK) searching algorithms using the non-redundantdatabase.

7-2I. Pull-down and immunoprecipitation assaysRL34 cells incubated with and without 50 mM Bio-

DPA in a control medium for 1 h were washed withPBS, harvested, and lysed in 10 mM Tris-HCl (pH 7.4),5 mM EDTA, 150 mM NaCl, 1% NP-40, and proteaseinhibitors. Cell lysates containing 0.5 mg of proteinwere incubated batch-wise with 25 mL of StreptAvidin-Plus beads overnight at 4∞C with constant shaking. Thebeads were rinsed five times with lysis buffer by cen-trifugation at 3,000 rpm for 5 min. The proteins wereeluted by boiling the beads in SDS-sample buffer for 5min and analyzed by SDS-PAGE followed byimmunodetection with anti-b-actin polyclonal anti-body. In addition, the cell lysates were incubated with2 mg of anti-b-actin polyclonal antibody overnight at4∞C. The mixture was then treated with 20 mL of Pro-tein G-Sepharose and incubated for 1 h at 4∞C. Themixture was then centrifuged (3,000 rpm, 5 min), rinsedfive times with lysis buffer, and subsequently boiledwith SDS-sample buffer. The biotinylated proteins werethen subjected to immunoblot and detection with HRP-avidin and ECL.

7-3. Results

7-3A. Modification behavior of DPA to sulfhydrylsTo investigate the modification behavior of catechol-

type polyphenols on intracellular sulfhydryls such asGSH and protein, we used DPA as a model catecholcompound (Fig. 28A). We first examined the modifi-cation behavior of DPA with GSH, which is the mostabundant intracellular sulfhydryl, in the presence oflaccase. As shown in Fig. 28B, incubation of 0.1 mMGSH with DPA (100 mM) or laccase (30 units) alonefor 1h resulted in no significant GSH loss. On the otherhand, the co-incubation of both DPA and laccase withGSH resulted in the complete loss of GSH (Fig. 28B).Approximately 60% of GSH was lost after a 1-h incu-bation with 20 mM DPA (Fig. 28C), suggesting thatDPA not only reacted with GSH but also oxidized GSHinto GSSG possibly due to the generation of reactiveoxygen species (ROS) accompanied by the formationof DPA-quinone. To confirm the formation of the GSH-DPA adduct, the reaction mixture with laccase wasanalyzed by ESI-LC-MS. As shown in Fig. 28D, a new

peak was detected on the HPLC chromatogram by UV(215 nm) detection. As expected, the molecular ionpeak at m/z 474.0, which corresponded to the GSH-DPA adduct, was detected at the same UV peak reten-tion time (Fig. 28D). In addition, this major adductwas identified as 6-glutathionyl-DPA by its NMR spec-tral data (Fig. 28E). These results indicate that DPAreacts with GSH through the formation of an o-qui-none structure accompanied by the oxidation of thecatechol moiety. To gain further details about thesulfhydryl modification by DPA, we preliminarily usedboth the cysteine peptide CaMKII (MHRQETVDC)and its S-carbamidomethyl-derivative. The DPA-modi-fied peptide was significantly detected, whereas thereactivity remarkably decreased by the S-carbamidomethylation of the cysteine residue (data notshown). Thus, using DPA and the peptides containinga sulfhydryl, we showed the high reactivity and selec-tivity for the cysteine residues of the catechol com-pound.

7-3B. Detection of the protein modification by Bio-DPA

Typical flavonoids such as quercetin and catechins,lacking of these functional groups besides hydroxylgroups, are insufficient for biotinylation. In addition,we previously demonstrated that PCA is converted intothe reactive quinone intermediate(s) by phenol oxidase,which can bind nucleophilic residues of proteins andalter the cellular immune functions (Nakamura et al.2001a, 2014). This study supported the idea that a sim-ple catechol structure is the minimum determinant forprotein modification by polyphenol compounds. There-fore, the preliminary study was undertaken to make abiotin probe using PCA. However, the reactivity ofPCA with the biotinylation reagent is too low to pro-vide the desired compounds efficiently. Therefore, inthis study, we used DPA as a model catechol compoundcontaining a carboxyl group which has a good reactiv-ity with the biotinylation reagent. To directly detectthe protein modification by catechol-type polyphenols,Bio-DPA was synthesized (Fig. 29A), and the modifi-cation behavior was characterized by applications thatexploit the specificity of the avidin-biotin complex.Hence, we examined the potential reactivity of Bio-DPA toward sulfhydryl enzymes, using GAPDH as aconvenient model protein that contains four sulfhydrylgroups per subunit, and is known to be highly sensi-tive to modification by electrophiles in vitro (Ishii etal. 2004; Loecken et al. 2008). As shown in Fig. 29B,upon incubation of GAPDH (1 mg/mL) with 50 mMBio-DPA in the presence of laccase, the free sulfhydrylgroups of GAPDH were remarkably decreased. In ad-dition, the Bio-DPA-induced loss of the free sulfhydrylgroups was accompanied by an increase in ELISA-posi-tive rates probed with an HRP-avidin, due to the bind-ing of Bio-DPA (Fig. 29C). To evaluate the binding of



Fig. 29. Covalent binding of Bio-DPA to protein sulfhydryls.(A) Chemical structure of biotin-tagged DPA (Bio-DPA). (B)Loss of sulfhydryl groups in GAPDH. GAPDH (1 mg/mL)was incubated with 50 mM Bio-DPA in 100 mM sodium phos-phate buffer (pH 7.4) at 37∞C for 30 min with or withoutlaccase. After the incubation, an aliquot (0.1 mL) was takenfrom the reaction mixture, and the amount of sulfhydrylgroups was determined by DTNB assay as described in “Ex-perimental Procedures”. (C) ELISA of Bio-DPA-modifiedGAPDH probed with HRP-avidin. GAPDH (1 mg/mL) wasincubated with 0–100 mM Bio-DPA in 100 mM sodium phos-phate buffer (pH 7.4) at 37∞C for 30 min with (�) or with-out laccase (�). (D) Western blot analysis of Bio-DPA-modi-fied GAPDH probed with HRP-avidin. GAPDH (1 mg/mL)was incubated with 0–50 mM Bio-DPA in 100 mM sodiumphosphate buffer (pH 7.4) at 37∞C for 30 min with (left panel)or without laccase (right panel). (E) Modification behaviorof Bio-DPA to GAPDH in the presence of a protein sulfhydrylalkylating agent. GAPDH (1 mg/mL) was incubated with50 mM Bio-DPA and laccase in 100 mM sodium phosphatebuffer (pH 7.4) at 37∞C for 30 min with (upper panel) orwithout iodoacetoamide (lower panel). (F) Modificationbehavior of Bio-DPA to proteins. The protein samples(GAPDH, b-LG, and RNase) were incubated with 0–100 mMBio-DPA in 100 mM sodium phosphate buffer (pH 7.4) at37∞C for 30 min with laccase. Bio-DPA-modified proteinswere detected by Western blot analysis with HRP-avidin.Reprinted from Chemical Research in Toxicology, 22(10),Ishii et al., Catechol type polyphenol is a potential modifierof protein sulfhydryls: development and application of a newprobe for understanding the dietary polyphenol actions,1689–1698, Copyright (2009), with permission from Ameri-can Chemical Society.

Bio-DPA to the sulfhydryl groups of the enzyme, weexamined the incorporation of Bio-DPA to the enzymewith or without laccase by SDS-PAGE. Sulfhydrylmodification by Bio-DPA was confirmed only in thepresence of laccase by HRP-avidin blot analysis (Fig.29D), suggesting the oxidation of the catechol moietyon Bio-DPA is important for protein modification.Moreover, the incorporation of Bio-DPA was signifi-cantly inhibited by co-incubation with iodoacetoamide,which is an alkylating agent of protein sulfhydryls (Fig.29E). This result strongly suggests that Bio-DPA ex-clusively binds to the cysteine residues of GAPDH toform the covalent binding. In an attempt to gain sup-port for our hypothesis, that the protein modificationby Bio-DPA might be achieved through the directelectrophile scavenger function of the sulfhydrylgroups, we compared the Bio-DPA incorporating rateof GAPDH (contains four sulfhydryl groups persubunit) with other proteins, such as b-LG (contains asulfhydryl group per subunit), and bovine RNase (con-tains no sulfhydryl group per subunit). The Bio-DPAincorporating rate was assessed by an HRP-avidin blotanalysis at various concentrations during the incuba-tion. As shown in Fig. 29F, among the proteins tested,GAPDH showed the highest incorporating rate of Bio-DPA. In contrast to the remarkably fast rate of scav-enging by GAPDH, Bio-DPA had still not been incor-porated into other proteins at 10 mM. On the other hand,Bio-DPA was incorporated into b-LG and bovineRNase at 100 mM. Oxidation of polyphenols with acatechol structural motif leads to the formation of apolyphenol quinone, which rapidly reacts withsulfhydryls in GSH or protein cysteine residues to formcysteinyl polyphenol adducts. Although the reactivityis lower than that of the sulfhydryl group, amino andimidazol moieties are targets of polyphenol quinoneto form stable adducts (Bolton et al. 1997). Bio-DPAmay react with lysine and histidine residues in the pro-teins. These data suggest that the sulfhydryl groups mayrepresent the direct sensors of the catechol-typepolyphenols.

7-3C. Identification of target protein of Bio-DPA inRL34 cells

To detect cellular proteins that undergo cysteine-tar-geted modification through the oxidation of the cat-echol moiety, Bio-DPA, having the similar redox po-tential to DPA (data not shown), was utilized as a mo-lecular probe. As shown in Fig. 30A, when RL34 cellswere treated with 100 mM Bio-DPA for 1 h, a signifi-cant incorporation of Bio-DPA into the cells was ob-served. After exposure to Bio-DPA, the whole celllysates were blotted onto a PVDF membrane, and thebiotin-labeled proteins were analyzed by dot blot andWestern blot analyses probed with HRP-avidin. WhenRL34 cells were treated with 0–100 mM Bio-DPA for1 h, the biotin-labeled proteins were detected in a dose-



Fig. 30. Detection of Bio-DPA-modified proteins in RL34 cells (A) Immunocytochemical detection of Bio-DPA incorporationinto the cells. RL34 cells were treated with 100 mM Bio-DPA for 1 h. (B) Dot blot analysis of Bio-DPA-protein adductsprobed with HRP-avidin. (C) Western blot analysis of Bio-DPA-protein adducts probed with HRP-avidin. In panels B and C,cells were treated with 50 mM Bio-DPA for 1 h. The protein samples were direct-transblotted or separated by SDS-PAGEfollowed by Western blotting. (D) The cells were lysed with a sonicater in PBS with protease inhibitors. After separation bycentrifuge, the supernatant was removed and reserved as the soluble cytoplasmic fraction. The precipitated pellet was thendissolved in SDS-sample buffer and kept as the insoluble fraction containing membranes, organelles and cytoskeleton. Eachprotein samples were separated by SDS-PAGE followed by Western blotting. Reprinted from Chemical Research in Toxicol-ogy, 22(10), Ishii et al., Catechol type polyphenol is a potential modifier of protein sulfhydryls: development and applicationof a new probe for understanding the dietary polyphenol actions, 1689–1698, Copyright (2009), with permission from Ameri-can Chemical Society.



dependent manner (Fig. 30B). As shown in Fig. 30C,the Bio-DPA-modified proteins were detected around40-kDa (p-1) and 70-kDa (p-2) when assayed by West-ern blotting. In addition, this 40-kDa band was promi-nently detected in the insoluble fraction containingmembranes, organelles and cytoskeleton (Fig. 30D).The p-1 was identified to be b-actin by PMF analysis.Using MASCOT, the probability based MOWSE scorewas 126 for b-actin (p < 0.05). In addition, p-2 wasidentified as bovine serum albumin (BSA) by the analy-sis (data not shown), which is a contaminant from acell culture medium with FBS. Bio-DPA can bind toBSA, since serum albumin is one of the major targetsof interaction by polyphenols (Bae et al. 2009). Onthe other hand, GAPDH, reactive with DPA in vitro,was not detected as a Bio-DPA target, assumed thatbiotin tag could sterically hinder the ability of DPA toaccess protein cysteine residues.

