Role of Reactive Oxygen Intermediates in Cellular Responses to Dietary Cancer Chemopreventive Agents

Abstract

Epidemiological studies continue to support the premise that diets rich in fruits and vegetables may offer protection against cancer of various anatomical sites. This correlation is quite persuasive for some vegetables including Allium (e.g., garlic) and cruciferous (e.g., broccoli and watercress) vegetables. The bioactive food components responsible for cancer chemopreventive effects of various edible plants have been identified. For instance, anticancer effects of Allium and cruciferous vegetables are attributed to organosulfur compounds (e.g., diallyl trisulfide) and isothiocyanates (e.g., sulforaphane and phenethyl isothiocyanate), respectively. Bioactive food components with anticancer activity are generally considered antioxidants due to their ability to modulate expression/activity of anti-oxidative and phase 2 drug metabolizing enzymes and scavenging free radicals. At the same time, more recent studies have provided convincing evidence to indicate that certain dietary cancer chemopreventive agents cause generation of reactive oxygen species to trigger signal transduction culminating in cell cycle arrest and/or programmed cell death (apoptosis). Interestingly, the ROS generation by some dietary anticancer agents is tumor cell specific and does not occur in normal cells. This review summarizes experimental evidence supporting involvement of ROS in cellular responses to cancer chemopreventive agents derived from common edible plants.

Introduction

Nutrition has long been suspected to play an important role in cancer etiology. Accumulating evidence suggests that diet, in addition to cancer causing substances, also contains many cancer preventive agents [1], [2], [3]. This implies that many cancers can be prevented by changes in dietary habits. It is becoming increasingly clear that many dietary agents can retard or prevent the process of carcinogenesis with reduced cancer outcomes by multiple mechanisms including (a) enhanced detoxification of the carcinogenic intermediates through induction of phase 2 enzymes, (b) reduced carcinogen activation due to suppression of cytochrome P450-dependent monooxygenases, (c) selective promotion of apoptosis (cell death) in cancer cells but not in normal epithelial cells, and (d) perturbations in cell cycle progression in cancer cells [3], [4], [5], [6]. Cancer protective effect of food constituents is also supported by epidemiological data. For example, the epidemiological evidence for cancer protective effect of cruciferous vegetables (e.g., broccoli) and Allium vegetables (e.g., Garlic) is quite strong [7], [8]. In recent years, many studies have demonstrated strong chemopreventive and possibly cancer chemotherapeutic effects of whole food and bio-active food components against cancers of the skin, lung, breast, colon, liver, stomach, prostate and other sites [3], [9]. The promising dietary chemopreventive compounds, which have demonstrated anticancer effects in more than one tumor model, include (-)-epigallocatechin gallate in green tea, resveratrol in grapes, lupeol in fruits like mango, delphinidin in pigmented fruits, curcumin in turmeric, sulforaphane and other isothiocyanates (ITCs) in cruciferous vegetables, organosulfur compounds (OSCs) in Allium vegetables, lycopene in tomato, and genistein in soy among many others [3], [5], [9].

The dietary cancer chemopreventive agents are considered antioxidants due to their ability to modulate level/activity of anti-oxidants and phase 2 drug metabolizing enzymes [3], [7], [9], [10]. For example, cancer chemopreventive OSCs derived from Allium vegetables not only cause induction of phase 2 drug metabolizing enzymes including glutathione transferases and quinone reductase but also increase glutathione levels and glutathione peroxidase activity in the liver as well as extra-hepatic tissues of mice [11], [12], [13]. More recent studies have indicated that many bioactive food components suppress growth of cancer cell in vitro as well as in vivo by causing cell cycle arrest and apoptosis induction [9], [10], [14], [15], [16]. Interestingly, cellular responses to some dietary cancer chemopreventive agents correlate with generation of reactive oxygen species (ROS). This review focuses on the role of ROS in signal transduction leading to cell cycle arrest and apoptosis induced by various dietary cancer chemopreventive agents.

Allium derived organosulfides

Occurrence

Cancer chemopreventive effect of Allium vegetables (e.g., garlic, onions, scallions, chives, leeks) is attributed to OSCs that are generated upon processing (cutting or chewing) of these vegetables [17]. The primary sulfur compound in intact garlic is γ-glutamyl-S-alk(en)yl-L-cysteine that is hydrolyzed and oxidized to yield alliin [17], which is an odorless precursor of the OSCs. Processing of garlic bulb releases a vacuolar enzyme alliinase, which acts upon alliin to give rise to extremely unstable and odoriferous compounds including allicin. Allicin and other thiosulfinates decompose to oil-soluble OSCs including diallyl sulfide (DAS), diallyl disulfide (DADS), diallyl trisulfide (DATS), dithiins and ajoene [17]. Several water-soluble OSCs including S-allyl cysteine (SAC) and S-allylmercaptocysteine (SAMC) are also derived from Allium vegetables.

