The risk for coronary heart disease (CHD), the number one cause of death in Western societies, is increased in individuals with elevated concentrations of plasma LDL cholesterol.1 National Cholesterol Education Program (NCEP) and European Atherosclerosis Society (EAS) guidelines for prevention of CHD emphasize the importance of reducing LDL cholesterol levels in patients at risk for CHD as well as in the general population.123 If dietary and lifestyle changes are unsuccessful in bringing plasma cholesterol concentrations down to acceptable ranges, drug therapy is often warranted. Although several types of lipid-regulating drugs are available, combination drug therapy may be necessary in some patients to achieve target plasma cholesterol levels.4 3-Hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors effectively reduce plasma cholesterol levels in patients with hypercholesterolemia.56 These drugs decrease cholesterol synthesis by competitively inhibiting HMG-CoA reductase, the enzyme that catalyzes the rate-limiting step in the cholesterol biosynthetic pathway. Doses of 20 mg/d lovastatin or pravastatin or 10 mg/d simvastatin generally reduce plasma LDL cholesterol levels by about 20% to 30%.5789 Higher doses of these drugs can reduce LDL cholesterol levels by as much as 40%.5 Treatment with HMG-CoA reductase inhibitors also produces increases of about 5% to 10% in plasma HDL cholesterol and reductions of about 10% to 20% in triglycerides.5 Millions of patients have taken HMG-CoA reductase inhibitors to lower plasma cholesterol levels over the past 10 years. Atorvastatin, a recently synthesized member of the HMG-CoA reductase inhibitor class of lipid-modifying drugs, is currently being evaluated in clinical trials. Atorvastatin is a chiral, calcium salt of a pentasubstituted pyrrole.10 In laboratory animals, atorvastatin effectively lowers plasma LDL cholesterol as well as VLDL cholesterol and triglyceride levels.11 Acute- and multiple-dose (13-week) toxicology evaluations indicate that atorvastatin has an acceptable margin of safety between doses causing little or no toxicity at the anticipated human dose.12 In early clinical dose-ranging studies with healthy human volunteers, atorvastatin in a single dose or 2-week multiple doses was well tolerated.13 The dose-ranging study reported here is the first study of atorvastatin in patients with primary hypercholesterolemia. MethodsPatientsOutpatients with elevated LDL cholesterol (>4.14 but <6.21 mmol/L) and normal levels of triglycerides (<3.39 mmol/L) entered an 8-week, dietary, placebo-baseline phase at one Canadian and five US centers. Eligible patients were aged 18 to 70 years with a body mass index ≤30 kg/m2. Patients were ineligible if they had uncontrolled hypertension (diastolic blood pressure >95 mm Hg), diabetes mellitus and/or other metabolic endocrine disease, active liver disease, or hepatic or renal dysfunction. Women of childbearing potential were also ineligible. Study participants could consume no more than 14 oz/wk of ethanol equivalents and could not concurrently take drugs known to affect lipid levels or known to interact with the study medication. Informed ConsentIdentical protocols were submitted to and approved by an institutional review board for each center. Prior to entering the study, each patient provided witnessed, written informed consent. Dietary Counseling and MonitoringUpon entering the baseline phase and continuing throughout the double-blind phase of the study, patients were counseled on the use of the National Institutes of Health (NIH) NCEP Step 1 Diet.1 This diet limits dietary cholesterol to <300 mg/d and total fats to <30% of total calories, with <10% of total calories from saturated fats, 10% from polyunsaturated fats, and 10% to 15% from monounsaturated fats. During the week before selected clinic visits, patients recorded their daily food and drink intakes in a diary for 3 consecutive days. Food record rating (FRR) scores14 were determined from these patient diaries at weeks −8, −2, 2, and 6 to evaluate dietary compliance. Analysis of the average American diet yields an FRR score of >20, whereas a score of <10 is expected for Step 1 diets. The Chicago Center for Clinical Research, Chicago, Ill, coordinated the dietary aspects of the study. At weeks −2 and 6, the Chicago Center performed a Nutritional Data System dietary constituent analysis using information from the 3-day dietary diary. Baseline PhaseUpon entering the placebo-baseline phase, patients were instructed to take two placebo capsules once daily at bedtime. Each patient’s plasma lipid profile was determined at weeks −4, −2, and −1. To qualify for randomization into the double-blind period at week 0, patients had to have LDL cholesterol values >4.14 and <5.69 mmol/L at both weeks −2 and −1, with the lower value within 15% of the higher value; triglyceride values <3.39 mmol/L at both visits; and an FRR score <15 at week −2 or −1. Double-blind Treatment PhaseAt the end of the 8-week, placebo-baseline phase, eligible patients were randomly assigned to receive placebo or 2.5-, 5-, 10-, 20-, 40-, or 80-mg doses of atorvastatin once daily for 6 weeks. Patients were assigned to treatment according to a randomization code prepared by the Parke-Davis Biometrics Department. Patients received either one bottle containing 2.5-, 5-, 10-, 20-, or 40-mg atorvastatin capsules and one bottle with matching placebo capsules; two bottles with atorvastatin 40-mg capsules; or two bottles with placebo capsules. Patients were instructed to take one capsule from each bottle once a day at bedtime. The appearance of the capsules did not change throughout the baseline and double-blind phases of the study. Compliance with the study medication was judged by a capsule count at each clinic visit. During the active treatment phase, both patients and investigators were blinded to the study medication and to plasma lipid concentrations. Clinic visits took place at 2-week intervals during the double-blind phase, and patient lipid profiles were determined at each visit. Patients prepared dietary diaries for visits at weeks 2 and 6. Complete clinical laboratory determinations were done at the randomization visit and at the final visit at week 6. To monitor safety, clinical laboratory tests (aspartate aminotransferase [AST], alanine aminotransferase [ALT], creatine phosphokinase [CPK], alkaline phosphatase, and total bilirubin) were performed at every visit. Because there was no extension to the treatment period, investigators could return patients to their standard therapy at study completion. Laboratory AnalysesUsing standardized procedures, Medical Research Laboratories, Cincinnati, Ohio, performed lipid and clinical laboratory measurements for all sites. This laboratory was certified for standardization of lipid analyses during this study, as specified by the Standardization Program of the Centers for Disease Control and Prevention and the National Heart, Lung, and Blood Institute.15 After the patients fasted overnight for a minimum of 12 hours, blood was drawn for lipid profiles and collected in evacuated tubes (Vacutainers) containing EDTA. Total plasma cholesterol and triglycerides were determined enzymatically with the Hitachi 737 analyzer.16 Plasma HDL cholesterol was determined enzymatically after LDL cholesterol and VLDL cholesterol were selectively removed from the plasma sample by heparin and magnesium chloride precipitation.17 LDL cholesterol concentration was estimated by the Friedewald formula (LDL cholesterol=total cholesterol−HDL cholesterol−triglycerides/5).18 Direct LDL cholesterol was determined by ultracentrifugation (β-quantification) at weeks −1 and 6.19 Long-term precision was monitored by using a stabilized plasma pool, and the coefficient of variation (CV) was 2.5% during this study. Apo A-I and apo B were determined at weeks −1 and 6 by fixed-rate nephelometry.2021 The precision for both assays was measured by using two frozen serum pools. For apo A-I the between-run CVs were approximately 3.4% and 3.2% for the high and low pools, respectively. For apo B, the CVs were 3.2% and 2.6%, respectively. Lp(a) was qualitatively assessed at weeks −1 and 6 by competitive enzyme-linked immunosorbent assay,22 and long-term precision was monitored with two frozen serum pools. The CV of the low pool was 7%, whereas the high pool had a CV of 8.5%. Safety EvaluationBefore entering the placebo-baseline phase, patients received a complete physical examination and clinical laboratory evaluation. At each visit, patients were asked about their health status and adverse events. Each patient’s blood pressure and weight were determined, and clinical laboratory data were evaluated. Data AnalysisThe sample size for this study was chosen to detect a significant linear dose effect for a 25% difference between the mean percent changes of placebo and the highest dose of atorvastatin. Statistical analyses were performed with the SAS statistical package.23 Analyses included data from all randomized patients, with at least one baseline and one double-blind measurement of the parameter of interest regardless of patient compliance with the protocol. ANOVA was used to evaluate the effect of atorvastatin on the percent change from baseline in LDL cholesterol, the primary efficacy parameter. Baseline was defined as the mean of each patient’s LDL cholesterol values at weeks −2, −1, and 0, with the analysis of percent change from baseline being performed at the last visit of the double-blind period. On the basis of this model, a sequential, “step-down” trend test was performed to determine the significance of the drug effect. Dunnett’s test was used to compare each atorvastatin dose group to placebo when the percent change was not monotonic across the dose levels. An additional model for LDL cholesterol by ANCOVA added baseline LDL cholesterol as a covariate and tested the interaction of the treatment group and covariate. All analyses were done using a two-sided significance level of 5%. The same analyses were performed for secondary efficacy parameters except that the baseline value for apo A-I, apo B, and Lp(a) was the mean of measurements at weeks −1 and 0 and that for LDL cholesterol, as measured by β-quantification, was the measurement at week −1. ResultsPatient Characteristics and DispositionAlthough overall patient characteristics were similar for the seven treatment groups, the small number of patients randomized to each treatment group resulted in some differences. Median patient body mass index was 26 kg/m2, with a range of treatment group median values from 25 to 28 kg/m2. Treatment group median age ranged from 49 to 63 years with an overall patient median age of 54 years. Neither the placebo nor the 40-mg treatment group included patients aged 65 years or older compared with two or three in each of the other treatment groups. Sixteen of the 81 patients were women. All treatment groups included two or three women except the 80-mg treatment group that contained only men. Most patients (95%) in the study were white. Mean baseline LDL and total cholesterol levels were 4.86 and 6.98 mmol/L, respectively. Of the 81 patients randomized to double-blind treatment, 78 (96%) completed the study. One patient who was receiving 2.5 mg atorvastatin withdrew on day 13 due to indigestion and flu (not thought by the investigator to be related to the study drug), and two patients were withdrawn after 2 days because they had been incorrectly entered. The efficacy analyses included data from the 79 patients with double-blind data. No patients were excluded from the efficacy evaluations because of protocol variations. Most patients (97%) were judged to have been compliant with the study medication, as determined by capsule counts at clinic visits. Dietary compliance as judged by FRR scores was acceptable. Treatment group FRR scores were relatively constant from week −2 to week 6 except in the 20-mg atorvastatin treatment group, whose mean FRR score increased from 7.3 at week −2 to 12.1 at week 6. Plasma Lipids, Lipoproteins, and ApolipoproteinsMean percent reductions from baseline in LDL cholesterol increased with increasing doses of atorvastatin. Patients