Compartmental and noncompartmental modeling of ¹³C-lycopene absorption, isomerization, and distribution kinetics in healthy adults^,,

Abstract

Background: Lycopene, which is a red carotenoid in tomatoes, has been hypothesized to mediate disease-preventive effects associated with tomato consumption. Lycopene is consumed primarily as the all-trans geometric isomer in foods, whereas human plasma and tissues show greater proportions of cis isomers.

Objective: With the use of compartmental modeling and stable isotope technology, we determined whether endogenous all-trans-to-cis-lycopene isomerization or isomeric-bioavailability differences underlie the greater proportion of lycopene cis isomers in human tissues than in tomato foods.

Design: Healthy men (n = 4) and women (n = 4) consumed ¹³C-lycopene (10.2 mg; 82% all-trans and 18% cis), and plasma was collected over 28 d. Unlabeled and ¹³C-labeled total lycopene and lycopene-isomer plasma concentrations, which were measured with the use of high-performance liquid chromatography–mass spectrometry, were fit to a 7-compartment model.

Results: Subjects absorbed a mean ± SEM of 23% ± 6% of the lycopene. The proportion of plasma cis-¹³C-lycopene isomers increased over time, and all-trans had a shorter half-life than that of cis isomers (5.3 ± 0.3 and 8.8 ± 0.6 d, respectively; P < 0.001) and an earlier time to reach maximal plasma concentration than that of cis isomers (28 ± 7 and 48 ± 9 h, respectively). A compartmental model that allowed for interindividual differences in cis- and all-trans-lycopene bioavailability and endogenous trans-to-cis-lycopene isomerization was predictive of plasma ¹³C and unlabeled cis- and all-trans-lycopene concentrations. Although the bioavailability of cis (24.5% ± 6%) and all-trans (23.2% ± 8%) isomers did not differ, endogenous isomerization (0.97 ± 0.25 μmol/d in the fast-turnover tissue lycopene pool) drove tissue and plasma isomeric profiles.

Conclusion: ¹³C-Lycopene combined with physiologic compartmental modeling provides a strategy for following complex in vivo metabolic processes in humans and reveals that postabsorptive trans-to-cis-lycopene isomerization, and not the differential bioavailability of isomers, drives tissue and plasma enrichment of cis-lycopene. This trial was registered at clinicaltrials.gov as NCT01692340.

INTRODUCTION

The familiar red color of tomatoes is provided by lycopene, which is a hydrocarbon carotenoid, of which American adults consume 4.5–6.5 mg/d (1). Increasingly, carefully controlled laboratory investigations have been characterizing diverse biological impacts of consumed phytochemicals such as lycopene (2–4), which have supported epidemiologic associations between estimated intakes of tomato products or blood lycopene concentrations and health outcomes (5–9). However, there remains much to be learned regarding the bioavailability, tissue distribution, and metabolism of lycopene as we attempt to elucidate the biological impacts in humans. The metabolism of lycopene is of particular interest because lycopene occurs in many geometric isomeric forms in the body (10, 11), a number of cleavage products have been identified in plasma (12, 13), and growing evidence suggests that the bioactivity of different isomers and metabolites may differ (4, 14–18).

The source, formation, and bioactivity of geometric lycopene isomers in blood and tissues are some of the critical uncertainties regarding lycopene metabolism (10, 11). Indeed, 60–96% of lycopene in raw- and processed-tomato food products is present as the all-trans isomer (19), whereas there is generally a lower proportion of all-trans-lycopene in human blood (30–41%) (19, 20) and tissues (21–65%) (10, 11, 21). However, there has been much debate over the timing, mechanism, and location of lycopene isomerization. Kinetic differences have been observed that suggested that cis-lycopene isomers are more bioavailable from foods (22, 23), have a longer plasma half-life (t_1/2)¹¹ than all-trans-lycopene (24), and may be the preferred substrates for the β-carotene-9’,10’-oxygenase enzyme (25). Of particular interest is the finding that, ≤24 h after the consumption of a single dose of ¹⁴C-lycopene consisting of 92% all-trans isomer, the proportion of plasma all-trans-¹⁴C-lycopene decreased to 50% (24), and other experiments have suggested that all-trans-to-cis isomerization may occur in the gastric milieu, enterocyte, or hepatocyte (26–29).

An investigation of the kinetics of isomerization during digestive and distributive processes in humans has proven difficult because of sparse isotope-labeled lycopene-tracer data, which are necessary to follow the metabolic fate of a single oral dose in vivo. Highly enriched ¹³C-lycopene produced from tomato cell–suspension culturing methods (30, 31) provides a source of stable isotope–labeled lycopene for studies in humans. The application of pharmacologic principles with the use of physiologic compartmental kinetic modeling (32) can lead to estimates of lycopene bioavailability, tissue distribution, compartmental masses, and rates of clearance by integrating empirically obtained ¹³C-lycopene plasma concentrations with a priori knowledge of the physiologic processes involved in lycopene assimilation (32–35). In this study, we tested the hypotheses that biolabeled ¹³C-lycopene can be used to define lycopene and isomer kinetic variables in healthy adult men and women and that early lycopene isomer absorption and isomerization variables defined by compartmental modeling explain the enrichment of cis-lycopene in plasma and tissues. We use both noncompartmental and compartmental modeling approaches to analyze plasma ¹³C-lycopene data to define ¹³C-lycopene kinetic variables, thereby showing an informative application of stable isotope technology.

METHODS

Participants

Healthy adult men and women between 21 and 70 y of age, who were not currently taking carotenoid-containing dietary supplements, were recruited by word of mouth and met the following criteria that are indicative of a normal metabolic status and clotting ability during the enrollment screening visit: BMI (kg/m²) between 18 and 27 (inclusive), Eastern Cooperative Oncology Group status of zero, and normal kidney function, liver function, lipids, blood counts, and clotting times as assessed by standard clinical laboratory procedures. On day 28, the complete blood count and liver-function assays were repeated. Subjects were excluded if they had a known allergy or intolerance to tomatoes, a history of nutrient malabsorption or metabolic disorders with special diet recommendations, uncontrolled hyperlipidemia (total cholesterol concentration >200 mg/dL, LDL concentration >160 mg/dL, and serum triglyceride concentrations >200 mg/dL), smoked tobacco products, had a history of endocrine disorders other than diabetes or osteoporosis that required hormone administration, took medications that interfered with dietary fat absorption, or took any complementary or alternative medication that may have interfered with carotenoid absorption or metabolism. The study was conducted in compliance with the ethical standards of and was approved by The Ohio State University Institutional Review Board (2009C0104), and written informed consent was obtained from all subjects. This trial was registered at clinicaltrials.gov as NCT01692340.

