Original Article

Apolipoprotein B metabolism and the distribution of VLDL and LDL subfractions

Correspondence to: C. J. Packard

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

Apolipoprotein B (apoB) metabolism was investigated in 20 men with plasma triglyceride 0.66;–2.40 mmol/l and plasma cholesterol 3.95;–6.95 mmol/l. Kinetics of VLDL₁ (S_f 60;–400), VLDL₂ (S_f 20;–60), IDL (S_f 12;–20), and LDL (S_f 0;–12) apoB were analyzed using a trideuterated leucine tracer and a multicompartmental model which allowed input into each fraction. VLDL₁ apoB production varied widely (from 5.4 to 26.6 mg/kg/d) as did VLDL₂ apoB production (from 0.18 to 8.4 mg/kg/d) but the two were not correlated. IDL plus LDL apoB direct production accounted for up to half of total apoB production and was inversely related to plasma triglyceride (r = -0.54, P = 0.009). Percent of direct apoB production into the IDL/LDL density range (r = 0.50, P < 0.02) was positively related to the LDL apoB fractional catabolic rate (FCR). Plasma triglyceride in these subjects was determined principally by VLDL₁ and VLDL₂ apoB fractional transfer rates (FTR), i.e., lipolysis. IDL apoB concentration was regulated mainly by the IDL to LDL FTR (r = -0.71, P < 0.0001). LDL apoB concentration correlated with VLDL₂ apoB production (r = 0.48, P = 0.018) and the LDL FCR (r = -0.77, P < 0.001) but not with VLDL₁, IDL, or LDL apoB production. Subjects with predominantly small, dense LDL (pattern B) had lower VLDL₁ and VLDL₂ apoB FTRs, higher VLDL₂ apoB production, and a lower LDL apoB FCR than those with large LDL (pattern A).

Thus, the metabolic conditions that favored appearance of small, dense LDL were diminished lipolysis of VLDL, resulting in a raised plasma triglyceride above the putative threshold of 1.5 mmol/l, and a prolonged residence time for LDL. This latter condition presumably permitted sufficient time for the processes of lipid exchange and lipolysis to generate small LDL particles.—Packard, C. J., T. Demant, J. P. Stewart, D. Bedford, M. J. Caslake, G. Schwertfeger, A. Bedynek, J. Shepherd, and D. Seidel. Apolipoprotein B metabolism and the distribution of VLDL and LDL subfractions. J. Lipid Res. 2000. 41: 305;–317.

Supplementary key words: stable isotopes, multicompartmental modelling, synthesis, catabolism, mass spectrometry, small dense LDL

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

ApoB-100 is the major protein component of very low, intermediate, and low density lipoproteins (VLDL, IDL, LDL). These particles are linked in a delipidation cascade in which triglyceride-rich VLDL, released from the liver, is converted to cholesterol-rich LDL which is in turn catabolized by specific cell-membrane receptors throughout the body. Abnormalities in the metabolism of apolipoprotein B (apoB)-containing lipoproteins are responsible for the generation of hyperlipidemia and the associated increased risk of developing coronary heart disease. Risk has been known for a long time to be strongly related to elevated LDL levels but recently a syndrome, the atherogenic lipoprotein phenotype (ALP), has been identified as a common dyslipidemia in patients with CHD (1) (2) (3). It is characterized by a normal cholesterol concentration with a moderate increase in plasma triglyceride, a decrease in high density lipoprotein (HDL) cholesterol, and a shift in the size profile of LDL towards small dense species. All three features may contribute to risk; moderately raised plasma triglyceride (i.e., VLDL) promotes thrombosis and can affect endothelial function (4), low HDL compromises reverse cholesterol transport, while small dense LDL has properties that enhance its atherogenicity (5) (6). Recently, we described a quantitative model for the development of ALP (7). Based on epidemiological data from others and ourselves (1) (7) (8) (9) it was proposed that the appearance of small dense LDL (a pattern B phenotype on gel electrophoresis (1) (6) or a high LDL-III concentration on density gradient centrifugation (3)) was associated with a plasma triglyceride in excess of 1.5 mmol/l and a hepatic lipase level in the range seen in humans. The aim of the present report was to examine, using stable isotope methods, how apoB metabolism differs in men with a preponderance of small dense LDL in order to gain further insight into the factors that give rise to this atherogenic state. Earlier studies by Teng et al. (10) into the metabolic abnormalities seen in subjects with dense LDL in hyperapobetalipoproteinemia revealed that subjects with this disorder had a reduced clearance rate and higher production of VLDL. Comparison of the kinetics of "heavy" versus "light" LDL in controls and hyperapobetalipoproteinemic subjects demonstrated also that heavy LDL had a slower clearance rate than its lighter counterpart.

In general, VLDL apoB production rates measured by the incorporation of amino acids into apoprotein (11) (12) (13) (14) have agreed well with those obtained using radioiodinated tracers (12) (13) (15). Furthermore, results appear to be independent of the amino acid given, the protocol of administration, and the dose of tracer. There are drawbacks to the method however: in particular, tracer recycling and the low level of incorporation of tracer into slowly turning over lipoproteins such as LDL. We have developed sensitive detection methods that allow quantification of VLDL₁ (S_f 60;–400), VLDL₂ (S_f 20;–60), IDL, and LDL-apoB production and catabolic rates (15) (16). The metabolic behavior of these fractions is again very similar to that observed with radioiodinated tracers. The objectives of the study were i) to evaluate how VLDL₁ and VLDL₂ production and clearance rates varied across a range of lipid levels and so to understand the metabolic factors that regulate plasma triglyceride concentration, ii) to examine the consequences for IDL and LDL apoB metabolism of variation in precursor VLDL kinetics, and iii) to understand further the metabolic environment in which formation of small, dense LDL is favored.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

Subjects were healthy men, laboratory staff and medical students in Glasgow and Munich. They were screened for disorders of thyroid, liver and kidney function and hematological abnormalities. Their characteristics and plasma lipid levels are given in Table 1. Lipid and lipoprotein levels were measured according to the Lipid Research Clinic's protocol using an enzymatic assay for cholesterol and triglyceride (17). Apolipoprotein E phenotyping was performed on plasma essentially as described by Havekes et al. (18). LDL subfraction profile was determined by electrophoretic separation on 2;–16% polyacrylamide gradient gels of d 1.019;–1.063 g/ml LDL isolated from plasma stored at -70°C (19). Previous studies have documented the preservation of the LDL electrophoretic pattern in frozen plasma (20). Subjects were classified as phenotypes A, I, or B following the divisions used by Krauss and Dreon (21). Phenotype A was assigned when the LDL peak diameter on calibrated gradient gels was in excess of 26.0 nm. Phenotype B was assigned when the LDL peak diameter was less than 25.0 nm. Individuals who did not fall into A or B were typed as I (intermediate) in accordance with the revised nomenclature of Krauss and Dreon (21).

