INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN HUMAN BEINGS (3, 30, 47) and in rabbits (10, 19, 21, 29, 35, 36, 53), atherosclerosis develops preferentially in characteristic arterial regions. Most cholesterol in atherosclerotic aortas derives from plasma lipoproteins (11, 54), suggesting that the interaction of lipoproteins with arteries plays an important role in atherogenesis. Degradation of lipoproteins provides cells with cholesterol, which they must either store as cholesterol ester or release via cholesterol efflux (12, 33).

Evidence suggests that some low-density lipoprotein (LDL) within atherosclerotic lesions is associated with the extracellular matrix, whereas the remaining LDL appears to be present in a form(s) that could more readily exchange with plasma LDL (26, 44). The association of LDL with the arterial extracellular matrix would be expected to both increase aortic undegraded LDL concentrations and increase LDL retention within atherosclerotic lesions, both of which have been observed in earlier studies (34, 46). The functional significance of the association of LDL with the arterial extracellular matrix components in vivo remains to be determined. However, studies in vitro suggest that such an association could promote the degradation of LDL and cholesterol accumulation in cells within atherosclerotic lesions (17, 18, 49, 50).

Several studies (35, 36) showed that compared with the uniform abdominal aorta, LDL degradation rates were increased for abdominal branch sites even in normal rabbits. However, aortic LDL degradation rates were not increased to the same extent as aortic undegraded LDL concentrations (35, 36), suggesting the possibility that some aortic undegraded LDL may not be available for cellular degradation even in the aortas of normal rabbits. Two studies (37, 46) reported that aortic LDL retention did not differ between the branch and uniform abdominal aorta of normal rabbits. However, neither of these studies considered whether LDL might be metabolically "sequestered" in the aorta of normal rabbits in the absence of atherosclerosis. Thus the physical state of the increased concentrations of undegraded LDL in the branch sites of the abdominal aorta of normal rabbits (35, 36) is not clear. Furthermore, it is possible that the designs of the earlier studies (37, 46) may have prevented detection of increased LDL retention in the abdominal branch sites of normal rabbits.

This study investigated the following hypotheses. First, even in normal rabbits, LDL interacts with the abdominal aorta in vivo as if part of the LDL were sequestered in a form that poorly exchanges with plasma LDL. Second, compared with the atherosclerosis-resistant abdominal aorta, more LDL is sequestered in abdominal branch sites, increasing LDL concentrations at branch sites. Third, the mean residence time of LDL within the aorta is increased at abdominal branch sites.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Rabbits. These studies used sexually mature, young female New Zealand White rabbits from Robinson Services (Winston-Salem, NC). All rabbits were acclimated to the animal facility for 1 wk before the metabolic studies. Each day, rabbits were fed 100 g of a cholesterol-free rabbit chow (Prolab, Agway, Ithaca, NY). Rabbits weighed about 2.5 kg when they were received and 2.58 ± 0.03 kg (mean ± SE, n = 33) at the end of the metabolic studies.

Isolation of LDL for metabolic studies. LDL (1.020 < d < 1.060 g/ml, where d is density of a given lipoprotein fraction) for labeling was isolated in the presence of a cocktail of antioxidants and proteolytic inhibitors and washed as described (32, 33). Female New Zealand White rabbits fed rabbit chow were used as plasma donors (32, 33). Isolated LDLs were dialyzed for 72 h against six changes of 2,000 volumes of 0.15 M NaCl and 20 mmol/l sodium phosphate (pH 7.4) containing 2 mmol/l disodium EDTA (buffer A) (32, 33, 36, 37) in the dark. After dialysis, protein was determined (27).