Study design and specimen collection

On enrollment, the first 2 subjects (one man and one woman) were instructed to follow a self-selected, controlled lycopene diet (10–20 mg/d) for 2 wk before consuming the ¹³C-lycopene oral dose and to document compliance to this run-in diet with the use of a previously-described daily checklist (36). However, after these 2 subjects had completed the study, and plasma native lycopene and ¹³C-lycopene were measured, the remaining subjects (3 men and 3 women) were instructed to follow a diet with ≤5 mg lycopene/d to enhance the ability to detect low-abundance ¹³C-lycopene metabolic products by decreasing the circulating native lycopene in these 6 subjects. On the day of ¹³C-lycopene dosing (day 0), fasted subjects reported to the Clinical Research Center and provided a baseline blood sample. Subjects consumed the test meal of 10.2 mg ¹³C-lycopene in oil (details below) on an English muffin, which was followed by a low-carotenoid breakfast (either standard or vegetarian) (Supplemental Table 1) that provided 15–15.5 g fat, 20–22 g protein, and 66–68 g carbohydrate. Blood was drawn hourly (21 mL) for 10 h after dosing, and there were fasted blood draws 1, 3, and 28 d after dosing to follow the absorption, distribution, isomerization, and clearance kinetics over what a previous report indicated to be ∼5 plasma lycopene-clearance half-lives (24). Blood samples were collected into evacuated tubes coated with EDTA (BD) under subdued lighting and transported on ice for centrifugation (1000 × g, 10 min, 4°C), plasma was collected under red light, and samples were immediately stored at −80°C. Subjects had free access to water for the rest of the day and had a low-carotenoid snack (11 g fat, 16 g protein, and 43 g carbohydrate) 3 h after dosing and a lunch (20–22 g fat, 10–24 g protein, and 48–59 g carbohydrate) 5 h after dosing. Subjects consumed a standard daily multivitamin (One Daily; CVS) for the duration of the study with the exception of the day of ¹³C-lycopene test-meal consumption. Subjects were asked to discontinue taking any dietary nutrients or botanical supplements and were provided with a daily multivitamin to ensure an adequate nutrient status across subjects. The multivitamin provided ∼100% of the US Food and Drug Administration Daily Value for most micronutrients and 2500 IU vitamin A with 80% from retinyl acetate and 20% from β-carotene, which was 50% of the Daily Value. Participants were instructed to abstain from lycopene-containing foods for the evening meal after the ¹³C-lycopene test dose and to follow a lycopene-controlled diet (which provided 10–20 mg lycopene/d from foods) for the next 4 wk. Participants self-selected lycopene-containing foods to incorporate into their diets and recorded all lycopene consumption for the duration of the study with the use of a previously described lycopene-intake tracking worksheet (36). Any adverse events or unintended effects were evaluated by the study staff with the use of the National Cancer Institute’s Common Terminology Criteria for Adverse Events, version 4.0, and were reported when appropriate to the Ohio State University Institutional Review Board as required.

¹³C-Lycopene dose

Isotopically labeled lycopene was produced and isolated from tomato cell suspension cultures as previously described (30). Biolabeling yielded a distribution of mass isotopologues, and ¹³C-lycopene was of 91% ¹³C atomic purity and 30% uniformly labeled (Figure 1) as determined with the use of HPLC–mass spectrometry (MS) as previously described (30). On the morning of dosing, ∼10.2 mg ¹³C-lycopene [an amount similar to that provided by a 0.5-cup (132-g) serving of tomato-based pasta sauce or a 6-fl oz (177 mL) serving of tomato juice (∼16 mg lycopene in each) (37)] was dissolved in submicron-filtered methylene chloride (7.5 mL) and mixed with 10.2 mL light (carotenoid-free) olive oil (Pompeian Extra Light Olive Oil; Pompeian Inc.) under subdued lighting. The mixture was warmed to 40°C, and methylene chloride was removed under a stream of nitrogen. A lyophilizer was used to briefly expose the sample to reduced pressure, and the removal of methylene chloride was confirmed gravimetrically as previously described (38–40). The ¹³C-lycopene in oil and an olive-oil rinse of the dose container were transferred (final olive-oil mass: 13.65 g; 1 tablespoon) onto an English muffin (Thomas’ Original; Bimbo Bakeries) and served under subdued lighting. Any residual lycopene that was not ingested was collected, quantitated, and accounted for in subsequent kinetic analyses. The delivered ¹³C-lycopene in oil masses was recorded, samples of each dose were retained for analysis, and a portion of each ¹³C-lycopene dose was reserved before mixing with oil for use in standard calibration curves for plasma ¹³C-lycopene analyses. Subjects consumed the dose in <10 min, and breakfast was completed <20 min after dose ingestion. The timing for blood draws began after dose ingestion.

FIGURE 1.

Structure of lycopene (A), representative chromatogram of ¹³C-lycopene monitored at 470 nm (B), and the calculated ¹³C-isotopologue distribution in ¹³C-lycopene used for dosing (C). The percentage of isotopologue abundance was calculated from mass chromatogram peak areas of all-E + 5-Z-lycopene isotopologues masses from an m/z = 535.5 (the mass of unlabeled lycopene + 1 mass unit for H⁺ as a result of the analysis being conducted in positive ion mode) to m/z = 577.5 (the mass of uniformly labeled ¹³C-lycopene + 1 mass unit for H⁺ as a result of the analysis being conducted in positive ion mode). AU, arbitrary units.

¹³C-Lycopene dose mass confirmation and isomeric profile analysis

The mass of ¹³C-lycopene delivered in the dose was confirmed by a HPLC–photodiode array (PDA) analysis with the use of an authentic all-trans-lycopene standard (isolated in-house) for calibration. The system consisted of a Waters 2695 pump system with a Waters 996 PDA detector and a YMC C30 column (3 μm; 4.6 × 250 mm) (Waters Corp.) held at 35°C. The mobile phase consisted of A [60% methanol, 35% methyl tert-butyl ether (MtBE), 5% of 2% aqueous ammonium acetate, and 0.05% triethanolamine)] and B (78% MtBE, 20% methanol, 2% of 2% aqueous ammonium acetate, and 0.05% triethanolamine) applied at 1.3 mL/min beginning with 0% B and was linearly increased to 35% B over 9 min and to 100% B over 6.5 min, after which 100% B was held for 2.5 min, and conditions were returned to 0% B over 3.5 min.

