OUP user menu

Novel concepts of broccoli sulforaphanes and disease: induction of phase Ii antioxidant and detoxification enzymes by enhanced-glucoraphanin broccoli

Don James, Sridevi Devaraj, Prasad Bellur, Shantala Lakkanna, John Vicini, Sekhar Boddupalli
DOI: http://dx.doi.org/10.1111/j.1753-4887.2012.00532.x 654-665 First published online: 1 November 2012


Consumption of broccoli has long been considered to play a role in a healthy diet. Broccoli accumulates significant amounts of the phytonutrient glucoraphanin (4-methylsulfinylbutyl glucosinolate), which is metabolized in vivo to the biologically active sulforaphane. The preponderance of evidence available from in vitro, animal, and human studies supports the association of sulforaphane with phase II enzyme induction. This has provided impetus for developing varieties of broccoli, both sprouts and whole heads, that are rich in glucoraphanin. The cancer-preventive properties of cruciferous vegetables, especially broccoli, have been studied for decades. However, evidence of broccoli directly affecting cancer incidence or progression is ambiguous, in part because of the presence of substantial polymorphisms in enzymes that metabolize sulforaphane. Since broccoli sulforaphane is one of the most potent inducers of phase II enzymes, exploration into broccoli's impact on other areas of human health, such as cardiovascular health and upper airway immunity, has been suggested. This review provides an update on evidence supporting phase II enzyme induction by sulforaphanes, with implications for breeding broccoli varieties with enhanced amounts of glucoraphanin. Early-stage human studies of consumption of broccoli with enhanced glucoraphanin are also discussed.

  • broccoli
  • glucoraphanin
  • phase II enzymes
  • sulforaphane


Consumption of fruits and vegetables is recognized as elemental to a healthy diet.1,2 Notably, cruciferous vegetables such as broccoli have been associated with better health and, in particular, a lower incidence of certain cancers.36 In addition, broccoli products have been associated with a reduction in markers of oxidative stress and hypertension in cardiovascular and kidney tissues as well as with induction of protective mechanisms against pollution-related upper airway inflammation.79 Glucosinolates (GLS) are sulfur-based compounds found naturally in Brassica vegetables such as broccoli, Brussels sprouts, cabbage, and cauliflower and have, therefore, been a main focus of research to explain the putative health benefits of these vegetables. Specifically, GLS are β-thioglycoside N-hydroxysulfates with a side chain R and a sulfur-linked β-D-glycopyranose moiety (Figure 1).10 Two aliphatic GLS that accumulate in broccoli florets are glucoraphanin (4-methylsulfinylbutyl) and glucoiberin (3-methylsulfinylpropyl). No overt phytonutrient bioactivity has been associated with GLS in vegetables per se; however, glucoraphanin and glucoiberin are converted to the bioactive isothiocyanates (ITCs) sulforaphane (SF) and iberin (IB), respectively, either by the action of myrosinase, an intracellular broccoli thioglucosidase,11,12 or by thioglucosidases in the microbiota of the human colon.13 SF is the predominant ITC derived from heading broccoli (Calabrese), and IB is the major ITC derived from sprouting broccoli.14 Broccoli hybrids of both types have been bred with substantially higher levels of SF or IB compared with current market standards, and early investigations are demonstrating an enhanced functional effect of these products, both in in vitro and human clinical studies.

Figure 1

Basic structure of glucosinolate.

This review will summarize the effects of SF on the induction of phase II (PII) enzymes and will highlight the evidence of a dose-response effect when sources providing differing levels of the bioactive SF are compared. The importance of PII enzyme induction and the direct impact of broccoli SF on human health and the detoxification of dietary toxins, free radicals, and reactive oxygen species in the body are also presented.

Biomarkers of exposure and sulforaphane bioavailability

Evidence from animal studies indicates that SF is metabolized by the mercapturic acid pathway, in which SF initially reacts with glutathione (GSH) but ultimately appears in the urine as the N-acetyl-L-cysteine (NAC) conjugate of SF (SF-NAC).1517 Studies have also revealed that free SF and its cysteine conjugate were more abundant than other conjugates in plasma and that significant quantities of free SF and cysteine conjugate were present in the urine in addition to the NAC conjugate. It has already been noted that the thiol conjugates of SF as well as those of other ITCs serve as carriers of ITC, and SF-NAC has been shown to exhibit equally if not more potent chemopreventive potential in cell and animal studies in comparison with SF.16 It is likely that thiol conjugates of SF, acting as carriers, make the bioavailable levels of SF higher than is apparent from serum SF levels.18

