点击显示 收起
1 From the Institute of Clinical Pharmacology, HELIOS Klinikum Wuppertal, University of Witten/Herdecke, Wuppertal, Germany (PAT and UT), and the Department of Human Nutrition and Health, DSM Nutritional Products Ltd, Kaiseraugst, Switzerland (WS, J-CA, and WC)
2 Supported by DSM Nutritional Products and Kemin Foods. 3 Address reprint requests to W Cohn, Department of Human Nutrition and Health, DSM Nutritional Products Ltd, PO Box 3255, CH-4002 Basel, Switzerland. E-mail: willy.cohn{at}dsm.com.
ABSTRACT
Background:Adequate intake of lutein is postulated to reduce the risk of age-related macular degeneration, but kinetic information for developing a dosing regimen is sparse.
Objective:The objective was to characterize lutein plasma kinetics in a multiple dosing design and to assess the effects of lutein intake on concentrations of other plasma carotenoids.
Design:After a run-in period of 7 d, 19 healthy volunteers were assigned to receive daily oral doses of 4.1 mg lutein (n = 8; group 1) or 20.5 mg lutein (n = 8; group 2) for 42 d or no lutein (n = 3; control group). The supplement contained 8.3% zeaxanthin relative to lutein (100%). The time profiles of plasma xanthophyll concentrations were monitored over the dosing phase, and samples were collected frequently on day 42 and for 24 d after dosing.
Results:Average plasma all-E-lutein concentrations increased from 0.14 to 0.52 ± 0.13 and 1.45 ± 0.69 µmol/L in groups 1 and 2, respectively. Dose-normalized lutein bioavailability in group 2 was 60% of that in group 1. Kinetic disposition half-life did not differ significantly between groups. On average, dosing for 18 d was required to reach a >90% fraction of the steady state concentration, which is consistent with an effective half-life for accumulation of 5.6 d. Plasma kinetics of all-E-lutein were paralleled by those of all-E-3-dehydro-lutein. Kinetic analysis indicated formation of all-E-3-dehydro-lutein from lutein. Lutein was well tolerated and did not affect the concentrations of other carotenoids.
Conclusion:Long-term supplementation with 4.1 and 20.5 mg lutein as beadlets increased plasma lutein concentrations 3.5- and 10-fold, respectively.
Key Words: Xanthophylls • carotenoids • lutein • zeaxanthin • all-E-3-dehydro-lutein • multiple oral dose kinetics • macular pigment • age-related macular degeneration
INTRODUCTION
An adequate supply of the dietary carotenoids lutein and zeaxanthin depends on regular intakes of fruit and green and yellow vegetables, which represent major natural sources of xanthophylls (1). The combined daily dietary intake of lutein plus zeaxanthin ranges, on average, between 2 and 26 mg, for which a lutein-to-zeaxanthin ratio of 5:1 is generally assumed (2–4). In the human eye, lutein and zeaxanthin are specifically located in the center of the retina, where they form the "yellow spot," or macula lutea (5). Because of this specific location and their physicochemical properties, such as the absorption of high-energy blue light and their capability to quench reactive oxygen species, a protective action of the xanthophylls in the retina has been postulated (6). Evidence indicates that a low intake of lutein and zeaxanthin is related to an increased risk of age-related macular degeneration (7, 8). Moreover, an increased intake of lutein and zeaxanthin appears to be associated with a lower risk of cataract (9, 10). Although the supplemental intake of lutein, zeaxanthin, or both has been suggested to be useful in persons at high risk of macular degeneration (11–13), critical comment asks for additional prospective trials before general recommendations should be made (14).
The development of appropriate dosing regimens depends on adequate information on the pharmacokinetic properties of a compound. However, little is known about the kinetics of lutein. Several studies have dealt with the comparative availability of lutein in plasma, providing xanthophyll as lutein or as lutein ester in various foodstuffs or in formulated supplements (15–18). Landrum et al (19) treated 2 volunteers with a 30-mg daily dose for 140 d. Plasma concentrations plateaued after 20–40 d, with a 10-fold increase from baseline, and returned to baseline concentrations 40–50 d after supplementation was discontinued. Depletion studies estimate the terminal half-life of lutein to be 15 d in patients with type 1 diabetes (20) and to be 76 d in healthy subjects (21).