Extraction and HPLC-MS analysis of plasma native and ¹³C-lycopene

Lycopene was extracted from 1 mL plasma by mixing with 1 mL ethanol containing 0.1% (weight:volume) butylated hydroxytoluene and 5 mL hexane:ethanol:acetone:toluene solvent mixture (10:6:7:7; volume:volume:volume:volume) followed by brief probe sonication (Fisher Scientific Model 150I, immersion probe 4C15; Fisher Scientific). After centrifugation (2000 × g for 2 min at ambient temperature), the upper phase was reserved and the lower phase was re-extracted with 5 mL hexane:ethanol:acetone:toluene solvent mixture and centrifuged, and the upper layers were pooled and evaporated under a stream of nitrogen to dryness. Native and ¹³C-lycopene isomers were separated and analyzed by HPLC-MS. Samples were reconstituted in 300 μL MtBE:methanol (1:1) and centrifuged (21,130 × g for 2 min at ambient temperature), the supernatant fluid was transferred to an autosampler vial, and 20 μL was injected. Native and ¹³C-lycopene isomers were separated and analyzed by HPLC-MS with the use of the HPLC system previously described and interfaced with a Q-ToF Premier hybrid mass spectrometer (Micromass UK Ltd.) via an atmospheric pressure chemical ionization probe operated in negative-ion mode. Additional MS variables were as follows: corona current, 30 μA; collision energy, 8 eV; cone voltage, 40 V; source block temperature, 110°C; probe temperature, 600°C; and desolvation gas flow, 400 L/h. All MS experiments were run in Enhanced Duty Cycle mode for increased sensitivity with V-optics enabled (7500 mass resolution). The mobile phases were slightly modified to maximize the MS sensitivity (A: 60% methanol, 35% MtBE, and 5% water; B: 78% MtBE, 20% methanol, and 2% water) and were eluted as previously described. HPLC and MS data were acquired with the use of MassLynx software (v. 4.1; Waters Corp.). Plasma ¹³C-lycopene concentrations were calculated from the peak area of ¹³C-lycopene in plasma at the monoisotopic mass of uniformly labeled lycopene (¹³C₄₀H₅₆ m/z: 576.57) (the ¹³C-lycopene isotopologue distribution is shown in Figure 1) with the use of an external calibration curve of ¹³C-lycopene. Total (native plus tracer) plasma lycopene concentrations were determined by PDA with the use of an unlabeled lycopene-calibration curve. Although native lycopene could be identified and distinguished from ¹³C-lycopene by the accurate mass, because of the saturation of the MS response, native plasma lycopene was calculated to be the difference between total lycopene determined by PDA and ¹³C-lycopene determined by MS. Lycopene-isomer identities were assigned on the basis if the retention time, accurate mass, and characteristic UV-visible spectra (19, 41).

Primary kinetic variable data analysis

The mean time to reach maximal plasma concentration (T_max) and the mean maximal plasma ¹³C-lycopene concentration (C_max) were calculated from the observed values of all 8 subjects because the kinetics did not consistently differ between the first 2 subjects and the last 6 subjects. The plasma ¹³C-lycopene AUC was calculated with the use of the log-linear trapezoidal rule (34).

Physiologic compartmental model of lycopene kinetics

The compartmental modeling approach was conducted in 2 main stages. In the first stage, the data for total lycopene were modeled on the basis of the previously published lycopene model of Diwadkar-Navsariwala et al. (32). In the second stage, a model was developed for the cis and all-trans isomers on the basis if the total lycopene model constructed in the first stage. Compartmental modeling calculations were performed with the use of WinSAAM software (v. 3.0.8; NIH, www.winsaam.com).

Compartmental modeling of total lycopene kinetics

Physiologic compartmental modeling was used to define the kinetics of ¹³C-lycopene over 28 d. In physiologic compartmental modeling, a compartment represents a kinetically homogenous amount of material (34) and can relate to a particular body pool such as lycopene in liver, or it can represent several physiologic spaces such as lycopene in extrahepatic tissues. These compartments are connected by differential equations to represent the transfer of material between compartments, which may consist of several processes occurring in the same, kinetically distinct time frame (34). In this study, the movement of lycopene between compartments was expressed as a fraction per day, and the variables describing that movement are called fractional transfer coefficients (FTCs). FTCs are kinetic constants that describe the fraction of analyte transferred from a donor compartment (J) to a recipient compartment (I) per unit of time and are referred to as L(I,J) in WinSAAM modeling notation. The flow rate of analyte between compartments J and I is calculated as the product of the L(I,J) and the mass of analyte in the donor compartment, which results in the mass transferred per unit time (32). Initial FTCs in the total lycopene compartmental model were based on those reported previously (32) and were adjusted within physiologic constraints on the basis of the carotenoid and lipid metabolism literature until the model-predicted data visually fit the experimentally observed data. At that point, the least-squares iterative fitting process was used to arrive at a final model for each subject, in which the model fractional SDs (FSDs) for each variable were deemed acceptable when <0.5, and the correlation between variables was generally <80%. The bioavailability of lycopene across dietary sources and from the ¹³C-lycopene in oil was determined by the compartmental model by simultaneously analyzing labeled and unlabeled plasma lycopene data. Bioavailability was determined from the compartmental model by solving for the percentage of lycopene entering the gastrointestinal tract (through the gastrointestinal delay compartment preceding the enterocyte compartment) instead of being irreversibly lost from the gastrointestinal tract. Differences in the run-in diet protocol were not deemed to consistently affect the variables; thus, the population kinetic variables that were based on all 8 subjects were estimated with the use of a WinSAAM population analysis.

The initial model was identical to a previously published 7-compartment model describing the absorption, distribution, and clearance of lycopene (32) with exception that lycopene was modeled to be irreversibly lost from the slow-turnover tissue compartment rather than from the plasma lipoproteins (Figure 2). This change was made to more-accurately represent the location from which lycopene is eliminated because carotenoid cleavage enzymes are expressed in tissues (25, 42). This initial model structure fitted the empirical data well, and therefore, the structure was not changed although some variables and model constraints were modified (as later described). Subjects’ average native plasma lycopene masses in the total plasma pool were calculated on the basis of HPLC-PDA analyses and plasma volumes and were included to ensure that the model predictions were conducive to maintaining the observed plasma native lycopene concentrations.