Biomarkers such as urinary SF conjugates and plasma SF have been utilized successfully to assess exposure to bioactive SF in small human clinical trials. For example, Conaway et al.15 examined the metabolic fate of GLS after ingestion of a single dose of steamed or fresh broccoli in 12 subjects and found that the relative apparent bioavailability of SF from fresh broccoli is three times greater than that of steamed broccoli, even though the initial content of GLS in both types was the same. Hanlon et al.19 conducted a feeding study with six subjects for 10 days to evaluate plasma pharmacokinetic characteristics of SF following single and repeated intake of raw broccoli. The study found no difference in pharmacokinetic parameters of plasma SF following single and repeated intake of broccoli. SF was absorbed rapidly following consumption, but additive accumulation did not occur from repeated intake. The study authors attributed the lack of accumulation partly to the very low levels of SF from the commercial broccoli used in the test. Shapiro et al.20 conducted a phase I randomized, double-blind, placebo-controlled trial with 12 healthy volunteers given doses of broccoli sprout extracts containing either GLS or ITC. In this study, it was found that, while the bioavailability of the SF preparation (extrapolated from total urinary ITC excretion) is extremely high and the interindividual variability of SF absorption and metabolism is small, the bioavailability of the glucosinolate preparation is much lower and there is a large difference between subjects in terms of total excretion of ITC (hence the large interindividual variability in the bioavailable SF). Gasper et al.21 conducted a randomized crossover dietary trial in 16 subjects to examine the serum bioavailability of SF after consumption of a single meal of standard broccoli versus a meal of broccoli containing three times as much glucoraphanin as a single meal of standard broccoli. The authors compared SF metabolism in glutathione S-transferase mu 1 (GSTM1)-null and GSTM1-positive subjects following consumption of standard broccoli or high-glucoraphanin broccoli using liquid chromatography linked to tandem mass spectrometry to quantify SF and its thiol conjugates in plasma and urine. GSTM1-null subjects had slightly higher SF metabolite concentrations in plasma, more rapid urinary excretion of SF metabolites in the first 6 h after consumption, and a higher percentage of SF excretion 24 h after ingestion compared with GSTM1-positive subjects. Most importantly, consumption of the high-glucoraphanin broccoli resulted in area under the curve (AUC) and maximum concentration (Cmax) of SF metabolites in plasma (AUC = 35.0–40.2 µmol · h/L; Cmax = 7.3–7.4 µM) that were threefold the control values.

Overall, the findings from these and other clinical trials with broccoli sprouts, florets, and extracts of GLS or ITC reflect significant interindividual differences in the conversion from GLS to SF. These differences are most likely due to differences in gut microflora and glutathione S-transferase (GST) polymorphisms, which may contribute to variable bioavailability, transformation, and excretion of broccoli SF in the body. The determinants of the efficiency in the microfloral conversion from GLS to ITC are complex and are affected by diet, host genetic factors, ethnicity, gender, gastrointestinal transit, and even enterohepatic circulation.20 Both intrinsic and extrinsic factors are important determinants of the degree of conversion from glucoraphanin in broccoli to SF and, hence, its availability for cellular activity. Finally, the Gasper et al.21 study demonstrates a glucoraphanin dose-dependent exposure profile that suggests the potential for a differentiable effect on PII enzyme induction with a high-glucoraphanin broccoli.

Induction of phase II antioxidant enzymes by sulforaphane: in vitro and animal evidence

Based on in vitro and animal model data, broccoli SF appears to mitigate some of the effects of carcinogens, toxins, reactive oxygen species, and other inducers of inflammation and cell damage by inhibition of phase I (PI) activation enzymes, such as cytochrome P450s, and induction of PII antioxidant and detoxification enzymes, including quinone reductase, thioredoxin reductase 1, and heme oxygenase 1 (HO-1).1622 PII enzymes regulate the redox state in mammalian cells by providing a precise balance between the level of reactive oxygen species (ROS) and endogenous thiol buffers present in the cell, particularly, GSH and thioredoxin (Trx), which protect cells from oxidative damage. PII enzymes also play a critical role in maintaining the balance of NAD+/NADH and NADP+/NADPH.23

SF is one of the most potent inducers of PII enzymes identified and is thought to act in humans via the nuclear factor E2-related factor 2 (Nrf2), which is a transcription factor that acts by binding to an antioxidant response element (ARE), a cis-acting enhancer sequence upstream of the genes coding for these PII enzymes.24,25 This mechanistic link with SF activity has led to a rapid growth in research. As a result, there are numerous published reports focused on SF-induced health benefits in in vitro, animal model, and human clinical studies.930 Results of these studies have made it apparent that three cellular components work together to regulate the gene expression of PII enzymes: 1) the AREs; 2) Nrf2, a basic leucine zipper transcription factor; and 3) Kelch-like ECH-associated protein 1 (Keap1), the sensor and chemical target for inducers (Figure 2). Keap1 binds Nrf2 to the actin cytoskeleton, thus controlling the function of the Nrf2 transcription factor.31 This putative chemical target of inducers such as ITC, Keap1, in conjunction with Cullin 3, binds Nrf2 tightly, leading to the ubiquitination and proteasomal degradation of Nrf2. The sulfur-exposed cysteine residues of Keap1 interact with electrophilic inducers (e.g., SF), resulting in conformational changes that are reported to switch ubiquitination from Nrf2 to Keap1 through a conformational change to Keap1. Degradative loss of Keap1, in effect, leads to Nrf2 nuclear accumulation and heterodimerization with small Maf transcription factors following elimination of Keap1's repression of Nrf2.32 This then results in binding of the complex to the ARE and recruiting of basal transcriptional machinery to activate transcription of PII enzymes. This mechanism is supported by recent experiments with Nrf2 knockout mice, providing important conceptual proof for the protective role of Nrf2-dependent gene products as modulated by electrophilic xenobiotic molecules.25 SF-mediated PII enzyme induction through the Nrf2-Keap1 pathway has been well documented in numerous review articles over the past decade.436

Figure 2

Induction of phase II enzymes by sulforaphane (SF) via the Nrf2-mediated pathway. Abbreviations: ARE, antioxidant response element.

The story of SF's putative benefits is evolving, and evidence supporting the association of SF with PII antioxidant and detoxifying enzyme levels continues to accumulate. Angeloni et al.37 investigated the time-dependent induction of gene transcription, protein expression, and enzyme activity of the antioxidant PII enzymes glutathione reductase (GR), GST, glutathione peroxidase, NADPH:quinone oxidoreductase-1 (NQO1), and thioredoxin reductase (TrxR) by SF in cultured rat neonatal cardiomyocytes. SF was also found to decrease production of intracellular reactive oxygen species (ROS), increase cell viability, and decrease DNA fragmentation after 48 h of treatment.

