Caco-2 cell iron uptake from meat and casein digests parallels in vivo studies: use of a novel in vitro method for rapid estimation of iron bioavailability.

We developed a model for assessing iron bioavailability from foods which combines simulated peptic and intestinal digestion with measurement of iron uptake by Caco-2 cell monolayers. Our objective was to further validate this model by determining if meat enhances Caco-2 cell iron uptake relative to casein. Caco-2 cell monolayers were covered with Hank's balanced salt solution (HBSS) buffered with HEPES, pH 7.4. An upper chamber was created over the cells by fitting the bottom of a Costar Transwell insert with a 12,000-14,000 molecular weight cut-off dialysis membrane. This membrane allowed low molecular weight iron complexes to diffuse into the media bathing the cells and prevented damage to the cells from the digestive enzymes. Prior to digestion, each sample (homogenate of beef, chicken, fish or casein) was mixed with 59FeCl3 to achieve an iron concentration of 10 mumol/L. Following pepsin digestion (pH2), pH was adjusted to 7.4, pancreatic enzymes and bile extract were added to each digest, and an aliquot was then introduced into the upper chamber of the culture dish. During this intestinal digestion period, 59Fe uptake occurred from iron that dialyzed into the lower chamber. The 59Fe uptake from beef, chicken and fish digests was 300-400% of the 59Fe uptake from a casein digest. Our results parallel human absorption studies indicating that meat enhances iron absorption. The results suggest that digestion products of the meat proteins were at least partially responsible for the enhancement of iron uptake. Overall, this study supports the usefulness of our model as a means of assessing iron bioavailability.

Bathophenanthrolene disulfonic acid and sodium dithionite effectively remove surface-bound iron from Caco-2 cell monolayers.

Iron uptake by Caco-2 cell monolayers is commonly assessed by incubating the cells under radiolabeled iron solutions, removing the radiolabeled solution, rinsing to stop uptake and measuring the radioactivity retained by the cells. It is therefore essential to differentiate between iron that is nonspecifically bound to the cell surface from that which has been taken up by the cell. We report here on a method for removal of surface-bound iron from Caco-2 cell monolayers. We used a 140 mmol/L NaCl, 10 mmol/L PIPES, pH 6.7 solution containing 5.0 mmol/L sodium dithionite (Na2S2O4) and 5.0 mmol/L bathophenanthroline disulfonic acid to reduce, remove and chelate iron bound to the cell surface. We validated our method by demonstrating the removal of 97% of an insoluble iron complex from the apical surface of Caco-2 cell monolayers. Our data indicate that the removal solution does not damage the apical membrane and thereby does not have access to intracellular iron; thus only surface bound iron is removed. The remaining cell-associated iron represents that which has been transported into the cell. We present data on the uptake and nonspecific binding of iron from iron complexes of both ferrous and ferric forms, and show that iron removal treatment resulted in uptake measurements that agree more closely with accepted principles of iron uptake by intestinal epithelium. The iron removal method used in this study should provide investigators with a valuable tool for accurately determining iron uptake by epithelial cells in culture.

Enhanced Fe(3+)-reducing capacity does not seem to play a major role in increasing iron absorption in iron-deficient rats.

Some eucaryotic organisms, including many plants, yeast and mice, have a higher iron uptake during iron deficiency because the capacity to reduce Fe3+ from the environment to Fe2+ is greatly enhanced. To determine whether this occurs in rats, a common experimental model for iron absorption in humans, we compared the in vivo capacity to reduce intraluminal Fe3+ in iron-deficient and normal rats. We also measured potential Fe(3+)-reducing components within the intestinal lumen and on the mucosal surface. Iron-reducing capacity was higher in iron-deficient rats, by a significant (P = 0.026) but modest 20%, in parallel with higher mucosal weight (R2 = 0.501, P = 0.003). In vitro iron reduction by lumen contents was correlated with mucosal weight, even though mucosal tissue was not present in the assays. This capacity was not related to ascorbic acid, glutathione or other nonprotein sulfhydryls. Mucosal ferric reductase activity was higher in iron-deficient rats in parallel with higher tissue weight, but the specific activity did not differ and the higher total activity was not associated with the brush border fraction. The role of endogenous Fe3+ reduction in regulating iron absorption should be investigated in humans and in other experimental models.

Ferric iron absorption in rats: relationship to iron status, endogenous sulfhydryl and other redox components in the intestinal lumen.

