Extrinsic Labeling of Staple Food Crops with Isotopic Iron Does Not

Oct 11, 2015 - Robert Holley Center for Agriculture and Health, Agricultural Research Service, U.S. Department of Agriculture, 538 Tower Road, Ithaca,...
0 downloads 4 Views 2MB Size
Download PDF
Article pubs.acs.org/JAFC

Extrinsic Labeling of Staple Food Crops with Isotopic Iron Does Not Consistently Result in Full Equilibration: Revisiting the Methodology Raymond P. Glahn,*,† Zhiqiang Cheng,‡ and Shree Giri‡ †

Robert Holley Center for Agriculture and Health, Agricultural Research Service, U.S. Department of Agriculture, 538 Tower Road, Ithaca, New York 14853, United States ‡ Department of Food Science, Cornell University, Stocking Hall, 411 Tower Road, Ithaca, New York 14853, United States ABSTRACT: Extrinsic isotopic labeling of food Fe has been used for over 50 years to measure Fe absorption. This method assumes that complete equilibration occurs between the extrinsic and the intrinsic Fe prior to intestinal absorption. The present study tested this assumption via in vitro digestion of varieties of maize, white beans, black beans, red beans, and lentils. Prior to digestion, foods were extrinsically labeled with 58Fe at concentrations of 1, 10, 50, and 100% of the intrinsic 56Fe. Following an established in vitro digestion protocol, the digest was centrifuged and the Fe solubilities of the extrinsic 58Fe and the intrinsic 56Fe were compared as a measure of extrinsic/intrinsic equilibration. In the beans, significantly more of the extrinsic Fe (up to 2−3 times, p < 0.001) partitioned into the supernatant. The effect varied depending upon the seed coat color, the harvest, and the concentration of the extrinsic Fe. For lentils and maize the extrinsic Fe tended to partition into the insoluble fraction and also varied depending on variety and harvest. There was no crop that consistently demonstrated full equilibration of the extrinsic Fe with the intrinsic Fe. These observations challenge the accuracy of Fe absorption studies in which isotopic extrinsic Fe was used to evaluate Fe absorption and bioavailability. KEYWORDS: extrinsic iron labeling, iron bioavailability, isotopic exchange

■

INTRODUCTION The elemental properties of iron make it insoluble in free form at pH >3 unless it associates with compounds such as citrate, fumarate, phytic acid, or polyphenols.1 Most Fe absorption from the diet occurs in the upper small intestine, that is, the duodenum, where luminal pH is approximately 5 at fasting and between 6 and 7 post food consumption.2−4 Absorption of Fe from soluble forms therefore depends on the binding constant and chemical properties of the compounds that interact with Fe in the environment of the intestinal lumen. These properties make iron a challenging mineral for humans to absorb, particularly from diets in which staple food crops such as rice, wheat, beans, and maize form the bulk of the diet and provide most, if not all, of the dietary Fe.5 In such diets, phytic acid and polyphenolic compounds are in high molar excess relative to Fe, and the Fe is predominantly bound by these compounds, thus rendering it poorly available for exchange with the duodenal Fe transport mechanisms. Populations that exist on these diets are therefore prone to Fe deficiency anemia, and it is not uncommon for anemia rates to reach 60−70% in these areas.6 Despite decades of study and interventions, iron deficiency remains the leading nutritional deficiency worldwide, affecting approximately a third of the world’s population, most of whom are resource-poor women and children.7 Strategies to enhance Fe intake include fortification of food with extrinsic Fe and more recently, in the past decade, via plant biofortification efforts utilizing traditional plant breeding to increase Fe content in crops such as beans and pearl millet.8 Regardless of the strategy to alleviate Fe deficiency, research that requires measurement of Fe uptake from foods must be reasonably accurate. This is essential whether one is developing © 2015 American Chemical Society

a fortified food product or evaluating a biofortified variety of common beans that will be released in a targeted region of Fe deficiency. Traditionally, the techniques to measure Fe absorption in such studies involve the use of an isotopic label of the food Fe, followed by measurement of labeled Fe uptake from a meal coupled with knowledge of the Fe content of the meal. The bioavailability (i.e., absorbability) of the Fe is often expressed as a percentage, dividing the amount of Fe absorbed by the amount consumed.9 Isotopic labeling of food Fe can be done by intrinsic or extrinsic techniques. For intrinsic labeling of staple food crops, the plant must be grown hydroponically, a process that is inherently expensive and requires greenhouse facilities.10 It is also limited in the amount of material that can be produced, with 1−2 kg of material considered by many to be demanding to produce. Extrinsic labeling is an attractive alternative as it is involves the simple addition of a known amount of isotope to the test meal prior to consumption. This extrinsic Fe is then assumed to “label” the intrinsic Fe by exchanging and equilibrating fully with the known amount of intrinsic Fe of the test meal. The test meal is then consumed and usually 2 weeks later, a venous blood sample is taken from the subject and the concentration of the extrinsically added isotope is measured via mass spectroscopy. A series of calculations is then utilized to quantify the amount of Fe absorbed from the test meal.11 Received: Revised: Accepted: Published: 9621

