| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Lipids, Lipoproteins, and Cardiovascular Risk Factors |
1 Centre for Nutrition & Food Safety, School of Biomedical & Molecular Sciences, University of Surrey, Guildford, Surrey GU2 7XH, UK.
2 Department of Biomolecular Sciences, Liverpool John Moores University, Liverpool L3 3AF, UK.
aAuthor for correspondence. Fax 44-1483-576978/300374; e-mail B.Griffin{at}surrey.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: LDL subclasses were separated in 3 h from prestained plasma on a self-forming density gradient of iodixanol. LDL subclass profiles were generated by digital photography and gel-scan software. Plasma samples from 106 normo- and dyslipidemic individuals were used to optimize the gradient for the resolution of LDL heterogeneity. A subgroup of 47 LDL profiles were then compared with LDL subclasses separated by salt DGUC.
Results: The peak density of the predominant LDL band correlated significantly with the relative abundance (as a percentage) of sdLDL as resolved by salt DGUC (P <0.001). As shown previously, LDL isolated at a lighter density in iodixanol compared with salt gradients. A predominance of sdLDL corresponded to a peak density on iodixanol of 1.028 kg/L. This density and the area under the LDL profile lying above this density were sensitive and specific markers for the prediction of a predominance of sdLDL (P <0.001) and showed predictable associations with plasma triglycerides (r = 0.59; P <0.001) and HDL (r = -0.4; P <0.001).
Conclusions: This simple method for the detection of sdLDL can differentiate a predominance of sdLDL, is highly reproducible, and can be used preparatively to isolate sdLDL.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although this phenotypic classification greatly simplified the interpretation of LDL heterogeneity, in nearly all cases its measurement required either extended ultracentrifugation or electrophoresis and was thus unsuitable for routine clinical analysis. These techniques have now been superseded by newer, more rapid procedures, including 3% polyacrylamide gel electrophoresis [Quantimetrix LipoprintTM (7)] and nuclear magnetic resonance (8). A method for the separation of the predominant LDL subclass with HDL and VLDL on a single-step iodixanol gradient has also been described recently (9). The present study reports original data on LDL subclasses separated by iodixanol gradient centrifugation. The procedure was optimized to resolve LDL by direct comparison with salt DGUC. The purity of lipoproteins was checked by agarose gel electrophoresis and the determination of cholesterol in fractions taken across the gradient. Interpretation of LDL subclasses on iodixanol was based on the identification and validation of a diagnostic cutoff that delineated a predominance of sdLDL (pattern B) and on the strength of association between this diagnostic index and other lipid indices.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
materials
OptiprepTM was supplied by Axis-Shield. Cholesterol and TG calibrators, cholesterol, TG assays, and serum quality controls Level I and Level II were supplied by Randox Laboratories Ltd. Hydragel Lipo and lipoprotein(a) [Lp(a)] agarose gel electrophoresis reagents sets were obtained from Analytical Technologies.
procedures
Iodixanol DGUC.
The iodixanol gradient and visualization of separated LDL subclasses were based on a procedure for the separation of the major lipoprotein classes originally described by Graham and coworkers (10)(11), coupled with the prestaining of plasma to detect LDL subclasses as pioneered by Swinkels et al. (12). In addition, the distribution of LDL subclasses was validated against an established method, discontinuous density gradient centrifugation, on a KBr gradient as described by Griffin et al. (5). In the original procedure, lipoproteins were separated on a two-step gradient consisting of a lower layer of plasma (5 mL) adjusted to 120 g/L with respect to iodixanol in a 11.2-mL polycarbonate OptisealTM centrifuge tube and a clear Tris-buffered saline solution (5 mL) adjusted to 60 g/L iodixanol as the upper layer. In the present study, this gradient was adapted by reducing the volume of the lower layer to 3 mL and increasing both the volume and concentration of the upper layer to 7.9 mL and 90 g/L iodixanol, respectively.
