Broccoli or Sulforaphane: Is It the Source or Dose That Matters?

1The collective retrospective (observational), prospective, and interventional evidence for the beneficial effects of broccoli on health is robust. The former two categories will be briefly summarized herein, and the latter make up the bulk of this review. Broccoli was thought to have been domesticated in the 1500s, brought to the UK in the early 1700s, and to the future United States in the late 1700s, but was little-known in the USA until the 1920s. Thus, as a relatively new crop, one can only point to about a century of relatively widespread consumption. The epidemiology of broccoli′s impact on health covers precisely half of that century, beginning with Graham′s early work showing a dose-response relationship between the consumption of broccoli and other cruciferous vegetables, on colon cancer. Since then there have been impressive demonstrations of risk reduction associated with cruciferous vegetables and/or broccoli for bladder cancer and prostate cancer just to name a few. We and others have recently reviewed the growing body of epidemiologic and mechanistic work implicating cruciferous vegetables in general, broccoli specifically, and sulforaphane, with respect to its association with neurologic, neoplastic, dermatologic, and other conditions (e.g.,). Organizations including the World Cancer Research Fund and the American Institute for Cancer Research have chosen to highlight lifestyles (a Mediterranean type diet) and food groups (non-starchy vegetables or fruit) as featuring strongly in cancer prevention, and they have moved away from recommending specific fruits or vegetables vis à vis cancer risk. However, all evidence points to broccoli and more specifically to sulforaphane from broccoli, and its biogenic precursor glucoraphanin, as being protective against a variety of chronic, and even infectious (e.g., Helicobacter pylori) conditions.

2Sulforaphane was described in the middle of the last century, as an antibiotic, and was isolated from red cabbage, and from the western USA rangeland weed hoary cress. Various groups have since synthesized it, but Talalay and Zhang were the first to isolate it from broccoli and to demonstrate its cancer protective properties. Its biogenic precursor, glucoraphanin, was then found in abundance in broccoli sprouts and sulforaphane was confirmed to be active in animal carcinogenesis models. Structure-activity evaluation of a series of over 100 synthetic analogs made by Posner and colleagues did not find a more potent inducer of cytoprotective (Phase 2) enzymes than sulforaphane, and indeedeme sulforaphane remains one of, if not, the most potent naturally occurring inducers yet discovered. We and many others have identified subsequently multiple pathways and metabolic consequences of this molecule’s presence in mammalian cells, tissues, and in the human body. While far too numerous to review herein, we note this multiplicity of effects has been well reviewed by others. Our early work showed that in the undamaged plant there is very little if any free sulforaphane, all of it being present in the form of its biologically inactive precursor glucoraphanin. Thus, it is important to consider the factors controlling glucoraphanin biosynthesis in plants.

3The glucosinolate glucoraphanin is derived from the amino acid methionine. The biosynthetic pathway leading to glucoraphanin formation in natura is elaborate and includes three separate stages (Scheme 1). The first stage results in the elongation of the methionine side chain by two methylene groups. The second stage forms the core glucosinolate structure. The third stage constitutes a secondary modification of the glucosinolate side chain.

4During the first stage (Scheme 1a), a cytoplasmic branched chain amino acid aminotransferase (BCAT) catalyzes the transamination of methionine to give 2-oxo-4-methylthiobutanoic acid. In turn, this α-keto acid is elongated by two methylene groups in a cyclical mechanism which involves two rounds of three successive transformations that take place in the chloroplast: (i) a condensation with acetyl-CoA, which is catalyzed by a methylthioalkyl malate synthase (MAM); (ii) an isomerization catalyzed by aconitase (AC); (iii) an oxidative decarboxylation catalyzed by an isopropylmalate dehydrogenase (IPMDH). The final product of these transformations, 2-oxo-6-methylthiohexanoic acid, is converted to dihomomethionine in a transamination reaction catalyzed by a BCAT.

5During the second stage (Scheme 1b), dihomomethionine undergoes a cytochrome P450 (CYP)-mediated conversion to an aldoxime, which is then further oxidized to a nitrile oxide and subsequently conjugated to a sulfur donor, such as glutathione (GSH). The conjugation reaction can occur non-enzymatically or can be catalyzed by a glutathione S-transferase. The resulting S-alkyl-thiohydroxymate is converted to a thiohydroxymate in a reaction catalyzed by the C-S lyase SUR1.

6Because this enzyme requires a free amino group within the substrate, an intermediate step is required: The hydrolytic removal of the γ-glutamyl residue within the conjugate between glutathione and the activated aldoxime, a reaction catalyzed by γ-glutamyl peptidase (GGP). The last steps of the second stage are a glucosyltransferase-mediated S-glucosylation resulting is the formation of a desulfoglucosinolate, followed by sulfation catalyzed by a sulfotransferase (ST). Thus, the parent glucosinolate, 4-methylthiobutyl glucosinolate (glucoerucin) is produced.

