Synthesis and secretion of nascent lipoprotein particles

TNF-alpha and IL-6 inhibit apolipoprotein A-IV production induced by linoleic acid in human intestinal Caco2 cells

Background

Apolipoprotein A-IV (apoA-IV) is a protein mainly synthesized by enterocytes in the intestine. Its gene expression is suppressed during fasting and stimulated during active fat absorption. Chronic feeding of a high-fat (HF) diet abolishes the differential expression between fasting and fat-feeding and therefore may contribute to diet-induced obesity since apoA-IV is a potent satiety factor. It is well established that the circulating pro-inflammatory cytokines TNF-α and IL-6 are increased by HF feeding.

Methods

To determine whether pro-inflammatory cytokines are involved in the diminished response of apoA-IV gene expression to fat-feeding, different concentrations of linoleic acid (LA), an important dietary fatty acid, was used to stimulate apoA-IV expression in human intestinal Caco2 cells. Cells were pre-treated with or without human recombinant TNF-α, IL-6 or their combination before the addition of LA. Real-time PCR and ELISA were used to detect and quantify RNA transcripts and proteins of apoA-IV and the cytokines.

Results

LA stimulated gene and protein expression of apoA-IV in a dose and time dependent manner. Pre-treatment with the cytokines for 72 h significantly inhibited the increased expression of apoA-IV gene and protein induced by LA. Furthermore, the cytokines, especially TNF-α, also positively up-regulate the cytokine themselves in Caco2 cells.

Conclusions

Our data indicate that the pro-inflammatory cytokines may be responsible for the reduced apoA-IV production in response to fat feeding. Because of apoA-IV’s role in satiety, we propose the inhibitory effect of circulating pro-inflammatory cytokines on apoA-IV production contributes to diet-induced obesity.

Retina expresses microsomal triglyceride transfer protein: implications for age-related maculopathy

The principal extracellular lesions of age-related maculopathy (ARM), the leading cause of vision loss in the elderly, involve Bruch's membrane (BrM), a thin vascular intima between the retinal pigment epithelium (RPE) and its blood supply. With age, 80-100 nm solid particles containing esterified cholesterol (EC) accumulate in normal BrM, and apolipoprotein B (apoB) immunoreactivity is detectable in BrM- and ARM-associated lesions. Yet little evidence indicates that increased plasma cholesterol is a risk factor for ARM. To determine if RPE is capable of assembling its own apoB-containing lipoprotein, we examined RPE for the expression of microsomal triglyceride transfer protein (MTP), which is required for this process. Consistent with previous evidence for apoB expression, MTP is expressed in RPE, the ARPE-19 cell line, and, unexpectedly, retinal ganglion cells, which are neurons of the central nervous system. De novo synthesis and secretion of neutral lipid by ARPE-19 was supported by high levels of radiolabeled EC and triglyceride in medium after supplementation with oleate. Lipoprotein assembly and secretion is implicated as a constitutive retinal function and a plausible candidate mechanism involved in forming extracellular cholesterol-containing lesions in ARM. The pigmentary retinopathy and neuropathy of abetalipoproteinemia (Mendelian Inheritance of Man 200100; Bassen-Kornzwieg disease), which is caused by mutations in the MTP gene, may involve loss of function at the retina.

Trans polyunsaturated fatty acids have more adverse effects than saturated fatty acids on the concentration and composition of lipoproteins secreted by human hepatoma HepG2 cells

