Evidence for decreased interaction and improved carotenoid bioavailability by sequential delivery of a supplement

Abstract Despite the notable health benefits of carotenoids for human health, the majority of human diets worldwide are repeatedly shown to be inadequate in intake of carotenoid‐rich fruits and vegetables, according to current health recommendations. To address this deficit, strategies designed to increase dietary intakes and subsequent plasma levels of carotenoids are warranted. When mixed carotenoids are delivered into the intestinal tract simultaneously, competition occurs for micelle formation and absorption, affecting carotenoid bioavailability. Previously, we tested the in vitro viability of a carotenoid mix designed to deliver individual carotenoids sequentially spaced from one another over the 6 hr transit time of the human upper gastrointestinal system. We hypothesized that temporally and spatially separating the individual carotenoids would reduce competition for micelle formation, improve uptake, and maximize efficacy. Here, we test this hypothesis in a double‐blind, repeated‐measure, cross‐over human study with 12 subjects by comparing the change of plasma carotenoid levels for 8 hr after oral doses of a sequentially spaced carotenoid mix, to a matched mix without sequential spacing. We find the carotenoid change from baseline, measured as area under the curve, is increased following consumption of the sequentially spaced mix compared to concomitant carotenoids delivery. These results demonstrate reduced interaction and regulation between the sequentially spaced carotenoids, suggesting improved bioavailability from a novel sequentially spaced carotenoid mix.

Despite the notable health benefits, the majority of human diets worldwide are repeatedly shown to be inadequate in intake of carotenoid-rich fruits and vegetables, according to current health recommendations (Blanck, Gillespie, Kimmons, Seymour, & Serdula, 2008;Murphy, Barraj, Spungen, Herman, & Randolph, 2014;Murphy et al., 2012;Tennant, Davidson, & Day, 2014;WHO, 2002). One strategy to improve the variety and quantity of carotenoids included in dietary regimens is through fortification or supplementation. Consumption of a mixed variety of carotenoids is suggested to be more beneficial than focusing on supplementation of a single carotenoid in lieu of others (Astley, Ruan, Archer, & Southon, 2004;Liu, 2003;Tapiero, Townsend, & Tew, 2004). In addition to increased dietary carotenoid consumption, a mixed carotenoid supplement should deliver a diversity of carotenoids contained within healthy diets comprised of adequate fruit and vegetable content.
When considering the formulation of a mixed carotenoid supplement, absorption and bioavailability are the key variables influencing efficacy. The small intestine is responsible for absorbing hydrophobic substances, such as carotenoids, so they can be subsequently delivered through the blood stream to peripheral tissues. During digestion in the upper intestine, carotenoids are incorporated into mixed micelles composed of phospholipids, lipids, cholesterol, and bile salts and absorbed into the enterocyte through passive and facilitated diffusion (Hollander & Ruble, 1978;Valacchi, Sticozzi, Lim, & Pecorelli, 2011;Van Bennekum et al., 2005;Yonekura & Nagao, 2007). After absorption, carotenoids are packaged into plasma and secreted into the lymph system for transport to the bloodstream.
However, when various mixed carotenoids are delivered into the intestinal tract together at higher concentrations, the compounds compete for micelle formation and subsequent intestinal absorption (Maiani et al., 2009;van den Berg, 1998van den Berg, , 1999van het Hof, West, & Westerate, 2000). Carotenoids may also compete with one another for uptake and metabolism in the enterocyte and for incorporation into the plasma (Kostic, White, & Olsen, 1995;Paetau, Huiping, Goh, & White, 1997;van den Berg & Van Vliet, 1998). This competitive interaction can hinder carotenoid absorption and transportation to target end sites.
In this regard, a method delivering mixed carotenoids that are temporally and spatially separated within the gastrointestinal tract may minimize micelle competition, consequently improving overall bioavailability of individual carotenoids. The hypothesis tested here is that a mix utilizing carotenoid spacing will have better absorption due to decreased competition among the individual carotenoids for improved bioavailability and uptake.
In this study, a mix of natural carotenoids was tested incorporating a sequential release of individual carotenoids across gut transit time of approximately 6 hr to elucidate the impact on uptake.
The designed carotenoid mix is composed of the primary carotenoids found in human plasma that are associated with a variety of fruit and vegetable intake: β-and α-carotene, lutein, lycopene, and zeaxanthin (Maiani et al., 2009). In light of its reported health benefits but general exclusion in the average diet, astaxanthin was also included in the mix to determine if competitive interaction exists with the common dietary carotenoids. (Fassett & Coombes, 2012;Yuang, Peng, Jin, & Wang, 2011).
The mix of carotenoids were included in a set ratio, designed by incorporating key aspects of well-accepted dietary guidelines and considering typical and wide-spread dietary inadequacies, such as those exposed by other recent results (Murphy et al., 2012(Murphy et al., , 2014. These results suggested a carotenoid intake amount that would be found in a "prudent diet" following the World Health Organization's recommendation of at least five servings of fruits and vegetables per day (Joint FAO/WHO Workshop on Fruit and Vegetables for Health, 2004). The amount and proportion of carotenoids found in this "prudent diet" served as the baseline to create a ratio of carotenoids that, when consumed, would help improve carotenoid intake. Notably, the ratio was not designed to simply increase the amounts of carotenoids but rather to promote a mix of carotenoids as would be found in a healthy diet with a minimum of five servings of fruit and vegetables (Blanck et al., 2008;Gellenbeck, Salter-Venzon, Lala, & Chavan, 2012;Murphy et al., 2012). This ratio is shown in Table 1.
The designed spacing sequence was informed by the interactions between specific carotenoids in addition to evidence of when each carotenoid might first appear in the triacylglycerol-rich fraction of plasma after oral administration in order to approximate when peak absorption might occur (Bierer, Merchen, & Erdman, 1995;Cardinault et al., 2003;Johnson, Qin, Krinsky, & Russell, 1997;O'Neill & Thurnham, 1998;Tysandier et al., 2002;van den Berg, 1999;Van het Hof et al., 2000;Zaripheh & Erdman, 2002). Previously, a carotenoid mix was designed and tested in vitro for its ability to deliver individual carotenoids sequentially spaced from one another over the 4 to 6 hr of transit time through the upper intestinal system of the human gut (Gellenbeck et al., 2012). This study expands on the previous work by investigating the change of plasma carotenoid concentrations in adult humans after oral consumption of a mix of carotenoids that are sequentially spaced from one another, in comparison to a matcheddose mix of carotenoids that are delivered to the intestinal system concomitantly. The aim of this study was to determine if separating individual carotenoids spatially and temporally throughout the transit time of the upper gastrointestinal tract lends absorption and bioavailability benefits over a concomitantly delivered carotenoid mix.

