The use of solubility, dialyzability, and Caco-2 cell methods to predict iron bioavailability

An update on in vitro methods employed to assess iron bioavailability is presented. Solubility and dialyzability are not useful predictors of iron absorption, whereas Caco-2 cell models provide useful comparisons on the availability of iron from different sources. Strengths and weaknesses of in vitro approaches are briefly described, including the need to characterize Caco-2 cells in order to interpret results correctly. Further developments are required to refine Caco-2 model systems, including optimization and standardization of methods.     

Caco-2 cells can be used to assess human iron bioavailability from a semipurified meal

A Caco-2 cell model with extrinsic radioiron was used to evaluate the effect of dietary factors on nonheme iron bioavailability from a semipurified meal. Study 1 was conducted to evaluate the effect of enhancers (ascorbic acid) and inhibitors (bran, phytate and tea) on iron bioavailability when added to semipurified meal containing egg albumen as a protein source. The effect of various proteins [bovine serum albumin (BSA), casein, beef and soy] on iron bioavailability was evaluated in Study 2 by substituting the above protein sources for egg albumen. Protein solubilization following in vitro digestion for individual test meals was not significantly different from the control. On the other hand, nonheme iron solubilization (0.8+/-0.0 to 5.9+/-0.3 vs. 4.9+/-0.8 mg/L) varied significantly. The total iron uptake for each meal was calculated based on the percentage of radioiron taken up and transported by Caco-2 cells and the amount of nonheme iron present in uptake solutions. Iron uptake ratios represent test/control values. With the exception of BSA and ascorbic acid, the effect of dietary factors was similar to that found in humans reported in the literature. A significant correlation (r = 0.97; P<0.0001) was found between the published human absorption data and the iron uptake by the Caco-2 cells. The results of our study indicate the usefulness of Caco-2 cells in assessing human iron absorption and the feasibility of this cell model in studying iron bioavailability from various food combinations, otherwise not easily performed in humans.     

Histidine content of low-molecular-weight beef proteins influences nonheme iron bioavailability in Caco-2 cells

The objective of this study was to isolate and characterize beef muscle proteins that enhance nonheme iron bioavailability. Beef sirloin was cooked, lyophilized and reconstituted with water before in vitro digestion. After centrifugation, the digest supernatant was sequentially ultrafiltered using 10- and 1-kDa molecular weight cut-off membranes. Nonheme iron bioavailability was assessed by Caco-2 cell monolayer (59)Fe uptake using an extrinsic labeling method. All ultrafiltration fractions significantly (P < 0.001) increased iron solubility at pH 6.0, compared with the blank. However, iron uptake was significantly (P < 0.001) greater than the blank only in the presence of the 1-kDa retentate (1KR). Therefore, the 1KR was chosen for further analysis. Immobilized metal affinity chromatography (IMAC) of the 1KR yielded four fractions, i.e., three distinct fractions (F1, F3, F4) and one fraction (F2) comprised of a few closely associated peaks. All four IMAC fractions resulted in significantly (P < 0.001) greater (two- to fivefold) iron solubility at pH 6.0, compared with the blank. Iron uptake with F2 and F4 was significantly greater than the blank (P < 0.001 and P < 0.05, respectively). Gel electrophoresis and matrix-assisted laser desorption/ionization analysis illustrated that F1-F4 contained many peptides ranging from 1- to 7-kDa. Amino acid composition analysis revealed that histidine concentration increased progressively from F1 to F4, corresponding to a general, but not parallel increase in iron solubility and uptake. Our results suggest that the enhancement of nonheme iron absorption by beef may be due to peptides produced during gastrointestinal digestion and that histidine content may be important.     

Profile and bioavailability analysis of myo-inositol phosphates in rye bread supplemented with phytases: a study using an in vitro method and Caco-2 monolayers

Commercial preparations of 6-phytase A alone and in combination with phytase B were used in rye breadmaking. Determination of bioavailability of myo-inositol phosphates from bread was performed by an in vitro digestion method followed by the measurement of an uptake by Caco-2 cells in culture. In bread supplemented with a combination of 6-phytase A and phytase B, a significant reduction in phytate content was observed from 3.62?μmol/g in the control to 0.7?μmol/g. Bioavailability of phytate estimated by an in vitro method simulating digestion in the human alimentary tract was 9% in the bread supplemented with phytase B, 7% (6-phytase A) and 50% in the control bread. In cell culture, the bioaccessibilities of inositol triphosphates from bread baked with the addition of 6-phytase A was higher by 36% as compared to the samples baked with phytase B and by 32% in breads baked with combination of both phytases.     