The objective of this study was to assess the relative long-term effects of linoleic (cis, cis 18:2), linolelaidic (trans, trans 18:2), and palmitic (16:0) acids on hepatic lipoprotein production in HepG2 cells. All fatty acids increased the mass of triglycerides (TG) in the medium and the incorporation of [(3)H]-glycerol into secreted TG; the increase was more pronounced with linoleic acid than with linolelaidic and palmitic acids. The net accumulation in the medium of apolipoprotein (apo) A-I was not affected by the fatty acids tested and moderate changes in that of apoB resulted in apoB/apoA-I mass ratios of 1.05, 1.27 and 0.86 with linoleic, linolelaidic and palmitic acids, respectively. The incorporation of [(14)C]-acetate into cellular plus secreted total sterols was 9.1%, 33.6% and 17.4% of total [(14)C]-labeled lipids with linoleic, linolelaidic and palmitic acids, respectively. Relative to linoleic acid, palmitic acid, and to a greater extent (P < 0.05) linolelaidic acid, increased the secretion and cellular accumulation of [(14)C]-labeled free cholesterol (FC) and cholesteryl esters and decreased those of TG and phospholipids (PL). Compared with linoleic acid, linolelaidic acid increased LDL-cholesterol (C) and HDL-C by 154% (P < 0.001) and 50% (P = 0.016), respectively, whereas palmitic acid increased LDL-C by 17% (P > 0.1) and did not affect HDL-C. The LDL-C to HDL-C ratios were 0.70, 1.18 and 0.96 with linoleic, linolelaidic and palmitic acids, respectively. These differences were not due to altered LDL receptor activity. The PL to C ratios of HDL particles were 1.61, 0.40 and 0.77 with linoleic acid, linolelaidic acid and palmitic acid, respectively. These results suggest that relative to cis polyunsaturated and saturated fatty acids, trans PUFA more adversely affect the concentration and composition of apoA-I- and apoB-containing lipoproteins secreted by HepG2 cells.

Perturbation of lipid metabolism by linoleic acid hydroperoxide in CaCo-2 cells

Dietary hydroperoxides are being discussed as potential health hazards contributing to oxidative stress-related diseases. However, how food-born hydroperoxides could exert systemic effects remains elusive in view of the limited chances to be absorbed. Therefore, the metabolic fate of 13-HPODE (13-hydroperoxy octadecadienoic acid), 13-HODE (13-hydroxy octadecadienoic acid) and linoleic acid (LA) was investigated in a CaCo-2 cell monolayer as a model of the intestinal epithelium. [1-14C]-13-HPODE, up to a non-cytotoxic concentration of 100 microM, did not cross the CaCo-2 cell monolayer unreduced if applied to the luminal side. The [1 -14C]-HPODE-derived radioactivity was preferentially recovered from intracellular and released diacylglycerols (DG), phospholipids (PL) and cholesterol esterified with oxidized fatty acids (oxCE). A similar distribution pattern was obtained with 13-HODE. In contrast, LA is preferentially incorporated into triacylglycerols (TG), cholesteryl esters (CE) and PL (but mainly released as TG). 13-HPODE dose-dependently decreased the incorporation of LA into released TG, while LA accumulated in cellular and released DGs, effects similarily exerted by 13-HODE. We concluded that food-born hydroperoxy fatty acids are instantly reduced by the gastrointestinal glutathione peroxidase, which was previously shown to persist in selenium deficiency. Accordingly, modulation of the glutathione peroxidases by selenium deprivation/repletion did not modify the disturbance of the lipid metabolism by 13-HPODE. Thus, hydroperoxy fatty acids disturb intestinal lipid metabolism by being esterified as hydroxy fatty acids into complex lipids, and may render lipoproteins synthesized thereof susceptible to further oxidative modifications.

Reduced secretion of triacylglycerol in CaCo-2 cells transfected with intestinal fatty acid-binding protein

The fatty acid-binding proteins are hypothesized to be involved in cellular fatty acid transport and trafficking. We established CaCo-2 cells stably transfected with intestinal fatty acid-binding protein (I-FABP) and examined how the expression of this protein may influence fatty acid metabolism. I-FABP expression was detectable in I-FABP-transfected cells, whereas parent CaCo-2 cells as well as mock-transfected cells failed to express detectable levels of I-FABP mRNA or protein at any stage of differentiation. For studies of lipid metabolism, cells were incubated with [14C]oleic acid in taurocholate micelles containing monoolein, and distribution of labeled fatty acid in cellular and secreted lipids was examined. In one transfected cell clone, expressing the highest level of I-FABP, labeled cellular triacylglycerol increased approximately twofold as compared to control cells. The level of intracellular triacylglycerol in two other I-FABP-transfected clones resembled that of control cells. However, secretion of triacylglycerol was markedly reduced in all the I-FABP-expressing cell lines. Our data suggest that increased expression of I-FABP leads to reduced triacylglycerol secretion in intestinal cells.