7During the final stage (Scheme 1c), a secondary modification of the glucosinolate side chain occurs. This is accomplished by an S-oxygenation reaction, which is catalyzed by a flavin monooxygenase (FMO). Together, these elaborate biotransformations, which involve 13 enzymes, result in the synthesis of glucoraphanin. Notably, in Arabidopsis the biosynthesis of glucosinolates is regulated by light (being downregulated in prolonged periods of darkness and greatly increased following exposure to light) and shows diurnal variations that are coordinated with those of general sulfur metabolism.

8Glucoraphanin is chemically stable and biologically inert. However, following plant tissue injury, such as biting or chewing, glucoraphanin comes in contact with the enzyme myrosinase, a ß-thioglucosidase, which in the intact plant is physically separated from its substrate. Myrosinase catalyzes the hydrolysis of glucoraphanin to liberate glucose and form an unstable aglucone (Scheme 1d) that spontaneously rearranges to give rise to a range of products, the most reactive of which is the isothiocyanate sulforaphane. Importantly, mammalian cells do not produce myrosinases; however, the conversion of glucoraphanin to sulforaphane still occurs in mammals. It is carried out by the bacterial microflora of the gastrointestinal tract and can be greatly reduced by antibiotic treatment or mechanical bowel cleansing. Of note, this microbially-mediated conversion of glucoraphanin has been exploited in a recent study to generate high concentrations of sulforaphane locally in the colon of mice.

9In plants, the main function of the glucosinolates is considered to be defense against pathogens and herbivores, and this has largely been attributed to the isothiocyanate hydrolytic products. Resistance to pathogens positively correlates with the glucosinolate content of the plant, which also affects the fungal species composition in the soil. In addition, there is a correlation between alterations in the glucosinolate profile of the plant and basic physiological processes, such as photosynthesis and growth, that occur during abiotic stress, including drought, extreme temperatures, light, salinity, and nutrient deprivation. Taken together, these correlations suggest that the glucosinolates function for protection of the plant against a wide array of environmental challenges. It is therefore not surprising that domesticated lines of Brassica oleracea have lower glucosinolate levels compared to wild species; conversely, the glucoraphanin content of certain wild Brassica species (B. villosa) is high, and this species has been used to generate broccoli hybrids with enhanced concentrations of glucoraphanin. In addition to their role in plant defense, high levels of isothiocyanates have been shown to reduce the biomass of the plant, and interestingly, mutants deficient in glutathione biosynthesis are more susceptible than their wild-type counterparts to the growth-inhibitory effect sulforaphane, suggesting similarities in sulforaphane metabolism in plants and mammals.

10Glucoraphanin occurs in all tissues of broccoli plants, though it is most abundant in the aerial portions and the developing florets (flower buds) and ultimately the seeds, are richest in this compound (Figure 1). Although two or three other edible cruciferous (Brassica) species contain significant amounts of glucoraphanin, it is not as widespread as popular culture would have it. In one sampling of 31 fresh, uncooked broccoli (Brassica oleracea var. italica) heads, each from different Baltimore area supermarkets, we found a mean of 0.38 μmol glucoraphanin per gram fresh weight, but with a range of from less than 0.005 to 1.13 μmol/gram. Levels of glucoraphanin when tested in over 75 different genotypes of field-grown hybrid broccoli averaged 0.88 and 1.10 μmol glucoraphanin per gram fresh weight when grown in the same fields in two consecutive years. Further work with 32 different genotypes grown across three different years, in the field and the greenhouse produced a similar range of values and a mean glucoraphanin content in broccoli heads of 0.36 μmol per gram fresh weight. Glucoraphanin level in broccoli seeds was judged to be largely determined by genotype, though the environment in which the plants are grown (e.g., location, year, drought, pollution, and disease pressure) also plays a clear and significant role.

11We surveyed the now vast literature on the evaluation of sulforaphane as an agent for disease prevention in mouse and rat models. While some studies reported feeding animals with broccoli (typically lyophilized) incorporated into rodent diets, the large majority have examined efficacy (monitored through molecular, biochemical, biological or pathological endpoints) of sulforaphane as a discrete, commercially available research-grade chemical. Of note, while R-sulforaphane is found naturally, the synthetically derived R,S-sulforaphane has been likely used in most animal studies due to historical matters of availability and relative cost. Most publications do not directly account for the form of sulforaphane used. There is some suggestion that R-sulforaphane shows increased effectiveness compared to its racemic R,S counterpart in various models. This review focuses on animal studies using sulforaphane as the test article as few of the dietary studies with broccoli-based preparations attempted to accurately determine the administered bioactive dose.

12Animal studies have principally used three routes of administration for sulforaphane: Oral, intraperitoneal and topical. Figure 2 highlights the distributions of doses selected by investigators for oral or intraperitoneal dosing to mice. Oral administration is the route typically used by the NCI in chemopreventive agent development. Yet, somewhat paradoxically as depicted in Figure 2, intraperitoneal administration has been the most commonly employed route for studies with sulforaphane. Presumably this choice reflects relative ease of administration to animals rather than attempted mimicry of a route most appropriate for administration of a dietary compound or matrix to humans.