Exposure of human adult retinal pigment epithelial (RPE-19) cells to ultraviolet light in the presence of all-trans-retinaldehyde results in photo-oxidative cytotoxicity.38 Protection from cell death of RPE-19 cells was demonstrated, whereby cell variability was 50% greater following prior treatment with SF. In the same study, embryonic fibroblasts from mice with knockouts of Keap1−/−, Nrf2−/−, and Nrf2−/−:Keap1−/− were isolated and subjected to photo-oxidative stress from long-wave ultraviolet light in the presence of retinaldehyde. The resultant damage was found to be directly dependent on whether the transcription factor Nrf2 was suppressed or missing; the magnitude of resistance to photo-oxidative damage was also found to parallel the basal levels of the antioxidant GSH and the PII enzyme NQO1, both products of Nrf2-based induction. Results of these studies demonstrate the requirement for an active Nrf2 transcription factor for the induction of PII enzymes and strengthen the case for the direct intervention of SF in this process. In a related study, the authors compared the effects of treatment of human RPE-19 cells with a series of concentrations of SF (0–5 µM) for 24 h on GSH concentrations. GSH concentrations increased with increased SF concentrations.39 This clearly indicates that SF increases GSH levels in a dose-dependent manner and at a level of exposure that has been shown to be achievable in humans.21

In an animal study using Nrf2+/+ (wild type) and Nrf2−/− (Nrf2 mutant) mice, it was demonstrated that administration of broccoli seeds (26.4 mmol/kg glucoraphanin + glucoiberin) as 15% w/w of diet led to an approximate 1.5-fold increase in NQO1, GST, and glutamate cysteine ligase catalytic subunit activities in the stomach, small intestine, and liver of wild-type mice but not in mutant mice (Nrf2−/−).40 Unfortunately, it is impossible to assess levels of exposure of the mice to ITCs from the data. In the same study, it was demonstrated that an aqueous broccoli seed extract containing 61% IB and 30% SF (effective concentration 0.6 µmol/L total ITCs) induced glutathione S-transferase alpha (GSTA)1/2, GSTA3, NQO1, and glutamate cysteine ligase catalytic subunit between threefold and tenfold in murine Hepa-1c1c7 and rat liver RL-34 cells.

In another animal study, the pharmokinetics of a single 150 µmol oral dose of SF was examined in rats (approximately 210 g each), while at the same time assessing transcription and translation of key PII enzymes.41 The authors found increased SF metabolites in urine, plasma, and breast tissue, along with increases in NQO1 and HO-1 transcripts by reverse transcription polymerase chain reaction (RT-PCR) in the breast tissue. They also detected a threefold increase in NQO1 enzymatic activity and a fourfold increase in HO-1 protein using immunostaining in breast tissue. In a similar dose-dependence study, male Fischer 344 rats were treated with increasing doses of SF by oral gavage using pair-fed controls.42 In this set of experiments, SF induced NQO1 and total GST in the liver, pancreas, or colon of the rats. There was a clear dose response evident in this PII enzyme induction study, implying that sources of higher levels of glucoraphanin could provide a higher level of antioxidant enzyme-inducing power in the diet (Figure 3).

Figure 3

Effect of sulforaphane (SF) on induction of the phase II enzyme quinone reductase (QR), expressed as mean ± SE reduction of 2,6-dichlorophenol-indophenol (DPIP). The dose-related response is evident. An asterisk (*) indicates the mean is significantly greater than that for a pair-fed control (t test, P < 0.05). Reproduced with permission from Matusheski and Jeffery.42

At least two studies have examined the effect of broccoli SF on cardiac injury. Mukherjee et al.43 fed rats for 30 days with broccoli slurry, and then isolated hearts were perfused via working mode and made ischemic for 30 min followed by 2 h of reperfusion. Broccoli significantly improved postischemic ventricular function, reduced myocardial infarct size, and decreased cardiomyocyte apoptosis and also reduced cytochrome c release and increased pro-caspase 3 activities. Ischemia/reperfusion reduced both RNA transcripts and protein levels of the thioredoxin superfamily, including Trx1, Trx2, glutaredoxin 1, glutaredoxin 2, and peroxiredoxin, which were either restored or enhanced with broccoli. Broccoli enhanced the expression of Nrf2, a cytosolic suppressor of Keap1, suggesting a role of ARE in the induction of Trx. Additionally, broccoli induced the expression of another cardioprotective protein, HO-1, which could be transactivated by Trx. Thus, consumption of broccoli appears to be cardioprotective via redox cycling of thioredoxins.43 In a subsequent study,44 this group showed that the cardioprotective effects were superior in steamed broccoli compared with cooked broccoli. Two groups of rats were fed either fresh (steamed) broccoli or cooked broccoli for 30 days, and similar ischemia-reperfusion studies were conducted.

Both cooked and steamed broccolis displayed significantly improved post-ischemic ventricular function and reduced myocardial infarction and cardiomyocyte apoptosis compared with control, but steamed broccoli showed superior cardioprotective effects compared with the cooked broccoli. The expressions of proteins of the thioredoxin (Trx) superfamily, including Trx1 and its precursor sulforaphane, Trx2, and Trx reductase, were enhanced only in the steamed broccoli group. Thus, steamed broccoli is better than cooked broccoli due to its enhanced redox signaling of Trx.