Labeling and characterization of LDL. Five experiments were done, each with LDL isolated from a different pool of plasma. For three experiments, LDL (24.2-28.3 mg protein) was directly labeled with ¹³¹I (0.71-0.83 mCi/mg protein) (7, 14, 32, 33) and then covalently coupled to ¹²⁵I-labeled tyramine cellobiose (TC) (¹²⁵I-TC,¹³¹I-LDL; 4.7-6.2 nmol TC and 0.71-0.83 mCi/mg protein) with cyanuric chloride (7, 28, 32, 33). For two experiments, LDL (10.8-11.2 mg protein) was coupled to ¹²⁵I-TC only (¹²⁵I-TC-LDL; 7.1-7.4 nmol TC and 0.89-0.92 mCi/mg protein) (7, 28, 32, 33). Labeled LDLs were dialyzed for 35-43 h with five to six changes of 1,000-2,000 volumes of buffer A (32, 33, 36, 37). Labeled LDLs were sterilized by filtration (35) before injection 5 to 6 days after isolation of LDL. Specific activities were 648 ± 138 (n = 5) cpm/ng for ¹²⁵I and 126 ± 28 (n = 3) cpm/ng for ¹³¹I. Polyacrylamide gel electrophoresis (22, 39, 40) of delipidated (13) LDL showed only 2.42 ± 0.1% (n = 5) and 2.44 ± 0.12% (n = 3) of ¹²⁵I-TC and ¹³¹I labels to be associated with apoprotein E, as in earlier studies with LDL isolated from normal rabbits (32, 33, 35-37). Agarose gel electrophoresis (25) was performed on a few LDL preparations. For these and other similar preparations (32, 33, 35-37), ~95% of ¹²⁵I-TC and ¹³¹I comigrated with unlabeled LDL. Only 1.4 ± 0.2% (n = 5) and 2.1 ± 0.6% (n = 3) of ¹²⁵I-TC and ¹³¹I, respectively, were soluble in 10% trichloracetic acid (TCA) (39, 40), whereas 6.9 ± 1.1% (n = 5) and 3.5 ± 0.5% (n = 3) of these labels could be extracted with chloroform/methanol (13). Iodination, use, and disposal of labeled lipoproteins followed procedures recommended by the Wake Forest University School of Medicine Office of Environmental Health and Safety.

Metabolic studies. The time-dependent aortic accumulation of undegraded LDL and aortic LDL degradation products was determined up to 120 h after injecting radiolabeled LDL in 33 normal rabbits. Twenty-five rabbits were injected with ¹²⁵I-TC,¹³¹I-LDL (5.2 ± 0.3 × 10^{8 125}I cpm/kg and 1.2 ± 0.1 × 10^{8 131}I cpm/kg). Blood samples were collected after 5 min and at increasing intervals after injection (33, 35, 36) until perfusion (below) at 2.30 ± 0.11 (n = 4), 6.36 ± 0.19 (n = 3), 15.85 ± 0.09 (n = 2), 24.03 ± 0.40 (n = 3), 30.63 (n = 1), 40.40 ± 0.12 (n = 2), 48.25 ± 0.41 (n = 4), 72.28 ± 0.51 (n = 2), 96.38 ± 0.28 (n = 2), and 119.81 ± 0.27 h (n = 2) after injection. To determine the sum of aortic undegraded LDL and aortic LDL degradation products during the first hour after injection (when aortic LDL degradation should be minimal) (51), eight rabbits were injected intravenously with ¹²⁵I-TC-LDL (5.2 ± 0.9 × 10^{8 125}I cpm/kg). Blood samples were collected as described above (four during 0.6 h and five during 1.2 h), and rabbits were perfused after 0.66 ± 0.03 (n = 4) or 1.26 ± 0.01 h (n = 4).

Aortic sampling. After fixation in situ, the aorta was removed (32, 33). Fixation was continued overnight in half-strength Karnovsky's fixative (35, 36) to preserve the TC label present on the undegraded LDL and aortic LDL degradation products (7). The thoracic and abdominal aortas were separated (31). After adventitial tissue was removed from the abdominal aorta, this aortic segment was opened longitudinally, pinned flat, and photographed. Samples of the abdominal aorta distal to orifices of the celiac, superior mesenteric, and right and left renal arteries (branch sites) were collected (31, 32). Branch sites and the remaining atherosclerosis-resistant uniform abdominal aorta were photographed again. Fixed aortic samples were weighed and counted for radioactivity.

Lipids and lipoproteins. Blood samples collected just before euthanization were immediately mixed with disodium EDTA (final concentration 2.7 mmol/l). Very-low-density lipoprotein plus intermediate-density lipoprotein, LDL, and high-density lipoprotein were isolated from these plasma samples at d < 1.020, 1.020 < d < 1.060, and d > 1.060 g/ml (16, 33). Plasma lipoproteins were also separated by electrophoresis (25, 36, 37). Plasma cholesterol concentrations were determined (23) in the Centers for Disease Control-standardized lipid analytical laboratory of the Wake Forest University School of Medicine. LDL cholesterol concentrations were determined as described (36).