August 11, 2015 October 10, 2015 October 11, 2015 October 11, 2015 DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry

Figure 1. Summary of the extrinsic 58Fe to intrinsic 56Fe ratios in the supernatant of in vitro digests of various staple food crops. Extrinsic 58Fe was added at 1, 10, 50, and 100% of the total intrinsic Fe. Dashed lines represent the expected ratio of extrinsic to intrinsic Fe if complete exchange occurs. Asterisk indicates a significant difference (p < 0.001) >10% from the expected ratio. Values are expressed as the mean ± SEM, n = 6 independent replications. samples, one small red bean sample, and a red “kidney” bean sample were selected from the shelves at a local supermarket and were the Goya brand label. Normal- and high-Fe beans from the same varieties and harvests used for a human efficacy trial in Rwanda were also selected for study.14 These lentil samples were supplied from the Crop Development Center at the University of Saskatchewan and were red lentil varieties known as Maxim and Robin. These samples were both dehulled prior to the study. The maize samples were from varieties developed for high and low Fe bioavailability.15 To cook the samples, the samples were autoclaved for 15 min. The autoclaved samples were freeze-dried afterward and then ground into powder. The samples were wet-digested in a mixture of HNO3 and HClO4 (6:4 v/v), and the Fe contents were determined by inductively coupled argon plasma emission spectrometry (ICAP model 61E trace analyzer, Thermo Jarrell Ash Corp., Franklin, MA, USA). In Vitro Digestion. The in vitro digestion protocol was conducted as per an established in vitro digestion model.12 Exactly 1 g of each sample was used for each sample digestion. To initiate the gastric phase of digestion, 10 mL of fresh saline solution (0.9% sodium chloride) was added to each sample and mixed. As it was deemed possible that the degree of exchange and equilibration could be related to concentration of extrinsic Fe, the samples were extrinsically labeled with 58Fe in 0, 1, 10, 50, and 100%, respectively, on the basis of the Fe concentrations in the sample. Each labeling group contained six subsamples. The samples were then mixed and remained at room temperature for 30 min. The pH was then adjusted to 2.0 with 1.0 mol/L HCl, and 0.5 mL of the pepsin solution (containing 1 g pepsin per 50 mL; certified >250 U per mg protein; Sigma P7000) was added to each mixture. In some experiments the gastric pH was adjusted to 4.0, as it is known that many meals never reach low pH in the stomach.3,4 The mixtures were under gastric digestion for 1 h at 37 °C on a rocking platform (model RP-50, Laboratory Instrument, Rockville, MD, USA) located in an incubator. After 1 h of gastric digestion, the pH of the sample mixture was raised to 5.5−6.0 with 1.0 mol/L NaHCO3 solution, and then 2.5 mL of the pancreatin/bile extract solution was added to each mixture. The pancreatin/bile extract solution contained 0.35 g of pancreatin (Sigma P1750) and 2.1

The assumption that the extrinsically added Fe equilibrates fully with the intrinsic Fe of the test meal is the focus of this paper. Clearly, this assumption must be valid for extrinsic labeling to be an accurate and useful method. To test this assumption, we conducted experiments using various staple food crops, such as maize, common beans, and lentils, and we included different varieties within these crops. The extrinsically added Fe isotope (58Fe) was added to the cooked samples of each test crop, and the sample was then subjected to an established in vitro digestion protocol.12 Four different levels of extrinsically added Fe isotope were utilized, representing 1, 10, 50, and 100% of the intrinsic Fe. Measurement of the actual versus expected extrinsic (58Fe) to intrinsic Fe (56Fe) ratio in the supernatant of the in vitro digest was used as the proxy of Fe equilibration. This study represents one of only two studies known to directly test the primary assumption of extrinsic Fe labeling and should prompt renewed discussion of the accuracy and application of this method to measure Fe absorption from staple food crops.

■

MATERIALS AND METHODS

Chemicals, Enzymes, and Hormones. All chemicals, enzymes, and hormones were purchased from Sigma Chemical Co. (St. Louis, MO, USA) unless stated otherwise. Stable isotope 58Fe in elemental form was purchased from Isoflex USA (San Francisco, CA, USA) for extrinsic labeling experiments. Upon using, the stable isotopes were solubilized using a mixture of sulfuric acid and hydrochloric acid to make a stock solution, and appropriate dilutions were made prior to each experiment. Extrinsically Labeled Crop Sample Preparation. Bean, corn, and lentil samples were selected for 58Fe extrinsic labeling experiment. For the bean samples, high- and low-Fe black bean varieties were selected as these same harvests were used in a human feeding trial conducted in Mexico in 2011.13 Two Great Northern white bean 9622

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry

Table 1. Iron Concentrations, Average Percent Differences in Soluble 58Fe/56Fe versus Expected Values, and Characteristics of Staple Food Crops Used in This Study % difference of 58Fe:56Fe ratio versus expected valuesa [Fe] μg g−1