The effects of bioactive SF on cellular physiology may change with both dose and cell type.17 From cell culture studies, it has been shown that very low concentrations of SF in the 0.4–0.8 µM range are able to induce a number of Nrf2-dependent changes in PII enzymes.45 Whether the concentrations necessary to cause the observed changes in cultured cells reflect the necessary plasma levels for similar activities to occur in vivo is not definitively known.17 Epidemiological and investigational studies in humans focused on dietary glucoraphanin or SF are much less numerous in the literature, but recent publications have demonstrated a clear link between glucoraphanin intake and PII enzyme induction and provide some clues as to doses of glucoraphanin and, ultimately, of SF necessary for induction of PII enzymes and their effect on biomarkers relevant to health.

Epidemiological studies in humans

Since GSH conjugation of SF is a major pathway for elimination from the body, it has been hypothesized that GSTM1-null individuals (i.e., null for the GSTM1 allele) would excrete less dietary SF and have more available for inducing PII enzymes via the Nrf2 pathway. Hence, a more beneficial effect would most likely be seen in GST-null individuals.46,47 The epidemiological evidence does not universally confirm this hypothesis.

The effect of broccoli intake among GSTM1-null or GSTM1-positive subjects was examined in seven case-control studies. In three of these studies, there was an association between intake of either broccoli or mixed cruciferous vegetables and cancer risk among GSTM1-null subjects.36 On the contrary, three other studies found an association between the intake of broccoli or cruciferous vegetables and cancer risk among GSTM1-positive subjects.4850 Three case-control studies and one cohort study examined the effect of total ITC or Brassica intake among GSTM1-null subjects. Specifically, London et al.51 found that ITC consumption was associated with reduced risk of lung cancer among male subjects from Shanghai, particularly among GSTM1-null subjects. Moreover, Seow et al. found that high dietary ITC was associated with a significantly lower risk of colorectal cancer among Singaporean subjects who were both GSTM1 null and glutathione S-transferase tau 1 (GSTT1) null.52 Fowke et al. found that urinary ITC levels were inversely associated with breast cancer risk among Chinese women and that this association was more consistent for women who have homozygous deletion of GSTM1 or GSTT1.53 Finally, Seow et al.54 reported a significant positive association between overall consumption of cruciferous vegetables and urinary total ITC level among Singaporean Chinese subjects. The urinary excretion of ITC was much greater for GSTT1-positive subjects, relative to GSTT1-null subjects, and was dependent on the level of cruciferous vegetable intake in subjects. The differences in the results reported by the studies summarized above are likely due to several factors. In particular, the studies, covering different ethnic populations, used different food frequency questionnaires to collect the intake estimates. In addition, estimates obtained via the food frequency questionnaires may include possible inaccuracies due to respondent recall biases. Relative nutrient absorption is difficult to estimate under the best of circumstances without direct measures of serum concentrations, and combined with the difficulties of estimating food intakes, this concern makes the assessment of SF impact from observational studies difficult. While such studies imply an association between GLS content and health benefit, appropriate human treatment studies utilizing sources of glucoraphanin or SF address more directly the causative factors in broccoli supporting xenobiotic and oxidant detoxification.

Phase II enzyme induction by sulforaphane: human intervention studies

An antioxidant is defined by the FDA as “a substance for which there is scientific evidence that, following absorption from the gastrointestinal tracts, the substance participates in physiological, biochemical, and/or cellular processes that inactivate free radicals or that prevent free radical-initiated chemical reactions.”55 By this definition, PII enzymes are potent antioxidants, which act catalytically (are not consumed in their antioxidant action), have relatively long half-lives, and catalyze a wide variety of chemical reactions that lead to detoxification. There are over 130 published studies – including 20 intervention studies – with broccoli as a test article as of early 2011. There are five single- or multi-dose human clinical studies that provide supporting evidence for PII antioxidant enzyme induction by SF from broccoli consumption (see Table 1). In one such example, Petri et al.56 conducted a controlled study with 11 volunteers, six of whom received a single-pass perfusion of the proximal jejunum of broccoli and onion extract at 2 mL/min and the other five of whom received a control perfusion. The broccoli extract contained 11 µM SF; however, the total dose administered could not be determined because the duration of the perfusion was not given. PII enzyme-encoding mRNA in exfoliated enterocytes from these subjects was measured. The broccoli extract rapidly induced a twofold increase in glutathione S-transferase alpha 1 (GSTA1) and UDP-glucuronosyltransferase 1A1 (UGT1A1) mRNA when compared with cells collected before and after the perfusion. These enzymes were not significantly induced in subjects receiving the control perfusion. To provide supporting data, Petri et al. 56 also demonstrated that SF from broccoli was responsible for induction of GSTA1 and that quercetin from onions was responsible for induction of UGT1A1 in Caco-2 cells.

View this table:
Table 1

Broccoli sulforaphane studies supporting induction of phase II enzymes in humans.

Type of study/subjectsConclusionsReference
Multiweek randomized, placebo-controlled feeding (2 wks): 100 µmol glucoraphanin infusion; 200 healthy Chinese adultsSulforaphane is absorbed and induces phase II detoxification enzymesKensler et al. (2005) 58
Multiweek randomized, crossover feeding (4 wks): =160 g/day crucifers vs. micronutrient + fiber; 20 adults in good health, 36–80 years of ageIntake of Brassica is associated with lower levels of an indicator of oxidative stressFowke et al. (2006) 57
Single-dose perfusion of proximal jejunum: 2 mL/min broccoli extract; 11 subjectsSulforaphane is efficiently absorbed in the small intestine (74%) and rapidly induces phase II mRNAPetri et al. (2003) 56
Single-dose randomized, 3-phase, crossover trial with soup made with standard broccoli and soup made with broccoli containing threefold amounts of glucoraphanin; 16 subjects, 18–46 years of ageSulforaphane is bioavailable and is correlated with dose; rapidly induces phase II mRNAGasper et al. (2007)22
Placebo-controlled, dose-escalation trial (25–100 g broccoli sprout homogenate); 65 healthy subjects, =18 years of ageBroccoli sulforaphane rapidly induces phase II mRNA in upper airways of humansRiedl et al. (2009)8
4-week placebo-controlled, randomized, double-blind, 2-dose trial: 5 g/day and 10 g/day broccoli sprout powder; 63 subjects with type II diabetesBiomarkers of oxidative stress were reduced at both doses, with oxidized low-density lipoprotein reduced by 5% with the 10 g/day treatmentBahadoran et al. (2011)60

