Fenofibrate-associated changes in renal function and relationship to clinical outcomes among individuals with type 2 diabetes: the Action to Control Cardiovascular Risk in Diabetes (ACCORD) experience.

AIMS/HYPOTHESIS: Fenofibrate has been noted to cause an elevation in serum creatinine in some individuals. Participants in the Action to Control Cardiovascular Risk in Diabetes Lipid Study were studied to better characterise who is at risk of an increase in creatinine level and to determine whether those with creatinine elevation have a differential risk of adverse renal or cardiovascular outcomes. METHODS: A fenofibrate-associated creatinine increase (FACI) was defined as an increase in serum creatinine of at least 20% from baseline to month 4 in participants assigned to fenofibrate. Baseline patient characteristics, and baseline and 4-month drug, clinical, laboratory characteristics and study outcomes were examined by FACI status. RESULTS: Of the sample, 48% of those randomised to receive fenofibrate had at least a 20% increase in serum creatinine within 4 months. In multivariable analysis, participants who were older, male, used an ACE inhibitor at baseline, used a thiazolidinedione (TZD) at 4 months post-randomisation, had baseline CVD, and had lower baseline serum creatinine and LDL-cholesterol levels were all more likely to meet the criteria for FACI. Participants in the FACI group were also more likely to have a decrease in their serum triacylglycerol level from baseline to 4 months. No differences in study outcomes were seen by FACI criteria. CONCLUSIONS/INTERPRETATION: Several characteristics predict a rapid rise in serum creatinine upon starting fenofibrate. Participants who met the criteria for FACI also had a greater change in triacylglycerol levels. In the setting of careful renal function surveillance and reduction of fenofibrate dose as indicated, no increase in renal disease or cardiovascular outcome was seen in those individuals demonstrating FACI. TRIAL REGISTRATION: ClincalTrials.gov: NCT00000620. FUNDING: The ACCORD Trial was supported by grants (N01-HC-95178, N01-HC-95179, N01-HC-95180, N01-HC-95181, N01-HC-95182, N01-HC-95183, N01-HC-95184, IAA-Y1-HC-9035 and IAA-Y1-HC-1010) from the National Heart, Lung, and Blood Institute; by the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute on Aging, and the National Eye Institute; by the Centers for Disease Control and Prevention; by General Clinical Research Centers and by the Clinical and Translational Science Awards. Abbott Laboratories, Amylin Pharmaceutical, AstraZeneca Pharmaceuticals LP, Bayer HealthCare LLC, Closer Healthcare, GlaxoSmithKline Pharmaceuticals, King Pharmaceuticals, Merck, Novartis Pharmaceuticals, Novo Nordisk, Omron Healthcare, sanofi-aventis US and Takeda Pharmaceuticals provided study medications, equipment or supplies.

HIV protease inhibitors protect apolipoprotein B from degradation by the proteasome: a potential mechanism for protease inhibitor-induced hyperlipidemia.

Highly active anti-retroviral therapies, which incorporate HIV protease inhibitors, resolve many AIDS-defining illnesses. However, patients receiving protease inhibitors develop a marked lipodystrophy and hyperlipidemia. Using cultured human and rat hepatoma cells and primary hepatocytes from transgenic mice, we demonstrate that protease inhibitor treatment inhibits proteasomal degradation of nascent apolipoprotein B, the principal protein component of triglyceride and cholesterol-rich plasma lipoproteins. Unexpectedly, protease inhibitors also inhibited the secretion of apolipoprotein B. This was associated with inhibition of cholesteryl-ester synthesis and microsomal triglyceride transfer-protein activity. However, in the presence of oleic acid, which stimulates neutral-lipid biosynthesis, protease-inhibitor treatment increased secretion of apolipoprotein B-lipoproteins above controls. These findings suggest a molecular basis for protease-inhibitor-associated hyperlipidemia, a serious adverse effect of an otherwise efficacious treatment for HIV infection.

Combined lipase deficiency (cld): a lethal mutation on chromosome 17 of the mouse.

Two triglyceride lipases, lipoprotein lipase and hepatic triglyceride lipase, participate in the metabolism of plasma lipoproteins. A single recessive mutation, cld, on mouse chromosome 17 causes an apparent deficiency of both lipoprotein lipase and hepatic triglyceride lipase activities. Mice homozygous for this defect develop lethal hyperchylomicronemia within 2 days postpartum as a consequence of nursing. Plasma triglyceride values in affected mice often reach 20,000 milligrams per deciliter (100 times higher than that in normal littermates), and total lipase activity in plasma or tissues is 5 to 20 percent of that in controls.

Regulation of the production and catabolism of plasma low density lipoproteins in hypertriglyceridemic subjects. Effect of weight loss.