Turnover protocol
Two protocols for administration of tracer were used: bolus and primed constant infusion. Subjects fasted from 8.00 PM the previous night and were given L-[5,5,5,^-2H₃]leucine (d₃-leucine, Cambridge Isotope Laboratory, Woburn, MA) starting at 8.00 AM either in the form of a large bolus (6.0 mg d₃-leucine/kg body weight; subjects 1, 2, 3, 5, 7, 11, 12, 14) or a small bolus (0.6 mg d₃-leucine/kg body weight) followed by a continuous infusion (0.6 mg d₃-leucine/kg; subjects 4, 6, 8, 9, 10, 13, 15;–20) for 10 h. During the initial 10-h period they continued to fast but were allowed non-caloric drinks and remained ambulatory. At 6:00 PM they were given a light snack and allowed to go home. Blood samples (10 ml), collected in EDTA, were taken before administration of tracer (0 h) and then frequently (at 0.16, 0.33, 0.5, 0.75, 1, 2, 4, 6, 8, 10, 12, 15, 24 h) over the next 24 h.

Further fasting samples were obtained at approximately 8.00 AM, daily for the next 14 days. A complete description of the procedures for conducting the turnover is given in an earlier publication (15).

Lipoprotein isolation and preparation of apoB
The preparation of VLDL subfractions VLDL₁ (S_f 60;–400) and VLDL₂ (S_f 20;–60), IDL (S_f 12;–20,) and LDL (S_f 0;–12) has been described in a previous publication (22) and is based on the procedure of Lindgren, Jensen, and Hatch (23). In this method, which depends on flotation rate, there is little contamination of LDL with lipoprotein[a] (M. J. Caslake and C. J. Packard, unpublished observations), however, any Lp[a] protein present would co-precipitate with apoB in LDL and potentially contaminate the fraction. To each fraction was added an equal volume of isopropanol at room temperature (24). This led to the precipitation of apoB and lipids while other apoproteins remained in solution. The soluble proteins were removed and the pellicle was delipidated thoroughly with repeated ethanol/ether washes at 4°C until apoB remained as a fine white protein pellet. It was important that delipidation was exhaustive prior to subsequent analysis.

Preparation and analysis of leucine in apoB and plasma
ApoB was hydrolyzed in glass tubes (Labco, High Wycombe, UK) in the presence of 0.5;–1.0 ml 6 N HC1 at 110°C for 24 h and the resultant amino acid mixture was concentrated in a vacuum concentrator centrifuge (Gyrovap, VA Howe, Banbury, UK) and aliquotted into microvials (Chromacol, Herts, UK). After complete removal of HCl, samples were ready for derivatization and mass spectrometric analysis.

Proteins were precipitated from 1 ml plasma by adding 1 ml trichloroacetic acid (10%) and amino acids prepared from the supernatant by cation exchange chromatography using 2 ml columns filled with Dowex AG-50W-X8 resin (H⁺-form, 50–100 mesh; Bio-Rad, Richmond, CA).

The amino acids that bound to the resin were desorbed by 4 M NH₄OH which was subsequently removed by evaporation in a vacuum concentrator (Gyrovap, VA Howe, Banbury, UK), transferred into microvials, and dried again ready for derivatization.

The method used for the analysis of d₃-leucine enrichment in protein hydrolyslates and plasma amino acids is presented in detail elsewhere (15) (16). Ion mass fragments at m/z 277, m/z 276, and m/z 274 were monitored by selective ion recording (SIR mode). The advantages in terms of enhanced sensitivity and precision of monitoring higher mass fragments have been described (15) (16). The m/z 277:m/z 276 ratio showed a linear relationship with isotopic enrichment over the range 0.0;–10.0% atom percent excess (APE). The ratios of m/z 277:m/z 276 were multiplied by an average value for the constant ratio of m/z 276:m/z 274 (determined repeatedly throughout the analytical run, it shows no change over the range 0.0;–10% APE) and the resulting m/z 277: m/z 274 values were used to calculate specific isotopic enrichments and tracer/tracee ratios (15) (25).

Measurement of apoB masses in lipoprotein fractions
Compositions of VLDL₁, VLDL₂, IDL, and LDL were determined by measurements of total and free cholesterol, triglycerides, phospholipids, and total protein (22). The apoB content was calculated as the difference between total and isopropanol-soluble protein (24). ApoB plasma pools were determined for each lipoprotein fraction from apoB concentrations (in mg/ml of plasma) and an estimate of the plasma volume (taken as 4% of the body weight). These were corrected for centrifugal losses (26) and then the leucine content of the apoB pool was calculated from the apoB amino acid composition (27).

Kinetic analysis and multicompartmental modelling
The change in tracer/tracee ratios with time in plasma and in apoB in the four lipoprotein fractions, together with the measured apoB pool size for VLDL₁, VLDL₂, IDL, and LDL, were used as the data set for the derivation of kinetic parameters. A multicompartmental model was constructed using the SAAM II modelling program (SAAM Institute, Seattle, WA), the development of which has been described (15). Briefly, the model ( Fig 1) has the following features: plasma leucine is represented by a compartment (compartment 1) in rapid equilibrium with an intracellular compartment (compartment 2) which is the immediate source of leucine for apoB synthesis. Compartments 3 and 4 are required to account for the uptake and subsequent slow release of leucine by slowly turning over body protein pools. Continuous measurement of plasma leucine tracer/tracee ratios throughout the 14-day turnover permitted modelling of this recycling process and hence its impact on IDL and LDL apoB metabolism could be taken into account when deriving kinetic parameters. Compartment 5 is a delay component which was usually set at 0.5 h but could be adjusted if required. Direct input of apoB occurs into VLDL₁ (compartment 6), VLDL₂ (compartment 9), IDL (compartment 12), and LDL (compartment 14). Compartments 6, 7, 9, 10, 12, and 14 form a delipidation chain, a feature found in most models of apoB metabolism (28) (29) (30) while compartments 8, 11, and 13 represent "remnant populations" of particles which are cleared relatively slowly from each density interval (28) (29) (30). The need for remnant pools is demonstrated in Fig 2. In panel A the lack of a remnant pool as represented by compartment 8 led to an inability to fit the tail of the VLDL₁ apoB curve (i.e., from 10 to 24 h post tracer administration) even though the rise, peak, and initial decay portion of the curve fitted well. A similar situation was seen for VLDL₂ (not shown). The requirement for a delipidation chain within VLDL₂ is seen in panel B. Omission of compartment 10 and use of a single compartment resulted in an unsatisfactory fit. The rise portion of the VLDL₂ apoB curve fitted well but the peak was not "broad" enough to fit the observed data until compartment 10 was included.

View larger version (21K): [in this window] [in a new window]   Figure 1. Multicompartmental model for apoB metabolism. Both bolus and primed constant infusion data were analyzed with the same model. Plasma leucine was represented by a compartment (comp. 1) which received the tracer and distributed it to body protein pools (comps. 3 and 4) and an intracellular compartment (comp. 2) which was the precursor to apoB synthesis. After a delay (comp. 5), tracer appeared throughout the delipidation chain, i.e., in VLDL1 (k6,5), VLDL2 (k9,5), in IDL (k12,5), and in LDL (k14,5). Parameter dependencies were k2,1 = k1,2; k8,6 = k11,9; k0,11 = k0,8; k0,13 = k13,10; k3,4 = 0.1 x k4,3; k14,15 = 2.5 x k15,14. A fuller description of the model is given in reference 15.

