Allium vegetables and organosulfur compounds: do they help prevent cancer?

Allium vegetables have been shown to have beneficial effects against several diseases, including cancer. Garlic, onions, leeks, and chives have been reported to protect against stomach and colorectal cancers, although evidence for a protective effect against cancer at other sites, including the breast, is still insufficient. The protective effect appears to be related to the presence of organosulfur compounds and mainly allyl derivatives, which inhibit carcinogenesis in the forestomach, esophagus, colon, mammary gland, and lung of experimental animals. The exact mechanisms of the cancer-preventive effects are not clear, although several hypotheses have been proposed. Organosulfur compounds modulate the activity of several metabolizing enzymes that activate (cytochrome P450s) or detoxify (glutathione S-transferases) carcinogens and inhibit the formation of DNA adducts in several target tissues. Antiproliferative activity has been described in several tumor cell lines, which is possibly mediated by induction of apoptosis and alterations of the cell cycle. Allium vegetables and organosulfur compounds are thus possible cancer-preventive agents. Clinical trials will be required to define the effective dose that has no toxicity in humans.

The Allium genus includes approximately 500 species, the most widely used of which are onions (Allium cepa), garlic (Allium sativum), leeks (Allium porrum), chives (Allium schoenoprasum), and shallots (Allium ascalonicum). Garlic is widely cultivated and consumed worldwide, and its beneficial effects have been known for thousands of years. It has been considered to increase longevity and to confer stamina and physical strength, and it has been used empirically as a vermifuge, antiseptic, antimicrobic, antipyretic, and analgesic. The scientific community has now become interested in the pharmacologic properties of Allium vegetables and their chemical constituents, particularly with regard to their effects on the cardiovascular system and in the prevention of cancer. Garlic has antihypertensive and antiarrythmic properties and exerts an antithrombotic effect through fibrinolytic activity and the reduction of platelet aggregation. Ingestion of garlic has also been reported to lower the concentration of triglycerides, cholesterol, and low-density lipoproteins and to increase the concentration of high-density lipoproteins in blood. These findings suggest that garlic has a preventive effect against atherosclerosis and its complications, including stroke, myocardial infarction, and thrombotic disorders. The findings in relation to cancer prevention that have been accumulated over the last 20 years are summarized here.
We performed a systematic Medline search of the published literature on epidemiologic studies involving the relationship between consumption of Allium vegetables-garlic, onions, and leeks-and neoplasms/tumors. For the animal studies and for studies on mechanisms of action, we conducted searches for review articles. These lists of references were then updated using references from review papers and original research articles.

Active Compounds in Allium Vegetables
Fresh garlic contains water, carbohydrates, proteins, fiber, and fat, as well as essential amino acids, vitamins, and minerals. When garlic is cut, chopped, or crushed, the clove's membrane is disrupted and S-allylcysteine sulfoxide (an odorless compound called alliine) is transformed enzymatically into allicin by alliinase (1). Allicin is responsible for the typical odor of garlic, but it is unstable and converts readily into mono-, di-, and trisulfides and other compounds such as ajoene. Onions mainly contain S-propenylcysteine sulfoxide (1), but also other sulfoxides, including S-propylcysteine sulfoxide and S-methylcysteine sulfoxide (2). S-Propenylcysteine sulfoxide, positional isomer of alliin, is called lacrimatory precursor, because it is tranformed by alliinase into the lacrimatory factor propanethial S-oxide. The lacrimatory factor is highly reactive and hydrolyzes to propionaldehyde, sulfuric acid, and hydrogen sulfide; it is also the precursor of several sulfur derivatives. Organosulfur compounds present in Allium vegetables, which are either lipid or water soluble (Table 1), are considered responsible for the beneficial effects of these herbs. Garlic derivatives generally have a thioallyl moiety, whereas onion extracts contain a thiopropyl group with somewhat different chemical properties. The amounts of volatile compounds present in garlic bulbs are reported in Table 2.