13That said, dose ranges for intraperitoneal administration appears to provide roughly equivalent pharmacological efficacy to oral dosing with sulforaphane, presumably reflecting the excellent bioavailability of this agent. Doses selected for oral administration have spanned a greater than 4-log range and those for intraperitoneal administration an overlapping 3-log range. The median effective dose of sulforaphane in the published literature by oral administration is 175 µmol/kg body weight and by intraperitoneal administration is 113 µmol/kg. Likely reflecting publication bias, most studies irrespective of route of administration report a positive outcome on efficacy. Nonetheless, also informative, is the observation that studies, albeit far fewer in number, report a lack of efficacy in some models: E.g., tumor xenografts or chemical carcinogenesis, bacterial infections, airways inflammation and cerebral ischemia. For most of the listed “not effective” responses, significant responses were reported for study endpoints at higher doses as part of dose-response evaluations. Very few studies report harmful outcomes but are notable as discussed later. Topical application of sulforaphane has been used effectively in limited settings for mitigation of skin erythema, inflammation or photocarcinogenesis.

14The literature surveyed is listed and annotated for studies employing oral (Table S1), intraperitoneal (Table S2), and topical (Table S3) administration of sulforaphane to mice as well as oral administration to rats (Table S4).

15Relatively few studies have rigorously examined the mechanisms of action in vivo of sulforaphane that underlie the reported biological outcomes (Figure 3). This experimental paucity lies in stark contrast to the multitude of studies reporting on mechanistic actions identified in cell culture systems as reviewed elsewhere. A recent bibliometric review indicates that sulforaphane is the most cited natural product activator of Nrf2 signaling. Indeed, several studies have compared the action of sulforaphane on Nrf2 signaling and disease prevention by comparing outcomes in wild-type and Nrf2-knockout mice. In several cases, Nrf2-dependent effects have been linked to anti-inflammatory actions. In both settings (with some overlapping studies), a wide range of doses have been employed, indicating at least lower bounds for efficacy (3–10 µmol/kg). Studies examining the induction of apoptosis in vivo, principally in xenograft anti-tumorigenesis experiments have utilized high doses (>60 µmol/kg). A similar reliance on higher doses is reflected in studies on the inhibition of histone deacetylase (HDAC) activities (>100 µmol/kg). Absence of data in these cases should be distinguished from absence of effect at lower doses. Further experiments are required. To this point, relatively few dose-response studies have been conducted in vivo, so it is not possible to appreciate whether a plot of the relationship of response(s) to sulforaphane dose may take the form of “S”, “∩” or possibly some other shape. No maximally effective doses have been established—They are likely to vary depending upon the endpoint.

16Several studies conducted with higher doses of sulforaphane in mice do describe toxicities that require careful attention to risk-benefit analyses and determination of therapeutic or prophylactic indices. Socala et al. examined the toxicity profile of sulforaphane in mice after intraperitoneal injection of single doses. High doses of sulforaphane produced marked sedation (at 150–300 mg/kg), hypothermia (at 150–300 mg/kg), impairment of motor coordination (at 200–300 mg/kg), decrease in skeletal muscle strength (at 250–300 mg/kg), and deaths (at 200–300 mg/kg). The LD50 value of sulforaphane in mice was estimated to be 213 mg/kg i.p. (1203 µmol/kg). This value is about 10-fold higher than the median dose reported for efficacy outcomes in mice. Sulforaphane (at 100 mg/kg) potentiated the anticonvulsant efficacy of carbamazepine in a seizure test. This drug interaction could have been pharmacokinetic in nature, a form of interaction also considered but not observed in humans. Shorey et al. observed increased morbidity and no reduction in lung tumorigenesis in offspring born to mothers receiving transplacental and lactational exposure to the carcinogen dibenzo[def,p]chrysene and supplemented with dietary sulforaphane (400 ppm) or its primary whole food source, broccoli sprouts (10% wt/wt), contrasting to many reports of chemoprotection in adult animal models. Of potentially greatest concern is the study of Tao et al.. They utilized a chemical carcinogenesis model with vinyl carbamate (A/J mice) and a genetic model (LSL-K-rasG12D/+ mice) to induce lung cancers. Mice were treated intraperitoneally with 12.5 mg/kg (75.5 µmol/kg) (5 doses, once every three days during tumor induction (pre-treatment) or 12.5 mg/kg every three days, starting one week after tumor induction for 13 weeks (post-treatment). In the chemical carcinogenesis model, pre-treatment reduced the number of tumors, and post-treatment slightly promoted tumors. In the genetic model, pre-treatment with SF had no effect on tumor number, but post-treatment increased tumor number and size. Given observed efficacy in humans at doses <0.5 µmol/kg of sulforaphane, and providing allowance for allometric scaling from mice and the improved bioavailability with i.p. dosing, there is only a 10-fold difference in safety margin. It should be noted that Kombairaju et al. reported that prolonged sulforaphane treatment (0.5 mg, 5d/wk for 3 mo. by means of a nebulizer) did not enhance tumorigenesis in the same LSL-K-rasG12D/+ mouse model. In settings of long-term preventive interventions with sulforaphane or a broccoli-based formulation, careful consideration will need to be given to matters of dose, schedule and duration.

