In terms of biofortification, target levels for bean Fe concentration have been set at approximately 90 μg/g or higher, which should likely represent a 40 μg/g differential from more typical common bean Fe levels
[29, 30]. That target value is based on calculations that assume the percent bioavailability of the Fe is similar between the normal and enhanced lines, hence delivering more absorbable Fe. This assumption is potentially disastrous in terms of planning and ultimately moving forward with a human efficacy trial, as such trials can easily cost $500,000 or more. Given this expense, it is prudent to utilize inexpensive screening tools that are validated to predict relative differences in Fe bioavailability.
A strong case can now be made for utilizing the widely accepted and inexpensive in vitro digestion/Caco-2 cell culture model to determine if biofortified lines warrant advancement to human study
[7–9]. Furthermore, this in vitro approach has recently been coupled with a poultry model for Fe bioavailability and it has confirmed the in vitro predictions
[7–9]. Such observations certainly provide confidence in advancing lines for the more expensive human trials. If consistent correlations between these screening tools and human results can be established, then future work can be guided by these models thus enhancing the effectiveness of research dollars. Moreover, such lines will have to be monitored once released, and these models represent tools that will be needed to do so.
In the present study, the in vitro results predicted that the biofortified (high Fe) bean will not provide additional absorbable Fe. The in vivo results reflect this prediction, as no significant differences in liver ferritin and liver Fe were observed, and only slight differences were observed in Hb (higher only on day 42 in the high Fe bean group) and total body Hb (also higher on day 42 for the high Fe bean group). Hemoglobin maintenance efficiency was higher in the low Fe group at days 14th, 21st and 28th, indicating an adaptive response (ie. a relative up regulation of absorption) to less absorbable dietary Fe. The duodenal relative mRNA abundance analysis of DMT-1, DcytB and ferroportin that showed no significant differences between the standard (Low Fe) group vs. the biofortified (High Fe) group (P > 0.05), suggests a lower dietary Fe bioavailability in both groups
[5–9]. Taken together, all of the above mentioned measures of Fe status indicate that the two dietary treatment groups were of relatively similar Fe status, with only a hint of more absorbable Fe being provided by the high Fe black beans.
The strong inhibition of Fe uptake reflected in the in vitro results is typical of a strong polyphenolic inhibition of Fe bioavailability. Similar results (ie. when the Caco-2 cell ferritin formation is at or below the untreated baseline values), have been seen many times in the use of this in vitro model
[7, 8, 21]. The polyphenolic profile and measurement of concentration indicates that the high Fe beans were also higher in polyphenolic content, specifically, catechin, myricetin, kaempferol 3-glucoside and quercetin 3-glucoside (Table
4), than the low Fe beans. Thus it appears that the potential nutritional benefit from the higher Fe content was offset by the increased levels of polyphenolics. These observations clearly suggest that further research should be done to explore the metabolic and potential genetic link between Fe content and polyphenols.
A wealth of information is available on the phenolic and polyphenolic profile of foods
[31–34], and specifically in regards to beans
[27, 35]. In addition, there are a considerable number of papers that focus on the Fe-binding properties of polyphenols as related to antioxidation
. Perron et al.
 investigated the kinetics of Fe oxidation upon binding to polyphenols and observed that gallol-containing polyphenols have faster rates of Fe oxidation (ie. Fe+2 to Fe+3) than their catechol analogs. This faster rate of oxidation was believed to be related to the strength of the Fe binding. Thus, compounds such as epicatechin and quercitin would be expected to have slower rates of Fe oxidation and less strength of Fe binding than their respective gallol analogs, epigallocatechin and myricitin. These differences could be translated into different effects on Fe bioavailability as greater oxidation and stronger binding of Fe is generally correlated with lower bioavailability. However it is important to note that in the above studies, conditions for the measured reactions were not designed to match the conditions of the intestinal lumen, hence the oxidation and binding of Fe could differ.