Further transformation of organosulfur compounds can occur after interaction with free sulfhydryl groups, including those present in cysteine, glutathione, or proteins (4,5). Incubation of cystein with allyldisulfide or diallylsulfide groups produces Sallylmercaptocysteine and allylmercaptane, respectively (6). S-Allylmercaptocysteine is further transformed to allylmercaptane following incubation with fresh blood. One study (7) in the isolated, perfused rat liver showed that allicin is first metabolized to diallyldisulfide, which is further reduced to allylmercaptane. Another study identified allylmercaptane and allylmethylsulfide as metabolites of diallyldisulfide and diallylsulfide in primary rat hepatocytes (8). Urine from healthy individuals consuming garlic or onions contains N-acetyl-S-allylcysteine and N-acetyl-S-(2-carboxylpropyl)-cysteine (9)(10)(11). Allylmercaptane and diallyldisulfide have been found in human breath after ingestion of garlic (12,13).

Epidemiologic Studies
The association between consumption of Allium vegetables and risk for cancer has been assessed in several epidemiologic, mainly case-control, studies. One of the first indications of a role of Allium vegetables in the prevention of stomach cancer came from China. The 10-fold difference in the death rate from stomach cancer in two Chinese provinces was suggested to be attributable to garlic consumption, which was approximately 20 g/day in the low-risk area and < 1 g/day in the high-risk area (14). This hypothesis was further supported by the results of an ecologic study in China (15), which reported that 82.3% of men and 74.7% of women in a low-risk area consumed garlic more than three times per week, whereas only 0.5% of the population in a high-risk area did so. Similar differences were found for the consumption of Welsh onions, onions, and Chinese chives.
Several case-control studies in Asia and Europe have been conducted to assess the effect of Allium vegetables on the risk of stomach cancer [ Table 3 (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)]. Most reported a decreased risk with increasing consumption. In one study (17) a higher risk for stomach cancer was found with increasing frequency of onion consumption, but the controls had severe gastric diseases, including chronic or acute gastritis, gastric ulcer, and polyps. No association with onion consumption was found in two other studies (19,23). In a multicenter case-control study in areas of high and low risk for stomach cancer in Italy (20), use of garlic and onions as condiments decreased the risk of stomach cancer. When questions on the frequency of garlic consumption were added to a questionnaire given to a subgroup of subjects, an inverse association was observed between consumption of raw garlic and cancer risk, persons in the highest tertile of intake having only 40% of the risk of those in the lowest. The effects of different Allium vegetables have been evaluated separately in a few studies. You et al. (21) showed that patients with gastric cancer consumed approximately 3.5 kg less Allium vegetables per year than controls (15.5 vs. 19). Each Allium vegetable was associated with a decreased risk, which was most marked for scallions, garlic, and Chinese chives. Similar results were obtained when the usual diet 25-30 years before diagnosis (before the Cultural Revolution) was considered. In a population-based study, the age-and sex-adjusted odds ratio for stomach cancer decreased inversely with the frequency of consumption of garlic, onions, Welsh onions, and Chinese chives (26). This study also showed a decreased risk for esophageal cancer with a higher frequency of consumption of individual Allium vegetables [ Table 4 (26,28,29)].
A cohort study in the Netherlands addressed the effect of daily use of onions, leeks, and garlic supplements for at least 1 year in the 5 years before baseline on the risk for stomach cancer (27) and for colorectal, breast, and lung cancers. Overall, cases were associated with lower consumption of onions and leeks, but slightly greater use of garlic supplements (9.2% vs. 8.8%). Onion consumption was significantly inversely correlated with the risk of stomach cancer, with a statistically significant test for trend (Table  3). Slightly different results were obtained in stratified and multivariate analyses, but the trends remained statistically significant. Leek consumption was associated with a decreased risk only for persons in the highest category of consumption, but there was no clear decreasing trend. The relative risk associated with use of garlic supplements was slightly greater than 1.0. Similar results were obtained when cases occurring in the first year of follow-up were excluded from the analysis and when only subjects without a history of gastric disorders were considered. This study has the advantages of good completeness of follow-up and adjustment for several dietary and nondietary confounding factors, although no distinction was made between raw and cooked onions and leeks.