We recently reported on the pharmacokinetics of zeaxanthin after multiple dosing in healthy subjects and showed that all-E-3-dehydro-lutein is formed from zeaxanthin (22). The present study aimed to acquire additional kinetic data on lutein in a multiple-dose design study in healthy subjects. The dose was chosen to provide an amount in the range of 2- to 4-fold the average daily intake (4.1 mg/d; low dose) and to facilitate the assessment of pharmacokinetic parameters (20.5 mg/d; high dose). The kinetic data to be generated were plasma concentration dose response at steady state, time to attain steady state, index of accumulation, effective half-life for accumulation, and dose proportionality. Moreover, the plasma kinetics of all-E-3-dehydro-lutein (3R,6R-3-hydroxy-ß,-carotene-3-one) and its possible origin from lutein was investigated, because such a metabolic pathway has been postulated by Khachik et al (23) and the metabolite has also been identified in the retina (23, 24). Because the lutein preparation administered in the present trial contained a small fraction of zeaxanthin, the kinetics of zeaxanthin under the condition of excess lutein was investigated as well and compared with the data from our previous study with zeaxanthin (22). Finally, the potential influence of lutein supplementation on plasma concentrations of other carotenoids was also assessed. The structural formula of all-E-lutein, all-E-zeaxanthin, and all-E-3-dehydro-lutein are presented in Figure 1.
FIGURE 1.. Structural formula and nomenclature of all-E-zeaxanthin, all-E-3-dehydro-lutein, and all-E-lutein.
SUBJECTS AND METHODS
Subjects
Nineteen healthy men and women gave their written informed consent to participate in this monocentric, randomized, open-label, parallel-group study. All subjects were healthy according to the results of a physical examination that included blood pressure and heart rate measurements, an electrocardiogram, and clinical chemistry, hematology, and serology (hepatitis B and C, HIV-1, and HIV-2) tests. Written informed consent was obtained from each subject after they received an adequate explanation of the aims, methods, objectives, and potential hazards of the study. The study protocol was in accordance with the German drug law and the Declaration of Helsinki and was approved by the local ethics committee of the University of Witten/Herdecke, Germany. Demographic data and information on nutritional habits are provided in Table 1.
View this table:
TABLE 1. Demographic data and nutritional habits of the subjects1
Study design
Lutein, extracted from marigold, was formulated in beadlets (5%Lutein, tablet grade; DSM Nutritional Products, Ltd), which were incorporated in hard gelatin capsules containing 4.1 mg lutein (of which 93% was all-E-lutein) and 0.34 mg was zeaxanthin (of which 97% was all-E-zeaxanthin).
Eight subjects (4 women, 4 men) were randomly assigned to receive 4.1 mg lutein/d (one capsule; group 1), 8 (4 women, 4 men) were randomly assigned to receive 20.5 mg lutein/d (5 capsules; group 2), and 3 (2 women, 1 man) were randomly assigned to served as control subjects to estimate baseline fluctuations of plasma lutein and zeaxanthin concentrations. For the entire study, subjects were asked to avoid lutein- and zeaxanthin-rich vegetables and fruit such as kale, Brassica oleracea, spinach, carrots, corn, tomatoes, nectarines, and peaches. Daily fat intake was restricted to 100 g. The subjects received nutritional diaries and were required to complete a 1-d dietary record thrice weekly. The capsules were ingested with 150 mL water and a light breakfast. Compliance was controlled by counting the remaining capsules at each visit.
On day 67 a final physical examination and laboratory tests were performed following the same protocol used at the inclusion visit. After a 1-wk run-in period, blood samples for the measurement of baseline plasma lutein concentrations were drawn on 3 consecutive d and thereafter at weekly intervals until week 5 (day 35). Additional blood samples were taken on days 38, 39, 40, and 41. Except for the blood specimens collected on day 42, all blood specimens were drawn in the morning, before dosing, after the subjects had fasted overnight. On day 42 blood samples were obtained before dosing and 2, 4, 6, 8, 12, and 24 h after dosing (24-h kinetic profile). On this "pharmacokinetic study day," the capsules were taken with a standardized breakfast consisting of one roll of bread with cheese and coffee or tea. In the postdosing period, additional blood samples were drawn in the morning on days 43, 44, 48, 53, 58, 62, and 67. For the control subjects, blood sampling was performed only at the end of the run-in period (3 samples) and on days 28, 48, and 67.