Evidence for cancer chemoprevention

Initial evidence for anticancer effect of Allium vegetables came from population-based epidemiological studies [18], [19], [20]. For example, You et al. [18] reported a significant reduction in gastric cancer risk with increasing intake of Allium vegetables in a population-based case-control study. Similarly, Steinmetz et al. [19] observed an inverse correlation between intake of fruits and vegetables and colon cancer risk in the Iowa Women’s Health Study. Preclinical animal studies have indicated that OSCs can offer protection against chemically-induced cancers [21], [22], [23], [24], [25]. For example, Belman [21] was the first to show that topical application of essential oil of garlic inhibited incidence of tumor promoted by phorbol-myristate-acetate. Cancer chemoprevention by OSCs has been observed against benzo[a]pyrene-induced forestomach and pulmonary cancer in mice, dimethylhydrazine-induced colon cancer, and N-nitrosomethylbenzylamine-induced esophageal cancer in rats and azoxymethane-induced colon carcinogenesis in rats [22], [23], [24]. Elucidation of the mechanisms by which OSCs may prevent carcinogenesis has been a favorite subject of research for the past twenty years. On one hand, cancer chemoprevention by OSCs is, at least in part, believed to be due to their ability to modulate levels of anti-oxidants and phase 2 drug metabolizing enzymes. In HepG2 cells, OSC-mediated induction of NQO1 and heme oxygenase 1 was mediated by Nrf2 [25]. The antioxidant response element (ARE) activation and Nrf2 protein accumulation by OSCs correlated with phase 2 gene expression induction [25]. The structure-activity analysis indicated a critical role for the third sulfur in the OSC for its bioactivity [25]. In experiments involving transient transfection of HepG2 cells with mutant Nrf2, the DATS-induced ARE activity was inhibited by dominant-negative Nrf2 Kelch-like ECH-associating protein 1 and constructs [25]. At the same time, several OSCs are shown to suppress cancer cell growth in vitro and in vivo in association with ROS generation leading to cell cycle arrest and apoptosis induction [14], [26]. The anti- and pro-oxidant effects of OSCs are summarized below.

Anti-oxidative effects

The Allium vegetable-derived OSCs exhibit anti-oxidant activity [11], [12], [13], [27], [28], [29], [30], [31]. Some of them are simply free radicals scavengers, while others increase cellular anti-oxidant potential by inducing anti-oxidant enzymes. For example, treatment of mice with DAS, DADS and DATS resulted in elevation of glutathione levels and induction of glutathione transferase and quinone reductase expression in the liver, lung and/or forestomach [11], [12], [13]. The DADS administration ameliorated gentamicin-induced oxidative stress and nephropathy in rats in vivo [27]. The hydrogen peroxide mediated injury in neuronally differentiated PC12 cells was significantly attenuated by pretreatment with 20 μM DADS, which correlated with increased cell survival, activation of Akt, inactivation of GSK-3 and inhibition of cytochrome c release and caspase-3 activation [28]. Attenuation of carbon tetrachloride-induced oxidative stress and pulmonary fibrosis in rats by oral administration of SAC has also been reported [29]. SAC administration prevented carbon tetrachloride-induced depletion of glutathione levels, increased inducible nitric oxide synthase expression, infiltration of leukocytes, and ROS generation [29]. SAC-mediated protection against cerebral ischemic injury due to scavenging of endogenously and exogenously generated peroxynitrite and inhibition of extracellular signal-regulated kinase (ERK) has also been reported [30]. Kim et al. [31] showed that SAC, but not DAS, DADS or DATS, offered neuroprotection in an in vitro ischemia model. DATS and DADS, but not DAS, were shown to reduce lipopolysaccharide-induced expression of inducible nitric oxide synthase, nitric oxide production, oxidative stress and nuclear factor-kappaB (NF-kappaB) activation in RAW 264.7 macrophages [32]. NF-kappaB is a transcription factor that not only regulates gene expression of a number of pro-survival (anti-apoptotic) proteins but also is constitutively activated in a variety of hematological and solid tumor cells [33]. However, studies are needed to determine whether the anti-carcinogenic effects of OSCs are, at least in part, mediated by suppression of NF-kappaB activation. Some of the anti-oxidative effects of OSCs are summarized in Table 1.

Table 1.