To avoid the potentially deleterious effects of staining plasma at low pH (4.5), we prestained the plasma at physiologic pH (7.4). Briefly, we mixed plasma with Optiprep (iodixanol at 60 g/L) and prestained it with Coomassie Blue R 250 [200 µL of 50 g/L in phosphate-buffered saline (PBS)] to provide a final concentration of 120 g/L with respect to iodixanol (final density = 1.067 kg/L). For example, plasma (2.6 mL) was added to Optiprep (0.7 mL), and 200 µL of Coomassie Blue R 250 was added to the mixture to provide a working sample of 3.5 mL.
For the analysis of gradient fractions, a nonstained plasma sample was prepared using 200 µL of PBS instead of stain. For the upper layer, Optiprep was mixed with PBS to provide a 90 g/L iodixanol solution (final density = 1.050 kg/L). Aliquots (7.9 mL) of this solution were dispensed into Beckman Optiseal centrifuge tubes (11.2 mL), and 3 mL of the working sample was carefully underlayered with a syringe and cannula. The centrifuge tubes were housed in a Beckman NVT65 near-vertical rotor and centrifuged at 341 000g(av) and 16 °C for 3 h (at speed) in a Beckman Optima XL-100 ultracentrifuge, with acceleration program 5 and deceleration program 5.
The gradient characteristics and LDL separation in the 8-pocket NVT65 rotor were also compared with results obtained with the 16-pocket NVT65.2. Although the concentration of plasma and stain in the 120 g/L iodixanol (working sample) was identical to that described for the NVT65, because the NVT65.2 uses a smaller, 4.9-mL tube, 1.5 mL of working sample was underlayered beneath 3.4 mL of 90 g/L iodixanol.
Discontinuous DGUC.
To validate the iodixanol method, LDL subclasses were co-isolated by a discontinuous KBr gradient (5). Briefly, plasma (3 mL) was adjusted to a density of 1.090 kg/L by the addition of KBr. The sample was layered over a dense "cushion" of density 1.182 kg/L (0.5 mL) in 13-mL polycarbonate Ultra-Clear centrifuge tubes followed by successive overlayering of the following densities: 1.060 kg/L (1 mL), 1.056 kg/L (1 mL), 1.045 kg/L (1 mL), 1.034 kg/L (2 mL), 1.024 kg/L (2 mL), and 1.019 kg/L (1 mL). Tubes were centrifuged in a Beckman SW40 Ti rotor at 200 000g(av) at 23 °C for 23 h (acceleration program 7 and deceleration without brake) in a Beckman Optima XL-100 ultracentrifuge. The separated LDL bands were converted into LDL subclass profiles by upward displacement of the gradient and continuous monitoring of the gradient effluent at 280 nm, as described previously (5).
Measurement of gradient density and validation of lipoprotein purity.
To determine the density profile of the gradient, we prepared a blank gradient using iodixanol adjusted to 120 g/L with PBS; the blank was centrifuged alongside working plasma samples. After centrifugation, the contents of the blank tube were eluted by upward displacement in a Beckman Fraction Recovery System and a modified Optiseal tube plug, and fractionated into 0.5-mL fractions. The outside of the blank tube was marked with calibrations to denote the fraction number. Nonstained tubes were fractionated in the same way. The refractive index of each 0.5-mL fraction was determined by refractometry and converted to density by use of the following formula: 
= 
a - b, where a = 3.2984, b = 3.3967, = refractive index, and = density (13). The calibrations denoting fraction number on the blank tube were then assigned a density value. A cholesterol profile of the lipoprotein separation was determined by measuring the cholesterol concentration on each gradient fraction. The nature and positions of the principal lipoprotein classes in the gradient (VLDL, LDL, and HDL) and Lp(a) were also confirmed by agarose gel electrophoresis (14).
Generation of LDL subclass profiles by digital photography.
Immediately after centrifugation, Optiseal tubes containing the stained LDL bands were photographed against a vertical light box with a Nikon D1X digital camera set at the highest resolution. To standardize the conditions for photographing the LDL bands, we placed the tubes in a rack at a fixed distance from the camera, which was held in a fixed position relative to the tube-rack by means of a stereotactic clamp. Photographs were downloaded to a personal computer and incorporated into Total-Lab 1D gel-scan software (Pharmacia, UK). The software converted the photographs of the stained LDL bands into LDL profiles with an x axis of distance (mm) against a y axis of pixel intensity. The "gel-scan" software automatically assigned relative electrophoretic migration distance (Rf) values to the primary LDL peak and any secondary peaks. These values were then converted to density by cross-reference to a photograph of the blank tube calibrated in density intervals.
