Antioxidant Functions of Sulforaphane: a Potent Inducer of Phase II Detoxication Enzymes

1. Introduction

The widespread belief that oxidative damage plays a major role in cancer, ageing, and in a number of chronic diseases has focused scientific and public attention on the possibility that antioxidants could prevent or at least retard these processes. Antioxidants are of two types: direct and indirect. Direct antioxidants [e.g. glutathione (GSH), tocopherols, ascorbic acid and carotenoids] are substances that can participate in physiological, biochemical or cellular processes that inactivate free radicals or that prevent free-radical initiated chemical reactions. Direct antioxidants may also exhibit prooxidant effects under some experimental conditions, but whether these properties play a significant role in vivo remains unclear. In contrast, indirect antioxidants are not able to participate in radical or redox reactions as such, but they boost the antioxidant capacity of cells by a variety of mechanisms described below, and thereby afford protection against oxidative stress.

There is substantial and mounting evidence that Phase II enzymes [e.g. glutathione transferases (GSTs), NAD(P)H:quinone reductase (QR), epoxide hydrolase, heme oxygenase, UDP-glucuronosyltransferase] play important roles in the detoxication of electrophiles and their induction protects animals and their cells against carcinogenesis and mutagenesis (Benson et al., 1978, Benson et al., 1980, Kensler, 1997; Prestera et al., 1995; Talalay, 1992). The prevailing view of the mechanisms responsible for the protective functions of Phase II enzyme inducers has focused on the abilities of these enzymes to detoxify reactive electrophiles, and the inductive process has been termed the “electrophile counterattack response” (Prestera et al., 1993).

In this paper, we marshall evidence that a very important but largely ignored consequence of Phase II enzyme induction is the enhancement of cellular antioxidant capacity. There is now considerable information on the chemistry of inducers. Eight classes of chemical agents have been identified as inducers of Phase II enzymes (Prestera et al., 1993; Talalay et al., 1988): (1) diphenols, phenylenediamines, and quinones; (2) Michael reaction acceptors—compounds containing olefin or acetylene linkages conjugated to electron-withdrawing groups; (3) isothiocyanates; (4) hydroperoxides and hydrogen peroxide; (5) 1,2-dithiole-3-thiones; (6) vicinal dimercaptans; (7) trivalent arsenicals; (8) divalent heavy metals. These inducers vary greatly in potencies and show little structural similarity, but are all chemically reactive with sulfhydryl groups, and most are also substrates for GSTs (Prestera et al., 1993; Spencer et al., 1991). Recently, a ninth class of inducers comprising carotenoids, curcumins and related polyenes has been added (Dinkova-Kostova and Talalay, 1999; Khachik et al., 1999). Inducers of Phase II enzymes also raise the activity of γ-glutamylcysteine synthetase, the rate-limiting enzyme of GSH synthesis, and thereby elevate tissue GSH levels (Mulcahy et al., 1997). It is of interest that several of the types of aforementioned inducers of Phase II enzymes are both indirect and direct antioxidants, for example, phenolic antioxidants, mercaptans, carotenoids and curcumin.

Isothiocyanates are an especially interesting class of anticarcinogenic Phase II enzyme inducers. Many isothiocyanates are present in substantial quantities in the human diet (primarily originating from the ingestion of Cruciferae). More than 20 isothiocyanates inhibit the formation of tumours of several animal organ targets elicited by a variety of carcinogens (Hecht, 1995; Zhang and Talalay, 1994). Much evidence suggests that the major mechanisms of chemoprotection by isothiocyanates depend on the induction of Phase II detoxication enzymes and the inhibition of Phase 1 enzymes that are involved in the activation of certain carcinogens. Isothiocyanates are powerful electrophiles, a property attributable to the central carbon atom of the −N=C=S group, which reacts readily with sulfur-, nitrogen- and oxygen-based nucleophiles. However, there is no known chemical or biochemical evidence that this group can participate in oxidation or reduction reactions as a direct-acting antioxidant under physiological conditions (see Barton and Ollis, 1979). Furthermore, to our knowledge there is no evidence that the isothiocyanate group can behave as a prooxidant, unlike some direct-acting antioxidants.