The chemopreventive effects of Allium vegetables against stomach and esophageal cancers may be related to their antibacterial properties. Inhibition of bacterial growth in the gastric cavity may result in less conversion of nitrate to nitrite in the stomach (14), a decreased probability of endogenous formation of carcinogenic N-nitroso compounds, and reduction in Helicobacter pylori infection specifically. In vitro, garlic oil and extracts of individual organosulfur compounds strongly inhibit the bacterial growth of H. pylori (30)(31)(32)(33). In contrast, some studies in humans showed no correlation between seropositivity to H. pylori and the consumption of garlic (34) or use of garlic to treat H. pylori infection (35).
An intervention trial is being conducted in China in which capsules containing steam-distilled garlic oil and aged garlic extract (in addition to amoxicillin and omeprazole or antioxidants) are given to determine whether they reduce the prevalence of dysplasia and other preneoplastic lesions in the stomach (36). No results have yet been published.
Studies of the effects of Allium vegetables against colorectal cancer are summarized in Table 5 (17,(37)(38)(39)(40)(41)(42)(43)(44)(45)(46). Apart from the study of Tajima and Tominaga (17) discussed above, the case-control studies generally found decreased risks for colorectal cancer. The effect was particularly significant for consumption of cooked onions and leeks in Belgium (38), for a combination of garlic, onions, and pepper in Argentina (40), and for garlic in Switzerland (43). In a case-control study in Australia, Steinmetz and Potter (41) reported a lower risk for both sexes, with a more pronounced decrease for women and for cancer of the proximal compared with the distal colon. Nevertheless, two of the three cohort studies reported decreased risks for cancer of the distal colon with higher consumption of garlic (44,45). The third cohort study showed no significant effect of consumption of onions, leeks, and garlic supplements (46).
Studies of the association between consumption of Allium vegetables and risk for cancers at other sites are summarized in Table 7 (51-57). In case-control studies, significantly decreased risks with onion consumption for cancer of the lung in India (51) and of the brain in China (56) were reported. A significantly decreased risk for cancer of the larynx with garlic consumption was reported in China (53). A case-control study in England (55) reported a significant protective effect against prostate cancer when garlic was eaten as a food and as a food and supplement, but no effect was seen with onions. However, adjustment for social class reduced the association with garlic to nonsignificance. A moderate protection of endometrial cancer was also reported with Allium vegetable consumption (57).
Studies summarized above can suffer from some limitations, including use of hospitalbased controls, low response rates among cases or controls, low rates of histologic confirmation of the tumors, use of recent diet as a reference, use of frequency instead of amount consumed, bias due to differential recall of dietary intake associated with awareness of disease, and absence of information on vegetable preparation (i.e., raw or cooked). These factors could explain some inconsistencies in Reviews • Allium vegetables and cancer prevention Environmental Health Perspectives • VOLUME 109 | NUMBER 9 | September 2001 895 the results. Publication bias, with limited number of published data with no apparent protective effect, could also be operating and detracting from the evidence of an effect. However, studies assessing the role of Allium vegetables in cancer prevention have generally not been designed for this purpose, and it is therefore possible that finding no effect for Allium vegetables did not really affect the publication. For calculation of odds ratios, the data were generally adjusted for age and sex and other variables, including socioeconomic status, anthropometry, and family history of gastric disorders and cancer. Alcohol, smoking, and energy intake were also taken into account in some studies. However, consumption of other food items, including vegetables, was adjusted for in few studies (23,26,27,43,46,50,52). Such adjustment seems essential in studies of gastric cancer, because low consumption of green vegetables and fruits has been described as an important risk factor for this cancer. Epidemiologic studies summarized here are generally very consistent in reporting a protective effect of Allium vegetables in gastric cancer, and the adjustment for fruits and vegetables was made in studies showing either protection (26,27) or lack of effect (23). It seems, therefore, that such adjustment is not the only factor explaining differences in results.