Anti-oxidative effects of Allium vegetable-derived organosulfur compounds

Compound	Anti-oxidative effect	Reference
DAS, DADS, DATS	Increase in glutathione level and glutathione redox-cycle enzymes, induction of phase 2 enzymes (glutathione transferases and quinone reductase) in liver, lung and/or forestomach of mice	11-13
DADS	Protection against gentamicin-induced oxidative stress and neuropathy in rats	27
DADS	Protection against hydrogen peroxide-mediated cytotoxicity in PC12 cells neuronally differentiated by nerve growth factor	28
DATS	Reduction of lipopolysaccharide-induced expression of iNOS, nitric oxide production, oxidative stress and activation of nuclear factor-kappa B in RAW 264.7 macrophages	32
SAC	Protection of cerebral ischemia in vitro and neuroprotection in vivo	30,31
SAC	Protection against carbon tetrachloride-induced oxidative stress and pulmonary fibrosis in rats	29

Role of ROS in signal transduction

More recent studies have suggested a critical role for ROS in cellular responses to OSCs. For instance, Filomeni et al. [34] showed that DADS-induced apoptosis in SH-SY5Y neuroblastoma cell line correlated with ROS generation. The oxidative insult resulted in protein and lipid damage and activation of redox-sensitive c-Jun N-terminal kinase (JNK) [34]. DADS-mediated activation of JNK in SH-SY5Y cells was linked to ROS-dependent dissociation of glutathione transferase from JNK, since activation of JNK and apoptosis were attenuated by a spin trap and overexpression of copper/zinc-superoxide dismutase [34]. These results suggested a critical role for ROS in JNK activation and apoptosis by DADS [34]. Studies from our own laboratory have shown that apoptosis induction by DATS in human prostate cancer cells is associated with ROS-dependent activation of JNK [35]. DATS-mediated ROS production, JNK activation, and/or apoptosis in human prostate cancer cells were significantly attenuated by overexpression of catalase or pretreatment with NAC [35], [36]. It is interesting to note that a normal prostate epithelial cell line (PrEC) is significantly more resistant to growth inhibition and apoptosis induction by DATS compared with prostate cancer cells [35], [36]. The precise mechanism for selective killing of prostate cancer cells by DATS is not clear, but the possibility that resistance of normal prostate epithelial cells to apoptosis by DATS is due to the lack of ROS generated can’t be ignored. ROS-mediated activation of stress kinases, including JNK, in apoptotic response to DADS has also been observed in human glioblastoma T98G and U87MG cells [37]. Even though several studies suggest a tight link between ROS generation, JNK activation and apoptosis by OSCs [34], [35], [36], [37], the precise mechanism by which OSCs cause JNK activation is not fully understood. Similarly, the signaling events down-stream of ROS generation and JNK activation in OSC-mediated apoptosis remain elusive.

Structure-activity studies from our laboratory have revealed important roles for allyl groups as well as the oligosulfide chain length in cellular responses to OSCs [35], [38]. We found that the growth inhibition by OSCs against prostate cancer cells increases with an increase in the number of sulfur atoms [35]. For example, DATS is significantly more potent suppressor of prostate cancer cell proliferation compared with DADS whereas DAS, dipropyl sulfide and dipropyl disulfide are practically inactive [35]. Studies have indicated that exposure of cancer cells to OSCs results in G2/M phase cell cycle arrest. The above structure-activity relationship is also observed for OSC-mediated G2/M phase cell cycle arrest [38]. The G2/M phase cell cycle arrest has been observed for DADS and DATS in human colon, prostate, and lung cancer cells [34], [38], [39], [40], [41]. DADS and DATS-induced G2/M phase cell cycle arrest was markedly suppressed in the presence of NAC, suggesting that cell cycle arrest by garlic compounds is also dependent on ROS production [38], [41]. However, it is unclear if the protection offered by NAC is due to suppression of ROS generation or because of chemical reactivity between NAC and DATS. Even though the precise mechanism by which ROS contribute to DADS-mediated cell cycle arrest is not clear, we have shown previously that DATS-mediated cell cycle arrest and down-regulation and Ser-216 phosphorylation of the cell cycle regulatory phosphatase Cdc25C were significantly attenuated by pretreatment with NAC [38]. These results suggest that ROS generation may cause oxidative modification of Cdc25C to regulate its degradation. However, further studies are needed to substantiate this possibility. Hosono et al. [40] have documented that DATS can cause specific oxidative modification of cysteine residues Cys-12 and Cys-354 of beta-tubulin in a cell-free system. DATS treatment has been shown to disrupt the microtubule network in human colon cancer cells [40]. However, it remains to be seen whether DATS causes oxidative modification of beta-tubulin cysteine in vivo. More recent studies from our laboratory have revealed that DATS treatment disrupts the microtubule network in human prostate cancer cells [42]. Moreover, DATS treated human prostate cancer cells were arrested in prometaphase in a checkpoint kinase 1-dependent manner [42], [43]. Since checkpoint kinase 1 is an intermediary of DNA damage checkpoints [44], it is possible that DATS-mediated activation of checkpoint kinase 1 is caused by ROS-dependent DNA damage. Additional work is needed to experimentally verify this speculation. Pinto and co-workers have also suggested that beta-elimination reactions with cysteine S-conjugates in garlic may modify cancer cell growth by targeting redox-sensitive signal proteins thereby regulating cellular responses [45]. Examples of ROS-linked cellular responses to selected OSCs are summarized in Table 2.