View larger version (25K): [in this window] [in a new window]   Figure 2. Characteristics of the multicompartmental model of apoB metabolism. The model in Fig 1 appears complex. It was constructed as detailed in reference 15. The need for multiple mathematical compartments in each of the lipoprotein density ranges is illustrated in panels A;–D (data from subject 12 is shown in panels A and B, data from subject 17 in panel C and from subject 15 in panel D). In panel A, compartment 8 representing VLDL1 ‘remnants’ was omitted from the model. While the rise and initial decay in the tracer/tracee ratio in VLDL1 apoB could be fitted using compartments 6 and 7, the terminal portion of the curve did not fit (as indicated by the dashed line); inclusion of compartment 8 gave a satisfactory fit as shown by the solid line. In panel B a remnant compartment was included in VLDL2 but the second compartment in the short, internal delipidation chain (compartment 10 in Fig 1) was omitted. As shown by the dashed line (note: linear y axis) the reduced model generated a fit to the rise portion of the curve but not to the decay portion (5;–10 hours). The breadth of the peak could only be fitted using the chain of compartments 9 and 10 (solid line). Panel C demonstrates that omission of the second compartment in IDL which represents slowly metabolized particles (compartment 13) gave an unsatisfactory fit (dashed line). Its inclusion allowed the terminal exponential (after 50 h) to be fitted satisfactorily. Panel D shows that omission of an extravascular LDL compartment (compartment 15 in Fig 1) restricted the LDL decay curve to a mono-exponential function (dashed line) while its inclusion gave a satisfactory fit to the observed bi-exponential decay (solid line). Panels E and F show the effect of including a rapid lipolysis compartment (Vt) in VLDL1 rather than having direct VLDL2 secretion from the liver, i.e., omitting the transfer from compartment 5 to compartment 9 in Fig 1. If the fractional transfer rate of Vtcompartment 9 (VLDL2) was permitted to rise to up to 10 times the fractional transfer rate of VLDL1 to VLDL2 (compartment 7compartment 9) no satisfactory fit was obtained (dashed line). Application of direct VLDL2 production as in the Fig 1 model did give a good fit to observed data (solid line) (data from subject 16).

It was this need to generate a broad peak in VLDL that led to the formulation of the delipidation chain in earlier studies (28) (29) (30). In nearly all the normal subjects reported here, a delipidation chain was optional for VLDL₁, but when the current model is applied to patients with elevated plasma triglyceride levels, the same phenomenon is seen for VLDL₁ as for VLDL₂ in panel B (data not shown). For consistency and to allow comparison of results, we used a short delipidation chain in VLDL₁ and VLDL₂ in all types of subjects. IDL apoB, which is largely restricted to the plasma space, is modelled as two intravascular compartments, one of which is delipidated to LDL (compartment 12). In panel C, it can be seen that if a single IDL compartment is included in the model, the rise, peak, and initial down slope can be accounted for satisfactorily but again the model cannot fit the terminal exponential (seen after 50 h) until a further compartment (compartment 13) with a slower decay rate is added. LDL apoB, on the other hand, is represented by a single plasma compartment and the bi-exponential nature of the d₃-leucine decay curve is accounted for by intra-extravascular exchange. The effect of omitting the extravascular compartment 15 can be seen in Fig 2 panel D. With only compartment 14 present, LDL decay is monoexponential whereas there are clearly two exponential decays present in the observed data. A satisfactory fit is obtained with the inclusion of an extravascular LDL compartment. In other studies using radioiodinated lipoprotein tracers, it has been possible to resolve the LDL apoB decay further with the inclusion of two plasma compartments, one with a more rapid catabolic rate than the other. This required either injection of dual tracers (30) or collection of excreted radioactivity (7). Administration of an amino acid labels all body proteins and it is not possible to distinguish the metabolic fate of individual LDL sub-species within a density interval.

In many subjects, apoB input was mainly into the VLDL₁ density interval. However, some had an early rapid rise in tracer/tracee ratio in VLDL₂, IDL, or LDL apoB. In the model presented in Fig 1, this latter phenomenon is accounted for by permitting direct input from the delay compartment, 5, into compartments 9, 12, and 14. This gives a satisfactory fit in all subjects.

An alternative model, examined in our earlier evaluation of the methodology, had a rapidly turning over pool V_T (see panel F, Fig 2) in VLDL₁ which had outputs into VLDL₂, IDL, and LDL. This arrangement is favored by some investigators (29) who prefer the concept of rapid lipolysis to that of secretion of smaller lipoproteins from the liver. In the 6 subjects studied in reference 15, the majority (about 70%) of apoB appeared in VLDL₁ and there was limited need for input of apoB into VLDL₂, IDL, or LDL. However, the present series includes subjects in whom VLDL₂, IDL, and LDL account for over half of total apoB input. To illustrate this, data are presented for subject 16 in panel E (Fig 2). Using the model in Fig 1, a satisfactory fit was obtained with approximately equal input of apoB into VLDL₁ and VLDL₂ (see Table A1). In the alternate model, no input to VLDL₂ was allowed (i.e., no transfer 59) but a rapidly turning over compartment was included in VLDL₁ (V_T, panel F in Fig 2). If a lipolysis rate for V_T of up to 10 times the lipolysis rate of the rest of VLDL₁ was permitted, a poor fit was obtained; the VLDL₁ apoB calculated curve was displaced above the observed data while the VLDL₂ apoB calculated points fell below the observed data during the first 2;–3 hours. Allowing V_T to turn over at a rate in excess of 100 pools/d did permit the data from subject 16 and others to fit such a model. However, it is difficult to conceive how one VLDL₁ particle can be delipidated at 1;–2 pools/d and another at 100 pools/d in the same person. Conceptually we prefer a system in which the liver secretes a range of particles varying in size from VLDL₁ to LDL. HepG2 cells (a human-derived liver cell (31)) and perfused rabbit livers in recent studies from our laboratory (32) secrete particles the size of small VLDL, IDL, and LDL, thus the model is physiologically plausible. Further, new evidence has emerged that human cardiac muscle can secrete LDL directly (33).

The model was constrained by introducing parameter dependencies to reduce the number of unknowns to within the limit necessary for system identifiability (see legend to Fig 1). The model was tested and found to be a priori globally identifiable (15), i.e., all rate constants could be determined (34) if the data were of sufficient precision.

Steady state production rates (in mg/d) and transport rates of apoB (in mg/d) from one lipoprotein fraction to another were obtained by the application of the model to the observed data. Fractional transfer rates (FTR) were calculated as follows: for VLDL₁ to VLDL₂ the transport rate from compartments 7 to 9 was divided by the VLDL₁ apoB mass (combined masses of compartments 6 + 7 + 8); for VLDL₂ to IDL transport rates from compartment 10 to 12 and 10 to 13 were combined and divided by the VLDL₂ apoB mass (compartments 9 + 10 + 11); for IDL to LDL the transport rate from compartment 12 to 14 was divided by IDL apoB mass (compartments 12 + 13). Fractional rates of direct catabolism were derived similarly for VLDL₁ by dividing the sum of outputs from compartments 7 and 8 (in mg/day) by the VLDL₁ apoB mass, for VLDL₂ by dividing the sum of outputs from compartments 10 and 11 by the VLDL₂ mass, and for IDL by dividing the sum of outputs from compartments 12 and 13 by the IDL apoB mass. The overall fractional catabolic rate (FCR) for VLDL₁, VLDL₂, and IDL-apoB was the sum of the FTR and the fractional rate of direct catabolism for each lipoprotein. The LDL apoB fractional catabolic rate was calculated as the output from compartment 14 divided by its mass. The residence time (RT) of a lipoprotein was the reciprocal of the FCR for each lipoprotein class.