In subjects with hypertriglyceridemia, plasma concentrations of low density lipoprotein (LDL) cholesterol are often normal or reduced. Perturbations that alter plasma very low density lipoprotein (VLDL) concentrations are associated with opposite changes in plasma LDL levels. To determine the mechanisms regulating plasma LDL levels, we used 131I-VLDL and 125I-LDL to measure the fractional catabolic rates (FCR), production rates (PR), and rates of interconversion of apoprotein B (apo B) in VLDL, intermediate density lipoprotein, and LDL in six hypertriglyceridemic subjects pre- and post-weight reduction. [2-3H]glycerol was used to quantitate VLDL triglyceride PR. All data are presented as mean +/- SD. Percent ideal body weight fell from 132 +/- 17.9 to 119 +/- 15.9% in the group, P less than 0.05. After weight loss, plasma VLDL triglyceride (486.0 +/- 364.1 vs. 191.3 +/- 65.4 mg/dl, P less than 0.05) and VLDL apo B (32.2 +/- 12.0 vs. 14.8 +/- 6.8 mg/dl, P less than 0.05) concentrations were reduced. VLDL triglyceride PR also fell after weight reduction (56.6 +/- 39.0 vs. 28.6 +/- 23.1 mg/kg per h, P less than 0.05), as did VLDL apo B PR (47.9 +/- 41.4 vs. 19.0 +/- 14.1 mg/kg per d, P less than 0.05). Pre-weight loss, plasma LDL cholesterol and apo B levels were low-normal or reduced (64.0 +/- 12.6 and 58.4 +/- 11.9 mg/dl, respectively) despite normal or elevated LDL apo B PR (17.4 +/- 7.2 mg/kg per d). The reduced cholesterol and apo B levels were associated with increased FCRs (0.68 +/- 0.29 d-1) and reduced cholesterol/protein ratios (1.01 +/- 0.18) in LDL. The plasma levels of LDL cholesterol and apo B rose after weight reduction (84.8 +/- 24.9, P less than 0.05; and 69.5 +/- 14.3 mg/dl, P less than 0.05, respectively, vs. base line). These increased concentrations resulted from a combination of events. First, the FCR for LDL apo B fell in five of six subjects with a significant reduction for the group as a whole (0.48 +/- 0.11 d-1, P less than 0.05 vs. base line). Second, the cholesterol/protein ratio increased in all six subjects with a significantly greater mean after weight loss (1.25 +/- 0.27, P less than 0.05 vs. base line). In contrast, the LDL apo B PR fell or was essentially unchanged in the six subjects after weight loss (mean, 14.4 +/- 2.8 mg/kg per d; NS vs. pre-weight loss). The changes in LDL catabolism and composition were associated with changes in the source of LDL apo B. Pre-weight loss, 73.3% of LDL was derived from VLDL, while 26.7% was directly secreted into plasma. Post-weight reduction, VLDL-derived LDL fell to 46.8% of total, while direct secretion accounted for 53.2% of LDL production. These changes were significant; P < 0.95. Thus, all subjects had direct secretion of LDL apo B and the magnitude of this source of VLDL triglyceride secretion. These results indicate that the regulation of plasma LDL levels in hypertriglyceridemic subjects is quite complex and that the rise in LDL levels after weight loss results from reduction in the fractional catabolism of this lipoprotein. The fall in the FCR is associated with changes in the source of LDL and in its composition.

Association of plasma lipoproteins with postheparin lipase activities.

Studies were designed to explore the association of lipoprotein lipase (LPL) and hepatic triglyceride lipase (HTGL) activities with lipoproteins in human postheparin plasma (PHP). The major peak of LPL activity after gel filtration of PHP eluted after the triglyceride-rich lipoproteins and just before the peak of low density lipoprotein (LDL) cholesterol. When PHP contained chylomicrons, an additional peak of LPL activity eluted in the void volume of the column. Most HTGL activity eluted after the LDL and preceded the elution of high density lipoprotein cholesterol. LPL activity in preheparin plasma eluted in the same position, relative to lipoproteins, as did LPL in PHP. Gel filtration of purified human milk LPL mixed with plasma or isolated LDL produced a peak of activity eluting before LDL. During gel filtration of PHP in high salt buffer (1 M NaCl) or after isolation of lipoproteins by ultracentrifugation in high salt density solutions, most of the lipase activity was not associated with lipoproteins. LPL activity was removed from PHP by elution through immunoaffinity columns containing antibodies to apolipoprotein (apo) B and apo E. Since lipoproteins in PHP have undergone prior in vivo lipolysis, LPL activity in PHP may be bound to remnants of chylomicrons and very low density lipoproteins.