17Upon entry into the mammalian cell, sulforaphane is conjugated with GSH in a GST-catalyzed reaction, entering the mercapturic acid pathway (Scheme 1e). The glutathione conjugate of sulforaphane is subjected to a series of sequential conversions catalyzed by γ-glutamyltranspeptidase (γGT), followed by cysteinylglycinase (CGase), and N-acetyltransferase (NAT). The final product is the N-acetylcysteine conjugate of sulforaphane (mercapturic acid). In addition, sulforaphane can undergo an interconversion to erucin [1-isothiocyanato-4-(methylthio)butane], which is then metabolized in an identical manner to that for sulforaphane.

18Sulforaphane and its metabolites (dithiocarbamates) can be quantified collectively by cyclocondensation with 1,2-benzenedithiol, with sensitivity in the picomolar range. This highly sensitive, simple and convenient method has been widely used to measure the levels of sulforaphane and its metabolites in blood, plasma, urine and tissues following sulforaphane administration to rodents and humans. The use of this method revealed that sulforaphane crosses the placental barrier based on detection of dithiocarbamates in embryos 2 h post-treatment of pregnant mice with a single (5 μmol) dose of sulforaphane. In addition, methods have been developed to analyze the individual metabolites following their separation by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). Furthermore, the use of mass spectrometry coupled with stable isotope-labeled internal standards of sulforaphane [1-isothiocyanato-4-methyl-sulfinyl(1,1,2,2,3,3,4,4-2H8)butane] and its corresponding mercapturic acid pathway conjugates allows for quantitative, precise, sensitive, and specific analysis of sulforaphane and its metabolites.

19With these analytical tools in hand, a number of pharmacokinetic studies have been conducted in rodents and humans. Thus, following oral administration of an exceedingly high dose of 150 μmol sulforaphane to 10-week-old female Sprague–Dawley rats, the concentration of dithiocarbamates in the plasma of the animals increased rapidly reaching a peak (Cmax) of 60 μM 1 h after dosing, with area under the concentration curve (AUC) of 491 h μmol/L, elimination constant (Kel) of 0.1 h−1, and elimination half-life of 6.7 h. Similarly, following oral administration of 200 μmol broccoli sprout isothiocyanates to four healthy human volunteers, the peak plasma dithiocarbamate concentration (Cmax) was 1.91 ± 0.24 (0.943–2.27) μM 1 h after dosing, with half-life of 1.77 ± 0.13 h, and clearance of 369 ± 53 mL/min. A study in 20 participants administered 200 μmol sulforaphane as sulforaphane-rich powder in capsules reported a Cmax of 0.7 ± 0.2 µM at 3 h, with a half-life of 1.9 ± 0.4 h for elimination of sulforaphane equivalents measured by mass spectrometry. Another pharmacokinetic study, in which a single dose of broccoli soup delivering the equivalent of either 16 μmol or 52 μmol of sulforaphane was administered, reported Cmax of 2.2 ± 0.8 µM and 7.3 ± 2.9 µM at 1.5 h and 2 h for the low and the high dose, respectively. A double-blinded, randomized crossover trial with broccoli soups (prepared from plants with increased glucoraphanin content) delivering a single dose of 84, 280, or 452 μmol of glucoraphanin documented peak plasma concentrations (Cmax) of 0.17 ± 0.12, 0.37 ± 0.26, and 0.61 ± 0.40 μM, respectively. Another study reported plasma dithiocarbamate levels of 0.92 ± 0.72 μM and mean epithelial-/stromal-enriched breast tissue dithiocarbamate concentration of 1.45 ± 1.12 and 2.00 ± 1.95 pmol/mg tissue for the right and the left breast, respectively in eight healthy women undergoing reduction mammoplasty who had received a single dose of a broccoli sprout preparation delivering 200 μmol sulforaphane 1 h prior to surgery.

20In a double-blind randomized placebo-controlled trial in men presenting for prostate biopsy, plasma levels of 0.12 µM of sulforaphane and its metabolites were detected after an intervention period of 4–8 weeks with two daily doses of 100 μmol sulforaphane administered 12 h apart. Isothiocyanate levels of 2.2 µM and 500 nM were detected in the plasma and synovial fluid, respectively of patients with osteoarthritis undergoing knee replacement surgery following consumption of glucosinolate-rich diets for 2 weeks.

21A study in healthy subjects who received single oral doses of broccoli sprout extracts containing the equivalent of 111 µmol of glucosinolates or isothiocyanates showed cumulative urinary dithiocarbamate excretion of 88.9 ± 5.5 μmol and 13.1 ± 1.9 μmol for the isothiocyanate and the glucosinolate preparation, respectively. This study further revealed that for the isothiocyanate preparation, excretion was consistent and linear over a 25–200 μmol dose range, whereas for the glucosinolate preparation, excretion was highly variable among individuals. These observations are in close agreement with results from a randomized, placebo-controlled, double-blind Phase I clinical trial, in which isothiocyanate (25 μmol)- or glucosinolate (25 μmol or 100 μmol)-rich preparations were orally administered to three cohorts of three healthy human subjects at 8-h intervals for 7 days; one other subject in each cohort received placebo. Notably, this study showed no evidence of clinically significant adverse events based on 32 types of hematology and chemistry tests, including liver (transaminases) and thyroid (TSH, T3, and T4) function tests. In agreement, a recent analysis of biochemical parameters of thyroid function in serum collected from 45 female volunteers that had participated in a randomized clinical trial revealed no alterations compared to baseline following an intervention for 12 weeks with a broccoli sprout beverage containing a combination of 40 μmol sulforaphane and 600 μmol of glucoraphanin.