Gasper et al.22 conducted a randomized, crossover dietary trial in 17 subjects to examine the effect of standard broccoli and high-glucoraphanin (HG) broccoli (broccoli with three times more glucoraphanin than standard broccoli) as a single meal, compared with water. This study involved a 21-day washout period between treatments, after which subjects consumed 150 mL of a soup made from the test broccoli, which delivered 102 µmol SF in the regular broccoli treatment and 344 µmol SF in the high-glucoraphanin (HG) broccoli treatment. Gastric mucosa biopsies were performed before the meal and 6 h after consumption of the meal (4 samples per timepoint). Changes in gene expression were evaluated using genome microarrays, which showed that consumption of HG broccoli resulted in upregulation of multiple genes for PII antioxidant and xenobiotic metabolizing enzymes, including TrxR, several aldoketoreductases, and glutamate cysteine ligase modifier subunit (GSH synthesis). Only one oxidoreductase-type gene (i.e., hydroxysteroid [11-β] dehydrogenase 2) was upregulated with standard broccoli. Real-time polymerase chain reaction (PCR) showed there was a positive dose relationship for expression of TrxR and glutamate cysteine ligase modifier subunit when comparing water, standard broccoli, and HG broccoli: HG broccoli showed a twofold induction of these gene products over water and standard broccoli controls (Figure 4).

Figure 4

Expression of thioredoxin reductase 1 (Tr1) (A) and glutamate-cysteine ligase modifier subunit (GCLM) (B) as measured by real-time reverse transcriptase polymerase chain reaction in a subjects 6 h after consumption of high-glucosinolate (HG) broccoli, standard broccoli (S), and water (W). Data are expressed as log2-fold changes and are means ± SEM, n = 10 (A), or n= 6 (B). Asterisks indicate difference from preintervention: *P ≤ 0.05; **P ≤ 0.001. Reproduced with permission from Gasper et al.22

Riedl et al.8 conducted a safety, efficacy, and dose-effect trial of repeat doses of oral SF on the induction of PII enzyme expression in upper airway cells in humans. A single-blind escalation block dose design was used. Sixty-five healthy volunteers were recruited from the campus of the University of California, Los Angeles (UCLA) campus and surrounding communities to participate in the study. Subjects gave baseline bronchial lavage samples and then ingested broccoli sprout homogenate (BSH) once daily for 3 days. Five subjects each consumed doses of 25, 50, 75, 100, 125, 150, 175, and 200 g of BSH per day (equivalent daily dose of SF at 13, 25, 38, 51, 64, 76, 89, and 102 µmol/day, respectively). The control subjects were given a daily dose of 200 g alfalfa sprout homogenate that provided 0 µmol/day of SF. Expression of GSTM1, glutathione-S-transferase pi 1 (GSTP1), NQO1, and HO-1 in the upper airway was evaluated. Expression of PII enzymes significantly increased in nasal lavage fluid samples at a BSH dose of =100 g/day. At this dose, NQO1 had the greatest induction, followed by HO-1. There was no change in PII enzyme induction in the controls. There was a significant dose-response effect of BSH on mucosal PII enzyme expression (P < 0.001) for all PII enzymes measured. Doubling to tripling of baseline enzyme expression rates was observed at the highest dose of BSH. Linear correlations by dose were apparent for all PII enzymes measured, consistent with the common pathway of induction by SF, i.e., Nrf2 activation. The study demonstrated that oral SF at doses of ≥ 64 µmol/day from BSH can significantly induce PII enzyme expression in the human airway.

Fowke et al.57 conducted a randomized crossover trial (n = 20) to compare the effects of a Brassica vegetable intervention against a micronutrient and fiber supplementation intervention on urinary F2-isoprostane levels (F2-iP), a stable biomarker of systemic oxidative stress. Subjects were generally healthy and were instructed to eat =160 g/day of mostly raw Brassica vegetables (most common: broccoli, cauliflower, Brussels sprouts, and cabbage) for 4 weeks during the Brassica vegetable intervention phase. During the micronutrient and fiber supplementation phase, subjects were given vitamin, mineral, and fiber supplements to match the increase in these nutrients from 2 cups/day of Brassica vegetables. Actual consumption of the vegetables during the Brassica vegetable intervention was between 220 g and 250 g/day; however, the equivalent intakes of GLS or SF were not reported. The Brassica vegetable intervention resulted in decreased levels of F2-iP in urine when compared with baseline or the micronutrient and fiber intervention (P < 0.05 for both). The authors speculated that Brassica consumption reduces the oxidative stress marker through a reduction in GSH reserves, although specific data were not provided. F2-iP synthesis is dependent on the reduction of prostaglandin H2 by GSH. It was hypothesized that GSH reserves were depleted sufficiently to inhibit F2-iP synthesis through either the induction of phase I/II enzymes that catalyze GSH conjugation to reactive oxygen species or the sequestration of GSH through conjugation with ITCs such SF. As ITCs also induce GSH synthesis, the mechanism for ITC-induced reduction of F2-iP is not fully understood. In summary, the findings of this study support an antioxidative effect of Brassica vegetables as measured by urinary F2-iP levels; however, the mechanism of this effect and the role of PII enzymes and GSH were not elucidated clearly.