Intestinal transcription and synthesis of apolipoprotein AI is regulated by five natural polymorphisms upstream of the apolipoprotein CIII gene.

To understand the factors contributing to the synthesis of human apolipoprotein AI (apoAI), relative apoAI synthesis was measured from endoscopic biopsy samples obtained from 18 healthy volunteers. The relative amount of apoAI synthesis was directly correlated with steady state intestinal apoAI mRNA levels and a 10-fold within-group variability was observed. Analysis of genomic DNA from the subjects revealed five polymorphic sites which defined two haplotypes in the intestinal enhancer region of the apoAI gene located upstream of the apolipoprotein CIII gene transcriptional start site (+ 1): (-641 C to A, -630 G to A, -625 T to deletion, -482 C to T, and -455 T to C). The population frequencies of the wild-type and mutant alleles were 0.53 and 0.44, respectively. Mean steady state apoAI mRNA levels and mean relative apoAI synthesis were 49 and 37% lower, respectively, in homozygotes for the mutant allele and 28 and 41% lower, respectively, in heterozygotes than in homozygotes for the wild-type allele (P < 0.05 for both). Site-directed mutants of apoAI gene promoter/reporter constructs containing the above mutations were transfected into Caco-2 cells and showed a 46% decrease in transcriptional activity compared with the wild type (P < 0.001); however, no significant differences were observed in HepG2 cells. Electrophoretic mobility shift assays showed that the mutated sequences from -655 to -610 bound Caco-2 cell nuclear protein(s) while the wild type did not. These results indicate that intestinal apoAI gene transcription and protein synthesis are genetically determined and are reduced in the presence of common mutations which induced binding of nuclear protein(s), possibly a transcriptional repressor.

Effect of somatostatin-induced suppression of postprandial insulin response upon the hypertriglyceridemia associated with a high carbohydrate diet.

In an attempt to define the relationship between plasma insulin and triglyceride concentrations, we have studied the effect of suppression of the postprandial insulin response upon the secretion and plasma concentration of very low density lipoprotein (VLDL)-triglycerides. Eight nondiabetic subjects with a wide range of fasting plasma triglyceride levels (100-358 mg/dl) were studied during three dietary periods: base line, high carbohydrate (80% calories), and high carbohydrate (80% calories) with a daily intravenous infusion of somatostatin (SRIF) (1.3 micrograms/min) between 800 and 2,100 h. The significant increase in postprandial insulin response observed during high carbohydrate vs. base line was completely abolished during high carbohydrate-SRIF. However, plasma triglyceride levels rose in all subjects during each high carbohydrate period (with/without SRIF) vs. base line and the mean values reached during each period were the same (476 +/- 165 vs. 482 +/- 152 mg/dl, respectively). The secretion of VLDL-triglyceride into plasma was higher in four subjects, the same in two subjects, and lower in one subject during high carbohydrate-SRIF vs. high carbohydrate alone. The mean production rate of VLDL-triglyceride (mg/kg per h) was 25.6 +/- 4.9 during the high carbohydrate and 40.9 +/- 28.1 during the high carbohydrate-SRIF periods. These values were not significantly different. Postprandial glucose levels were slightly increased during high carbohydrate-SRIF, but overnight glucose concentrations were not affected. Plasma FFA levels were not different during the two high carbohydrate periods. Plasma glucagon levels did not appear to affect the results either. This study indicates that postprandial hyperinsulinemia during a high carbohydrate diet is not necessary for induction of hypertriglyceridemia.

Suppression of apolipoprotein B production during treatment of cholesteryl ester storage disease with lovastatin. Implications for regulation of apolipoprotein B synthesis.

Cholesteryl ester storage disease (CESD) is characterized by the deficient activity of lysosomal cholesteryl ester (CE) hydrolase, accumulation of LDL-derived CE in lysosomes, and hyperlipidemia. We studied the kinetics of VLDL and LDL apolipoprotein B (apoB), using 125I-VLDL and 131I-LDL, in a 9-yr-old female with CESD and elevated total cholesterol (TC) (271.0 +/- 4.4 mg/dl), triglyceride (TG) (150.0 +/- 7.8 mg/dl), and LDL cholesterol (184.7 +/- 3.4 mg/dl). These studies demonstrated a markedly elevated production rate (PR) of apoB, primarily in LDL, with normal fractional catabolism of apoB in VLDL and LDL. Urine mevalonate levels were elevated, indicative of increased synthesis of endogenous cholesterol. Treatment with lovastatin, a competitive inhibitor of hydroxymethylglutaryl coenzyme A reductase, resulted in significant reductions in TC (196.8 +/- 7.9 mg/dl), TG (100.8 +/- 20.6 mg/dl), and LDL cholesterol (102.0 +/- 10.9 mg/dl). Therapy reduced VLDL apoB PR (5.2 vs. 12.2 mg/kg per d pretreatment) and LDL apoB PR (12.7 vs. 24.2 mg/kg per d pretreatment). Urine mevalonate levels also decreased during therapy. These results indicate that, in CESD, the inability to release free cholesterol from lysosomal CE resulted in elevated synthesis of endogenous cholesterol and increased production of apoB-containing lipoproteins. Lovastatin reduced both the rate of cholesterol synthesis and the secretion of apoB-containing lipoproteins.