1%

10%

50%

100%

descriptionb

red lentil 1 red lentil 2

54.1 60.5

21.1 −9.6

−12.4 −21.4

−8.9 −24.8

−10.8 −26.0

dehulled red lentil, ‘Robin’ variety; (source) Crop Development Center, University of Saskatchewan, Saskatoon, Canada

white bean 1 white bean 2 white bean 3 pH 2 pH 4

64.0 62.0 42.3

38.5 126.6

30.3 71.6

−1.3 50.7

11.9 36.7

22.0 28.3

24.8 15.3

21.2 18.9

21.4 20.0

black bean 1 (variety MIB465) black bean 2 (variety DOR500)

88.0

339.5

258.7

109.0

97.4

59.0

225.7

154.2

40.2

42.2

black bean 3 pH 2 pH 4 red kidney bean small red bean

76.5

57.5 58.3

74.3 128.2 −5.9 66.1

17.4 27.7 24.4 18.4

8.8 22.6 6.1 7.4

1.3 24.4 4.6 3.7

maize 1 maize 2

18.8 19.5

−44.4 −6.4

−27.3 −22.0

−21.0 −31.9

−15.3 −31.2

Rwanda bean gastric pH 2 gastric pH 4

51.0 −8.2 21.1

−17.4 0.8

−4.9 −9.5

−4.1 12.4

Rwanda bean gastric pH 2 gastric pH 4

87.2 19.3 13.8

−0.9 8.9

−5.9 0.7

−6.7 14.6

sample

Goya Foods (Great Northern beans); Wegemans Supermarket, Ithaca, NY, USA

from CIAT Germplasm Collection (CIAT.cgiar.org/crops/bean); from same harvest as used in refs13 and 35

Goya Foods; Wegemans Supermarket, Ithaca, NY, USA

research variety (maize 1 = low-bioavailable Fe; maize 2 = high-bioavailable Fe) developed from IBM population15 CIAT G4825, aka “low Fe”, a normal Fe landrace, brown carioca color

CIAT SMC, aka “high Fe”, a line selected for high Fe content, brown carioca color

Percent soluble intrinsic Fe post completion of standard in vitro digestion protocol where “gastric” pH is 2, followed by the “intestinal” pH 7 (12). Description of variety including source, variety name (if known), and relevant research papers related to this sample/variety. Values represent the mean of six independent replicates. a b

Statistical Analysis. Data were analyzed using the software package GraphPad Prism (GraphPad Software, San Diego, CA, USA). Two-tailed t tests were performed by comparing the observed values to the expected values for the 58Fe:56Fe ratio at each concentration of Fe added to the samples. Due to the high number of t tests performed and to account for the Bonferroni principle, a p value of 0.001 was selected to denote statistical significance. Moreover, because from a practical standpoint a difference >10% is potentially significant, only differences with p values 10% from the expected values were noted in the results. Values are expressed as the mean ± SEM, n = 6 independent replications. Calculations. The natural abundances of 56Fe and 58Fe are 91.754 and 0.282%, respectively. Therefore, given that extrinsic 58Fe was added at 1, 10, 50, and 100% of the total Fe in the samples, the expected ratios of 58Fe:56Fe can be calculated to be 0.010899, 0.10899, 0.5449, and 1.0899, respectively.

g of bile extract (Sigma B8631) in a total volume of 245 mL. The pH of the mixture was then adjusted to approximately 7.0, and the final volume of each mixture was adjusted to 15.0 mL by weight using a salt solution of 140 mmol/L of NaCl and 5.0 mmol/L of KCl at pH 6.7. At this point, the mixture was referred to as a “digest”. The samples were then incubated for an additional 2 h at 37 °C, at which point the digests were centrifuged, and supernatants and pellet fractions were collected and transferred to tubes for analysis. Six independent replications of the in vitro digestion procedure were for all of the food samples. Analysis of Fe Isotopes. Analyses of Fe were conducted using an inductively coupled plasma mass spectrometer (Agilent 7500c ICP-MS Agilent Technologies, Santa Clara, CA, USA) after wet-ashing with HNO3 and HClO4. To separate sample digest supernatant and pellet, the sample digests were centrifuged at 4000g for 20 min. Exactly 5.0 mL of the supernatant was transferred into quartz tubes for wet-ashing and analysis by ICP-MS. The pellet of each sample digest was collected after removal of all of the supernatant, dried in a freeze-dryer, weighed into quartz tubes for wet-ashing, and analyzed by ICP-MS. However, due to the inherent difficulty in accurately separating all of the supernatant from the pellet and the relatively soft texture of the insoluble pellet fraction, only the supernatant results are presented in this paper. The isotopes that were quantified were 56Fe and 58Fe, where 56Fe was used to represent the intrinsic Fe and 58Fe was added as the extrinsic form.

■

RESULTS As stated in the methodology, only values pertaining to the supernatants of the in vitro digests are presented. This was deemed a reasonable approach for reasons both technical (i.e., the insoluble fraction proved to be substantially more difficult to handle due to its soft texture and thus was less accurately quantified) and practical as the results were complementary 9623

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry

Figure 2. Summary of the extrinsic 58Fe to intrinsic 56Fe ratios in the supernatant of in vitro digests of white and black bean samples where the gastric digestion pH values of 2 and 4 were compared. Extrinsic 58Fe was added at 1, 10, 50, and 100% of the total intrinsic Fe. Dashed lines represent the expected ratio of extrinsic to intrinsic Fe if complete exchange occurs. Asterisk indicates a significant difference (p < 0.001) >10% from the expected ratio. Values are expressed as the mean ± SEM, n = 6 independent replications.