Based on the hypothesis that Fe+3 must be reduced before Fe absorption, we investigated luminal factors that might participate in the physiological Fe+3 reduction. Rats were fed diets containing 7 mg Fe/kg diet [adequate iron (+Fe]) for 3 wk prior to a 10-min test of 59Fe absorption from an in vivo ligated duodenal segment. During absorption of 59Fe, the oxidation-reduction potential became more reducing and the pH rose in segment contents. There were small but significant differences between the -Fe and +Fe rats. In one experiment, the lumen environment was modified by bile duct ligation and/or intestinal perfusion prior to the absorption test. Ascorbic acid, nonprotein sulfhydryl compounds, Fe+2 and total ionizable Fe were measured in luminal contents. Nonprotein sulfhydryl concentration was positively correlated with, and the best predictor of, Fe absorption in -Fe rats.

Mucus and iron absorption regulation in rats fed various levels of dietary iron.

We tested two hypotheses: (1) that iron binding by secreted mucus enhances iron absorption (Quarterman, Digestion 37: 1, 1987) and (2) that iron binding by secreted mucus prevents excess iron absorption. Rats were fed diets containing 6, 200 or 500 mg Fe/kg diet (Fe-0, Fe-200 and Fe-500 rats, respectively) for 3 wk. Iron absorption was measured in fasted rats using 59FeCl3 in a 10-min in situ duodenal ligated-segment procedure. After draining the segment contents, the mucus layer was separated from the under-lying mucosal surface using Quarterman's agar cast technique. In comparison with that in Fe-200 rats, iron absorption in Fe-0 rats was markedly increased, but the 59Fe and the total mucus in the mucus layer were decreased. The 59Fe absorption and total mucus and total iron in the mucus layer were similar in Fe-500 rats and Fe-200 rats, but the 59Fe in the mucus layer was marginally lower in Fe-500 rats. There was no evidence that mucus enhanced iron absorption; it appeared to trap or bind iron proportionally to the amount of secreted mucus, suggesting protection against excess absorption. Mucus secretion and possibly synthesis were decreased in the Fe-0 rats.

Apparent absorption and retention of Ca, Cu, Mg, Mn, and Zn from a diet containing bran.

To establish conditions for comparisons of mineral bioavailability from plant sources, seven male subjects consumed a constant diet containing bran fiber and phytate. Absorption and retention of Ca, Cu, Mg, Mn, and Zn were measured for 7-day periods through wk 2-7. Intakes of Mg, Mn, and Zn significantly exceeded the RDA; Ca and Cu intakes were only slightly in excess of RDA. All mineral retentions fluctuated from week to week but only Mg and Mn showed a consistent positive trend over time. Phytate excretions showed characteristic individual patterns, but did not appear to change with time. In contrast to previous observations fecal recovery of polyethyleneglycol (PEG) (MW = 4000) was consistently lower than recovery of simultaneously ingested Cr. Only five of the seven subjects returned close to 100% of Cr within 7 days. It was concluded that at least 4 wk were needed for adaptation in investigations involving more than one mineral when the experimental diet is adequate in the nutrients under investigation, that measurements of responses to treatment required 2-3 wk each, and that successive isotopically labeled test meals may overlap if they are spaced at 7-day intervals.

Comparison of in vitro and in vivo measurements of dietary Ca exchangeability and bioavailability.

An in vitro method for measuring dietary mineral exchangeability (miscibility with an extrinsic isotopic tracer) was tested by comparison with in vivo measurements in rats. Collards, spinach and soybeans, intrinsically labeled with 45Ca, were fed to rats together with extrinsically added 47Ca. Absorption of both tracers was determined by measuring their concentrations in the femur 2 days after consumption of the labeled test meals. The same intrinsically and extrinsically labeled foods were digested in vitro for estimation of Ca exchangeability and solubility after peptic-pancreatic digestion. Corresponding in vivo and in vitro estimates of exchangeability agreed closely for the three foods tested. Solubility after in vitro peptic-pancreatic digestion (potential bioavailability) showed discrepancies. In vitro values were somewhat higher for collards than in vivo fractional absorption. The reverse was true for Ca in soybeans and spinach. Thus, the in vitro procedure appears adequate for measuring intrinsic dietary Ca exchangeability but requires modification if it is to be a reliable substitute for in vivo measurements of Ca bioavailability.

Serum vitamin A, carotene and cholesterol levels in Nigerian women using various types of contraceptives.

The diet in southern Nigeria is unusually high in carotenoids and low in preformed vitamin A. We studied the changes in serum vitamin A carotene and cholesterol induced in women in this region by the use of 3 types of contraceptives: IUD, injectable progestogen (INJ), and oral combination estrogen-progestogen (OC). The mean serum vitamin A and carotene levels were high in all groups. As expected, the serum vitamin A level in OC users was higher than in the other groups, but unexpectedly, the serum vitamin A was lower in IUD users. Serum carotene was lower in OC users than in other groups. Serum cholesterol was lower in the IUD and INJ groups than in the control and OC groups. A more detailed study of plasma transport forms of vitamin A is needed to determine if the very high serum vitamin A levels seen in some OC users in this population are potentially harmful.

