Anal. Chem. 1994,66, 4177-4185
Stable Isotope Methods for the Study of /?-Carotene-&Metabolism in Humans Utilizing Tandem Mass Spectrometry and High-Performance Liquid Chromatography Stephen R. Dueker,t A. Daniel Jones,, Gary M. Smith,* and Andrew J. Clifford*lt
Departments of Nutrition and Food Science and Technology and Facility for Advanced Instrumentation, University of California, Davis, California 956 16
This report presents analytical methods for the isolation and quantifcation of u~~-truns-/?-carotene-ds in human plasma following a 73 pmol oral dose. Retinol-& derived from /?-carotene-ds was also determined in the same plasma. Plasma samples drawn over a 24 day period were analyzed. /?-Caroteneand retinol were isolated and purified for analysis using a solid phase extraction protocol with aminopropyl-bonded silica sorbent. Ratios of /?-carotene-d$/?-carotene were determined using reversedphase HPLC with spectrophotometric detection, which fully resolved the isotopomers, and by tandem mass spectrometry (MS/MS) with electron ionization. Results obtained from MS/MS and HPLC analysis showed close agreement and demonstrated improved selectivity relative to analysis using a single mass analyzer. Retinol-& was converted to its tert-butyldimethylsilylether and analyzed by gas chromatography/mass spectrometry usingselected ion monitoring. The ability to resolve the /?-carotene isotopomers by HPLC makes it feasible for investigators without mass spectrometers to investigate the dynamics of absorption and metabolism of /?-carotene-dain humans. p-Carotene is a metabolic precursor of vitamin A and is believed to have other biological functions including antioxidant activity. Intense interest in P-carotene developed after it was proposed that it might inhibit carcinogenesis in humans.' Despite the recognized importance of carotenoids in human physiological functions, many aspects of the absorption and metabolism of P-carotene and other carotenoids in humans are not well understood. Progress here has been hindered by lack of a suitable animal model that mimics human p-carotene metabolic dynamics, scarcity of carotenoids labeled with stable isotopes, and technical difficulties associated with development of analytical methods for monitoring the distribution and fate of carotenoids labeled with stable isotopes. Parker et al. recently reported results from a human study using per-13C-labeledp-carotene and high-precision isotope ratio mass spectrometry. They reported that 64%,21%,and 14%of the absorbed 13C entered the plasma as retinyl ester, retinol, and Department of Nutrition. Department of Food Science and Technology. 9 Facility for Advanced Instrumentation. (1) Peto, R; Doll, R; Buckley, J. D.; Spom, M. B. Nature 1981,290,201-208. +
0003-2700/94/0366-4177$04.50/0 0 1994 American Chemical Society
unaltered karotene, respectively, following a 1.87pmol oral dose.2 Two previous experiments using radioactiveBarotene found that 8.7-52.5% of the radioactivity of an administered dose of radioactive ,&carotene was recovered in l y n ~ p h .Of ~ ~this, ~ 2-30% was associated with p-carotene, and most of the remaining label was associated with retinyl esters. No other human studies using labeled ,&carotene appear in the literature. Carotenoids consist of a diverse group of hydrocarbons including numerous cis-isomers, and many carotenoids and carotenoid oxidation products have been identified in human p l a ~ m a .Studies ~ of the uptake and metabolism of stable isotope labeled /?carotene require that it be isolated from carotenoids and other endogenous substances that could interfere with the analysis. Concentrations of D-carotene in human blood plasma are -0.3 pmol/L. ,&Carotene exhibits a strong characteristic absorbance at 453 nm (in EtOH) which facilitates its determination in biological fluids and tissues using reversed-phase HPLC. Mass spectrometric analysis of ,!?carotene requires more rigorous isolation procedures to remove other lipids, such as fatty acid esters of glycerol and cholesterol, which occur in high concentrations in human plasma. This report describes stable isotope methods for the analysis of /?