The isothiocyanate sulforaphane [1-isothiocyanato-(4R)-(methylsulfinyl)butane: CH₃S(O)(CH₂)₄−N=C=S] has attracted much recent interest since it was isolated in our laboratory from broccoli by monitoring QR inducer activity, and found to be the most potent naturally-occurring inducer of Phase II enzymes (Prochaska et al., 1992; Zhang et al., 1992). Furthermore, sulforaphane inhibits mammary tumour formation in female rats treated with single oral doses of 9,10-dimethyl-1,2-benz[a]anthracene (DMBA) (Zhang et al., 1994). Interest in sulforaphane as a chemoprotector has been further stimulated by the recent finding that certain varieties of 3-day-old broccoli sprouts contain remarkably high levels of glucoraphanin, the glucosinolate precursor of sulforaphane and may provide an efficient source of dietary glucoraphanin and consequently of sulforaphane (Fahey et al., 1997).

2. Isothiocyanates, glucosinolates and myrosinase

Upon disruption of plant cells, isothiocyanates are one of the major products formed from the hydrolysis of glucosinolates (β-thioglucoside N-hydroxysulfates) by the action of the co-existing but physically segregated enzyme myrosinase (thioglucoside glucohydrolase; EC 3.2.3.1). Glucosinolates are water-soluble, anionic, non-volatile and heat-stable. They are not believed to possess significant direct biological activity. When plants containing glucosinolates are damaged by predators, or by food preparation or chewing, myrosinase is released and catalyses hydrolysis of the β-thioglucoside linkage. The resultant aglycones then undergo non-enzymatic, intramolecular (Lossen) rearrangement to yield isothiocyanates, thiocyanates, nitriles or epithionitriles. To date, a total of 104 glucosinolates have been reported from about 450 species of higher plants belonging to 16 families. There is a voluminous literature on the glucosinolates from the family Cruciferae, particularly those of the genus Brassica. This genus contains a large number of commonly consumed vegetables, the glucosinolates of which have been extensively documented and most recently reviewed by Rosa et al. (1997). Although there have only been two unconfirmed reports of the presence of glucosinolates in non-vascular plants, myrosinase-like (β-thioglycosidase) activity is quite commonly found in fungi and bacteria, including enteric bacteria. In contrast to their glucosinolate precursors, the isothiocyanates are typically lipophilic and may be highly reactive, volatile, malodorous, taste sharp or bitter, and some of them can be hepatotoxic, antibacterial, antifungal, antiprotozoal, nematocidal, allelopathic, or produce goitrogenic degradation products (Fenwick et al., 1983; Rosa et al., 1997).

3. Discovery, isolation, chemistry and biosynthesis of sulforaphane

The name glucoraphanin was initially coined for an antibacterial principle isolated from radish (Raphanus sativus) (Ivanovics and Horvath, 1947). Close examination of this report suggests that glucoraphenin, the glucosinolate of sulforaphene (CH₃S(O)CH=CHCH₂CH₂−N=C=S) may have been the compound actually isolated. Subsequent to this report, glucoraphanin (4-methylsulfinylbutyl glucosinolate) has only been identified in one of 16 higher plant families examined, the Cruciferae. In this family it is found in most of the commonly eaten vegetables in the genus Brassica (e.g. cole crops such as broccoli, kale, cauliflower, Brussels sprouts, cabbage, collards, turnips) as well as in two other genera containing such widely-consumed vegetables as radish (Raphanussp.) and arugula (Eruca sp.), Arabidopsis spp. in which much of the genetics of its production has been worked out, and in the following “non-food” genera: Alyssoides, Biscutella, Cakile, Cardaria, Choriopspora, Diplotaxis, Erysimium, Euzomodendron, Farestia, Fibigia, Isatis, Lepidium, Lesquerella, Lunaria, Malcomia, Physaria, Streptanthella and Thelypodium. Kjaer (1960)describes the early isolation of many of these compounds.