We then attempted to detect the modification of b-actin in cells exposed to Bio-DPA. To this end, RL-34cells were treated with 50 mM Bio-DPA for 1 h, andthe cell lysates were incubated with NeutrAvidin beads.After washing with lysis buffer, proteins bound to theresin through Bio-DPA were eluted with SDS-PAGEsample buffer, and b-actin was detected by Western

blot analysis using the anti-b-actin polyclonal antibody(Fig. 31, upper panel). Alternatively, the cell lysateswere subjected to immunoprecipitation with the anti-body, and the presence of Bio-DPA-modified proteinswas detected by Western blotting probed with HRP-avidin (Fig. 31, lower panel). Thus, it appears that DPAreacted, to an appreciable extent, with endogenous b-actin in intact RL34 cells.

7-3D. Modification of b-actin by catechol-t ypepolyphenols in RL34 cells

Previously, Böhl et al. reported that b-actin is a tar-get protein for the flavonol quercetin (2005). To ex-amine the reactivity of the catechol-type polyphenols,such as a flavonoid, to endogenous b-actin, RL34 cellswere treated with 50 mM Bio-DPA after pretreatmentwith DPA and flavonoids (quercetin, EC, and ECg) for1 h, and its modifications were analyzed by Westernblotting probed with HRP-avidin. As shown in Fig.32A, DPA (0–500 mM) pretreatment of RL34 cells re-sulted in a dose dependent inhibition of the Bio-DPAmodification. Furthermore, the formation of the Bio-DPA-modified b-actin was significantly decreased bypretreatment with the 50 mM flavonoids (Fig. 32B).On the other hand, vitamin C or vitamin Epretreatments did not significantly inhibit the Bio-DPAmodification (Fig. 32C). This indicated that the anti-oxidant action of DPA or flavonoids does not contrib-ute to the inhibition of the reactivity of Bio-DPA forb-actin. These results strongly suggest that b-actin is apotential target of the catechol-type polyphenols incellular proteins.

7-3E. Effect of DPA on cytoprotective enzyme in-duction in RL34 cells

It has been reported that dietary polyphenols have apotential to induce phase II drug metabolizing enzymeexpression (Na and Surh 2006). As described above,electrophilic reaction of Bio-DPA to cellular proteinswas inhibited by the flavonoid pretreatment (Fig. 32B).Therefore, to investigate whether the electrophilicityof the catechol-type polyphenol is involved in thecytoprotective enzyme induction, we examined theGSTP1- and NQO1-inducing ability of DPA. The in-cubation of RL34 cells with 0-100 mM DPA for 8 h ledto a significant increase in the level of GSTP1 andNQO1 mRNA (Fig. 33). Human Keap1 contains 27cysteines, some of which are purported to be the tar-gets of electrophiles and ROS that, when modified,facilitate the derepression of Nrf2 and give rise to en-zyme induction (Itoh et al. 2003). This leads us to as-sume that direct modification of Keap1 of DPA mayregulate Keap1 and induce the gene expression. How-ever, the modification of Keap1 by Bio-DPA was notdetected by either the proteomic approach (Fig. 30) orthe pull-down assay (data not shown). These observa-tions suggested that Keap1 is an additional target of

Fig. 31. Modification of b-actin by Bio-DPA in RL34 cells.RL34 cells were treated with 50 mM Bio-DPA for 1 h. Celllysates were incubated with Immobilized NeutrAvidin orwith anti-b-actin, as indicated. The presence of the Bio-DPA-modified b-actin was detected by immunoblot analysis (up-per panel), and the incorporation of Bio-DPA into the b-ac-tin immunoprecipitates was detected with HRP-avidin andECL (lower panel). Reprinted from Chemical Research inToxicology, 22(10), Ishii et al., Catechol type polyphenol isa potential modifier of protein sulfhydryls: development andapplication of a new probe for understanding the dietarypolyphenol actions, 1689–1698, Copyright (2009), with per-mission from American Chemical Society.



Fig. 32. Modification of b-actin by catechol-type polyphenolsin RL34 cells. RL34 cells were pretreated with DPA andflavonoids (quercetin, EC, and ECG) for 1 h. The flavonoid-containing medium was removed by washing with PBS andreplaced by the control medium containing 50 mM Bio-DPAfor 1 h, and its modification was analyzed by Western blot-ting probed with HRP-avidin. (A) Pretreatment with DPA(0–500 mM) for 1 h. (B) Pretreatment with 50 mM flavonoids(quercetin, EC, and ECG) for 1 h. (C) Pretreatment with 50mM vitamin C or vitamin E for 1 h. Reprinted from Chemi-cal Research in Toxicology, 22(10), Ishii et al., Catechol typepolyphenol is a potential modifier of protein sulfhydryls:development and application of a new probe for understand-ing the dietary polyphenol actions, 1689–1698, Copyright(2009), with permission from American Chemical Society.

Fig. 33. Induction of the Keap1/Nrf2-regulated enzyme genesby DPA in RL34 cells. Incubation of RL34 cells with 0–100mM DPA for 8 h. Dose-dependent effect of DPA on mRNAlevels of NQO1 (upper), GSTP1 (middle), and GAPDH(lower). Reprinted from Chemical Research in Toxicology,22(10), Ishii et al., Catechol type polyphenol is a potentialmodifier of protein sulfhydryls: development and applica-tion of a new probe for understanding the dietary polyphenolactions, 1689–1698, Copyright (2009), with permission fromAmerican Chemical Society.

DPA through other mechanism than covalent bindingsuch as oxidation of intracellular sulfhydryls includ-ing GSH or protein cysteine residues.

7-4. Discussion

Much attention has recently focused on the identifi-cation of a promising target protein which can interactwith food chemicals. Quinones act as electrophiles andform covalent bonds with protein nucleophiles, irre-versibly altering key cellular proteins, and individualquinones exhibit this reactivity to varying degrees (Bol-ton et al. 2000). Dopamine quinone, which is an oxi-

dation metabolite of dopamine, has been shown tocovalently bind to reduced cysteine residues of pro-teins in vivo and in vitro (LaVoie et al. 2005; Akagawaet al. 2006). Hence, we also note that quinone-medi-ated covalent binding through oxidation may also oc-cur in other polyphenolic compounds. In this study,we demonstrated that the known reactions of catecholmoieties with protein nucleophiles, such as thiolgroups, can be exploited to study protein-polyphenolinteractions and, furthermore, to screen for unknownprotein targets of polyphenols of physiological andpharmacological interest. We have definitively indi-cated that the catechol-type polyphenols, includingBio-DPA, have a high reactivity and selectivity forcysteine residues (Figs. 28, 29). The target identifica-tion was based on the proteomic approach (Fig. 30)and pull-down technique (Fig. 31). The inherent limi-tation of the technique lies in the required structuralproperties of polyphenol and protein that allow the useof chemical synthesis techniques. However, in thepresent experiments using Bio-DPA, we found that theapproach may be successfully undertaken for a widerange of protein-polyphenol interactions (Fig. 32).Further studies will address these problems, and anattempt will be made to identify all major target pro-teins of the catechol-type polyphenols in human cells.

Several lines of evidence indicate that ROS andelectrophiles act as second messengers, representingan integral part of the cellular signal transduction net-work. The downstream effect of ROS and electrophilesis the oxidation of redox-sensible proteins through thedirect modification of the thiol group of reactivecysteines (Na and Surh 2006; Rhee 2006). Therefore,the modification of sulfhydryl groups in cellular pro-



teins by catechol-type polyphenols may cause the sig-nal transductions through structural or functional al-terations. We demonstrated here that b-actin is one ofthe major molecular targets of protein modification,not only by Bio-DPA (Figs. 30, 31), but also byflavonoids (Fig. 32). Actin is one of the most abun-dant proteins in human tissues and serves essentialfunctions as the cytoskeletal component in muscle andnonmuscle cells. Actin may exist as a monomer (globu-lar actin (G-actin)) in the cell, but readily polymerizesto form microfilaments (filamentous actin (F-actin)),rendering actin the major molecular player in the con-trol of cell shape, cell adhesion, and cell motility (Pol-lard et al. 2000). The dynamic reorganization of cellu-lar actin is highly regulated, and ROS appear to be onevital regulatory element. Recent studies indicated thatactin could constitute a direct target for oxidative modi-fication (Fratelli et al. 2002; Fiaschi et al. 2006). Ac-tin glutathionylation has already been reported in re-sponse to epidermal growth factor and in hepatocytesand hepatoma cells exposed to oxidative stress (Fratelliet al. 2002), and Cys272 and Cys374 were identifiedas reactive cysteines (Lassing et al. 2007). Moreover,actin is oxidatively modified in human pathophysi-ological states suggestive of oxidation as a cause ofmechanical dysfunction (Pastore et al. 2007). Thus, an

explanation of the molecular basis of redox-regulatedmicrofilament processes requires an understanding ofthe mechanism of assembly/disassembly and relatedchanges of actin during redox conditions, as well asits regulation by associated proteins.

Previously, actin cytoskeleton organization, i.e.,stress fibers, in adherent fibroblasts was changed upontea catechin treatment (Hung et al. 2005). Furthermore,a microscopic analysis of quercetin-targets in livingcells revealed that F-actin rings, as found in Drosophilafollicles, fluoresce brightly (Gutzeit et al. 2004). Inthe light of these results, it seems likely that theflavonoids bind to actin in this special cytoplasmicstructure, even though actin is apparently present in F-actin. This interpretation may also explain the observedquercetin-induced disruption of actin rings and of boneresorption in osteoclast-like cells (Woo et al. 2004).Although the authors speculated that quercetin mightact on the signaling pathway involved in the assemblyof actin rings, the present results may indicate that di-rect interaction of the catechol moiety with actin alsomodulates other signal transduction pathways.