treated with 2.5 mg atorvastatin had mean reductions of 25%; those treated with 80 mg atorvastatin had mean reductions of 61% (Table 1). Approximately 90% of the maximum reduction in plasma LDL cholesterol levels was achieved by week 2 of the double-blind phase in atorvastatin-treated patients (Figure). Values of LDL cholesterol as estimated by the Friedewald formula were generally in agreement with those determined by ultracentrifugation. At the last visit of the double-blind phase, mean percent changes from baseline in LDL cholesterol for patients treated with 2.5, 5, 10, 20, 40, or 80 mg atorvastatin or placebo were −22%, −28%, −37%, −45%, −48%, −59%, and 3%, respectively, as determined by ultracentrifugation, compared with −25%, −29%, −41%, −44%, −50%, −61%, and 8%, respectively, as estimated with the Friedewald formula. Atorvastatin-treated patients had dose-related reductions from baseline in total plasma cholesterol and apo B (Table 1). Patients treated with 2.5 and 80 mg atorvastatin had reductions in total cholesterol of 17% and 46%, respectively, and reductions in apo B of 17% and 50%, respectively. Atorvastatin reduced plasma triglyceride concentrations at every dose level, but without any consistent dose trend. Reductions in triglycerides from baseline values were 25% or greater for patients treated with 5, 20, 40, and 80 mg atorvastatin. There was no consistent pattern in the percent changes from baseline for HDL cholesterol, apo A-I, or Lp(a). SafetyThe study treatments were well tolerated. All patients who entered the double-blind phase were included in the safety evaluations. Thirty-four (42%) of the 81 patients (4 [33%] on placebo; 30 [43%] on atorvastatin) had adverse events. Of these patients, 85% reported adverse events of mild or moderate intensity. For patients treated with atorvastatin, the most frequent adverse event was the common cold (5.8%), followed by headache (4.3%) (Table 2). There was no consistent atorvastatin dose relationship for adverse events. One patient who was receiving 2.5 mg atorvastatin had a serious adverse event (broken ankle) that was not considered drug related. In addition, one patient who was also in the 2.5-mg treatment group withdrew from the study at week 2 due to adverse events (mild indigestion and moderate flu symptoms) not attributed to the study drug. Clinically significant changes in laboratory parameters were not dose related. One patient in each of the atorvastatin-treatment groups had bilirubin values one to two times the upper limit of the reference range (1.1 mg/dL) at week 4 or 6; prior to randomization, three of these patients had levels in this elevated range. No patients had clinically significant elevations of CPK. One patient treated with 40 mg atorvastatin had elevations of three to four times the upper limit of the reference range (22 U/L for AST; 25 U/L for ALT) in AST and ALT values that returned to normal 2 to 3 weeks after the end of the study. There was a dose-related increase in the number of patients with mild elevations of ALT and AST. At week 6, ALT elevations of one to two times the upper limit of the reference range were observed for 1 patient in both the 5- and 10-mg treatment groups, for 3 in the 20-mg treatment group, for 4 in the 40-mg treatment group, and for 6 in the 80-mg group. While fewer patients had AST elevations, the trend was similar. No other dose-related changes in laboratory parameters were seen. DiscussionAtorvastatin has a rapid onset of action; approximately 90% of the LDL cholesterol reduction from baseline occurred within the first 2 weeks of treatment (Figure). Increasing doses of atorvastatin produced progressive increases in efficacy. Reductions of 25% to 61% in LDL cholesterol were achieved with once-daily doses of 2.5 to 80 mg atorvastatin. In addition to LDL cholesterol, apo B, the major protein component of LDL cholesterol, was reduced from baseline by 18% to 51% in a dose-related manner. As with other drugs in this class, there were no clinically important changes in HDL cholesterol, apo A-I, or Lp(a). Although the patients who were selected for this study had elevated LDL cholesterol levels, most patients had normal plasma triglyceride concentrations (<3.39 mmol/L) at randomization. Triglycerides were reduced from baseline by 9% to 32%, but with no apparent dose trend. Atorvastatin was well tolerated in this study. Of 81 patients randomized to treatment, 78 completed the study. No drug-related serious adverse events were reported. There was a dose-related increase in the number of patients with mild elevations (one to two times the reference range) of AST and ALT. Similar transaminase elevations have been reported with other HMG-CoA reductase inhibitors5 and lipid-lowering drugs such as cholestyramine24 and may result from changes in hepatic lipid metabolism. In this study, one atorvastatin-treated patient had clinically significant AST and ALT elevations of three to four times the reference range, but these values returned to normal at study follow-up. The recently released NCEP II recommendations include reducing LDL cholesterol to <3.36 or <2.59 mmol/L for patients with elevated LDL cholesterol and two other risk factors or with elevated LDL cholesterol and preexisting CHD, respectively. Adequate treatment often requires combination therapy, with the associated risk of increased side effects. In this clinical trial, greater reductions from baseline in LDL cholesterol levels were observed in atorvastatin-treated patients than have been previously reported in patients treated with other lipid-regulating drugs.25 At the end of this study (week 6), patients treated with 80 mg atorvastatin achieved a 61% mean reduction from baseline in LDL cholesterol. Although a 50% to 60% reduction in LDL cholesterol can be achieved by combining several lipid-modifying drugs,25 no single agent has been reported to produce this result. This study suggests that atorvastatin, with its enhanced efficacy, may provide adequate therapy for a large number of dyslipidemic patients, including those previously treated with multiple therapies. Reprint requests to Donald M. Black, MD, Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Co, 2800 Plymouth Rd, Ann Arbor, MI 48105.  Figure 1. Line plot of reductions in LDL cholesterol by atorvastatin as a function of dose and time.
This study was supported by a research grant from Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Co, Ann Arbor, Mich. The authors would like to thank David Canter, MD, Evan Stein, MD, PhD, Cynthia Sevilla, PhD, and Linda Shurzinske, MS, for their assistance with the design, performance, and reporting of this trial. References
Page 2Tangier disease is a rare, autosomal recessive disorder of cellular lipid and lipoprotein metabolism. Clinically, it is characterized by severe reduction of serum HDL and the major apolipoproteins (apo) of HDL, apo A-I and apo A-II. Concomitant with the reduction of HDL plasma levels, there is cholesteryl ester deposition in various tissues, particularly in the reticuloendothelial system.123 Most Tangier patients also have decreased levels of LDL and low total serum cholesterol, whereas serum triglycerides are elevated in many cases. The diagnosis of Tangier disease is supported by the combination of the lipoprotein abnormalities described above with hyperplastic yellow-orange tonsils and hepatosplenomegaly.3 Whether there is an increased cardiovascular risk in Tangier disease is still a matter of debate.34 However, patients with classic Tangier disease with no additional risk factors such as the patients reported in this article do not show signs of premature atherosclerosis.5 Metabolic studies in Tangier patients showed that the reduced levels of HDL, apo A-I (<1% of normal), and apo A-II (5% to 10% of normal)6 are due to rapid catabolism of HDL and its apolipoproteins in Tangier patients, whereas synthetic rates are within the normal range.7 Structural defects of apo A-I and apo A-II have been excluded as the cause for hypercatabolism of HDL.489 These data suggest an abnormality in the interaction of cells with HDL leading to hypercatabolism. A key experiment showed that Tangier mononuclear phagocytes (MNPs) degrade internalized HDL completely in lysosomes, rather than resecrete the internalized HDL particles as observed in control MNPs.101112 This was the first evidence for an abnormality of cellular lipid metabolism in Tangier disease. Further analysis of cellular lipid metabolism showed that Tangier MNPs have increased rates of synthesis for phospholipids, triglycerides, and cholesteryl esters compared with normal.13 At the same time, catabolism of cellular phospholipids is enhanced, whereas the catabolism of triglycerides and cholesteryl esters is normal. This may account for the observed lipid storage in these cells. The biochemical abnormalities of Tangier MNPs are accompanied by distinct morphological abnormalities. These affect mainly the Golgi apparatus and the lysosomal compartment and are found in MNPs as well as in fibroblasts.14 Tangier MNPs erroneously target HDL to a lysosomal compartment. If the genetic defect specifically disturbs retroendocytosis of HDL, one would expect that Tangier fibroblasts, which do not internalize HDL,15 would interact normally with HDL. Interaction of normal skin fibroblasts with HDL results in a net cholesterol efflux if the cells have not been cholesterol depleted. This effect is apparently related to a pathway involving activation of protein kinase C (PKC). Evidence exists that PKC mediates translocation of cholesterol to the cell membrane after specific binding of HDL3 to the cell membrane.16 HDL thereby increases the amount of newly synthesized sterol in the membrane, which will increase the effective concentration gradient for desorption. This effect of HDL appears to be mediated by its protein moiety.16 The same investigators showed that PKC activators such as 1,2-diacylglycerol and phorbol myristate induce translocation of intracellular cholesterol to the plasma membrane and cholesterol efflux, whereas inhibition of PKC by sphingosine reduces cholesterol efflux.16 In the present investigation, HDL3-mediated efflux of cholesterol from Tangier fibroblasts was analyzed. The experiments were designed to study efflux from different cellular cholesterol pools to determine whether specific transport routes are affected in Tangier disease. MethodsPatientsCutaneous fibroblasts were obtained from two patients homozygous for Tangier disease: Patient 1 (E.G.) was a 60-year-old woman (triglycerides, 2.94 to 4.89 mmol/L; cholesterol, 2.02 to 2.67 mmol/L); patient 2 (J.S.) was a 57-year-old man, brother of patient 1 (triglycerides, 1.58 to 2.24 mmol/L; cholesterol, 1.16 to 1.50 mmol/L). Niemann-Pick type C (NPC) fibroblasts were obtained from an 11-year-old male patient (D.O.). Four lines of control fibroblasts (G.M., T.L., N.F., R.W.) were cultured from the cutis of normolipemic individuals who underwent abdominal surgery. MaterialsCell culture media were obtained from Gibco-BRL. [14C]Cholesterol (51 mCi/mmol), [14C]mevalonolactone (54.1 mCi/mmol), and [3H]cholesteryl linoleate (71.4 Ci/mmol) were purchased from NEN. All other chemicals and solvents were from Merck. All other biochemicals, including antibodies, were from Sigma. Cell CultureFibroblasts were cultured according to standard conditions in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with l-glutamine, nonessential amino acids, and 10% fetal calf serum in a humidified 5% CO2 atmosphere at 37°C. All fibroblast cultures were used between passages 5 and 15 and cultured for 7 days after splitting 1:2 to ensure that they were confluent for at least 2 days. LipoproteinsHuman LDL (d=1.006 to 1.063 g/mL), HDL3 (d=1.125 to 1.21 g/mL), and lipoprotein-deficient serum (LPDS, d>1.23 g/mL) were isolated from serum of individual normolipemic volunteers by sequential ultracentrifugation in a Beckman L-70 ultracentrifuge equipped with a 70-Ti rotor at 4°C.17 Serum was prepared from recalcified plasma to prevent release of growth factors and cytokines by white blood cells into the serum during clotting. The lipoprotein fractions were extensively dialyzed against a