Blood samples of 7.5 mL each were drawn into precooled monovettes (Sarstedt, Nuembrecht, Germany) containing EDTA and immediately centrifuged for 10 min at 4°C and 2500 x g. Plasma was separated under light protection, immediately transferred into polypropylene tubes, and stored at –35°C.
Analytic methods
Plasma samples were analyzed for the E and Z isomers of lutein, all-E-3-dehydro-lutein, and all-E-zeaxanthin and for the sums of Z-zeaxanthin isomers (typically: 85% 13-Z-zeaxanthin, 9-Z-zeaxanthin, and 15-Z-zeaxanthin) and Z-lutein isomers (9-Z-lutein, 9-Z-lutein, 13-Z-lutein, 13-Z-lutein, and 15-Z-lutein isomers) as previously described (22). Recovery rates of lutein and zeaxanthin were 100% and 99%, respectively. Plasma concentrations in the range 0.005–2.8 µmol/L could be measured with intraday CVs of 4.8% (lutein) and 5.6% (zeaxanthin) and interday CVs of 2.6% (lutein) and 6.8% (zeaxanthin), respectively. The lower limit of detection was 0.002 µmol/L; the lower limit of quantification was 0.007 ± 0.002 µmol/L.
ß-Carotene, lycopene, ß-cryptoxanthin, -carotene, and retinol concentrations were measured according to Aebischer et al (25). Cholesterol and triacylglycerol concentrations in plasma were measured according to Richmond (26), Fossati and Prencipe (27), and McGowan et al (28) by using the CHOD-PAP method (Merck AG, Dietikon, Switzerland), adapted to a centrifugal analyzer (Cobas-Bio, Roche Diagnostics, Basel, Switzerland).
Pharmacokinetic analyses
Lutein and zeaxanthin
Baseline plasma xanthophyll concentrations (CB) were calculated individually as the mean value of the 3 predosing samples on days –3 to –1. The approach to plateau was monitored by recording lutein or zeaxanthin trough concentrations over the dosing period, and the corresponding plasma concentration-time profiles were modeled according to an equation of the same form as that describing the kinetics following a constant-rate intravenous infusion (29, 30). The data were fitted according to the following equation:
RESULTS
All subjects completed the study according to the trial protocol, and no adverse events occurred that were related to lutein supplementation. Compliance was excellent (97–100%). Although the subjects in this study were required to avoid the intake of lutein-rich vegetables and fruit, the residual lutein intake was determined by differing eating habits. Thus, average dietary lutein consumption was highly variable, but median values were lower in all groups (Table 1) compared with data from nutritional surveys that showed a population mean value of 1–2 mg lutein/d. This finding indicates that the subjects complied with the dietary restrictions of the present study.
Plasma kinetics of all-E-lutein and all-E-zeaxanthin
Mean concentrations of all-E-lutein increased to plateau concentrations in response to administration of multiple oral doses of 4.1 mg lutein (group 1) or 20.5 mg lutein (group 2) and subsequently decreased after the cessation of dosing on day 42 (Figure 2 A). all-E-Lutein represented the major fraction of lutein throughout the study, contributing to 85% at baseline and during the postdosing period and to 90% during dosing (data not shown). The remaining lutein contributions were determined as the sum of the 9-Z, 9-Z, 13-Z, 13-Z, and 15-Z isomers. Concentration-time profiles for all-E-lutein on day 42 are presented in Figure 2B. Control subjects did not ingest lutein supplements; therefore, baseline plasma all-E-lutein concentrations were maintained (Figure 2A).