(i.e., conversely high or low) to what was found in the supernatant. Furthermore, the soluble fraction is likely to best represent the bioavailable Fe; thus, it seems most relevant to focus on the soluble Fe fraction and not overload the paper with unnecessary data. Figure 1 and Table 1 document that in multiple crops, and at numerous Fe concentrations, differences in the extrinsic to intrinsic Fe ratio (58Fe/56Fe ratio) can be substantially different from the expected (i.e., full equilibration or complete exchange values) values, ranging as high as 340%. Moreover, in 31 of the 40 conditions shown in Figure 1, significant differences (p values 10% were measured; moreover, 26 of 40 exhibited values differing from expected values by >20%. In general, these differences were prevalent regardless of the concentration of the extrinsic 58Fe added to the in vitro digest. The black beans demonstrated the greatest differences from the expected 58Fe/56Fe ratio, consistently several-fold higher, indicating that more of the extrinsic 58Fe became soluble, relative to the intrinsic 56Fe. The white beans, the red kidney bean, and the small red bean also exhibited differences higher than the expected values for 58Fe/56Fe ratio and did not exhibit any values below the expected ratios. Interestingly, the Rwanda beans did not follow the trends of the other bean samples, exhibiting fewer and lesser differences from the expected values (Table 1 and Figure 3). The effects of gastric pH were tested on white, black, and Rwanda beans (i.e., a brown carioca) samples (Table 1,; Figures 2 and 3). No effect of the higher pH (pH 2 vs pH 4) was noted

for the white beans; however, the higher gastric pH did create a higher 58Fe/56Fe ratio for the black beans. For the Rwanda beans, the higher gastric pH did not have a consistent effect on the 58Fe/56Fe ratio (Figure 3), demonstrating a clear pH effect onlyat the 100% level of extrinsic Fe.

■

DISCUSSION In the present study, the four levels of extrinsic Fe were selected as they span the historic range of extrinsic labeling of food Fe. Most recently, studies in beans have been done in the 7−30% range,9,16 whereas older radioisotopic studies have been closer to the 1% level.17,18 Some recent studies have also used extrinsic Fe equal to approximately 50% of the intrinsic Fe,19 and several studies have been published in which the amount of added extrinsic Fe was >50% of the intrinsic Fe and reasonably close to but not exceeding 100%.18 Overall, the results of the present study clearly demonstrate that extrinsic labeling of the food Fe has potential for significant inaccuracy and that the degree of incomplete exchange can be different not only among color classes, such as in beans, but also between varieties and harvests of such crops. High polyphenol content such as the black beans can have strong effects on Fe exchange; however, the white bean samples (i.e., low polyphenol content) used in the present study also demonstrated significant levels of incomplete exchange, although not to the same extent as the black beans (Figures 1 and 2). Interestingly, the red bean and the Rwanda bean samples demonstrated significant nonexchange only at the 1 and 10% levels of extrinsic Fe, and not at the higher levels of 50 9624

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry

Figure 3. Summary of the extrinsic 58Fe to intrinsic 56Fe ratios in the supernatant of in vitro digests of low-Fe and high-Fe bean samples from Rwanda, where the gastric digestion pH values of 2 and 4 were compared. Extrinsic 58Fe was added at 1, 10, 50, and 100% of the total intrinsic Fe. Dashed lines represent the expected ratio of extrinsic to intrinsic Fe if complete exchange occurs. Asterisk indicates a significant difference (p < 0.001) >10% from the expected ratio. Values are expressed as the mean ± SEM, n = 6 independent replications.

exchange of extrinsic Fe with intrinsic. The higher gastric pH resulted in less exchange of the extrinsic Fe with the intrinsic; thus, in conditions that are more likely to exist in the human stomach, the accuracy of extrinsic Fe labeling appears even more questionable. In 1983, a critical review of extrinsic labeling of food Fe thoroughly summarized a multitude of studies, conducted in the 1970s, that compared extrinsic and intrinsic Fe labeling.18 They concluded that extrinsic tag technique cannot be considered proven with regard to all types of foods, that it is not known how completely the different nonheme forms of iron are labeled by an extrinsic tag, and that the Fe exchange process was tested only indirectly, using comparison of absorption of the extrinsic and intrinsic isotopes. This observation is surprising as in vitro methods for Fe bioavailability were developed in the late 1970s and early 1980s.27,28 These methods could certainly have been adapted to measure Fe exchange; hence, this aspect of Fe bioavailability methodology appears to have been overlooked. Since 1983 only two studies have revisited the issue of validity and accuracy of extrinsic labeling of Fe in foods, with both studies comparing varieties of beans.22,29 In one study, a human absorption trial comparison of absorption of Fe from the intrinsic Fe and the extrinsic Fe isotope was used as the measure of Fe exchange.22 In this study Fe absorption was very low (i.e., 1−1.8%), and no statistical difference in absorption was observed between intrinsic and extrinsic Fe. It appears that from this lack of statistical difference, the authors broadly concluded that extrinsic labeling may be used to screen various varieties of beans for Fe bioavailability in humans. In the other