Radioassay. Total and TCA-soluble ¹²⁵I and ¹³¹I radioactivity in the aortic samples, plasma, and lipoprotein fractions were determined in a well-type gamma counter with a 3-inch crystal (Cobra II autogamma, Packard) as described (32, 33).

Total body LDL fractional catabolic rate. The total body fractional catabolic rate (FCR) of LDL was calculated for each rabbit euthanized 6 h after injection as described (33, 35, 36). The aggregate FCR of LDL for all 25 rabbits injected with ¹²⁵I-TC,¹³¹I-LDL was determined similarly from combined plasma data for these rabbits.

LDL transport into intima media. Intima-media permeability to LDL (µl plasma · h¹ · g aorta¹) was calculated as described (32, 33, 37) from aortic data from rabbits perfused 0.66 ± 0.03 or 1.26 ± 0.01 h after injection of ¹²⁵I-TC-LDL. Intima-media permeability to LDL was also calculated by multiplying the fractional rate constant for transfer of LDL from plasma into the aorta (determined from fits of the models to aortic data, below) by the mean plasma volume of rabbits in this study. The mean plasma volume of the rabbits (108.5 ± 2.4 ml, 42.2 ± 0.8 ml/kg, n = 31) was determined by dilution of the injected doses.

Determination of rate constants for transport of LDL into and out of aorta and metabolism of LDL by aorta. Fractional rate constants defining the aortic transport and metabolism of LDL were determined by fitting compartmental models to aortic data. These analyses were done with pooled aortic data from all 33 rabbits using the simulation, analysis, and modeling program (SAAM 31.0) (5). For rabbits given LDL labeled with ¹³¹I and ¹²⁵I-TC, aortic data used in these analyses were data for the time-dependent aortic accumulation of undegraded LDL (traced by protein-bound ¹³¹I) and the time-dependent aortic accumulation of products of aortic LDL degradation (determined by subtracting protein-bound aortic ¹³¹I from aortic total ¹²⁵I-TC, each expressed in terms of a fraction of the injected doses) (8, 33, 35, 36). For rabbits given LDL labeled only with ¹²⁵I-TC, total aortic ¹²⁵I-TC, representing the sum of aortic undegraded LDL and aortic LDL degradation products, was used in this analysis. Fractional rate constants defining the transfer of LDL into and out of the aorta and between aortic compartments and the metabolism of LDL within the aorta were assumed identical for all rabbits. However, because data for the time-dependent specific activity of radiolabeled LDL in plasma of individual rabbits were collected, and because there were small differences among rabbits, ¹²⁵I-TC,¹³¹I-LDL (or ¹²⁵I-TC-LDL) was considered to enter the aortas of the rabbits at the time-dependent specific activity of ¹²⁵I-TC,¹³¹I-LDL (or ¹²⁵I-TC-LDL) in plasma of those individual rabbits. In these analyses, aortic data were weighted by the inverse of the standard deviation for aortic data for replicate rabbits studied at the same time after injection. For this purpose, aortic data from the single rabbit studied 30 h after injection were assumed to have standard deviations equal to the average for corresponding aortic data of rabbits studied 6-48 h after injection.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Illustration of the compartmental models used to analyze aortic data. All models show that low-density lipoprotein (LDL) enters aorta from plasma. Left, 1-pool model of aortic undegraded LDL; all LDL enters one homogeneous aortic compartment (pool 1) from which efflux and degradation occur. Middle, 2-pool model of aortic undegraded LDL. Plasma LDL enters two aortic compartments. Aortic undegraded LDL in pool 1 represents the LDL that is subject to cellular degradation (thus "metabolically active pathway") and also effluxes from the aorta. Aortic LDL in pool 2 represents LDL "sequestered" so that it is not available for degradation and effluxes slowly from aorta (thus "sequestered pathway"). Right, 3-pool model of aortic undegraded LDL. Plasma LDL enters three aortic compartments. As before, aortic undegraded LDL in pool 1 represents the LDL that is subject to cellular degradation (metabolically active pathway) and could efflux from the aorta. Aortic LDL in pools 2 and 3 represent LDL that is sequestered in two metabolically different forms that are not available for degradation and that efflux slowly from the aorta (sequestered pathways). All aortic models include a compartment (pool 4, Deg Prod) representing the accumulated products of aortic LDL degradation. Pathways of transport or metabolism of LDL are indicated by arrows. Fractional rate constants (k) for transfer into compartment i from compartment j (kij) are shown.