FIGURE 2.

Conceptual compartmental pharmacokinetic model of total lycopene absorption, distribution, and clearance on the basis of Diwadkar-Navsariwala et al. (32). Ovals represent kinetically homogeneous, physiologic compartments. The empirically measured plasma compartment within the plasma compartment is delineated by a dashed rectangle. The rectangle enclosing “Gastric Motility Delay” represents a delay element to account for the interindividual variability in the time between ingestion and plasma appearance. ¹³C-lycopene tracer introduction is denoted by an asterisk, bold numbers inside ovals represent compartment numbers used in the compartmental analysis, italicized lettering next to arrows represents the L(I,J), which is the fractional transfer coefficient, used in the model development to describe the rate of transfer between compartments for the transfer of lycopene from compartment J to compartment I. GI, gastrointestinal.

The masses of lycopene estimated to be in the fast- and slow-turnover tissue compartments at initial conditions were based on those previously published, which were calculated from human biopsy and cadaver tissue lycopene-concentration data from multiple studies and with the use of estimated tissue masses (43). The fast-turnover tissue lycopene compartment represents any extravascular lycopene that is rapidly cleared from chylomicrons and recycled to and from the plasma as lipoproteins and is generally assumed to largely reside in the liver, and thus, the initial pool-size estimate for men and women was calculated as 6.4 and 5 μmol, respectively. The slow-turnover tissue lycopene pool can be interpreted as a kinetically slower compartment that exchanges lycopene with plasma lipoproteins later than the fast turnover pool does and more slowly and is the compartment from which lycopene is irreversibly lost. The slow-turnover compartment is conceptualized to primarily represent adipose tissue but could include any other slow-turnover extravascular pool of lycopene and, thus, was initially estimated to be 21.8 μmol in men and 24.4 μmol in women (43). These compartment masses were used both as initial conditions and to evaluate the ability of the model to predict physiologically plausible tissue pool sizes. Initial estimates for the average daily intake of native lycopene included in the tracee (unlabeled lycopene from the diet) model were determined from lycopene-intake records reported from each subject.

The data used for model fitting included the ¹³C-lycopene in plasma, the native lycopene in plasma, selected FTCs obtained from the model of Diwadkar-Navsariwala et al. (32), and kinetic data related to chylomicron clearance (44). Data related to FTCs obtained from the model of Diwadkar-Navsariwala et al. (32) were necessary because our limited plasma ¹³C-lycopene–concentration data between days 3 and 28 could not accurately characterize tissue-to-plasma lycopene recycling when these tissue-exchange variables were loose as in the initial model. Therefore, the tissue-uptake and -clearance variables were constrained from the initial model reported to ensure physiologically relevant model predictions by incorporating the FTCs and statistical constraints calculated for the subset of men who consumed 10 mg lycopene in the Diwadkar-Navsariwala et al. study [L(6,7): 0.32, FSD: 0.23; L(7,6): 9.24, FSD: 0.49; and L(7,0): 0.0167, FSD: 0.55] (32). The use of these variables from the 10-mg subset group provided narrower initial variable constraints and permitted greater certainty (FSD <0.5 and variable correlation <0.8) for the tissue related kinetic variables. Subject plasma ¹³C-lycopene concentrations were converted to the fraction of ingested dose in the whole plasma compartment by multiplying the concentration by the total estimated plasma volume that was calculated as previously described (32), and native plasma lycopene concentrations were converted to total plasma lycopene masses by multiplying by the estimated plasma volume. Without the inclusion of the data of the average native plasma lycopene mass in the model, the estimated slow-turnover tissue lycopene pool size was unreasonably large. Exact times of blood draws were used. Some adjustments to the initial variables of the 7-compartment model published by Diwadkar-Navsariwala et al. (32) were necessary to improve the predicted fits to the observed data. Initially, the mean ± SEM chylomicron-clearance rate was estimated as t_1/2 = 16.7 ± 9.8 min because this was the variable used by Diwadkar-Navsariwala et al. (32), but the initial variable estimate was adjusted to t_1/2 = 7.1 ± 0.7 [L(4,3): 140.5, FSD: 0.33] on the basis of a report of chyomicron clearance in healthy adults who were fed olive oil (45), which was the oil carrier for the ¹³C-lycopene provided to the current subjects. Other variables were adjusted to improve the physiologically plausible estimates of the absorptive and distributive kinetic variables.

Modeling of lycopene isomers

To study the absorption and interconversion kinetics of cis- and all-trans-lycopene, the model to describe total ¹³C-lycopene kinetics was modified by constructing 2 parallel models for cis- and all-trans-lycopene (Figure 3), each of which were identical in structure to the final model for total lycopene. All FTCs were fixed to be equal to the final FTCs for each subject for the total lycopene compartmental model, with the exception of the bioavailability term and terms to describe all-trans-to-cis isomerization, which was tested in several pools (enterocyte, fast turnover tissue, and slow-turnover tissue lycopene compartments). Adjustments for bioavailability and isomerization (the transfer of all-trans-lycopene to cis-lycopene) were made until the tracer (labeled lycopene) and tracee (unlabeled lycopene) model predictions were in agreement with the empirical data. The percentages of labeled and unlabeled lycopene isomers absorbed were calculated with the use of WinSAAM as previously described for total lycopene.

FIGURE 3.

Final compartmental model describing all-trans and cis-lycopene absorption, distribution, isomerization, and clearance kinetics. Ovals represent kinetically homogeneous, physiologic compartments. The empirically measured plasma values represent the sum of 2 model compartments as delineated by the dashed rectangles. The solid rectangles enclosing “Gastric Motility Delay” represent a delay element to account for interindividual variability in the time between ingestion and plasma appearance. ¹³C-lycopene introduction to the system is denoted by the arrows connecting “cis-lycopene” and “all-trans-lycopene” with the GI tract compartment, bold numbers inside ovals represent the compartment number used in compartmental analysis, and italicized lettering next to arrows represents the L(I,J), which is the fractional transfer coefficient, used in model development to describe the rate of transfer between compartments for the transfer of lycopene from compartment J to compartment I. GI, gastrointestinal.