An in vitro system for measuring intrinsic dietary mineral exchangeability: alternative to intrinsic isotopic labeling.

The primary purpose for the in vitro system described here was to provide an alternative method to intrinsic isotopic labeling for determining exchangeability of intrinsic food mineral with extrinsic inorganic mineral or one of its isotopes. In this system, foods or food mixtures extrinsically tagged with an isotope of the mineral of interest are incubated successively in media simulating peptic or pancreatic digestion. Radioactivity and mineral measurements are made on a)the mixture before digestion and after b)peptic and c)successive peptic pancreatic (P-Pa) digestion. Exchangeability is determined by comparing the specific activity in the peptic and P-Pa digest supernatants to that in the mixture before digestion. In addition, determination of the total soluble mineral after P-Pa digestion provides a measurement of potential bioavailability. The procedure was tested for calcium (Ca) exchangeability and bioavailability from a number of foods and food combinations with 45Ca as the tracer. It could be used for any mineral, however, including non-nutrient or toxic elements. It is likely to be most useful in association with human studies carried out with extrinsic stable (non-radioactive) or short-lived radioactive isotopes that are unsuitable for intrinsic labeling.

Use of nonabsorbable markers for gastrointestinal contents in in vivo measurement of magnesium bioavailability.

Magnesium absorption from five leafy vegetables was measured in rats using 28Mg as an extrinsic label and polyethylene glycol (PEG) and chromium-mordanted vegetable fibers (cr-mordants) as unabsorbable markers of soluble and particulate gut and fecal contents, respectively. The test meals used in this investigation were identical to those used in a previous test in which we found: a) that stable 26Mg biologically incorporated into the vegetables was freely exchangeable with an extrinsic 28Mg tracer; b) PEG consistently preceded Cr-mordants in transit through the gut; and c) fecal Mg isotope excretion 12 hours after the test meal was inadequate for measurement of Mg absorption. Accordingly, in the present investigation, all measurements in feces and gut contents were made 24 hours after the test meal. By that time an average of greater than 60% of the ingested PEG had been excreted. Magnesium-28 excretions, estimated under these conditions, were 4-5% less than they would have been had fecal PEG excretion been 100%. The discrepancy was due to partial separation of PEG and Mg in the lower gastrointestinal tract (GIT), possibly due to sequestration of Mg by microorganisms. That Mg must have entered an insoluble phase in the lower GIT was born out by the observation that little or no Mg absorption occurred in the cecum or colon. As before, Cr-modants lagged behind PEG and unabsorbed 28Mg, and thus do not seem suitable to monitoring intestinal transit of Mg.

Magnesium absorption from leafy vegetables intrinsically labeled with the stable isotope 26Mg.

Five leafy vegetables were grown in nutrient solutions in which the natural magnesium was replaced by the stable isotope, 26Mg. They were fed to rats in a test meal together with the extrinsic tracer 28Mg: a) to determine to what extent the instrinsic tracer (26Mg) was exchangeable with extrinsic 28Mg during the digestion and absorption processes, and b) measure the relative Mg availability from the different vegetables. The two tracers, 26Mg and 28Mg, were close to 100% exchangeable, as judged by the ratio of 26Mg/28Mg in the livers. Mean relative Mg absorption from the various vegetables ranged from 108 to 118% of the Mg absorbed from a standard test meal containing MgSO4. There were no statistically significant differences between the rates of Mg absorption from the five vegetables although two of the vegetables tested contained oxalate. The usefulness of stable 26Mg as a tracer in Mg bioavailability tests is discussed.

Vitamin B6 status of Nigerian women using various methods of contraception.

The vitamin B6 status of Nigerian women, and the effects of three contraceptive methods--intrauterine contraceptive device, injectable progestogen, and combination estrogen-progestogen oral contraceptive pills--on the vitamin B6 status of these women were assessed by measuring the erythrocyte alanine aminotransferase (Ala-AT) activity both with and without in vitro stimulation by pyridoxal-5'-phosphate. The unstimulated Ala-AT activity and the Ala-AT index (ratio of pyridoxal-5'-phosphate-stimulated:unstimulated activity) were used as indications of vitamin B6 status. The criterion for vitamin B6 deficiency was an Ala-AT index greater than 1.25. Of the 238 women in the total sample, 11% were judged to be deficient. There were no significant differences in the mean Ala-AT activity, mean Ala-AT index or rate of deficiency foun in the control and three contraceptive groups. The absence of an effect of oral contraceptives on vitamin B6 status is in contrast with other cross-sectional studies, but in agreement with controlled longitudinal studies. The packed cell volume and reticulocyte count were also measured and were significantly affected by contraceptive steroids, but these effects are not thought to be of clinical importance.