FIGURE 2.. Mean (±SD) plasma all-E-lutein concentration-time profiles for group 1 (; 4.1 mg/d, or 7.21 µmol/d), group 2 (; 20.5 mg/d, or 36.04 µmol/d), and the control group (•) over the entire study period (A) and on day 42 (B; last day of supplementation). A: Values immediately before lutein intake (day 0), during the dosing period (days 1–42), and during the postdosing period (days 43–67) are shown.
Attainment of lutein steady state was tested by evaluating the plasma concentration profile of predose samples (ie, blood specimens collected at the end of a dosing interval and just before ingestion of the next dose) over days 38–43 (Figure 3) by linear regression. The resulting mean slopes were 0.010 ± 0.031 and 0.003 ± 0.027 for groups 1 and 2, respectively, ie, statistically not distinct from 0. This finding indicated that steady state had been reached at day 38 or before.
FIGURE 3.. Attainment of apparent steady state all-E-lutein concentrations for representative women (n = 2) and men (n = 2) in group 2. Concentrations in plasma sampled after the subjects fasted overnight and before administration of the daily dose of lutein in the morning are presented. The data for each individual were fitted by linear regression (solid lines). The slopes of individual fits were not significantly different from 0 (P < 0.05).
In addition to lutein, the dosage form contained a small amount of zeaxanthin (8.3% with respect to lutein). Although the kinetic profiles for all-E-zeaxanthin are not presented, the kinetic parameters for both all-E-xanthophylls are summarized in Table 2. Baseline plasma all-E-lutein and all-E-zeaxanthin concentrations were not significantly different between dosing groups. On day 42 peak xanthophyll plasma concentrations were reached 10–11 h after dosing. There was little fluctuation in lutein and zeaxanthin concentration on day 42 (Figure 2B), as evident from PTF values (Table 2). Steady state all-E-lutein concentrations exceeded baseline concentrations by 3.5- and 10-fold for groups 1 and 2, respectively. The increases from baseline to steady state concentrations and the corresponding Cmax were consistently greater in group 2 than in group 1 for both xanthophylls. The two-sided 95% CIs for the ratio of the geometric means (group 1/group 2) for all-E-lutein ranged from 0.19 to 0.45 and from 0.21 to 0.50 for Css and Cmax, respectively. The corresponding 95% CIs for all-E-zeaxanthin were 0.17–0.42 and 0.30–0.46, respectively. The steady state concentrations of the sums the E- and Z-lutein isomers were 0.59 ± 0.14 and 1.64 ± 0.77 µmol/L for groups 1 and 2, respectively.
View this table:
TABLE 2. . Kinetic parameters for all-E-lutein and all-E-zeaxanthin1
Dose proportionality was examined by comparing dose-normalized Cmax and Css values between group 1 and group 2. Dose-normalized Css and Cmax values for all-E lutein were significantly higher in the low-dose group (0.054 ± 0.017 and 0.063 ± 0.027 L–1, respectively) than in the high-dose group (0.036 ± 0.016 and 0.039 ± 0.018 L–1, respectively). The pertaining two-sided 95% CI for the ratio of geometric means (group 1/group 2) ranged from 1.01 to 2.25 and from 1.02 to 2.52 for Css and Cmax, respectively. For all-E-zeaxanthin, dose-normalized incremental, steady state concentrations did not differ significantly between groups 1 and 2 (Css: 0.045 ± 0.020 and 0.034 ± 0.016 L–1, respectively); the 95% CI for the ratio of geometric means (group 1/group 2) ranged from 0.83 to 2.10). Similarly, dose-normalized Cmax values for all-E-zeaxanthin concentrations did not differ significantly between groups 1 and 2 (0.058 ± 0.022 and 0.039 ± 0.018 L–1, respectively). The corresponding 95% CI for the ratio of geometric means (group1/group2) ranged from 0.96 to 2.29.