Garlic contains compounds that inhibit the initiation step of carcinogenesis at a variety of sites in different animal species and with different initiators. Diallylsulfide (DAS) generally inhibited cancers of the forestomach, colon, esophagus, mammary gland, and lung (58,60,64,65). Contradictory results were obtained with regard to the prevention of liver cancer (62,69). Other allyl derivatives, including diallyldisulfide, allylmethyldisulfide, and allylmercaptan (DADS, AMDS, and AM, respectively), decreased tumor formation in the forestomach and lung, but the related saturated propyl derivatives propylmethyldisulfide, propylmethyltrisulfide, dipropylsulfide, and dipropyltrisulfide (PMDS, PMTS, DPS, and DPTS, respectively) did not (58,59). Diallyltrisulfide (DATS) and allylmethylsulfide (AMTS) inhibited forestomach cancer, DATS, which contains two allyl groups, being more potent than AMTS, the analogue derivative with only one allyl group (58). This finding underlines the importance of the allyl group for cancer-preventive activity. These trisulfide derivatives, AMTS and DATS, were not effective against pulmonary adenoma formation, suggesting that the number of sulfur atoms in the molecule is also important, possibly determining the organ sites at which protection is achieved against carcinogenesis (58). DADS was also protective against cancers of the colon, mammary gland, and kidney (62,63,65,67). S-Allylcysteine (SAC) inhibited the incidence and frequency of 1,2-dimethylhydrazine (DMH)-induced colon tumors (61). Contradictory results were reported with SAC for N-methylnitrosourea (MNU)-induced mammary carcinogenesis (67,68). These inconsistencies could be explained by differencies in the doses of MNU and SAC used and the type of diet administered. Garlic powder in the diet inhibited mammary tumors (65)(66)(67), and a garlic extract decreased the incidence of cervical carcinoma (70). DAS and garlic oil applied topically during the initiation phase of skin cancer reduced the number of tumors per mouse (72,74). In addition, garlic extract and garlic and onion oil decreased the promotion step of skin carcinogenesis induced by 12-O-tetradecanoylphorbol 13acetate (71,73). The protection provided by garlic against carcinogenesis may be modulated by various dietary components, including selenium, vitamin A, and lipids (65,75,76).
The finding of lack of cancer-preventive effect of organosulfur compounds could have affected the publication of the study; however, the protective effect seems to be very consistent, and publication bias appears unlikely.

Mechanisms of Cancer Prevention
Several mechanisms have been proposed to explain the cancer-preventive effects of Allium vegetables and related organosulfur compounds. These include inhibition of mutagenesis, modulation of enzyme activities, inhibition of DNA adduct formation, free-radical scavenging, and effects on cell proliferation and tumor growth. Although there is evidence supporting these mechanisms for organosulfur compounds, they are still speculative, and further research is needed to support a causality between such properties and the cancer-preventive activity in experimental animals.
Inhibition of mutagenesis. Aqueous and methanolic garlic extracts inhibited the mutagenic activity of aflatoxin B 1 in Salmonella typhimurium (77). Aqueous garlic extract also decreased the mutagenicity of 4-nitroquinoline-1-oxide in Escherichia coli (78) and the mutagenicity of γ-radiation, hydrogen peroxide, cumene, and t-butyl hydroperoxides in S. typhimurium (79). Onion juice protected S. typhimurium against cooked mutagens present in hamburgers (80).
Modulation of enzyme activities. Organosulfur compounds have been shown to modulate the activity of glutathione Stransferases (GST), a family of enzymes important in detoxification of carcinogens, and cytochromes P450 (CYP), a family of enzymes that activate many chemical carcinogens in experimental animals.
Sparnins and co-workers (81) first showed that AMTS increased the activity of  GST in the forestomach, small-bowel mucosa, liver, and lung of mice. Other allyl derivatives also increased GST activity in these tissues (58). Derivatives with a propyl instead of an allyl group were less active or inactive. The induction of GST paralleled the inhibition of benzo[a]pyrene-induced carcinogenesis in the forestomach, but not in the lung, suggesting that increased carcinogen detoxification is only one of the factors responsible for the cancer-preventive effects of organosulfur compounds. These results were partially confirmed by Sumiyoshi and Wargovich (61), who found a greater effect of thioallyl than thiopropyl derivatives in inducing hepatic and colonic GST in mice. In contrast, DAS did not increase GST activity in mouse liver (60) or in a culture of rat hepatocytes (69). The activity of mammary and liver GST was increased by the addition of garlic powder to the diet of rats (66). The maximum activity of GST did not coincide with maximum inhibition of carcinogenesis, however, further indicating that increased GST activity does not account fully for the protection provided by garlic powder against carcinogenesis. Thus, the effects on enzymes that activate chemical carcinogens are not sufficient to explain the cancer-preventive activity. For example, an oral dose of DAS suppressed

Reviews • Allium vegetables and cancer prevention
Environmental Health Perspectives • VOLUME 109 | NUMBER 9 | September 2001 897 esophageal carcinogenesis induced by Nnitrosomethylbenzylamine in rats and significantly reduced the microsomal conversion of this nitrosamine in liver but not in esophagus (64). In addition, the prevention of benzo[a]pyrene-induced forestomach cancer in mice by organosulfur compounds is not attributable to a reduction in the activity of CYP1A1 (82). DADS in the diet increased not only the activity of GST but also that of other detoxifying enzymes, including reduced nicotinamide adenine dinucleotide phosphate [NAD(P)H]dependent quinone oxidoreductase, which is involved in detoxification of activated quinone metabolites of benzo[a]pyrene, and of uridine diphosphate (UDP)-glucuronosyl transferase in rat tissues (63,83).