| MATERIALS AND METHODS
Twelve healthy, nonsmoking men and women volunteers between the ages of 18 and 60 were recruited for this study. Prior to entry, written consent was obtained from each volunteer. All procedures adopted were in accordance with the Helsinki Declaration and this study was approved by Alpha Independent Review Board (San Clemente, CA).
The volunteers were judged to be in good health, on the basis of a medical exam, with no history of chronic disease, blood cholesterol levels between 120 and 219 mg/dl, triglycerides between 35 and 164 mg/dl, body mass index between 20 and 28 kg/m 2 , and not taking any dietary supplements in the 2 months prior to the study. The volunteers were instructed to maintain their normal diet and exercise habits throughout the duration of the study, with the exception of avoiding consumption of carotenoid-rich fruit or vegetables in the 3 days prior to each visit. Subjects were verbally instructed on carotenoid-rich foods and given a written list of foods to avoid during this time frame. Volunteers arrived at the testing center on the day of experiment following a 12-hr overnight fast. Compliance was verified through a three-day diet history recall. Each subject was randomly assigned to one of two matched-dose, but differently delivered, mixes of carotenoid beadlets in the form of a two-piece capsule containing noncompressed powder (as opposed to a compressed tablet or soft gel delivery format): one in which the individual carotenoids are sequentially separated from one another (sequentially spaced carotenoid mix, SSCM) or one in which each carotenoid is delivered concomitantly (concomitant carotenoid mix, CCM). The mix of carotenoids were consumed 20 min after eating a carotenoid-free, high-fat meal (≈50% of calories from fat). The meal, composed of white biscuits, white gravy, hash browns, white bread, coffee and/or water, contained 1270 calories and 68 grams of dietary fat. A baseline fasting blood sample was taken (time 0), followed by an allotted 20 min for the consumption of the carotenoid-free meal, and additional blood samples taken 1, 2, 4, 6, and 8 hr after consumption of the carotenoid mix. Subjects did not receive a second meal, or any other food or drink, other than water, for the remaining 8 hr of the postprandial experiment.
Fourteen to sixteen days later, the subjects returned to the research center to repeat the procedure in a double-blinded, repeatedmeasure, crossed-over design such that each subject received the opposite matched-dose carotenoid mix from the previous visit. It is known that individual subjects can absorb carotenoids differently in dynamics and magnitude, leading to large intersubject variability despite the study being well-controlled (Brown et al., 1989;O'Neill & Thurnham, 1989). Commonly, these differences are noted by referring to individuals as being either strong or weak responders, and indicates that carotenoid responses are, at least in part, under the control of specific genetic variants (Wang, Edwards, & Clevidence, 2013