Induction of a family of cytoprotective enzyme genesencoding for proteins that protect against the damageof ROS and electrophiles is potentially a major strat-egy for reducing the risk of cancer and chronic degen-

Fig. 34. The mechanism underlying dose-dependent switching between beneficial and harmful phenomena induced by food-derived electrophiles. Reprinted from Bioscience, Biotechnology and Bioscience, 74(2), Nakamura and Miyoshi, Electrophilesin foods: the current status of isothiocyanates and their chemical biology, 242–255, Copyright (2010), with permission fromJapan Society for Bioscience, Biotechnology and Agrochemistry.



erative diseases. Many cytoprotective enzyme genesare regulated by upstream antioxidant response ele-ments that are targets of the leucine zipper transcrip-tion factor Nrf2 (Itoh et al. 2003). Under basal condi-tions, Nrf2 mainly resides in the cytoplasm bound toits cysteine-rich Keap1, which is itself anchored to theactin cytoskeleton and represses Nrf2 activity. Induc-ers disrupt the Keap1-Nrf2 complex by oxidizing re-active cysteine residues of Keap1 (Eggler et al. 2005).The identification of b-actin as the molecular target ofcatechol-type polyphenols led us to the assumption thatactin scaffold proteins might also be involved in thepolyphenol-induced modulation of the Keap1/Nrf-2system. This assertion is based on the following ob-servations; (i) simple catechol compounds, as well ascatechol-type polyphenols, have been reported to in-duce cytoprotective enzyme expression and activity(Na and Surh 2006), (ii) catechol-type flavonoids bindto protein sulfhydryls, (iii) Keap1 binds to the actincytoskeleton in the cytoplasm, and (iv) the modifica-tion of Keap1 by Bio-DPA was not detected by thepresent approach. Detection of modified Keap1 is dif-ficult because the expression of Keap1 in cells is muchlower than that of housekeeping proteins such as ac-tin, one of the major targets of protein modification bycatechol-type polyphenols. In addition, the thiol-polyphenol adduct is reversible, which has also beenreported for the GSH or GSTP1 adduct (Boersma etal. 2000; van Zanden et al. 2003). As a similar exam-ple, the Keap1 adduct of sulforaphane, one of the mostfamous electrophilic inducers of cytoprotective en-zymes from cruciferous vegetables, has never beendirectly detected from the in vivo system because ofits reversibility to hydrolysis and transacylation reac-tions (Hong et al. 2005). As another possibility, wepropose the participation of the pro-oxidant actions ofpolyphenols. Most plant polyphenols possess both anti-oxidant as well as pro-oxidant properties (Sergedieneet al. 1999). In vitro studies also indicated that teacatechins are unstable in several cell culture media(Nakayama et al. 2002). The oxidation of polyphenolsand the resulting production of ROS have been con-sidered responsible for biological activities such asreceptor inactivation and telomerase inhibition(Naasani et al. 2003; Hou et al. 2005). Conversely, ithas been suggested that although free quinones areshort-lived in vivo, the binding of quinones to a pro-tein may dramatically extend the half-life of these re-active species (Graham et al. 1978; Paz et al. 1991).Furthermore, the intracellular-ROS production bycysteine- and protein-bound polyphenols was reported(Izumi et al. 2005; Akagawa et al. 2006). Therefore,modification of the intracellular actin by catechol-typepolyphenols may induce the oxidation of GSH andother protein sulfhydryls, such as Keap1, through pro-oxidant actions. In fact, down-regulation of the intrac-ellular GSH level enhanced the cytoprotective enzyme

induction by flavonoids, supporting the pro-oxidanttheory (Lee-Hilz et al. 2006). This finding is in linewith other literature reports demonstrating that the pro-oxidant action of catechol-type polyphenols rather thantheir antioxidant activity is important for their health-promoting property (Azam et al. 2004). Future studiesneed to clarify this point and address the question ofwhether the catechol group is able to inhibit the differ-ential functions of actin in the nucleus and cytoplasm,such as with the Keap1/Nrf-2 system.

In conclusion, this study indicates that b-actin is oneof the major targets of protein modification by catechol-type polyphenols, and provides an alternative approachto understand that catechol-type polyphenol is a po-tential modifier of redox-dependent cellular eventsthrough sulfhydryl modification. The results also en-courage further investigation into the biological activi-ties of polyphenols. In addition, the present results maynot reflect bindings of dietary polyphenols in vivo en-tirely, because a biotinylation of DPA may hinder theability of DPA to access protein cysteine residues. It isstill uncertain and needs to be elucidated that catechol-type polyphenol first targets b-actin in cultured cellsand that these bindings are associated with itsbioactivities. Searching for high affinity proteins thatbind to the polyphenolic flavonoid is the first step tounderstanding the molecular and biochemical mecha-nisms of the functional effects of dietary polyphenols.Our results provide a foundation for future studiesabout the mechanism of action of dietary polyphenols.

8. Conclusion and perspective

Undoubtedly, polyphenols in foods possess poten-tial to exhibit various beneficial effects in health pro-motion. Indeed, the present study confirms that a lowdose of a catechol-type phenolic compound PCA sig-nificantly decreases both the incidence and the multi-plicity of skin tumors (Section 2). PCA at high dosesenhances tumorigenesis in an in vivo experiment, eventhough it is a promising candidate as anti-carcinogen-esis (Tanaka et al. 1993). This finding is consistent witha previous report showing that some phenolic antioxi-dants have tumor promoting activity in rat forestom-ach carcinogenesis (Ito and Hirose 1989). The presentstudies also clearly demonstrate that topical treatmentwith high doses of phenolic compounds significantlyenhances mouse skin inflammation and oxidative stressin an oxidative enzyme-dependent manner (Sections2, 3, and 4). The puzzling results showing that the typi-cal phenolic antioxidants unexpectedly enhanceoxidative stress, even as the application dose is in-creased, are the topic of this monograph. Applicationof PCA alone in large amounts causes not only the dis-turbance of the GSH-dependent detoxification systembut also immunoinflammatory responses such as con-tact hypersensitivity (Section 3), strongly suggesting



the electrophilic reactivity of the quinone metabolitesis involved in these toxic mechanisms. The in vitrostudies provide supporting evidence showing that cat-echol-type phenolic compounds form electrophilicquinone-type intermediates which bind to intracellu-lar nucleophiles such as GSH and protein sulfhydryls.In addition to the skin, possible toxic effects on theliver and kidney with GSH depletion were observed inmice administered with high doses of a phenolic anti-oxidant (Section 5). Consistent with this, hepatotoxic-ity in mice was induced by the flavonoid EGCg, pro-pyl gallate and nordihydroguaiaretic acid (NDGA)when administered i.p. to mice (Galati and O’Brien2003). Taken together, the data in this monograph sup-port the idea that intake of excess amounts ofpolyphenols provokes electrophilic and/or pro-oxidantreactions dependently on biotransformation, which cancontribute to promotion of the inflammatory process.It has been reported, in a review by Mennen and col-leagues, that tissues rich in oxidative enzymes may beespecially vulnerable to pro-oxidant toxicity bypolyphenols (2005), supporting this hypothesis.

Transcriptional induction of phase II cytoprotectivegenes involved in cellular defense against chemical car-cinogens and oxidative stress substantially contributesto their preventive activity, as shown in Sections 6 and7. The present studies conclude that electrophilic oxi-dation products of polyphenols that react with intrac-ellular nucleophiles including protein thiol or GSH playa major role in the phase II cytoprotective gene ex-pression. As mentioned in Section 2, it is possible tospeculate that electrophilic quinones also play someimportant roles in anti-carcinogenic activity of phe-nolic antioxidants in lower doses. Since the phase IIcytoprotective genes are highly inducible, phase II pro-tective systems can be up-regulated by a wide varietyof chemically diverse inducers (Nakamura and Miyoshi2010). To date, several epidemiological studies haveindicated that the protective effects of high levels ofvegetable consumption against the risk of developingcancer may be at least partly attributable to inductionof phase II genes (Talalay et al. 2003). It has recentlybecome clear that phase II gene inducers also act asindirect antioxidants in a manner different from radi-cal-scavenging direct antioxidants (Dinkova-Kostovaand Talalay 2008), based on the following reasons: (i)The family of phase II enzymes includes, in additionto the classical conjugating enzymes (e.g., GSTs andUDP-glucuronosyltransferases), many other proteinssuch as antioxidant enzymes (e.g., heme oxygenase-1;HO-1 and catalase) and other protective proteins thatare not enzymes but can affect oxidative status (e.g.,ferritin and thioredoxin). Hence it is more accurate torefer to them as cytoprotective proteins. (ii) Function-ally, many cytoprotective proteins play multifunctionalroles. For example, all GSTs catalyze the detoxifica-tion of a wide array of endogenous and exogenous

electrophiles, whereas some GSTs have GSH peroxi-dase activity towards hydroperoxides. (iii) Somecytoprotective proteins catalyze reactions whose prod-ucts are small direct antioxidant molecules. For exam-ple, HO-1 generates carbon monoxide and biliverdin/bilirubin, and g-glutamylcysteine synthetase catalyzesthe rate-limiting step in the synthesis of GSH, the mostabundant cellular small-molecule antioxidant. (iv) Thephase II enzymatic systems are responsible for regen-eration of small molecule direct antioxidants. GSHreductase plays a major role in regenerating reducedGSH, which participates in the reduction ofdehydroascorbate, which maintains ascorbate in itsreduced state. NQO1 regenerates ubiquinol and toco-pherol in their reduced forms. (v) The genes encodingfor cytoprotective proteins are coordinately induced bya common molecular mechanism. Three cellular com-ponents participate in the gene expression ofcytoprotective proteins: ARE, Nrf2, and Keap1. Thegeneration of nrf2-knockout mice provided an un-equivocal demonstration of the cytoprotective role ofNrf2-dependent gene products in a series of elegantexperiments using various models of electrophile andoxidant toxicities (Dinkova-Kostova and Talalay 2008).In addition to the Keap1/Nrf2/ARE-dependent mecha-nism, recent studies have implicated cysteine thiolspresent in various key factors, including inhibitor kB(IkB)/nuclear factor (NF)-kB/activator protein (AP)-1 and p53, as redox sensors in the transcriptional regu-lation of many genes essential for maintaining cellularhomeostasis. Also, some electrophilic agents have beenfound to target cysteine thiols present in transcriptionalfactors or their regulators (Na and Surh 2006). Al-though elucidation of the molecular and cellular tar-gets of polyphenols remains a big challenge, pharma-ceutical and dietary modulation of thiol-containing keyregulatory factors directly via covalent modificationby electrophiles in foods should be a unique and po-tentially effective strategy for protection of cells againstoxidative and inflammatory damage.

In addition to the beneficial aspect, it should be notedthat Nrf2 activation can enhance the resistance of can-cer cells to chemotherapeutic drugs (Wang et al. 2008).As mentioned above, the transgenic mice withoverexpression of GPx or both GPx and superoxidedismutase showed an enhanced tumorigenic response(Lu et al. 1997). Several food-derived compounds in-cluding electrophiles have also been identified assensitizers of allergic contact dermatitis, possiblythrough a non-classical carbon-centered radical-de-pendent pathway as well as classical electrophile-nucleophile-dependent pathway (Merckel et al. 2010).Moreover, several sensitizers of allergic contact der-matitis can induce Nrf2-dependent gene expression inhuman monocyte dendritic cells (Migdal et al. 2013).However, it is still unclear whether the Nrf2-depend-ent cytoprotective responses, occurring in the differ-



ent tissues, completely share the same chemistry, tar-get molecule, and molecular recognition system. Be-sides Nrf2, the toxicity induced by oxidizedpolyphenols (quinones) is also limited by a number ofefficient cytoprotective mechanisms that include meth-ylation by catechol-O-methyltransferase (COMT) andreduction reactions with either ascorbate or NQO1(Galati et al. 2006) under normal cellular conditions.COMT is being expressed in numerous mammalian tis-sues and catalyzes the O-methylation of endogenousand xenobiotic compounds including flavonoids. Forexample, COMT metabolizes EGCG to 4≤-O-methyl-EGCG and 4¢,4≤-di-O-methyl-EGCG, both of whichhave significantly less biological activity than theunmetabolized EGCG (Fang et al. 2003). NQO1 is aubiquitous cytosolic phase II detoxification enzymethat efficiently catalyzes the two-electron reduction ofpolyphenol quinones to their reduced catechol or

hydroquinone form (Gliszcynska-Swigl o et al. 2003).Under abnormal conditions with weaker activities ofsuch cytoprotective enzymes, the doses of polyphenolsthat are required for harmful phenomena might be low-ered. Since these findings indicate the difficulty inhealth promotion by electrophilic chemicals, furtherstudies are necessary to reveal more precise molecularmechanisms of absorption, metabolism, tissue distri-bution, signal transduction as well as molecular tar-gets of polyphenols.

