Evaluation of the bioaccessibility of a carotenoid beadlet blend using an in vitro system mimicking the upper gastrointestinal tract

Abstract The release characteristics of a unique blend of carotenoid beadlets designed to increase bioavailability were tested using the dynamic gastrointestinal model TIM‐1. Individual carotenoid bioaccessibility peaks were observed over approximately 3–4 hr in the order of lutein and zeaxanthin first, followed by lycopene, and then finally α‐ and β‐carotene; when tested as a beadlet blend or when the beadlets were compressed into tablets. Bioaccessibility measurements of 7%–20% were similar to those previously reported in literature and comparable between the two formulations, beadlet blend and tablet formulations. Total recovery of carotenoids from all compartments ranged from 70% to 90% for all carotenoids, except lycopene where almost 50% was unrecoverable after digestion in the TIM system.

time of the digestive system. Gellenbeck et al., 2012 reported in vitro testing of a multilayered beadlet design that resulted in a 2-3 hr separation between lycopene, β-, and α-carotene peak release under simulated gastrointestinal conditions (Gellenbeck et al., 2012). In a separate clinical evaluation, a formulation designed to deliver a sequential release of a series of carotenoids was shown to limit interactivity one from another, resulting in improved carotenoid bioavailability in human subjects (Salter-Venzon et al., 2017).
Here, we present the release profile of a blend of carotenoid beadlets designed to separate individual carotenoids one from one another (Table 2) when tested using a gut simulation model (TIM). The TIM is an in vitro model that provides a unique ability to sample and measure foods, formulations and compounds throughout the digestion process, providing insight into their predicted digestibility and bioaccessibility within a human intestinal tract.

| TIM model and test conditions
Experiments were conducted at The TIM Company (TIM B.V., The Netherlands). The dynamic, multi-compartmental in vitro system of the stomach and small intestine (TIM-1) has been described in detail in several publications (Bellmann et al., 2014;Domoto et al., 2013;Helbig et al., 2013;Minekus et al., ,1995Minekus et al., , , 2005Van Loo-Bouwman et al., 2014). Briefly, the TIM-1 system consists of a stomach compartment and three small intestinal compartments, the duodenum, jejunum, and ileum. Each compartment is composed of two glass units with a flexible silicone inner wall enclosing the luminal material. The space between the inner and outer walls is filled with water.
Peristaltic mixing of the chyme results from alternate compression and relaxation of the flexible inner wall. The compartments are connected by peristaltic valve pumps that successively open and close, allowing the chyme to transit over time through the compartments.
In the stomach segment hydrochloric acid, α-amylase, pepsin, and lipase are added while in the small intestinal compartments bicarbonate, electrolytes, pancreatic juice, and bile are added as described previously (Bellmann et al., 2016). The experimental test conditions including fluids, meal, pH, and timing are summarized in Table 1.
Prior to the performance of each experiment detailed here, the secretion fluids (e.g., gastric juice with enzymes, electrolytes, bile, and pancreatic juice) were freshly prepared, the pH electrodes calibrated, and semipermeable membrane (hollow fiber) units installed.
All experiments were performed under yellow light with exclusion of daylight.
For simulation of the fed state conditions, a high-fat standard meal was used as recommended by the U.S Food and Drug Administration (FDA) and the Center for Drug Evaluation and Research (CDER) for food-effect bioavailability and fed bioequivalence studies (USDHHS_FDA_CDRE, 2002). This meal contains approximately 50% energy from fat, 20% energy from protein, and 30% energy in the form of carbohydrates. The meal is composed of eggs, bacon, toast bread, potatoes, milk, butter, and margarine. The meal was prepared as one batch, divided into portions of 150 g and stored at <−18°C. One portion of the meal was used for each TIM run.
For experiments with the tablet format, a sinker basket (Japanese_Pharmacopoeia, 2016) was used to allow the tablets to move together through the TIM system with full fluid contact. To mimic the housekeeper wave, the basket with the tablets was manually moved from the stomach compartment to the duodenum compartment where the tablets then disintegrated almost immediately.