Plasma lipids and interpretation of LDL subclasses.
Whole plasma was analyzed for cholesterol, TGs, and HDL-cholesterol by commercially available assays on a SPACE automated analyzer (Schiapparelli Biosystems, Inc., ENI Diagnostics Division, USA). GGE (2–16%) was carried out as described previously (5) on a few selected samples to illustrate the relationship between LDL subclasses separated by density (iodixanol and salt DGUC) and LDL particle size. In a subgroup of individuals (n = 47), the patterns of LDL heterogeneity produced on iodixanol gradients were interpreted by direct comparison with LDL subclasses defined by an established density gradient procedure (5). This allowed determination of the correspondence between density intervals on the iodixanol gradient with LDL subclasses I-IV, as described previously (4)(5)(6). In addition, to establish a diagnostic cutoff density on iodixanol that corresponded to a predominance of LDL-III (LDL subclass pattern B), we related LDL densities on iodixanol to the relative percentage abundance of sdLDL-III as measured by the comparison method.
analytical performance
Precision.
Larger blood samples (100 mL) were taken from volunteers to examine the variability of LDL separation on the iodixanol gradient by calculating the CV within and between rotors. For within-rotor variability, eight replicate LDL profiles were prepared from four individuals, each with a different LDL phenotype, in four separate runs. For between-rotor variability, eight additional replicate LDL profiles were prepared from two of these individuals in two separate runs. All runs were performed without blank tubes.
Statistics.
The analytical performance of the iodixanol gradient in terms of specificity and sensitivity was assessed relative to its ability to predict a predominance of sdLDL-IIIDGUC (>50%). Correlation and ROC plots were produced to evaluate the diagnostic power of the cutoff density and area under the LDL profile lying above this cutoff density to differentiate a predominance of sdLDL. The statistical significance of the cutoff density on iodixanol was tested by construction of 2 x 2 contingency tables and calculation of 
2. Pearson correlation and product–moment (r2) coefficients were determined to examine the relationships between LDL density and other lipid indices.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Cholesterol and agarose gel profiles showed two major peaks corresponding to HDL (fractions 2–5; density, 1.109–1.056 kg/L) and LDL (fractions 11–17; density, 1.041–1.021 kg/L). Evidence of a shoulder at fractions 18–19 indicated traces of lighter lipoproteins of intermediate density (density, 1.021–1.017 kg/L). The appearance of VLDL at the very top of the gradient was noted in fraction 21 (density <1.009 kg/L). Lp(a) is a heterogeneous particle with a density lying between those of HDL and LDL and a pre-ß migration on agarose gels above that of VLDL. Traces of Lp(a) were detectable in fractions 6–12 (density, 1.056–1.036 kg/L).
|
comparison between nvt65 and nvt65.2 rotors
A comparison between rotors revealed that the larger 16-pocket NVT65.2 rotor produced a density gradient that was very similar in shape and its LDL separation characteristics to that of the 8-pocket NVT65, but in a run time of only 2.5 h. LDL peak densities between these two rotors were highly correlated (P <0.001), with density from the NVT65.2 requiring a linear adjustment of an additional 0.006 kg/L.