Kensler et al.58 conducted a randomized, placebo-controlled study to examine the effect of glucoraphanin consumption on the disposition of aflatoxin, a food-borne carcinogen. Healthy adults (n = 200) were randomized to two treatments and drank infusions containing either 400 or < 3 µmol glucoraphanin daily for 2 weeks. This population was chosen because they had a high, natural exposure to aflatoxin. At the end of treatment, urinary levels of the aflatoxin B1 DNA adduct, AFB1-N7-Gua, were similar between the two intervention arms; however, measurements of urinary levels of SF metabolites had large and significant interindividual differences in bioavailability. The AFB1-N7-Gua is a DNA adduct formed by the carcinogen aflatoxin B1. Excretion serves as a biomarker of a biologically active dose, and elevated levels are associated with increased risk of cancer. The authors found that there were individual differences in the rates of hydrolysis of glucoraphanin into SF by the intestinal microflora in the study subjects. When the data analysis was controlled for this large interindividual variability in the bioavailability of SF from broccoli glucoraphanin by monitoring for dithiocarbamate excretion in each individual, a statistically significant inverse association was found between the levels of ITC excretion and the AFB1-N7-Gua adducts. GLS are hydrolyzed to ITCs through the action of broccoli myrosinase. Given that myrosinase is not heat stable, it is most likely destroyed during the thermal preparation of the broccoli infusion. In animal models, induction of GST is associated with reduced AFB1-N7-Gua. Kensler et al.58 suggested that the reduction of AFB1-N7-Gua adducts in this study is likely due to the induction of GST by SF, thereby shunting the reactive epoxide intermediate away from nucleophiles in DNA and toward GSH and reduction. Hence, this study, through evidence of the reduction of the aflatoxin-DNA adduct in urine associated with bioavailable SF from broccoli intake, suggests an association between SF exposure and the induction of PII enzymes such as GST.

Riso et al.59 assessed 20 subjects for their GSTM1 and GSTT1 genotypes using a crossover design in which subjects were fed 200 g/day of steamed broccoli for 10 days compared with a crucifer-free diet. The subjects consumed about 200 µmol/day of non-indolyl GLS, which would be mostly glucoraphanin, about 100 mg vitamin C, and about 5 mg of the carotenoids lutein and β-carotene. Increased plasma concentrations of carotenoids and relevant ITCs were found during the broccoli treatment, which were independent of GST genotype; however, GST activity in plasma, with 1-chloro-2, 4-dinitrobenzene as a substrate, was measured and was not affected by broccoli consumption. In this study, the broccoli preparation would have destroyed the endogenous broccoli myrosinase, while conversion to ITC would likely have occurred in the gut. The serum ITC levels measured in this study were relatively low compared with other published studies (0.4 µmol/L), possibly due to the heat-processing regimen of the treatment samples. This might explain why the broccoli intervention did not show an effect on plasma GST activity, a product of PII enzyme induction.

Most recently, Bahadoran et al.60 conducted a double-blind, placebo-controlled, randomized clinical trial to investigate the effects of a broccoli sprout powder (BSP) on biomarkers of oxidative stress in 63 subjects with type 2 diabetes. The 63 volunteers were assigned randomly to one of three treatment groups for 4 weeks, receiving 10 g/day BSP, 5 g/day BSP, or placebo. Serum total antioxidant capacity, total oxidant status, oxidative stress index, malondialdehyde, and oxidized low-density lipoprotein (LDL) cholesterol were measured at baseline and at completion of the treatment regimen. Three-day dietary recalls showed no significant differences between the groups in total energy and nutrient intakes. The doses used in this study provided 225 µmol and 112 µmol SF daily per 10 g and 5 g BSP doses, respectively. After 4 weeks, consumption of BSP resulted in significant decreases in malondialdehyde (8.9%; P = 0.001), oxidized LDL cholesterol (4.9%; P = 0.03), and oxidative stress index (13.7%; P = 0.001) and a significant increase in total antioxidant capacity (15.9%; P = 0.001) at the highest BSP dose. Although the authors did not measure serum SF concentration or assess any PII enzymes that might have been induced, they explained their results on the basis of the induction of PII enzymes.

Breeding a high-glucoraphanin broccoli for delivery of enhanced sulforaphane benefit

It is clear from preclinical in vitro and animal model studies that the impact of SF on induction of enzyme synthesis via the Nrf2 pathway is dose dependent (Table 2). Broccoli is already the richest common source of glucoraphanin, the precursor of SF, and there are several ways to increase the intake of SF. The first way is to increase total cruciferous vegetable intake or alter vegetable consumption to favor higher-glucoraphanin vegetables. The most likely way for this to occur would be through increased consumption of broccoli, since broccoli is high in glucoraphanin and represents almost 50% of all crucifers consumed in the United States.61 The second way is to increase conversion of glucoraphanin to SF, which could result from concomitantly consuming myrosinase from a second food source, such as daikon radish. A third way is to consume foods specifically developed to contain greater amounts of glucoraphanin. This would include, for example, the consumption of vegetables harvested at a point in maturation that is optimal for glucoraphanin levels. Alternately, plant parts that contain the highest levels of glucoraphanin could be utilized. Examples of this are found in broccoli sprouts, which are a very rich source of glucoraphanin,11 and broccoli florets, which are enriched in glucoraphanin relative to broccoli stalks. Broccoli with a greater content of glucoraphanin has been bred successfully.62 Gasper et al.,22 led by R. Mithen, demonstrated that a wild relative of broccoli, Brassica villosa, contained greater concentrations of aliphatic GLS than commercial broccoli cultivars. Using plant breeding methods, they introgressed three segments of the B. villosa genome into commercial broccoli, resulting in the HG variety SB1. This was accomplished by hybridizing a single plant of B. villosa to the double haploid breeding line of B. oleracea var. italica, which was crossed with the commercial cultivar Marathon, the result being optimized by phenotypic and marker-assisted selection. Several clinical studies in humans have already been conducted using SB1.22 Mithen's broccoli contains two- to three-fold more glucoraphanin compared with commercial check varieties of broccoli grown under the same conditions. A benefit of this high-glucoraphanin variety is that it enables nutritional intervention studies that concentrate on a specific phytonutrient from a whole food.