Stearoyl-CoA desaturase 1 coding sequences and antisense RNA affect lipid secretion in transfected chicken LMH hepatoma cells

Hepatic stearoyl CoA desaturase (SCD) activity in chickens from a fat line is higher than that of chickens from a lean line and correlates with plasma triacylglycerol concentrations. Furthermore, in these lines, the hepatic SCD1 mRNA level is positively correlated with the adipose tissue weight. To analyze the contribution of the SCD1 gene in the regulation of adiposity in the early stages of triacylglycerol secretion, SCD1 coding sequence and antisense RNA expression vectors were transfected in LMH cells. After selection, these cells were analyzed with regard to SCD1 expression and lipid secretion. The amounts of secreted triacylglycerols and phospholipids were shown to be higher in LMH cells transfected with the SCD1 gene, but reduced in those transfected with the SCD1 antisense sequences when compared to cells transfected with the vector alone (without SCD1 sequences). These results provide direct evidence that the expression of the SCD1 gene plays a major role in the triacylglycerol and phospholipid secretion process.

A proposed model for the assembly of chylomicrons

The intestine synthesizes very low density lipoproteins (VLDL) and chylomicrons (CM) to transport fat and fat-soluble vitamins into the blood. VLDL assembly occurs constitutively whereas CM assembly is a characteristic property of the enterocytes during the postprandial state. The secretion of CM is specifically inhibited by Pluronic L81. CM are very heterogeneously-sized particles that consist of a core of triglycerides (TG) and cholesterol esters and a monolayer of phospholipids (PL), cholesterol and proteins. The fatty acid composition of TG, but not PL, in CM mirrors the fatty acid composition of fat in the diet. CM assembly is deficient in abetalipoproteinemia and CM retention disease. Abetalipoproteinemia results due to mutation in the mttp gene and is characterized by the virtual absence of apoB-containing lipoproteins in the plasma. Patients suffer from neurologic disorders, visual impairment, and exhibit acanthocytosis. CM retention disease, an inherited recessive disorder, is characterized by chronic diarrhea with steatorrhea in infancy, abdominal distention and failure to thrive. It is caused by a specific defect in the secretion of intestinal lipoproteins; secretion of lipoproteins by the liver is not affected. Besides human disorders, mice that do not assemble intestinal lipoproteins have been developed. These mice are normal at birth, but defective in fat and fat-soluble vitamin absorption, and fail to thrive. Thus, fat and fat-soluble vitamin transport by the intestinal lipoproteins is essential for proper growth and development of neonates. Recently, differentiated Caco-2 cells and rabbit primary enterocytes have been described that synthesize and secrete CM. These cells can be valuable in distinguishing between the two different models proposed for the assembly of CM. In the first model, the assembly of VLDL and CM is proposed to occur by two 'independent' pathways. Second, CM assembly is proposed to be a product of 'core expansion' that results in the synthesis of lipoproteins of different sizes. According to this model, intestinal lipoprotein assembly begins with the synthesis of 'primordial' lipoprotein particles and involves release of the nascent apoB with PL derived from the endoplasmic reticulum (ER) membrane. In addition, TG-rich 'lipid droplets' of different sizes are formed independent of apoB synthesis. The fusion of lipid droplets and primordial lipoproteins results in the formation of different size lipoproteins due to the 'core expansion' of the primordial lipoproteins.

Assembly and secretion of chylomicrons by differentiated Caco-2 cells. Nascent triglycerides and preformed phospholipids are preferentially used for lipoprotein assembly