Thus, electrophiles experimentally possess both ben-eficial and harmful effects in a dose-dependent man-ner (Fig. 34). Although dietary consumption ofpolyphenols from the food supply is high, the risk oftoxicity is relatively low, largely due to poor absorp-tion. This is because most polyphenols are present asglycosides in plant vacuoles and apoplasts, cellularstructures that require softening by cooking process tomaximize polyphenol release (Miglio et al. 2008).However, concentrations and bioavailability of dietarybioactive agents can be significantly increased throughpolyphenol enrichment, alterations in absorption effi-cacy, and supplementation with purified agents or mix-tures rendering polyphenol consumption potentiallyproblematic (Martin and Appel 2010). Therefore, moreattention should be paid to the dose administered as asupplement and, hence, carefully designedpharmacokinetic studies are needed before clinical test-ing of polyphenols. Further studies on the optimiza-tion of their dose or dose scheme for clinical study arealso necessary. In addition, strategies that use multi-ple agents with different modes of action rather thanindividual single agents should produce results withhigher efficacy and lower toxicity. Thus there is a needfor investigation concerning the combinatory use ofcertain food chemicals with electrophiles to achievesynergistic interaction for health promoting effect andlowered harmful effects.

AcknowledgmentsWe wish to express our sincere gratitude to Prof. Emer.

Hajime Ohigashi and Prof. Akira Murakami of Kyoto Uni-versity and Prof. Emer. Toshihiko Osawa and Prof. KojiUchida of Nagoya University for guidance and encourage-ment throughout the course of this study. We thank all ofour co-workers, especially Prof. Yoshiyuki Murata ofOkayama University, for their collaborative work and heart-felt support. We are also deeply thankful to our colleaguesand students for their technical support, helpful advice, andencouragement. This study was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Educa-tion, Culture, Sports, Science, and Technology of Japan.

ReferencesAinbinder E, Bergelson S, Daniel V. Regulatory mechanisms

involved in activator-protein-1 (AP-1)-mediated activa-tion of glutathione-S-transferase gene expression bychemical agents. Eur. J. Biochem. 1995; 243: 49–57.

Akagawa M, Ishii Y, Ishii T, Shibata T, Yotsu-Yamashita M,Suyama K, Uchida K. Metal-catalyzed oxidation of pro-tein-bound dopamine. Biochemistry 2006; 45: 15120–15128.

Ames BN, Shigenaga MK, Hagen TM. Oxidants, antioxi-dants, and the degenerative diseases of aging. Proc. Natl.Acad. Sci. USA 1993; 90: 7915–7922.

Arranz S, Chiva-Blanch G, Valderas-Martínez P, Medina-Remón A, Lamuela-Raventós RM, Estruch R. Wine, beer,alcohol and polyphenols on cardiovascular disease andcancer. Nutrients 2012; 4: 759–781.

Astill BD, Mills J, Rassett RL, Roundabush RL, Terhaar CJ.Fate of butylated hydroxyanisole in man and rat. J. Agric.Food Chem. 1962; 10: 315–318.

Awad HM, Boersma MG, Boeren S, Van Bladeren PJ,Vervoort J, Rietjens IM. Quenching of quercetin quinone/quinone methides by different thiolate scavengers: stabil-ity and reversibility of conjugate formation. Chem. Res.Toxicol. 2003; 16: 822–831.

Azam S, Hadi N, Khan NU, Hadi SM. Prooxidant propertyof green tea polyphenols epicatechin and epigallocatechin-3-gallate: implications for anticancer properties. Toxicol.in Vitro 2004; 18: 555–561.

Bae MJ, Ishii T, Minoda K, Kawada Y, Ichikawa T, Mori T,Kamihira M, Nakayama T. Albumin stabilizes (–)-epigallocatechin gallate in human serum: binding capac-ity and antioxidant property. Mol. Nutr. Food Res. 2009;53: 709–715.

Berhane K, Widersten M, Engstrom A, Kozarich JW,Mannervik B. Detoxication of base propenals and otheralpha, beta-unsaturated aldehyde products of radical re-actions and lipid peroxidation by human glutathionetransferases. Proc. Natl. Acad. Sci. USA 1994; 91: 1480–1484.

Bernofsky C. Nucleotide chloramines and neutrophil-medi-ated cytotoxicity. FASEB J. 1991; 5: 295–300.

Boersma MG, Vervoort J, Szymusiak H, Lemanska K,Tyrakowska B, Cenas N, Segura-Aguilar J, RietjensIMCM. Regioselectivity and reversibility of the glutath-ione conjugation of quercetin quinone methide. Chem. Res.Toxicol. 2000; 13: 185–191.

Böhl M, Czupalla C, Tokalov SV, Hoflack B, Gutzeit HO.Identification of actin as quercetin-binding protein: an



approach to identify target molecules for specific ligands.Anal. Biochem. 2005; 346: 295–299.

Bolton JL, Turnipseed SB, Thompson JA. Influence of qui-none methide reactivity on the alkylation of thiol andamino groups in proteins: studies utilizing amino acid andpeptide models. Chem. Biol. Interact. 1997; 107: 185–200.

Bolton JL, Trush MA, Penning TM, Dryhurst G, Monks TJ.Role of quinines in toxicology. Chem. Res. Toxicol. 2000;13: 135–160.

Bravo L. Polyphenols: chemistry, dietary sources, metabo-lism, and nutritional significance. Nutr. Rev. 1998; 56:317–333.

Cerutti PA, Trump BF. Inflammation and oxidative stress incarcinogenesis. Cancer Cell. 1991; 3: 1–7.

Chen C, Okayama H. High-efficiency transformation ofmammalian cells by plasmid DNA. Mol. Cell. Biol. 1987;7: 2745–2752.

Collins SJ, Gallo RC, Gallagher RE. Continuous growth anddifferentiation of human myeloid leukaemic cells insuspensionculture. Nature 1977; 270: 347–349.

Cooksey CJ, Land EJ, Rushton FAP, Ramsden CA, RileyPA. Tyrosinase-mediated cytotoxicity of 4-substitutedphenols: Use of QSAR to forecast reactivities of thiolstowards the derived ortho-quinones. Quant. Struct.-Act.Relat. 1996; 15: 498–503.

Dahlin DC, Miwa GT, Lu AY, Nelson SD. N-acetyl-p-ben-zoquinone imine: a cytochrome P-450-mediated oxidationproduct of acetaminophen. Proc. Natl. Acad. Sci. USA1984; 81: 1327–1331.

Daniel V. Glutathione S-transferases: gene structure andregulation of expression. CRC Crit. Rev. Biochem. Mol.Biol. 1993; 28: 173–207.

Dawley RM, Flurkey WH. 4-Hexylresorcinol, a potent in-hibitor of mushroom tyrosinase. J. Food. Sci. 1993; 58:609–610.

De Long MJ, Prochaska HJ, Talalay P. Induction ofNAD(P)H:quinone reductase in murine hepatoma cells byphenolic antioxidants, azo dyes, and otherchemoprotectors: a model system for the study ofanticarcinogens. Proc. Natl. Acad. Sci. USA 1986; 83: 787–791.

Diccianni MB, Imagawa M, Muramatsu M. The dyadpalindromic glutathione transferase P enhancer binds mul-tiple factors including AP1. Nucleic Acid Res. 1992; 20:5153–5158.

Dinkova-Kostova AT, Talalay P. Direct and indirect antioxi-dant properties of inducers of cytoprotective proteins. Mol.Nutr. Food Res. 2008; 52: S128–S138.

Dinkova-Kostova AT, Holtzclaw WD, Cole RN, Itoh K,Wakabayashi N, Katoh Y, Yamamoto M, Talalay P. Directevidence that sulfhydryl groups of Keap1 are the sensorsregulating induction of phase 2 enzymes that protectagainst carcinogens and oxidants. Proc. Natl. Acad. Sci.USA 2002; 99: 11908–11913.

Eggler AL, Liu G, Pezzuto JM, van Breemen RB, MesecarAD. Modifying specific cysteines of the electrophile-sens-ing human Keap1 protein is insufficient to disrupt bind-ing to the Nrf2 domain Neh2. Proc. Natl. Acad. Sci. USA2005; 102: 10070–10075.

El-Rashidy R, Niazi S. A new metabolite of butylatedhydroxyanisole in man. Biopharm. Drug Dispos. 1983; 4:389–396.

Ermakova S, Choi BY, Choi HS, Kang BS, Bode AM, DongZ. The intermediate filament protein vimentin is a newtarget for epigallocatechin gallate. J. Biol. Chem. 2005;280: 16882–16890.

Fang MZ, Wang Y, Ai N, Hou Z, Sun Y, Lu H, Welsh W,Yang CS. Tea polyphenol (–)-epigallocatechin-3-gallateinhibits DNA methyltransferase and reactivates methyla-tion-silenced genes in cancer cell lines. Cancer Res. 2003;63: 7563–7570.

Fiaschi T, Cozzi G, Raugei G, Formigli L, Ramponi G,Chiarugi P. Redox regulation of b-actin during integrin-mediated cell adhesion. J. Biol. Chem. 2006; 281: 22983–22991.

Fischer SM, Baldwin JK, Jasheway DW, Patrick KE,Cameron GS. Phorbol ester induction of 8-lipoxygenasein inbred SENCAR (SSIN) but not C57BL/6J mice corre-lated with hyperplasia, edema, and oxidant generation butnot ornithine decarboxylase induction. Cancer Res. 1988;48: 658–664.

Fischer SM, Jasheway DW, Klann RC, Butler AP, PatrickKE, Baldwin JK, Cameron GS. Correlation of phorbol esterpromotion in the resistant C57BL/6J mouse with sustainedhyperplasia, but not with ornithine decarboxylase or pro-tein kinase C. Cancer Res. 1989; 49: 6693–6699.

Fong TA, Mosmann TR. The role of IFN-g in delayed-typehypersensitivity mediated by Th1 clones. J. Immunol.1989; 143: 2887–2893.

Fraga CG. Plant polyphenols: how to translate their in vitroantioxidant actions to in vivo conditions. IUBMB Life2007; 59: 308–315.

Fratelli M, Demol H, Puype M, Casagrande S, Eberini I,Salmona M, Bonetto V, Mengozzi M, Duffieux F, MicletE, Bachi A, Vandekerckhove J, Gianazza E, Ghezzi P. Iden-tification by redox proteomics of glutathionylated proteinsin oxidatively stressed human T lymphocytes. Proc. Natl.Acad. Sci. USA 2002; 99: 3505–3510.

Galati G, O’Brien PJ. Potential toxicity of flavonoids andother dietary phenolics: significance for theirchemopreventive and anticancer properties. Free RadicalBiol. Med. 2003; 37: 287–303.