DAS acted as a competitive inhibitor of N-dimethylnitrosamine demethylase activity (84). It also decreased the activity of CYP2E1 in a time-and dose-dependent manner and induced the activities of CYP2B1 and pentoxy-and ethoxyresorufin dealkylases in hepatic microsomes (85). An increase in CYP2B1 mRNA was also observed. Treatment with the DAS metabolites diallyl sulfoxide (DASO) and diallylsulfone (DASO 2 ) had similar effects on rat hepatic monooxygenase activities (85,86). Reicks and Crankshaw (87) reported that DAS, DADS, and AMS decreased p-nitrophenol hydroxylase activity and CYP2E1 protein concentration in rat liver. When the diet of rats was supplemented with DAS or DADS (88), DADS increased the activities of several monooxygenases and transferases in intestine and liver; the protein levels of epoxide hydrolase and CYP2B1/2 were also increased. DADS also decreased CYP2E1 level in liver. The effects of DAS were similar to those of DADS in liver, but only epoxide hydrolase activity and CYP2B1/2 protein levels were increased in the intestine.
In a study of the effect of garlic oil, DAS, and DADS on the activities of several metabolizing enzymes in the liver of rats fed high-fat diets (89), GST activity was increased by all treatments. Garlic oil induced the expression of the placental form of GST and CYP2B1 and decreased the expression of CYP2E1. DAS and DADS also modulated these enzymes, but DAS increased mainly CYP2B1, whereas DADS increased mainly GST activity; similar effects were observed on CYP2E1 expression.
DAS and its oxidation derivatives DASO and DASO 2 are conjugated with glutathione, in rats (90). To our knowledge, no study has investigated the effects of possible GST polymorphisms in the deactivation of these Allium vegetable-derived compounds, although this could provide some explanations of differential effects in humans.
Modulation of the activity of arylamine N-acetyltransferase, a polymorphic enzyme that deactivates arylamines and activates some heterocyclic dietary amines, was addressed in a few studies. The slow and fast acetylator phenotypes have been associated with increased risk for cancers of the bladder and colon, respectively. DAS and DADS decreased the activity of this enzyme in strains of H. pylori from peptic ulcer patients (31) and inhibited its activity in a human colon tumor cell line (91) and in human bladder tumor cells (92) in a dose-dependent manner.
Inhibition of DNA adduct formation. DNA adducts are believed to be an initial step in carcinogenesis by chemicals. In rat mammary gland, garlic powder decreased the occurrence of 7,12-dimethylbenz[a]anthracene (DMBA)-DNA adducts in vivo and the amounts of total and individual adducts correlated positively with mammary tumor incidence (66). Garlic powder, garlic water extract, a deodorized garlic powder, a garlic powder with a high sulfur content, and SAC were also effective against mammary DMBA-DNA binding (93).
DNA adducts induced by incubation of human bladder tumor cells with 2-aminofluorene were inhibited by DAS and DADS (92). In contrast, a water extract of raw garlic and SAC, but not DAS, significantly inhibited benzo[a]pyrene-DNA adduct formation in simulated human peripheral blood lymphocytes in vitro (94).