Initially, the data used for model fitting included only plasma ¹³C-all-trans-lycopene and ¹³C-cis-lycopene concentrations. With the use of only these data, the model could provide good fits of model-predicted plasma concentrations with experimentally measured plasma data when isomerization occurred in different locations. However, this approach revealed that the location of isomerization dramatically affected predicted concentrations of all-trans- and cis-lycopene that were accumulated in body tissues, which led to predicted concentrations in body tissues that were not in agreement with previously published reports of lycopene isomer concentrations in tissues. Therefore, to define plausible kinetic variables and the lycopene isomer distribution and concentrations in the body, more-empirical information was added to the model. Specifically, parallel tracee models of unlabeled cis-lycopene and all-trans-lycopene, which included data for unlabeled cis- and all-trans-lycopene concentrations in each subject’s plasma, were incorporated into the overall model for fitting.

Estimated initial condition masses of all-trans- and cis-lycopene in the fast- and slow-turnover tissue compartments were calculated with the use of the compartment masses previously described for the total lycopene model and then by multiplying by the proportion of cis- and all-trans-lycopene previously reported in the liver (∼55% cis and ∼45% all-trans) and adipose tissue (61% cis and 39% all-trans) (10, 21). Initial estimates for daily intakes of native all-trans- and cis-lycopene included in the tracee model were calculated to be 80% and 20% of lycopene, respectively, on the basis of previous reports of the lycopene isomer contents of foods (19).

As with the total lycopene model, parameter values were adjusted until the model prediction was in good accord with the observed data. Final variable values were determined with the use of a least-squares fitting routine. Model FTCs, flow rates, percentages of absorption, and steady state lycopene isomer masses were calculated as described for the total lycopene model. Differences in run-in diet protocols were evaluated and were not shown to consistently affect the parameters; thus, the population kinetic parameters for all 8 subjects were estimated with the use of a WinSAAM population analysis.

RESULTS

Subjects and controlled-diet compliance

Before ¹³C-lycopene dosing, subjects (Table 1) consumed a controlled lycopene diet for 2 wk (run-in period). The first 2 subjects consumed an average of 12.0 mg lycopene/d, which resulted in a mean fasted native lycopene plasma concentration of 925 nmol/L (713 and 1138 nmol/L) before test-meal consumption on day 0. After preliminary analyses of the first 2 subjects’ plasma samples, the remaining 6 subjects were asked to limit their intakes to 0–5 mg lycopene/d during the run-in–diet period to facilitate the detection of low amounts of lycopene cleavage products during the absorptive phase (0–10 h). These 6 subjects consumed an average of 1.3 ± 0.3 mg lycopene during the run-in period, which resulted in a fasted plasma lycopene concentration of 486 ± 78 nmol/L. However, other than baseline plasma lycopene concentrations, noncompartmental and compartmental kinetic variables for the first 2 subjects compared with the last 6 subjects did not differ, and thus, subsequent variable calculations were performed on the pooled data from all 8 subjects. From days 1–28 after ¹³C-lycopene dosing, all subjects were asked to consume 10–20 mg lycopene/d and reported consuming an average of 11.8 ± 0.7 mg lycopene/d, which resulted in a final native plasma lycopene concentration of 868 ± 57 nmol/L, which was an 80% increase from the average plasma concentration on day 0. Only one subject consumed less than the targeted range with an average reported intake of 9.4 mg lycopene/d. The population mean fasting plasma native-lycopene concentration (n = 8) was 0.60 ± 0.10 μmol/L on day 0 and 0.86 ± 0.06 μmol/L on day 28 (Table 2). Subjects reported no toxicity from the ¹³C-lycopene dose, no adverse symptoms were reported, complete blood count analyses were normal, and liver-function variables were unchanged from enrollment to day 28 of the study.

TABLE 1.

Baseline subject characteristics¹

	Subjects
	Women (n = 4)	Men (n = 4)	Pooled (n = 8)
Age, y	26 ± 2	23 ± 2	24 ± 1
Body mass, kg	55 ± 1	79 ± 2	66 ± 5
BMI, kg/m2	21 ± 1	24 ± 0.2	23 ± 1
Total cholesterol, mg/dL	165 ± 13	156 ± 16	160 ± 10
LDL, mg/dL	88 ± 10	92 ± 11	90 ± 7
HDL, mg/dL	64 ± 8	52 ± 7	58 ± 6
Triglycerides, mg/dL	50 ± 3	67 ± 17	58 ± 9

¹All values are means ± SEMs measured at study enrollment.

TABLE 2.

Noncompartmental, single-dose kinetics after oral ingestion of ¹³C-lycopene in 8 healthy adults¹

Analyte and parameter	Total	All-trans	Total cis isomers	5-cis
13C-Lycopene
Cmax, μM	0.14 ± 0.022	0.10 ± 0.02	0.04 ± 0.01	0.03 ± 0.004
Tmax, h	30 ± 6	28 ± 7	48 ± 9	42 ± 9
AUC (0–672), μmol · h−1 · L−1	31 ± 3	17 ± 3	11 ± 2	7 ± 1
Total AUC, %	–	55	35	23
t1/2, d	6.2 ± 0.3	5.3 ± 0.3	8.8 ± 0.6	9.8 ± 0.8
Native lycopene
Baseline (day 0) plasma concentration, μmol/L	0.60 ± 0.10	0.23 ± 0.05	0.33 ± 0.08	0.20 ± 0.04
Final (day 28) plasma concentration, μmol/L	0.86 ± 0.06	0.35 ± 0.03	0.45 ± 0.07	0.27 ± 0.04

¹C_max, time to reach maximal plasma ¹³C-lycopene concentration; T_max, time to reach maximal plasma concentration; t_1/2, half-life.

²Mean ± SEM (all such values).

Noncompartmental pharmacokinetics

Plasma ¹³C-lycopene was first quantifiable 1–3 h after tracer ingestion and was quantifiable for the remainder of the study. All-trans-, 5-cis-, and other cis-¹³C-lycopene isomers were detected in plasma (Figure 4A). The ¹³C-lycopene test meal increased total circulating lycopene concentrations (native plus tracer) by up to 23% over the native-lycopene concentrations at 24 h (Figure 4B). Plasma ¹³C-lycopene–concentration responses varied widely, with the C_max ranging from 81 to 223 nmol/L; however, when evaluated on the basis of the plasma ¹³C-lycopene mass (i.e., when plasma concentrations were converted to plasma masses), the range of maximal masses (310–585 nmol) deviated less from the mean ± SEM amount (387 ± 53 nmol) (Figure 4C). Differences in all-trans and cis-lycopene kinetics were apparent in the concentration-time curves (Figure 4D, E) in which cis-¹³C-lycopene plasma concentrations plateaued between days 1 and 3, whereas all-trans-¹³C-lycopene concentrations decreased over this time period. The kinetic variables of all-trans- and cis-¹³C-lycopene differed from those of total ¹³C-lycopene, such that the maximal concentration and AUC for all-trans-¹³C-lycopene were greater than for total cis-lycopene, the t_1/2 of all-trans-¹³C-lycopene was shorter, and the all-trans-¹³C-lycopene T_max occurred earlier than that of total cis-¹³C-lycopene (Table 2). Although the proportion of cis-¹³C-lycopene isomers in plasma increased over the 28 d, whether individual cis-¹³C-lycopene isomers in the plasma differed from those in the dose could not be fully discerned from the chromatograms because some cis isomers co-eluted in several peaks during the gradient method used. Nevertheless, it was clear that the proportion of cis isomers relative to all-trans isomers increased over the 28-d period.