To develop a cell culture model for chyclomicron (CM) assembly, the apical media of differentiated Caco-2 cells were supplemented with oleic acid (OA) together with either albumin or taurocholate (TC). The basolateral media were subjected to sequential density gradient ultracentrifugations to obtain large CM, small CM, and very low density lipoproteins (VLDL), and the distribution of apoB in these fractions was quantified. In the absence of OA, apoB was secreted as VLDL/LDL size particles. Addition of OA (>/=0.8 mM) with TC, but not with albumin, resulted in the secretion of one-third of apoB as CM. Lipid analysis revealed that half of the secreted phospholipids (PL) and triglycerides (TG) were associated with CM. In CM, TG were 7-11-fold higher than PL indicating that CM were TG-rich particles. Secreted CM contained apoB100, apoB48, and other apolipoproteins. Secretion of large CM was specifically inhibited by Pluronic L81, a detergent known to inhibit CM secretion in animals. These studies demonstrate that differentiated Caco-2 cells assemble and secrete CM in a manner similar to enterocytes in vivo. Next, experiments were performed to identify the sources of lipids used for lipoprotein assembly. Cells were labeled with [3H]glycerol for 12 h, washed, and supplemented with OA, TC, and [14C] glycerol for various times to induce CM assembly and to radiolabel nascent lipids. TG and PL were extracted from cells and media and the association of preformed and nascent lipids with lipoproteins was determined. All the lipoproteins contained higher amounts of preformed PL compared with nascent PL. VLDL contained equal amounts of nascent and preformed TG, whereas CM contained higher amounts of nascent TG even when nascent TG constituted a small fraction of the total cellular pool. These studies indicate that nascent TG and preformed PL are preferentially used for CM assembly and provide a molecular explanation for the in vivo observations that the fatty acid composition of TG, but not PL, of secreted CM reflects the composition of dietary fat. It is proposed that in the intestinal cells the preformed PL from the endoplasmic reticulum bud off with apoB as primordial particles and the assembly of larger lipoproteins is dependent on the synthesis and delivery of nascent TG to these particles.

Assembly and secretion of VLDL in nondifferentiated Caco-2 cells stably transfected with human recombinant ApoB48 cDNA

Intestinal cells secrete apoB48-containing very low density lipoproteins (VLDLs) and chylomicrons for the transport of biliary and dietary lipids. The molecular mechanisms regulating the assembly of intestinal lipoproteins are not known due to a lack of reliable and specific cell culture models. Caco-2 (a human colon carcinoma) cells have been used to study intestinal lipid metabolism. These cells have been shown to secrete both apoB100- and apoB48-containing triglyceride (TG)-rich lipoproteins only after differentiation into enterocyte-like cells. To study lipoprotein assembly in nondifferentiated Caco-2 cells, we stably expressed human recombinant apoB48 cDNA under the control of a constitutive cytomegalovirus promoter. Pulse-chase analysis revealed that the majority (> 50%) of apoB48 synthesized was degraded intracellularly in the presence or absence of oleic acid. Transfected nondifferentiated cells secreted lipoproteins with flotation densities similar to those of plasma HDL or LDL when cultured in serum-free or serum-containing media, respectively. Incubation of cells with media containing serum and oleic acid resulted in the secretion of VLDL-like particles. Secretion of VLDL was inhibited (> 80%) by triacsin C due to > 60% inhibition of oleate-induced TG synthesis. However, inhibition of cholesteryl ester synthesis by 70% with an acyl coenzyme A:cholesterol acyltransferase inhibitor did not affect VLDL secretion. Efficient assembly of lipoproteins usually requires the microsomal TG transfer protein (MTP). The presence of MTP in transfected Caco-2 cells was investigated by measuring TG transfer activity in microsomal fractions. Microsomal fractions had 0.2% TG transfer activity per hour per microgram of protein, which corresponds to 30% to 60% of the MTP activity present in liver-derived cells. To determine whether MTP activity was required for lipoprotein assembly, transfected cells were incubated in the presence of the MTP inhibitor CP-10,447. This compound completely abolished the secretion of apoB. These data show that the transfected cell lines secrete lipoproteins of different densities under different culture conditions and that the assembly of larger VLDL particles requires active TG synthesis and MTP activity. Thus, in nondifferentiated Caco-2 cells, the amount of apoB secreted and not the MTP activity is the limiting factor for lipoprotein assembly.