Organic acid bioavailability from banana and sweet potato using an in vitro digestion and Caco-2 cell model

Introduction  

Organic acids from plant food have been shown to play an important role in the prevention of chronic diseases (osteoporosis, obesity), inherent to western diets, but little is known about their bioavailability in the small intestine, information that needs to be determined in order to quantify likely effects on human health.     

The concept of bioavailability as it relates to iron nutrition

Bioavailability is discussed in conceptual terms; the definition proposed for all nutrients is that it is a measure of the proportion of the total in a food or diet that is digested, absorbed and metabolised by normal pathways. The availability of Fe is affected by a number of dietary and physiological variables, and various techniques used to quantify these effects and thereby predict availability are reviewed. Emphasis is placed on the modifying influence of physiological factors on Fe absorption and hence measurements of availability. Methods of estimating Fe status as a means of incorporating the utilisation element of Fe availability are described.     

Effect of retinol on iron bioavailability from Iranian bread in a Caco-2 cell culture model

OBJECTIVE: Iron and vitamin A deficiencies are among the most prevalent nutritional problems. There are no data on iron bioavailability in Iran. In addition, interaction of vitamin A with iron bioavailability is not well documented, so we determined iron bioavailability from the most widely consumed food in Iran, lavash bread, and the effect of vitamin A on this bioavailability. METHODS: In vivo human studies for determining iron bioavailability are cumbersome, time consuming, and costly to perform, so we used an in vitro model in Caco-2 cells. Bread samples were digested with or without vitamin A (10 microg/1.0 g of dried bread). We used an iron solution containing vitamin C as a positive control. Iron absorption was measured using 59FeCl3. Bioavailability was defined as the percentage of radiolabeled iron taken up and transferred by Caco-2 cells after 1.5 h of incubation. Experiments were carried out two to four times. RESULTS: Mean +/- standard deviations for iron bioavailability in bread samples digested without or with additional vitamin A and in controls were 2.53 +/- 1.55%, 6.62 +/- 3.40%, and 20.80 +/- 2.30%, respectively (P < 0.001). Vitamin A caused a 2.62-fold increase in iron absorption from bread samples. CONCLUSIONS: Iranian lavash bread has low iron bioavailability, but this can be increased by vitamin A supplementation. Increasing vitamin A intake can be considered as a method for increasing iron bioavailability, thus combating iron and vitamin A deficiencies simultaneously.     

A pilot study to investigate effects of inulin on Caco-2 cells through in vitro metabolic fingerprinting

Metabolic fingerprints are novel measurement tools to evaluate the biochemical status of a living organism by using 1H NMR and multivariate data analysis (MVDA). In this way, a quick evaluation of changes in health or diseased state can be made, reflected in alterations of metabolic patterns. Normally, metabolic fingerprinting is based on in vivo studies. These studies often represent a labor-intensive and expensive manner of investigation. In vitro studies are not hampered by these disadvantages, thus constituting an interesting alternative. In this research, results are presented of a pilot experiment in which metabolic fingerprinting was combined with an in vitro model. For this purpose, differentiated Caco-2 cells were exposed to inulin and its fermentative metabolites, both dissolved in culture medium. Cells were incubated for 0 or 48 h. Cell fractions were analyzed by NMR, then subsequently with MVDA. Differences in treatment provided detectable variations in the time of metabolic patterns of cell contents. Results indicated that glucose metabolism linked to glutamate was of major importance in the effects of inulin and its metabolites on Caco-2 cells under the conditions of our study. Metabolic fingerprinting in combination with an in vitro model appears to be a feasible technique with which to visualize metabolic patterns of cell contents and provides an efficient place for the generation of hypotheses about the metabolic pathways involved. In vitro metabolic fingerprinting may be of great benefit in the future for a better understanding of the relationship between nutrition and health.     