The cholesterol-normalized concentration-time profiles of lutein resembled those shown in Figure 2A and had similar intersubject variations (data not shown). As described in Subjects and Methods, empirical modeling was used to fit both the time courses of the approach to plateau in response to carotenoid dosing and the decay of plasma concentrations during the post-dosing phase (Figure 4). The monoexponential plus constant model was applied to fit the decay curves of all-E-lutein in both groups and for all-E-zeaxanthin in the high-dose group. Model parameters for all-E-zeaxanthin in the low-dose group could not be accurately estimated because the increase in plasma concentrations above baseline was too small. The precision of all-E-lutein disposition half-life estimates for each subject was acceptable, because the CVs were <20% and 12% for groups 1 and 2, respectively. The half-life of apparent lutein disposition (td) for all-E-lutein and all-E-zeaxanthin ranged between 5 and 7 d (Table 2), and the half-lives for all-E lutein were not significantly different between groups. Again, the precision of the parameters characterizing all-E-lutein plasma accumulation for each subject was acceptable, because the CVs were <16% and 10% for groups 1 and 2, respectively. However, for 2 subjects in group 1 and for 1 subject in Group 2, at least one kinetic parameter resulted in 95% CIs, which included 0, and thus by inference, the parameter was not established. Data for these subjects were not considered for further data compilation in Table 2. The mean half-lives of accumulation (teff) for plasma all-E-lutein were, in a formal statistical sense, different between groups 1 and 2, because the 95% CI for the arithmetic mean differences between groups were –3.53 and –0.73 d, respectively. This finding indicated that 15 (group 1) or 22 (group 2) d were required to reach 90% of steady state concentrations. Accordingly, the calculated accumulation factors for all-E-lutein differed between groups. By contrast, the td values for all-E-lutein were not significant different between groups (95% CI for corresponding mean differences: –2.37, 1.17 d).
FIGURE 4.. Estimates of kinetic parameters for plasma concentrations of all-E-lutein during and after daily doses of lutein for 42 d. The concentration-time profiles are for the same subjects in group 2 as shown in Figure 3. Model predictions were generated as outlined in Subjects and Methods and reflect the approach to plateau (broken lines) and the postdosing period (solid lines).
As shown by paired t tests, half-lives determined from lutein accumulation (teff) were consistently not different from those from lutein plasma decay (td) for group 1 (95% CI: –0.54, 0.38 d) and group 2 (95% CI: –0.57 to 1.54 d) and for the pooled data from both groups (95% CI: –0.32, 0.77 d).
Mean values of the all-E-lutein parameter B, representing the constant of the kinetic disposition model, were 0.158 ± 0.062 and 0.226 ± 0.223 µmol/L for groups 1 and 2, respectively, and consistently exceeded baseline concentrations (Table 2). However, the difference between constant term B and basal concentrations accounted to 8 ± 5% of the incremental dose response for group 2, and these differences were significant (paired t test) on the basis of the corresponding 95% CIs for the geometric mean ratios (baseline concentration/parameter B) for group 1 (0.60, 1.24) and group 2 (0.49, 0.70), respectively.
Baseline lutein concentrations are, to some extent, a predictor of the increase in lutein from baseline to steady state plasma concentrations, as shown in Figure 5. To emphasize the sigmoid behavior, data were fitted according to a Hill equation (data for group 2 only). There was a significant positive correlation between the dose-normalized increments in plasma all-E-lutein concentrations and the dose-normalized increase in plasma all-E-zeaxanthin concentrations (r2 = 0.68, P < 0.05; data not shown).
FIGURE 5.. Baseline-corrected steady state concentrations of all-E-lutein after dosing plotted against predosing baseline lutein concentrations for group 2. Lutein dosing increased plasma lutein concentrations, which were related to predosing concentrations. The solid line was generated by curve fitting according to the Hill equation as an empirical approach.
Plasma kinetics of all-E-3-dehydro-lutein
After lutein dosing, lutein concentrations of all-E-3-dehydro-lutein increased significantly by factors of 1.7 (group 1) and 4.2 (group 2) over mean baseline concentrations, as shown by paired t tests. (Figure 6 A). Kinetic parameters of all-E-3-dehydro-lutein are summarized in Table 3. The plasma concentration time profiles on day 42 were unvarying for both dosing groups (Figure 6B). The dose-normalized Cmax and AUC were not significantly different between groups 1 and 2.