Galati G, Lin A, Sultan AM, O’Brien PJ. Cellular and invivo hepatotoxicity caused by green tea phenolic acids andcatechins. Free Radical Biol. Med. 2006; 40: 570–580.

García-Molina F, Muñoz JL, Varón R, Rodríguez-López JN,García-Cánovas F, Tudela J. A review on spectrophoto-metric methods for measuring the monophenolase anddiphenolase activities of tyrosinase. J. Agric. Food Chem.,2007; 55: 9739–9749.

Gibson JD, Pumford NR, Samokyszyn VM, Hinson JA.Mechanism of acetaminophen-induced hepatotoxicity:covalent binding versus oxidative stress. Chem. Res.Toxicol. 1996; 9: 580–585.

Gliszcynska-Swiglo A, van der Woude H, de Haan L,Tyrakowska B, Aarts, JMMJG, Rietjens IMCM. The roleof quinone reductase (NQO1) and quinone chemistry inquercetin cytotoxicity. Toxicol. in Vitro 2003; 17: 423–431.

Gorman CM, Moffat LF, Howard BH. Recombinant genomeswhich express chloramphenicol acetyltransferase in mam-malian cells. Mol. Cell. Biol. 1982; 2: 1044–1051.

Graham DG. Oxidative pathways for catecholamines in thegenesis of neuromelanin and cytotoxic quinones. Mol.



Pharmacol. 1978; 14: 633–643.Gutzeit HO, Henker Y, Kind B, Franz A. Specific interac-

tions of quercetin and other flavonoids with target pro-teins are revealed by elicited fluorescence. Biochem.Biophys. Res. Commun. 2004; 318: 490–495.

Habig WH, Jakoby WB. Assays for differentiation of glu-tathione S-transferases. Methods Enzymol. 1981; 77: 398–405.

Habig WH, Pabst MJ, Jakoby WB. Glutathione S-transferase.The first enzymatic step in mercapturic acid formation. J.Biol. Chem. 1974; 249: 7130–7139.

Halliwell B. Dietary polyphenols: Good, bad, or indifferentfor your health? Cardiovasc Res. 2007; 73: 341–347.

Hayes JD, Pulford DJ. The glutathione S-transferase super-gene family: regulation of GST and the contribution ofthe isoenzymes to cancer chemoprotection and drug re-sistance. CRC Crit. Rev. Biochem. Mol. Biol. 1995; 30:445–600.

Hearing VJ, Jiménez M. Mammalian tyrosinase, the criticalregulatory control point in melanocyte pigmentation. Int.J. Biochem. 1987; 19: 1141–1147.

Henderson CJ, Smith AG, Ure J, Brown K, Bacon EJ, WolfCR. Increased skin tumorigenesis in mice lacking pi classglutathione S-transferases. Proc. Natl. Acad. Sci. USA1998; 95: 5275–5280.

Hirakawa K, Oikawa S, Hiraku Y, Hirosawa I, Kawanishi,S. Catechol and hydroquinone have different redox prop-erties responsible for their differential DNA-damagingability. Chem. Res. Toxicol. 2002; 15: 76–82.

Hong F, Freeman ML, Liebler DC. Identification of sensorcysteines in human Keap1 modified by the cancerchemopreventive agent sulforaphane. Chem. Res. Toxicol.2005; 18: 1917–1926.

Hou Z, Sang S, You H, Lee MJ, Hong J, Chin KV, Yang CS.Mechanism of action of (–)-epigallocatechin-3-gallate:auto-oxidation-dependent inactivation of epidermalgrowth factor receptor and direct effects on growth inhi-bition in human esophageal cancer KYSE 150 cells. Can-cer Res. 2005; 65: 8049–8056.

Huang MT, Smart RC, Wong CQ, Conney AH. Inhibitoryeffect of curcumin, chlorogenic acid, caffeic acid, andferulic acid on tumor promotion in mouse skin by 12-O-tetradecanoylphorbol-13-acetate. Cancer Res. 1988; 48:5941–5946.

Hung CF, Huang TF, Chiang HS, Wu WB. (–)-Epigallocatechin-3-gallate, a polyphenolic compound fromgreen tea, inhibits fibroblast adhesion and migrationthrough multiple mechanisms. J. Cell Biochem. 2005; 96:183–197.

Hurst JK, Barrette WC. Leukocyte oxygen activation andmicrobicidal oxidative toxin. Crit. Rev. Biochem. Mol.Biol. 1989; 24: 271–328.

Imagawa M, Osada S, Okuda A, Muramatsu M. Silencerbinding proteins function on multiple cis-elements in theglutathione transferase P gene. Nucleic Acids Res. 1991;19: 5–10.

Ishii T, Tatsuda E, Kumazawa S, Nakayama T, Uchida K.Molecular basis of enzyme inactivation by an endogenouselectrophile 4-hydroxy-2-nonenal: identification of modi-fication sites in glyceraldehyde-3-phosphate dehydroge-nase. Biochemistry 2003; 42: 3474–3480.

Ishii T, Mori T, Tanaka T, Mizuno D, Yamaji R, Kumazawa

S, Nakayama T, Akagawa M. Covalent modification ofproteins by green tea polyphenol (–)-epigallocatechin-3-gallate through autoxidation. Free Radical Biol. Med.2008; 45: 1384–1394.

Ishii T, Ishikawa M, Miyoshi N, Yasunaga M, Akagawa M,Uchida K, Nakamura Y. Catechol type polyphenol is apotential modifier of protein sulfhydryls: development andapplication of a new probe for understanding the dietarypolyphenol actions. Chem. Res. Toxicol. 2009; 22: 1689–1698.

Ito N, Hirose M. Antioxidants-carcinogenic andchemopreventive properties. Adv. Cancer Res. 1989; 53:247–302.

Itoh K, Chiba T, Takahashi S, Ishii T, Igarashi K, Katoh Y,Oyake T, Hayashi N, Satoh K, Hatayama I, Yamamoto M,Nabeshima Y. An Nrf2/small Maf heterodimer mediatesthe induction of phase II detoxifying enzyme genesthrough antioxidant response elements. Biochem. Biophys.Res. Commun. 1997; 236: 313–322.

Itoh K, Ishii T, Wakabayashi N, Yamamoto M. Regulatorymechanisms of cellular response to oxidative stress. FreeRadic. Res. 1999a; 31: 319–324.

Itoh K, Wakabayashi N, Katoh Y, Ishii T, Igarashi K, EngelJD, Yamamoto M. Keap1 represses nuclear activation ofantioxidant responsive elements by Nrf2 through bindingto the amino-terminal Neh2 domain. Genes Dev. 1999b;13: 76–86.

Itoh K, Wakabayashi N, Katoh Y, Ishii T, O’Connor T,Yamamoto M. Keap1 regulates both cytoplasmic-nuclearshuttl ing and degradation of Nrf2 in response toelectrophiles. Genes Cells 2003; 8: 379–391.

Izumi Y, Sawada H, Yamamoto N, Kume T, Katsuki H,Shimohama S, Akaike A. Iron accelerates the conversionof dopamine-oxidized intermediates into melanin and pro-vides protection in SH-SY5Y cells. J. Neurosci. Res. 2005;82: 126–137.

Ji C, Marnett LJ. Oxygen radical-dependent epoxydation of(7S,8S)-dihydroxy-7,8-dihydrobenzo[a]pyrene in mouseskin in vivo. Stimulation by phorbol esters and inhibitionby antiinflammatory steroids. J. Biol. Chem. 1992; 267:17842–17878.

Kalergis AM, Lopez CB, Becker MI, Diaz MI, Sein J,Garbarino JA, De Ioannes AE. Modulation of fatty acidoxidation alters contact hypersensitivity to urushiols: roleof aliphatic chain beta-oxidation in processing and acti-vation of urushiols. J. Invest. Dermatol. 1997; 108: 57–61.

Kampa M, Nifli AP, Notas G, Castanas E. Polyphenols andcancer cell growth. Rev. Physiol. Biochem. Pharmacol.2007; 159: 79–113.

Kawamoto Y, Nakamura Y, Naito Y, Torii Y, Kumagai T,Osawa T, Ohigashi H, Satoh K, Imagawa M, Uchida K.Cyclopentenone prostaglandins as potential inducers ofphase II detoxification enzymes. 15-deoxy-D12,14-pros-taglandin J2-induced expression of glutathione S-transferases. J. Biol. Chem. 2000; 275: 11291–11299.

Kensler TW, Egner PA, Moore KG, Taffe BG, Twerdok LE,Trush MA. Role of inflammatory cells in the metabolicactivation of polycyclic aromatic hydrocarbons in mouseskin. Toxicol. Appl. Pharmacol. 1987; 90: 337–346.

Kensler TW, Egner PA, Taffe BG, Trush MA. Role of freeradicals in tumour promotion and progression. In: Slaga



TJ, Klein-Szanto AJP, Boutwell RK, Stevenson DE,Spitzer HL, D’Motto B (eds). Skin Carcinogenesis, Mecha-nisms and Human Relevance. Alan R. Liss, Inc., New York.1989; 233–248.

Kim MC, Pratt DE. Thermal degradation of phenolic anti-oxidants. In: Ho CT, Lee CY, Huang MT (eds). PhenolicCompounds in Food and Their Effects on Health II. Ameri-can Chemical Society, Washington, D.C. 1992; 200–218.

Klebanoff SS. Myeloperoxidase: occurrence and biologicalfunction. In: Everse J, Everse KE, Grisham MB (eds).Peroxidases in Chemistry and Biology. CRC Press, BocaRoton. 1991; 1–35.

Koga S, Nakano M, Tero-Kubota S. Generation of superoxideduring the enzymatic action of tyrosinase. Arch. Biochem.Biophys. 1992; 292: 570–575.

Kong ANT, Yu R, Chen C, Mandlekar S, Primiano T. Signaltransduction events elicited by natural products: role ofMAPK and caspase pathways in homeostatic response andinduction of apoptosis. Arch. Pharm. Res. 2000; 23: 1–16.

Laemmuli UK. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1970;227: 680–685.

Lahti A, Pylvanen V, Hannuksela M. Immediate irritant re-actions to benzoic acid are enhanced in washed skin ar-eas. Contact Dermatitis 1995; 33: 177–182.

Lam LK, Garg P. Tumorigenicity of di-tert-butyl-substitutedhydroquinone and hydroxyanisoles in the forestomach ofSyrian golden hamsters. Carcinogenesis 1991; 12: 1341–1344.

Lassing I, Schmitzberger F, Björnstedt M, Holmgren A,Nordlund P, Schutt CE, Lindberg U. Molecular and struc-tural basis for redox regulation of b-actin. J. Mol. Biol.2007; 370: 331–348.

LaVoie MJ, Ostaszewski BL, Weihofen A, SchlossmacherMG, Selkoe DJ. Dopamine covalently modifies and func-tionally inactivates parkin. Nat. Med. 2005; 11: 1214–1221.

Lee JM, Moehlenkamp JD, Hanson JM, Johnson JA. Nrf2-dependent activation of the antioxidant responsive elementby tert-butylhydroquinone is independent of oxidativestress in IMR-32 human neuroblastoma cells. Biochem.Biophys. Res. Commun. 2001; 280: 286–292.

Lee-Hilz YY, Boerboom AM, Westphal AH, Berkel WJ, AartsJM, Rietjens IM. Pro-oxidant activity of flavonoids in-duces EpRE-mediated gene expression. Chem. Res.Toxicol. 2006; 19: 1499–1505.