Lipoprotein metabolism during acute inhibition of lipoprotein lipase in the cynomolgus monkey.

To clarify the role of lipoprotein lipase (LPL) in the catabolism of nascent and circulating very low density lipoproteins (VLDL) and in the conversion of VLDL to low density lipoproteins (LDL), studies were performed in which LPL activity was inhibited in the cynomolgus monkey by intravenous infusion of inhibitory polyclonal or monoclonal antibodies. Inhibition of LPL activity resulted in a three- to fivefold increase in plasma triglyceride levels within 3 h. Analytical ultracentrifugation and gradient gel electrophoresis demonstrated an increase predominantly in more buoyant, larger VLDL (Sf 400-60). LDL and high density lipoprotein (HDL) cholesterol levels fell during this same time period, whereas triglyceride in LDL and HDL increased. Kinetic studies, utilizing radiolabeled human VLDL, demonstrated that LPL inhibition resulted in a marked decrease in the catabolism of large (Sf 400-100) VLDL apolipoprotein B (apoB). The catabolism of more dense VLDL (Sf 60-20) was also inhibited, although to a lesser extent. However, there was a complete block in the conversion of tracer in both Sf 400-100 and 60-20 VLDL apoB into LDL during LPL inhibition. Similarly, endogenous labeling of VLDL using [3H]leucine demonstrated that in the absence of LPL, no radiolabeled apoB appeared in LDL. We conclude that although catabolism of dense VLDL continues in the absence of LPL, this enzyme is required for the generation of LDL.

Combined hyperlipidemia in transgenic mice overexpressing human apolipoprotein Cl.

We have generated transgenic mice over-expressing human apolipoprotein CI (apo CI) using the native gene joined to the downstream 154-bp liver-specific enhancer that we defined for apo E. Human apo CI (HuCI)-transgenic mice showed elevation of plasma triglycerides (mg/dl) compared to controls in both the fasted (211 +/- 81 vs 123 +/- 52, P = 0.0001) and fed (265 +/- 105 vs 146 +/- 68, P < 0.0001) states. Unlike the human apo CII (HuCII)- and apo CIII (HuCIII)-transgenic mouse models of hypertriglyceridemia, plasma cholesterol was disproportionately elevated (95 +/- 23 vs 73 +/- 23, P = 0.002, fasted and 90 +/- 24 vs 61 +/- 14, P < 0.0001, fed). Lipoprotein fractionation showed increased VLDL and IDL + LDL with an increased cholesterol/triglyceride ratio (0.114 vs 0.065, P = 0.02, in VLDL). The VLDL apo E/apo B ratio was decreased 3.4-fold (P = 0.05) and apo CII and apo CIII decreased in proportion to apo E. Triglyceride and apo B production rates were normal, but clearance rates of VLDL triglycerides and postlipolysis lipoprotein "remnants" were significantly slowed. Plasma apo B was significantly elevated. Unlike HuCII- and HuCIII-transgenic mice, VLDL from HuCI transgenic mice bound heparin-Sepharose, a model for cell-surface glycosaminoglycans, normally. In summary, apo CI overexpression is associated with decreased particulate uptake of apo B-containing lipoproteins, leading to increased levels of several potentially atherogenic species, including cholesterol-enriched VLDL, IDL, and LDL.

Lipoprotein metabolism during acute inhibition of hepatic triglyceride lipase in the cynomolgus monkey.