FIGURE 6.. Mean observed and model-predicted plasma all-E-3-dehydro-lutein concentrations over the entire study period (A) and observed plasma all-E-3-dehydro-lutein concentrations on day 42 (B) in group 1 (n = 6; , ) and group 2 (n = 8; , ). Predicted concentrations (, ) were generated by nonlinear regression with the use of linearly interpolated all-E-lutein concentrations as input function. SDs are provided only for the observed concentrations; the SDs of the predicted concentrations were 10% smaller than those of the experimental concentrations.
View this table:
TABLE 3. . Kinetic parameters for all-E-3-dehydro-lutein1
Because the rise in plasma lutein concentrations was paralleled by an increase in all-E-3-dehydro-lutein concentrations, a possible parent-compound metabolite relation was investigated, assuming that the rate of 3-dehydro-lutein formation is proportional to the lutein plasma concentration. This amounted to postulating first-order kinetics (ie, one step or a series of first-order reactions) for the conversion. For 2 subjects (out of 8) in group 1, at least one kinetic parameter resulted in 95% CIs that included 0, and, thus, by inference, the parameter was not established. The parameters for those 2 subjects were not considered for further data compilation in Table 3. The precision of the parameters for individuals in group 2 was adequate; the CV was 7% for parameters kf and ke. Because the increase in 3-dehydro-lutein above baseline was not substantial for group 1, a somewhat lower individual precision was found, which was still adequate for ke (CV: 20%) and kf (CV: 22%). The kinetic parameters kf and ke and the related half-lives did not differ significantly between groups. Further evidence for a joint increase in all-E-lutein and all-E-3-dehydro-lutein concentrations was established by the highly significant correlation between baseline-corrected steady state plasma concentrations of these compounds on day 42 (Figure 7).
FIGURE 7.. Correlation of baseline-corrected steady state all-E-lutein concentrations and baseline-corrected steady state concentrations of all-E-3-dehydro-lutein on day 42. Linear-regression (R2 = 0.96, P = 1.8 x 10–11), derived parameters: intercept (–0.009 µmol/L) and slope (0.086).
Plasma concentrations of other carotenoids and retinol
To establish whether or not there was an effect of xanthophyll dosing on concentrations of other carotenoids and retinol, mean baseline concentrations values and time-averaged (over 24 h) Css concentrations on day 42 were compared. The sum of the E and Z isomers of lycopene, - and ß-carotene, and ß-cryptoxanthin before dosing and at the end of the dosing phase (day 42) are presented in Table 4. Lutein dosing did not affect concentrations of these compounds because there were no significant differences between baseline and steady state concentrations. The baseline concentration of ß-cryptoxanthin in group 1 exceeded that of group 2 (arithmetic difference: 0.181; 95% CI: 0.027, 0.336). There were no further differences between baseline or Css values between groups.
View this table:
TABLE 4. Plasma concentrations of various carotenoids and retinol at baseline and on day 421
DISCUSSION
The findings of the present study complement those of a previous trial of the chemically related zeaxanthin (22), ie, the pharmacokinetics of lutein and zeaxanthin showed many similarities. The inclusion of a small fraction of zeaxanthin into the dosing preparation provided an opportunity to study the kinetics of this carotenoid under conditions of lutein excess. In the absence of supplemented lutein, dose-normalized Css values for all-E-zeaxanthin were 0.086 and 0.050 L–1 for zeaxanthin doses of 1.76 and 17.6 µmol/d (22), which exceeded the dose-normalized concentrations of 0.045 and 0.034 L–1 observed for group 1 (dose = 0.58 µmol) and group 2 (dose = 2.90 µmol) in the present investigation. Therefore, lutein dosing impaired zeaxanthin bioavailability but did not affect plasma concentrations of other carotenoids and retinol (Table 4). Lutein was shown previously to reduce ß-carotene absorption in single-dose studies (35–37), but such an effect was not manifested in the present multiple-dose study.
Evidence for the attainment of apparent steady state was 2-fold: 1) predose samples from days 38 to 43 (Figure 3) indicated that plasma lutein concentrations were maintained, and 2) based on teff values (Table 2), the calculated fraction of steady state was >99% for this interval. The time required to attain a >90% fraction of plasma steady state concentration was 15–22 d for lutein. The corresponding time observed for synthetic zeaxanthin was in the same range, 17 d (22).