ldl heterogeneity on iodixanol
Shown in Fig. 2
are the LDL subclasses separated from the prestained plasma of eight different individuals by iodixanol gradient ultracentrifugation (top panel) and polyacrylamide GGE (bottom panel). The clear variability in LDL banding patterns between and, to a lesser extent, within individuals indicated that the iodixanol gradient resolves structural heterogeneity in LDL that corresponds to that achieved with electrophoresis. Closer examination of the LDL subclass profiles obtained by the two comparison methods and the iodixanol gradient in Fig. 3
revealed many similarities and some subtle differences in patterns of heterogeneity. Profiles of LDL density and particle size as measured by the established methods of salt DGUC and GGE, respectively, have been shown previously to be closely related (5). Although the shapes and peak densities of the LDL profiles on iodixanol were, in general, comparably consistent with the LDL profiles of individuals with LDL subclass patterns A (Fig. 3
, traces PB, RB, and GV) and B (Fig. 3
, traces LW, BE, and PD), the LDL profiles on iodixanol differed from those obtained with the comparison methods by showing a greater proportion of lighter LDL, most notably in individuals with an intermediate (I) phenotype (Fig. 3
, traces JW and RG). From a comparison of the peak LDL densities on iodixanol with the density intervals on salt DGUC, it was possible to calculate LDL density intervals on iodixanol that were equivalent to that defined for LDL subclasses I–IV (Table 1
).
|
|
|
As reported previously (10)(11), LDL densities in iodixanol are lighter than those defined in salt gradients, and the LDL subclasses are distributed over narrower density intervals. This is primarily because high salt concentrations cause water to dissociate from the LDL particle, making LDL more dense in salt. An LDL density on iodixanol of >1.028 kg/L was identified as a critical cutoff point that delineated sdLDL-III from the lighter LDL-II on salt DGUC and GGE (note that LDL-IV is a relatively minor fraction in terms of its relative abundance and expression in healthy individuals).
iodixanol density >1.028 KG/L as a predictor of SDLDL-III (pattern b)
We examined the diagnostic characteristics of a density of >1.028 kg/L in terms of its power to differentiate a predominance of sdLDL (pattern B) by plotting both the distribution of LDL peak density and the area under the LDL profile on iodixanol lying above this density against the percentage abundance of sdLDL and LDL subclass pattern as determined by salt DGUC (Fig. 4
). Both LDL peak density and the area under the LDL profiles lying above 1.028 kg/L in iodixanol correlated significantly (P <0.001) with the relative percentage abundance of LDL by salt DGUC. The distribution of LDL density showed a bimodal clustering of LDL density into predominantly light and dense populations representative of LDL subclass patterns A (LDL-III <40%) and B (LDL-III >50%), respectively, with three individuals showing the intermediate pattern I. The cutoff density of 1.028 kg/L and a relative percentage area under the LDL profile above this density of >50% were both highly significant as predictors of sdLDL-III (2 = 42.6 and 42.9, respectively; P <0.001). An area under the LDL profile of >51% (density >1.028 kg/L) was shown to give 100% specificity and sensitivity in differentiating a predominance of sdLDL-III (pattern B). This was marginally better as a predictor of sdLDL-III than the cutoff density of 1.028 kg/L alone (94% sensitivity; 92% specificity; Fig. 5
).
|
|
reproducibility of ldl separation on iodixanol gradient
Within-rotor variability.
The within-rotor CV for LDL peak density for four different individuals (8 samples/rotor) were 2.1% (density = 1.025 kg/L), 0.6% (density = 1.026 kg/L), 0.3% (density = 1.027 kg/L), and 0.4% (density = 1.034 kg/L).
Between-rotor variability.
There was nonsignificant variation in Rf values for LDL from two individuals (peak densities = 1.027 and 1.034 kg/L) between rotors [2 x 8 samples; mean (SD) Rf = 0.377 (0.001) vs 0.377 (0.003); CV = 0.6%; and 0.511 (0.001) vs 0.516 (0.002); CV = 0.4%]. On conversion of Rf values into density, LDL peak densities were identical for each sample between rotors.
Relationship with other lipids and HDL.