Table 2.

Examples of ROS-linked cellular responses to organosulfides

Agent	Cell line	ROS-linked cellular responses	Reference
DADS	SH-SY5Y neuroblastoma	ROS generation JNK activation, apoptosis Protein and lipid damage	34
DATS	PC-3 and DU145 human prostate cancer	ROS generation JNK activation, apoptosis Protection against JNK activation and apoptosis by overexpression of catalase and in the presence of NAC	35
DATS	LNCaP human prostate cancer	ROS generation Apoptosis induction Protection against ROS generation and apoptosis in the presence of a combined mimetic of catalase and SOD	36
DADS/DATS	T98G and U87MG human glioblastoma	ROS generation Activation of JNK and p38 MAPK Apoptosis Increase in ER stress	37
DATS	HCT-15 and DLD-1 human colon cancer	G2/M phase cell cycle arrest followed by apoptosis Activation of caspase-3 Disruption of microtubule network	40
DATS	PC-3 and DU145 human prostate cancer	ROS generation G2/M phase and mitotic arrest Degradation and increased S216 phosphorylation of Cdc25C Protection of G2/M arrest, and degradation and S216 phosphorylation of Cdc25C in the presence of NAC Activation of checkpoint kinase 1 (Chk1) Attenuation of mitotic arrest by knockdown of Chk1 ROS generation linked to increased labile iron pool due to degradation of ferritin	38,42,46

The mechanism of ROS generation by DADS is not clear, but we have shown recently that the DATS-mediated ROS production, at least in human prostate cancer cells, is linked to an increase in labile iron pool [46]. DATS-induced ROS generation, G2/M phase cell cycle arrest and degradation and hyperphosphorylation of Cdc25C were significantly attenuated in the presence of EUK134, a combined mimetic of superoxide dismutase and catalase [46]. Interestingly, DATS-induced ROS generation and G₂-M phase cell cycle arrest were also inhibited significantly in the presence of desferrioxamine (DFO), an iron chelator, but this protection was not observed with iron-saturated DFO. DATS treatment caused a marked increase in the level of labile iron that was accompanied by degradation of light chain of the iron storage protein ferritin [46]. To our surprise, DATS-mediated degradation of ferritin, increase in labile iron pool, ROS generation and/or cell cycle arrest were significantly attenuated by ectopic expression of a catalytically inactive mutant of JNK kinase 2 as well as RNA interference of stress-activated protein kinase/extracellular signal-regulated kinase 1 [46]. These results provided convincing evidence to suggest the existence of a novel pathway involving the JNK signaling axis in regulation of DATS-induced ROS generation. However, further studies are needed to determine whether this pathway is specific for prostate cancer cells and unique to DATS. In an in vitro study, di-, tri-, and tetrasulfides generated hydrogen peroxide in the presence of GSH and hemoglobin and caused oxidative damage to erythrocytes [47]. The activity decreased in the order of tetra- > tri- > disulfide [47]. It remains to be seen if a similar mechanism is operative in cancer cells.

Cruciferous vegetable-derived isothiocyanates

Occurrence

Like Allium vegetables, evidence is quite strong for cancer protective effect of the vegetables of the family Cruciferae [7], [15], [47]. Commonly known edible cruciferous vegetables include broccoli, watercress, cabbage, kale, horseradish, radish, turnip, and gardencress. The cancer chemopreventive effect of cruciferous vegetables is attributed to ITCs, which occur naturally as thioglucoside conjugates (glucosinolates) and distinguish them from other vegetables [48]. The ITCs are hydrolysis products of glucosinolates and generated through catalytic mediation of myrosinase, which is released upon processing (cutting or chewing) of cruciferous vegetables from a compartment separated from glucosinolates [48]. Evidence exists for conversion of glucosinolates to ITCs in the gut [48]. ITCs have a common basic skeleton but differ in their terminal R-group, which can be an alkyl, alkenyl, alkylthioalkyl, aryl, beta-hydroxyalkyl or indolylmethyl group [48]. At least 120 different glucosinolates have been identified. The widely studied ITCs include sulforaphane, phenethyl isothiocyanate (PEITC), benzyl isothiocyanate (BITC), and allyl isothiocyanate (AITC).