| Carotenoid mix composition
All active ingredients contained within the carotenoid mixes were derived from natural sources and combined with standard inert ingredients to form the carotenoid mixes as described previously (Gellenbeck et al., 2012). Briefly, the carotenoid mix to include sequential spacing was prepared by Omniactive Health Technologies (Morristown, NJ) using proprietary technology. Layers of carotenoid extracts are applied to an inner sugar core along with functional release coats and an outer protective coat. The final form is a small beadlet suitable for inclusion in functional foods, capsules, or tablets. In addition, a matched carotenoid mix was prepared without the sequential spacing for use as a comparator. The active ingredient sources are as follows: Natural β-carotene derived from Dunaliella algae and palm fruit, natural α-carotene derived from palm fruit, lutein from marigold flowers, lycopene from tomato, and astaxanthin from Haemotococcus algae. The amounts contained in SSCM and CCM could be closely, but not perfectly, matched due to the nature of the methods used to sequentially space the carotenoids in SSCM. Therefore, β-carotene and astaxanthin were chosen as the carotenoids used as reference to match the dosage provided from each formulation, that is, β-carotene and astaxanthin doses were exactly the same in both formulations and the small differences in the other carotenoids were taken into account during analysis. The compostion, ratio, and amount of mixed carotenoids provided to subjects is shown in  (Johnson et al., 1997;Maiani et al., 2009;O'Neill & Thurnham, 1989).

| Statistical analysis
Data are expressed as means ± standard error of the means (SEMs). To