Lipid, apolipoprotein, and lipoprotein synthesis and secretion during cellular differentiation in Caco-2 cells

Although Caco-2 cells are frequently employed for the study of enterocyte lipid metabolism, variable results have been reported regarding their ability to synthesize and secrete lipids and apolipoproteins. The major goal of this investigation is to examine the capacity of Caco-2 cells to elaborate and secrete lipids, lipoproteins, and apolipoproteins at different degrees of morphological and functional differentiation. Cells were cultured in medium with 5% fetal bovine serum (FBS), on permeable polycarbonate filters from 2 to 30 d in the presence of 14C-oleate or 35S-methionine. Cellular differentiation, as assessed by morphology (light and electron microscopy), transepithelial resistance, free fatty acid flux, and sucrase activity, progressed steadily up to 20 d of culture. Caco-2 cells esterified oleic acid mainly into phospholipids, triglycerides (TG), and smaller amounts of cholesterol esters. Lipid synthesis began as early as 2 d, and TG secretion was enhanced with increased duration of culture. However, very low efficiency of lipid export was observed at all levels of differentiation, reaching a maximum of only 6% of intracellular lipids. VLDL and LDL were the dominant lipoproteins secreted, with HDL comprising < 20% of the total. VLDL secretion increased, while LDL decreased, whereas the lipid composition of lipoproteins varied little with increasing duration of culture. Apoprotein B and A-I synthesis and secretion increased markedly from 11 to 20 d of culture. The ratio of apo B-100/B-48 decreased between 11 and 30 d, consistent with enhanced apo B editing of more mature enterocytes. Taken together, our data suggest that from 20 d of culture, Caco-2 cells are morphologically and functionally mature, capable of lipid esterification, and lipoprotein and apolipoprotein synthesis. However, despite their functional and morphological similarities to mature enterocytes, Caco-2 cells have a very limited lipid export capacity.

Differences in hypolipidaemic effects of two statins on Hep G2 cells or human hepatocytes in primary culture

1. The objective of this study was to compare in cultured human hepatocytes or Hep G2 cells, changes in the fate of unesterified low density lipoprotein (LDL)-cholesterol induced by crilvastatin, a new cholesterol lowering drug and a reference statin, simvastatin. 2. The experiments were carried out for 20 h, each well contained 4.2 x 10(5)/cm2 Hep G2 cells or 0.5 x 10(5)/Cm2 human hepatocytes, 130 microM ursodeoxycholate, 0.68 microCi or 1.59 microCi unesterified human [14C]-LDL-cholesterol, crilvastatin or simvastatin at 0 or 50 microM (both cell types) or 300 microM (Hep-G2 cells). Incubation with the two drugs resulted in increased amounts of unesterified [14C]-LDL-cholesterol taken by the two cell types, compared to control. 3. Crilvastatin 50 microM led to significantly higher quantities of [14C]-glyco-tauro-conjugated bile salts, compared to simvastatin. Statins reduced the apo B100 level secreted by the two cell types (simvastatin) or human hepatocytes (crilvastatin). Crilvastatin enhanced both the level of apo A1 secreted by the Hep G2 cells and the level of APF, a high density lipoprotein (HDL) and biliary apoprotein. 4. Crilvastatin not only acts by stimulating LDL-cholesterol uptake by hepatocytes, but also by enhancing the catabolism of LDL-cholesterol in bile salts and probably by stimulating HDL and/or bile component secretion. Such a mechanism was not previously described for HMG CoA reductase inhibitors. Our results on APF show that this apoprotein could be considered also as an indicator of changes in bile and/or HDL compartments. 5. The human hepatocyte model appeared to be a suitable and relevant model in the pharmacological-metabolic experiments carried out in this study. It led to more consistent data than those obtained with Hep G2 cells.

Chylomicron assembly and catabolism: role of apolipoproteins and receptors

Chylomicrons are lipoproteins synthesized exclusively by the intestine to transport dietary fat and fat-soluble vitamins. Synthesis of apoB48, a translational product of the apob gene, is required for the assembly of chylomicrons. The apob gene transcription in the intestine results in 14 and 7 kb mRNAs. These mRNAs are post-transcriptionally edited creating a stop codon. The edited mRNAs chylomicrons from the shorter apoB48 peptide remains to be elucidated. In addition, the roles of proteins involved in the assembly pathway, e.g. apobec-1, MTP and apoA-IV, needs to be studied. Cloning of enzymes involved in the intestinal biosynthesis of triglycerides will be crucial to fully appreciate the assembly of chylomicrons. There is a need for cell culture and transgenic animal models that can be used for intestinal lipoprotein assembly. The catabolism of chylomicrons is far more complex and efficient than the catabolism of VLDL. Even though the major steps involved in the catabolism of chylomicrons are now known, the determinants for apolipoprotein exchange, processing of remnants in the space of Disse, as well as the mechanism of uptake of these particles by extra-hepatic tissue needs further exploration.