buffer containing 0.15 mol/L NaCl and 5 mmol/L Na2EDTA (pH 7.4) at 4°C. The final dialysis step was performed against a 0.15 mol/L NaCl without EDTA. Determination of HDL3-Binding to Cultured FibroblastsBinding of HDL3 to fibroblasts was determined by use of 125I-labeled HDL3 as described previously.11 Binding was performed at 4°C in DMEM containing 1 mg/mL bovine serum albumin and the labeled ligand. After the incubation period, cells were washed five times and lysed by addition of 600 μL of 0.3 mol/L NaOH. Cell-associated radioactivity was determined in an LKB-Pharmacia gamma-counter. Flow Cytometric Determination of Uptake of DiI-Labeled LipoproteinsLabeling of lipoproteins with the fluorescent dye 1,1′-dioctadecyl-3,3,3′3′-tetramethyl-indocarbocyanine perchlorate [DiI(3)-C18] was carried out as described earlier.18 Fibroblasts were incubated with DiI-labeled HDL3 for 2 hours at 37°C. DiI-lipoprotein labeled cells were analyzed by flow cytometry in a Becton Dickinson FACScan. Cellular accumulation of DiI was measured at 580 nm in samples containing at least 10 000 cells. Autofluorescence of unlabeled cells was subtracted. Labeling of LDL With [3H]Cholesteryl LinoleateAliquots of human LDL (1.9 mg protein) were lyophilized in the presence of potato starch (ratio of starch to LDL protein, 12:1 wt/wt) in Siliclad-treated glass tubes. Neutral lipids were removed by two extractions with 5 mL heptane at −18°C. The heptane-extracted LDL was then mixed with 200 μL heptane containing 6 mg (9.2 μmol, 51 μCi) [3H]cholesteryl linoleate and kept at −18°C for 2 hours with intermittent vortexing.19 Finally, heptane was evaporated, and the radiolabeled LDL was resuspended in 1 mL of 10 mmol/L tricine buffer, pH 8.4, for 36 hours at 4°C. Soluble reconstituted LDL was separated from the potato starch by centrifugation at 2000 rpm for 10 minutes at 4°C. Metabolic Labeling of FibroblastsAfter reaching confluence for at least 2 days, fibroblasts were rinsed and then incubated with DMEM containing 10% LPDS for 48 hours to deplete the cells of cholesterol. Thereafter, cells were rinsed again and incubated for 3 hours with either 0.5 μCi/mL [14C]cholesterol to label membrane cholesterol or 2.0 μCi/mL [14C]mevalonolactone to label newly synthesized cholesterol. Incubation was performed at 15°C to minimize intracellular cholesterol transport. For homogeneous labeling of cellular cholesterol, fibroblasts were incubated without prior incubation in LPDS in the presence of 10% fetal calf serum with 2.0 μCi/mL [14C]cholesterol for 48 hours at 37°C.20 To label lysosomal cholesterol, cells were incubated with reconstituted LDL (0.29 μCi/mL containing 52 nmol cholesteryl linoleate) for 3 hours at 37°C.21 After incubation with the radioactive tracer, the medium was removed and cells were rinsed five times with 3 mL phosphate-buffered saline (PBS). Cells from three dishes were harvested to determine mevalonate uptake and cholesterol synthesis in cells pulsed with [14C]mevalonolactone or cholesterol incorporation in cells pulsed with [14C]cholesterol or reconstituted LDL. Determination of Cholesterol Efflux From FibroblastsTo determine sterol efflux, cells were incubated in DMEM containing 1% BSA supplemented with increasing concentrations of HDL3 as indicated. Aliquots of the medium were taken at the time points specified. Radioactivity in the medium was determined by liquid scintillation counting. Specific HDL3-mediated efflux is defined as the difference between efflux in the presence of HDL3 and 1% BSA minus the efflux in the presence of 1% BSA only. After 24 hours, the cells were carefully rinsed three times with PBS, harvested, and resuspended in 1 mL PBS. The cell suspension was centrifuged at 10 000 rpm for 10 minutes, the supernatant was removed, and the cells were resuspended in 800 μL PBS. Finally, cells were sonicated on ice for 15 seconds with a Branson sonifier. Aliquots were taken for protein determination and lipid extraction. Activation of PKCActivation of PKC during the efflux experiments was achieved by incubation of cells in the presence of 10−5 mol/L of the membrane-permeable 1,2-dioctanoylglycerol (1,2-DOG) as described by Mendez et al.16 This was added after the 3-hour labeling period and again after 4 and 8 hours of incubation. Determination of Cellular LipidsLipid extractions were performed according to the method of Bligh and Dyer.22 Cellular lipids were separated by high-performance thin-layer chromatography (HPTLC) with cholesteryl formate (Sigma) used as an internal standard.23 Samples were dissolved in 30 μL of the solvent used for chromatography. External standards containing free cholesterol, cholesteryl stearate, and cholesteryl formate and the samples were applied to 10×20-cm silica gel HPTLC plates (Merck) with a capillary dispenser (Camag). Separation conditions for neutral lipids have been described previously.23 HPTLC plates were developed with manganese chloride/sulfuric acid. Quantification of cellular free cholesterol and cholesteryl esters was carried out by scanning the plates with a fluorescence scanner (Camag). The amount of radioactivity in specific cellular lipids was determined by scraping the respective spots from the HPTLC plates, solubilization in scintillation liquid, and counting in a beta-counter. Protein DeterminationProtein was determined according to the method described by Smith et al.24 ResultsAnalysis of HDL3 Binding and Uptake by FibroblastsBinding of HDL3 to cultured fibroblasts was determined by use of 125I-labeled HDL3. Tangier fibroblasts have a slightly higher specific binding of HDL3 than control fibroblasts (Bmax, 70 versus 52 ng/mg cell protein) and a similar affinity (Kd, 8.8 versus 10.6 μg/mL) (Fig 1). This indicates that HDL binding capacity and affinity of Tangier fibroblasts are not reduced. Uptake of HDL3 was determined with DiI-labeled HDL3. Neither by fluorescence flow cytometry nor by confocal laser scan microscopy could appreciable uptake of DiI-labeled HDL3 into Tangier or control fibroblasts be demonstrated. Hepatocytes, which were used as control, internalized HDL3 as expected (data not shown). This confirms previous data on fibroblasts1516 and indicates that cholesterol efflux from control and Tangier fibroblasts must take place at the cell surface. Cholesterol Efflux After Homogeneous Labeling of FibroblastsFibroblasts were labeled by incubation with [14C]cholesterol for 48 hours at 37°C. Under these conditions, cellular cholesterol pools are homogeneously labeled.25 Efflux to BSA increased from 3.14±0.29% of uptake after 1 hour to 5.84±0.39% after 8 hours in Tangier cells and from 3.28±0.46% of uptake after 1 hour to 6.44±0.36% after 8 hours in control fibroblasts, with no significant difference between the two cell types. With 200 μg/mL HDL3 in the medium, HDL3-specific efflux increased from 2.63±1.27% of uptake after 1 hour to 12.34±1.59% after 8 hours in Tangier fibroblasts. In control fibroblasts, there was an increase from 3.43±0.36% after 1 hour to 30.65±2.21% after 8 hours. Specific HDL3-mediated efflux of cellular cholesterol from Tangier fibroblasts was only about 50% compared with controls for almost all concentrations and time points (Fig 2). Efflux of Plasma Membrane Cholesterol From FibroblastsThe plasma membrane cholesterol pool was labeled with [14C]cholesterol for 3 hours at 15°C, and cholesterol efflux was measured by incubation of the cells with increasing amounts of HDL3. Under these conditions, a concentration- and time-dependent efflux of cholesterol from the cell membrane occurred (Fig 3). HDL3-mediated cholesterol efflux of Tangier fibroblasts was similar to that of control fibroblasts for all concentrations. This indicates that membrane desorption of cholesterol is not disturbed in Tangier disease and does not account for the reduced efflux of cholesterol. Efflux of LDL-Derived Cholesterol From FibroblastsEfflux of LDL-derived cholesterol was measured in control fibroblasts, Tangier fibroblasts, and as an internal control in NPC fibroblasts, which are known to have a defect in the release of lysosomal cholesterol. Cells were labeled by incubation with reconstituted [3H]cholesteryl linoleate-LDL for 3 hours at 37°C. To exclude differences in the activity of the LDL-receptor pathway, uptake of labeled cholesterol and cholesteryl esters was measured. Total uptake after the pulse period was similar in control and Tangier fibroblasts, whereas it was significantly lower in NPC fibroblasts (Table 1). Incubation with different concentrations of HDL3 increased cholesterol efflux compared with 1% BSA in both control and Tangier fibroblasts in a concentration-dependent manner. Only during the first 4 hours was efflux from Tangier fibroblasts apparently lower than that from control fibroblasts. Thereafter, effluxes from control and Tangier fibroblasts were similar (Fig 4). As expected, in NPC fibroblasts, HDL3-mediated efflux was low. After 24 hours of chase, 32% of the cholesterol taken up into control cells was recovered in the medium. In Tangier cells, efflux was 31%, which is not significantly different from normal. Efflux from NPC fibroblasts was reduced significantly, to 17% of the total radioactivity taken up (Table 1). These data suggest that efflux of cholesterol incorporated into fibroblasts by uptake of reconstituted LDL for 3 hours is similar between control and Tangier fibroblasts. Efflux of Newly Synthesized Sterol From FibroblastsTo analyze efflux of newly synthesized sterols, cells were labeled with radioactive mevalonolactone. Since differences in sterol synthesis between control and Tangier fibroblasts might lead to apparent differences in sterol efflux, de novo cholesterol synthesis was determined after incubation for up to 24 hours with [14C]mevalonolactone (0.5 μCi/mL). There were no statistically significant differences between the two cell types. This indicates that uptake of mevalonolactone and synthesis of cholesterol are similar in the two cell types. When control fibroblasts were labeled at 15°C for 3 hours with [14C]mevalonolactone as precursor of endogenous sterol synthesis, HDL3 increased sterol efflux at 37°C in a concentration-dependent manner. Specific HDL3-mediated efflux was calculated as the increase over efflux in the presence of 1% BSA only. The medium was analyzed after 4 hours by lipid extraction and HPTLC to identify the radioactive sterols or sterol precursors. By this procedure, only late precursors after the lanosterol step cannot be separated from cholesterol. Of the unspecific efflux to BSA, 88±6% was water soluble. This could represent either nonmetabolized mevalonolactone or early nonsterol precursors of sterol synthesis. Total efflux to medium containing 100 μg/mL HDL3 was composed of 75±7% water-soluble cholesterol precursors (60±5% of uptake). The water-insoluble radioactive products all migrated in the cholesterol position, indicating that they were cholesterol or potentially late precursors of cholesterol, such as zymosterol and desmosterol. Specific HDL3-mediated efflux was completely accounted for by cholesterol (or cholesterol and desmosterol/zymosterol). This demonstrates that the reduction in specific HDL3-mediated efflux is due to newly synthesized cholesterol (or cholesterol and zymosterol/desmosterol) and not to any other precursors. In control fibroblasts, specific HDL3-mediated sterol efflux ranged between 3% and 18% for HDL3 concentrations between 10 and 100 μg/mL (Fig 5). Most of the HDL3-specific efflux occurred within the first 4 hours of incubation. In Tangier fibroblasts, the efflux to 1% BSA was not appreciably different from control fibroblasts (data not shown). Specific HDL3-mediated efflux of sterol was almost nonexistent (0% to 2.5%) during the whole incubation time in the cell cultures of both patients. In fact, there was no overlap between HDL3-specific efflux between the four control and two Tangier fibroblast cultures with 50 and 100 μg/mL of HDL3 (Fig 4). This indicates that the lack of HDL3-mediated efflux of newly synthesized sterols is specific for Tangier fibroblasts. Effect of PKC Stimulation on Efflux of Newly Synthesized CholesterolHDL3 has been shown to induce sterol translocation to the cell membrane by