The role of the enzyme hepatic triglyceride lipase was investigated in a primate model, the cynomolgus monkey. Antisera produced against human postheparin hepatic lipase fully inhibited cynomolgus monkey posttheparin plasma hepatic triglyceride lipase activity. Lipoprotein lipase activity was not inhibited by this antisera. Hepatic triglyceride lipase activity in liver biopsies was decreased by 65-90% after intravenous infusion of this antisera into the cynomolgus monkey. After a 3-h infusion of the antisera, analytic ultracentrifugation revealed an increase in mass of very low density lipoproteins (S(f) 20-400). Very low density lipoprotein triglyceride isolated by isopycnic ultracentrifugation increased by 60-300%. Analytic ultracentrifugation revealed an increase in mass of lipoproteins with flotation greater than S(f) 9 (n = 4). The total mass of intermediate density lipoproteins (S(f) 12-20) approximately doubled during the 3 h of in vivo enzyme inhibition. While more rapidly floating low density lipoproteins (S(f) 9-12) increased, the total mass of low density lipoproteins decreased after infusion of the antibodies. The changes in high density lipoproteins did not differ from those in control experiments. In order to determine whether the increases of plasma concentrations of very low density lipoproteins were due to an increase in the rate of synthesis or a decrease in the rate of clearance of these particles, the metabolism of radiolabeled homologous very low density lipoproteins was studied during intravenous infusion of immunoglobulin G prepared from the antisera against hepatic triglyceride lipase (n = 3) or preimmune goat sera (n = 3). Studies performed in the same animals during saline infusion were used as controls for each immunoglobulin infusion. There was a twofold increase in the apparent half-life of the very low density lipoprotein apolipoprotein-B tracer in animals receiving the antibody, consistent with a decreased catabolism of very low density lipoproteins. Concomitantly, the rise in low density lipoprotein apoprotein-B specific activity was markedly delayed. None of these changes were observed during infusion of preimmune immunoglobulin G.Hepatic triglyceride lipase participates with lipoprotein lipase in the hydrolysis of the lipid in very low density lipoproteins, intermediate density lipoproteins, and the larger low density lipoproteins (S(f) 9-12). Thus, hepatic triglyceride lipase appears to function in a parallel role with lipoprotein lipase in the conversion of very low density and intermediate density lipoproteins to low density lipoproteins (S(f) 0-9).

Effect of heparin-induced lipolysis on the distribution of apolipoprotein e among lipoprotein subclasses. Studies with patients deficient in hepatic triglyceride lipase and lipoprotein lipase.

In normal subjects, apolipoprotein E (apo E) is present on very low density lipoproteins (VLDL) (fraction I) and on particles of a size intermediate between VLDL and low density lipoproteins (LDL) (fraction II). The major portion of apo E is, however, on particles smaller than LDL but larger than the average high density lipoproteins (HDL) (fraction III). To investigate the possible role of the vascular lipases in determining this distribution of apo E among the plasma lipoproteins, we studied subjects with primary deficiency of either hepatic lipase or of lipoprotein lipase and compared them with normal subjects. Subjects with familial hepatic triglyceride lipase deficiency (n = 2) differ markedly from normal in that fraction II is the dominant apo E-containing group of lipoproteins. When lipolysis of VLDL was enhanced in these subjects upon release of lipoprotein lipase by intravenous heparin, a shift of the apo E from VLDL into fractions II and III was observed. In contrast, apolipoproteins CII and CIII (apo CII and CIII, respectively) did not accumulate in intermediate-sized particles but were shifted markedly from triglyceride rich lipoproteins to HDL after treatment with heparin. In subjects with primary lipoprotein lipase deficiency (n = 4), apo E was confined to fractions I and III. Release of hepatic triglyceride lipase by heparin injection in these subjects produced a shift of apo E from fraction I to III with no significant increase in fraction II. This movement of apo E from large VLDL and chylomicron-sized particles occurred with little hydrolysis of triglyceride and no significant shift of apo CII or CIII into HDL from triglyceride rich lipoproteins. When both lipoprotein lipase and hepatic triglyceride lipase were released by intravenous heparin injection into normal subjects (n = 3), fraction I declined and the apo E content of fraction III increased by an equivalent amount. Either moderate or no change was noted in the intermediate sized particles (fraction II). These data strongly support the hypothesis that fraction II is the product of the action of lipoprotein lipase upon triglyceride rich lipoproteins and is highly dependent on hepatic triglyceride lipase for its further catabolism. In addition, the hydrolysis by hepatic triglyceride lipase of triglyceride rich lipoproteins in general results in a preferential loss of apo E and its transfer to a specific group of large HDL.

Increased plasma and renal clearance of an exchangeable pool of apolipoprotein A-I in subjects with low levels of high density lipoprotein cholesterol.