Ethical considerations
All subjects participating in this study gave written, informed consent. The study met the requirements of the Ethics Committees of the Klinikum Grosshadern in Munich and of Glasgow Royal Infirmary University NHS Trust.

Statistical analysis
Relationships between kinetic variables and plasma lipid, lipoprotein and apolipoprotein B concentrations were tested by regression analysis using Minitab Version 10 (Minitab Inc., State College, PA).

Concentrations rather than pool sizes were used to identify possible associations between production and transport rates (which were expressed per kg body weight) and the plasma level of an apoB-containing lipoprotein. This overcomes possible exaggeration of the strength of a relationship due to the inclusion of body weight as a factor in both response and predictor variables. Variables that were not normally distributed (as tested by generating normality plots in Minitab) were transformed prior to use in regression analysis. Specifically, plasma triglyceride, VLDL₁ FTR, VLDL₂ FTR, and the fractional rates of VLDL₁, VLDL₂ direct catabolism, and the fraction of IDL + LDL apoB direct production were normalized by taking their logarithm. Best subset and stepwise regression were performed to ascertain the potential metabolic determinants of plasma triglyceride, IDL apoB- and LDL apoB-plasma concentrations. Analysis of variance (ANOVA) was used to compare kinetic parameters in subjects in the three LDL-size pattern groups (A, I, B). Fisher's pairwise comparisons were made if the ANOVA was significant (P < 0.05) and a t-test was used to generate P values for the comparison.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

The subjects of this study were 20 males aged 21 to 59 years. They were chosen initially not by LDL size phenotype but to represent a wide range of plasma lipid levels: 0.66 to 2.40 mmol/l for plasma triglyceride and 3.95 to 6.95 mmol/l for plasma cholesterol (Table 1), in order to uncover the principal kinetic parameters that governed VLDL metabolism across the putative threshold of a plasma triglyceride of 1.5 mmol/l and so influenced the appearance of small dense LDL. Body weight varied from 58 to 111 kg and was related positively to plasma triglyceride (r = 0.46, P = 0.023), negatively to HDL cholesterol (r = -0.45, P = 0.028) but not to plasma or LDL cholesterol.

VLDL₁ and VLDL₂ apoB kinetics and the regulation of plasma triglyceride
Across the plasma triglyceride range seen in these subjects, VLDL₁ and VLDL₂ apoB concentration and pool size varied about 10-fold ( Table 2, Table A1). The levels of the two lipoproteins were strongly correlated, r = 0.70, P < 0.0001, but as described below, they had distinct kinetic properties. Typical tracer/tracee curves for apoB in the four lipoprotein fractions are given in Fig 3 for subjects with pattern A versus pattern B LDL. In VLDL₁ and VLDL₂ the initial rate of rise in tracer/tracee ratio was more rapid for the subject with pattern A than for pattern B subject, indicating a higher turnover rate (i.e., FCR). Similarly, in IDL apoB and LDL apoB, the rate of clearance of tracer from these fractions was faster in the pattern A than pattern B subjects. Multicompartmental modeling generated kinetic rate constants which are summarized for subjects with pattern A, I, and B in Table 2 and Table 3 and presented individually in Table A1 and Table A2.

View larger version (24K): [in this window] [in a new window]   Figure 3. Apolipoprotein B metabolism in pattern A and pattern B subjects. Tracer/tracee curves are given for subjects with a pattern A LDL size phenotype (subject 6, triangles with dashed line) and a pattern B (subject 20, squares with solid line).

VLDL₁ apoB production rate in the whole group of subjects varied from 5.4 to 26.6 mg/kg/d (Table 2; 452 to 2074 mg/day, Table A1). There was no significant correlation between VLDL₁ apoB concentration and the rate of VLDL₁ apoB production, but there was a positive association between the residence time of the particle and its concentration (r = 0.80, P < 0.0001). As the disappearance of tracer from VLDL₁ was followed over a prolonged period, it was possible to distinguish two fates for the apoB of this fraction. The first was transfer to VLDL₂, the second, direct catabolism of the lipoprotein. The fractional rate of the former diminished as plasma triglyceride increased (log VLDL₁ FTR vs. log plasma triglyceride, r = -0.83, P < 0.0001). The rate of the latter was highly variable and showed no univariate association with plasma triglyceride or VLDL₁ apoB concentration.

Direct production of apoB in the VLDL₂ fraction showed substantial variation, from 0.2 mg/kg/d to 8.4 mg/kg/d (Table 2, 15 mg/d, Table A1). This variable tended to increase with increasing plasma triglyceride; the association was not significant when production was expressed as mg/kg/d but was in mg/d (VLDL₂ apoB production vs. log plasma triglyceride, r = 0.50, P = 0.015). VLDL₁ apoB and VLDL₂ apoB production rates were not related to one another (r = 0.0, NS). There were correlations between the VLDL₂ apoB concentration in plasma and the VLDL₂ apoB direct production rate (r = 0.51, P = 0.013) as well as the VLDL₂ FTR (r = -0.74, P < 0.0001) but not between VLDL₂ apoB concentration and the VLDL₁ to VLDL₂ apoB transport rate (r = 0.0, NS) or the fractional rate of VLDL₂ apoB direct catabolism (r = -0.32, P = 0.093).

Plasma triglyceride was strongly correlated with variation in both VLDL₁ and VLDL₂ apoB concentrations (r = 0.84 and r = 0.92, respectively, both P < 0.0001). Best subset and stepwise regression analysis revealed that the most important predictors of plasma triglyceride were VLDL₁ apoB FTR (explaining 71% of the variation) followed by the fractional rate of VLDL₁ apoB direct catabolism (explaining an additional 7%) and VLDL₂ apoB production (an additional 3%). VLDL₁ and VLDL₂ apoB FTR were highly correlated (r = 0.76, P < 0.0001) and could be used interchangeably in the regression analysis. Together the above parameters of VLDL₁ and VLDL₂ apoB metabolism explained 81% of the observed variation in plasma triglyceride levels. Most subjects were homozygotes for the E₃ allele of apoE and there was no relationship between apoE phenotype and VLDL₁– or VLDL₂–apoB kinetics in this group.

Apolipoprotein B kinetics in IDL and LDL
D₃-leucine enrichment in IDL and LDL apoB rose over the first 10 h of the turnover (Fig 3) and these lipoprotein fractions exhibited peak incorporation of tracer many hours after its administration, e.g., in LDL apoB at 30;–50 h. The general validity of the stable isotope technique can be seen by comparing the results for the LDL apoB FCR obtained in the present study ( Fig 4) with those derived using a ¹²⁵I-labeled-LDL tracer in normolipemic subjects in an earlier publication (35). The values obtained for the LDL FCR and the relationship to plasma triglyceride were similar using both methods.

View larger version (19K): [in this window] [in a new window]   Figure 4. Relationship between plasma triglyceride and LDL apoB FCR. LDL apoB FCR values in the present study were compared to those seen in an earlier group of subjects in whom the apoLDL FCR was determined using a tracer of 125I-labeled-LDL (35). The correlation for the previous data set was r = -0.54, P < 0.01 (35); ()-previous study; (•)-present subjects.