activation of PKC. In parallel experiments, we observed that incubation with HDL3 does not lead to normal activation of PKC in Tangier fibroblasts (W. Drobnik, MD, et al, unpublished data). The lack of HDL3-induced activation of PKC might be responsible for the reduced HDL3-mediated efflux of newly synthesized sterol observed here. Therefore, PKC was activated by addition of 10−5 mol/L 1,2-DOG to the medium, and the effect on sterol efflux to BSA and HDL3 was determined. In unstimulated Tangier fibroblasts (patient J.S.), HDL3-mediated sterol efflux was again virtually nonexistent. PKC stimulation considerably increased specific HDL3-mediated efflux of newly synthesized sterols (Fig 6). In fact, specific HDL3-mediated sterol efflux from Tangier fibroblasts after PKC stimulation was similar to that from control cells without or with additional PKC stimulation. BSA-mediated sterol efflux was also increased after PKC stimulation. Therefore, in control cells, PKC stimulation resulted in a slight net reduction of specific HDL3-mediated sterol efflux (data not shown). Cholesterol and Cholesteryl Ester Content of Tangier and Control CellsTo investigate whether the reduced efflux of newly synthesized sterol leads to an enrichment of cellular cholesterol, lipids were extracted from Tangier and control cells and quantified. Cells were incubated for 48 hours with DMEM containing 10% LPDS to reduce cellular cholesterol stores and to induce de novo sterol synthesis. Cells were incubated consecutively in DMEM supplemented with 100 μg/mL HDL3 for 24 hours. Tangier fibroblasts showed a significant enrichment in cholesteryl esters under these conditions (Table 2). This indicates that all or part of the intracellular cholesterol pool is not available for transport to the cell membrane but rather is esterified by acyl coenzyme A:cholesterol acyltransferase. DiscussionIt has been shown previously by our group that fibroblasts from Tangier patients similar to Tangier MNPs have an abnormal Golgi apparatus and vesicular compartment.14 This was taken as evidence that the genetic defect in cellular lipid metabolism and traffic shown in Tangier MNPs10 is also expressed in fibroblasts. To test this hypothesis, a detailed analysis of cholesterol traffic was performed in Tangier fibroblasts. Analysis of cellular cholesterol traffic in normal cells has revealed that there are different transport mechanisms for cholesterol from cholesterol stores or cholesterol-poor intracellular membranes, eg, the endoplasmic reticulum, to the cell membrane. These obviously depend on the origin of cholesterol (for review see References 25 and 262526 ). DeGrella and Simoni27 showed that when cells are pulsed with precursors of sterol synthesis, newly synthesized cholesterol is labeled within minutes. Transport of newly synthesized cholesterol from the endoplasmic reticulum is energy dependent. It is completely abolished by temperatures <15°C. At 37°C, transport takes between 10 and 60 minutes. In contrast to these findings, transport of cholesterol to the cell membrane taken up via the LDL-receptor and the lysosomal route is not inhibited by energy poisons, indicating that it is not energy dependent. Lysosomal cholesterol appears to be somewhat faster in the cell membrane than newly synthesized cholesterol. However, transport time (2 to 40 minutes) is similar to that for newly synthesized cholesterol.252829 Evidence for a specific transport route of lysosomal cholesterol to the cell membrane also derives from NPC fibroblasts, in which the transport of lysosomal cholesterol to the cell membrane is disturbed, whereas the transport of newly synthesized cholesterol appears to be normal.2125 Thus, there is evidence for two, at least in part independent, transport routes of cellular cholesterol to the cell membrane. Defects in either of these pathways might affect cholesterol homeostasis of the cell and reverse cholesterol transport. In the present study, sterol transport was determined by measuring efflux to an extracellular acceptor. Cellular cholesterol pools were labeled in four different ways: (1) homogeneous labeling of cellular cholesterol by long-term incubation with [14C]cholesterol, (2) incorporation of labeled cholesterol into the cell membrane lipid pool by diffusion, (3) uptake of labeled cholesteryl esters by the LDL-receptor pathway, and (4) incorporation of labeled mevalonolactone into newly synthesized sterols. Homogeneous labeling of all cellular cholesterol pools showed a reduction of specific HDL3-mediated cholesterol efflux to approximately 50% of control. In further experiments, it could be shown that this reduction is not caused by disturbances in membrane desorption or transport of lysosomal cholesterol, which were shown to be normal. The slight reduction of efflux of LDL-derived cholesterol observed during the first 4 hours disappears after longer incubation (Fig 4) and cannot account for the overall reduction in cholesterol efflux. The normal efflux of LDL-derived cholesterol after the relatively short labeling procedure is perhaps because cholesterol taken up via LDL rapidly exchanges with other cholesterol pools, particularly the cell membrane, before it becomes accessible to acyl coenzyme A:cholesterol acyltransferase, as has been shown previously.30 These data imply that under the labeling conditions used, most of the radioactive cholesterol will be in the cell membrane rather than in intracellular pools and, in particular, in intracellular cholesteryl esters. The major result of this investigation is the almost complete absence of the concentration-dependent specific HDL3-mediated efflux of newly synthesized sterol from Tangier fibroblasts. The most likely explanation for this observation is a defect in the transport of sterols from the endoplasmic reticulum to the cell membrane. The reason for the disturbed cholesterol translocation could be either a defect in one or more steps in the transport process itself or a defect in the regulation of transport. It has been shown recently that HDL apolipoproteins induce sterol transport to the cell membrane for desorption by activating PKC.1631 Desorption itself appears to depend solely on the physicochemical properties of the lipid acceptor available.3233 That means that HDL3 serves a dual function as cholesterol acceptor and activator of transport processes that provide cholesterol to the membrane for desorption. This suggested to us that PKC activation or another signal induced by HDL leading to translocation of cellular cholesterol to the cell membrane might be defective in Tangier fibroblasts. Therefore, the effects of PKC activation on HDL3-mediated efflux of newly synthesized cholesterol were analyzed. When PKC was activated by 1,2-DOG, there was no difference in HDL3-mediated efflux between control and Tangier fibroblasts. This is evidence that the genetic defect in Tangier disease leads to an inadequate stimulation of PKC by HDL3, resulting in retention of cholesterol in cellular pools. The reduced PKC activation could not be correlated to a reduction of specific binding sites for HDL on Tangier fibroblasts. This may be interpreted in two ways: (1) HDL binding to the signal-transducing receptor is not affected by the genetic defect or (2) binding to the signal-transducing receptor is defective but responsible for only a minor fraction of specific binding of HDL. The present study supports the concept that Tangier disease is caused by a cellular defect leading to abnormal regulation of lipid transport. We show for the first time that the interaction of HDL3 with Tangier fibroblasts is not followed by normal efflux of newly synthesized sterol and that this defect can be overcome by pharmacological PKC activation. Further studies are needed to identify the cellular defect at the molecular level. Skin fibroblasts, even though not a major player in lipoprotein metabolism, will be a useful tool for these studies, as has been the case for other disorders of lipid metabolism. Figure 1. Graph showing binding of 125I-HDL3 to Tangier and control fibroblasts (Scatchard plots inserted). Cells were incubated with 2.5, 5, 10, 20, 30, and 50 μg/mL for 2 hours at 4°C. Tangier fibroblasts showed Bmax of 70 ng/mg cell protein (Kd, 8.8 μg/mL), whereas the binding to control fibroblasts was slightly lower (Bmax, 52 ng/mg cell protein; Kd, 10.6 μg/mL). Solid symbols indicate Tangier fibroblasts; open symbols, control fibroblasts; •/○, specific binding of HDL3 to Tangier/control fibroblasts; ▪/□, nonspecific binding of HDL3 to Tangier/control fibroblasts; ▾/▿, Scatchard plots of Tangier/control fibroblasts. B indicates bound HDL3 (ng/mg cell protein); B/F, ratio of bound to free (not bound) HDL3. Figure 2. Graphs showing specific HDL3-mediated efflux of cholesterol after homogeneous labeling of cellular cholesterol with [14C]cholesterol from (A) Tangier fibroblasts (patient J.S.) and (B) control fibroblasts (patient N.F.). Cells were incubated with 2.0 μCi/mL [14C]cholesterol for 48 hours at 37°C followed by incubation with Dulbecco’s modified Eagle’s medium supplemented with 1% bovine serum albumin (BSA) and with 10 μg/mL (•), 50 μg/mL (▿), 100 μg/mL (▾), or 200 μg/mL (□) HDL3. Specific HDL3-mediated efflux is given as difference between total efflux with HDL minus unspecific efflux to BSA in percent of cellular [14C]cholesterol. Labeling and unspecific efflux were not different in the two cell cultures. Specific HDL3-mediated efflux was reduced to about 50% of control in Tangier fibroblasts. Data represent mean±SD from three independent experiments per fibroblast culture. Figure 3. Graphs showing efflux of cholesterol incorporated into the cell membrane by diffusion from two Tangier (•) and four control (○) fibroblast cultures after 1-, 4-, 12-, and 24-hour chase periods. Cells were prelabeled with [14C]cholesterol for 3 hours at 15°C and incubated in Dulbecco’s modified Eagle’s medium with 1% bovine serum albumin (BSA) or 1% BSA plus 10, 50, or 100 μg/mL HDL3 as indicated for up to 24 hours. There is a dose- and time-dependent increase in cholesterol efflux from all six fibroblast cultures. The Tangier fibroblasts behave similarly to the control fibroblasts. The data points represent the mean±SEM from three independent experiments with two Tangier and four control fibroblast cultures each. Figure 4. Graphs showing efflux of LDL-derived cholesterol from (A) Tangier, (B) control, and (C) Niemann-Pick type C fibroblasts. Cells were labeled with [3H]cholesteryl linoleate LDL for 3 hours at 37°C and chased in the presence of 1% bovine serum albumin (BSA) (○) or 1% BSA plus 10 μg/mL (•), 50 μg/mL (▿), or 100 μg/mL (▾) HDL3. The difference between nonspecific efflux to 1% BSA and efflux mediated by 1% BSA supplemented with HDL3 is similar in Tangier and control fibroblasts. Data represent mean±SD from six experiments per fibroblast culture. Figure 5. Graphs showing specific HDL3-mediated efflux of newly synthesized sterol from two Tangier (•) and four control (○) fibroblast cultures after 1-, 4-, 12-, and 24-hour chase periods. An HDL3 concentration–dependent efflux could be demonstrated for control cells, whereas HDL3-specific sterol efflux is almost absent in both Tangier cultures. It is important to note that there is no overlap between Tangier and control fibroblasts. The data points represent the means from six independent experiments with two Tangier and four control fibroblast cultures each. Figure 6. Graph showing HDL3-mediated efflux of newly synthesized cholesterol from Tangier fibroblasts (patient J.S.) after pharmacological stimulation of protein kinase C with 1,2-dioctanoylglycerol (1,2-DOG). Cells were prelabeled with [14C]mevalonolactone for 3 hours at 15°C and subsequently incubated in Dulbecco’s modified Eagle’s medium without (solid symbols) and with (open symbols) 10−5 mol/L 1,2-DOG together with 1% bovine serum albumin (BSA) (•, ○) or 1% BSA plus 100 μg/mL HDL3 (▾, ▿) for 12 hours. 1,2-DOG was added every 4 hours to the incubation medium. Tangier fibroblasts showed normalization of HDL3-mediated cholesterol efflux after incubation with 1,2-DOG. Data represent mean±SD from three independent experiments per fibroblast culture.