LDL peak density on iodixanol correlated positively with plasma TG [r = 0.59 (r2 = 35%); P <0.001] and inversely with HDL-cholesterol [r = -0.42 (r2 = 18%); P <0.001; Table 2
]. The area under the LDL profile (density >1.028 kg/L) showed equally significant correlations with plasma TG and HDL-cholesterol concentrations (data not shown). When grouped according to LDL subclass pattern as determined by LDL density intervals on iodixanol, LDL subclass patterns consisting of smaller and denser LDL were characterized by increased plasma TGs and decreased HDL-cholesterol but not total plasma cholesterol. These correlations were consistent in sign and order of magnitude to that described with use of the established comparison methods.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Iodixanol, a self-generating gradient medium, revolutionized the separation of serum lipoproteins by centrifugation by simplifying the preparation of density gradients, dramatically reducing run times and in the process increasing sample throughput. The original application of this technique for the separation of the principal serum lipoproteins by Graham et al. (10) has been recently adapted into a single-step method capable of resolving the predominant LDL subclass (9), although in this case the gradient was not developed specifically to resolve LDL subclasses or validated against salt DGUC. The present study builds on this and on the original work by describing a highly reproducible density gradient that has been optimized to resolve LDL subclasses from prestained plasma in either 2.5 or 3 h (NVT65.2 or NVT65 rotor, respectively). Further developments over this earlier work include the analysis of LDL subclass profiles by digital photography and gel-scan software and the characterization and diagnostic evaluation of LDL profiles by direct comparison with a previously established comparison method (5).
The procedure for the prestaining of plasma with Coomassie Blue was modified from that described by Swinkels et al. (12), chiefly to avoid having to decrease the pH of plasma. A low pH increases staining affinity for LDL so that the prestaining of plasma requires less stain. In the iodixanol gradient method, a relatively greater amount of stain was required to achieve a similar degree of staining of LDL but at a physiologic pH. The detection of stained LDL bands by photography effectively removed the need for time-consuming spectroscopic monitoring or compositional analysis of eluted LDL fractions and at the same time eliminated problems associated with lipoprotein recovery. The iodixanol gradient was stable and highly reproducible, with respect to LDL peak density, within and between runs. Although not assessed in this study, the LDL banding patterns were remarkably stable on long standing. Variation in LDL subclass patterns among individuals was comparable to that achieved by the comparison methods, making it possible to visually recognize LDL patterns on iodixanol consisting of either predominantly light, intermediate, or dense LDL. The appearance of secondary LDL peaks and points of inflexion or shoulders in the LDL profile provided clear evidence of single sample heterogeneity. In many cases this was as markedly distinct as that resolved by the comparison method of salt DGUC, although in some cases the patterns between the two methods were dissimilar, with iodixanol appearing to recover lighter, more buoyant LDL.
The high salt concentration and cumulative g force required by other methods can produce adverse effects that can include the dissociation of lipids and apoproteins from the surface of LDL (15). In contrast, iodixanol is a nonionic, isoosmotic medium that requires a relatively low cumulative g force (648 x 105 g · min vs 288 x 106 g · min for salt DGUC) that, in addition to preserving the hydration of LDL, is more likely to maintain the structural integrity and conformation of lipoprotein particles in general. These different conditions, together with the nature of the detection systems (protein staining vs continuous monitoring of absorbance at 280 nm) and the characteristics of the gradient, could explain the lack of concordance in LDL profiles between salt and iodixanol gradients. Moreover, in preserving the chemical composition and functional properties of LDL as well as being physiologically inert, iodixanol gradients could be favored for the preparative isolation of LDL subfractions for use in metabolic studies, although this must be from unstained plasma because stains can alter the physicochemical properties of LDL. The preparative potential of this gradient offers an advantage over tube gel electrophoresis (7) and nuclear magnetic resonance imaging (8).
LDL subclass pattern B has been associated with increased CHD risk in case–control and prospective cohort studies (16)(17)(18). This abnormal LDL pattern or phenotype is determined by the relative proportion of sdLDL, making it essentially a qualitative trait. An absolute increase in the number of sdLDL particles as measured by the concentration of total serum or, preferably LDL apoprotein B (hyperapo B), has also been consistently associated with increased CHD risk (19). Conversely, LDL subclass pattern B can exist in the absence of an increased number of LDL particles (20). These observations highlight the need for a quantitative measure of sdLDL in addition to LDL subclass pattern to provide a more accurate assessment of cardiovascular risk. Electrophoretic methods have attempted to quantify the trait by generating a LDL score based on area under the LDL profile as a function of its Rf (6)(21), or alternatively, by making compositional measurements of total lipoprotein (LDL) mass or the concentration of apoprotein B in eluted LDL fractions (5)(20). In the present study, LDL was characterized on iodixanol by its peak density and the relative percentage area under the LDL profile (density >1.028 kg/L). The latter, as an area corresponding to sdLDL that should be analogous to LDL subclass pattern B as determined by salt DGUC, was only marginally better at differentiating sdLDL than an LDL peak density of 1.028 kg/L. In addition, the absolute pixel volume under the LDL profile was also determined as a semiquantitative estimate of LDL concentration or particle number (data not shown). However, this measure did not provide any additional information for predicting a predominance of sdLDL, thus confirming, at least in this population, the relative superiority of a qualitative description of sdLDL.