Evidence for cancer chemoprevention

In addition to epidemiological evidence [7], [49], [50], laboratory studies have indicated that ITCs are highly effective in protection against cancer in animal models, induced by a variety of chemical carcinogens, including tobacco smoke-derived chemicals [15], [47]. For example, anticancer effects of sulforaphane [(-)-1-isothiocyanato-(4R)-(methylsulfinyl)-butane, naturally occurring L-isomer], which is present in rather high concentrations in broccoli, are well documented [51], [52], [53], [55], [56]. Cancer chemoprevention by L-sulforaphane or its synthetic analogue (racemic D,L-isomer) has been observed against 9,10-dimethyl-1,2-benzanthracene-induced mammary cancer in rats, azoxymethane-induced colonic aberrant crypt foci in rats, benzo[a]pyrene-induced forestomach cancer in mice, and 7,12-dimethylbenz(a)anthracene/12-O-tetradecanoylphorbol 13-acetate-induced skin tumorigenesis in mice [51], [52], [53], [55]. More recently, sulforaphane administration was shown to significantly inhibit lung metastasis induced by B16F-10 melanoma cells in C57BL/6 mice [56]. The mechanism by which sulforaphane inhibits chemically-induced cancers may involve impairment of carcinogen activation through inhibition of cytochrome-P450 and/or acceleration of inactivation of carcinogenic intermediates through induction of phase 2 enzymes [57], [58]. Similarly, BITC and PEITC have been shown to inhibit chemically-induced cancers in animal models. For example, BITC is a potent inhibitor of rat mammary and mouse lung carcinogenesis induced by the polycyclic hydrocarbons 7,12-dimethylbenz[a]anthracene and benzo[a]pyrene, respectively [59], [60]. Studies have indicated that even a subtle change in ITC structure could have a significant impact on its anticancer activity. For instance, pulmonary tumor induction by benzo[a]pyrene in A/J mice is inhibited significantly by BITC but not by PEITC, which is a close structural analogue of BITC [15], [47]. Thus, caution must be exercised in extrapolation of results from one ITC compound to the others. Although certain ITC compounds have also been shown to promote bladder tumorigenesis in rats [61], available epidemiological and preclinical data are mostly in favor of a cancer protective effect of dietary ITCs [7], [15], [47], [50]. In addition to preventing chemically-induced cancers, several ITC compounds have also been shown to inhibit growth of cancer cells in vivo. The in vivo anticancer effects have been reported for sulforaphane [62], [63], PEITC [64], [65] and AITC [66]. The in vivo growth inhibitory effects of SFN, PEITC and AITC correlate with apoptosis induction [62], [64], [65], [66]. As discussed below, recent studies have indicated that ROS act as key signaling intermediates in apoptosis induction by ITCs.

Role of ROS in signal transduction

Similar to Allium vegetable-derived organosulfides, the ITCs (e.g., sulforaphane) act indirectly to boost antioxidant capacity of the cells by increasing expression of phase 2 enzymes [67], [68]. The enzymes up-regulated by sulforaphane include Nrf2-regulated genes, including glutathione transferase, quinone reductase, and heme oxygenase-1, which can function as protectors of oxidative stress [67], [68]. In addition, sulforaphane was shown to offer powerful and prolonged protection of human retinal pigment epithelial cells, keratinocytes, and mouse leukemia cells against cytotoxic effects of several oxidants including menadione, tert-butyl hydroperoxide, 4-hydroxynonenal, and peroxynitrite [69]. Sulforaphane and its close structural analogues erucin and iberin have been shown to up-regulate thioredoxin reductase 1, which has broad substrate specificity and can reduce many low molecular weight compounds including hydrogen peroxide and lipid hydroperoxides, in MCF-7 human breast cancer cells [70]. The same group of investigators showed subsequently that sulforaphane is not only a potent inducer of thioredoxin reductase 1, but can also up-regulate its substrate thioredoxin [71]. BITC treatment inhibited excessive superoxide generation in inflammatory leukocytes [72]. Thus, it is reasonable to conclude that sulforaphane and possibly other ITCs may function as indirect antioxidants by modulating cellular redox status.