This study was supported in part by a grant to Prof Dr Schmitz by the Deutsche Forschungsgemeinschaft within the SFB 310. The expert technical assistance of Renate Glätzl is greatly appreciated. The authors are indebted to Dr David Bowyer, Cambridge, for critically reading and commenting on the manuscript. This study was possible only with the continuing cooperation of the patients. FootnotesReferences
Page 3
Abstract HDLs encompass structurally heterogenous lipoproteins that fulfill specific functions in reverse cholesterol transport. Two-dimensional nondenaturing gradient gel electrophoresis (2D-PAGGE) of normoalphalipoproteinemic plasma and subsequent immunoblotting with anti–apoA-I-antibodies differentiates pre-β1-LpA-I, pre-β2-LpA-I, pre-β3-LpA-I, α-LpA-I2, and α-LpA-I3. Immunodetection with anti-apoE antibodies differentiates γ-LpE and α-LpE. Pulse-chase incubations of plasma with [3H]unesterified cholesterol ([3H]UC)–labeled fibroblasts and subsequent 2D-PAGGE revealed that cell-derived [3H]UC is taken up by pre-β1-LpA-I and γ-LpE. From these initial acceptors, [3H]UC is transferred to LDL via pre-β2-LpA-I→pre-β3-LpA-I→α-LpA-I. Some UC is esterified in pre-β3-LpA-I, and some is esterified in α-LpA-I after its retransfer from LDL. In this study we investigated the effect of various forms of familial HDL deficiency on reverse cholesterol transport. Plasma samples of patients with various forms of HDL deficiency are characterized by the lack of specific HDL subclasses. ApoE-containing HDLs, including γ-LpE, are present in all kinds of HDL deficiency. However, all forms of LpA-I are absent in apoA-I–deficient plasma, pre-β3-LpA-I and α-LpA-I from the plasma of patients with Tangier disease (TD), and pre-β3-LpA-I and large α-LpA-I from the plasma of patients with lecithin:cholesterol acyltransferase (LCAT) deficiency and fish-eye disease (FED). After a 1-minute pulse with labeled fibroblasts, efflux of [3H]UC into HDL-deficient plasmas decreased, compared with normal plasma, by 49% (apoA-I deficiency), 36% (TD), 21% (LCAT deficiency), and 28% (FED). In apoA-I deficiency, only γ-LpE takes up cell-derived [3H]UC. In the three other HDL-deficiency states, cell-derived [3H]UC is initially taken up by both pre-β1-LpA-I and γ-LpE. The four HDL deficiencies are also characterized by differences in the esterification of cell-derived [3H]UC. No esterification occurs in LCAT-deficient plasma. In FED plasma, [3H]UC is esterified in LDL. In apoA-I deficiency and TD, however, [3H]UC is esterified in lipoproteins free of apoA-I and apoB. In the two latter cases, the transfer of [3H]cholesteryl ester to LDL is enhanced compared with normal plasma. The lack of specific HDL subclasses and the consequent changes in reverse cholesterol transport pathways differently affect net mass efflux of cholesterol from fibroblasts into HDL-deficient plasma. Compared with normoalphalipoproteinemic plasma, net cholesterol efflux from fibroblasts into plasma is reduced by 48%, 12%, 60%, and 34% in apoA-I deficiency, TD, LCAT deficiency, and FED, respectively. Removal of apoB-containing lipoproteins from plasma of patients with apoA-I deficiency, TD, LCAT deficiency, and FED further decreased net cholesterol efflux rates by 77%, 84%, 72%, and 64%, respectively, compared with a reduction of 39% in normoalphalipoproteinemic control plasma. In conclusion, various quantitatively minor HDL subfractions and LDL also present in HDL-deficient plasma effectively contribute to reverse cholesterol transport. Several epidemiological and clinical studies have revealed an inverse correlation between the plasma concentration of HDL cholesterol and the risk of myocardial infarction (reviewed in Reference 11 ). The ability of HDL to protect the vessel wall from atherosclerosis has usually been explained by the reverse cholesterol transport model (reviewed in References 2 through 4234 ), in which HDL mediates the flux of excess cholesterol from peripheral cells to the liver. HDL, however, includes structurally and functionally heterogenous lipoproteins that can be differentiated on the basis of density, size, charge, and apolipoprotein composition.5678 Pulse-chase incubations of plasma with [3H]cholesterol-labeled fibroblasts and subsequent nondenaturing two-dimensional gradient gel electrophoresis (2D-PAGGE) have helped to assign distinct roles to the various HDL subclasses in reverse cholesterol transport. From cell membranes, cholesterol is initially taken up by a subgroup of HDL that contains apoA-I as its only apolipoprotein and is termed pre-β1-LpA-I because of its electrophoretic pre-beta mobility9 and by another subclass of HDL that contains only apoE and is termed γ-LpE because of its electrophoretic gamma mobility.10 From pre-β1-LpA-I, cell-derived cholesterol is rapidly transferred to other lipoproteins in the order pre-β2-LpA-I→pre-β3-LpA-I→α-LpA-I→LDL.11 Details of cholesterol transfer subsequent to its uptake by γ-LpE are not fully understood, except that cholesterol ultimately accumulates in LDL. In normolipoproteinemic human plasma, lecithin:cholesterol acyltransferase (LCAT) directly esterifies a minor portion of cell-derived cholesterol during its passage through pre-β3-LpA-I; most cholesterol, however, is esterified in α-LpA-I after recycling from LDL.111213 Several inborn errors of metabolism interfere with the formation of normal HDL.14 Some mutations in the apoA-I gene prevent synthesis and secretion of apoA-I. Clinically, the patients may present with xanthomatosis, atherosclerosis, and/or corneal opacifications.151617181920 Mutations in the LCAT gene lead to the expression of two different clinical and biochemical phenotypes. Familial LCAT deficiency results from the complete failure to esterify cholesterol in the plasma compartment and is characterized by increases in the ratio of unesterified cholesterol (UC) to cholesteryl ester (CE).212223 Affected patients suffer from corneal opacifications and nephropathy with proteinuria and renal insufficiency. By contrast, in fish-eye disease (FED), LCAT fails to esterify cholesterol in HDL but not in the apoB-containing lipoproteins (LpB). Clinically, this selective loss of α-LCAT activity is characterized by the presence of massive corneal opacifications, which provide the name of the disease.242526 In another form of partial LCAT deficiency characterized by the presence of corneal opacities and a normal UC-CE ratio, the ability of the patient’s plasma to esterify radiolabeled cholesterol in VLDL, LDL, and HDL was reduced because of a decrease in LCAT mass.27 The pathogenesis of Tangier disease (TD) is as yet unknown but is thought to mainly involve a disturbance of intracellular lipid transfer processes in macrophages and Schwann cells. Patients with TD present with abnormal tonsils, neuropathy, and hepatosplenomegaly.28 Despite the absence or a severe reduction of HDL, most patients with these familial HDL deficiency syndromes appear not to be at increased risk for coronary disease. Therefore, we hypothesized that the maintenance of reverse cholesterol transport in both HDL-deficient and normal plasma does not depend on the major part of HDL but on the presence of subfractions that might compensate for one another. To prove this hypothesis, we performed pulse-chase incubations of plasma samples from HDL-deficient patients by using [3H]cholesterol-labeled fibroblasts. After separation of these plasma samples by nondenaturing 2D-PAGGE, we monitored the occurrence of radioactive cholesterol and CEs in the various HDL subclasses. This helped us to identify those lipoproteins that are involved in the initial uptake of cell-derived cholesterol as well as its subsequent transfer and esterification. Three normolipidemic probands and four patients with different forms of primary HDL deficiency were included in this study. Characteristics of their lipid metabolism are summarized in Table 1. The 32-year-old Italian woman with apoA-I deficiency is homozygous for a nonsense mutation in codon 32 of the apoA-I gene.29 The truncated protein could not be detected in her plasma. She was not affected by premature atherosclerosis. The 60-year-old German patient with TD has been reported previously.30 The patient with familial LCAT deficiency is homozygous for a mutation in codon 321 of the LCAT gene, which leads to replacement of a Thr by an Ile.21 FED was diagnosed in the 30-year-old German woman, who presented with typical corneal opacifications and selective loss of α-LCAT activity. Like most of the German patients with FED who have been described,2425 this woman was homozygous for a missense mutation in the LCAT gene, which leads to a Thr→Ile substitution at residue 123. Blood samples were taken after the subjects had fasted overnight and immediately placed on ice. Plasma samples and sera were obtained by centrifugation at 4°C (2000g, 15 minutes), divided into aliquots, and frozen at −70°C. In former studies we found that freezing and thawing did not affect the ability of normal plasma to take up, esterify, and transfer cell-derived cholesterol. Serum was used for the quantification of lipids. LCAT and CE transfer protein activities were determined in EDTA-plasma. For experiments in which plasma was incubated with cells, streptokinase (Sigma Chemical Co) was used as the anticoagulant at a final concentration of 150 U/mL. Serum concentrations of triglycerides and cholesterol were quantified with an autoanalyzer (Hitachi/Boehringer). HDL cholesterol concentrations were measured after precipitation of LpB with phosphotungstic acid/MgCl2 (Boehringer). LDL cholesterol was calculated with the Friedewald formula.31 Concentrations of apoA-I and apoB were determined with a modified commercially available turbidimetric assay (Boehringer Mannheim).32 LCAT activity was determined as the amount of esterified [3H]cholesterol that was incorporated into apoA-I–containing proteoliposomes.33 The plasma activity of CE transfer protein was determined as the amount of [14C]cholesteryl oleate transferred from artificial apoA-I–containing proteoliposomes to LDL, as reported previously.3435 LDL (d=1.019 to 1.063 g/mL) and HDL (d=1.063 to 1.21 g/mL) were isolated from fresh normal human plasma by standard preparative ultracentrifugation techniques and dialyzed against 10 mmol/L sodium phosphate buffer (PBS, pH 7.4) containing 0.15 mol/L NaCl.36 In some experiments as indicated, apoB-free plasma was obtained by precipitation of LpB by phosphotungstic acid/MgCl2 as recommended by the manufacturer (Boehringer Mannheim). The apoB-free supernatant was subsequently dialyzed against PBS (pH 7.4) and used in experiments for determining net cholesterol efflux. Complete removal of LpB was ascertained by immunoturbidimetry of apoB.32 The distribution of apoA-I– and apoE-containing lipoproteins in the plasma from normoalphalipoproteinemic and HDL-deficient probands was determined by nondenaturing 2D electrophoresis, in which agarose gel electrophoresis was followed by polyacrylamide gradient gel electrophoresis (2D-PAGGE).910 Briefly, in the first dimension, 20 μL of normal or HDL-deficient plasma was separated by electrophoresis at 4°C on a 0.75% agarose gel with a 50 mmol/L merbital buffer (pH 8.7, Serva). Agarose gel strips containing the preseparated lipoproteins were then transferred to a 2% to 20% polyacrylamide gradient gel. Separation in the second dimension was performed at 40 mA for 4 to 5 hours at 4°C in a buffer system that has been described by Altland and coworkers.37 After separation, the proteins in 2D-PAGGE gels were electroblotted onto a nitrocellulose membrane. ApoA-I– and apoE-containing lipoproteins were detected by the use of sheep antibodies against human apoA-I and human apoE, respectively (Boehringer Mannheim), which were biotinylated according to the manufacturer’s recommendations (Sigma). The antigen-antibody complexes were visualized with a streptavidin-biotinylated horseradish peroxidase complex (Amersham) at a dilution of 1:1000. 