IDL apoB concentration (Table 3; Table A2) correlated with the VLDL₂ apoB FTR (log VLDL₂ FTR vs. IDL apoB concentration, r = 0.69, P < 0.001) but not the amount of VLDL₂ converted to IDL apoB (r = 0.0, NS) or the IDL apoB production rate (r = 0.0, NS). Variation in IDL apoB concentration was also related to IDL apoB FTR (r = -0.71, P < 0.0001). In regression analyses, IDL apoB FTR was the most important determinant of IDL apoB concentration (accounting for 53% of the variation), followed by IDL apoB fractional rate of direct catabolism (explaining a further 17% of the variation) and VLDL₂ apoB FTR (explaining a further 8%).

Variation in LDL apoB concentration was related in univariate regression to LDL apoB FCR (r = -0.77, P < 0.001) and VLDL₂ apoB production (r = 0.48, P = 0.018) but not significantly to VLDL₁ apoB production (r = 0.0, NS), IDL apoB production (r = 0.38, P = 0.058), or LDL apoB production (r = 0.0, NS). In multivariate analysis, LDL apoB concentration was best explained by two variables, LDL apoB FCR (accounting for 61% of the variation) and VLDL₂ apoB production (explaining a further 7%). LDL apoB FCR was related reciprocally to plasma triglyceride (r = -0.60, P = 0.002, Fig 4) but positively to both the amount of IDL + LDL apoB production (r = 0.59, P = 0.004) and the fraction of total apoB input into the IDL/LDL density range due to direct production (r = 0.65, P = 0.001, Fig 5A).

View larger version (16K): [in this window] [in a new window]   Figure 5. Relationship between IDL/LDL apoB direct production, LDL apoB FCR, and plasma triglyceride. In panels A and B, the fraction of direct production was calculated from the data in Table A1 and Table A2. It is the amount of IDL + LDL apoB direct production divided by the total apoB input. Plasma triglyceride and the fraction of IDL/LDL apoB direct production were transformed to loge values while LDL apoB FCR was used untransformed in the regression analyses. Separate symbols are used for subjects of differing LDL size phenotype: (•)-phenotype A; ()-intermediate phenotype; ()-phenotype B. Regression lines were generated using all subjects.

In both the IDL and LDL density intervals it was necessary using the model in Fig 1 to invoke apoB direct production in order to generate a satisfactory fit to the observed data. The amount of apoB entering the metabolic cascade by these routes was unrelated to the circulating mass of IDL and LDL but the percent of total apoB production that was directly released into these density ranges was inversely associated with the plasma triglyceride level (r = -0.54, P = 0.009, Fig 5B), and positively related to HDL cholesterol (r = 0.59, P = 0.004) and to the VLDL₁ to VLDL₂ apoB FTR (r = 0.48, P = 0.018) but not to the VLDL₂ to IDL apoB FTR (r = 0.18, NS).

ApoB metabolism in subjects with LDL size pattern A vs. pattern B
As might be predicted from the above findings and the known association of LDL subfraction distribution with plasma triglyceride, LDL apoB kinetic parameters differed in individuals with differing LDL size phenotypes (Table 1, Table 2, Table 3). Pattern B subjects on average had a low mean LDL apoB FCR (0.36 ± 0.14) and a low percent IDL/LDL derived by direct production (13 ± 8%) (Fig 5B). In contrast, pattern A subjects had a higher mean LDL apoB FCR (0.55 ± 0.20) and a tendency to derive a higher proportion of their LDL from IDL/LDL direct production (25 ± 16%). Further significant differences were seen in the VLDL density range with higher VLDL₁ and VLDL₂ apoB concentrations in pattern B due principally to reduced FTRs for these lipoproteins (Table 2). VLDL₂ apoB production was significantly higher in pattern B compared to pattern A subjects. Plasma triglyceride levels in the pattern B group were above 1.5 mmol/l principally because of a prolonged residence time for VLDL particles ( Fig 6) in these subjects. In general those with an intermediate LDL size (pattern I) gave results intermediate between patterns A and B.

View larger version (16K): [in this window] [in a new window]   Figure 6. VLDL1 residence time in subjects of differing LDL size phenotype. Relationship between residence time (RT = reciprocal of the sum of the VLDL1 apoB FTR plus the fractional rate of VLDL1 apoB direct catabolism) and plasma triglyceride. Symbols as in Fig 5. Regression line was derived using all subjects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

Most subjects who suffer a coronary event have plasma lipid levels that fall within the range of values commonly seen in the population. Hence the present study examined the differences in apoB metabolism in men with normal to moderately elevated cholesterol and triglyceride levels. More particularly, we attempted to understand the transport of apoB-containing lipoproteins in subjects with ALP who were at particularly elevated CHD risk. Two protocols were used for the administration of deuterated leucine, bolus and infusion, and the resultant kinetic data merged in the present report. Theoretically, using a multicompartmental modeling approach to data analysis, derived kinetic parameters should not be influenced by tracer administration protocol. In confirmation of this Parhofer et al. (13) (36) in two separate studies directly compared the results obtained by bolus versus infusion protocols in four normal and five hyperlipidemic subjects. The latter were studied on two occasions (on and off drug). In all instances the same derived rate constants fitted both sets of experimental data. In an earlier publication we reproduced these findings for VLDL₁, VLDL₂, and IDL apoB in a single individual given bolus versus infused amino acid. There was a discordancy in the LDL apoB FCR with the two protocols but this we attributed to measurement difficulties rather than an inherent systematic difference.

Before focussing on ALP, a number of general observations can be made about apoB kinetics in the men studied. First, VLDL₁ and VLDL₂ apoB production rates, although varying widely, were not correlated. This finding indicates that the liver can regulate independently the assembly and secretion of these subclasses of VLDL, a suggestion that receives further support from recent in vivo observations in which a) estrogen administration was shown to stimulate VLDL₁ apoB but not VLDL₂ apoB production (37) and b) the acute infusion of insulin in a hyperinsulinemic, euglycemic clamp in normal subjects was shown to suppress the production of VLDL₁ apoB but not affect that of VLDL₂ apoB (38). The mechanism by which the liver is able to vary the amount of large versus small VLDL secreted is unknown.