1

1

1

Undegraded LDL concentrations. Aortic undegraded LDL concentrations were calculated from fractional rate constants determined by the fitting of kinetic models to the aortic data (5). These calculations also used data for the plasma volume (above) and mean plasma LDL cholesterol concentrations (below) for these rabbits. Concentrations of aortic undegraded LDL were expressed both in terms of absolute amounts of LDL cholesterol and as percentages of the plasma LDL cholesterol concentration (33, 35, 36).

LDL mean residence time within aortas. LDL mean residence times within the aortas were calculated for the kinetic models using the fractional rate constants determined by the fitting of these models to aortic data (5, 37). LDL mean residence times were also determined by an independent method, stochastic analysis (41). This method makes no assumptions about the metabolic or transport processes that LDL undergoes within the aorta. The only assumption required to calculate LDL mean residence times by this method is that LDL enters the aorta at a constant rate at the mean specific activity of plasma LDL (41). LDL mean residence times determined by this method were calculated as the area under the curve of the time dependence of aortic undegraded LDL (traced by protein-bound ¹³¹I) from injection until infinite time divided by cumulative delivery of ¹³¹I-LDL from plasma to the aorta at infinite time (41). The area under the curve of the time dependence of aortic undegraded LDL from 0 to 15.85 h was calculated by the trapezoidal rule. The area under the aortic curve of undegraded LDL from 15.85 h to infinity was calculated by integrating (from 15.85 h to infinity) a monoexponential equation fitted to the aortic data from 15.85 to 120 h after injection. Cumulative delivery of ¹³¹I-LDL from the plasma to the aorta was determined by multiplying aortic permeability to LDL (determined from the 0.66 h period of uptake, above) by the area under the time-dependent ¹³¹I-LDL in plasma. The area under the time-dependent ¹³¹I-LDL in the plasma was calculated by integrating a biexponential equation fitted to composite data for clearance of ¹³¹I from plasma of all rabbits injected with ¹²⁵I-TC,¹³¹I-LDL (Fig. 2).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Composite data for plasma clearance of 131I for 25 rabbits injected with 125I-labeled tryamine cellobiose, 131I-labeled LDL (125I-TC,131I-LDL). Each symbol represents an individual datum. FCR, fractional catabolic rate.

Statistical methods. After analysis with the SAAM program, residual sums of squares from the fitting of the models to aortic data were compared using the F statistic (1). This provides a statistical basis for choice of one model over another (at P < 0.05). Data for branch sites and the uniform abdominal aorta were fitted in the same SAAM analysis to facilitate estimation of the differences between the corresponding parameters for these aortic sites for each model. Differences between the corresponding parameters determined for branch sites and the uniform abdominal aorta were evaluated by paired t-tests (43). All values are presented with standard errors.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Lipids and lipoproteins. In these normal rabbits, plasma cholesterol averaged 1.25 ± 0.06 mmol/l. LDL cholesterol was 0.39 ± 0.02 mmol/l, and the LDL cholesterol pool size was 39 ± 3 µmol.

Whole body LDL metabolism. For the 21 rabbits killed 6 h after injection, the FCR of LDL calculated from ¹²⁵I-TC was 0.075 ± 0.010 pools/h, whereas that calculated from ¹³¹I was 0.086 ± 0.010 pools/h. The ratio of the FCR calculated from ¹³¹I to that calculated from ¹²⁵I was 1.04 ± 0.06, not significantly different from unity and similar to earlier results (33, 36). This suggests that the concentration of EDTA used here and previously (33, 36) was sufficient to prevent the oxidation and subsequent rapid plasma clearance of LDL that occurred when ¹³¹I-LDL was dialyzed with much lower concentrations of EDTA (20). The FCR for LDL, determined from composite data for plasma clearance of ¹³¹I for all rabbits given ¹²⁵I-TC,¹³¹I-LDL, was 0.0878 ± 0.0002 pools/h (Fig. 2).