4-Chloro-1-naphthol was used as the chromogen. Cell CultureNormal human skin fibroblasts were cultured in Dulbecco’s modification of Eagle’s minimum essential medium containing 10% fetal calf serum as described previously.38 After 5 to 10 passages, cells were plated on 3.5-cm-diameter dishes for the pulse-chase experiments. When they were nearly confluent, the cells of some dishes were incubated for 72 hours at 37°C with 0.5 mCi [1,2-3H]cholesterol (51.7 Ci/mmol, New England Nuclear), which was complexed with fetal calf serum. Before the incubations with plasma, fibroblasts were washed six times with PBS (pH 7.4). The final specific radioactivity in the labeled cells then amounted to 5.2±1.4×108 cpm/mg cell protein, or 1.7±0.8×107 cpm/μg cell cholesterol (mean±SD). Pulse-Chase Incubations With FibroblastsIn pulse-chase experiments, 1 mL of complete plasma or apoB-free plasma was first incubated with labeled fibroblasts (pulse). After either 1 or 5 minutes of incubation, the plasma was removed and used for the chase incubations, which were performed in the absence of cells. Conditions and time intervals were varied as indicated in “Results.” In some instances, LDL (50 μg protein) or 5,5′-dithio-bis(2-nitrobenzoic acid) at a final concentration of 1.5 mmol/L (DTNB, Sigma) was added to complete or apoB-free plasma before starting the chase incubations. After the chase incubation, the plasma samples were either delipidated by the addition of chloroform/methanol (2:1, vol/vol)39 or used for nondenaturing 2D electrophoresis. Typically, an unlabeled sample from a normoalphalipoproteinemic control subject and a labeled sample from a patient were run in parallel on one gel. One half of the gel containing the 2D electrophoretic pattern of the patient was stored at 4°C. The other half of the gel containing the pattern of the normoalphalipoproteinemic control subject was electroblotted onto a nitrocellulose membrane to immunolocalize the lipoproteins containing apoA-I and apoE. The immunoblot was then used as a template to localize the corresponding lipoproteins in the other half of the gel (Fig 1). These lipoproteins were cut out, and their lipids were extracted with chloroform/methanol (2:1, vol/vol) for 72 hours. In some experiments, the total radioactive cholesterol in various lipoprotein fractions was determined. To separately count their radioactivities, in other experiments UC and CE were first separated by thin-layer chromatography using silica gel plates (Merck) as the immobile phase and hexane/ether (6:4, vol/vol) as the mobile phase. Determination of Cholesterol Net Mass TransferNet cholesterol mass transfer from fibroblasts describes the difference in the plasma concentrations of UC after a 2-hour incubation of plasma with and without fibroblasts and was determined as described previously.1140 This test is based on the assumption that esterification of cholesterol will decrease the concentration of UC in both the presence and absence of fibroblasts. However, efflux of cellular cholesterol into the medium causes less of a decrease in UC in the sample that has been incubated with fibroblasts. In brief, after they were washed four times with PBS, dishes with confluently growing fibroblasts that had been preloaded with cholesterol for 48 hours41 and dishes without cells were incubated with 2 mL of 5% plasma or apoB-free plasma in PBS. The media were removed after a 2-hour incubation at 37°C. Lipids in 1 mL of total medium were extracted with chloroform/methanol (2:1, vol/vol) three times for 6 hours each. Cholesterol mass in the extracts was quantified by the use of a modified fluoroenzymatic assay as reported previously.40 General ProceduresTotal protein concentrations were measured according to the method of Lowry et al42 using bovine serum albumin as the standard. Every experiment was performed three times on plasma samples from each proband. In some instances percent values are presented. They represent the amount of [3H]UC in one particle as a percentage of total [3H]UC in all lipoproteins (ie, γ-LpE+pre-β-LpA-I+α-LpA-I+LDL). VLDL was not included in this calculation, because this large lipoprotein does not migrate into the polyacrylamide gradient gel. ResultsCharacterization of ApoA-I– and ApoE-Containing Lipoproteins in Normoalphalipoproteinemic and HDL-Deficient Plasma SamplesFig 1 presents the distribution of apoA-I– and apoE-containing lipoproteins after nondenaturing 2D-PAGGE of plasma from a normoalphalipoproteinemic proband. As described previously,911 apoA-I can be immunochemically detected by 2D-PAGGE in a major HDL subfraction with α-mobility (α-LpA-I) as well as in three minor subfractions with pre-β mobility that differ by size (pre-β1-LpA-I, pre-β2-LpA-I, and pre-β3-LpA-I). According to Fielding and colleagues,43 we differentiated α-LpA-I into smaller α-LpA-I3 with a Stokes’ diameter of 7.5 to 9.5 nm and larger α-LpA-I2 with a Stokes’ diameter of 9.6 to 12 nm (Fig 1c). LDL also reacted with anti–apoA-I antibodies. ApoE is immunolocalized in two particles; the bulk of apoE is present in a particle that mainly colocalizes with α-LpA-I. A minor subfraction of apoE, however, is present in a distinct lipoprotein with γ-mobility and a Stokes’ diameter of 12 to 14 nm, which we previously described as γ-LpE.10 In some Western blots, LDL immunoreacted with anti-apoE. Fig 2 shows the distribution of apoA-I–containing lipoproteins in the plasma of patients with various forms of familial HDL deficiency. ApoA-I–containing lipoproteins were undetectable in apoA-I deficiency (Fig 2a) but present in TD, FED, and familial LCAT deficiency (Fig 2b through 2d). Plasma of the TD patient contained pre-β1-LpA-I and pre-β2-LpA-I at apparently normal concentrations but no pre-β3-LpA-I or α-LpA-I (Fig 2b). Pre-β1-LpA-I and pre-β2-LpA-I were also present in the plasma of patients with FED and familial LCAT deficiency; in the latter group, these lipoprotein fractions were present in higher concentration (Fig 2c and 2d). Pre-β3-LpA-I was undetectable in FED and LCAT deficiency. In contrast to TD and apoA-I deficiency, in LCAT deficiency and FED some small α-LpA-I particles with a Stokes’ diameter of 7.5 to 9.5 nm, ie, α-LpA-I3, were visible. By contrast with normal plasma, LDL in HDL-deficient plasma did not react with anti–apoA-I antibodies. Fig 3 presents the distribution of apoE-containing lipoproteins in the plasma of the HDL-deficient patients. γ-LpE was present in all plasma samples. α-LpE was very heterogenous in the various HDL deficiency conditions. In apoA-I deficiency, α-LpE covered a considerably larger area of the gel than did α-LpE in normal plasma, indicating a greater heterogeneity in size (Fig 3a). TD plasma contained an anti-apoE immunoreactive particle with β-mobility rather than α-mobility on agarose gel electrophoresis (Fig 3b). By contrast with normoalphalipoproteinemic plasma and despite its heterogeneity in size and charge, α-LpE never colocalized with anti–apoA-I immunoreactive particles in HDL-deficient plasma. Moreover, LDL was not detected by the use of anti-apoE antibodies. Uptake and Transfer of Cell-Derived Cholesterol in Normal and HDL-Deficient PlasmaFor practical reasons we performed all subsequently described studies on plasma samples that had been frozen at −70°C, because in a pilot study we had found that freezing and thawing of plasma did not affect the distribution of apoE- and apo A-I–containing subclasses. To determine their ability to release cholesterol from cells, plasma samples from normoalphalipoproteinemic probands as well as HDL-deficient patients were incubated with [3H]UC-labeled cells for 1 minute (Table 2). Compared with that from normoalphalipoproteinemic plasma, the efflux of cell-derived [3H]cholesterol from HDL-deficient patients was reduced by 49%, 36%, 21%, and 28% in apoA-I deficiency, TD, FED, and LCAT deficiency, respectively (Table 2). Anti–apoA-I and anti-apoE immunoblots of 2D electrophoretograms of normal plasma were used as templates to localize apoA-I– and apoE-containing lipoproteins in native 2D electrophoretograms of normal and HDL-deficient plasma samples that had been pulsed with radiolabeled cellular cholesterol. Because the plasma samples of patients with FED and familial LCAT deficiency contained only small α-LpA-I (Fig 2c and 2d), α-LpA-I was divided into α-LpA-I3 (Stokes’ diameter, 7.5 to 9.5 nm) and α-LpA-I2 (Stokes’ diameter, 9.6 to 12.0 nm). α-LpE was not considered separately for two reasons. First, in normoalphalipoproteinemic plasma, most α-LpE colocalizes with α-LpA-I2. Second, α-LpE is heterogenous in HDL deficiency. Table 3 summarizes the recoveries of radioactivity extracted from the gels compared with the radioactivity in total plasma, supernatants, and infranatants after precipitation of LpB with phosphotungstic acid/MgCl2. These recoveries ranged from 80% to 90% and did not differ significantly between plasma samples from HDL-deficient patients and normoalphalipoproteinemic control subjects. Experiments on every sample were performed in triplicate and in independent series. The interassay coefficients of variation of the radioactivity recovered from plasma or the various HDL subfractions were below 20%. Figs 4 and 5 show the presence of [3H]cholesterol in the various lipoproteins of normal and HDL-deficient plasma after a 1-minute pulse with radiolabeled fibroblasts (open bars) and an additional 1-minute chase without cells (hatched bars). Fig 5 represents the amount of [3H]UC in one particle as a percentage of total [3H]UC in all lipoproteins (ie, γ-LpE+pre-β-LpA-I+α-LpA-I+LDL). VLDL was not included in this calculation because this large lipoprotein does not migrate into the polyacrylamide gradient gel. Fig 4 gives absolute numbers as counts per minute in the various lipoproteins. Table 4 presents the initial efflux of cholesterol into the various lipoproteins as a percentage of [3H]cholesterol in the cells. After a 1-minute pulse of normal plasma (Fig 5a), the percentages of [3H]cholesterol in γ-LpE, pre-β1-LpA-I, pre-β2-LpA-I, pre-β3-LpA-I, α-LpA-I3, α-LpA-I2, and LDL amounted to 20±3%, 9±2%, 4±1%, 5±1%, 32±4%, 15±2%, and 15±2%, respectively. After an additional 1-minute chase, the radioactivity in γ-LpE and pre-β1-LpA-I decreased to 7±1% and 5±1% and simultaneously increased in α-LpA-I3, α-LpA-I2, and LDL to 40±4%, 18±2%, and 26±3%, respectively. As reported previously,91011 the occurrence of [3H]cholesterol in the cholesterol-poor pre-β1-LpA-I and γ-LpE during the pulse and its disappearance from these particles and increase in α-LpA-I during the chase indicate that in normal plasma, considerable proportions of cell-derived cholesterol are taken up first by both γ-LpE and pre-β1-LpA-I and then transferred to α-LpA-I (mainly α-LpA-I3) and LDL. In all forms of HDL deficiency, γ-LpE took up cell-derived [3H]cholesterol in normal amounts (Figs 4b through 4e and 5b through 5e and Table 4). Because of the reduced efflux of [3H]cholesterol into HDL-deficient plasma, the percentages of radioactivity in γ-LpE were higher in HDL-deficient plasma than in