In regards to Fe bioavailability there is clear evidence that polyphenolic compounds related to seed coat color in beans have a strong inhibitory effect on Fe bioavailability. Several in vitro and in vivo studies have demonstrated that Fe bioavailability from black, red and pinto beans is significantly less than that of white beans
[5–8, 21, 38]. To date there has only been one published study that has linked specific bean seed coat polyphenols to inhibition of Fe bioavailability
. In that study the flavonoids kaempferol and quercitin identified in red, pinto and black bean hulls were shown to inhibit Fe bioavailability. This work did not exclude any other compounds from having an effect on Fe bioavailability, it merely documented inhibitory effects of these compounds which were found in high quantities in the red and pinto bean seed coats, to a lesser extent in black beans, and absent from white beans.
An interesting study by Lin et al.
 profiled but did not quantify the polyphenolics in 24 bean samples representing 17 varieties and 7 “generic off the shelf” samples, with the total group belonging to 10 commercial US market classes. Their work showed that the beans could be organized into 6 classes based on their polyphenolic profiles. The classes were black, pinto, pink, light red kidney bean, small red and navy. In general, these groups contained similar hydroxycinnaminic acid derivatives as their main phenolic profile with the flavonoid content less prominent but more varied and distinctive for each group. For each group there were 10 or less flavonoid compounds that were predominant. If one were to screen these for effects on Fe bioavailability then certainly a relatively high throughput system would be needed, as concentration and interactive effects would also need to be considered. In our opinion, the in vitro digestion/Caco-2 model
 or a variation thereof is capable of performing this work.
Recently, we were able to demonstrate that specific polyphenols found in black bean seed coats are inhibitors of Fe uptake whereas others are promoters of Fe uptake
. This work uses a modification of the in vitro digestion/Caco-2 cell model
 and LC/MS to measure relative amounts of flavonoids and hydroxycinnamic acids in black bean seed coats that were previously identified as major polyphenolic components of common bean
. The aforementioned research has enabled us to generate a key list of phenolic and polyphenol compounds in black beans that predominate and may be related to Fe bioavailability. Those compounds are reported in Table
4. Comparisons of the polyphenols between the two lines used in this study certainly indicate that the higher polyphenols in the high Fe line may have negated the nutritional benefit of higher Fe. These differences could be environmental in origin, or possibly linked to the effort to breed for higher Fe. Regardless, these results indicate that the polyphenolic profile of beans should be considered for future human efficacy studies of high Fe beans.
The list of polyphenols and flavonoids reported here represent the predominant compounds that were detected in the tested black beans, however, these compounds are part of a major group of compounds observed in black beans as previously reported
. As it was shown before, several of the detected compounds present Fe+3 binding ability that results in decreased Fe bioavailability and absorption
[34–37, 40]. For example, quercetin, which is one of the most common flavonols present in nature and is the predominant flavonoid in the diet, and was detected in increased concentrations in the biofortified Fe bean variety. Quercetin was shown to be a strong Fe+3 chelator, hence, suggested to decrease Fe bioavailability in foods
. In addition the compound catechin, that was also detected in increased concentration in the biofortified Fe bean variety was also reported to have strong binding affinity to Fe+3, specifically in a less acidic environment, as in the duodenum where most absorption of Fe occurs
. Increased concentrations of kaempferol 3-glucoside were also detected in the biofortified Fe black bean variety. Previously, it was demonstrated that kaempferol 3- glucoside in red and pinto bean seed inhibited Fe bioavailability in vitro. Thus, seed coat kaempferol was identified as a potent inhibitory factor affecting iron bioavailability in the red and pinto beans studied. Results comparing the inhibitory effects of kaempferol, quercitrin, and astragalin on iron bioavailability suggest that the 3′,4′-dihydroxy group on the B-ring in flavonoids contributes to the lower iron bioavailability
In summary, the current study suggests that increasing Fe concentrations in black beans by about 30 μg/g provides a very small increase in the amount of bioavailable and therefore absorbable Fe in vivo. The increased polyphenol concentration in the high Fe beans indicates that these compounds must also be considered when advancing high Fe lines for human studies in order to achieve significant nutritional benefits.