Aortic accumulation of undegraded LDL and LDL degradation products. Figure 3 shows the time dependence of aortic undegraded LDL traced by protein-bound ¹³¹I (Fig. 3A), LDL degradation products (difference between ¹²⁵I-TC and ¹³¹I-labels, Fig. 3B), and aortic undegraded LDL plus the aortic LDL degradation products traced by total ¹²⁵I-TC (Fig. 3C). Compared with the uniform abdominal aorta, branch sites accumulated more undegraded LDL (ratio 2.31 ± 0.15, P < 0.00002), more LDL degradation products (ratio 1.73 ± 0.15, P < 0.00005), and more total ¹²⁵I-TC (ratio 1.85 ± 0.09, P < 0.00002).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Time dependence of aortic undegraded LDL and aortic LDL degradation products for abdominal aorta (n = 33 normal rabbits). Rabbits were injected with 125I-TC,131I-LDL (n = 25) or 125I-TC-labeled LDL (125I-TC-LDL; n = 8) to quantitate undegraded LDL and aortic LDL degradation products. A: protein-bound 131I representing undegraded LDL (n = 25). B: difference between 125I-TC and 131I representing aortic LDL degradation products (n = 25). C: 125I-TC representing undegraded LDL plus aortic LDL degradation products (n = 33). All aortic data are presented as %injected dose/g aorta. Data are presented as means ± SE for n = 2-4 rabbits, except for 30 h when n = 1. In some cases, error bars do not extend beyond the boundaries of symbols. Solid lines, data for atherosclerosis-susceptible branch sites (). Dotted lines, data for atherosclerosis-resistant uniform abdominal aorta ().

Comparison of ability of models with one, two, and three compartments of aortic undegraded LDL to fit aortic data. Table 1 compares the goodness of fit of models with one, two, and three pools of aortic undegraded LDL to the aortic data for undegraded LDL and the LDL degradation products. As shown here, for the two-pool model, eliminating pathways of transfer between metabolically active and sequestered LDL (model 2B) did not significantly alter the residual sums of squares (P > 0.25). Eliminating other pathways of transport or metabolism of LDL reduced the ability of the two-pool model to fit aortic data (data not shown). Because residual sums of squares for models both with (model 2A, Table 1) and without (model 2B, Table 1) pathways of transfer between the metabolically active and sequestered LDL are indistinguishable, the simpler reduced model (model 2B) is the best statistically justifiable two-pool model. For the three-pool models, residual sums of squares for the model including all transport pathways (model 3A, Fig. 1, right) were not significantly different (P > 0.25) from those for the reduced model lacking the pathway for efflux from metabolically active LDL (k₀₁) and with the fractional rate constant for efflux from one pool of sequestered LDL (k₀₂) fixed at the value for the terminal slope of the time dependence of aortic undegraded LDL (model 3B). However, eliminating other transport or metabolic pathways impaired the ability of model 3B to fit the data for aortic undegraded LDL. Thus model 3B is the best statistically justified three-pool model for aortic undegraded LDL.

View larger version (38K): [in this window] [in a new window]   Fig. 4.   Observed and fitted values for the time dependence of aortic undegraded LDL and aortic LDL degradation products for branch (left) and uniform (right) abdominal aorta. Data (n = 33 normal rabbits) from Fig. 3 are plotted together with values calculated from fits of 3 models. A and B: protein-bound 131I representing undegraded LDL. C and D: difference between 125I-TC and 131I representing aortic LDL degradation products. E and F: total 125I-TC representing undegraded LDL plus aortic LDL degradation products. All aortic data are presented as %injected dose/g aorta. , Data that were fitted to the compartmental models during the 120-h interval after injection (A-D) and up to 1.26 h after injection (E and F). , Observed values for the sum of undegraded LDL and LDL degradation products not fitted to the models but compared with the sum of fitted values for undegraded LDL and LDL degradation products. All data are means ± SE of observed values (n = 2-4 rabbits, except for 30 h when n = 1). In some cases, error bars do not extend beyond the boundaries of symbols. Thin broken lines, fits with a model with 1 pool of aortic undegraded LDL (model 1A). Dotted lines, best model with 2 pools of aortic undegraded LDL (model 2B). Thick dashed lines, fits with the best model with 3 pools of aortic undegraded LDL (model 3B). For abdominal branch sites, residual sums of squares for undegraded LDL were reduced 47% (P < 0.01) with model 3B vs. model 1A and 30% (P < 0.025) for model 3B vs. model 2B. For uniform abdominal aorta, residual sums of squares for undegraded LDL were reduced 69% (P < 0.005) with model 3B vs. model 1A and 26% (P < 0.025) for model 3B vs. model 2B. This provides evidence for the presence of aortic undegraded LDL in at least three metabolically distinct forms within both branch and uniform abdominal aorta. For branch sites, residual sums of squares for degradation products were reduced 53% (P < 0.005) for model 2B vs. model 1A. Similarly, for uniform abdominal aorta, residual sums of squares for degradation products were reduced 32% (P < 0.05) for model 2B vs. model 1A. For both aortic sites, there was no significant additional reduction in sums of squares for degradation products for model 3B vs. model 2B.