Peak steady state concentrations on day 42 were reached, on average, between 9 and 12 h after dosing; this finding agrees with the tmax values observed for other compounds that enter the circulation via the lymphatics incorporated into chylomicrons and eventually get recirculated within hepatic VLDL (38). Mean dose-normalized Cmax all-E-lutein concentrations and AUC values for group 2 were 64% of those of group 1. Because the disposition parameter kd was not significantly different between the 2 lutein-supplemented groups, the observed nonlinearity was not related to dose-dependent disposition kinetics. Therefore, as discussed for other carotenoids, the capacity for intestinal absorption or recirculation within VLDLs may become limiting with increasing dose (39). For all-E-zeaxanthin, which was provided at much lower dosages, the plasma response appeared to be dose proportional, because dose-normalized Cmax and AUC concentrations did not differ significantly. By contrast, at dosages of 1 and 10 mg zeaxanthin (22), nonlinearity was similar to that of lutein in the present trial.
The observation that the increases in plasma lutein concentrations were sigmoidally related to baseline concentrations (Figure 5) confirmed similar findings by Olmedilla et al (40). Several factors, including the disposition parameter kd, plasma cholesterol and triacylglycerol concentrations (as indexes of plasma lutein cotransport by lipoproteins), and sex were tested to account for the observed dependency in the present study. None of these parameters was significantly correlated with baseline concentrations or with the increase in Css (data not shown). It appears that the variability in the response of plasma concentrations may be related to individuality in intestinal absorption or other metabolic events.
The half-life of apparent lutein disposition was comparable with effective half-life (teff), characterizing lutein accumulation during the approach to plateau. Such accumulation is characterized by an effective half-life, which is a weighted average of the absorption half-life and the half-lives describing the disappearance from plasma (41). The effective half-life is always shorter than the terminal half-life. The similarity of td and teff values indicates that td represents an estimate of the effective half-life rather than that of the terminal half-life. This is further corroborated by the finding that the constant term B, obtained from data fitting with the monoexponential plus constant term model, consistently exceeded the baseline lutein concentrations. The monoexponential plus constant term model does not represent the general integrated form of a corresponding compartmental model, and the prediction that plasma concentrations would not return to baseline concentrations could point to a biexponential model with a vanishing exponential term. Thus, a monoexponential plus constant term model may be interpreted as an approximation to the integrated form of a 2-compartment model (2 exponential terms) for the extreme case when the second rate constant approaches zero, ie, the terminal half-life becomes very long and may not be evaluated from data obtained within the selected time window. Such interpretation implies that a very slow plasma disappearance will become indistinguishable from a constant plasma concentration, particularly when the quality of the concentration data are corrupted by measurement errors. In fact, much longer lutein half-lives of 15 d (20) and 76 d (21) were previously published.
Furthermore, the mean td (5.5 d) was very similar to the teff (5.2 d) for zeaxanthin found in our previous study and was distinct from the reported terminal half-life of 12 d (22). Estimation of the terminal half-life of zeaxanthin in that study was critically dependent on the extended sampling interval to up to 76 d. An alternative explanation to account for the preference of the monoexponential plus constant term model might be the lack of adequate control of the dietary restrictions of lutein ingestion during the trial, which could have resulted in lutein intakes that surpassed prestudy intakes. However, this explanation appears to be highly unlikely on the basis of the subjects' dietary records and on the basis of the finding that the constant term B exceeded baseline concentrations for all subjects. Moreover, the plasma lutein concentrations of the control subjects did not increase.
The accumulation index, which relates exposure at steady state to that after the first dose, is determined by the teff. The accumulation index was comparable for all-E-lutein (8.5 on average) and all-E-zeaxanthin ( 8, assuming that the td reflects teff), and there was excellent agreement with the accumulation index previously determined for zeaxanthin (7.5) (Table 2) (22).