Recent experiments in cell culture (31) (39) suggest that VLDL assembly is more complex than was originally thought, involving a two-step process in which a small lipoprotein particle containing at first little triglyceride is formed in the rough endoplasmic reticulum (ER), and then the bulk of the triglyceride core is added to this at the junction of the rough and smooth ER. It is possible that the release of small VLDL follows the addition of a relatively small quantity of triglyceride (or of cholesteryl ester) to the nascent particle while large VLDL is formed by the addition of a substantial triglyceride core in a second, quantum step. Second, input of apoB into IDL and LDL was substantial and inversely related to plasma triglyceride. The nature of ‘direct’ LDL production has been controversial with some investigators suggesting that it is the consequence of a very fast lipolytic pathway (29) (40). However, on the basis of the arguments rehearsed in Methods, we favor the concept that direct IDL/LDL production occurs; cultured hepatic cells (HepG2 [31]), perfused rabbit livers (32), and cardiac tissue (33) all appear capable of producing apoB-containing particles of the size of IDL/LDL. The link with plasma triglyceride concentration suggests that in subjects with a low circulating concentration of this lipid, the liver has insufficient triglyceride to assemble even a VLDL₂-sized particle and IDL/LDL are secreted. Third, in a population survey conducted previously in normolipemic subjects, 75% of the variation in plasma triglyceride was determined by the VLDL₁ concentration (9). The present study shows that the level of this lipoprotein subclass is regulated mainly by its fractional transfer rate to VLDL₂. The substantial decrease in clearance rates of both VLDL₁ and VLDL₂ as plasma triglyceride rises in the normal population is in good agreement with the findings of Cortner et al. (41) who observed a fall in total VLDL (i.e., S_f 20;–400) apoB FCR from 27 pools/d at a plasma triglyceride of 0.4 mmol/l to 4.6 pools/d at 2.2 mmol/l. This fall-off in FTR is likely to reflect the rates of lipolysis and could be due to alteration in lipoprotein lipase (LpL) activity; however, in a separate survey in normal subjects we found, in confirmation of the results of Lamarche et al. (42), only a weak, negative correlation between post-heparin plasma LpL and plasma triglyceride (r = -0.21, P = 0.007, n = 137; C. J. Packard, C. E. Tan, and J. Shepherd, unpublished observations). Thus, we suggest that other factors such as the apoC-II content or the apoCII/CIII ratio in VLDL (43) may determine the rate of VLDL lipolysis in our subjects.

The atherogenic lipoprotein phenotype which is characterized by moderately elevated plasma triglyceride (generally >1.5 mmol/l), low HDL cholesterol, and the presence of small, dense LDL is associated with a 3- to 7-fold risk of CHD (1) (2) (3). Of the present group, six subjects had pattern B LDL (i.e., small dense particles) and most of these had associated alterations in plasma triglyceride and HDL cholesterol. These were compared to nine pattern A subjects and five in whom the LDL size was intermediate. ApoB metabolism was altered throughout the VLDL–LDL cascade in the group with pattern B versus pattern A LDL. Specifically, in the former there was an accumulation of VLDL₁ and VLDL₂ due to a reduced FTR (i.e., lipolysis rate) for these lipoproteins (Fig 6), and a prolonged residence time for LDL. The findings here are in concordance with the earlier characterization of apoB kinetics in hyperapobetalipoproteinemics who exhibit a preponderance of dense LDL (10). The relationship between LDL apoB FCR and plasma triglyceride is complex. We have shown that, compared to subjects with low plasma triglyceride levels, those with high normal triglyceride or moderate hypertriglyceridemia have a reduced LDL FCR (35). On the other hand, patients with more severe hypertriglyceridemia exhibit higher than normal LDL clearance rates, possibly due to accelerated, receptor-independent catabolism (7). A full discussion of this topic is presented in a recent review (7).

The metabolic conditions described for pattern B subjects in Table 2 and Table 3, we suggest, provide the environment needed for the formation of small, dense LDL ( Fig 7). The key step in reducing LDL size in this model is the removal of core triglyceride and surface phospholipid by hepatic lipase (HL). This enzyme has been shown to have an important role in determining small dense LDL levels in particular in men, who have substantially higher levels of activity than women (7) (9). Because HL cannot act on cholesteryl ester, this lipid must be exchanged out of LDL and replaced with triglyceride for a particle to be susceptible to the enzyme's action. Cholesteryl ester transfer protein (CETP) is the agent which facilitates this core lipid exchange. From the results of the present investigation it is proposed that LDL has to have a residence time in excess of 3 days and be exposed to a high level of VLDL triglyceride (generated either by over-production of VLDL as in diabetics (44) or inefficient lipolysis as presented here) before enough triglyceride is present in the LDL particle so that the next time it is exposed to HL, a size reduction is effected. The model suggests that the cholesteryl ester content of the most abundant, intermediate-sized LDL has to be reduced (by CETP-mediated transfer) to that seen in small LDL before the conversion step can occur. In pattern A subjects there is insufficient time for a LDL particle to lose this cholesteryl ester and acquire triglyceride before it is catabolized by receptors. Further, in these subjects VLDL triglyceride availability is limited while HDL cholesteryl ester is relatively abundant and available via CETP to transfer to LDL to maintain its core composition. In pattern I subjects the residence time for LDL is prolonged and VLDL triglyceride high is enough to generate some small LDL, while in pattern B individuals most of LDL is remodelled due to the long residence time and the higher ratio of VLDL triglyceride to HDL cholesteryl ester. It has been shown in vitro (45) that small dense LDL binds less well to LDL receptors, in accord with our findings that the LDL FCR is reduced by about 60% in subjects with pattern B and with the data of Teng et al. (10). The model in Fig 7 suggests that the reason for this is that LDL with a prolonged residence time will be preferentially converted to small-sized particles.

View larger version (21K): [in this window] [in a new window]   Figure 7. Metabolic model for formation of small dense LDL. Mean residence time for LDL was calculated for each group (A, I, B). Likewise, the weighted mean residence time for VLDL particles was determined as: (VLDL1 apoB RT x VLDL1 apoB pool + VLDL2 RT x VLDL2 apoB pool)/(VLDL1 + VLDL2 apoB pool). Estimates of mean plasma levels of VLDL triglyceride (µmol/l), HDL cholesteryl ester (µmol/l), and LDL cholesteryl ester and triglyceride (in µmol/l) are given inside the circles for each group. LDL cholesteryl ester and triglyceride contents were determined in a previous compositional analysis of isolated intermediate size LDL (d 1.034;–1.044 g/ml; LDL-II in reference 3) and small dense LDL (d 1.045;–1.060 g/ml; LDL-III in reference 3) in 37 controls whose plasma lipid levels were in the range of the subjects in the present study (M. J. Caslake and C. J. Packard, unpublished observations). Intermediate-sized LDL was composed (g/100 g) of 24.1% protein (of which 97% was apoB), 9.2% free cholesterol, 37.5% cholesteryl ester, 10.5% triglyceride, and 18.6% phospholipid, whereas small dense LDL was 24.4% protein (97% apoB), 7.3% free cholesterol, 27.6% cholesteryl ester, 12.9% triglyceride, and 17.1% phospholipid. Cholesteryl ester content of HDL was taken as 70% of total HDL cholesterol (46). VLDL triglyceride levels were calculated from published compositions of VLDL1 and VLDL2 and the apoB concentrations for these subfractions in the three LDL-size phenotype groups (see reference 9: VLDL1 triglyceride:protein weight ratio = 4.30; VLDL2 triglyceride:protein ratio = 2.29; in the present group of subjects apoB was 36 ± 3% of total protein in VLDL1 and 70 ± 8% of total protein in VLDL2). LDL is subject to continuous inter-particle lipid exchange while in the plasma compartment. It is suggested that the cholesteryl ester content of intermediate-sized LDL (1930 µmol/l) has to be reduced to that normally seen in small LDL (1400 µmol/l) before the conversion in size to a small dense particle can occur. This requires the replacement of 530 µmol/l CE with 530 µmol/l TG. The activity of cholesteryl ester transfer protein (CETP) at about 85 µmol CE/h/liter plasma (47) may also be rate-limiting under some circumstances.