Fractional rate constants for LDL transport and metabolism in abdominal aorta. Table 2 presents fractional rate constants determined from the fits to aortic data for the models that are shown in Fig. 4. Fractional rate constants for influx into the metabolically active pool as determined by all models were greater for abdominal branch sites. This difference was significant for model 1A and for model 3B, which fitted aortic data best. The fractional rate constant for efflux of LDL from the metabolically active pool was significantly greater for branch sites compared with the uniform abdominal aorta only for model 1A. No other fractional rate constants differed significantly between the abdominal branch sites and the uniform abdominal aorta for any of the models.

Aortic permeability to LDL. Figure 5 shows aortic permeability to LDL calculated from the 0.66 ± 0.03-h or 1.26 ± 0.01-h periods of aortic uptake and determined from the models fitted to all of the aortic data, including that for the 0.66- and 1.26-h time points. Aortic permeability to LDL as determined by all methods averaged 36-63% (P < 0.04 to P < 0.001) greater for branch sites compared with the uniform abdominal aorta. These differences were significant for models 1A, 3A, and 3B and for the 1.26-h period of uptake. If one can assume that the most accurate measure of aortic permeability derives from model 3B, one can calculate that values of aortic permeability determined after 0.66 and 1.26 h of aortic uptake underestimate aortic permeability by 7-12% and 27-32%, respectively, with no difference between the branch and uniform abdominal aorta.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   Permeability to LDL for abdominal aorta of normal rabbits. Values are shown ± SE. Solid bars, branch sites; cross-hatched bars, uniform abdominal aorta; 0.7 and 1.3 h, aortic permeability determined from 0.7- and 1.3-h periods of aortic uptake. 1A, 2A, 2B, 3A, and 3B represent aortic permeability determined from 1-pool (model 1A), 2-pool (models 2A and 2B), and 3-pool (models 3A and 3B) compartmental models, respectively. For 0.7- and 1.3-h periods of aortic uptake, n = 4. In comparison, n = 33 rabbits were used to determined aortic permeability by each of the compartmental models. *P < 0.04, **P < 0.025, and ***P < 0.001 for comparison between branch sites and uniform abdominal aorta.

Undegraded LDL concentrations in abdominal aorta. Figure 6 shows aortic undegraded LDL concentrations determined from fractional rate constants (Table 2) characterizing the fits of the models to aortic data shown in Fig. 4. Aortic concentrations of sequestered undegraded LDL differed between the aortic sites for both the two-pool and three-pool models. For models 2B and 3B, concentrations of sequestered aortic undegraded LDL averaged 124% and 109% higher, respectively (P < 0.0002 and P < 0.0001, respectively) for branch sites (Fig. 6A). Sequestered LDL, as determined by models 2B and 3B, accounted for most (>87% and >98%, respectively) of the aortic undegraded LDL. Thus total aortic undegraded LDL concentrations calculated by models 2B and 3B were 117% and 109% greater, respectively, for abdominal branch sites (P < 0.0002 and P < 0.00005, respectively; Fig. 6C). For each aortic site, rank order for concentrations of sequestered aortic undegraded LDL were three pool > two pool, whereas rank order for concentrations of total aortic undegraded LDL were three pool > two pool > one pool. Concentrations of metabolically active LDL did not differ between the aortic sites for models with two or three compartments of aortic undegraded LDL. However, concentrations of metabolically active LDL were increased 143% (P < 0.00002) for model 1A, where metabolically active aortic undegraded LDL constituted all of the aortic undegraded LDL (Fig. 6B).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Concentrations of undegraded LDL for abdominal aorta of normal rabbits. A: sequestered aortic undegraded LDL (sum of pools 2 and 3). B: metabolically active LDL (pool 1). C: total aortic undegraded LDL (sum of pools 1, 2, and 3). Solid bars, branch sites; cross- hatched bars, uniform abdominal aorta. 1A, 2A, 2B, 3A, and 3B represent aortic undegraded LDL concentrations determined from 1-pool (model 1A), 2-pool (models 2A and 2B), and 3-pool (models 3A and 3B) compartmental models, respectively. Data from 33 rabbits were used to determine aortic undegraded LDL concentrations. *P < 0.001, **P < 0.0002, and ***P < 0.00005 for comparison between branch sites and uniform abdominal aorta.