| RESULTS
Study Population: Subject baseline characteristics are presented in

| Effect of spacing carotenoids on postprandial plasma response
The changes in individual postprandial plasma carotenoid concentrations after either CCM or SSCM are shown in Figures 1 and 2. Figure 1 illustrates the astaxanthin, lutein, and zeaxanthin plasma responses.
Astaxanthin response to SSCM begins 1 hr following consumption, peaks at 4 hr, and plateaus for the remainder of the experimental measures. In contrast, after consumption of CCM, the same subjects demonstrate a decrease in astaxanthin for the first 2 hr, which reverses to peak by 6 hr, then quickly plummets over the remaining measured time points. Astaxanthin measures were significantly higher after SSCM than CCM at 2 and 4 hr, however, the statistical difference in incremental change did not remain over 8 hr. The SSCM AUC was 12.75 ng/ml*hr ± 15.7 greater than CCM AUC (p = .157).
Lutein and zeaxanthin show similar plasma responses over 8 hr after consumption of SSCM and CCM, which is not surprising since these two carotenoids have similar chemical structure (Bierer et al., 1995;Kotake-Nara & Nagao, 2011;Nagao, 2011  dynamics appear differently shaped, there were no statistically significant differences at any time points between the two carotenoid mixes, and the overall mean incremental AUCs also were not significantly different for either β-and α-carotene. The mean incremental AUC for αcarotene was 9.5 ng/ml* hr ± 24.8 less after SSCM consumption than CCM consumption (p = .645), while that for β-carotene was 23.0 ng/ ml*hr ± 61.2 greater after SSCM than after CCM (p = .357). Following consumption of SSCM, the incremental AUC of α-and β-carotene did not demonstrate much change over the course of the measured time points. The predominate difference between SSCM and CCM regarding β-and α-carotene was that the drop below baseline seen after CCM was not apparent after SSCM. Additionally, CCM appeared to peak at 6 hr while response of α-and β-carotene following SSCM was relatively flat. The final measurement (8 hr) of SSCM appeared to be that of an increasing response, so perhaps the peak of plasma carotenoid change was later than our experiment was designed to measure, suggesting that a longer experimental period would be appropriate.
An additional variable would be the inclusion of a secondary meal that could affect the uptake of carotenoids from the initial experimental dosage.
The sum of the mean incremental plasma responses of all carotenoids following consumption of SSCM or CCM is shown in Figure 3.
The postprandial plasma responses demonstrate a significant difference in the total carotenoid content at hours one-three between SSCM and CCM. Interestingly, the mean incremental plasma response showed a marked decrease from baseline following consumption of CCM for the first 4 hr, a peak level at 6 hr, followed by a steep decline.  F I G U R E 3 Effect of delivery on postprandial plasma response-total carotenoids. Error bars are constructed using 1 standard error from the mean. Two-sided paired t test illustrates total carotenoid response after SSCM is significantly higher at 1 and 2 hr than after concomitant carotenoid mix (CCM). Note the large dip from baseline carotenoid values for the first 4-5 hr following CCM delivery that may be indicative of increased competition among the individual carotenoids for intestinal absorption. Overall incremental areas under the curve (AUC) is not significantly different between delivery types, α=.05 level of significance

CCM SSCM
was not statistically significant at p = .193, most likely due to the large variability inherent in these type of small-scale experiments with carotenoids.

| Interactions and co-regulations of plasma carotenoid responses
Raw concentration values (original data, not transformed) were assessed using nonparametric Spearman's correlation ranking for each carotenoid from SSCM and CCM and compared to one another.  notion that mixed carotenoids inhibit one another in absorption and early metabolism dynamics (Maiani et al., 2009;van den Berg, 1998van den Berg, , 1999Van het Hof et al., 2000). In comparison, the muted color scheme apparent on the Spearman's heat map for SSCM (Fig. 4, right) illustrates that plasma carotenoid responses are less affected by one another. Carotenoid co-regulation following SSCM consumption is also less likely to be inhibitory than after CCM, as only three Spearman correlation coefficients were mildly negative with this type of carotenoid delivery; astaxanthin with lutein (−0.140, p = .2413), α-carotene with zeaxanthin (−0.176, p = .1397), and β-carotene with zeaxanthin (−0.195, p = .1015). The results indicate that a mix composed of naturally sourced carotenoids is taken up into the plasma after consumption by human volunteers. Although both SSCM and CCM elicited measurable increases in postprandial plasma carotenoid responses after consumption, the response curves between the two forms were noticeably different in both shape and magnitude in the same subject.