and 100% (Figures 1 and 3). Color classes of beans are known to differ in polyphenolic profiles and contents,20 and new studies indicate that not all polyphenols are inhibitors of Fe bioavailability.21 Thus, it seems likely that the concentration of the extrinsic Fe relative to the profile and the content of Fe binding components in the food are having a role in Fe exchange. In short, these results show that not all varieties and harvests of beans, maize, and lentils can be considered the same when it comes to extrinsic Fe labeling. This is an extremely important observation as investigators who have used extrinsic Fe labeling appear to assume that all varieties and harvests of a crop will respond the same to extrinsic Fe labeling.22−24 On the basis of a multitude of research showing different polyphenolic profiles and phytate levels in staple food crops, it seems prudent to consider that such differences affect extrinsic Fe labeling.20,21,25,26 The standardized in vitro conditions used in the present study should be optimal for Fe exchange as the gastric pH in the digest is set at pH 2 for 1 h prior to titration to near-neutral pH and addition of intestinal enzymes and bile salts. Iron should be highly soluble and thus exchangeable at pH 2. Physiological studies in humans utilizing gastric and/or duodenal pH electrodes demonstrate that upon ingestion of a meal the gastric pH rapidly rises from pH 1−2 to pH 4−6 and then gradually declines over a period of 2−3 h.3,4 Studies also document that a duodenal pH 6−7 is common and that emptying of stomach contents to the duodenum begins to occur within 20 min for a more liquid type of meal.4 In light of these human studies, the in vitro gastric conditions were modified to determine the effect of a higher gastric pH of 4 on 9625

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry

experiments comparing absorption of extrinsic and intrinsic Fe and concluded that the extrinsic tag provided a valid measure of absorption. Furthermore, the authors conducted eight studies with maize, one with wheat, one with soybean, one with black bean, one with the combination of maize and veal, and two with a “complete meal” that contained maize, potatoes, beef, bread, margarine, peaches, and milk. The authors presented a convincing figure showing how the extrinsic to intrinsic ratio tracked slightly higher at 1.10 but reasonably close to a value of 1 (see Figure 1 of ref 23); however, to get this figure, the authors excluded three studies with the highest variability and a fourth study that had the lowest variability. This was done to achieve similar variability across studies and “permit statistical comparison”. Thus, 9 of the remaining 10 studies were all using the same maize sample and one study used wheat. Excluded studies included one with soybean, one with black bean, and two with maize. Upon further examination of the appendix to that manuscript by Cook et al. (see Tables IV and V of ref 23, which list the individual absorption ratios of each subject), one sees that in 54% of the subjects the extrinsic to intrinsic absorption ratio was ≥1.10 and that in 28% of the subjects the ratio was ≥1.20. To summarize, this validation study appears to have used only one harvest or variety of maize, selectively excluded soybean and bean samples from their overall analysis, and documented that in 28% of the overall subjects the absorption ratio was off by >20%. Similar evidence of incomplete exchange is also documented in numerous studies summarized by Consaul and Lee.18 The authors of the present paper encourage readers to closely examine these validation studies and consider this information in the evaluation of the extrinsic Fe labeling method. Aside from the present study and one other,29 all of the validation studies tested the methodology indirectly via absorption. None have provided evidence of exchange using simple, more direct, methods such as solubility. Also, absorption of Fe, whether it be extrinsic label or intrinsic Fe, is often in the range of 0.1−8% for meals based on staple food crops.18,24,34 Therefore, with such low absorption values it is possible and likely that in many cases similar low percent absorption of extrinsic and intrinsic Fe could occur without complete exchange. Given this possibility, the question remains as to when extrinsic labeling can be considered a valid methodology. If there is doubt, what should be done to verify that exchange is complete? In vitro methods such as utilized in the present study represent a reasonable approach to assess Fe exchange and perhaps adjust the estimate of absorption. At the very least, the in vitro method could help decide if the extrinsic labeling approach is proper for the study design. A practical question to ask in regard to extrinsic labeling is, What are the ramifications if the accuracy is off by 10−30%? This is relevant as iron absorption studies are particularly useful in estimating the absorption efficiency of staple food crops targeted for biofortification. These measures of Fe bioavailability are used by nutritionists to establish target values of seed Fe concentration for plant breeders. Thus, if bioavailability estimates are in error, then breeding efforts could be misled and insufficient differential between common and biofortified lines could result. This could lead to several years of research effort that ultimately may not prove fruitful for this variety of beans. Indeed, recent in vitro and in vivo research on a line of highiron black beans suggests that such a scenario would be likely as black beans 1 and 2 used in this study were from the same harvest and batch of bean samples used in a human efficacy