-carotene& and its metabolite, retinol-d4,in human plasma following a 73 pmol oral dose. The isotopic ratios of ,&carotenedslp-carotene in the isolated plasma Bcarotene are determined using two independent methods: (1) reversed-phase HPLC with spectrophotometric detection and (2) tandem mass spectrometry (MS/MS) with electron ionization and selected reaction monitoring. A solid phase extraction (SPE) protocol for isolating total ,karotene and total retinol from human plasma is also presented in this paper. An application of the extraction and analysis protocols is demonstrated by following the appearance and disappearance of p-carotene-& and retinol-d4 in plasma. The availability of stable isotopelabeled p-carotene and routine analytical methods for measuring it in plasma offer great promise for investigating the absorption and metabolism of P-carotene in humans. Protocols for the chemical synthesis of /3carotene-& have been described recently.'j (2) Parker, R S.;Swanson,J. E.; Mannor, B.; Goodman, K. J.; Spiel", A B.; Brenna, J. T.; Viereck, S. M. Ann. N. E: Acad. Sci. 1994, 691, 86-95. (3) Goodman, D. W.; Blomstrand, R; Wemer, B.; Huang, H. S.; Shiratori, T.]. Clin. Inoestig. 1966, 45, 1615-1623. (4) Blomstrand, R;Werner, B. Scand. /. Clin. Lab. Invest. 1967,19,339-345. (5) Khachik, F.;Beecher, G. R; GoIi, M. B.; Lusby, W. R; Smith, J. C., Jr. Anal. Chem. 1992, 64,2111-2122.
Analytical Chemistry, Vol. 66, No. 23, December 1, 1994 4177
EXPERIMENTAL SECTION Apparatus. Extract-Clean NH2 solid phase extraction cartridges (500 mg, 2.8 mL) and Adsorbosphere NH2 and Adsorbosphere HS ODS HPLC columns (150 mm long x 4.6 mm id., 3 pm particle size) were from Alltech Associates Inc. (Deerfield, IL). A l&port vacuum manifold (Alltech Associates Inc.) was used to regulate the flow rate of solvents through the SPE cartridges at < 5 mL/min. The liquid chromatograph was a Varian Model 5000 (Varian Associates, Palo Alto, CA) connected to a Lee Model 501 UV-visible detector (Lee Scientific, Salt Lake City, UT). 'H NMR spectra of P-carotene-d8 were acquired in C2HC13at 5 "C using a 70" flip angle and a 5.02 s recycle time with a GE Q-300 spectrometer (GE NMR Instruments, Fremont, CA). A 8192-point time domain data block was used, and the data were zero-filled once during processing, which employed a line broadening of 0.1 Hz. Tandem mass spectrometry experiments were performed on a ZAB-2F-HS mass spectrometer (VG Analytical, Wythenshawe, U.K.) with 100 eV electron ionization (El) and an accelerating potential of 8000 V. The source temperature of the mass spectrometer was kept at 250 "C to minimize memory effects due to condensation of p-carotene in the source. The slits of the mass spectrometer were adjusted to provide only slight attenuation of the ion beam, and the resolution ( M / h M ) of the mass spectrometer was determined to be 500 (10%valley definition) at the singlefocus detector and 1000 at the double-focus detector. Retinol-& and retinol were measured using a Hewlett Packard Model 5890 gas chromatograph (Hewlett-Packard, Palo Alto, CA) fitted with a 15 m long x 0.15 mm i.d. capillary column with a 0.15 p m film OB-23, J & W Scientific, Folsom, CA) and directly interfaced to a Trio-2 quadrupole mass spectrometer (VG Masslab, Altrincham, U.K.) operating in the selected ion monitoring mode following a protocol developed in this l a b o r a t ~ r y . ~ , ~ Chemicals: albtruns-~-Carotene10,10',19,19,19,1Y J Y, 1Y-d8 @-carotene-d8) was purchased