Beta-carotene and inhibitors of iron absorption modify iron uptake by Caco-2 cells

A National fortification program instituted in Venezuela in 1993 reduced iron deficiency and anemia by half in only 1 y. The fortification mixture contained ferrous fumarate, vitamin A and other vitamins. We conducted experiments to characterize ferrous fumarate uptake by Caco-2 cells. Increasing amounts of ferrous fumarate, vitamin A, phytate, tannic acid and beta-carotene were added to incubation mixtures using a range of concentrations that included the molar ratios used in the Venezuelan fortification program. Cells were incubated for 1 h at 37 degrees C with 37 kBq (59)Fe and the compound to be evaluated. They were then rinsed, trypsinized and counted to measure uptake. Effects of ascorbic acid, days in culture and use of flasks or inserts were also evaluated. Optimal conditions for uptake experiments were pH 5.5, in the presence of ascorbic acid and at 16 d in culture. Use of flasks or inserts did not affect uptake. Vitamin A did not significantly increase iron uptake under the experimental conditions employed. However, beta-carotene (6 micromol/L) significantly increased iron uptake compared to no beta-carotene addition (114.9 +/- 6.3 and 47.2 +/- 5.9 pmol/mg cell protein, respectively). Moreover, in the presence of phytates or tannic acid, beta-carotene generally overcame the inhibitory effects of both compounds depending on their concentrations. We conclude that beta-carotene improves iron uptake and overcomes the inhibition by potent inhibitors of iron absorption. These experiments also validated the usefulness of Caco-2 cell model system to evaluate iron metabolism.     

An in vitro method for predicting the bioavailability of iron from foods.

An in vitro method for the determination of availability of nonheme iron from foods and diets was investigated. Food was extracted with pepsin-HCl at pH 1.35 and subsequently the pH was adjusted to pH 7.5 and filtered. Ionizable iron was determined in the pH 7.5 filterate by the alpha, alpha-dipyridyl method. The percent iron absorption from the same diets observed in the adult males. Ionizable iron at pH 7.5 was shown to increase in presence of ascorbic acid and meat extract while it decreased in presence of phytate and tannins, similar to the effects of these factors on iron absorption in human subjects. Based on these observations it is proposed that ionizable iron at pH 7.5 determined as described in this study can be used as a reliable measure of bioavailability of nonheme iron in foods.     

An in vitro digestion/Caco-2 cell culture system accurately predicts the effects of ascorbic acid and polyphenolic compounds on iron bioavailability in humans

We developed a rapid in vitro digestion/Caco-2 cell culture model for assessing relative bioavailabilities of iron in foods and meals. The objective of the present study was to determine how closely our Caco-2 model reflects the human response. Meals described in published reports of studies on effects of varying levels of ascorbic acid (AA) and tannic acid (TA) on iron absorption by human subjects were carefully replicated. Iron absorption ratios (iron absorption from meals containing AA or TA divided by iron absorption from identical meals without these enhancers or inhibitors) were determined using the Caco-2 model. Ferritin formation by the Caco-2 cells was used as an indicator of iron absorption. Response patterns of effects of AA and TA on absorption ratios (AR) calculated from Caco-2 and human data were very similar: AA increased ARs in a dose-response manner and TAs decreased AR. The natural logs of the ARs determined in Caco-2 and human studies were correlated: R = 0.935 (P = 0.012) and 0.927 (P = 0.007) for AA and TA, respectively. When results from meals with AA and TA were pooled, a linear relation between the natural logs of ARs from Caco-2 and human studies was observed (R = 0.968, P < 0.001). We conclude that our Caco-2 model accurately predicts the human response to AA and TA in the meals we tested. If future work reproduces the precision and accuracy shown in this paper for predicting iron bioavailability to humans, then the implications for saving time and resources in iron bioavailability measurements are considerable.     

Effect of dietary protein on heme iron uptake by Caco-2 cells

Objective  

To study heme iron bioavailability and the role of dietary protein (animal and vegetable) on iron uptake using an in vitro model (Caco-2 cell line).     

Carbohydrate fractions from cooked fish promote iron uptake by Caco-2 cells

The objective of this study was to isolate and characterize the meat factor(s) that enhances nonheme iron bioavailability using various analytical and in vitro cell culture techniques. Nonheme iron bioavailability was measured via radiolabeled iron uptake or ferritin formation in Caco-2 cells. Fish haddock fillet was cooked and lyophilized to be used as the muscle tissue of choice because of its low intrinsic iron content. It was demonstrated that the low pH of the stomach (pH 2.0) was the primary factor responsible for initiating the enhancing effect of fish on nonheme iron uptake. Subsequently, cooked fish samples were titrated with HCl to pH 2.0 and incubated for 1 h without digestive enzymes to release the factor(s) from the fish. The supernatant of this acidic digest was then used as a starting material for the meat factor isolation procedures. Fractions generated through Sephadex G-25 size exclusion increased Caco-2 cell iron uptake approximately 9-fold. Subsequent chromatography of these fractions via C18 reverse-phase HPLC were conducted, and enhancing activity was observed only in the "injection peak." This observation coupled with protein measurement and amino acid composition analysis revealed that the active fractions contained negligible amounts of proteins or amino acids. Active fractions were highly enriched with carbohydrates. Subsequent chromatography via high performance anion exchange chromatography with pulsed amperometric detection yielded 3 active peaks that increased Caco-2 cell iron uptake 3.4- to 4.9-fold. Our results indicate that specific carbohydrates contribute to the enhancing effect of meat on iron uptake by the enterocyte. These carbohydrates may be oligosaccharides originating from glycosaminoglycans in the extracellular matrix of muscle tissue.     