Li P, Wang XQ, Wang HZ, Wu YN. High performance liq-uid chromatographic determination of phenolic acids infruits and vegetables. Biomed. Environ. Sci. 1993; 6: 389–398.

Li Y, Trush MA, Yager JD. DNA damage caused by reactiveoxygen species originating from a copper-dependent oxi-dation of the 2-hydroxy catechol of estradiol. Carcino-genesis 1994; 15: 1421–1427.

Loecken EM, Guengerich FP. Reactions of glyceraldehyde3-phosphate dehydrogenase sulfhydryl groups with bis-electrophiles produce DNA-protein cross-links but notmutations. Chem. Res. Toxicol. 2008; 21: 453–458.

Long LH, Clement MV, Halliwell B. Artifacts in cell cul-ture: rapid generation of hydrogen peroxide on additionof (–)-epigallocatechin, (–)-epigallocatechin gallate, (+)-

catechin, and quercetin to commonly used cell culturemedia. Biochem. Biophys. Res. Commun. 2000; 25: 50–53.

Lu YP, Lou YR, Yen P, Newmark HL, Mirochnitchenko OI,Inouye M, Huang MT. Enhanced skin carcinogenesis intransgenic mice with high expression of glutathione per-oxidase or both glutathione peroxidase and superoxidedismutase. Cancer Res. 1997; 57: 1468–1474.

Maeda K, Fukuda M. Arbutin: Mechanism of i tsdepigmenting action in human melanocyte culture. J.Pharmacol. Exp. Ther. 1996; 276: 765–769.

Martin KR, Appel CL. Polyphenols as dietary supplements:A double-edged sword. Nutr. Diet. Supple. 2010; 2: 1–12.

Maruyama H, Williams GM. Hepatotoxicity of chronic highdose administration of acetaminophen to mice. A criticalreview and implications for hazard assessment. Arch.Toxicol. 1988; 62: 465–469.

Mayer AM. Polyphenol oxidases in plants: recent progress.Phytochemistry 1987; 26: 11–20.

Mennen LI, Walker R, Bennetau-Pelissero C, Scalbert A.Risks and safety of poyphenol consumption. Am. J. Clin.Nutr. 2005; 81: 326S–329S.

Merckel F, Bernard G, Mutschler J, Giménez-Arnau E,Gerberick GF, Lepoittevin JP. Effect of a microemulsionsystem on hapten-peptide reactivity studies: examples ofhydroxycitronellal and citral, fragrance skin sensitizers,with glutathione. Chem. Res. Toxicol. 2010; 23: 1433–1441.

Meyers DG, Maloley PA, Weeks D. Safety of antioxidantvitamins. Arch. Int. Med. 1996; 156: 925–935.

Migdal C, Botton J, El Ali Z, Azoury ME, Guldemann J,Giménez-Arnau E, Lepoittevin JP, Kerdine-Römer S,Pallardy M. Reactivity of chemical sensitizers towardamino acids in cellulo plays a role in the activation of theNrf2-ARE pathway in human monocyte dendritic cells andthe THP-1 cell line. Toxicol. Sci. 2013; 133: 259–274.

Miglio C, Chiavaro E, Visconti A, Fogliano V, Pellegrini N.Effects of different cooking methods on nutritional andphysicochemcial characteristics of selected vegetables. J.Agric. Food Chem. 2008; 56: 139–147.

Mills KJ, Smart RC. Comparison of epidermal protein ki-nase C activity, ornithine decarboxylase induction andDNA synthesis stimulated by TPA or dioctanoylglycerolin mouse strains with different susceptibility to TPA-in-duced tumor promotion. Carcinogenesis 1989; 10: 833–838.

Mitchell JR, Jollow DJ, Potter WZ, Davis DC, Jillette JR,Brodie BB. Acetaminophen-induced hepatic necrosis. I.Role of drug metabolism. J. Pharmcol. Exp. Ther. 1973;187: 185–194.

Monks TJ, Lau SS. Biological reactivity of polyphenolic-glutathione conjugates. Chem. Res. Toxicol. 1997; 10:1296–1313.

Monks TJ, Highet RJ, Lau SS. 2-Bromo-(diglutathion-S-yl)hydroquinone nephrotoxicity: physiological, biochemi-cal, and electrochemical determinants. Mol. Pharmacol.1988; 34: 492–500.

Monks TJ, Hanzlik RP, Cohen GM, Ross D, Graham DG.Comtemporary issues in toxicology. Quinone chemistryand toxicology. Toxicol. Appl. Pharmacol. 1992; 112: 2–16.

Monks TJ, Lo HH, Lau SS. Oxidation and acetylation as



determinants of 2-bromocystein-S-ylhydroquinone-medi-ated nephrotoxicity. Chem. Res. Toxicol. 1994; 7: 495–502.

Morgan BDG. Recent results of a clinical pilot study of in-tra-arterial 4-hydroxyanisole chemotherapy in malignantmelanoma. In: Riley PA (ed). Hydroxyanisole: RecentAdvances in Anti-melanoma Therapy. IRL Press Ltd.,Oxford. 1984; 233–241.

Mori T, Ishii T, Akagawa M, Nakamura Y, Nakayama T.Covalent binding of tea catechins to protein thiols: therelationship between stability and electrophilic reactivity.Biosci. Biotechnol. Biochem. 2010; 74: 2451–2456.

Moridani MY, Cheon SS, Khan S, O’Brien PJ. Metabolicactivation of 4-hydroxyanisole by isolated rat hepatocytes.Drug Metab. Dispos. 2002; 30: 1063–1069.

Morimitsu Y, Nakagawa Y, Hayashi K, Fujii H, Kumagai T,Nakamura Y, Osawa T, Horio F, Itoh K, Iida K, YamamotoM, Uchida K. A sulforaphane analogue that potently acti-vates the Nrf2-dependent detoxification pathway. J. Biol.Chem. 2002; 277: 3456–3463.

Muñoz-Muñoz JL, Garcia-Molina F, Varon R, Garcia-RuízPA, Tudela J, Garcia-Cánovas F, Rodríguez-López JN.Suicide inactivation of the diphenolase and monophenolaseactivities of tyrosinase. IUBMB Life 2010; 62: 539–547.

Murakami A, Nakamura Y, Koshimizu K, Ohigashi H.Glyceroglycolipids from Citrus hystrix, a traditional herbin Thailand, potently inhibits the tumor promoting activ-ity of 12-O-tetradecanoylphorbol-13-acetate in mouseskin. J. Agric. Food Chem. 1995; 43: 2779–2783.

Murakami A, Ohura S, Nakamura Y, Koshimizu K, OhigashiH. 1¢-Acetoxychavicol acetate, a superoxide anion gen-eration inhibitor, potently inhibits tumor promotion by 12-O-tetradecanolyphorbol-13-acetate in ICR mouse skin.Oncology 1996; 53: 386–391.

Murakami A, Nakamura Y, Tanaka T, Kawabata K, TakahashiD, Koshimizu K, Ohigashi H. Suppression by sitrusauraptene of phorbol ester- and endotoxin-induced inflam-matory responses: role of attenuation of leukocyte activa-tion. Carcinogenesis 2000a; 21: 1843–1850.

Murakami A, Nakamura Y, Torikai K, Tanaka T, Koshiba T,Koshimizu K, Kuwahara S, Takahashi Y, Ogawa K, YanoM, Tokuda H, Nishino H, Mikami Y, Sashida Y, KitanakaS, Ohigashi H. Inhibitory effect of citrus nobiletin onphorbol ester-induced skin inflammation, oxidative stress,and tumor promotion in mice. Cancer Res. 2000b; 60:5059–5066.

Na HK, Surh YJ. Transcriptional regulation via cysteine thiolmodification: a novel molecular strategy forchemoprevention and cytoprotection. Mol. Carcinog.2006; 45: 368–380.

Naasani I, Oh-Hashi F, Oh-Hara T, Feng WY, Johnston J,Chan K, Tsuruo T. Blocking telomerase by dietarypolyphenols is a major mechanism for limiting the growthof human cancer cells in vitro and in vivo. Cancer Res.2003; 63: 824–830.

Nakae D, Mizumoto Y, Yoshiji H, Andoh N, Horiguchi K,Shiraiwa K, Kobayashi E, Endoh T, Shimoji N, TamuraK, Tsujiuchi T, Denda A, Konishi Y. Different roles of 8-hydroxyguanine formation and 2-thiobarbituric acid-re-acting substance generation in the early phase of liver car-cinogenesis induced by a choline-deficient, L-amino acid-defined diet in rats. Jpn. J. Cancer Res. 1994; 85: 499–

505.Nakae D, Denda A, Kobayashi Y, Akai H, Kishida H,

Tsujiuchi T, Konishi Y, Suzuki T, Muramatsu M. Inhibi-tion of early-phase exogenous and endogenous liver car-cinogenesis in transgenic rats harboring a rat glutathioneS-transferase placental form gene. Jpn. J. Cancer Res.1998; 89: 1118–1125.

Nakamura Y, Miyoshi N. Electrophiles in foods: the currentstatus of isothiocyanates and their chemical biology.Biosci. Biotechnol. Biochem. 2010; 74: 242–255.

Nakamura Y, Murakami A, Koshimizu K, Ohigashi H. In-hibitory effects of pheophorbide a, a chlorophyll-relatedcompound, on skin tumour promotion in ICR mouse skin.Cancer Lett. 1996; 108: 247–255.

Nakamura Y, Murakami A, Ohto Y, Torikai K, Tanaka T,Ohigashi H. Suppression of tumor promoter-inducedoxidative stress and inflammatory responses in mouse skinby a superoxide generation inhibitor 1¢-acetoxychavicolacetate. Cancer Res. 1998a; 58: 4832–4839.

Nakamura Y, Ohto Y, Murakami A, Ohigashi H. Superoxidescavenging activity of rosmarinic acid from Perillafrutescens Britton var. acuta f. viridis. J. Agric. FoodChem. 1998b; 46: 4545–4550.

Nakamura Y, Ohto Y, Murakami A, Osawa T, Ohigashi H.Inhibitory effects of curcumin and tetrahydrocurcuminoidson the tumor promoter-induced reactive oxygen speciesgeneration in leukocytes in vitro and in vivo. Jpn. J. Can-cer Res. 1998c; 89: 361–370.

Nakamura Y, Kawamoto N, Ohto Y, Torikai K, MurakamiA, Ohigashi H. A diacetylenic spiroketal enol ether epox-ide, AL-1, from Artemisia lactiflora inhibits 12-O-tetradecanoylphorbol-13-acetate-induced tumor promotionpossibly by suppression of oxidative stress. Cancer Lett.1999; 140: 37–45.

Nakamura Y, Morimitsu Y, Uzu T, Ohigashi H, MurakamiA, Naito Y, Nakagawa Y, Osawa T, Uchida K. A glutath-ione S-transferase inducer from papaya: rapid screening,identification and structure-activity relationship ofisothiocyanates. Cancer Lett. 2000a; 157: 193–200.

Nakamura Y, Murakami A, Ohigashi H. Search for naturally-occurring antioxidative chemopreventors on the basis ofthe involvement of leukocyte-derived reactive oxygenspecies in carcinogenesis. Asian Pacif. J. Cancer Prev.2000b; 1: 115–120.