from Cambridge Isotope Labs (Woburn, MA) where it was synthesized as previously described.6 Unlabeled all-trans-@carotene was obtained from Sigma Chemical Co. (St. Louis, MO). HPLC grade solvents were obtained from Fisher Scientific (St. Louis, MO). lH N M R Analysis ofp-Carotene-ds. @-Carotene-&in C2HCls was analyzed by 'H NMR spectroscopy at 300 MHz and 5 "C using a GE Q-300 spectrometer. Integration was performed on 'H spectra taken at 300 MHz with the integral of the 16,16',17,17' methyl group resonances at 1.03 ppm assigned a value of 12 protons. Unlabeled p-carotene in C2HC13was also analyzed by lH NMR spectroscopy at 500 MHz and 15 "C using a GE Q-500 spectrometer to veriry the assignments reported by Vetter et aL9 A double-quantum-filtered phase-sensitive COSY experimentlo showed that the doublet arising from the 10,lO' protons overlapped with the AB quartet of the 7,7' and 8,8' protons (data not shown). Reagent Purification and Preparation of Standards. Samples containing @-carotenewere handled in amber vials and shielded from light to minimize isomerization and degradation. (6) Bergen, H. R Methods Enzymol. 1992,213, 49-53. (7) Handelman, G. L.; Haskell, M. J.; Jones, A. D.; Clifford, A J. Anal. Chem. 1993,65,2024-2028.
(8) Dueker, S. R; Lunetta. J. M.; Jones, A D.; Clifford, A J. Clin. Chem. 1993, 39,2318-2322. (9) Vetter, W.; Englert, G.; Rigassi, N; Schwieter, U. In Carotenoids; Isler, O., Gutman, H., Solms, U., Eds.; Birkhauser Verlag: Basel, Switzerland, 1971; pp 189-266. (10) Rance, M.; Sorensen, 0. W.; Bodenhausen, G.; Wagner, G.; Emst, R R; Wuthrich, IC Biochim. Biophys. Acta 1983,I 1 7, 479-486.
4178
Analytical Chemistry, Vol. 66, No. 23,December 7, 1994
The fl-carotene-& was -98% chemically pure as purchased as determined by HPLC, but it did contain P-carotene-d7 (16.4%)and P-carotene-& (3.5%) based on FAB/MS data supplied by the vendor. The ,&carotene& was used as purchased without further purification. The unlabeled p-carotene was purified on an ODS phase Adsorbosphere HS column with an isocratic mobile phase of CH3CN:2-propanol:CH3OH (68:20:12 v:v:v) pumped at 0.9 mL/ min. The all-trans-@-carotene peak was collected, concentrated to near dryness under argon, and resuspended in EtOH to a final concentration of 1.8-3.6 pmol/L the extinction coefficient used was 140 642 M-lcm-' at 453 nm. Five p-carotene isotopomer standards were made by adding a k e d amount of B-carotene to varying amounts of P-carotene-&; the molar ratios of pcarotene&/@-carotene in the five carotene isotopomer standards were 0.0426, 0.1278, 0.453, 1.278, and 2.557. Extraction of p-Carotene and Retinol from Plasma. A 1 mL volume of EtOH was added to 1 mL of plasma in a 16 x 125 mm screw-cap culture tube, the mixture was vortexed for 30 s, and lipophilic substances were extracted with 4 mL of hexane. The hexane layer was transferred to a new culture tube and dried under argon. Saponification of the lipid extract was then performed to facilitate the separation of /3-carotene and retinol from lipids with similar physical properties. The residue was suspended in 1.2 mL of EtOH, to which was added 500 pL of 5.3 mol/L KOH in H2O and 100 pL of 5.5 mmol/L pyrogallol in EtOH. The tube was purged with argon and capped, and the sample was heated at 70 "C for 1 h. The tube was allowed to cool to room temperature, 1 mL of H2O was added, and the saponified mixture was extracted with 4 mL of hexane. The hexane layer was transferred to a new tube, concentrated to -100 pL under argon, and applied to an NH2 SPE cartridge previously conditioned with hexane. The sorbent was rinsed with 4 mL of hexane; the first 1.2 mL (of rinse) was discarded, and carotenes that were eluted in the following (remaining) 2.8 mL (of rinse) were collected in a glass vial containing 45 nmol of butylated hydroxytoluene (BHT) and stored at -20 "C until being analyzed for B-carotene-& and p-carotene by HPLC (without further purification) or by MS/MS (after further purification). The NH2 SPE cartridge was then washed with 20 mL of hexane:ethyl acetate (95:5 v:v), and the retinol was eluted with 3 mL of hexane:ethyl acetate (75:25 v:v) into a glass vial containing 45 nmol of BHT and dried under argon. The residue was resuspended in 500 pL of CH&N, transferred to a reaction vial, and concentrated to dryness under argon. The sample was then derivatized and analyzed for retinold4 and retinol as previously described.'~' Since each plasma sample was saponified, the retinold4 and retinol values include both free and esterified retinol. The concentrations of total P-carotene and retinol were determined" in a 100 pL aliquot from the plasma pool before it was processed for analysis of isotopic enrichment. The concentrations of total p-carotene and retinol were also determined in the 2.8 mL hexane rinse and the 3.0 mL hexane:ethyl acetate (75:25 v:v) eluant, respectively, from the NH2 SPE after a 1 mL aliquot of the plasma pool was processed. Recovery was calculated as the amounts of p-carotene and retinol rinsed and eluted, respectively, from the NH2 SPE cartridge expressed as percentages of those present in the plasma prior to processing. R A; Lunetta, J. M.; Corso, F. A; Dueker, S. R; Schneider, P. D.; Joyce, V.; Rippon, M. B.; Wolfe. B. M.; Cliiord, A J. Cancer, submitted.
(11) Zulim,
Analysis of/?-CaroteneIsotopomers by HPLC. The 2.8 mL eluant from the NH2 SPE cartridge containing plasma carotenes was dried under argon, redissolved in 80 pL of CH3CN, and separated on a reversed-phase HPLC system of two columns of Adsorbosphere HS ODS connected in series using an isocratic mobile phase of CH3CN:HzO:ammonium acetate (82:180.01 v:v: w) at 0.9 mL/min. The absorbance at 450 nm was monitored, and peak areas for /?-carotene-&and /?-carotene were integrated using a Lab Calc software package (Chrom 1-AT, Galactic Industries Corporation, Salem, NH). The five B-carotene//?carotene& calibration standards were also analyzed on this HPLC system, and peak areas were integrated to establish a standard /?-carotene isotopomer HPLC calibration curve. Further Purification of Plasma Carotene for MS/MS Analysis. The 2.8 mL eluant from the NH2 SPE cartridge containing plasma carotenes was concentrated to -100 p L under argon and further purified (to remove interfering lipids) by normalphase HPLC on an Adsorbosphere NH2 HPLC column using hexane for 2 min, followed by a 4 min linear gradient to 2-propanol hexane 25:75 (v:v) at 1.2 mWmin. /?-Caroteneeluted between 1.5 and 1.8 min and was collected in an amber vial containiig 45 nmol of BHT. It was dried under argon and rechromatographed on an Adsorbosphere HS ODS column using CH3CN2-propanol: MeOH (68:20:12 v:v:v) at 0.9 mL/min. The all-truns-/?-carotene isotopomers eluted between 13 and 14 min, and this fraction was collected, concentrated to 20 pL under argon, and stored in an amber vial at -20 "C until being analyzed for /?-carotene-&and B-carotene by MS/MS. Analysis of/?-CaroteneIsotopomers by MS/MS. Aliquots (2-4 pL of the further purified plasma extract) were loaded onto the filament of a direct exposure probe which was introduced directly into the mass spectrometer. Fragmentation of both isotopomers was established by performing linked metastable scans at constant B/E to establish the principal fragments (m/z 452 and m/z 444) formed from the molecular ions of B-carotened8 and /?-carotene, respectively. Rapid evaporation of the loaded sample was effected by passing a current of 1A through the probe filament immediately upon introduction to the ion source of the mass spectrometer. These conditions caused the /?