22The finding that compared to isothiocyanates, oral administration of glucosinolates results in lower bioavailability, slower elimination, and greater inter-individual variation in excretion was further strengthened by a larger (50 participants) crossover clinical trial that involved 5-day baseline period followed by daily administration of broccoli sprout beverages delivering either glucosinolates or their corresponding isothiocyanates for 7 days, 5-day washout period, and 7-day administration of the opposite intervention. Using fecal sample collections from five subjects with high 24-h urinary excretion profiles (‘high converters’) and five subjects with low excretion profiles (‘low converters’), it was found that ex vivo, the degradation of glucoraphanin was greater in cultures of fecal bacteria derived from the ‘high converters’ in comparison to the ‘low converters’. These observations are consistent with earlier work showing that mechanical cleansing or antibiotic treatment greatly reduce the glucosinolate conversion in healthy human subjects and indicate that the gastrointestinal microflora represents a critical factor in determining the extent of gl ucosinolate hydrolysis. In addition to the inter-individual variations, there are also diurnal variations in the conversion of glucosinolates to dithiocarbamates, whereby conversion is greater during the day. By contrast, the conversion of isothiocyanates to dithiocarbamates is higher during the night.

23Overall, in humans, sulforaphane is rapidly absorbed and eliminated with small inter-individual variations and typical urinary excretion of 70% to 90% of the dose. By contrast, the conversion of glucoraphanin is slow and with high inter-individual variations. The urinary excretion of sulforaphane metabolites following intervention with glucoraphanin-containing preparations typically range from 2% to 15% of the dose, being 1% to 45% at the extremes. The differences in inter-individual variations between sulforaphane and glucoraphanin make, at first glance, the use of sulforaphane much more attractive for the purposes of dose precision. However, in contrast to its stable glucosinolate precursor, sulforaphane is unstable, which has prompted the development of stabilized preparations, such as an α-cyclodextrin-encapsulated form of sulforaphane and a stabilized version of pure plant-derived sulforaphane, known as Prostaphane® (Nutrinov, Noyal sur Vilaine Cedex, France). Alternatively, glucoraphanin-rich preparations containing active myrosinase have also been used. As formulations differ in their bioavailability (which provides a possible explanation for the differences in pharmacokinetic parameters reported in the various human studies), the excreted amount of sulforaphane metabolites in the urine, and not the amount in the administered preparation, provides a more reliable measure of the actual dose.

24Similar to the studies of the pharmacokinetics of sulforaphane, nearly all human studies addressing the pharmacodynamics of sulforaphane have used glucoraphanin- or sulforaphane-rich broccoli-based preparations. Although there is currently no direct evidence for specific target engagement by sulforaphane in humans, there is clear evidence for its pharmacodynamic action. Thus, increased levels of the Nrf2-target enzymes A-class GSTs and NQO1 have been reported in plasma and saliva of human subjects consuming cruciferous vegetables. In agreement, administration of glucoraphanin/sulforaphane-rich preparations to healthy volunteers resulted in increased mRNA or protein levels of NQO1 and GSTs in PBMC, skin punch biopsies, as well as in nasal and buccal scrapings.

25Broccoli-based glucoraphanin/sulforaphane-rich preparations have been shown to accelerate the detoxication and excretion of potentially carcinogenetic food contaminants and air pollutants, offering a very attractive strategy for population-wide reduction in cancer risk due to unavoidable exposures to pollution. A cross-over clinical trial with 50 human volunteers, which was conducted in Qidong, China, found statistically significant increases of 20–50% in the urinary excretion levels of glutathione-derived conjugates of the air pollutants acrolein and benzene following consumption of sulforaphane- and/or glucoraphanin-rich broccoli sprout-derived beverages. A subsequent 12-week placebo-controlled, randomized clinical trial involving 291 participants from the same area using broccoli sprout beverages containing a combination of 40 μmol sulforaphane and 600 μmol glucoraphanin confirmed and extended these findings by showing that the excretion levels of the glutathione-derived conjugates of benzene and acrolein were significantly increased, by 61% and 23%, respectively in the volunteers who received the broccoli sprout beverage compared with placebo. Very recently, a randomized, placebo-controlled, multidose intervention trial of a broccoli sprout beverage, which was conducted in the same area of China, showed a dose-dependent excretion of the urinary metabolites of sulforaphane, and further found that a treatment regime with daily doses of 40 μmol sulforaphane and 600 μmol glucoraphanin for 10 days, resulting in an urinary excretion of ∼25 μmol sulforaphane metabolites per day, promotes the detoxication of benzene.