N-Nitroso compounds, a class of potential human carcinogens that can be synthesized in humans from precursors present in the diet, are metabolized to alkylating agents that can bind to DNA. Shenoy and Choughuley (95) showed that onion and garlic juices inhibit the nitrosation reactions in vitro in a dose-dependent manner. The occurrence of 7-methyldeoxyguanosine (7-MedG) and O 6 -methyldeoxyguanosine (O 6 -MedG) was decreased in rat liver when garlic powder was added to a diet containing aminopyrine and sodium nitrite (96). Garlic powder also decreased DNA methylation in the livers of rats treated with N-nitrosodimethylamine and in mammary tissue of rats treated with N-methylnitrosourea. Garlic, SAC, and DADS also decreased the formation of 7-MedG and O 6 -MedG induced by N-methylnitrosourea in mammary DNA; this decrease correlated with the inhibition of mammary tumors by these compounds (67).

Reviews • Bianchini and Vainio
898 VOLUME 109 | NUMBER 9 | September 2001 • Environmental Health Perspectives Free-radical scavenging. Free radicals have been related to several age-related diseases, including cancer (97). Reduced glutathione (GSH) is not only a cofactor for GST but also serves as a reductant for glutathione peroxidase (GPX), an enzyme involved in natural protection by free radicals, in addition to superoxide dismutase and catalase. Garlic and onion oils stimulated the activity of GPX and inhibited the decreased ratio of reduced to oxidized glutathione produced by 12-O-tetradecanoylphorbol-13acetate in epidermal cells (98). GPX activity was also increased in animal tissues with DAS (99), DADS and garlic oil (89). DAS and DADS also increased the activity of glutathione reductase, and garlic oil increased the activity of superoxide dismutase (89). In contrast, DAS and garlic homogenates decreased catalase in the livers of rats and mice (100). S-Allylmercaptosysteine (SAMC) and SAC increased the synthesis of GSH in human prostate cancer cells (101).
Aged garlic extract, SAC, and SAMC exhibited radical scavenging activity (102). DAS, DADS, and AMS showed selective actions on different markers in tests for their ability to react with carbon tetrachloridederived free radicals (103). DADS also inhibited carbon tetrachloride-induced lipid peroxidation. The antioxidant properties of Allium vegetables might therefore result from the contributions of various sulfur components at different steps of the process.
Effects on cell proliferation and tumor growth. Inhibition of tumor cell proliferation by organosulfur compounds has been reported in several studies using different cell cultures, including canine mammary tumor cells (104), human colon, lung, and skin tumor cell lines (105,106), human neuroblastoma cells (107), human and murine melanoma cells (108), and human prostatic carcinoma cells (101).
Contradictory results have been obtained with regard to modulation of the proliferative activity of non-neoplastic cell lines by organosulfur compounds, with some studies showing inhibition (108,109) and others no effect (106).
Garlic and onion oils caused a marked suppression of proliferation of human promyelocytic leukemia cells (110). Garlic powder and an alliin-enriched garlic extract inhibited the growth of a human lymphatic leukemia cell line in a dose-dependent manner, but inhibited the growth of human hepatoma and human colorectal carcinoma cells only when applied as a mixture. This finding indicates that the antiproliferative effect of garlic is due to breakdown products of alliin catalyzed by the alliinase enzyme system present in garlic powder (111).