| Carotenoid beadlets and tablets
The active ingredient sources contained in the beadlets and tablets are as follows: Natural β-carotene derived from Dunaliella algae and palm fruit, natural α-carotene derived from palm fruit, lutein and zeaxanthin from marigold flowers, lycopene from tomato, and astaxanthin from Haemotococcus algae. These extracts were combined with standard inert ingredients by Omniactive Health Technologies (Morristown, NJ) using proprietary technology to form three of the carotenoid beadlets described in Table 2. The fourth beadlet containing astaxanthin was provided by Beijing Ginkgo Group (Irvine, CA).
Experiments were completed for both a blend of the free beadlets and the beadlets when compressed into tablets, which is a common delivery format. For the blend, four different beadlets were combined, providing 6 different dietary carotenoids in a specified ratio (Table 2). Beadlets containing lycopene, or β-carotene and αcarotene were formulated to delay peak release within the intestinal tract. For the tablet design, 270 mg of the beadlet blend was compressed into 3 tablets, in a background of crystalline cellulose, TA B L E 1 Parameters simulated in the TIM-1 systems describing the average gastrointestinal physiological conditions of healthy young adults after intake of a high-fat meal (HFM)

| Sampling
During each TIM run, timed samples were taken at four points: jejunum filtrates, ileum filtrates, ileum effluent, and residue samples.
The filtrate samples from the jejunum and ileum compartments were
Postgut simulation samples were dissolved in tetrahydrofuran (THF) and saponified with THF-ethanol/potassium hydroxide-hexane overnight. The pH was adjusted to neutral and carotenoids were extracted with hexane. Hexane was evaporated and carotenoids were redissolved in acetonitrile/methanol/dichloromethane (ACN/ MeOH/DCM). Analysis of lutein, zeaxanthin, lycopene, α-carotene, and β-carotene was performed using reversed-phase chromatography with diode array detection, quantifying at 450 nm with DMT (rac-5,7-dimethyltocol) as the internal standard.
A different Triskelion method was used to quantify astaxanthin. injected into an HPLC system using a YMC C30 column for qualitative determination.

| Data calculation
The raw data were used for calculation of the bioaccessible frac- (1) The bioaccessibility was calculated by expressing the total amount of carotenoids recovered from the filtrate (jejunum plus ileum) as a percentage of the intake (Equation 3).
The results of the duplicate TIM-1 runs in the main study were presented as mean ± SD (n = 2).

| Bioaccessibility from the blend of free beadlets
The bioaccessibility (% of intake) of carotenoids released from the blend of free beadlets in jejunum and ileum filtrate combined are shown in Figure 1 (Note: Results from the individual compartments can be found in the Appendix S1).
The maximal bioaccessibility of α-and β-carotene was observed between 240 and 300 min. The accessibility profile shows a steady increase of release of these carotenes from the free beadlets until 300 min, at which point the accessibility decreases for the final hour of the experiment. However, the released carotenes did not approach zero at the termination of the run, suggesting that the total bioaccessibility of α-and β-carotene was not complete by the conclusion of the experimental period (360 min).
The maximal bioaccessibility of lycopene released within the jejunal-ileal compartment was observed at an earlier time point than that seen with α-and β-carotene, at between 120 and 180 min.
Notably, the amount of lycopene was observed in very low levels, making distinction between the measurements difficult to discern.
The maximal bioaccessibility was observed between 120 and 180 min for lutein and zeaxanthin.
Absolute astaxanthin measurements were, in almost all cases, below the detection limit for the analysis (Limit of Quantification=300 μg/kg). Thus, they were not analyzed for bioaccessibility in this report.

| Bioaccessibility from the beadlets compressed into tablets
The bioaccessibility (% of intake) of carotenoids from the tablet experiments in jejunum and ileum filtrate combined are shown in The maximal bioaccessibility of lycopene released within the jejunal-ileal compartment was observed at an earlier time point than that seen with α-and β-carotene, at between 180 and 300 min.
Again, as seen with the free carotenoid beadlets, the amount of lycopene was observed in low levels, making distinction between the measurements difficult to discern and the maximal plateau appears relatively flat.
In contrast, lutein and zeaxanthin showed a distinct peak in bioaccessibility between 180 and 240 min.
(3) Bioaccessibility ( % of intake) = ΣA filtrate (mg) A intake (mg) × 100 % F I G U R E 1 Time course of bioaccessibility profile of the free beadlet blend in the jejunum and ileum compartments combined

| Cumulative bioaccessibility
An overall comparison of cumulative carotenoid bioaccessibility (% intake) between the free beadlet blend and the compressed tablet is shown in Figure 3. The highest cumulative bioaccessibility was seen for lutein and zeaxanthin, with 20.5% of lutein and 18.6% of zeaxanthin recovered from the free beadlet blend, and 14.0% of lutein and 13.6% zeaxanthin recovered from the compressed tablet.
Cumulative bioaccessibility for α-carotene was 7.6% and 7.5% for βcarotene recovered from the free beadlet blend, and 7.0% and 6.9% α-and β-carotene, respectively, recovered from the compressed tablet. Lycopene bioaccessibility was significantly lower than the other carotenoids as only 1.0% and 0.8% of lycopene was recovered from the free beadlet blend or the compressed tablet, respectively.