normoalphalipoproteinemic plasma samples, namely, 38±5%, 32±3%, 36±5%, and 34±5% in apoA-I deficiency, TD, FED, and LCAT deficiency, respectively. After a 1-minute chase, the radioactivity in γ-LpE decreased to approximately 10% in all forms of HDL deficiency. In contrast to the similar degree of uptake of cell-derived [3H]cholesterol by γ-LpE in all HDL-deficient plasma, efflux into LpA-I-subfractions varied widely in the different forms of HDL deficiency. After a 1-minute pulse with apoA-I–deficient plasma (Figs 4b and 5b and Table 4), radioactivity was detectable only in γ-LpE, LDL, and a fraction with the mobility of α-LpA-I3. The latter particle, however, did not react with anti–apoA-I antiserum (cf Fig 2a). By contrast, no radioactivity was detected in the fraction with α-LpA-I2–like mobility, although this fraction included an apoE-containing lipoprotein (cf Fig 3a). After a 1-minute chase, [3H]cholesterol disappeared from γ-LpE and the α-mobile lipoprotein and accumulated in LDL. A 1-minute pulse incubation with TD plasma (Figs 4c and 5c and Table 4) led to the regular uptake of [3H]cholesterol by pre-β1-LpA-I and γ-LpE. No radioactivity was found in pre-β2-LpA-I, pre-β3-LpA-I, or α-LpA-I2, although α-LpA-I contained apoE. Like apoA-I–deficient plasma but unlike normal plasma, pulse incubation with TD plasma led to the occurrence of small amounts of [3H]cholesterol in an apoA-I–free fraction with α-LpA-I3–like mobility (Figs 4c and 5c; cf Figs 2b and 3b). During chase incubations, [3H]cholesterol disappeared from all of these initial acceptors and accumulated in LDL (Figs 4c and 5c and Table 4). Pulse incubations of plasma samples from patients with FED and LCAT deficiency led to efflux of cell-derived [3H]cholesterol into pre-β1-LpA-I and γ-LpE (Figs 4d, 4e, 5d, and 5e and Table 4) followed by a decrease in radioactivity in this fraction during chase incubation. In contrast to plasma from control subjects, apoA-I–deficient patients, and TD patients, high amounts of radioactivity accumulated in the pre-β2-LpA-I of both FED and LCAT-deficient plasma samples. As with the other two forms of HDL deficiency, only trace amounts of radioactivity were detected in fractions with the mobility of pre-β3-LpA-I and α-LpA-I2, although the latter fraction was anti-apoE immunoreactive. The percentage of radioactivity in LDL after a 1-minute chase was increased in HDL-deficient compared with normal plasma (26±3%), LCAT-deficient and FED plasma (45±5%), and apoA-I–deficient and TD plasma (60±8%). However, because exogenous [3H]cholesterol and endogenous unlabeled cholesterol within plasma lipoproteins equilibrate by diffusion and 75% of UC in normal plasma and more than 90% of UC in HDL-deficient plasma are present in LDL, higher amounts of [3H]cholesterol in LDL of patients with HDL deficiency may simply reflect the disproportionate distribution of UC among the various lipoproteins in normal versus HDL-deficient plasma. To investigate this possibility, we prolonged the chase incubation periods to various time intervals (0 to 60 minutes) and subsequently precipitated LpB with phosphotungstic acid/MgCl2. This procedure allowed us to separately determine the specific radioactivity (counts per minute of [3H]UC per microgram of UC) in the apoB-free supernatants and the apoB-containing infranatants (Fig 6). Without chase (time 0), the specific radioactivity of cell-derived [3H]UC in the supernatants of all plasma exceeded that in the infranatants (ie, LDL+VLDL) by a factor 2 to 3. During subsequent chases the specific radioactivity decreased in the supernatants but increased in the infranatants. In control plasma, specific radioactivity in the infranatant exceeded that in the supernatant after 22 minutes (Fig 6a). By contrast, in HDL-deficient plasma a higher specific radioactivity in the infranatant was already observed after 6 to 8 minutes (Fig 6b through 6e). This indicates that the transfer of cell-derived [3H]UC from the initial acceptors to LDL is enhanced in HDL-deficient compared with normal plasma. Esterification of Cell-Derived Cholesterol in Normal and HDL-Deficient PlasmaIn subsequent studies, we compared the esterification of cell-derived UC and the transfer of CE to various lipoproteins of normal and HDL-deficient plasma. To obtain amounts of labeled cholesterol sufficient to differentiate between [3H]UC and [3H]CE, pulse incubations of plasma with radiolabeled fibroblasts were prolonged to 5 minutes and chase incubations to 15 minutes (Table 5). To inhibit LCAT, all chase incubations were done in the presence of 1.5 mmol/L DTNB. After a 5-minute pulse incubation, normal plasma esterified 3.6% of the [3H]cholesterol released from cells into plasma. This fractional esterification was increased to 4.3% in apoA-I deficiency and to 5.5% in TD. In FED we observed a slight decrease to 3.1%. As expected, no detectable amounts of [3H]CE were formed in LCAT-deficient plasma. In control experiments, approximately 32% of [3H]CE was recovered in pre-β3-LpA-I, 57% in α-LpA-I, and 11% in LDL (5-minute pulse). A 15-minute chase in the presence of the LCAT inhibitor DTNB resulted in a 32% to 12% decrease of [3H]CE in pre-β3-LpA-I, a 57% to 69% increase in α-LpA-I, and an 11% to 19% increase in LDL. In the plasma of patients with FED, apoA-I deficiency, and TD, [3H]CEs were detectable only in LDL but not in pre-β-LpA-I or α-LpA-I (Table 5). After pulse incubation, the amount of [3H]CE in LDL was twofold to threefold higher in the plasma of patients with FED, TD, and apoA-I deficiency compared with normal plasma. In contrast to normal plasma, chase incubation with these HDL-deficient plasmas did not significantly increase the amount of [3H]CE in LDL (Table 5). The more rapid appearance of [3H]CE in LDL raises the possibility that in these HDL-deficient plasmas, cell-derived cholesterol is esterified in LDL or transferred to LDL after generation in lipoproteins that contain neither apoA-I nor apoB. Therefore, pulse-chase incubations were repeated with apoB-free plasma from normoalphalipoproteinemic subjects and HDL-deficient probands (Fig 7). No [3H]CE was generated in apoB-free FED and LCAT-deficient plasma, whereas the amount of [3H]CE gradually increased with prolonged chase incubation in apoB-free supernatants of control plasma, TD plasma, and apoA-I–deficient plasma. Esterification of cell-derived [3H]cholesterol in apoB-free plasma from patients with apoA-I deficiency and TD was increased twofold and threefold, respectively, compared with apoB-free plasma from normal control subjects. Thus, with the exception of FED and LCAT-deficient plasma, esterification of cell-derived [3H]cholesterol occurred in the absence of LpB (Fig 7). To further investigate the more rapid transfer of CE to LDL in apoA-I–deficient and TD plasma, apoB-free plasma samples were first pulsed with cell-derived [3H]cholesterol, then supplemented with exogenous LDL, and finally used for 15-minute chase incubations. Exogenous LDL was then precipitated by addition of phosphotungstic acid/MgCl2 to separately determine [3H]UC and [3H]CE in the apoB-free supernatants (Fig 7b) and in the LDL-containing infranatants (Fig 7c). More than one fourth (29±4%) of [3H]CE was found in the LDL-containing infranatant of normal plasma, whereas 76±3% and 73±3% were detected in the LDL-containing infranatants of TD plasma and apoA-I–deficient plasma, respectively. Net Cholesterol Efflux From Fibroblasts Into Normal and HDL-Deficient PlasmaWe also determined the net mass transfer of UC from fibroblasts to both native and apoB-free plasma during a 2-hour incubation (Table 6). During this period, a net amount of 7.46±1.14 nmol UC per milligram of cell protein was released into 100 μL plasma of normoalphalipoproteinemic probands. Compared with normal plasma, the net transfer into plasma from apoA-I–deficient, TD, FED, and LCAT-deficient patients was reduced to 53±7%, 88±11%, 66±11%, and 41±7%, respectively. Removal of LpB from normal plasma resulted in a 39% decrease of net released cellular cholesterol, from 7.46±1.14 to 4.55±0.73 nmol UC per milligram of cell protein. Removal of LpB from the plasma of patients with apoA-I deficiency, TD, FED, and LCAT deficiency, however, decreased the net cholesterol transport rates by 77%, 84%, 72%, and 64%, respectively. These findings indicate that LpB plays an important role in net cholesterol efflux in HDL-deficient plasmas of various origins. DiscussionThe antiatherogenic effect of HDL has generally been attributed to its ability to mediate the flux of excess cholesterol from peripheral cells to the liver.234 In normoalphalipoproteinemic plasma, this reverse cholesterol transport involves (1) uptake of cell-derived UC by pre-β1-LpA-I and γ-LpE; (2) subsequent transfer of UC to LDL via pre-β2-LpA-I, pre-β3-LpA-I, and α-LpA-I; (3) esterification of UC, mostly in α-LpA-I but to a lesser extent in pre-β3-LpA-I; and (4) transfer of CE to LDL.91011121343 Patients with various forms of familial HDL deficiency do not lack HDL completely but rather are deficient in distinct HDL subclasses.194445464748 In our study, nondenaturing 2D electrophoresis revealed the absence of pre-β3-LpA-I and α-LpA-I2 and the presence of γ-LpE in all four kinds of familial HDL deficiency. Furthermore, pre-β1-LpA-I and pre-β2-LpA-I were present in all forms of HDL deficiency except apoA-I deficiency; α-LpA-I3 was detectable in LCAT deficiency and FED but not in TD and apoA-I deficiency. Uptake, transfer, and esterification of cell-derived cholesterol in the plasma of patients with familial HDL deficiency syndromes were significantly different from those of normal plasma (Table 7). Efflux of Cell-Derived CholesterolCholesterol efflux from cells is the result of complex mechanisms that involve the synthesis of cholesterol in the endoplasmic reticulum, hydrolysis of CE in lysosomes and cytosolic lipid droplets, translocation of cholesterol to the cell membrane, and finally desorption of cholesterol from the cell membrane into the plasma compartment.449 These processes are regulated differently in different cell types and depend on different extracellular stimuli and acceptors of cholesterol, such as the various apolipoproteins and HDL subclasses.50 In apoA-I deficiency, cellular cholesterol was taken up by γ-LpE but not by pre-β1-LpA-I. The initial efflux of [3H]cholesterol and the net efflux of cholesterol into plasma were decreased by 50% to 60% compared with normoalphalipoproteinemic plasma. Similarly, Fielding and coworkers (Fielding and Moser51 and Kawano et al52 ) observed a 55% reduction in cholesterol efflux stimulation by plasma that had been depleted of LpA-I by immunoaffinity chromatography. These authors concluded that pre-β1-LpA-I contributed to more than half of a plasma’s ability to release cholesterol from cells and attributed the residual activity to nonspecific effects of albumin.95152 Our data, however, show that γ-LpE takes up to twofold more [3H]cholesterol than does pre-β1-LpA-I, suggesting that this lipoprotein is also a major contributor to the cholesterol efflux–stimulating activity of plasma. The importance of γ-LpE for the initial efflux of cell-derived cholesterol into plasma is also underlined by our observation that the initial efflux of cell-derived [3H]UC into TD plasma is only slightly higher than that by apoA-I–deficient plasma, although TD plasma contains pre-β1-LpA-I. Transfer of UC to LDLIn normoalphalipoproteinemic plasma, cell-derived cholesterol is taken up by a number of particles and then transferred