Polyamines, mainly spermine, play an important role in cell division and differention. SAMC, but not SAC, has been shown to alter polyamine concentrations in human prostate carcinoma cells, increasing that of spermidine and decreasing those of putrescine and spermine (101). Ornithine decarboxylase, a rate-limiting enzyme involved in the synthesis of polyamines, is also reduced by DAS (98,112,113), although there is evidence of an increase in the livers of rat not treated with initiators (114). The antiproliferative effect of organosulfur compounds appears to be related to the induction of apoptosis. Exposure to DADS and DATS caused cells to undergo apoptosis, as determined by morphologic changes and/or DNA fragmentation (105,106). A positive correlation was found between DADS-induced DNA fragmentation and increased intracellular free-calcium concentration, which may activate calcium-dependent endonucleases leading to apoptosis. A recent study (115) showed that DAS, DADS, and garlic extract increase the number of non-small-cell lung cancer cells in the apoptotic state. This increase followed the induction of p53 protein by DADS or the increase of the expression of Bax and decrease of the expression of Bcl-2 by DAS and garlic extract. Ajoene induced apoptosis in human leukemic cells but not in peripheral mononuclear blood cells from healthy donors (116). The cell cycle was also affected by DADS, which decreased the percentage of human colon tumor cells in the G 1 and S phases and concomitantly increased the percentage of those in the G 2 /M phase (117). These effects depended on the dose of DADS and the length of incubation. The ability of DADS to inhibit cell proliferation was related to induction of G 2 /M phase arrest and to inhibition of p34 cdc2 kinase activity, which modulates the progression of cells from G 2 into the M phase of the cell cycle. The suppression of the p34 cdc2 kinase activity by DADS resulted not from a direct interaction with the protein but from modulation of the factors involved in the formation and conversion of the enzyme to its active form (118). DADS also significantly inhibited the growth of H-ras oncogenetransformed tumors implanted in nude mice by suppressing the association of p21 H-ras with the cell membrane (119).

Conclusions
Overall, evidence shows that Allium vegetables, mainly garlic and related organic allyl sulfur components, have cancer-preventive effects. In the 37 case-control or cohort studies for cancer at any site reported here, 28 showed some protective effect of Allium vegetables. The evidence is particularly strong for stomach cancer (9 out of 12 studies) and colon cancer (9 out of 11 studies), while there is still insufficient evidence for an effect on breast, lung, and other cancers. A recent meta-analysis also showed that a high intake of garlic may be associated with decreased risks for stomach and colorectal cancer (120).
However, the epidemiologic data do have some limitations, which decrease the strength of the evidence currently available. For example, consumption is often reported as frequency instead of amount consumed, and the consumption categories differ widely among studies. In addition, the meta-analysis (120) indicated that the apparent protective effect found between Allium vegetable consumption and cancers of the stomach and colon may have been overestimated, suggesting a possible publication bias. Lack of adjustment in the statistical analyses also limits the strength of the conclusion. For example, the data for stomach cancer were generally not adjusted for fruit and vegetable consumption, which are known to be protective and could therefore be a confounder in the analysis. Clinical trials are not currently available, but they could be useful for assessing the ability of garlic and other Allium vegetables to prevent cancer or preneoplastic lesions. There are difficulties in standardizing the real intake of Allium vegetables, and in estimating the composition in organosulfur compounds and other chemicals. In addition, these trials cannot use cancer as the end point and should therefore rely on surrogate intermediate biomarkers (121). Such end points are not fully validated, and further research is needed to develop and validate suitable biomarkers.
Studies in experimental animals indicate that the benefits of Allium vegetables are not limited to one species, tissue, or carcinogen. Organosulfur compounds can hinder activation of a carcinogen from its precursor, increase its metabolic detoxification, or Reviews • Bianchini and Vainio 900 VOLUME 109 | NUMBER 9 | September 2001 • Environmental Health Perspectives prevent its reaction with vulnerable target cells. Additional mechanisms include a delay or a reversion of the expression of malignancy by antiproliferative activity in tumor cells and modification of signal transduction mechanisms. These speculative mechanisms in animal models should be verified in human studies to establish a causative link between some molecular properties and the cancer-preventive activity. Extrapolation of the doses of pure chemicals that are effective in animals to their equivalents in terms of Allium vegetables leads to unrealistic estimates of the amounts that would have to be consumed by humans to benefit from the antitumor effects of these substances. For example, because the dose of DAS that is effective in animals ranges from 50 to 400 mg/kg bw, a person weighing 70 kg would have to consume 3.5-28 g of DAS per day. The concentration of DAS in garlic ranges from 0.03 to 0.1 mg/g; thus, 40-100 g of garlic would have to be consumed per day, corresponding to 10-100 garlic cloves per day. Epidemiologic studies have found a protective effect of consumption of much lower doses, the mean intake being 18.3 g/week (120).
Little is known about the stability of organosulfur compounds during cooking or about the metabolism, pharmacokinetics, and toxicity of the active compounds in humans. All these aspects require further attention if the use of garlic and other Allium vegetables are to be recommended for cancer prevention.