Methods and options for estimating iron and zinc bioavailability using Caco-2 cell models: benefits and limitations

Micronutrient deficiencies are prevalent worldwide and have detrimental effects on human health. Complex interactions between micronutrients and other dietary components largely determine micronutrient bioavailability, and understanding these interactions is key to improving micronutrient status. A number of in vitro and in vivo methodologies are available for assessing micronutrient bioavailability. The purpose of this review is to highlight the usefulness of one of the in vitro models, the Caco-2 cell, as a predictive tool for human micronutrient bioavailability. The review focuses on current methods used with the Caco-2 cell line, their benefits and limitations, and the possibilities for the future development of this model.     

A low-molecular-weight factor in human milk whey promotes iron uptake by Caco-2 cells

The iron bioavailability of human milk (HM) is substantially greater than that of cow's milk (CM), but the factor responsible for this high bioavailability is unknown. This study evaluated the effects of various HM and CM fractions on iron bioavailability. Milk was separated into fat, casein and whey fractions by ultracentrifugation. Whey was further fractionated by ultrafiltration with a 10-kDa membrane to produce a 10-kDa retentate (10kR) and a 10-kDa filtrate (10kF). Samples were prepared by mixing various combinations of the fractions, bringing the samples to prefractionation weight with minimum essential medium (MEM), and adding iron (10 micro mol/L) as ferrous sulfate. Samples were divided into two aliquots: one was subjected to in vitro digestion, the other was not. Bioavailability was assessed by applying the samples to Caco-2 cell monolayers and incubating for 24 h. Ferritin formation in the cells was used as an index of iron uptake. Removing the fat from undigested HM samples doubled the ferritin formation, but removing the whey or casein had no effect. Results with digested HM samples were similar, except that removing the whey decreased ferritin formation by 48%. Removing the fat from digested CM samples had no effect, but removing the casein doubled the ferritin formation. Removing the 10kF from HM reduced ferritin formation by 60%, but removing the 10kR had no effect. These data suggest that a low-molecular-weight factor (<10 kDa) in human milk enhances iron absorption.     

The relative bioavailability in humans of elemental iron powders for use in food fortification

Summary Background Bioavailability data in humans of elemental iron powders is limited although elemental iron is a common form of iron when used as a fortificant. Aim of the study The relative bioavailability (RBV) of seven elemental iron powders, five commercially available and two developmental are evaluated. In addition, one commercial electrolytic iron powder given with ascorbic acid (AA) was examined. Methods Based on a validated method this double–blinded randomized crossover study included three groups of male blood donors (n = 3*16) who were served rolls fortified with different elemental iron powders or ferrous sulfate (FeSO₄) nine weeks apart. Blood samples were drawn every hour for six hours. RBV was obtained by comparing the increase in serum iron concentration induced by the elemental iron with the increase induced by FeSO₄. Results All elemental iron powders studied were significantly less well absorbed compared to FeSO₄. The electrolytic iron given with 50–mg AA was as well absorbed as FeSO₄ (molar ratio = 1:6, AA:Fe). The mean RBVs of the iron powders were: electrolytic (A–131, RBV = 0.65); electrolytic (Electrolytic, RBV = 0.59); carbonyl (Ferronyl, RBV = 0.58); H–reduced (AC– 325, RBV = 0.56); H–reduced (Hi–Sol, RBV = 0.50); carbonyl (CF, RBV = 0.37); reduced (Atomet 95SP, RBV = 0.36). The reduced iron was distinguished by having significantly lower RBV (0.36) although no significant overall ranking was possible. Conclusion Based on a validated method this doubleblinded cross–over study in humans showed that the evaluated elemental iron powders currently available for commercial use are significantly less well absorbed compared to FeSO₄. The results indicate that the reduced iron powder was absorbed to a lower extent compared to the other iron powders and only 36% compared to FeSO₄. Ascorbic acid seems to improve the bioavailability of elemental iron even though a rather low molar ratio is used. Thus, if confirmed, this enhancing effect of ascorbic acid on elemental iron when used as a fortificant could be used by co–fortifying them.     