LDL mean residence time within abdominal aorta. Figure 7 shows LDL mean residence times within branch sites and the uniform abdominal aorta. Mean residence times as determined by the different methods differed by up to a factor of two, with rank-order stochastic method > three-pool model > two-pool model > one-pool model for both the branch sites and the uniform abdominal aorta. Despite these differences in absolute values for mean residence times as determined by different methods, all methods indicated a similar prolongation (47-67%) of mean residence time in the abdominal branch sites. These differences were significant for model 1A (P < 0.02), for model 3B (P < 0.05) that closely fitted aortic data for undegraded LDL, and for the stochastic method (P < 0.05).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Mean residence time of LDL within abdominal aorta of normal rabbits. Solid bars, branch sites; cross-hatched bars, uniform abdominal aorta. 1A, 2A, 2B, 3A, and 3B represent mean residence time for LDL within aorta determined from 1-pool (model 1A), 2-pool (models 2A and 2B), and 3-pool (models 3A and 3B) compartmental models (n = 33), respectively. SM, mean residence times determined by the stochastic method (n = 29; n = 4 for permeability data and n = 25 for time dependence of aortic undegraded LDL data), as described in MATERIALS AND METHODS. Mean residence times are shown ± SE. *P < 0.05 and **P < 0.02 for comparison between branch sites and uniform abdominal aorta.

Absolute rates of LDL degradation. Figure 8 shows absolute rates (nmol LDL cholesterol · h¹ · g aorta¹) of entry, degradation, and efflux of LDL for the abdominal aorta determined by the compartmental models. Consistent with similar values for fractional rates for total transport of LDL into the aorta (k₁₅ + k₂₅ + k₃₅, Table 2), absolute transport of LDL into a given aortic site as determined by all compartmental models was similar (Fig. 8A). Absolute transport of LDL into abdominal branch sites as determined by models 1A and 3B were 63% (P < 0.0005) and 42% (P < 0.025) greater, respectively, than for the uniform abdominal aorta. Importantly, absolute rates of degradation of LDL were remarkably similar for all compartmental models and 86-88% (P < 0.00002) greater for branch sites than the uniform abdominal aorta (Fig. 8B), even though fractional rate constants for LDL degradation differed among the models (Table 2). Absolute rates of LDL efflux as determined by the different compartmental models were also quite similar for a given aortic site (Fig. 8C). Both the three-pool and two-pool models indicated no difference in absolute rates of efflux between the branch and uniform abdominal aorta. All compartmental models indicated that most of the LDL entering each aortic site was lost by efflux (compare Fig. 8, A and C).

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Absolute rates of transport of LDL into the aorta, LDL degradation within the aorta, and efflux of LDL from the aorta. Solid bars, branch sites; cross-hatched bars, uniform abdominal aorta. Absolute rates of LDL transport are given ± SE. 1A, 2A, 2B, 3A, and 3B represent rates of transport or metabolism determined from 1 pool (model 1A), 2-pool (models 2A and 2B), and 3-pool (models 3A and 3B) models of aortic undegraded LDL, respectively. A: total transport of LDL into the aorta (in steady state this is equivalent to total loss of LDL from the aorta by efflux + degradation). B: loss of undegraded LDL from the aorta by cellular degradation of LDL. C: loss of undegraded LDL from the aorta by efflux (sum of efflux from all aortic pools). *P < 0.025, **P < 0.0005, and ***P < 0.00002 for comparison between branch sites and uniform abdominal aorta.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The goal of this study was to obtain metabolic evidence that would either support or refute the hypothesis that some undegraded LDL within the aorta of normal rabbits is sequestered in a form that does not readily exchange with plasma LDL. I also considered whether more LDL might be sequestered in the atherosclerosis-susceptible abdominal aorta compared with atherosclerosis-resistant abdominal aorta and whether LDL mean residence time would be increased at these sites. To accomplish these aims, techniques that allowed tracing aortic undegraded LDL and aortic LDL degradation products were combined with a mathematical analysis. This approach allowed sequential investigation of a model that assumed all aortic undegraded LDL was homogeneous, followed by models that included aortic undegraded LDL in both metabolically active and sequestered forms.