Nakamura Y, Ohigashi H, Masuda S, Murakami A, MorimitsuY, Kawamoto Y, Osawa T, Imagawa M, Uchida K. Redoxregulation of glutathione S-transferase induction by ben-zyl isothiocyanate: correlation of enzyme induction withthe formation of reactive oxygen intermediates. CancerRes. 2000c; 60: 219–225.

Nakamura Y, Torikai K, Ohto Y, Murakami A, Tanaka T,Ohigashi H. A simple phenolic antioxidant protocatechuicacid enhances tumor promotion and oxidative stress infemale ICR mouse skin: dose- and timing-dependent en-hancement and involvement of bioactivation by tyrosinase.Carcinogenesis 2000d; 21: 1899–1907.

Nakamura Y, Torikai K, Ohigashi H. A catechol antioxidantprotocatechuic acid potentiates inflammatory leukocyte-derived oxidative stress in mouse skin via a tyrosinasebioactivation pathway. Free Radical Biol. Med. 2001a; 30:967–978.

Nakamura Y, Torikai K, Ohigashi H. Toxic dose of a simple



phenolic antioxidant, protocatechuic acid, attenuates theglutathione level in ICR mouse liver and kidney. J. Agric.Food Chem. 2001b; 49: 5674–5678.

Nakamura Y, Feng Q, Kumagai T, Torikai K, Ohigashi H,Osawa T, Noguchi N, Niki E, Uchida K. Ebselen, a glu-tathione peroxidase mimetic seleno-organic compound, asa multifunctional antioxidant. Implication for inflamma-tion-associated carcinogenesis. J. Biol. Chem. 2002a; 277:2687–2694.

Nakamura Y, Kawakami M, Yoshihiro A, Miyoshi N,Ohigashi H, Kawai K, Osawa T, Uchida K. Involvementof the mitochondrial death pathway in chemopreventivebenzyl isothiocyanate-induced apoptosis. J. Biol. Chem.2002b; 277: 8492–8499.

Nakamura Y, Kumagai T, Yoshida C, Naito Y, Miyamoto M,Ohigashi H, Osawa T, Uchida K. Pivotal role ofelectrophilicity in glutathione S-transferase induction bytert-butylhydroquinone. Biochemistry 2003a; 42: 4300–4309.

Nakamura Y, Watanabe S, Miyake N, Kohno H, Osawa T.Dihydrochalcones: evaluation as novel radical scaveng-ing antioxidants. J. Agric. Food Chem. 2003b; 51: 3309–3312.

Nakamura Y, Ishii T, Abe N, Murata Y. Thiol modificationby bioactivated polyphenols and its potential role in skininflammation. Biosci. Biotechnol. Biochem. 2014; 78:1067–1070.

Nakayama T, Ichiba M, Kuwabara M, Kajiya K, KumazawaS. Mechanisms and structural specificity of hydrogen per-oxide formation during oxidation of catechins. Food Sci.Technol. Res. 2002; 8: 261–267.

Nishino H, Iwashima A, Fujiki H, Sugimura T. Inhibition byquercetin of the promoting effect of teleocidin on skinpapilloma formation in mice init iated with 7,12-dimethylbenz[a]anthracene. Jpn. J. Cancer Res. 1984; 75:113–116.

Ogawa K, Futakuchi M, Hirose M, Boonyaphiphat P,Mizoguchi Y, Miki T, Shirai T. Stage and organ depend-ent effects of 1-O-hexyl-2,3,5-trimethylhydroquinone,ascorbic acid derivatives, N-heptadecane-8,10-dione andphenylethyl isothiocyanate in a rat multiorgan carcinogen-esis model. Int. J. Cancer 1998; 76: 851–856.

Ohnishi M, Tanaka T, Makita H, Kawamori T, Mori H, SatohK, Hara A, Murakami A, Ohigashi H, Koshimizu K.Chemopreventive effect of a xanthine oxidase inhihitor,1¢-acetoxychavicol acetate, on rat oral carcinogenesis. Jpn.J. Cancer Res. 1996; 87: 349–356.

Ohno K, Fukushima M, Fujiwara M, Narumiya S. Inductionof 68,000-dalton heat shock proteins by cyclopentenoneprostaglandins. Its association with prostaglandin-inducedG1 block in cell cycle progression. J. Biol. Chem. 1988;263: 19764–19770.

Okubo T, Nagai F, Ushiyama K, Kano I. Contribution ofoxygen radicals to DNA cleavage by quinone compoundsderived from phenolic antioxidants, tert-butylhydroquinone and 2,5-di-tert-butylhydroquinone.Toxicol. Lett. 1997; 90: 11–18.

Okuda A, Imagawa M, Maeda Y, Sakai M, Muramatsu M.Structural and functional analysis of an enhancer GPEIhaving a phorbol 12-O-tetradecanoate 13-acetate respon-sive element-like sequence found in the rat glutathionetransferase P gene. J. Biol. Chem. 1989; 264: 16919–

16926.Okuda A, Sakai M, Muramatsu M. The structure of the rat

glutathione S-transferase P gene and related pseudogenes.J. Biol. Chem. 1987; 262: 3858–3863.

Park BK, Pirmohamed M, Kitteringham NR. Role of drugdisposition in drug hypersensitivity: A chemical, molecu-lar, and clinical perspective. Chem. Res. Toxicol. 1998;11: 970–988.

Pastore A, Tozzi G, Gaeta LM, Bertini E, Serafini V, DiCesare S, Bonetto V, Casoni F, Carrozzo R, Federici G,Piemonte F. Actin glutathionylation increases in fibroblastsof patients with Friedreich’s ataxia: a potential role in thepathogenesis of the disease. J. Biol. Chem. 2007; 278:42588–42595.

Paz MA, Fluchiger R, Boak A, Kagan HM, Gallop PM. Spe-cific detection of quinoproteins by redox-cycling stain-ing, J. Biol. Chem. 1991; 266: 689–692.

Peters MM, Lau SS, Dulik D, Murphy D, van Ommen B,van Bladeren PJ, Monks TJ. Metabolism of tert-butylhydroquinone to S-substituted conjugates in the maleFischer 344 rat. Chem. Res. Toxicol. 1996; 9: 133–139.

Peters MM, Rivera MI, Jones TW, Monks TJ, Lau SS. Glu-tathione conjugates of tert-butyl-hydroquinone, ametabolite of the urinary tract tumor promoter 3-tert-butyl-hydroxyanisole, are toxic to kidney and bladder. CancerRes. 1996; 56: 1006–1011.

Petruska JM, Mosebrook DR, Jakab GJ, Trush MA.Myeloperoxidase-enhanced formation of (+–)-trans-7,8-dihydroxy-7,8-dihydrobenzo[a]pyrene-DNA adducts inlung tissue in vitro: a role of pulmonary inflammation inthe bioactivation of a procarcinogen. Carcinogenesis 1992;13: 1075–1081.

Pinkus R, Weiner LM, Daniel V. Role of oxidants and anti-oxidants in the induction of AP-1, NF-kappaB, and glu-tathione S-transferase gene expression. J. Biol. Chem.1996; 271: 13422–13429.

Pollard TD, Blanchoin L, Mullins RD. Molecular mecha-nisms controlling actin filament dynamics in nonmusclecells. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 545–576.

Pound PA. Further observations concerning the influence ofpreliminary stimulation by croton oil and acetic acid onthe intiation of skin tumours in mice by urethan. Brit. J.Cancer 1966; 8: 666–671.

Prestera T, Zhang Y, Spencer SR, Wilczak CA, Talalay P.The electrophile counterattack response: protection againstneoplasia and toxicity. Adv. Enzyme Regul. 1993; 33: 281–296.

Primiano T, Sutter TR, Kensler TW. Redox regulation ofgenes that protect against carcinogens. Comp. Biochem.Physiol. 1998; 118B: 487–497.

Prochaska HJ, De Long MJ, Talalay P. On the mechanismsof induction of cancer-protective enzymes: a unifying pro-posal. Proc. Natl. Acad. Sci. USA 1985; 82: 8232–8236.

Ramos CL, Pou S, Britigan BE, Cohen MS, Rosen GM. Spintrapping evidence for myeloperoxidase-dependent hy-droxyl radical formation by human neutrophils andmonocytes. J. Biol. Chem. 1992; 267: 8307–8312.

Reiners JJ Jr, Kodari E, Cappel RE, Gilbert HF. Assessmentof the antioxidant/prooxidant status of murine skin fol-lowing topical treatment with 12-O-tetradecanoylphorbol-13-acetate and throughout the ontogeny of skin cancer.



Part II: quantitation of glutathione and glutathionedisulfide. Carcinogenesis 1991; 12: 2345–2352.

Renaud S, de Lorgeril M. Wine, alcohol, platelets, and thefrench paradox for coronary heart disease. Lancet 1992;339: 1523–1526.

Rhee SG. Cell signaling. H2O2, a necessary evil for cellsignaling. Science 2006; 312: 1882–1883.

Riley PA. Hydroxyanisole depigmentation: in-vitro studies.J. Pathol. 1969; 97: 193–206.

Riley PA, Cooksey CJ, Johnson CJ, Land EJ, Latter AM,Ramsden CA. Melanogenesis-targeted anti-melanoma pro-drug development: Effect of side-chain variations on thecytotoxicity of tyrosinase-generated ortho-quinones in amodel screening system. Eur. J. Cancer 1997; 33: 135–143.

Roberts DW, Bucci TJ, Benson RW, Warbritton AR, McRaeTA, Pumford NR, Hinson JA. Immunohistochemical lo-calization and quantification of the 3-(cystein-S-yl)-aceta-minophen protein adduct in acetaminophen hepatotoxic-ity. Am. J. Pathol. 1991; 138: 359–371.

Robertson IGC, Guthenberg C, Mannervik B, Jernström B.Differences in stereoselectivity and catalytic efficiency ofthree human glutathione transferases in the conjugationof glutathione with 7b,8a-dihydroxy-9a,10a-oxy-7,8,9,10-tetrahydrobenzo(a)pyrene. Cancer Res. 1986a;46: 2220–2224.

Robertson IGC, Jensson H, Mannervik B, Jernström B. Glu-tathione transferases in rat lung: the presence of transferase7-7, highly efficient in the conjugation of glutathione withthe carcinogenic (+)-7b,8a-dihydroxy-9a,10a-oxy-7,8,9,10-tetrahydrobenzo[a]pyrene. Carcinogenesis1986b; 7: 295–299.

Rushmore TH, Pickett CB. Transcriptional regulation of therat glutathione S-transferase Ya subunit gene. Characteri-zation of a xenobiotic-responsive element controlling in-ducible expression by phenolic antioxidants. J. Biol. Chem.1990; 265: 14648–14653.

Rustin GJ, Stratford MR, Lamont A, Bleehen N, Philip PS,Howells N, Watfa RR, Slack JA. Phase I study of intrave-nous 4-hydroxyanisole. Eur. J. Cancer 1992; 28A: 1362–1364.

Saito S, Kawabata J. Synergistic effects of thiols and amineson antiradical efficiency of protocatechuic acid. J. Agric.Food Chem. 2004; 52: 8163–8168.

Sakai M, Okuda A, Muramatsu M. Multiple regulatory ele-ments and phorbol 12-O-tetradecanoate 13-acetate respon-siveness of the rat placental glutathione transferase gene.Proc. Natl. Acad. Sci. USA 1988; 85: 9456–9460.