A considerable accumulation of plasma 3-dehydro-lutein was observed in response to administration of the lutein preparation. Kinetics of 3-dehydro-lutein resembled that of lutein, and we postulated that 3-dehydro-lutein is formed from lutein. Substantiation of this hypothesis followed the same line of argument as described for formation of 3-dehydro-lutein from zeaxanthin (22). Plasma 3-dehydro-lutein concentrations were coupled to linearly interpolated all-E-lutein concentrations, which served as input function for driving the metabolite subsystem. This approach allowed for approximation of the kinetics of plasma 3-dehydro-lutein, assuming first-order kinetics for formation and elimination of the compound, as described by a one-compartment model. The predicted concentrations were not significantly different from experimental plasma concentrations of 3-dehydro-lutein (Figure 6). Furthermore, the increases in plasma lutein and 3-dehydro-lutein at steady state were found to be proportional (Figure 7). These findings provide strong evidence that the rise in 3-dehydro-lutein above baseline concentrations was closely related to plasma lutein concentrations and was, therefore, a consequence of lutein supplementation.
Because the lutein preparation also contained a small fraction of zeaxanthin, 3-dehydro-lutein formation was probably in part derived from zeaxanthin (22). Therefore, the probable contribution of 3-dehydro-lutein production from zeaxanthin to total metabolite was assessed by referring to the rate constants for formation of 3-dehydro-lutein from zeaxanthin and eventual elimination (22) and by using the average linearly interpolated plasma all-E-zeaxanthin concentrations of group 2 (present study) as input function for driving the conversion to all-E-3-dehydro-lutein. The simulation showed that 18% of the increase in 3-dehydro-lutein concentrations may be related to the added zeaxanthin. The contribution at steady state was calculated from the AUC for all-E-zeaxanthin of the present study and from the slope and intercept of Figure 5 of reference 22. Again, 18% of the all-E-3-dehydro-lutein was estimated to be formed from all-E-zeaxanthin. This confirms that lutein represented the major source of 3-dehydro-lutein formation.
The model parameter characterizing the formation of 3-dehydro-lutein, kf, is associated with a half-life of 46 d, which is much longer than the apparent half-life for lutein disposition (5.5–6 d). This implies that lutein is not exclusively eliminated via formation of 3-dehydro-lutein, but also by other, more efficient, pathways, which may include metabolic reactions as suggested by Khachik et al (23). Furthermore, this is consistent with the hypothesis that lutein is also eliminated by a comparatively slow process (with a half-life of 46 d) as postulated above.
The 3-dehydro-lutein formation rate constant from zeaxanthin (kf = 0.055 d–1) (22) significantly exceeded that from lutein (kf = 0.017 d–1) (Table 3). By contrast, the disposition constants of 3-dehydro-lutein were not significantly different: ke 0.28 d–1 (22) and ke 0.24 d–1 (present study) when zeaxanthin or lutein were supplemented, respectively. Therefore, plasma disposition does not depend on the parent compound of the metabolite. The half-life associated with ke is 2.8 d and, accordingly, the formation of 3-dehydro-lutein was limited by the rate of formation.
In conclusion, the present study showed that plasma lutein concentrations increased 3.5-fold and 10-fold on average, respectively, after the long-term intake of 4.1 and 20.5 mg lutein. Plasma concentrations of all-E-3-dehydro-lutein increased in parallel with those of all-E-lutein, and the increase was clearly related to lutein intake. Except for zeaxanthin, which was also present to a small portion in the preparation administered, other carotenoids remained unaffected by lutein administration.
ACKNOWLEDGMENTS
PAT was responsible for the clinical conduct of the study, was involved in the design of the protocol (principal investigator according to GCP guidelines), and wrote the first draft of the manuscript. J-CA was responsible for the carotenoid and lipid analyses. WS initiated and supervised the project. UT performed the clinical investigations and supervised the dietary instructions (clinical coinvestigator). WC designed the study, was responsible for both the kinetic and statistical data analyses, and prepared the final manuscript. PAT received research grants for the conduct of the clinical part of the study from DSM Nutritional Products Ltd (formerly Roche Vitamins Ltd). J-CA, WS, and WC are employees of DSM Nutritional Products Ltd. UT had no conflict of interest.
REFERENCES