Plasma levels of HDL apo A-I are reduced in individuals with low HDL cholesterol (HDL-C) concentrations as a result of increased fractional catabolic rates (FCRs). To determine the basis for the high apo A-I FCRs, seven subjects with low HDL-C levels (31.0 +/- 4.3 mg/dl) were compared with three subjects with high HDL-C levels (72.0 +/- 4.5 mg/dl). Each subject received autologous HDL that was labeled directly by the iodine-monochloride method (whole-labeled) and autologous HDL that was labeled by exchange with homologous radiolabeled apo A-I (exchange-labeled). Blood was obtained for 2 wk, specific activities determined, and FCRs (d-1 +/- SD) estimated. In every subject, whether in the low or high HDL-C group, the exchange-labeled FCR was greater than the whole-labeled FCR. The exchange-labeled FCR was higher in the low HDL-C group (0.339 +/- 0.043) versus the high HDL-C group (0.234 +/- 0.047; P < 0.009). The whole-labeled FCR was also greater in the low HDL-C group (0.239 +/- 0.023) versus the high HDL-C group (0.161 +/- 0.064; P < 0.02). In addition, in both low and high HDL groups ultracentrifugation resulted in more radioactivity in d > 1.210 (as percentage of total plasma counts per minute) with the exchange-labeled tracer than with the whole-labeled tracer (12.55 +/- 4.95% vs. 1.02 +/- 0.38%; P < 0.003). With both HDL tracers, more radioactivity was found in d > 1.210 in the low versus the high HDL-C groups. When apo A-I catabolism was studied by perfusing isolated rabbit kidneys with whole-labeled HDL, there was twice as much accumulation (cpm/g cortex) of HDL apo A-I isolated from subjects with low HDL-C than from subjects with high HDL-C (P < 0.0025). Finally, HDL that had been isolated from subjects with high levels of HDL-C was triglyceride enriched and exposed to partially purified lipases before perfusion through kidneys. Threefold more apo A-I from modified HDL accumulated in the cortex compared with the unmodified preparation (P < 0.007). The results of these in vivo and ex vivo studies indicate that individuals with low HDL-C levels have more loosely bound, easily exchanged apo A-I and that this exchangeable apo A-I is more readily cleared by the kidney.

Mechanism of hypertriglyceridemia in human apolipoprotein (apo) CIII transgenic mice. Diminished very low density lipoprotein fractional catabolic rate associated with increased apo CIII and reduced apo E on the particles.

Hypertriglyceridemia is common in the general population, but its mechanism is largely unknown. In previous work human apo CIII transgenic (HuCIIITg) mice were found to have elevated triglyceride levels. In this report, the mechanism for the hypertriglyceridemia was studied. Two different HuCIIITg mouse lines were used: a low expressor line with serum triglycerides of approximately 280 mg/dl, and a high expressor line with serum triglycerides of approximately 1,000 mg/dl. Elevated triglycerides were mainly in VLDL. VLDL particles were 1.5 times more triglyceride-rich in high expressor mice than in controls. The total amount of apo CIII (human and mouse) per VLDL particle was 2 and 2.5 times the normal amount in low and high expressors, respectively. Mouse apo E was decreased by 35 and 77% in low and high expressor mice, respectively. Under electron microscopy, VLDL particles from low and high expressor mice were found to have a larger mean diameter, 55.2 +/- 16.6 and 58.2 +/- 17.8 nm, respectively, compared with 51.0 +/- 13.4 nm from control mice. In in vivo studies, radiolabeled VLDL fractional catabolic rate (FCR) was reduced in low and high expressor mice to 2.58 and 0.77 pools/h, respectively, compared with 7.67 pools/h in controls, with no significant differences in the VLDL production rates. In an attempt to explain the reduced VLDL FCR in transgenic mice, tissue lipoprotein lipase (LPL) activity was determined in control and high expressor mice and no differences were observed. Also, VLDLs obtained from control and high expressor mice were found to be equally good substrates for purified LPL. Thus excess apo CIII in HuCIIITg mice does not cause reduced VLDL FCR by suppressing the amount of extractable LPL in tissues or making HuCIIITg VLDL a bad substrate for LPL. Tissue uptake of VLDL was studied in hepatoma cell cultures, and VLDL from transgenic mice was found to be taken up much more slowly than control VLDL (P < 0.0001), indicating that HuCIIITg VLDL is not well recognized by lipoprotein receptors. Additional in vivo studies with Triton-treated mice showed increased VLDL triglyceride, but not apo B, production in the HuCIIITg mice compared with controls. Tissue culture studies with primary hepatocytes showed a modest increase in triglyceride, but not apo B or total protein, secretion in high expressor mice compared with controls. In summary, hypertriglyceridemia in HuCIIITg mice appears to result primarily from decreased tissue uptake of triglyceride-rich particles from the circulation, which is most likely due to increased apo CIII and decreased apo E on VLDL particles. the HuCIIITg mouse appears to be a suitable animal model of primary familial hypertriglyceridemia, and these studies suggest a possible mechanism for this common lipoprotein disorder.

Apolipoprotein B metabolism in subjects with deficiency of apolipoproteins CIII and AI. Evidence that apolipoprotein CIII inhibits catabolism of triglyceride-rich lipoproteins by lipoprotein lipase in vivo.

