Submitted to: American Journal of Clinical Nutrition
Publication Type: Peer reviewed journal
Publication Acceptance Date: 4/10/2007
Publication Date: 8/1/2007
Citation: Beisiegel, J.M., Hunt, J.R., Glahn, R.P., Welch, R.M., Menkir, A., Maziya-Dixon, B.B. 2007. Iron bioavailability from maize and beans: a comparison of human measurements with Caco-2 cell and algorithm predictions. American Journal of Clinical Nutrition. 86:388-396. Interpretive Summary: A simple, economical method of assessing the bioavailability of iron from foods, estimating how well food iron is absorbed and utilized, is needed to evaluate and improve crop varieties to prevent human iron deficiency. A cell culture model may predict iron bioavailability to humans, but direct comparisons with human measurements are lacking. Based on previous cell culture measurements showing iron bioavailability differences between two maize varieties and between white vs. colored beans, we directly compared human and Caco-2 cell culture results using identical meals. In one experiment, iron absorption by 26 women was tested with two Nigerian maize varieties. In a second experiment iron absorption by 13 women was tested with great northern vs. pinto beans. Each food was served with and without orange juice to provide vitamin C. In contrast to results with previous harvests, the cell culture results detected no difference in iron availability between the maize varieties. Maize varieties did not affect the percent iron absorbed by the women, but tended to affect total iron absorption slightly because of a modest difference in iron content between the maize varieties. Iron bioavailability differed between bean varieties when measured in cell culture, but not when measured in humans. As cell culture measurements predicted for all but the pinto beans, vitamin C increased human iron absorption from all foods by at least 3 times. In conclusion, cell culture results correctly predicted enhancement by vitamin C of iron bioavailability to humans from maize and great northern beans, but incorrectly predicted color-associated differences between bean varieties and their interaction with vitamin C. Further refinement of the cell culture model to correspond with the human results would provide an efficient and economical means of screening crop varieties for nutritional value to help prevent iron deficiency in developing regions of the world.
Technical Abstract: Background: An in vitro digestion/Caco-2 cell model may predict iron bioavailability to humans, but direct comparisons are lacking. Objective: To confirm previous in vitro iron bioavailability differences between two maize varieties and between white vs. colored beans, directly comparing human and Caco-2 cell results. Design: Two randomized, 2 x 2 factorial, 29-d experiments compared women’s iron absorption (n=26) from two maize varieties (ACR vs. TZB) or great northern vs. pinto beans (n = 13). Each food was served with and without orange juice to provide ascorbic acid. Nonheme iron bioavailability was determined from 2-wk whole body and erythrocyte retention of extrinsically added radioiron tracers, and compared with Caco-2 cell results from identical meals. Results: In contrast to results with previous harvests, in vitro results predicted no difference in iron availability between the maize varieties. Maize varieties did not affect percent iron absorption, but women tended (p = 0.06) to absorb 4-7 µg more iron/meal from ACR compared to TZB, because of a slightly higher corn iron content (0.8 vs 0.7 mg/meal, respectively, p <0.0001). Contrary to in vitro predictions, iron absorption did not differ between bean varieties. As predicted in vitro for all but the pinto beans, Ascorbic acid increased iron absorption from all foods by at least 3 times (p<0.0001). Conclusion: Caco-2 cell results correctly predicted ascorbic acid enhancement of iron bioavailability to humans from maize and great northern beans, but incorrectly predicted color-associated differences between bean varieties and their interaction with AA.