| DISCUSSION
There were differences in curve shapes and peak timing of the postprandial plasma responses following SSCM in comparison to CCM. The CCM resulted in a plasma response that, in general, appeared to peak at the 6 hr measurement, followed by a quick decline toward baseline by the end of the experiment. The key significant differences between the two mixes are shown at the early time points. SSCM resulted in a significantly higher plasma response in the first 5 hr following consumption, primarily at hours one through four with astaxanthin, lutein, lycopene, and α-carotene, (Figs. 1 and 2) supporting the notion that sequentially spacing a mix of carotenoids could be a favorable method to improve dietary carotenoid absorption. Despite the significance at early time points, analysis of AUC of all carotenoids over the entire study duration ( Fig. 3) does not reach statistical significance (p = .193). However, the increased total AUC for lycopene trended toward significance (p = .056).
Further support of a decreased interaction among individual carotenoids for absorption following the SSCM comes from the results of the nonparametric Spearman's correlation ranking analysis. In agreement with previous findings, this analysis provides evidence that individual carotenoids released into the gut lumen all at one time are highly interactive and tend to inhibit the uptake of one another (Kostic et al., 1995;Maiani et al., 2009;Paetau et al., 1997;van den Berg, 1998van den Berg, , 1999van den Berg & Van Vliet, 1998;Van het Hof et al., 2000).
In contrast, following SSCM, the results show a decreased interactivity and co-regulation among individual carotenoids, presumably due to decreased uptake competition.
Because there is natural variation in an individual's ability to absorb carotenoids, large intersubject variability is a frequent issue even in well-controlled studies (Brown et al., 1989;Maiani et al., 2009;O'Neill & Thurnham, 1998). In this study, the intersubject variability was mitigated using a cross-over design where each subject served as its own comparator and incremental changes were evaluated from their baseline values. However, some variability in our study remained as evidenced by a larger standard error accompanying a few of the measures, underpowering its ability to achieve statistical significance.
The data from this work can be used to sufficiently power a future study with a larger number of subjects.
The total postprandial incremental AUC for β-and α-carotene following SSCM was flatter than expected. Individual analysis of each subject's β-and α-carotene response did not elucidate any clear nonresponders. Thus, the flat response following SSCM could be due to these two carotenoids not releasing as well in the human gut as in the previously conducted in vitro lab simulations (Gellenbeck et al., 2012), or that the timeline for measurements in this study was too short to capture the peak of the postprandial plasma response to β-and α-carotene. Investigation of the postprandial β-carotene plasma or the chylomicron fraction of plasma, the fraction that represents the first uptake from dietary sources, illustrates that peak response typically occurs 4-5 hr after consumption of this carotenoid (Bierer et al., 1995;Johnson et al., 1997;O'Neill & Thurnham, 1998;Paetau et al., 1997;Traber, Diamond, Lane, Brody, & Kayden, 1994;Tysandier et al., 2002;van den Berg & Van Vliet, 1998). Since the β-and α-carotene in the SSCM were spaced toward the end of the 6 hr time, the peak response may occur later than what was captured in the 8 hr experiment period, up to 10-13 hr after consumption. A follow-up study incorporating more subjects, focusing on the chylomicron fraction, and using longer measurement times that would allow a return of plasma concentrations to baseline is warranted to help answer this question.
The plasma carotenoid response following CCM consumption demonstrated a curious and relatively consistent decrease from baseline for the first 4 hr. A similar phenomenon has been noted in a few examples measuring β-carotene as well, following a mixed carotenoid or a higher dose (Johnson et al., 1997;Paetau et al., 1997), but the cause of this drop is not obvious. Movement across the mucosal cells of the intestinal wall has long been thought to take place through passive diffusion (El-Gorab, Underwood, & Loerch, 1975;Hollander & Ruble, 1978). However, recent lines of investigation show the presence of facilitated diffusion for the absorption and subsequent early metabolism of carotenoids (Valacchi et al., 2011;Van Bennekum et al., 2005). Therefore, in the case of CCM, the decrease in the carotenoid response from baseline values may be due to competition among the carotenoids, competition from facilitated diffusion proteins, or perhaps carotenoid efflux back to the lumen (Nagao, 2011). Because of the more dispersed release of carotenoids across time and the length of the upper intestinal tract, this competition effect may be mitigated following SSCM. Directed experimentation is required to clarify these observations.
In summary, the results obtained show that mixed carotenoids are bioavailable and result in a plasma carotenoid response after consumption in healthy volunteers. However, when the mixed carotenoids are delivered in a format that releases carotenoids sequentially across the gut lumen, competition for absorption appears to be lessened in comparison to a format that releases its contents at one time.
In general, the postprandial plasma response to a sequentially spaced carotenoid mix increased uptake, especially during the first hours following consumption. Further investigation using a larger subject group and a longer experimental time line is warranted.

ACKNOWLEDGMENTS
This work was funded by Access Business Group, LLC. Employees of Access Business Group LLC, Amway R&D, had a role in the design of the experimentation and the interpretation of the data.

FUNDING INFORMATION
This work was funded by Access Business Group, LLC.