The principle findings of the study were as follows. For the first time, it was possible to show metabolic evidence that for normal rabbits, some LDL in the abdominal aorta is sequestered in a form that is not available for cellular degradation and that exchanges slowly with plasma LDL (Table 1, Fig. 4). Second, most undegraded LDL within each aortic site was sequestered (Fig. 6). Third, undegraded LDL concentrations calculated by all compartmental models were consistently larger for branch sites (Fig. 6). Fourth, LDL mean residence times within the aorta calculated by compartmental models both with and without sequestered LDL were consistently larger for branch sites than the uniform abdominal aorta (Fig. 7). The stochastic method, which is independent of any assumptions regarding transport or metabolism of LDL within the aorta, also indicated a significantly longer mean residence time for LDL in the abdominal branch sites (Fig. 7).

Influence of the nature of the compartmental model considered on calculated parameters of aortic LDL transport and metabolism. For each aortic site, both arterial permeability to LDL (Fig. 5) and mass transport of LDL into the aorta (Fig. 8) were similar irrespective of the compartmental model that was considered. However, as one might expect due to differences in the structure of the compartmental models (Fig. 1), individual fractional rate constants for LDL entry into the various aortic pools, and for transport and metabolism of LDL within the aortic pools, differed among the models (Table 2).

Evidence for sequestration of LDL within the aorta. If the hypothesis that some aortic LDL is sequestered is correct, one would expect to find 1) evidence for two or more aortic pools of undegraded LDL and 2) that the fractional rate constant for total loss of LDL via the sequestered pathway might be less than that via the metabolically active pathway. Sequestered LDL could be readily detected within both the branch sites and the uniform abdominal aorta (Table 1, Fig. 4). As shown, this study provides metabolic evidence that LDL is present within both the branch sites and the uniform abdominal aorta in at least three forms, of which two are sequestered. For both aortic sites, and for both model 2B and 3B, fractional rate constants for LDL total loss via (each of) the sequestered pathway(s) (k₀₂, k₀₃) were only 1/6 to 1/100 that for total loss via the metabolically active pathway (model 2B, k₀₁ + k₄₁, model 3B, k₄₁; Fig. 1, Table 2).

Interpretation and relevance of sequestered LDL. Previous studies (2, 9) showed that arterial proteoglycans can bind LDL in vitro. Atherosclerotic lesions both contain increased total concentrations of proteoglycans (42, 45, 52) and higher concentrations of proteoglycans with high affinity for LDL (48). Even in macroscopically normal human arteries, the affinity of arterial glycosaminoglycans for LDL correlated with susceptibility to atherosclerosis (6). The binding properties of proteoglycans from branch sites and the uniform abdominal aorta of normal rabbits remain to be determined. However, preliminary data indicate that characteristics of proteoglycans may differ between branch sites and the uniform abdominal aorta and that concentrations of proteoglycans may be increased in the abdominal branch sites of normal rabbits (38). My working hypothesis, which will need to be tested by further experiments, is that increased concentrations of proteoglycans with increased affinity for LDL account for sequestration of increased concentrations of undegraded LDL and prolonged residence of undegraded LDL within the branch sites.

Aortic LDL retention. A previous study (37) used the same one-pool model as this paper (model 1A, Fig. 1, left) to calculate LDL mean residence times for branch sites and the uniform abdominal aorta. Mean residence times reported in that study for normal rabbits were ~3 h (37). A later study that used a stochastic method to calculate LDL mean residence times for normal rabbits reported values of 9.28 ± 0.57 and 7.09 ± 0.49 h for the branch and uniform abdominal aorta, respectively (46). LDL mean residence times determined in both of those studies (37, 46) are shorter than those determined in the present study (Fig. 7). Also, in contrast to the present results, neither of the earlier studies found LDL mean residence time to be prolonged in the branch sites of normal rabbits (37, 46).

Limitations of these studies. In these studies, the metabolic heterogeneity of aortic LDL and the lack of participation of some of the aortic LDL pools in aortic LDL degradation were considered to reflect extracellular sequestration of LDL. However, one could propose that either one or both of the metabolic pathways denoted as sequestered pathway(s) might reflect retroendocytosis. However, a previous study (15) with cultured cells suggests that the fractional rate constant for retroendocytosis of LDL would be greater than 1.0 h¹, considerably greater than the fractional rate constants for efflux of LDL from sequestered pools of both model 2B and 3B (k₀₂ and k₀₃; Table 2). Thus it is unlikely that LDL in sequestered pools reflects that destined to be exocytosed.