The iodixanol gradient was validated by use of plasma samples from a range of healthy males and females participating in dietary trials at the University of Surrey. The resulting distribution of LDL subclasses showed that 36% (38 individuals) had LDL subclass pattern B. This is a higher frequency of expression for this phenotype than would be expected for a mixed population [adult men, 30–35%; postmenopausal women, 15–25% (1)] and could reflect bias in the selection of participants for dietary intervention who have moderately increased plasma TGs and low HDL. At an area under the LDL profile (density >1.028 kg/L) of 51%, there was no misclassification of LDL subclass pattern B. With the notable exception of the lowest HDL-cholesterol occurring in individuals with an intermediate LDL pattern, the relationship of LDL density on iodixanol with plasma TGs and HDL was entirely consistent with that found in an atherogenic lipoprotein phenotype (20).
In summary, this report presents novel methodology for the rapid separation and identification of LDL subclasses by use of iodixanol gradient density centrifugation coupled with digital photography. This method offers the advantages of being simple to perform and avoiding the multiple overlayering associated with traditional salt gradients. It can also be used preparatively for the isolation of LDL subfractions and, with access to the appropriate rotors and equipment, is inexpensive to run. It produces patterns of LDL heterogeneity that correspond to those resolved by current methods and can be easily adapted to multiple-sample rotors, such as the NVT65.2, to give a realistically high throughput for routine clinical applications.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. S. Yee, D. V. Pavitt, T. Tan, S. Venkatesan, I. F. Godsland, W. Richmond, and D. G. Johnston Lipoprotein separation in a novel iodixanol density gradient, for composition, density, and phenotype analysis J. Lipid Res., June 1, 2008; 49(6): 1364 - 1371. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D Griffin, T. A. Sanders, I. G Davies, L. M Morgan, D J. Millward, F. Lewis, S. Slaughter, J. A Cooper, G. J Miller, and B. A Griffin Effects of altering the ratio of dietary n-6 to n-3 fatty acids on insulin sensitivity, lipoprotein size, and postprandial lipemia in men and postmenopausal women aged 45-70 y: the OPTILIP Study Am. J. Clinical Nutrition, December 1, 2006; 84(6): 1290 - 1298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Tricon, G. C Burdge, E. L Jones, J. J Russell, S. El-Khazen, E. Moretti, W. L Hall, A. B Gerry, D. S Leake, R. F Grimble, et al. Effects of dairy products naturally enriched with cis-9,trans-11 conjugated linoleic acid on the blood lipid profile in healthy middle-aged men. Am. J. Clinical Nutrition, April 1, 2006; 83(4): 744 - 753. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. L Hall, K. Vafeiadou, J. Hallund, S. Bugel, M. Reimann, C. Koebnick, H-J F. Zunft, M. Ferrari, F. Branca, T. Dadd, et al. Soy-isoflavone-enriched foods and markers of lipid and glucose metabolism in postmenopausal women: interactions with genotype and equol production Am. J. Clinical Nutrition, March 1, 2006; 83(3): 592 - 600. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. L. Hall, Y. M. Jeanes, and J. K. Lodge Hyperlipidemic Subjects Have Reduced Uptake of Newly Absorbed Vitamin E into Their Plasma Lipoproteins, Erythrocytes, Platelets, and Lymphocytes, as Studied by Deuterium-Labeled {alpha}-Tocopherol Biokinetics J. Nutr., January 1, 2005; 135(1): 58 - 63. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) content of each lipoprotein fraction. IDL, intermediate-density lipoprotein.