Satué-Gracia MT, Andrés-Lacueva C, Lamuela-RaventósRM, Frankel EN. Spanish sparkling wines (Cavas) as in-hibitors of in vitro human low-density lipoprotein oxida-tion. J. Agric. Food Chem. 1999; 47: 2198–2202.

Sazuka M, Itoi T, Suzuki Y, Odani S, Koide T, Isemura M.Evidence for the interaction between (–)-epigallocatechingallate and human plasma proteins fibronectin, fibrino-gen, and histidine-rich glycoprotein. Biosci. Biotechnol.Biochem. 1996; 60: 1317–1319.

Scalbert A, Manach C, Morand C, Rémésy C, Jiménez L.Dietary polyphenols and the prevention of diseases. Crit.Rev. Food Sci. Nutr. 2005; 45: 287–306.

Sergediene E, Jönsson K, Szymusiak H, Tyrakowska B,Rietjens IM, Cenas N. Prooxidant toxicity of polyphenolic

antioxidants to HL-60 cells: description of quantitativestructure-activity relationships. FEBS Lett. 1999; 462:392–396.

Shen XM, Dryhurst G. Further insights into the influence ofL-cysteine on the oxidation chemistry of dopamine: reac-tion pathways of potential relevance to Parkinson’s dis-ease. Chem. Res. Toxicol. 1996; 9: 751–763.

Shibata T, Yamada T, Ishii T, Kumazawa S, Nakamura H,Masutani H, Yodoi J, Uchida K. Thioredoxin as a molecu-lar target of cyclopentenone prostaglandins. J. Biol. Chem.2003; 278: 26046–26054.

Sisskin EE, Gray T, Barrett JC. Correlation between sensi-tivity to tumor promotion and sustained epidermal hyper-plasia of mice and rats treated with 12-O-tetradecanoylphorbol-13-acetate. Carcinogenesis 1982; 3:403–407.

Slaga TJ, Fischer SM. Strain differences and solvent effectsin mouse skin carcinogenesis experiments using carcino-gens, tumor initiators and promoters. Prog. Exp. TumorRes. 1983; 26: 85–109.

Spencer SR, Xue LA, Klenz EM, Talalay P. The potency ofinducers of NAD(P)H:(quinone-acceptor) oxidoreductaseparallels their efficiency as substrates for glutathionetransferases. Structural and electronic correlations.Biochem. J. 1991; 273: 711–717.

Steineck MJ, Khan AU, Karnovsky MJ. Intracellular singletoxygen generation by phagocytosing neutrophils in re-sponse to particles coated with a chemical trap. J. Biol.Chem. 1992; 267: 13425–13433.

Sugioka Y, Fujii-Kuriyama Y, Kitagawa T, Muramatsu M.Cloning and the nucleotide sequence of rat glutathione S-transferase P cDNA. Nucleic Acid Res. 1985; 13: 6049–6057.

Tachibana H, Koga K, Fujimura Y, Yamada K. A receptorfor green tea polyphenol EGCG. Nature Struct. Mol. Biol.2004; 11: 380–381.

Talalay P, Benson AM. Elevation of quinone reductase ac-tivity by anticarcinogenic antioxidants. Adv. EnzymeRegul. 1982; 20: 287–300.

Talalay P, Dinkova-Kostova AT, Holtzclaw WD. Importanceof phase 2 gene regulation in protection againstelectrophile and reactive oxygen toxicity and carcinogen-esis. Adv. Enzyme Regul. 2003; 43: 121–134.

Tanaka T, Kojima T, Suzui M, Mori H. Chemoprevention ofcolon carcinogenesis by the natural product of a simplephenolic compound protocatechuic acid: suppressing ef-fects on tumor development and biomarkers expressionof colon tumorigenesis. Cancer Res. 1993; 53: 3908–3913.

Tanaka T, Kawamori T, Ohnishi M, Okamoto K, Mori H,Hara A. Chemoprevention of 4-nitroquinoline 1-oxide-induced oral carcinogenesis by dietary protocatechuic acidduring initiation and postinitiation phases. Cancer Res.1994; 54: 2359–2365.

Tanaka T, Kawabata K, Kakumoto M, Makita H, MatsunagaK, Mori H, Satoh K, Hara A, Murakami A, Ohigashi H,Koshimizu K. Chemoprevention of azoxymethane-inducedrat colon carcinogenesis by a xanthine oxidase inhibitor,1¢-acetoxychavicol acetate. Jpn. J. Cancer Res. 1997a; 88:821–830.

Tanaka T, Kawabata K, Kakumoto M, Makita H, MatsunagaK, Mori H, Satoh K, Hara A, Murakami A, Ohigashi H,Koshimizu K. Chemoprevention of azoxymethane-induced



rat colon carcinogenesis by a xanthine oxidase inhibitor,1¢-acetoxychavicol acetate. Jpn. J. Cancer Res. 1997b; 88:821–830.

Thompson BY, Sivam G, Britigan BE, Rosen GM, CohenMS. Oxygen metabolism of the HL-60 cell line: compari-son of the effects of monocytoid and neutrophilic differ-entiation. J. Leukoc. Biol. 1988; 43: 140–147.

Thompson D, Constantin-Teodosiu D, Norbeck K, SvenssonB, Moldéus P. Metabolic activation of eugenol bymyeloperoxidase and polymorphonuclear leukocytes.Chem. Res. Toxicol. 1989; 2: 186–192.

Thurston CF. The structure and function of fungal laccases.Microbiology 1994; 140: 19–26.

Tseng TH, Hsu JD, Lo MH, Chu CY, Chou FP, Huang CL,Wang CJ. Inhibitory effect of Hibiscus protocatechuic acidon tumor promotion in mouse skin. Cancer Lett. 1998;126: 199–207.

Tsuda T, Ohshima K, Kawakishi S, Osawa T. Oxidation prod-ucts of cyanidin 3-O-beta-D-glucoside with a free radicalinitiator. Lipids 1996; 31: 1259–1263.

Tsuda T, Shiga K, Ohshima K, Kawakishi S, Osawa T. Inhi-bition of lipid peroxidation and the active oxygen radicalscavenging effect of anthocyanin pigments isolated fromPhaseolus vulgaris L. Biochem. Pharmcol. 1996; 52:1033–1039.

Tsuda, T, Horio F, Osawa T. Absorption and metabolism ofcyanidin 3-O-beta-D-glucoside in rats. FEBS Lett. 1999;449: 179–182.

Tsuji RF, Geba GP, Wang Y, Kawamoto K, Matis LA,Askenase PW. Required early complement activation incontact sensitivity with generation of local C5-dependentchemotactic activity, and late T cell interferon gamma: apossible initiating role of B cells. J. Exp. Med. 1997; 186:1015–1026.

Ueda J, Saito N, Shimazu Y, Ozawa T. A comparison of scav-enging abilities of antioxidants against hydroxyl radicals.Arch. Biochem. Biophys. 1996; 333: 377–384.

van Ommen B, Koster A, Verhagen H, van Bladeren PJ. Theglutathione conjugates of tert-butyl hydroquinone as po-tent redox cycling agents and possible reactive agentsunderlying the toxicity of butylated hydroxyanisole.Biochem. Biophys. Res. Commun. 1992; 189: 309–314.

van Zanden JJ, Ben Hamman O, van Iersel ML, Boeren S,Cnubben NH, Lo Bello M, Vervoort J, van Bladeren PJ,Rietjens IM. Inhibition of human glutathione S-transferaseP1-1 by the flavonoid quercetin. Chem. Biol. Interact.2003; 145: 139–148.

Venugopal R, Jaiswal AK. Nrf1 and Nrf2 positively and c-Fos and Fra1 negatively regulate the human antioxidantresponse element-mediated expression ofNAD(P)H:quinone oxidoreductase1 gene. Proc. Natl.Acad. Sci. USA 1996; 93: 14960–14965.

Verhagen H, Thijssen HH, ten Hoor F, Kleinjans JC. Dispo-sition of single oral doses of butylated hydroxyanisole inman and rat. Food Chem. Toxicol. 1989; 27: 151–158.

Wakabayashi N, Dinkova-Kostova AT, Holtzclaw WD, KangMI, Kobayashi A, Yamamoto M, Kensler TW, Talalay, P.

Formation of disulfide bond in p53 correlates with inhibi-tion of DNA binding and tetramerization. Proc. Natl. Acad.Sci. USA 2004; 101: 2040–2045.

Wang XJ, Sun Z, Villeneuve NF, Zhang S, Zhao F, Li Y,Chen W, Yi X, Zheng W, Wondrak GT, Wong PK, ZhangDD.Nrf2 enhances resistance of cancer cells to chemo-therapeutic drugs, the dark side of Nrf2. Carcinogenesis2008; 29: 1235–1243.

Wei H, Frenkel K. In vivo formation of oxidized DNA basesin tumor promoter-treated mouse skin. Cancer Res. 1991;51: 4443–4449.

Wei H, Frenkel K. Suppression of tumor promoter-inducedoxidative events and DNA damage in vivo by sarcophytolA: a possible mechanism of antipromotion. Cancer Res.1992; 52: 2298–2303.

Wei H, Frenkel K. Relationship of oxidative events and DNAoxidation in SENCAR mice to in vivo promoting activityof phorbol ester-type tumor promoters. Carcinogenesis1993; 14: 1195–1201.

Wei L, Wei H, Frenkel K. Sensitivity to tumor promotion ofSENCAR and C57BL/6J mice correlates with oxidativeevents and DNA damage. Carcinogenesis 1993; 14: 841–847.

Williams GM, Iatropoulos MJ. Inhibition by acetaminophenof intestinal cancer in rats induced by an aromatic aminesimilar to food mutagens. Eur. J. Cancer Prev. 1997; 6:357–362.

Woo JT, Nakagwa H, Notoya M, Yonezawa T, Udagawa N,Lee IS, Ohnishi M, Hagiwara H, Nagai K. Quercetin sup-presses bone resorption by inhibiting the differentiationand activation of osteoclasts. Biol. Pharm. Bull. 2004; 27:504–509.

Yamada M, Okigaki T, Awai M. Adhesion and growth of ratliver epithelial cells on an extracellular matrix with pro-teins from fibroblast conditioned medium. Cell Struct.Funct. 1987; 12: 53–62.

Yamamoto RS, Williams GM, Richardson HL, WeisburgerEK, Weisburger JH. Effect of p-hydroxyacetanilide onliver cancer induction by N hydroxy-N-2-fluorenylacetamide. Cancer Res. 1973; 33: 454–457.

Yoon HL, Marcus CB, Pfeifer RW. Induction of superoxideby 12-O-tetradecanoylphorbol-13-acetate andthapsigargin, a non-phorbol-ester-type tumor promoter, inperitoneal macrophages elicited from SENCAR andB6C3F1 mice: A permissive role for the arachidonic acidcascade in signal transduction. Mol. Carcinog. 1993; 7:116–125.

Yu R, Lei W, Mandlekar S, Weber MJ, Der CJ, Wu J, KongANT. Role of a mitogen-activated protein kinase pathwayin the induction of phase II detoxifying enzymes by chemi-cals. J. Biol. Chem. 1999; 274: 27545–27552.

Yu R, Mandlekar S, Lei W, Fahl WE, Tan TH, Kong ANT.p38 mitogen-activated protein kinase negatively regulatesthe induction of phase II drug-metabolizing enzymes thatdetoxify carcinogens. J. Biol. Chem. 2000; 275: 2322–2327.

plant polyphenols as a double-edged sword in health

Documents