View this table:
Table 2

Summary of human, animal, and in vitro studies of the effects of sulforaphane/Brassica extract on specific phase II enzymes.

Phase II enzymeDosage of sulforaphaneType of study; system usedFold changeEnzyme changeReference
NQO113 to 102 µmol/dayHuman clinical; nasal lavage cells0 to 3ExpressionRiedl et al. (2009)8
5 µM SFIn vitro; rat cardiomyocytes2.2ExpressionAngeloni et al. (2009)37
9 µmol SF/dayIn vivo mouse; small intestine cells1.6ActivityThimmulappa et al. (2002)25
50 µmol/LIn vitro; human Caco-2 cells2.5ExpressionTraka et al. (2005)66
5 µmol/LIn vitro; human Hepa 1c1c7 and RL-34 cells4.5ActivityMcWalter et al. (2004)40
0.16 to 5 µMIn vitro; human ARPE-19 cells2ActivityGao et al. (2001)39
100 nMIn vitro; mouse cortical neurons8ActivityVauzour et al. (2010)67
2.0 µmol/LIn vitro; human NHBE cells2ExpressionTan et al. (2010)68
5 µmol/LIn vitro; rat hepatic Clone 9 cells3.9ActivityLii et al. (2010)69
5 µMIn vitro; human BEAS-2B cells15.0ExpressionRitz et al. (2007)9
In vitro; human NHBEC cells3.0
40 and 160 µmol ITC per kg body weightIn vivo rat; bladder2.4 and 4.4ActivityZhang et al. (2006)70
4 and 8 µmol/LIn vitro; rat bladder NBT-II cells2.3 and 2.6ActivityZhang et al. (2006)70
0.534 to 681 nmolHuman clinical; skin1 to 2ActivityDinkova-Kostova et al. (2007)71
0.5 µmolIn vivo SKH-1 hairless mice; skin2.7ActivityDinkova-Kostova et al. (2007)71
40 µmol per kg body weight per dayIn vivo rat; urinary bladder cells1.9ActivityMunday and Munday (2004)72
3 µmol/gIn vivo Nrf2(+/+) mouse; intestinal cytosol1.5ActivityMcMahon et al. (2001)73
50 µMIn vitro; human LS-174 cells3.0ExpressionMcMahon et al. (2001)73
0.1 µMIn vitro; human prostatic epithelial cells1.35ActivityBrooks et al. (2001)74
Brussels sprouts extractIn vivo rat; hepatic cells2.6ActivitySorensen et al. (2001)75
TrxR5 µMIn vitro; rat cardiomyocytes0.4ExpressionAngeloni et al. (2009)37
102 or 344 µmol/dayHuman clinical; gastric mucosa1.5 or 2.0ExpressionGasper et al. (2007)22
50 µmol/LIn vitro; human Caco-2 cells8.8ExpressionTraka et al. (2005)66
50 mg per kg body weightIn vivo tub/tub mice; retinal cells2.4ProteinKong et al. (2007)76
0.1 µMIn vitro; mouse cortical neurons2.6ActivityVauzour et al. (2010)67
2 µmol/LIn vitro; human HepG2 cells1.5ActivityBacon et al. (2007)77
10 µmol/LIn vitro; human Caco-2 cells1.5
1 to 12 µMIn vitro; human MCF-7 cells2.1 to 4.5ExpressionWang et al. (2005)78
12 µM4.0Activity
HO-113 to 102 µmol/dayHuman clinical; nasal lavage cells0 to 4ExpressionRiedl et al. (2009)8
GST (or isoform not given)9 µmol /dayIn vivo rat; cardiomyocytes1.3ActivityThimmulappa et al. (2002)25
2.5 to 6Expression
0.1 µmol/LIn vitro; mouse cortical neurons1.7ActivityVauzour et al. (2010)67
5 µmol/LIn vitro; rat hepatic Clone 9 cells1.8ActivityLii et al. (2010)69
24 or 120 mg per kg body weightIn vivo rat; hepatic microsomes3.2ActivityBarillari et al. (2007)79
40 and 160 µmol ITC per kg body weightIn vivo rat; bladder1.4 and 2ActivityZhang et al. (2006)70
4 and 8 µmol/LIn vitro; rat bladder NBT-II cells2.3 and 2.6ActivityZhang et al. (2006)70
40 µmol per kg body weight per dayIn vivo rat; bladder2.5ActivityMunday and Munday (2004)72
Broccoli seeds at 15% (by weight)In vivo Nrf2+/+ mice; stomach, small intestine, and liver0.5ActivityMcWalter et al. (2004)40
Brussels sprouts extractIn vivo rat; hepatic cells1.3ExpressionSorensen et al. (2001)75
GSTM113 to 102 µmol/dayHuman clinical; nasal lavage cells0 to 2ExpressionRiedl et al. (2009)8
GSTP15 µMIn vitro; human NHBEC cells2.0ActivityRitz et al. (2007)9
3 µmol/gIn vivo Nrf2(+/+) mouse; intestinal cytosol1.4ProteinMcMahon et al. (2001)73
13 to 102 µmol/dayHuman clinical; nasal lavage cells0 to 2ExpressionRiedl et al. (2009)8
GSTA15 µmol/LIn vitro; rat hepatic Clone 9 cells8.1ActivityLii et al. (2010)69
5 µMIn vitro; rat cardiomyocytes0.9ActivityAngeloni et al. (2009)37
GSTA311 µM onion/broccoli extractHuman clinical; enterocytes2ExpressionPetri et al. (2003)56
10 µM SFIn vitro; human Caco-2 cells3
3 µmol/gIn vivo Nrf2(+/+) mice; intestinal cytosol0.4ActivityMcMahon et al. (2001)73
Broccoli seeds at 15% (by weight) for 7 daysIn vivo Nrf2+/+ mice; stomach, small intestine, and liver1.0ActivityMcWalter et al. (2004)40
GSTA43 µmol/gIn vivo Nrf2(+/+) mice; intestinal cytosol0.7ActivityMcMahon et al. (2001)73
3 µmol/gIn vivo Nrf2(+/+) mice; intestinal cytosol0.9ActivityMcMahon et al. (2001)73
GR0.1 µMIn vitro; mouse cortical neurons2ActivityVauzour et al. (2010)67
GCLCBroccoli seeds at 15% (by weight) for 7 daysIn vivo Nrf2+/+ mice; stomach and small intestine5ProteinMcWalter et al. (2004)40
GSH0.16 to 5 µMIn vitro; human ARPE-19 cells2ProteinGao et al. (2001)39
  • Abbreviations: ARPE, arising retinal pigment epithelial; BEAS, transformed human bronchial epithelial; GCLC; glutamate cysteine ligase catalytic; GR, glutathione reductase; GSH, glutathione; GST, glutathione S-transferase; HO-1, heme oxygenase 1; ITC, isothiocyanate; NHBE, normal human bronchial epithelial; NHBEC, normal human bronchial epithelial cells; NQO1, NADPH:quinone oxidoreductase-1; SF, sulforaphane; TrxR, thioredoxin reductase.