study, in vitro digestion was used, and extrinsic and intrinsic Fe levels were directly measured in supernatant and pellet fractions of the digests.29 These authors found that extrinsic Fe does not always equilibrate well with the intrinsic Fe of beans. Using an in vitro methodology similar to that used by Jin et al.,29 the present study essentially expands this work to include multiple white, black, red, and brown bean samples, plus dehulled red lentils and maize. Taken together, these observations clearly demonstrate that the primary assumption of extrinsic labeling may result in inaccurate assessment of Fe absorption from staple food crops. It has been suggested that the digestion methods used by Jin et al.29 were “too short and not vigorous enough to be a good predictor of the human digestion process”.23 More specifically, the suggestion was that the gastric phase of the in vitro digestion process should be 2 h and not just 1 h. This statement was based on recent research that demonstrated in a “solid” meal containing eggs and toast that the median half gastric emptying time was observed to be 127 min.23,30 For a liquid meal in that study, the median half gastric emptying time was about 80 min; however, it should be noted that numerous other factors can influence the gastric emptying time, and other studies have documented gastric emptying time to be much less even for solid meals.31,32 The same in vitro conditions used by Jin et al. were also used in the present study, so, in this regard, let us consider the following. Overall knowledge of food digestion has been well summarized.33 Pepsin activity in the stomach is known to be optimized at pH 2−3, but it only initiates the process of protein digestion, providing about 10− 20% of the total protein breakdown with the bulk of the protein and carbohydrate digestion occurring in the intestine. Upon ingestion of a meal, the pH values of the human stomach and duodenum have been directly measured in a number of studies, and within minutes gastric pH immediately goes up to the pH of the meal and does not fall back down to the pH 2−3 range until about 2−3 h post consumption.3,4 In the present study, the “meal” consists of 1 g of a finely ground cooked and freezedried bean sample that is subjected to gastric digestion in about an 11 mL volume of liquid (normal saline). The pH is immediately set to pH 2 during the gastric digestion and kept there for exactly 1 h; thus, pepsin activity should be optimized. After the addition of intestinal enzymes and bile salts that further break down the meal, the volume is adjusted to 15 mL and 2 h of intestinal digestion ensues. This meal is therefore high in liquid content, and digestion conditions are optimal for enzyme activity. Protein digestion post gastric and post intestinal digestion has been evaluated in this model system and confirmed to be thorough for staple food crops such as those used in the present study.27 Furthermore, in the references cited by Petry and Hurrell,23 it is stated that the 2 h gastric digestion was selected as it was deemed “convenient” and that a previous in vitro study documented that Fe release in pepsin incubations was the same for gastric digestion times ranging from 50 to 180 min.27,28 On the basis of all of the above, it does not appear that the postulated claim23 of insufficient gastric digestion can be a factor for the incomplete exchange of extrinsic Fe documented by Jin et al.29 and the present study. One could also argue that extrapolating gastric median emptying time to an in vitro assay is also irrelevant as the degree of digestion is the primary issue and not the overall time of the gastric digestion phase. One of the most widely cited validation studies was that of Cook et al.24 In this paper the authors reported on a total of 14 9626

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry study in Mexico in 2011.13 These harvests were simultaneously studied in vitro and in vivo with an animal feeding model,35 and although the black beans were significantly higher in Fe content, essentially meeting the target values in plant breeding, the fractional Fe bioavailability was significantly lower due to higher levels of polyphenols in the high-Fe beans. The human efficacy results with these black beans were in agreement with the in vitro and animal studies, showing a disappointing lack of effect of consuming the high-Fe beans.13,35 Bioavailability values are extremely useful in the proper design of efficacy trials, helping to ensure adequate subjects and duration of feeding.36 Obviously, for a strategy such as biofortification, the more precise an Fe bioavailability measurement can be for a given crop, meal, or diet, the more confidence one can have for a successful outcome. Given the expense of human Fe absorption studies (approximately $50,000−100,000) and efficacy studies in biofortification (approximately $300,000− 500,000), it would seem prudent to ensure that the determination of Fe bioavailability is accurate. As per the published history of the extrinsic Fe labeling technique, is it acceptable to have a primary assumption in methods that could be inaccurate by 10−30%? So if extrinsic labeling has accuracy concerns due to potential for incomplete labeling and intrinsic labeling is often costprohibitive and has limitations in terms of being representative of field-grown crops, then what is an alternative approach for the evaluation of foods and meal conditions that can improve human Fe absorption? One approach that shows promise is the combination of the in vitro digestion/Caco-2 cell model coupled with subsequent studies using a poultry model.31,33 Recent research has demonstrated that these tools can qualitatively match human efficacy studies with crops such as pearl millet and beans.13,35−37 Indeed, the most recent parallel studies between these screening tools and the human efficacy trial in Rwanda with biofortified beans (the same varieties and harvests used in this paper) demonstrated that the poultry model can be used to model the human meal plan, even one that utilizes a cafeteria-type setting.38 In conclusion, the present study suggests that, at the very least, renewed concern for the accuracy of extrinsic Fe labeling is warranted. The technique of extrinsic Fe labeling has certainly been useful in studies of Fe absorption; however, future studies using this method should consider more direct testing of the primary assumption of Fe exchange prior to conducting the human trial. Adjustments and interpretation of the results should therefore be discussed on the basis of the measured potential for incomplete equilibration of the extrinsic label.