Previous data suggest that apolipoprotein (apo) CIII may inhibit both triglyceride hydrolysis by lipoprotein lipase (LPL) and apo E-mediated uptake of triglyceride-rich lipoproteins by the liver. We studied apo B metabolism in very low density (VLDL), intermediate density (IDL), and low density lipoproteins (LDL) in two sisters with apo CIII-apo AI deficiency. The subjects had reduced levels of VLDL triglyceride, normal LDL cholesterol, and near absence of high density lipoprotein (HDL) cholesterol. Compartmental analysis of the kinetics of apo B metabolism after injection of 125I-VLDL and 131I-LDL revealed fractional catabolic rates (FCR) for VLDL apo B that were six to seven times faster than normal. Simultaneous injection of [3H]glycerol demonstrated rapid catabolism of VLDL triglyceride. VLDL apo B was rapidly and efficiently converted to IDL and LDL. The FCR for LDL apo B was normal. In vitro experiments indicated that, although sera from the apo CIII-apo-AI deficient patients were able to normally activate purified LPL, increasing volumes of these sera did not result in the progressive inhibition of LPL activity demonstrable with normal sera. Addition of purified apo CIII to the deficient sera resulted in 20-50% reductions in maximal LPL activity compared with levels of activity attained with the same volumes of the native, deficient sera. These in vitro studies, together with the in vivo results, indicate that in normal subjects apo CIII can inhibit the catabolism of triglyceride-rich lipoproteins by lipoprotein lipase.

Overexpression of apolipoprotein CII causes hypertriglyceridemia in transgenic mice.

We have generated transgenic mice expressing the human apolipoprotein CII (apoCII) gene under the transcriptional control of the human cytochrome P-450 IA1 (CYPIA1) promoter. Human apoCII transgenic (HuCIITg) mice exhibited significant basal expression of the transgene (plasma apoCII level = 26.1 +/- 4 mg/dl) and showed further induction of transgene expression after treatment with beta-naphthoflavone. Unexpectedly, HuCIITg mice were hypertriglyceridemic and human apoCII levels correlated strongly to triglyceride levels (R = 0.89, P < 0.0001). Triglyceride levels (mg/dl +/- SEM) were elevated compared to controls in both the fed (804 +/- 113 vs 146 +/- 18, P < 0.001) and fasted (273 +/- 39 vs 61 +/- 4, P < 0.001) states. HuCIITg mice accumulated triglyceride-rich very low density lipoproteins (VLDL) with an increased apoC/apoE ratio. Tracer kinetic studies indicated delayed clearance of VLDL-triglyceride, and studies using Triton inhibition of VLDL clearance showed no increase in VLDL production. Plasma from these mice activated mouse lipoprotein lipase normally and radiolabeled VLDL were normally hydrolyzed. However, HuCIITg VLDL showed markedly decreased binding to heparin-Sepharose, suggesting that apoCII-rich, apoE-poor lipoprotein may be less accessible to cell surface lipases or receptors within their glycosaminoglycan matrices. HuCIITg mice are a promising model of hypertriglyceridemia that suggests a more complex role for apoCII in the metabolism of plasma triglycerides.

In vitro metabolism of apolipoprotein E.

Apolipoprotein E plays a major role in the uptake of chylomicrons and of very-low-density lipoprotein (VLDL) remnants by the liver. It has also been clearly demonstrated that apolipoprotein E rapidly and spontaneously exchanges between lipoproteins. To assess whether all lipoprotein-bound apolipoprotein E is available to participate in spontaneous transfer and/or exchange, the present study followed the fate of radiolabeled apolipoprotein E in an in vitro system. The results show that in vitro, apolipoprotein E can be considered as having both a spontaneously exchangeable pool and a nonexchangeable pool. Based upon specific radioactivity data, only a limited amount of apolipoprotein E originating in VLDL or in high-density lipoproteins (HDL) was capable of in vitro exchange with that in other lipoprotein fractions. Lipolysis of VLDL triacylglycerol by milk lipoprotein lipase, however, resulted in complete transfer of VLDL apolipoprotein E mass and radioactivity to HDL, supporting the potential for transformation of exchangeable apolipoprotein to a transferable pool in vivo. The results of these studies indicate that during the course of lipoprotein metabolism, conformational changes occur which alter the accessibility of apolipoprotein E. Such dynamic heterogeneity may have implications for the regulation of lipoprotein metabolism.

Immunoaffinity isolation of apolipoprotein E-containing lipoproteins.