The antioxidant BHT normalizes some oxidative effects of iron + ascorbate on lipid metabolism in Caco-2 cells

We showed recently that iron + ascorbate can impair the assembly of intestinal lipoproteins. However, we could not determine whether these changes were caused by iron + ascorbate-mediated lipid peroxidation per se. We therefore conducted studies to evaluate how antioxidants antagonize the iron + ascorbate-induced derangements. To this end, Caco-2 cells, a reliable experimental intestinal model, were incubated with iron + ascorbate (0.2 mmol/L each) alone or with different concentrations of catalase, mannitol, tocopherol or BHT. Exposing Caco-2 cells to iron + ascorbate increased malondialdehyde levels fourfold (P < 0.0001); this effect was decreased markedly (P < 0.02) in the presence of BHT. Furthermore, BHT normalized the abnormal intracellular events involved in fat absorption, i.e., lipid esterification, cholesterol synthesis and apolipoprotein production. On the other hand, it did not fully restore the secretion of lipids and lipoproteins. Thus, our current data imply that iron + ascorbate-catalyzed lipid peroxidation is partially responsible for the disturbances observed in intestinal lipid transport.     

The precision of in vitro methods and algorithms for predicting the bioavailability of dietary iron

Three factors determine how much iron will be absorbed from a meal. They are the physiological mechanisms that regulate uptake by and transfer through the enterocytes in the upper small intestine, the quantity of iron in the meal, and its availability to the cellular iron transporters. Established methods exist for predicting the effect of physiological regulation and for measuring or estimating meal iron content. Three approaches to estimating bioavailability have been advocated. Two are in vitro screening procedures: measurement of dialyzable iron and Caco-2 cell uptake, both carried out after in vitro simulated gastric and pancreatic digestion. The third is the use of algorithms based on the predicted effects of specific meal components on absorption derived from isotopic studies in human volunteers. The in vitro procedures have been very useful for identifying and characterizing factors that affect non-heme iron absorption, but direct comparisons between absorption predicted from the in vitro tests and measurements in human volunteers have only been made in a limited number of published studies. The available data indicate that dialysis and Caco-2 cell uptake are useful for ranking meals and single food items in terms of predicted iron bioavailability, but may not reflect the magnitudes of the effects of factors that influence absorption accurately. Algorithms based on estimates of the amounts of heme iron and of enhancers and inhibitors of non-heme iron absorption in foods make it possible to classify meals or diets as being of high, medium, or low bioavailability. The precision with which meal iron bioavailability can be predicted in a population, for which a specific algorithm has been developed, is improved by measuring the content of the most important enhancers and inhibitors. However, the accuracy of such predictions appears to be much lower when the algorithm is applied to meals eaten by different populations.     

Soyasaponin I and sapongenol B have limited absorption by Caco-2 intestinal cells and limited bioavailability in women

A human study was conducted to evaluate soyasaponin bioavailability in humans. Eight healthy women ingested a single dose of concentrated soy extract containing 434 micromol of group B soyasaponins, the predominant form of soyasaponins in soybeans. Neither soyasaponins nor their metabolites were detected in a 24-h urine collection. Soyasapogenol B, a major metabolite of group B soyasaponins, was found (36.3 +/- 10.2 micromol) in a 5-d fecal collection but no group B soyasaponins were detected. A human colon cancer Caco-2 cell model was used to evaluate the absorbability of soyasaponins at the mucosal level. The mucosal transfers of soyasaponin I and soyasapogenol B were 0.5-2.9 and 0.2-0.8%, respectively, after 4-h incubation on the Caco-2 monolayer. The apical to basolateral absorptions of soyasaponin I and soyasapogenol B were low with P(app) of 0.9 to 3.6 x 10(-6) and 0.3 to 0.6 x 10(-6) cm/s, respectively. The transport rate and cell uptake of soyasaponin I were saturable and concentration-independent. In contrast, soyasapogenol B was taken up by Caco-2 cells in a concentration-dependent manner. Soyasaponin I had no apparent cytotoxic effect on Caco-2 cells at concentrations up to 3 mmol/L, whereas soyasapogenol B at 1 mmol/L or more significantly reduced cell viability. Therefore, ingested soyasaponins have low absorbability in human intestinal cells and seem to be metabolized to soyasapogenol B by human intestinal microorganisms in vivo and excreted in the feces.