26Global gene expression profiling to evaluate the transcriptional changes in the prostate of men at high risk for prostate cancer has revealed that consumption of broccoli-rich diets for 6- or 12 months associates with transcriptional changes in signaling pathways involved in inflammation and carcinogenesis in the prostate tissue; importantly, these changes are dose-dependently attenuated in subjects receiving the glucoraphanin-rich diet. Other pharmacodynamic effects of interventions with glucoraphanin/sulforaphane in humans include: Increase in the levels of reduced glutathione in brain, enhanced integration of fatty acid β-oxidation with TCA cycle activity, protection against skin erythema caused by exposure to ultraviolet radiation, reduction in plasma LDL-cholesterol, decrease in the levels of fasting blood glucose and glycated hemoglobin in obese patients with dysregulated type 2 diabetes, and improvements in social interaction, behavior, and verbal communication in young men with autism spectrum disorder. Overall, although the precise molecular mediators are not always known, it is clear that interventions with glucoraphanin/sulforaphane-rich broccoli preparations in humans lead to diverse beneficial effects.

27Broccoli-based preparations consist almost exclusively of either glucoraphanin, sulforaphane, glucoraphanin with added active myrosinase, the raw, cooked, or dried vegetables themselves (either broccoli or broccoli sprouts), or extracts of broccoli seeds or sprouts—glucoraphanin-rich, sulforaphane-rich, or both. These studies are, of course, not all straightforward, and measurement of the broccoli-derived ingredients can be a source of mystery, obfuscation, and confusion, that is readily exploited by the supplement industry. Nonetheless, many clinical studies have now been done, and some overarching learnings can begin to be formulated. We and others have spent much time assessing bioavailability, conversion of glucoraphanin to sulforaphane, safety, and the classic ADME (absorption, distribution, metabolism, and excretion) pharmacokinetic parameters discussed in a previous section of this review. Efficacy has been studied much less thoroughly, but a picture is beginning to emerge. The following areas (in alphabetical order) have received special attention:

28The levels of aflatoxins from stored, subsistence seeds (corn and peanuts) in the Qidong area of coastal China near Shanghai, have in recent history been exceedingly high, as has associated incidence of liver cancer. Broccoli sprouts have been successfully piloted as a preventive intervention, and much data has been generated on safety and efficacy [biomarkers], as well as tolerance and acceptance by the local population;

29The natural laboratory in Qidong further served to study the effects of broccoli spouts on detoxication of volatile organic air pollutants. Others, in other environments, have examined effects of diesel exhaust (particulate) and other air pollutants;

30Early work by Healy and collaborators suggested an effect on shear stress in chondrocytes. This has now been followed by the elegant demonstration that increasing broccoli intake results in isothiocyanate (e.g., sulforaphane) uptake into the joint, with concomitant changes in the joint. Much other work on anti-inflammatory properties of sulforaphane may ultimately be related to arthritis;

31A variety of relatively small but significant effects of sulforaphane have been demonstrated in a number of relatively complex experimental systems;

32A large variety of biomarkers have now been reported following sulforaphane treatment. The cancers being studied include: Breast, lung; gastric, colo-rectal, prostate, skin, head and neck, and liver;

33A 90-subject, three-center RCT was conducted with sulforaphane-rich broccoli sprout extract in which no effect was seen on patients with advanced COPD. Post-facto reasoning suggested that the already severely degraded condition of subjects’ airways may well not have permitted for responsiveness (Nrf2-related and anti-inflammatory) regardless of the agent used. A follow-up analysis suggests that there was compartmentalization of anti-oxidant and anti-inflammatory gene expression in current and former smokers with COPD;

34At least three different groups have reported effects that included biochemical markers, blood pressure and flow-mediated dilation, and plasma metabolite biomarkers;

35Many clinical studies from at least three different groups have now been published. They have shown effects on reduced gluconeogenesis, inflammatory markers, and insulin resistance;

36Extensive measures in a variety of studies have documented upregulation of key chemoprotective enzymes. To name only a few studies not mentioned elsewhere in this review:;

37Our initial observations in vitro and in an animal model have been translated to some eradication and more importantly, reduction of the levels of H. pylori colonization and gastric inflammation in some infected individuals;

38Positive changes in liver function markers have been observed and much interest continues, based upon encouraging pre-clinical studies;

39Small studies showing effects on GSH levels and mapping GSH elevations to specific brain regions have generated much excitement;

40One successful and highly visible study showing improvement of conditions of autism spectrum disorder (ASD) following consumption of sulforaphane from broccoli sprouts is now being followed up with at least five studies with ASD subjects, two with schizophrenia, and one with psychoses. Results of only one of these follow-on studies has published so far;

41Although only a single, small, dose escalation Phase 1 trial has been run, a limited number of disease-relevant pharmacodynamic endpoints were queried, and safety was evaluated through a relatively high calculated dose.

42Table 1 summarizes the published literature of clinical trials conducted with broccoli-based preparations, sulforaphane and/or glucoraphanin. Also included are clinical trials listed on ClinicalTrials.gov that are known to be underway. Pending trials are not included. Oral doses that delivered sulforaphane directly ranged from 9.9 to 847 µmol per person per day (median 100 µmol) and for glucoraphanin 25 to 800 µmol per person per day (median 190 µmol). The trend towards the use of higher doses of glucoraphanin versus sulforaphane reflect considerations on the part of study directors on the limited conversion of glucoraphanin to sulforaphane in the absence of exogenous myrosinase in beverage or dietary supplement preparations.