■

(2) Wheby, M. S. Site of iron absorption in man. Scand. J. Haematol. 1970, 7, 56−62. (3) Gardner, J. D.; Ciociola, A. A.; Robinson, M. Measurement of meal-stimulated gastric acid secretion by in vivo gastric autotitration. J. Appl. Physiol. 2002, 92, 427−434. (4) Tyssandier, V.; Reboul, E.; Dumas, J. F.; Bouteloup-Demange, C.; Armand, M.; Marcand, J.; Sallas, M.; Borel, P. Processing of vegetableborne carotenoids in the human stomach and duodenum. Am. J. Physiol. Gastrointest. Liver Physiol. 2003, 284, G913−G923. (5) Hotz, C.; McClafferty, B. From harvest to health: challenges for developing biofortified staple foods and determining their impact on micronutrient status. Food Nutr. Bull. 2007, 28 (2 Suppl.), S271−S279. (6) Bailey, R. L.; West, K. P.; Black, R. E. The epidemiology of global micronutrient deficiencies. Ann. Nutr. Metab. 2015, 66 (Suppl. 2), 22− 33. (7) Welch, R. M. Biotechnology, biofortification, and global health. Food Nutr. Bull. 2005, 26, 419−421. (8) De Moura, F. F.; Palmer, A. C.; Finkelstein, J. L.; Haas, J. D.; Murray-Kolb, L. E.; Wenger, M. J.; Birol, E.; Boy, E.; Peña-Rosas, J. P. Are biofortified staple food crops improving vitamin A and iron status in women and children? New evidence from efficacy trials. Adv. Nutr. 2014, 5, 568−570. (9) Petry, N.; Egli, I.; Gahutu, J. B.; Tugirimana, P. L.; Boy, E.; Hurrell, R. Phytic acid concentration influences iron bioavailability from biofortified beans in Rwandese women with low iron status. J. Nutr. 2014, 144, 1681−1687. (10) Sankaran, R. P.; Grusak, M. A. Whole shoot mineral partitioning and accumulation in pea (Pisum sativum). Front. Plant Sci. 2014, 23, 149. (11) Petry, N.; Egli, I.; Zeder, C.; Walczyk, T.; Hurrell, R. Polyphenols and phytic acid contribute to the low iron bioavailability from common beans in young women. J. Nutr. 2010, 140, 1977−1982. (12) Glahn, R. P.; Lee, O. A.; Yeung, A.; Goldman, M. I.; Miller, D. D. Caco-2 cell ferritin formation predicts nonradiolabeled food iron availability in an in vitro digestion/Caco-2 cell culture model. J. Nutr. 1998, 128, 1555−1561. (13) Haas, J. D.; Villalpando, S.; Beebe, S.; Glahn, R. P.; Shamah, T.; Boy, E. The effect of consuming biofortified beans on the iron status of Mexican school children. FASEB J. 2011, 25, 96.6. (14) Haas, J. D.; Luna, S.; Lung’aho, M.; Ngabo, F.; Wenger, M.; Murray-Kolb, L.; Beebe, S.; Gahutu, J. B.; Egli, I. Iron biofortified beans improve iron status in Rwandan university women: results of a feeding trial. FASEB J. 2014, 28 (1), Suppl. 646.1. (15) Tako, E.; Hoekenga, O. A.; Kochian, L. V.; Glahn, R. P. High bioavailablilty iron maize (Zea mays L.) developed through molecular breeding provides more absorbable iron in vitro (Caco-2 model) and in vivo (Gallus gallus). Nutr. J. 2013, 12 (3), 3. (16) Petry, N.; Egli, I.; Gahutu, J. B.; Tugirimana, P. L.; Boy, E.; Hurrell, R. Stable iron isotope studies in Rwandese women indicate that the common bean has limited potential as a vehicle for iron biofortification. J. Nutr. 2012, 142, 492−497. (17) Beiseigel, J. M.; Hunt, J. R.; Glahn, R. P.; Welch, R. M.; Menkir, A.; Maziya-Dixon, B. Iron bioavailability from maize and beans: a comparison of human measurements with Caco-2 cell and algorithm predictions. Am. J. Clin. Nutr. 2007, 86, 388−396. (18) Consaul, J. R.; Lee, K. Extrinsic tagging in iron bioavailability research. J. Agric. Food Chem. 1983, 31, 684−689. (19) Petry, N.; Egli, I.; Campion, B.; Nielsen, E.; Hurrell, R. Genetic reduction of phytate in common bean (Phaseolus vulgaris L.) seeds increases iron absorption in young women. J. Nutr. 2013, 143, 1219− 1224. (20) Lin, L.-Z.; Harnly, J. M.; Pastor-Corrales, M. S.; Luthria, D. L. The polyphenolic profiles of common bean (Phaseolus vulgaris L.). Food Chem. 2008, 107, 399−410. (21) Hart, J. J.; Tako, E.; Kochian, L. V.; Glahn, R. P. Identification of black bean (Phaseolus vulgaris L.) polyphenols that inhibit and promote iron uptake by Caco-2 cells. J. Agric. Food Chem. 2015, 63, 5950−5956.