Evidence also exists to indicate that treatment of cells with ITCs results in transient burst of ROS, which seem to signal cellular responses to this class of dietary phytochemicals. The ROS-dependent cellular responses to ITC exposure are summarized in Table 3. The ROS generation in response to ITC treatment was first documented for BITC in rat liver epithelial RL34 cells [73]. Interestingly, these investigators also showed that BITC-mediated induction of glutathione transferase Pi was tightly linked to ROS generation [73], since ROS generation and induction of glutathione transferase Pi expression were found to be significantly inhibited by pretreatment of cells with glutathione [73]. These observations suggest that even the antioxidant effect of ITCs may be, at least in part, related to cellular adaptive response to ITC induced ROS generation. In a follow-up study, the same group of investigators reported a strong correlation between ROS generation and apoptosis induction by BITC in RL34 cells [74]. BITC-mediated ROS production in RL34 cells was associated with inhibition of mitochondrial respiration, mitochondrial swelling, and release of cytochrome c [74]. This was the first published report to document a mitochondrial redox-sensitive mechanism in BITC-mediated apoptosis [74]. Sulforaphane was shown to increase the transcript and protein levels of multi drug resistance-associated protein 2, an efflux pump contributing to biliary secretion of xenobiotics, in primary rat and human hepatocytes in an ROS-dependent manner [75]. Our studies have indicated that ROS generation is a critical event in apoptosis induction by sulforaphane in human prostate cancer cells [76]. Exposure of PC-3 cells to growth-suppressive concentrations of sulforaphane, resulted in ROS generation that was accompanied by disruption of the mitochondrial membrane potential, cytosolic release of cytochrome c, and apoptosis. All these effects were significantly blocked on pretreatment with NAC as well as ectopic expression of catalase [76]. The sulforaphane-induced ROS generation was also significantly attenuated on pretreatment with mitochondrial respiratory chain complex I inhibitors, including diphenyleneiodonium chloride and rotenone [76]. These results suggested that mitochondria are involved in sulforaphane-induced ROS generation in human prostate cancer cells [76]. However, further studies are needed to systematically explore this possibility. Sulforaphane-mediated apoptosis in human pancreatic cancer cell lines MIA PaCa-2 and PANC-1 also correlated with ROS generation, as NAC treatment conferred significant protection against the cell death [77]. Likewise, sulforaphane-mediated DNA fragmentation in HepG2 cells was significantly attenuated by NAC and catalase [78]. Subtoxic doses of sulforaphane sensitized hepatoma cells to apoptosis induction by tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), which correlated with ROS generation and induction of death receptor 5 mRNA and protein [79]. Pretreatment with NAC and catalase overexpression attenuated sulforaphane-induced up-regulation of death receptor 5 and nearly completely blocked apoptosis by the sulforaphane and TRAIL combination [79]. These results indicated that sulforaphane-induced ROS production and subsequent up-regulation of death receptor 5 were critical for triggering and amplifying TRAIL-induced apoptotic [79]. In DU145 human prostate cancer cells, sulforaphane treatment resulted in rapid and transient burst of ROS followed by apoptotic cell death, which were blocked by pretreatment of cells with NAC [80]. The positive correlation between ROS generation and apoptosis induction has also been observed for PEITC in human prostate cancer cells [65] and for BITC in human breast cancer cells [81]. The ROS production as well as apoptosis by PEITC and BITC was significantly attenuated in the presence of the combined superoxide dismutase and catalase mimetic Euk134 [65], [81]. Trachootham et al. [82] have suggested that abnormal increases in ROS can be exploited to selectively kill cancer cells by PEITC. Oncogenic transformation of ovarian epithelial cells with H-Ras^V12 or expression of Bcr-Abl in hematopoietic cells not only caused ROS generation but also rendered the malignant cells highly sensitive to PEITC [82]. Excessive ROS generation resulted in oxidative mitochondrial damage, inactivation of redox-sensitive molecules, and cell death [82]. Collectively, these studies provide convincing experimental evidence to implicate ROS in signal transduction leading to programmed cell death by several structurally different ITC compounds in different cell types [76], [77], [78], [79], [80], [81]. It is interesting to note that the aromatic ITCs (PEITC and BITC) are relatively more potent inducers of ROS production and apoptosis than sulforaphane, a alkylthioalkyl ITC [65], [76], [81]. It remains to be seen whether differential potencies of aromatic ITCs and sulforaphane for apoptosis induction is solely related to differences in their abilities to cause ROS production. Further investigation is also desired to elucidate the signaling events down-stream of ROS generation in execution of apoptosis by ITCs as well as to gain insight into the mechanism of ITC-mediated ROS generation. For example, it remains to be determined whether alteration in iron homeostasis contributes to ITC-induced ROS generation. Nonetheless, it is reasonable to conclude that ROS serve as key signaling intermediates in apoptosis induction by ITC family of cancer chemopreventive agents.

