Isoflavones are contained in soybean or soy foods in two chemical forms, i.e., aglycones and glucosides. We investigated the difference in the absorption of soy isoflavone aglycones and glucosides in humans. After a single, low dose intake (0.11 mmol), the highest isoflavone concentrations in plasma were reached 2 and 4 h after ingestion of aglycones and glucosides, respectively; subjects were four men (41 y old) and four women (45 y old). The highest plasma concentration after aglycone intake was more than two times greater than that after glucoside ingestion. In a similar manner, we then compared the plasma isoflavone concentration profiles after intake of a single, high dose of isoflavones (1.7 mmol) in eight subjects (four men, 40 y old; four women, 47 y old) and found the highest plasma concentration after aglycone intake was more than five times higher than that after glucoside intake. In both high and low dose intake tests, the plasma concentration of genistein was significantly higher than that of daidzein despite the similar levels of intake. After long-term (4 wk) intakes (0.30 mmol/d), we also measured the plasma concentration of isoflavones (eight men, 45 y old). After 2 and 4 wk, these concentrations remained >100% higher after ingestion of aglycones than of glucosides. The isoflavone aglycones were absorbed faster and in greater amounts than their glucosides in humans. Isoflavone aglycone-rich products may be more effective than glucoside-rich products in preventing chronic disease such as coronary heart disease.

Soy isoflavonoids have attracted much attention because of their estrogenic activity. To study the intestinal absorption of the isoflavonoids, we investigated the cellular uptake and metabolism of genistein and daidzein and their glucosides, genistin and daidzin, by Caco-2 cell monolayers as a model of the human intestinal epithelium. When Caco-2 monolayers were incubated with genistein or daidzein at 10 micromol/L from the apical (mucosal) side, aglycone was lost from the apical solution for 2.0 h (P < 0.05) and the glucuronide/sulfates appeared at the level of 1-2 micromol/L. In the basolateral (serosal) solution, both intact aglycones and their glucuronide/sulfates increased (P < 0.05) with time and reached approximately 20 and 15% of the initial dose, respectively. The transport of their glucosides, genistin and daidzin, through Caco-2 monolayers was less than one tenth that of the aglycones. The cellular metabolism of genistein was compared with quercetin, kaempferol, luteolin and apigenin. Only genistein aglycone was transported intact to the basolateral solution. Transport was greater (P < 0.05) than that of flavonoid aglycones and was without an appreciable decrease of transepithelial resistance. Radical scavenging activity was not related to the susceptibility to conjugation of flavonoids/isoflavonoids. Affinity to the liposomal membrane was increased in the order of genistin = daidzin < daidzein < genistein << flavonoid aglycones. These results strongly suggest that isoflavone aglycones are taken up into enterocytes more efficiently than their glucosides because of their moderate lipophilicity. Furthermore, they are generally transported to the basolateral side in intact form in contrast to flavonoids, probably due to their unique isoflavonoid structure.

P-Glucosidases which hydrolyze isoflavone glucosides, daidzin and genistin, were partially purified from soybean cotyledon. Three isoforms of P-glucosidases (A,B,C) were separated on a CM-Sephadex C-50 ion exchange column. P-Glucosidase-B and C brought about nearly all of the hydrolyzing action on daidzin and genistin. P-Glucosidase-B and C were further purified with gel-filtration chromatography, and the partially purified enzymes were characterized at pH 5.0 and 45"C, (optimum conditions). Adding the p-glucosidase-B and C to soymilk at 45"C, both enzymes hydrolyzed daidzin and genistin and caused an increase in aglycones daidzein and genistein.

Daidzein and genistein, which are responsible for the objectionable flavor of soy milk, increased during the soaking of soybeans, the first step of soy milk manufacturing. The maximum production of these isoflavone compounds occurred at 50°C and at pH 6.0 depending upon the temperature and pH of the soaking water. The production of these substances was strongly inhibited by glucono-&lactone which was a competitive inhibitor of S-glucosidascs. It would appear, thcrcfore, that S-glucosidases were responsible for the production of daidzein and genistein during soaking.

The antiatherogenic effect of soy protein with intact isoflavones is well established, but the effects of isoflavones without soy protein have not been determined. We investigated the antiatherogenic effect of an isoflavone aglycone-rich extract (containing 429.4 mg/g isoflavone aglycones) without soy protein from fermented soy in cholesterol-fed rabbits. We fed 12-wk-old New Zealand white male rabbits diets containing 1 g/100 g cholesterol with 0, 0.33 or 1 g/100 g isoflavone aglycones for 8 wk. We also fed the rabbits a diet containing 1 g/100 g cholesterol with 1.09 g/100 g soy saponin-rich extract, a component other than isoflavone aglycones in the isoflavone aglycone-rich extract. Controls did not consume cholesterol, isoflavone aglycones or saponins. The isoflavone aglycone- and saponin-rich extracts did not affect the serum lipid profile of cholesterol-fed rabbits. The serum concentration of daidzein in its conjugated form was significantly higher in the high isoflavone group than in the low isoflavone group. The level of cholesteryl ester hydroperoxide (ChE-OOH) induced by CuSO(4) in plasma in the high isoflavone group was significantly less than that in the cholesterol group, and the ChE-OOH levels of LDL in the low and high isoflavone groups were significantly less than those in the cholesterol group. The ChE-OOH levels in plasma and LDL in the saponin group did not differ from the cholesterol group. In the aortic arch, the cholesterol concentration was significantly lower in the high isoflavone group, and malondialdehyde concentration was significantly lower in the low and high isoflavone groups compared with the cholesterol group; however these concentrations in the saponin group did not differ from those in the cholesterol group. The atherosclerotic lesion area of the aortic arch was significantly lower in the isoflavone groups (26.3% lower in the low isoflavone group and 36.9% lower in the high isoflavone group) than in the cholesterol group. The lesion areas were not different in the soy saponin and cholesterol groups. Immunohistochemical analysis revealed fewer oxidized LDL-positive macrophage-derived foam cells in atherosclerotic lesions in the aortic arch of isoflavone groups compared with that of the cholesterol group. These results suggest that the antioxidative action of isoflavones and their antioxidative metabolites inhibit the oxidation of LDL, thereby exerting an antiatherosclerotic effect.