In conclusion, it appears that in normal to moderately hyperlipidemic men, the nature of apolipoprotein B-containing particles secreted from the liver, which can vary across a spectrum from VLDL₁ to LDL, has an impact upon subsequent lipoprotein catabolism. In contrast (at least within this population) total apoB production is not a determinant of plasma cholesterol and triglyceride levels. The efficiency of VLDL clearance is the primary regulator of VLDL levels in normal men and this, in turn, is a major controlling factor in small dense LDL formation.

	ACKNOWLEDGMENTS

Miss Emma Airley provided excellent secretarial assistance in the preparation of this manuscript. The work was supported by grants from the British Heart Foundation (190/1242) and the Deutsche Forschungsgemeinschaft (De 3591/1;–2). C. Mathes carried out turnover studies and provided lipoprotein and free amino acid preparations from two subjects (subjects 15 and 9).

Manuscript received March 10, 1999; and in revised form November 22, 1999

Abbreviations: apo, apolipoprotein; FCR, fractional catabolic rate; FTR, fractional transfer rate; d₃-leucine, tri-deuterated leucine; VLDL, very low density lipoprotein; IDL, intermediate density lipoprotein; LDL, low density lipoprotein; RT, residence time in plasma

	APPENDIX 1

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

	APPENDIX 2

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX 1 APPENDIX 2 REFERENCES

1. Austin, M. A., King, M-C., Vranigan, K. M., Krauss, R. M. 1990. Atherogenic lipoprotein phenotype: a proposed genetic marker for coronary heart disease risk. Circulation. 82:495-506[Abstract/Free Full Text].

2. Swinkels, D. W., Demacker, P. N. M., Hendriks, J. C. M., van't Laar, A. 1989. Low density lipoprotein subfractions and relationships to other risk factors for coronary artery disease in healthy individuals. Arteriosclerosis. 9:604-613[Abstract/Free Full Text].

3. Griffin, B. A., Freeman, D. J., Tait, G. W., Thomson, J., Caslake, M. J., Packard, C. J., Shepherd, J. 1994. Role of plasma triglyceride in the regulation of plasma low density lipoprotein (LDL) subfractions: relative contribution of small, dense LDL to coronary heart disease risk. Atherosclerosis. 106:241-253[Medline].

4. Grundy, S. M., Vega, G. 1992. Two different views of the relationship of hypertriglyceridemia to coronary heart disease. Arch. Intern. Med. 152:28-34[Abstract].

5. De Graaf, J., Hak-Lemmers, H. L. M., Hectors, M. P. C., Demacker, P. N. M., Hendriks, J. C. M., Stalenhoef, A. F. H. 1991. Enhanced susceptibility to in vitro oxidation of the dense low density lipoprotein subfraction in healthy subjects. Arteriosclerosis. 11:298-306[Abstract/Free Full Text].

6. Krauss, R. M. 1994. Heterogeneity of low density lipoproteins and atherosclerosis risk. Curr. Opin. Lipidol. 5:339-349[Medline].

7. Packard, C. J., Shepherd, J. 1997. Lipoprotein heterogeneity and apolipoprotein B metabolism. Arterioscler. Thromb. Vasc. Biol. 17:3542-3556[Abstract/Free Full Text].

8. McNamara, J. R., Jenner, J. L., Li, Z., Wilson, P. W. F., Schaefer, E. J. 1992. Change in LDL particle size is associated with change in plasma triglyceride concentration. Arterioscler. Thromb. 12:1284-1290[Abstract/Free Full Text].

9. Tan, C. E., Forster, L., Caslake, M. J., Bedford, D., Watson, T. D. G., McConnell, M., Packard, C. J., Shepherd, J. 1995. Relations between plasma lipids and postheparin plasma lipases and VLDL and LDL subfractions in normolipemic men and women. Arterioscler. Thromb. Vasc. Biol. 15:1839-1848[Abstract/Free Full Text].

10. Teng, B., Sniderman, A. D., Soutar, A. K., Thompson, G. R. 1986. Metabolic basis of hyperapobetalipoproteinemia. Turnover of apolipoprotein B in low density lipoprotein and its precursors compared with normal and familial hypersholesterolemia. J. Clin. Invest. 77:663-672.

11. Cryer, D. R., Matsushima, T., Marsh, J. B., Yudkoff, M., Coates, P. M., Cortner, J. A. 1986. Direct measurement of apolipoprotein B synthesis in human very low density lipoprotein using stable isotopes and mass spectrometry. J. Lipid Res. 27:508-516[Abstract].

12. Cohn, J. S., Wagner, D. A., Cohn, S. D., Millar, J. S., Schaefer, E. 1990. Measurement of very low density and low density lipoprotein apolipoprotein B-100 and high density lipoprotein apoA-I production in human subjects using deuterated leucine. J. Clin. Invest. 85:804-811.

13. Parhofer, K. G., Barrett, P. H. R., Bier, D., Schonfeld, G. 1991. Determination of kinetic parameters of apolipoprotein B metabolism using amino acids labeled with stable isotopes. J. Lipid Res. 32:1311-1323[Abstract].

14. Foster, D. M., Barrett, P. H. R., Toffolo, G., Beltz, W. F., Cobelli, C. 1993. Estimating the fractional synthetic rate of plasma apolipoproteins and lipids from stable isotope data. J. Lipid Res. 34:2193-2205[Abstract].

15. Demant, T., Packard, C. J., Demelmair, H., Stewart, P., Bedynek, A., Bedford, D., Seidel, D., Shepherd, J. 1996. Sensitive methods to study human apolipoprotein B metabolism using stable isotope-labeled amino acids. Am. J. Physiol. 270:E1022-E1036.

16. Demant, T., Packard, C. J., Stewart, P., Bedynek, A., Calder, A. G., Shepherd, J., Seidel, D. 1994. A sensitive mass spectrometry technique for the measurement of human apolipoprotein B metabolism in-vivo. Clin. Chem. 40:1825-1827[Free Full Text].

17. Lipid Research Clinics Program Manual of Laboratory Operations. 1975. DHEW Publications, No. (NIH) 75;–628, Washington, DC.

18. Havekes, L. M., Knyff, P., Beisiegel, U., Habing, J., Smit, M., Klasen, E. 1987. A rapid micromethod for apolipoprotein E phenotyping directly in serum. J. Lipid Res. 28:455-463[Abstract].

19. Krauss, R. M., Burke, D. J. 1982. Identification of multiple subclasses of plasma low density lipoproteins in normal humans. J. Lipid Res. 23:97-104[Abstract].

20. McNamara, J., Campos, H., Ordovas, J. M., Peterson, J., Wilson, P. W. F., Schaefer, E. J. 1987. The effect of gender, age and lipid status on low density lipoprotein subfraction distribution: results of the Framingham Offspring Study. Arteriosclerosis. 7:483-490[Abstract/Free Full Text].

21. Krauss, R. M., Dreon, D. M. 1995. Low density lipoprotein subclasses and response to a low fat diet in healthy men. Am. J. Clin. Nutr. 62:4785-4875.

22. James, R. W. B., Martin, B., Pometta, D., Fruchart, J. C., Duriez, P., Puchois, P., Farriaux, J. P., Tacquet, A., Demant, T., Clegg, R. J., Munro, A., Oliver, M. F., Packard, C. J., Shepherd, J. 1989. Apolipoprotein B metabolism in homozygous familial hypercholesterolemia. J. Lipid Res. 30:159-169[Abstract].