Table 3.

ROS-linked cellular responses to cancer chemopreventive isothiocyanates

Agent	Cell line	ROS-linked cellular responses	Reference
BITC	Rat liver RL34 cell line	apoptosis induction depletion of glutathione level	74
Sulforaphane	Primary rat and human hepatocytes	increase in MRP2 mRNA and protein expression attenuation of ROS production and MRP2 induction by dimethyl sulfoxide	75
Sulforaphane	PC-3 and DU145 human prostate cancer cells	depletion of GSH levels loss of mitochondrial membrane potential (MMP) caspase-9/3 activation apoptosis induction attenuation of ROS generation, MMP collapse, caspase activation and apoptosis by pretreatment with NAC and/or overexpression of catalase	76
Sulforaphane	MIA PaCa-2 and PANC-1 human pancreatic cancer cells	ROS production loss of MMP depletion of glutathione level apoptosis induction protection against loss of plasma membrane integrity and MMP in the presence of NAC	77
Sulforaphane	HepG2 hepatoma cell line	apoptosis induction abrogation of cell death by pretreatment with NAC and catalase	78
Sulforaphane	Hep3B, Huh-7, and HepG2 human hepatoma cells	Sensitization of hepatoma cells to apoptosis induction by tumor necrosis factor-related apoptosis-inducing ligand induction of death receptor 5 (DR5) ROS production attenuation of ROS generation, DR5 induction and cell death by pretreatment with NAC and overexpression of catalase	79
Sulforaphane	DU145 human prostate cancer cell line	ROS generation Apoptosis attenuation of ROS generation and DNA fragmentation in the presence of NAC	80
BITC	MDA-MB-231 and MCF-7 human breast cancer cells	ROS generation apoptosis induction abrogation of ROS generation and apoptosis induction by pretreatment with a combined superoxide dismutase and catalase mimetic Euk134	81
PEITC	Oncogenically transformed hematopoietic cells	ROS production cell death oxidative damage to cardiolipin glutathione depletion attenuation of ROS production and cell death by preincubation with exogenous catalase abrogation of glutathione depletion by pre-incubation with NAC and catalase	82
PEITC	PC-3 human prostate cancer cell line	ROS generation MMP collapse apoptotic DNA fragmentation attenuation of ROS generation, MMP loss and DNA fragmentation by pre-incubation with Euk134	65

ROS-dependent signaling by other bioactive food components

ROS-linked cellular responses in cancer cells are not unique to Allium vegetable-derived OSCs or cruciferous vegetable-derived ITCs. Experimental evidences linking ROS generation with signal transduction in cancer cells by three highly promising and widely studied bioactive food components (genistein, capsaicin, and curcumin) are described below, which illustrates the importance of ROS in cancer chemoprevention by dietary agents.

Apoptosis and growth inhibition caused by fermented soy product was related to increased ROS generation, and both processes were inhibited by catalase and an iron chelator [83]. In the isolated rat liver mitochondria model, soy constituent genistein caused swelling, loss of mitochondrial membrane potential, and release of accumulated calcium [84]. Moreover mitochondrial permeability transition pore opening caused by this soy isoflavone was mediated by ROS due to their interaction with respiratory chain at the level of complex III [84]. Likewise, apoptosis induction by capsaicin (8-methyl-N-vanillyl-6-nonenamide), the major pungent component of red pepper, in Jurkat cells is linked to ROS generation and JNK activation [85]. Growth inhibitory effect of capsaicin on H-ras-transformed human breast epithelial cells was related to overexpression of Rac1 and NAC reversed capsaicin-induced growth inhibition [86]. Capsaicin treatment induced apoptosis in human cutaneous squamous cell carcinoma cell lines, and cell death and ROS generation caused by this agent were significantly decreased in respiration-deficient cells [87]. Curcumin (1,7 bis [4-hydroxy-3-methoxyphenol]-1,6-heptadiene-3,5-dione) is another widely studied bioactive food component derived from the rhizome of the plant Curcuma longa. Curcumin exhibits both anti-oxidant and pro-oxidative properties [88], [89], [90], [91], [92]. For instance, curcumin at concentrations ranging between 3-30 micromolar was able to induce ROS production in cancerous human sub-mandibular adenocarcinoma cells and to smaller extend in normal human primary gingival fibroblasts [92]. Antioxidants including glutathione and NAC or hydroxyl radical scavenger-mannitol significantly reduced curcumin-induced ROS generation [90], [91], [92]. Pro-oxidative effect of curcumin was observed in another study where increased ROS formation was associated with a decrease in GSH concentration [93]. To the contrary, curcumin at 12.5 to 25 micromolar concentrations reduced ROS formation in human myeloid leukemia cells but elevated ROS level at higher concentrations [94].