Discrete apolipoprotein E-containing lipoproteins can be identified when EDTA plasma is fractionated on columns of 4% agarose. The present study has demonstrated, by physical and metabolic criteria, that these apolipoprotein E-containing lipoprotein subclasses may be further isolated by immunoaffinity chromatography. Whole plasma was first bound to an anti-apolipoprotein E immunoadsorbent prior to gel filtration on 4% agarose. After elution from the affinity column and dialysis, the bound fraction was chromatographed on 4% agarose. Discrete subfractions of apolipoprotein E could be demonstrated within elution volumes similar to those observed in the original plasma. When whole plasma was first submitted to gel filtration and the apolipoprotein E-containing lipoproteins of either intermediate- or of high-density lipoprotein (HDL) size were subsequently bound to anti-apolipoprotein E columns, the bound eluted fractions maintained their size and physical properties as shown by electron microscopy and by rechromatography on columns of 4% agarose. The metabolic integrity of apolipoprotein E-containing very-low-density lipoproteins (VLDL) was examined by coinjection into a cynomolgus monkey of 125I-labeled apolipoprotein E-rich and 131I-labeled apolipoprotein E-deficient human VLDL which had been separated by immunoaffinity chromatography. The plasma specific activity time curves of the apolipoprotein B in VLDL, intermediate-density (IDL) and low-density (LDL) lipoproteins demonstrated rates of decay and precursor-product relationships similar to those obtained after injection of whole labeled VLDL, supporting the metabolic integrity of VLDL isolated by immunoaffinity chromatography.

Diabetic dyslipidemia: basic mechanisms underlying the common hypertriglyceridemia and low HDL cholesterol levels.

Elevated levels of plasma triglycerides (TG) and reduced concentrations of HDL cholesterol are very common in patients with diabetes, particularly NIDDM. Although regulation of the plasma concentrations of VLDL, the major TG-rich lipoprotein is extremely complex, it is clear from in vivo kinetic studies that increased rates of secretion of VLDL into plasma is almost uniformly present in patients with NIDDM and hypertriglyceridemia. Recent studies at the cellular level indicate that increased fatty acid flux to the liver, also common in NIDDM (and other insulin-resistant states associated with elevated plasma TG levels), will stimulate the assembly and secretion of apoprotein (apo) B-containing lipoproteins by targeting apoB for secretion rather than intracellular degradation. Increased rates of secretion of VLDL into plasma appears to drive the exchange of TG from these lipoproteins for HDL cholesteryl ester. This exchange, which occurs in plasma, is facilitated by cholesteryl ester transfer protein, and generates a TG-enriched HDL that is a substrate for either hepatic lipase or lipoprotein lipase. When the TG in HDL is hydrolyzed, the resultant particle is smaller, and this appears to affect the binding of the major HDL protein, apoA-I. ApoA-I dissociates from the smaller, lipid-poor HDL, and the free apoA-I (molecular weight 28,000) can be filtered by the glomerulus in the kidney and most likely is degraded in renal tubular cells after reabsorption. Thus, increased free fatty acid transport in plasma, a common abnormality in insulin-resistant states, may be the underlying driving force for the two common lipid abnormalities seen in diabetes.

Investigation of insulin sensitivity in treated subjects with ketosis-prone diabetes mellitus.

Two similar intravenous infucion techniques have been utilized to investigate insulin sensitivity in young subjects with recent-onset ketosis-prone diabetes mellitus. All subjects presented initially with mild to moderately severe ketoacidosis and had been treated with daily insulin therapy for two to eight weeks at the time of study. Six diabetics and 10 normal subjects (group A) received intravenous infusions of glucose (6 mg./kg./min), insulin (80 mU./min.) for 150 minutes. Steady-state plasma glucose (SSPG) and insulin levels (SSPI) were reached by 90 minutes and maintained through the end of the study. As all subjects achieve simolas SSPI while simultaneously receiving similar glucose loads, the SSPG can be used to measure individual insulin sensitivity. Under these conditions the diabetics in group A had a mean SSPG (+/-S.E.) of 99+/-26 mg./100ml., which was not different from the level of 98 +/-14mg./100 ml. for their control subjects. Six diabetics and six normal subjects (group B) received infusions of only glucose (6 mg./kg./min.) and insulin (80 mU./min.). Similar SSPI levels were attained in both the diabetic and control subjects, and their mean SSPG (+/-S.E.) levels were not significantly different (83+/- 13 vs. 61 +/-6 mg./100 ml.). One diabetic in group A and two diabetics in group B had SSPG levels above the highest values measured in their control groups. However, all three had markedly elevated fasting plasma glucose levels on the day of study. In contrast, nine well-controlled diabetics had normal insulin sensitivity. These results suggest that well-controlled subjects with ketosis-prone diabetes mellitus have normal sensitivity to insulin.