23. Lindgren, F. T., L. C. Jensen, and F. T. Hatch. 1972. The isolation and quantitation analysis of serum lipoproteins. In Blood Lipids and Lipoproteins: Quantitation, Composition and Metabolism. G. J. Nelson, editor. Wiley-Interscience, New York. 181;–274.

24. Egusa, G., Brady, D. W., Grundy, S. M., Howard, B. V. 1983. Isopropanol precipitation method for the determination of apolipoprotein B specific activity and plasma concentrations during metabolic studies of very low density lipoprotein and low density lipoprotein apolipoprotein B. J. Lipid Res. 24:1261-1267[Abstract].

25. Cobelli, C., Toffolo, G., Foster, D. M. 1992. Tracer-to-tracee ratio for analysis of stable isotope tracer data: link with radioactive kinetic formalism. Am. J. Physiol. 262:E968-E975[Abstract/Free Full Text].

26. Gaw, A., Packard, C. J., Lindsay, G. M., Griffin, B. A., Caslake, M. J., Lorimer, A. R., Shepherd, J. 1995. Over-production of small very low density lipoproteins (S_f 20;–60) in moderate hypercholesterolemia: relationships between apolipoprotein B kinetics and plasma lipoproteins. J. Lipid Res. 36:158-171[Abstract].

27. Hardman, D. A., Kane, J. P. 1986. Isolation and characterisation of apolipoprotein B-48. Methods Enzymol. 128:262-272[Medline].

28. Berman, M., Hall, M., III, Levy, I., Eisenberg, S., Bilheimer, D. W., Phair, R. D., Goebel, R. H. 1978. Metabolism of apoB and apoC lipoproteins in man: kinetic studies in normal and hyperlipoproteinemic subjects. J. Lipid Res. 19:38-56[Abstract].

29. Beltz, W. F., Kesaniemi, Y. A., Howard, B. V., Grundy, S. W. 1985. Development of an integrated model for analysis of the kinetics of apolipoprotein B in plasma, very low density lipoproteins, intermediate density lipoproteins and low density lipoproteins. J. Clin. Invest. 76:575-585.

30. Packard, C. J., Gaw, A., Demant, T., Shepherd, J. 1995. Development and application of a multicompartmental model to study very low density lipoprotein subfraction metabolism. J. Lipid Res. 36:172-187[Abstract].

31. Pease, R. J., Leiper, J. M. 1996. Regulation of hepatic apolipoprotein B-containing lipoprotein secretion. Curr. Opin. Lipidol. 7:132-138[Medline].

32. Ardern, H. A., Benson, G. M., Suckling, K. E., Caslake, M. J., Shepherd, J., Packard, C. J. 1999. Apolipoprotein B overproduction by the perfused liver of the St. Thomas' mixed hyperlipidemic (SMHL) rabbit. J. Lipid Res. 40:2234-2243[Abstract/Free Full Text].

33. Borén, J., Véniant, M. M., Young, S. G. 1998. ApoB100-containing lipoproteins are secreted by the heart. J. Clin. Invest. 101:1197-1202[Medline].

34. Cobelli, C., DiStefano, J., III. 1980. Parameter and structural identifiability concepts and ambiguities: a critical review and analysis. Am. J. Physiol. 239:R7-R24.

35. Caslake, M. J., Packard, C. J., Series, J. J., Yip, B., Dagen, M., Shepherd, J. 1992. Plasma triglyceride and low density lipoprotein metabolism. Eur. J. Clin. Invest. 22:96-104[Medline].

36. Parhofer, K. G., Barrett, P. H. R., Dunn, J., Schofeld, G. 1993. Effect of pravastatin on metabolic parameters of apolipoprotein B in patients with mixed hyperlipidemia. Clin. Invest. 71:939-946[Medline].

37. Walsh, B. W., Schiff, I., Rosner, B., Breenberg, L., Ravnikar, V., Sacks, M. 1991. Effects of postmenopausal estrogen replacement on the concentrations and metabolism of plasma lipoproteins. N. Engl. J. Med. 325:1196-1204[Abstract].

38. Malmström, R., Packard, C. J., Watson, T. D. G., Rannikko, S., Caslake, M., Bedford, D., Stewart, P., Yki-Järvinen, H., Shepherd, J., Taskinen, M-R. 1997. Metabolic basis of hypotriglyceridemic effect of insulin in normal men. Arterioscler. Thromb. Vasc. Biol. 17:1454-1464[Abstract/Free Full Text].

39. Borén, J., Rustaeus, S., Olofsson, S. O. 1994. Studies on the assembly of apolipoprotein B-100 and B-48 containing very low density lipoproteins in McA-RH7777 cells. J. Biol. Chem. 269:25879-25888[Abstract/Free Full Text].

40. Shames, D. M., Havel, R. J. 1991. De novo production of low density lipoproteins: fact or fancy. J. Lipid Res. 32:1099-1112[Abstract].

41. Cortner, J. A., Le, N-A., Coates, P. M., Bennett, M. J., Cryer, D. 1992. Determinants of fasting plasma triglyceride levels: metabolism of hepatic and intestinal lipoproteins. Eur. J. Clin. Invest. 22:158-165[Medline].

42. Lamarche, B., Després, J-P., Pouliot, M-C., Prud'homme, D., Moorjani, S., Lupien, P. J., Nadeau, A., Tremblay, A., Bouchard, C. 1993. Metabolic heterogeneity associated with high plasma triglyceride or low HDL cholesterol levels in men. Arterioscler. Thromb. 13:33-40[Abstract/Free Full Text].

43. Ginsberg, H. N., Le, N-A., Goldberg, I. J., Gibson, J. C., Rubinstein, A., Wang-Iverson, P., Norum, R., Brown, W. V. 1986. Apolipoprotein B metabolism in subjects with deficiency of apolipoproteins C-III and AI: evidence that apolipoprotein C-III inhibits catabolism of triglyceride-rich lipoproteins by lipoprotein lipase in vivo. J. Clin. Invest. 78:1287-1295.

44. Taskinen, M-R. 1995. Insulin resistance and lipoprotein metabolism. Curr. Opin. Lipidol. 6:153-160[Medline].

45. Nigon, F., Lesnik, P., Rouis, M., Chapman, M. J. 1991. Discrete subspecies of human low density lipoproteins are heterogeneous in their interaction with cellular LDL receptor. J. Lipid Res. 32:1741-1753[Abstract].

46. Shepherd, J., Packard, C. J., Stewart, J. M., Vallance, B. D., Lawrie, T. D. V., Morgan, H. G. 1980. The relationship between the cholesterol content and subfraction distribution of plasma high density lipoproteins. Clin. Chem. Acta. 101:57-62[Medline].

47. Guérin, M., Dolphin, P. J., Chapman, M. J. 1994. Preferential cholesteryl ester acceptors among the LDL subspecies of subjects with familial hypercholesterolemia. Arterioscler. Thromb. 14:679-685[Abstract/Free Full Text].

48. Demant, T., Mathes, C., Gutlich, K., Bedynek, A., Steinhauer, H. B., Bosch, T., Packard, C. J., Warwick, G. L. 1998. A simultaneous study of the metabolism of apolipoprotein B and albumin in nephrotic patients. Kidney Int. 54:2064-2080[Medline].