AUTHOR INFORMATION

Corresponding Author

*(R.P.G.) E-mail: [email protected]. Phone: (607) 255-2452. Fax: (607) 255-1132. Funding

This research was funded by USDA-ARS. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Engle-Stone, R.; Yeung, A.; Welch, R.; Glahn, R. Meat and ascorbic acid can promote Fe availability from Fe-phytate but not from Fe-tannic acid complexes. J. Agric. Food Chem. 2005, 53, 10276− 10284. 9627

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628

Article

Journal of Agricultural and Food Chemistry (22) Donangelo, C. M.; Woodhouse, L. R.; King, S. M.; Toffolo, G.; Shames, D. M.; Viteri, F. E.; Cheng, Z.; Welch, R. M.; King, J. C. Iron and zinc absorption from two bean (Phaseolus vulgaris L.) genotypes in young women. J. Agric. Food Chem. 2003, 51, 5137−5143. (23) Petry, N.; Hurrell, R. Reply to R. P. Glahn. J. Nutr. 2015, 145, 1026−1027. (24) Cook, J. D.; Layrisse, M.; Martinez-Torres, C.; Walker, R.; Monsen, E.; Finch, C. A. Food iron absorption measured by an extrinsic tag. J. Clin. Invest. 1972, 51, 805−815. (25) Harnly, J. M.; Pastor-Corrales, M. A.; Luthria, D. L. Variance in the chemical composition of dry beans determined from UV spectral fingerprints. J. Agric. Food Chem. 2009, 57, 8705−8710. (26) DellaValle, D. M.; Thavarajah, D.; Thavarajah, P.; Vandenberg, A.; Glahn, R. P. Lentil (Lens culinaris L.) as a candidate crop for iron biofortification: is there genetic potential for iron bioavailability? Field Crops Res. 2013, 144, 119−125. (27) Miller, D. D.; Schricker, B. R.; Rasmussen, R. R.; Van Campen, D. An in vitro method for estimation of iron availability from meals. Am. J. Clin. Nutr. 1981, 34, 2248−2256. (28) Rao, N.; Prabhavathi, T. An in vitro method for predicting the bioavailability of iron from foods. Am. J. Clin. Nutr. 1978, 31, 169−175. (29) Jin, F.; Cheng, Z.; Rutzke, M. A.; Welch, R. M.; Glahn, R. P. Extrinsic labeling method may not accurately measure Fe absorption from cooked pinto beans (Phaseolus vulgaris): comparison of extrinsic and intrinsic labeling of beans. J. Agric. Food Chem. 2008, 56, 6881− 6885. (30) Hellmig, S.; Von Schöning, F.; Gadow, C.; Katsoulis, S.; Hedderich, J.; Fölsch, U. R.; Stüber, E. Gastric emptying time of fluids and solids in healthy subjects determined by 13C breath tests: influence of age, sex and body mass index. J. Gastroenterol. Hepatol. 2006, 21, 1832−1838. (31) Bornhorst, G. M.; Singh, P. R. Gastric digestion in vivo and in vitro: how the structural aspects of food influence the digestion process. Annu. Rev. Food Sci. Technol. 2014, 5, 111−132. (32) Guerra, A.; Etienne-Mesmin, L.; Livrelli, V.; Denis, S.; BlanquetDiot, S.; Alric, M. Relevance and challenges in modeling human gastric and small intestinal digestion. Trends Biotechnol. 2012, 30, 591−600. (33) Hall, J. E. Chapter 66: Digestion and Absorption in the Gastrointestinal Tract. In Guyton and Hall Textbook of Medical Physiology, 13th ed.; Elsevier: Philadelphia, PA, USA, 2015. (34) Bjorn-Rasmussen, E.; Hallberg, L.; Walker, R. B. Food iron absorption in man. Isotopic exchange of iron between labeled foods and between a food and an iron salt. Am. J. Clin. Nutr. 1973, 26, 1311− 1319. (35) Tako, E.; Beebe, S. E.; Reed, S.; Hart, J. J.; Glahn, R. P. Polyphenolic compounds appear to limit the nutritional benefit of biofortified higher iron black bean (Phaseolus vulgaris L.). Nutr. J. 2014, 13, 28. (36) Finkelstein, J. L.; Mehta, S.; Udipi, S. A.; Ghugre, P. S.; Luna, S. V.; Wenger, M. J.; Murray-Kolb, L. E.; Przybyszewski, E. M.; Haas, J. D. A randomized trial of iron-biofortified pearl millet in school children in India. J. Nutr. 2015, 145, 1576−1581. (37) Tako, E.; Reed, S. M.; Budiman, J.; Hart, J. J.; Glahn, R. P. Higher iron pearl millet (Pennisetum glaucum L.) provides more absorbable iron that is limited by increased polyphenolic content. Nutr. J. 2015, 14, 11. (38) Tako, E.; Reed, S.; Anandaraman, A.; Beebe, S. E.; Hart, J. J.; Glahn, R. P. Studies of cream seeded carioca beans (Phaseolus vulgaris L.) from a Rwandan efficacy trial: in vitro and In vivo screening tools reflect human studies and predict beneficial results from iron biofortified beans. PLoS One 2015, 10, e0138479.

9628

DOI: 10.1021/acs.jafc.5b03926 J. Agric. Food Chem. 2015, 63, 9621−9628