to LDL. This transfer involves various lipoproteins, including pre-β2-LpA-I, pre-β3-LpA-I, and α-LpA-I. From LDL, UC is either retransferred to α-LpA-I for esterification or taken up by cells.1113 These transfer mechanisms of cell-derived cholesterol to LDL were found to be operative in the various HDL deficiency syndromes, suggesting first that γ-LpE may directly participate in the transfer of cell-derived cholesterol to LDL and second, that the specific absence of pre-β2-LpA-I, pre-β3-LpA-I, and α-LpA-I does not interfere with the transfer of UC from pre-β1-LpA-I to LDL. Esterification of Cholesterol and Transfer of CE to LDLIn normoalphalipoproteinemic plasma, most of the cell-derived cholesterol is esterified in α-LpA-I after retransfer from LDL. A smaller amount is esterified during passage through pre-β3-LpA-I.1113 In contrast to LCAT deficiency, in which plasma cholesterol esterification activity is completely lost, FED is characterized by the selective failure of plasma to esterify cholesterol in exogenous HDL or apoA-I–containing proteoliposomes while retaining the ability to esterify cholesterol in exogenous LpB.535455 In our experiments with FED plasma, CEs accumulated in LDL and were not found in other lipoproteins. These findings further suggest that LDL cholesterol can be directly esterified by the mutant LCAT in FED. By contrast, in apoA-I–deficient and TD plasma, CEs accumulated in LDL after undergoing esterification in lipoproteins containing neither apoA-I nor apoB. The nature of these lipoproteins is unclear at present. Other authors have also detected LCAT activity in the apoA-I– and apoB-free fractions of TD plasma304656 and in apoA-IV–containing particles.45 Obviously, the newly formed CEs can be effectively transferred from these abnormal particles to LDL. Familial HDL deficiency is very rare, so we were able to perform only exemplary studies on single patients and could not analyze the effect of possible heterogeneity within a given syndrome. For example, heterogeneity within apoA-I deficiency, LCAT deficiency, FED, or TD arises from allelic variation of the underlying defects in the genes of apoA-I, LCAT, or the as yet unidentified TD gene.1415161718192021222324252627282930 Further heterogeneity may originate from variations in the genetic background of an individual and other factors, such as gender, age, concomitant diseases, or genetic polymorphisms. Nevertheless, we believe that the phenomena described in this article can be extrapolated to other patients with the same HDL deficiency syndromes, because they generally cause qualitative rather than quantitative changes in the functioning of a given HDL subfraction. In this study, we searched only for those lipoproteins that contained apoA-I or apoE. We may, therefore, have overlooked two classes of HDL particles that may play a role in reverse cholesterol transport. First, there is some evidence for the involvement of lipoproteins containing neither apoE nor apoA-I, which are present even in normoalphalipoproteinemic plasma. Thus, lipoproteins that contain apoA-IV but no apoA-I or apoE can release cholesterol from cells.5758 LpA-IV also contains LCAT activity.57 Second, because apoA-I is the most abundant apolipoprotein in all subclasses, a lack or severe decrease in apoA-I gives rise to abnormal lipoproteins. Thus, α-LpE particles in apoA-I deficiency and TD are heterogenous in size and charge and do not entirely colocalize with apoA-I– or apoE-containing particles in normal plasma. These apoE-containing particles, however, do not appear to contribute substantially to reverse cholesterol transport, as neither pulse nor chase incubations led to the occurrence of radiolabel in those proportions of abnormal α-LpE that do colocalize with α-LpA-I of normal plasma (Figs 2 through 6). Abnormal HDL that contains apoA-II but no apoA-I has also been identified in some patients with apoA-I deficiency or TD.28294759 In one case, this apoA-II–containing particle was shown to promote cholesterol efflux from cells.47 Despite the uncertainty surrounding the role of these minor HDL subfractions in reverse cholesterol transport, they appear either to contribute little to the efflux of [3H]cholesterol from cells or to be entirely colocalized in those electrophoretic fractions that, in normal plasma, contain apoA-I or apoE, as 80% to 90% of the total plasma radioactivity was recovered in LpE, LpA-I, and LDL of both normal and HDL-deficient plasmas. In summary, our studies demonstrate that the two crucial steps in reverse cholesterol transport—efflux of cellular cholesterol into the plasma compartment and its transfer to LDL for final targeting to the liver—are maintained in all forms of HDL deficiency, although quantitatively and qualitatively modified (Table 7). This may explain why many forms of familial HDL deficiency do not put homozygous carriers at increased risk of coronary disease. Our data suggest that quantitatively minor plasma subfractions, eg, pre-β1-LpA-I and γ-LpE, together with LDL, are important contributors to reverse cholesterol transport. In particular, apoA-I does not play an exclusive role in reverse cholesterol transport, as shown by our in vitro findings in A-I deficiency and the absence of premature atherosclerosis in affected individuals. This is also highlighted by recent observations in transgenic animals that do not express apoA-I but fail to develop atherosclerosis.60 Figure 1. Two-dimensional (2D) electrophoresis and immunoblotting of apoA-I– (a) and apoE- (b) containing lipoproteins in normoalphalipoproteinemic human plasma. Nondenaturing 2D electrophoresis was performed in the sequence agarose gel electrophoresis→nondenaturing polyacrylamide gel electrophoresis. After electroblotting to nitrocellulose membranes, apoA-I– and apoE-containing lipoproteins were detected with biotinylated polyclonal sheep antisera against either human apoA-I or human apoE and streptavidin-biotinylated horseradish peroxidase complex. c, Schematic summary of the localization of apoA-I– and apoE-containing lipoproteins in the gel. Rectangular fields represent areas removed from the gel for extraction of lipids from the indicated lipoproteins. Area 1, pre-β1-LpA-I; 2, pre-β2-LpA-I; 3, pre-β3-LpA-I; 4, α-LpA-I3; 5, α-LpA-I2; 6, γ-LpE; 7, α-LpE; and 8, LDL. Figure 2. Nondenaturing two-dimensional electrophoresis and immunoblotting of apoA-I–containing lipoproteins in plasma of patients with apoA-I deficiency (a), Tangier disease (TD) (b), fish-eye disease (FED) (c), and familial lecithin:cholesterol acyltransferase (LCAT) deficiency (d). For further details, see the legend to Fig 1. Figure 3. Nondenaturing two-dimensional electrophoresis and immunoblotting of apoE-containing lipoproteins in plasma of patients with apoA-I deficiency (a), Tangier disease (TD) (b), fish-eye disease (FED) (c), and familial lecithin:cholesterol acyltransferase (LCAT) deficiency (d). For further details, see the legend to Fig 1. Figure 4. Bar graphs showing uptake and transfer of cell-derived [3H]cholesterol through various lipoproteins of plasma from normoalphalipoproteinemic probands and HDL-deficient patients. Pulse incubations with radiolabeled fibroblasts were performed for 1 minute (open bars); chase incubations were performed without cells for another 1 minute (hatched bars). Plasma samples were then separated by nondenaturing two-dimensional (2D) polyacrylamide gradient gel electrophoresis. Anti–apoA-I and anti-apoE immunoblots of 2D electrophoretograms of normal plasma were used to localize lipoproteins (cf Fig 1). These were removed from the native gels, their lipids were extracted, and radioactivity was counted. For further details see the “Methods” section. Panels a, b, c, d, and e represent results obtained with plasma from normoalphalipoproteinemic probands (normal) and patients with apoA-I deficiency (def.), Tangier disease (TD), fish-eye disease (FED), and lecithin:cholesterol acyltransferase (LCAT) deficiency, respectively. Each bar shows the mean and SD of three experiments as counts per minute released into the various lipoproteins. Figure 5. Bar graphs showing uptake and transfer of cell-derived [3H]cholesterol through various lipoproteins of plasma from normoalphalipoproteinemic probands and HDL-deficient patients. The figure shows the percent distribution of radioactivity among the various lipoproteins, as calculated from the data presented in Fig 4. See the legend to Fig 4 for details and explanation of abbreviations. Figure 6. Time course of changes in the distribution of cell-derived [3H]unesterified cholesterol (UC) in apoB-free (○) and apoB-containing (▵) lipoproteins of plasma from normoalphalipoproteinemic probands (normal) and various HDL-deficient (def.) patients. Pulse incubations with radiolabeled fibroblasts for 1 minute were followed by chase incubations without cells for the indicated times. ApoB-free and apoB-containing lipoproteins were separated by precipitation with phosphotungstic acid/MgCl2. Subsequent to lipid extraction, radioactive UC and cholesteryl ester were separated by thin-layer chromatography. Specific radioactivity is expressed as counts per minute of [3H]UC per microgram of UC in apoB-free and apoB-containing lipoproteins. Each point shows the mean value and standard deviation of four experiments. TD indicates Tangier disease; FED, fish-eye disease; and LCAT, lecithin:cholesterol acyltransferase. Figure 7. Effect of LpB on production and transfer of cholesteryl ester (CE) in normoalphalipoproteinemic (normal) and HDL-deficient (def.) plasma. a, Time-dependent appearance of cell-derived [3H]CE in apoB-free plasma; b and c, bar graphs showing effect of exogenous LDL on the transfer of [3H]CE from apoB-free lipoproteins to LDL. Each point or bar shows the mean and SD of three experiments. Experimental details: a, 1 mL apoB-free plasma from normoalphalipoproteinemic probands or HDL-deficient patients was incubated with radiolabeled fibroblasts for the indicated time intervals. At each indicated time point, 100 μL of sample was removed, and after lipid extraction, the radioactive unesterified cholesterol (UC) and CE were separated by thin-layer chromatography. Shown are data from the Tangier disease (TD) patient (▵), the apoA-I–deficient patient (⋄), the normoalphalipoproteinemic probands (○), and the patients with fish-eye disease and lecithin:cholesterol acyltransferase deficiency (□). b and c, After 60 minutes’ incubation with radiolabeled fibroblasts, 200 μL apoB-free plasma was removed and incubated for a further 15 minutes with 200 μg LDL. Subsequently, the apoB-free fraction (b) and LDL (c) were separated by precipitation with phosphotungstic acid/MgCl2. After lipid extraction, radioactive UC and CE in the apo B-free supernatants (b) and LDL-containing infranatants (c) were separated by thin-layer chromatography. Open and closed bars represent [3H]CE in lipoproteins before and after chase incubations, respectively.
This project was the topic of the Bennigsen-Foerder-Award from Ministerium für Forschung und Wissenschaft Nordrhein Westfalen to Dr von Eckardstein. Further support was provided by a fellowship from Boehringer Ingelheim Fonds to Dr Huang. We gratefully acknowledge the assistance of Dr Ali Chirazi in the determination of lipid transfer enzyme activities and the help of Dr Paul Cullen in editing the manuscript. FootnotesReferences
|
Which would the nurse tell the patient who is on atorvastatin in preparation for a lipid profile
Pos Terkait
Copyright © 2024 ketajaman Inc.