43The harsh taste (a.k.a. back-of-the-throat burning sensation) that is noticed by most people who consume higher doses of sulforaphane, must be acknowledged and anticipated by investigators. They must make accommodations for what some subjects may consider a highly objectionable taste. This is particularly so at the higher limits of dosing with sulforaphane, and not so much of a concern when dosing with glucoraphanin, or even with glucoraphanin-plus-myrosinase. It is this harsh or burning sensation that has led in part to the characterization of the glucosinolate-myrosinase-isothiocyanate system as “the mustard oil bomb”. This should not, however, be confused with the bitter descriptor leveled at most brassica vegetables and moringa, that appears to have more to do with the TAS2R class of taste receptors.

44The presence and/or enzymatic production of levels of sulforaphane in oral doses ranging above about 100 µmol, creates a burning taste that most consumers notice in the back of their throats rather than on the tongue. This prevents some from consuming broccoli/sulforaphane preparations. Higher doses of sulforaphane lead to an increased number of adverse event reports, primarily nausea, heartburn, or other gastrointestinal discomfort. We have conducted scientifically guided studies to mask or distract consumers from that very distinctive taste, and to facilitate the development of proper placebos.

45As an ultimate public health paradigm, eating plants directly, with or without cooking or minimal processing, must be the solution in most of the world. The wealthiest regions can afford supplements as preventive interventions. Attempting to turn a chemopreventive approach into a pharmaceutical strategy would be foolhardy, but to the extent that sulforaphane is proven to have curative or therapeutic properties, this could be a direction in which to turn. From the perspective of designing and conducting clinical studies to demonstrate preventive efficacy, many issues come to bear on the subject. Though we have dealt with them, certain of them bear repeating:

46If it is indeed plant foods that will be the ultimate delivery vehicle, then food companies which will share in an enormous upside to successfully enhanced health-span, must bear some of the costs of conducting these trials. Though much has been made of the size and lobbying muscle of “big pharma”, (roughly a trillion dollar industry worldwide), the food sector of the economy in the USA alone is now valued at about $1.1 trillion.

47Standardization of dose delivery is extremely difficult when conducting clinical studies with foods. As sulforaphane gains scientific credibility as a preventive phytochemical, the clinical studies aiming to determine how, how much, for how long, and how to deliver and measure it, will by definition grow larger, longer, and more sophisticated. Perhaps there will come a time when foods such as broccoli will be labeled with variety name (e.g., like the maddening little stickers on apples) and glucoraphanin titer to enhance the consumer’s ability to make direct health-span purchasing choices. In the meantime, the clinical studies that we have performed with broccoli and broccoli sprouts have already strained the academic system to the breaking point. The food industry needs to step up.

48We and others have switched to the use of sulforaphane-containing or sulforaphane-generating supplements as discussed herein. However, the use of commercial supplements is fraught with dangers since there are many poor supplements and many unscrupulous supplement purveyors. Robust quality control is essential. Thus, better industry monitoring and regulation, as well as rigorous validation by investigators, of the labeled “doses”, is critical. One cannot for the foreseeable future trust that what appears on the label is an accurate representation of “dose” (and sometimes plant entity), and in Table 1 we have indicated clinical studies in which label results have been used rather than making dose measurements prior to or during intervention.

49There must be clinician buy-in to the concept that “thy food is thy medicine, and thy medicine is thy food”. It has been attributed, perhaps fallaciously, to Hippocrates, but regardless of its origin, the physicians who are Hippocrates’ intellectual descendants, have not risen to take the bait. They have instead taken the bait of big pharma, hook-line-and-sinker, and it is threatening to sink our economy. Pharmaceuticals most certainly have their place, but we must recruit the physicians in our society, whom people still trust in matters of their critical and preventive care, to more firmly place at least one foot in the dietary side of the aisle.

50Several key challenges in the design of clinical chemoprevention trials, especially food-based trials, are the selection of the dose, formulation, and dose schedule of the intervention material. Dose schedule is typically delineated by the convenience of a daily schedule and knowledge of the short biological half-life of sulforaphane. The other factors are less circumscribed. Selection of dose is complicated by the very different bioavailability of sulforaphane when administered in the precursor form of glucoraphanin and when given as sulforaphane itself, as discussed earlier. Figure 4 summarizes our experiences from four clinical trials conducted in Qidong, China with broccoli-based interventions enriched with glucoraphanin, sulforaphane, or both.