FIGURE 4.

Mean ± SEM (n = 8) plasma ¹³C-lycopene and native lycopene concentrations. (A) Extracted ion chromatogram at an m/z of 576.57 showing plasma ¹³C-lycopene isomers. (B) Plasma native lycopene, ¹³C-lycopene, and total (summed native and ¹³C-lycopene) lycopene concentrations over 72 h. (C) Average, highest, and lowest plasma lycopene mass responses over 672 h. ¹³C-lycopene isomer concentrations over 24 (D) and 672 (E) h are shown.

Physiologic-compartmental kinetics of total lycopene

The final model parameters, estimated compartmental lycopene masses, bioavailability, FTCs, and flow rates of total lycopene between compartments are shown in Tables 3–5. Model-predicted fractions of tracer dose in the plasma, fast-turnover, and slow-turnover compartments for the population are shown in Figure 5C–E. A simultaneous evaluation of a parallel model of tracee kinetics showed that the model was predictive of both experimentally observed plasma ¹³C-lycopene and unlabeled lycopene concentrations (Figure 5A–D) along with plausible daily lycopene intakes (14.2 ± 2.6 μmol; 7.6 ± 1.4 mg).

TABLE 3.

Compartmental kinetic variables for total lycopene model¹

From donor compartment (n)	To recipient compartment (n)	Flow rate, μmol/d	Fractional transfer coefficient, d
Gastrointestinal tract (1)	Irreversible loss (0)	11.8 ± 2.7	154.10 ± 12.92
Gastrointestinal tract (1)	Delay compartment (2)	2.46 ± 0.43	45.80 ± 12.92
Delay compartment (2)	Enterocyte (3)	2.46 ± 0.43	22.65 ± 2.78
Enterocyte (3)	Chylomicrons (4)	2.46 ± 0.43	5.51 ± 1.30
Chylomicrons (4)	Fast-turnover tissue pool (5)	2.46 ± 0.43	183.53 ± 29.48
Fast-turnover tissue pool (5)	Lipoproteins (6)	2.85 ± 0.42	1.00 ± 0.25
Lipoproteins (6)	Fast-turnover tissue pool (5)	0.40 ± 0.06	0.225 ± 0.002
Lipoproteins (6)	Slow-turnover tissue pool (7)	6.23 ± 1.12	4.05 ± 0.80
Slow-turnover tissue pool (7)	Lipoproteins (6)	3.77 ± 0.90	0.10 ± 0.04
Slow-turnover pool of tissue lycopene (7)	Irreversible loss (0)	2.46 ± 0.43	0.056 ± 0.007

¹All values are means ± SEMs, n = 8.

TABLE 5.

Additional variables for total lycopene kinetics determined by compartmental modeling ¹

Variable	Value
Gastrointestinal delay (compartment 2), h	1.22 ± 0.19
Bioavailability, lycopene absorbed, %	23 ± 6
Estimated intake of trace, μmol/d	14.2 ± 2.6

¹All values are means ± SEMs. n = 8.

FIGURE 5.

Example of total lycopene compartmental model–predicted fraction of tracer dose in plasma (A) and mass of tracee lycopene in plasma (B) in a representative subject. The solid line represents the model-predicted tracer and tracee amounts, and the solid triangles represent the empirically measured amounts. Also shown are the population averages for the fraction of the ¹³C-lycopene dose in the total plasma compartment (C), fast-turnover (D) and slow-turnover (E) tissue lycopene pools, and the mass of tracee lycopene in the plasma (F) over 28 d after a single oral ¹³C-lycopene dose.

TABLE 4.

Steady state lycopene masses in physiologic compartments determined by compartmental modeling¹

Compartment (n)	Steady state lycopene mass, μmol
Chylomicrons (4)	0.017 ± 0.004
Lipoproteins (6)	1.75 ± 0.27
Fast-turnover tissue lycopene pool (5)	3.67 ± 0.82
Slow-turnover tissue lycopene pool (7)	48.45 ± 11.13

¹All values are means ± SEMs, n = 8.

Physiologic-compartmental kinetics of all-trans- and cis-lycopene isomers

To determine whether cis- and all-trans-lycopene have similar or dissimilar absorption, distribution, and clearance kinetic behaviors, the final model for total lycopene was applied simultaneously to all-trans- and to cis-¹³C-lycopene with the use of 2 parallel models, one model for cis and one model for all-trans. Data inputs were adjusted to be the amount of ingested cis- or all-trans-¹³C-lycopene and included data on plasma cis- or all-trans-¹³C-lycopene concentrations.