Sulforaphane, the cognate isothiocyanate of glucoraphanin, was synthesized in 1948 (Schmid and Karrer, 1948), and subsequently shown to be an abundant component of a weed known as hoary cress [Cardaria (Lepidium) draba L. (Desv.)] by Procházka (1959). Sulforaphane was also isolated from savoy and red cabbage by the same group (Procháska et al., 1959). Its antimicrobial activity was first identified by Procházka and Komersová (1959) and was much more extensively characterized by Dornberger et al. (1975). Hoary cress is described as a pot herb, and is apparently eaten as a peppery herb in central Europe, although details of its consumption are sketchy. It is also very widely distributed throughout the world as a rangeland weed (McInnis et al., 1993; Miller, 1991). We have found that hoary cress is very rich in glucoraphanin which can be readily converted to sulforaphane by the addition of exogenous myrosinase.

Glucoraphanin biosynthesis probably occurs via elongation of methionine (stepwise addition of methylene groups to form dihomomethionine), followed by addition of the glycoside moiety to provide 4-methylthiobutyl glucosinolate, followed by side-chain modification to produce 4-methylsulfinylbutyl glucosinolate (Faulkner et al., 1998). Many isothiocyanates, including sulforaphane (Schmid and Karrer, 1948) and its closely related sulfonyl isothiocyanate [α-methylsulfonylbutyl isothiocyanate; erysolin (Kjaer and Conti, 1954)] have been synthesized. Further recent improvements in the isolation of sulforaphane and related compounds have been reported (Kore et al., 1993).

We have isolated glucoraphanin from the pods, flowers, stems and leaves of hoary cress (Cardaria draba), also known as whitetop or heart-podded hoary cress (McInnis et al., 1993) and from broccoli (Brassica oleracea var. italica cv. DeCicco) and arugula (Eruca sativa) by a combination of chromatographic procedures. Hoary cress was collected from seven locations and extracts were prepared in the same manner as we have previously reported for broccoli sprouts (Fahey et al., 1997). In many of these extracts, glucoraphanin was responsible for most or all of the Phase II enzyme inducer activity as demonstrated by the development of such activity only after hydrolysis by addition of highly purified myrosinase (Shikita et al., 1999). Of the sites sampled, plants harvested from the Antelope and Leona Valleys, North of Los Angeles in 1993 and 1994 were by far the richest sources of detoxication enzyme inducer activity. The flowering tops of these plants rivalled broccoli sprouts (Fahey et al., 1997) in the potency of their inducer activity.

4. Inhibition of tumour formation

In 1992, after its isolation from broccoli, sulforaphane was shown to be a potent inducer of Phase II enzymes in isolated murine hepatoma cells and in various mouse tissues in vivo (Zhang et al., 1994). It was later shown to have potent activity in blocking the formation of mammary tumours in rats treated with the carcinogen DMBA (Zhang et al., 1994). The effect of the highest doses of sulforaphane tested on tumorigenesis was dramatic, resulting in a reduction in tumor incidence and multiplicity for all animals to 40.3% and 18.2% of control values, respectively, as well as a reduction in tumor weight and a significant delay in tumour development—arguably the most important effect. In subsequent studies with broccoli sprouts containing high levels of glucoraphanin as essentially the only glucosinolate/isothiocyanate, a similar protective effect against tumours was established at oral doses approximately equivalent to those of sulforaphane itself (Fahey et al., 1997). Other laboratories have shown that sulforaphane induces mouse mammary QR and GST activities and at very low concentrations reduces the incidence of preneoplastic lesions in carcinogen-treated mouse mammary glands in organ culture (Gerhauser et al., 1997).

Structure–activity studies (Zhang et al., 1992) have examined the effects of altering the oxidation state of the methylthio group (S, S=O, SO₂) and the length of the methylene chain (n=3, 4, or 5) separating the methylthio- and the −N=C=S groups of sulforaphane on the inducer potency for QR in murine hepatoma cells. All of these analogues were considerably less potent than the parent sulforaphane. Remarkably, replacement of the S=O by C=O produced an analogue that was equally potent to sulforaphane (Posner et al., 1994). A series of synthetic acetylnorbornyl isothiocyanates varied in inducer potencies depending on positions and steric relations of the acetyl and isothiocyanate groups. Some of these analogues were potent inducers, but in this series also, none was more potent than sulforaphane (Posner et al., 1994).