-carotene to evaporate rapidly, giving a detected peak with a width of 5 s or less. B-Carotene-& and p-carotene were detected by selected reaction monitoring of the first field-free region metastable transitions, m/z 544.5 452.4 and m/z 536.4 444.4, respectively, via linked switching of magnetic and electric field strengths under control of the VG 11/250 data system. Alternating detection of the two metastable reactions was performed using a sampling time of 0.200 s and a delay of 0.040 s between channels. The probe was cleaned between analyses by inserting the probe tip into the flame generated by a gas burner for 3-5 s. Analysis of blank samples containing no /?-carotene failed to show detectable memory effects. The five /?-carotene isotopomer standards were also analyzed on this MS/MS system, and the integrated areas of the signal for each channel were measured to establish a standard p-carotene isotopomer MS/MS assay curve. Human Plasma Specimens. An informed, consenting, healthy 53-year old male weighing 94 kg swallowed a gelatin capsule (No. 000, Eli Lilly and Co., Indianapolis, IN) containing 73 pmol of ,B-carotene-&dissolved in -2 g of olive oil with a light breakfast (a bran muffin and a cup of coffee). The subject ate lunch and dinner 3 and 8 h later, respectively. Blood samples
-
-
(-25 mL) were drawn into glass tubes containing &EDTA just before and again at 0.5, 1,2, 5, 7, 9, 12, and 24 h and at 2,3,4, 6, 8, 10, 12, 16, 20, 24, 36, 43, 57, 71, 85, 99, and 113 days after ingesting the B-carotene-ds. All blood draws after the first day were taken before breakfast. Extra blood (-50 mL) was drawn on day 16 for the recovery and reproducibility experiments. Plasma was separated by centrifugation and stored as 2 mL aliquots at -70 "C. Total /?-carotene and retinol was measured in all the plasma specimens." Reproducibility of Recovery and of HPLC Analysis of /?-CaroteneIsotopomers. Four aliquots of day 16 plasma were extracted, saponified, and processed through the NH2 SPE cartridges and analyzed for j3carotene-ds and /?-carotene each day for three days to determine the reproducibility of our HPLC method. Four additional aliquots from the same pool were extracted, saponified, and processed through the NH2 SPE cartridges and analyzed for total /?carotene and total retinol11each day for 3 days to determine the recovery of /?carotene and retinol. The amount of total /?-carotene and total retinol recovered from the NH2 SPE cartridges was expressed as a percent of amounts in the plasma aliquots prior to processing in calculating recovery. The reproducibility of our HPLC analysis of /?-carotene-&and /?-carotene in plasma was calculated using a 2-factor ANOVA to determine the within-day and between-day variability. The reproducibility of our sample preparation protocol for isolating Bcarotene and retinol from plasma was also calculated using a 2-factor ANOVA to determine the within-day and between-day variability. Repeated HPLC injections (n = 4) of an isotopomer standard solution containing 18.66 pmol of ,f?carotene-dsplus 597 pmol of /?-carotene (molar ratio of /?-carotene$8//?-carotene = 0.031) were performed to determine the lower limitsof measurement. As little as 1.87 pmol of /?-carotene&was injected onto the HPLC system to determine the lower limits of detection. Comparison of HPLC and MS/MS Methods for Plasma Analysis. Plasma samples from all 10 blood draws during the first 48 h after the subject ingested /?-carotene-& were analyzed for /?-carotene-& and P-carotene using our HPLC and MS/MS methods to provide a direct comparison of the two methods. These data were analyzed by correlation analysis. Calculationof /?-Carotene-d$/?-Caroteneand Retinol-&/ Retinol Ratios in Plasma. Integrated peak area ratios for B-carotene-dd/?carotenefor the five B-carotene isotopomer standards (x-variable) versus the molar ratios of /?