It was demonstrated in an early human trial that consumption of this glucoraphanin-enriched broccoli elevated circulating SF and SF conjugates.21 The first trials also demonstrated that SB1 broccoli induces transcription of certain antioxidant and metabolizing enzymes about twofold more than a standard check variety, which is in line with the twofold higher level of glucoraphanin. A commercial-quality broccoli (Beneforté®; Seminis Vegetable Seeds, Inc.) was bred to include Brassica villosa genetics with two- to three-fold higher glucoraphanin content compared with commercial standard hybrids measured over multiple years and agronomic environments (Figure 5).63 This broccoli was used in an ex vivo study by Devaraj et al.,64 who looked at the effects of broccoli extracts on the inhibition of LDL oxidation. Oxidized LDL has been linked with cardiovascular plaque formation, plaque instability, and increased risk of cardiovascular disease. In this study, the extract of the broccoli containing two- to three-fold higher glucoraphanin inhibited LDL oxidation twofold more than a standard broccoli extract. This result underscores the need for properly controlled dietary intervention trials in humans comparing high-glucoraphanin broccoli with standard broccoli hybrids currently available to the consumer in order to better understand the complimentary health benefits offered by a high-glucoraphanin broccoli. At the time of this writing, there is an ongoing trial to assess the cardiovascular health impact of the prototypic high-glucoraphanin broccoli SB1 in subjects with slightly elevated risk of cardiovascular disease.65

Figure 5

Fold glucoraphanin elevation in Beneforté® broccoli florets versus standard check hybrids. Broccoli were grown in 23 locations across California, Arizona, and Mexico over 3 years. Least-squares mean of glucoraphanin content in Beneforté® was 70 mg/100 g fresh weight compared with 23–29 mg/100 g fresh weight in checks.


Data from human clinical studies demonstrate that broccoli and SF can induce several PII antioxidant enzymes (e.g., GSTA1, TrxR, several aldoketoreductases, glutamate cysteine ligase modifier subunit, GSTM1, GSTP1, NQO1, and HO-1) and can likely modulate levels of GSH. In addition, many of these enzymes are known to play a role in the recycling of active forms of vitamins A, C, and E. As documented in this review, data from in vitro and animal studies and from most human clinical trials provide support for the importance of the Nrf2 pathway of PII enzyme induction by SF, which results from intake of broccoli glucoraphanin (see Table 2). The existing clinical data also provide support that consumption of broccoli provides bioavailable SF, although the amount available varies significantly between individuals, likely as a result of gut microbial conversion of broccoli glucoraphanin to SF. In addition, there is evidence of a dose-dependency for PII antioxidant and detoxifying enzyme induction and the putative health benefits linked to that induction. This dose dependency suggests that increased consumption of glucoraphanin could support human health. Broccoli hybrids such as the high-glucoraphanin prototype used in the study of Gasper et al.22 or the recently commercialized Beneforté® broccoli used in the study of Devaraj et al.64 provide useful vehicles for further study of the effects of natural levels of glucoraphanin on human health. The overall weight-of-evidence currently available supports a PII antioxidant enzyme-related function for broccoli and provides a strong argument for increasing the intake of broccoli. Multicenter, placebo-controlled trials are warranted to investigate the potential of long-term consumption of broccoli with enhanced SF to improve human health. With prototypical high-glucoraphanin broccoli now available, some of these trials are already under way.


Declaration of interest.  All authors either work for or have received research funding from Monsanto, a company with a commercial interest in broccoli seeds.


View Abstract