5. Cytotoxicity, mutagenicity and metabolism of sulforaphane

Although sulforaphane shows little cytotoxicity, some inhibition of growth of Hepa 1c1c7 cells was observed at concentrations 75-fold greater than those required (0.2 μm) to double quinone reductase activity (Gerhauser et al., 1997). Barcelo et al. (1996)were only able to demonstrate cytotoxicity against mouse hepatocytes at extremely high concentrations (>500 μm). Of even more direct relevance, sulforaphane has been shown to exert a dose-dependent cytotoxicity on human colon adenocarcinoma HT29 cells that was irreversible when administered to dividing cells; however, reversible inhibition of DNA synthesis was reported upon treatment of quiescent cells. Moreover, at the same doses (up to 30 μm), sulforaphane did not affect the viability of differentiated CaCo2 cells (Gamet-Payrastre et al., 1998). Additional evidence for the low toxicity of this compound is embodied in the demonstration by Barcelo et al. (1996)that sulforaphane is not genotoxic to hepatocytes, nor is it mutagenic to Salmonella typhimurium strain TA100 even at very high concentrations. Furthermore, sulforaphane does not induce unscheduled DNA synthesis in hepatocytes, but inhibits the mutagenicity of N,N-dimethylnitrosamine (NDMA) in the Salmonella assay, as well as inhibiting the unscheduled DNA synthesis evoked by this nitrosamine in hepatocytes.

Metabolic investigations of the fate of sulforaphane (administered either as pure compound or as a plant extract) in mammalian systems have shown that: (a) it is a good substrate for glutathione transferases (Kolm et al., 1995; Zhang et al., 1995), (b) it inhibits cytochromes P450 (Maheo et al., 1997; Morel et al., 1997), in particular, CYP2E1 (Barcelo et al., 1996), (c) it is conjugated to GSH and eliminated from mammalian system largely as its N-acetylcysteine conjugate (Kassahun et al., 1997); (d) it undergoes some reduction of the sulfoxide group to the sulfide to form erucin (Kassahun et al., 1997); and (e) it is excreted rapidly in humans (Shapiro et al., 1998).

6. Antioxidant functions of sulforaphane

Sulforaphane is not a direct-acting antioxidant or prooxidant, since it is very unlikely that the isothiocyanate group can participate in oxidation or reduction reactions under physiological conditions (see Barton and Ollis, 1979). There is, however, substantial and growing evidence that sulforaphane administration acts indirectly to increase the antioxidant capacity of animal cells, and their abilities to cope with oxidative stress.

7. Conclusions

There is ample evidence that sulforaphane is a very potent inducer of Phase II enzymes and also raises cellular glutathione levels. It is well known that Phase II enzymes play an important role in protecting cells against the toxicities of electrophiles. But we show here that some Phase II enzymes also protect cells against a wide variety of oxidative stresses, and the induction of these enzymes contributes to the mechanisms by means of which cells guard against the toxicities of reactive oxygen species and other forms of oxidative toxicity. Indirect antioxidants such as sulforaphane and other Phase II enzyme inducers are actually very efficient and rather versatile antioxidants for the following reasons: (a) unlike direct antioxidants, they are not consumed stoichiometrically while exerting antioxidant functions; (b) they have a longer duration of action and high cellular concentrations need not be maintained continuously since the induced enzymes have half-lives measured in days; (c) they support the functions of important naturally-occurring, direct-acting antioxidants such as the tocopherols and coenzymes Q; (d) they enhance the synthesis of glutathione, one of the most abundant intracellular direct antioxidants; and (d) they elevate enzymes that can cope with a wide variety of types of oxidants.

Talalay and Zhang, 1996, Zhang et al., 1993