51In a cross-over design study we observed recoveries of excreted sulforaphane metabolites (principally sulforaphane-N-acetylcysteine) to be about 5% when a glucoraphanin-rich beverage was administered with a range of 1 to 45% but about 70% with a much narrow range of inter-individual variation when a sulforaphane-rich beverage was used. Thus, bioavailability changes dramatically as a function of the source material—precursor or bioactive. Some of our subsequent studies used blends of sulforaphane- and glucoraphanin-rich preparations to improve bioavailability together with tolerability and to examine dose-response relationships. Higher intake was achieved, but with little impact on inter-individual variability. Formulation, which in turn reflects how broccoli sprout extracts are prepared (e.g., with or without exogenous myrosinase-catalyzed hydrolysis of glucoraphanin), strongly affects bioavailability. Using a dietary supplement formulation of glucoraphanin plus myrosinase (Avmacol®) in tablet form, we observed a median 20% bioavailability with greatly dampened inter-individual variability. Fahey et al. have observed approximately 35% bioavailability with this supplement in a different population. Thus, reporting of administered dose of glucoraphanin and/or sulforaphane is a poor measure of the bioavailable/bioactive dose of sulforaphane. As a consequence, we propose that the excreted amount of sulforaphane metabolites (sulforaphane + sulforaphane cysteine-glycine + sulforaphane cysteine + sulforaphane N-acetylcysteine) in urine over 24 h (2–3 half-lives), which is a measure of “internal dose”, provides a more revealing and likely consistent view of the delivery of sulforaphane to study participants. In turn, use of “internal dose” metrics will facilitate optimization of the linkage between formulation, dose, and schedule with determinants of efficacy and, importantly, allow more facile comparisons of results between different clinical trials.

52Epidemiological studies have strongly hinted at beneficial associations between crucifer or more specifically broccoli consumption and mitigation of disease risks. However, it is the animal studies that have provided a robust foundation for the imperative to conduct clinical trials with broccoli, broccoli sprouts and derived-preparations or sulforaphane itself. As summarized in the Supplemental Tables S1–S4, there are a breadth as well as depth of studies describing roles for sulforaphane in the prevention and treatment of diseases in mouse and rat models.

53Yet, animal studies have not delivered all that might be expected of them. It is clear from the data presented in Figure 5 that the pre-clinical experimentalists have not thought carefully about the selection of dose (or route) and its relevance to clinical utility. Using oral dosing in mice and rat studies as examples, over two-thirds of the animal studies have used doses that exceed the highest (and bordering on intolerable) doses of sulforaphane used in humans, even after accounting for allometric scaling between rodents and humans. Few studies have included a dose-response; thus, the greater than 4-log spread of doses used in mice appears to be driven by needs for effect reporting in publications rather than optimization of translational science. Authors of this review have contributed to this dose skewing, among many investigators.

54How were initial doses selected for clinical trials with broccoli-based preparations? Not from any animal studies, which were already in abundance when the first in human studies were conducted. Not from pharmacological studies in rodents to define peak plasma levels, half-lives, metabolic fate or tissue distribution of sulforaphane. Rather, after initial, careful quantitative dose-escalation pharmacokinetic studies by Talalay and colleagues with crucifers, clinical trialists quickly sought to push doses to the extremes of tolerability by attempting to mask the bitter taste of the simple enriched formulations in order to identify pharmacodynamic action. Only recently have there been attempts to define minimally effective doses in humans—an outcome made possible by the development of consistently formulated, stable, bioavailable broccoli-derived preparations. The disconnect between animal and human studies has been exacerbated because of the public health-based desire (and regulatory-driven need) to conduct the initial human studies with broccoli-based preparations, that in hindsight varied greatly in sulforaphane yield and bioavailability, and the expedient use of pure sulforaphane in most animal studies. Feeding animals lyophilized broccoli or powdered broccoli extracts were likely disfavored because of the need for high percentages within the diet, which could skew overall nutrient uptake. Preparations with stabilized sulforaphane now entering clinical use will close the gap in formulation and could be used to provide more quantitative insights into dose-response. Regardless, experimentalists need to now pay close attention to examining efficacy—and underlying mechanisms—using dose and schedule equivalencies to those that are attainable and tolerable in humans—and arguably with oral and not intraperitoneal routes of administration. This notion also extends to in vitro studies, whether in human or animal cell lines, that often report the use of sulforaphane concentrations in far excess of what is likely achievable in human tissues as peak levels, never mind the short biological half-life in vivo that quickly attenuates the duration of the exposure.

55While there has been substantial progress in the refinement of sourcing and formulation to optimize the pharmacokinetics of broccoli-based preparations now used in clinical trials, there is nonetheless limited information on the factors that contribute to the wide inter-individual variation in pharmacokinetics seen in many clinical trial participants. Host factors (genetic polymorphisms and epigenetics) as well as host-microbiome interactions likely play important roles in this variability. Additionally, there is a striking need to develop rigorous biomarkers of pharmacodynamic action. Better links between purported mechanisms of action delineated in the pre-clinical settings and functional assessments of clinical efficacy are needed to optimize interventions and to better identify those healthy or at-risk groups that might best benefit. With more than 50 papers published on clinical studies with broccoli, glucoraphanin-rich and/or sulforaphane-rich broccoli beverages, powders and dietary supplements, as well as stabilized formulations of sulforaphane, there is ample optimism for disease preventing or mitigating applications to continue evolving. The additional listings of more than 50 proposed/ongoing trials in ClinicalTrials.gov also bears witness to such expectations. Short-term markers of pharmacodynamic actions—likely supported by -omics platforms - are needed to serve as guideposts for rapid deployment of existing and emerging knowledge for the betterment of health using broccoli-based strategies.