For most subjects, the initial model closely predicted both cis- and all-trans-¹³C-lycopene plasma concentrations during the first 24 h after ingestion. However the model-predicted values markedly deviated from the observed 3- and 28-d plasma cis- and all-trans-¹³C-lycopene concentrations, which greatly overestimated plasma all-trans and underestimated cis unlabeled lycopene masses and overestimated the proportion of dietary lycopene present as cis relative to what was known to come from the red-tomato products that subjects reported consuming. Therefore, the inclusion of terms representing the isomerization of lycopene from the all-trans to cis forms was tested in compartments for the enterocytes, the fast-turnover pool, and the slow-turnover pool. First, to test the hypothesis that all-trans-to-cis isomerization occurs at a constant rate in all tissue compartments, isomerization rates were set as equal in the 3 tissue compartments. However, this adjustment led to a great underestimation of plasma all-trans-¹³C-lycopene by day 28, an overestimation of tracer and tracee cis-lycopene in the plasma, an underestimation of cis-lycopene being consumed in the diet, and extremely low proportions of all-trans-lycopene in the slow-turnover lycopene tissue pool. Isomerization in the enterocyte alone led to reasonable predictions of tracee plasma cis-lycopene masses but greatly overestimated tracee all-trans-lycopene plasma masses and plasma tracer cis-lycopene while underestimating dietary cis-lycopene intake. Isomerization in the fast-turnover tissue compartment alone led to model overestimation of plasma unlabeled all-trans-lycopene, and isomerization in both the fast- and slow-turnover pools led to too rapid of depletion of tracer all-trans-lycopene with a simultaneous overestimation of plasma tracee all-trans-lycopene and an underestimation of slow-turnover tissue unlabeled all-trans-lycopene. Isomerization in the slow-turnover tissue compartment alone led to an overestimation of all-trans-lycopene in the fast-turnover pool, an overestimation of cis-lycopene from the diet, an underestimation of native cis-lycopene in the plasma, and an underestimation of all-trans lycopene in the slow turnover pool. However, of these different isomerization locations, isomerization in the fast-turnover pool provided the closest model fit to the observed data. Next, because many reports have suggested that the bioavailability of geometric isomers differs (10, 23, 46), all-trans bioavailability and cis bioavailability were allowed to freely differ, and this adjustment, along with isomerization that occurred in the fast-turnover tissue pool, resulted in reasonable model predictions of the observed tracer and unlabeled tracee data as well as plausible pool sizes and estimated intakes of cis- and all-trans-lycopene. Plots of the model-predicted concentrations of ¹³C-lycopene isomers over 28 d show all-trans lycopene to be the predominant isomer at early time points in plasma and fast-turnover and slow-turnover tissue lycopene compartments, which shifted to cis isomers over time, whereas unlabeled plasma lycopene was predominantly present as cis isomers throughout (Figure 6).

FIGURE 6.

Lycopene isomer compartmental model–predicted fractions of dose in the plasma (A), fast-turnover (B) and slow-turnover (C) tissue lycopene pools, and model-predicted mass of tracee lycopene in plasma (D) over 28 d after a single oral ¹³C-lycopene dose and during controlled dietary lycopene intakes. All-trans lycopene is marked with a dashed line and total cis-lycopene with a solid line. The inset in panel A is an enlargement of the 0–3-d fraction of dose in plasma data.

The final model parameters, estimated compartmental lycopene isomer masses, bioavailability, FTCs, and flow rates of total lycopene between compartments are shown in Tables 6–8. Chylomicron lycopene was estimated to be mostly all-trans, lipoprotein lycopene was estimated to be more-evenly distributed between cis and all-trans forms, and both the fast- and slow- turnover extravascular tissue pools of lycopene were predicted to be composed of primarily cis-lycopene (Table 6). Furthermore, the compartmental model-calculated bioavailability of cis-lycopene was quite similar to that of all-trans (Table 7), and all-trans-lycopene was estimated to be the major form of lycopene consumed daily from the diet (71%). On the basis of the final model, in a lycopene intake steady state, the rate of all-trans-to-cis-lycopene isomerization was estimated to be ∼58% of the fast-turnover all-trans-lycopene pool per day, which equated to 0.97 μmol/d (Table 8). Although differing rates of enzymatic cleavage and nonenzymatic degradation and, therefore, an irreversible loss of cis compared with all-trans isomers may exist, the current data set did not have enough data points during the terminal slope to define these processes.

TABLE 6.

Compartmental kinetic variables for all-trans- and cis-lycopene isomers¹

Compartment	Steady state cis-lycopene mass, μmol (total lycopene, %)	Steady state all-trans-lycopene mass, μmol (total lycopene, %)
Chylomicrons	0.0035 ± 0.0008 (19)	0.015 ± 0.0045 (80)
Lipoproteins	1.12 ± 0.13 (55.5)	0.90 ± 0.17 (44.5)
Fast-turnover tissue lycopene pool	2.72 ± 0.61 (61.9)	1.68 ± 0.36 (38.1)
Slow-turnover tissue lycopene pool	30.19 ± 5.66 (59.2)	20.80 ± 4.61 (40.8)

¹All values are means ± SEMs. n = 8.

TABLE 8.

Rate of lycopene isomerization estimated by compartmental modeling¹

	Flow rate, μmol/d	Fractional transfer coefficient, d
trans-to-cis-Isomerization in the fast-turnover tissue lycopene pool	0.97 ± 0.25	0.58 ± 0.10

¹All values are means ± SEMs, n = 8.

TABLE 7.

Lycopene isomer bioavailability and estimated daily intake determined by compartmental modeling ¹

	cis-Lycopene	All-trans-lycopene
Bioavailability, lycopene absorbed, %	24.5 ± 6.2	23.2 ± 7.6
Estimated daily intake, μmol (total lycopene, %)	3.61 ± 0.92 (28.6)	12.61 ± 2.30 (71.4)

¹All values are means ± SEMs, n = 8.

DISCUSSION

With the use of stable isotope technology and plant cell culture, we produced sufficient ¹³C-lycopene from tomato cell cultures and used a compartmental model of human lycopene kinetics in a novel way, to our knowledge, to investigate lycopene isomer kinetics with the use of simultaneous isotope-tracer and lycopene tracee data. We showed that a model incorporating lycopene isomerization in the fast-turnover compartment along with an individualizable variable for bioavailability of cis- and all-trans-lycopene isomers was predictive of the empirical plasma data and was in accord with reported lycopene isomer tissue distribution. In particular, the inclusion of an endogenous isomerization term could both account for differences in lycopene isomer plasma kinetics and for greater proportions of cis-lycopene in plasma and tissues relative to that in tomato foods.