Concluding Remarks and Future Directions

ROS are at the crossroad of cancer chemoprevention and carcinogenesis. On one hand, as discussed above, pro-apoptotic signal transduction by many dietary cancer chemopreventive agents relies on ROS production. On the other hand, endogenous ROS production is relatively high in tumor cells compared to normal cells as a result of oncogenic signalling through the NADPH oxidase and by the mitochondria. Genetic knockdown of Cu,Zn-SOD or decreased Mn-SOD activity in mice promotes cancer development [95], [96]. In addition, catalase knockout mice are more susceptible to cancer development [97]. Finally, embryonic fibroblasts lacking peroxiredoxin-1 express higher level of c-myc oncogene and ROS generation [98]. The fact that tumor cells have higher ROS generation can be exploited for therapeutic purposes. One strategy is to increase ROS scavenging, thereby inhibiting ROS-mediated mitogenic signaling in cancer cells. Feasibility of this strategy has been confirmed by experiments where superoxide dismutase, glutathione peroxidase or catalase over-expression was found to inhibit tumor growth [99], [100].

Because oxidative stress is implicated in pathogenesis of many chronic diseases including cancer, the potential side effects of ROS production by bioactive food components can’t be ignored. For example, hydroxyl radicals can produce purine, pyrimidine, and deoxyribose oxidation products [101], [102], [103]. Oxidative DNA damage relevant to cancer development by other reactive oxygen intermediates (e.g., peroxynitrite) has also been suggested [104], [105], [106]. ROS can also exert a wide range of cellular effects through modulation of signaling pathways that influence neoplastic transformation as well as cancer cell proliferation. The steady-state level of different ROS represents a plausible determinant for their cellular responses. The cancer chemoprevention by dietary agents including DATS and ITCs may involve both protection against oxidative DNA modification due to their ability to boost antioxidant defense in normal cells/tissues [11], [12], [13], and signal transduction leading to growth arrest and apoptosis in cancer cells [34], [35], [36], [39]. Validation of this speculation awaits further investigation. It is also possible that normal epithelial cells are relatively more resistant to dietary cancer chemopreventive agent-induced ROS generation compared with malignant cells, which may explain resistance of normal epithelial cells toward cell death caused by these agents. Further studies are needed to systematically explore this possibility as well.

We are tempted to speculate that ROS generation by bioactive food components with proven cancer chemopreventive efficacy, at least in preclinical cellular and animal models, may not be deleterious because: (a) these chemicals (e.g., OSCs and ITCs) are derived from vegetables consumed by humans on a daily basis, yet epidemiological studies continue to support the idea that dietary intake of these vegetables may reduce cancer risk [7], [8], (b) long term administration of these phytochemicals to rodents does not cause weight loss or any other signs of toxicity [65], [107], and (c) normal epithelial cells appear resistant to ROS generation and/or apoptosis induction by some of these phytochemicals [35], [36], [82], [108]. It is possible that the ROS generation by these phytochemicals in cancer cells serves to trigger signaling cascade leading to growth arrest and cell death.

Even though it is clear that ROS serve as critical intermediates in signal transduction by several bioactive food components, the mechanism by which ROS production occurs is poorly understood for many of these phytochemicals. Also, the events down-stream of ROS generation in cellular responses to bioactive food components remains elusive. Understanding the processes leading to ROS formation and their specific action on signalling pathways will likely help optimize cancer chemopreventive regimens involving these agents.

Abbreviations

AITC

allyl isothiocyanate

ARE

antioxidant response element

BITC

benzyl isothiocyanate

DAS

diallyl sulfide

DADS

diallyl disulfide

DATS

diallyl trisulfide

DR5

death receptor 5

ERK

extracellular signal-regulated kinase

GSH

glutathione

ITCs

isothiocyanates

JNK

c-Jun N-terminal kinase

MMP

mitochondrial membrane potential

NAC

N-acetylcysteine

OSCs

organosulfur compounds

PEITC

phenethyl isothiocyanate

ROS

reactive oxygen species

SAC

S-allylcysteine

SAMC

S-allylmercaptocysteine