| Recovery of carotenoids
To calculate the total amount of each carotenoid recovered, the amount recovered from each bioaccessible filtrate (described above) was added to that recovered from the ileum effluent as well as the recovered amounts from each rinse fraction and residue from the gastric, duodenum, jejunum, and ileum compartments. The total values from both the Beadlet Blend and Tablet experiments are provided in Table 3.

| D ISCUSS I ON
It is widely understood that consumption of carotenoid-rich foods is beneficial for human health, yet most people are unable to meet the dietary recommendations through foods alone. For this reason, the addition of supplements containing carotenoids can be useful. As such, when multiple types of carotenoids are delivered into the gut, there can be interactions which inhibit bioavailability (Bohn, 2008). Here, we report the bioaccessibility profiles resulting from a blend of carotenoid beadlets when tested in an in vitro system designed to mimic the upper gastrointestinal tract (TIM). In the TIM-1 system, the tested beadlets, whether as a blend or compressed into a swallowable tablet, showed a bioaccessibility profile that matched the designed release profile (Table 2), with lutein and zeaxanthin released first, followed by lycopene, and α-and β-carotene shortly thereafter. This order aligns with the design of the beadlets, which were designed in a manner to promote a faster release of lutein and zeaxanthin, followed by lycopene, and later by α-and β-carotene.
Similar results have been shown previously from in vitro and clinical evaluations of a layered beadlet of analogous design as the blends tested here (Gellenbeck et al., 2012;Salter-Venzon et al., 2017).
The bioaccessibility release profile of the carotenoids contained in a free beadlet blend ( Figure 1) demonstrated a similar curve as that seen when the beadlets were compressed into a tablet formulation The total carotenoid recovery measurements shown in Table 3 illustrate a different release dynamic between the free beadlet blend recoveries and those measured from the compressed tablet formulation, with a small, but notable, higher recovery from the tablets for all carotenoids, except lycopene. The recovery of α-carotene, β-carotene, lutein, and zeaxanthin were 70%-90% from either formulation, but the tablet formulation illustrated ~5%-10% over that recovered from the free beadlet blend. Combining these results with the higher bioaccessibility seen with the beadlet blend over the tablet format, suggests carotenoid beadlet compression can hinder the release of the carotenoids within the intestine to a small degree.
Lycopene digestion and absorption is a complex topic (Arballo et al., 2021) and here we show recovery of lycopene was lowest for both formulations (~55% of intake). The low recovery resulted in difficulty determining a t max and suggests that the carotenoid may have degraded/oxidized or otherwise changed form during the experiment, so that the assay used to measure the base molecule no longer recognizes this altered form. A high amount of lycopene loss has been reported previously using a dynamic gastrointestinal model, particularly in the jejunum and ileum phases (Berni et al., 2020 Kopec, et. al., demon-strated metmyoglobin served as an oxidant and was able to degrade both lycopene and β-carotene (Kopec et al., 2017). The analytical method originally used in this experiment was not designed to detect lycopene oxidation products, but rather only the parent lycopene. To further investigate the loss of lycopene, reserve samples were assayed at Tufts University (courtesy of Dr. Xiang-Dong Wang).
While some polar metabolites were detected in these assays, it was not possible to determine if they originated from lycopene or other carotenoids. Further experiments designed with lycopene alone and with carotenoid combinations, using detailed analytical methods will be needed to conclusively determine the fate of lycopene and its possible oxidation during simulated gastrointestinal conditions. An understanding of the fatty acids available in the TIM in vitro digestion process model could also provide further insight into the complexity of the lycopene modification and uptake.

| CON CLUS IONS
This study has shown a unique beadlet blend formulation released carotenoids in the TIM dynamic gastrointestinal model with peak bioaccessibility separated over about 3-4 hr in the order of lutein and zeaxanthin first, followed by lycopene, then finally α-and βcarotene; both as a free beadlet blend and after the beadlets have been compressed into tablets. Bioaccessibility measurements were slightly higher for the free beadlet blend than those measured from the tablet formulation. Carotenoid recoveries, with the exception of lycopene, ranged from 7% to 20%, and are measurements similar to those reported previously in literature. Recovery of lycopene was low and requires further experimentation to determine its fate within a simulated digestion model.

ACK N OWLED G M ENTS
The authors would like to thank Dr. Xiang-Dong Wang and Kang-Quan Hu of Tufts University for their efforts in lycopene and lycopene derivative analysis.

E TH I C A L S TATEM ENT
C Hu, D Salter Venzon, and K Gellenbeck are employees of Access Business Group which provided the funding for this work.