Overall, the t_1/2 and plasma responses to ¹³C-lycopene were consistent with those previously reported for native and ¹⁴C-lycopene, supporting the stable isotope strategy for detailed studies of lycopene kinetics. The total ¹³C-lycopene t_1/2 of 6.2 ± 0.3 d was similar to the t_1/2 of 5 d previously reported for subjects who were consuming ¹⁴C-lycopene (24), although it was several days longer than that in men after the consumption of lycopene from a tomato-paste beverage (47). The observed ¹³C-lycopene T_max of 30 ± 6 h was more similar to the previously reported T_max of 17–32 h in subjects after the consumption of 10–120 mg lycopene from a tomato-paste–based beverage (47) than that reported in subjects after a ¹⁴C-lycopene microdose (6 h) (24). The C_max of 0.14 ± 0.02 μmol/L in the current study was greater than that in subjects who consumed 5 mg ²H-lycopene in corn oil (n = 2) or 10 mg ²H-lycopene from intrinsically labeled tomatoes (n = 2) (0.006–0.040 μmol/L) (48), possibly because of greater label stability afforded by ¹³C labeling compared with ²H because ²H may be lost from the lycopene molecule by rearrangement during metabolism or sample preparation (49). However, the C_max results were similar to those from individuals after the consumption of 10 and 20 mg native lycopene from a tomato-paste based beverage (0.075 and 0.16 μmol/L, respectively) (47). Differences in the C_max across studies that provided similar amounts of lycopene may have been due to differing amounts of oil consumed with the lycopene as suggested by previous carotenoid-absorption studies (50, 51). Four previous kinetic studies provided 9–10 mg lycopene with 0–29 g oil (24, 32, 47, 48), and the oil amount was positively correlated with the baseline-subtracted lycopene C_max [correlation coefficient: 0.88, P = 0.048 (Pearson correlation)].

The total ¹³C-lycopene compartmental model kinetic parameters agreed with previous lycopene and β-carotene model results (32, 33), which resulted in plausible estimates of lycopene intake and tissue and plasma distributions. The absorption of lycopene in this study (23 ± 6%) was lower than that previously reported for 10 mg lycopene from a tomato-paste drink (34%) (32), which was potentially due to dose isomer-profile differences, greater accuracy resulting from the tracer distinguishability from native lycopene, dose-delivery formulation differences, or the inclusion of the unlabeled lycopene from various food sources in the model. In addition to the model’s ability to predict plasma native lycopene concentrations (Figure 4), the model predictive power was shown by realistic estimates of daily lycopene intake (14.2 ± 2.6 μmol; 7.6 ± 1.4 mg), which were close to subjects’ reported daily lycopene intakes (11.8 mg). Furthermore, predicted fast- and slow-turnover tissue compartment lycopene masses (3.7 and 48.5 μmol, respectively) agreed with published values (43). Thus, the ¹³C-lycopene and native-lycopene plasma results were well characterized with the use of the previously established compartmental model.

The noncompartmental kinetic variables of all-trans-lycopene differed from those of cis-lycopene with all-trans-lycopene having a plasma t_1/2 of 5.5 d, whereas the combined cis-lycopene isomer t_1/2 was 8.8 d. These findings were very similar to the half-lives from a subject after the consumption of ¹⁴C-lycopene (24) and to the half-lives of 5 and 9 d for all-trans and cis-lycopene, respectively, in subjects who consumed a tomato-containing soup or lycopene tablets (52). In addition, the later T_max of cis-lycopene than that of all-trans-lycopene (Table 2) was consistent with findings from men after intake of a tomato-paste–based drink (47).

Previously, several cis-¹⁴C-lycopene isomers absent in the ¹⁴C-lycopene dose given to subjects were shown in the subjects’ plasma, which led the authors to propose that all-trans-lycopene was converted to cis isomers in vivo (24). Only isomerization in the fast-turnover tissue compartment was conducive to adequate predictions of plasma and tissue isomers as well as lycopene isomer intakes in the current study, which may be supported by experimental data suggesting that lycopene is isomerized in hepatic cells (27). Contrary to hypotheses suggested by previous clinical and in vitro studies (26, 47), isomerization in the enterocyte was not a major determinant of plasma isomer concentrations in this study. To reasonably predict plasma isomer concentrations, the simulations with enterocyte isomerization required very low dietary intakes of cis-lycopene and very low cis-lycopene bioavailability, both of which are in contrast to what is currently known about cis-lycopene intake. Isomerization in the larger, slow-turnover tissue pool was unlikely to be a major cause of isomeric profiles in the body because simulations resulted in extremely low predicted amounts of all-trans-lycopene in the slow-turnover pool that was interpreted to be adipose, which has nearly equal proportions of cis- and all-trans-lycopene (21). Whether endogenous all-trans-lycopene isomerization is primarily a thermodynamic process cannot be ascertained currently, but it is possible that the thermodynamic equilibration of all-trans- to cis-lycopene is coincident with the time frame in which lycopene passes through the fast-turnover pool before distribution to other tissues. Although carotenoid cis and trans isomerases present in plants facilitate isomerization in carotenoid biosynthesis (53), to our knowledge, there has not been a report of carotene cis and trans isomerases in animals. However, isomerization of retinal, which is a carotenoid metabolite, is a well-characterized step in the visual cycle (54). Lycopene isomerization may occur in oxidative conditions (55) that are present in the liver, which is a site of oxidative metabolism and the generation of reactive oxygen species. Whether thermodynamic equilibration, isomerases, or oxidative conditions contribute to trans-to-cis isomerization should be investigated.

In addition to isomerization, individual differences in the bioavailability of cis- and all-trans-lycopene were required in the model for good predictions of tracer and tracee cis- and all-trans-lycopene kinetics, intakes, and compartmental lycopene masses. On average the current results, in contrast with those of previous reports, did not indicate that cis-lycopene was more bioavailable than all-trans lycopene, although this finding varied between subjects. Most subjects (n = 5) absorbing a slightly greater percentage of cis-lycopene (16.6 ± 4.2%) than of all-trans lycopene (12.2 ± 2.1%), and fewer subjects (n = 3) absorbed both greater total amounts of lycopene and a somewhat greater percentage of all-trans-lycopene (43.7 ± 13.9%) than of cis-lycopene (36.6 ± 13.7%). Thus, consistent isomer-bioavailability differences did not clearly drive the greater proportion of tissue and plasma cis-lycopene shown in vivo. Whether interindividual lycopene-bioavailability differences result from genetic variation in transport and metabolic genes deserves additional investigation (43).

In conclusion, to our knowledge, this is the largest tracer study in men and women of lycopene kinetics to date, providing superior confidence in the observations and modeling results. Larger studies of variables believed to affect individuals’ kinetic responses to lycopene may improve the specificity of the model. The FTCs, flow rates, and bioavailability in this study apply not only to the ¹³C-lycopene but also to the unlabeled lycopene kinetic data. Therefore, the compartmental model findings may be generalizable across lycopene-containing foods. The lack of samples collected between days 3 and 28 reduced the ability to characterize differences in cis- and all-trans-lycopene tissue plasma recycling and irreversible loss. This study shows that ¹³C-lycopene from plant cell cultures is useful for defining lycopene kinetic variables, and endogenous all-trans-to-cis isomerization is a major source of greater proportions of tissue and plasma cis-lycopene than that in foods.