Nascent VLDL phospholipid composition is altered when phosphatidylcholine biosynthesis is inhibited: evidence for a novel mechanism that regulates VLDL secretion

Previous work has shown that inhibition of phosphatidylcholine biosynthesis inhibits very low density lipoprotein (VLDL) secretion by causing a decrease in the number of particles in the Golgi but not in the endoplasmic reticulum of rat liver (Verkade et al. (1993) J. Biol. Chem. 268, 24,990-24,996). One explanation for this observation was that VLDL from choline deficient livers was degraded in a post-endoplasmic reticulum compartment. This hypothesis was supported by experiments in which choline deficient (CD) or choline supplemented (CS) rat hepatocytes were incubated +/- Brefeldin A. In the presence of Brefeldin A, VLDL secretion was blocked, but no difference was observed in the degradation of apolipoprotein B (apoB) within the CD or CS cells. If increased catabolism of apoB were occurring in the endoplasmic reticulum of CD hepatocytes, enhanced degradation of apoB in CD cells might have been expected. Inhibition of phosphatidylcholine biosynthesis also caused decreases in the phosphatidylcholine content of membranes of the secretory pathway. The lipids of nascent VLDLs from the lumina of endoplasmic reticulum and Golgi prepared from CD rat liver were relatively enriched in phosphatidylethanolamine and depleted of phosphatidylcholine when compared to samples from CS liver. The changes in nascent VLDL phospholipid composition mimicked that of the organelle membranes from which the VLDLs were isolated. Possibly the phospholipid composition of the organelles is a factor in determining the final phospholipid composition of VLDLs. One hypothesis is that when phosphatidylcholine biosynthesis is impaired, nascent VLDL is assembled incorrectly and degraded by a quality control protease in a post-endoplasmic reticulum compartment.

Insulin regulation of triacylglycerol-rich lipoprotein synthesis and secretion

This review has considered a number of observations obtained from studies of insulin in perfused liver, hepatocytes, transformed liver cells and in vivo and each of the experimental systems offers advantages. The evaluation of insulin effects on component lipid synthesis suggests that overall, lipid synthesis is positively influenced by insulin. Short-term high levels of insulin through stimulation of intracellular degradation of freshly translated apo B and effects on synthesis limit the ability of hepatocytes to form and secrete TRL. The intracellular site of apo B degradation may involve membrane-bound apo B, cytoplasmic apo B and apo B which has entered the ER lumen. How insulin favors intracellular apo B degradation is not known. An area of recent investigation is in insulin-stimulated phosphorylation of intracellular substrates such as IRS-1 which activates insulin specific cellular signaling molecules [245]. Candidate molecules to study insulin action on apo B include IRS-1 and SH2-containing signaling molecules. Insulin dysregulation in carbohydrate metabolism occurs in non-insulin-dependent diabetes mellitus due to an imbalance between insulin sensitivity of tissue and pancreatic insulin secretion (reviewed in Refs. [307,308]). Insulin resistance in the liver results in the inability to suppress hepatic glucose production; in muscle, in impaired glucose uptake and oxidation and in adipose tissue, in the inability to suppress release of free FA. This lack of appropriate sensitivity towards insulin action leads to hyperglycemia which in turn stimulates compensatory insulin secretion by the pancreas leading to hyperinsulinemia. Ultimately, there may be failure of the pancreas to fully compensate, hyperglycemia worsens and diabetes develops. The etiology of insulin resistance is being intensively studied for the primary defect may be over secretion of insulin by the pancreas or tissue insulin resistance and both of these defects may be genetically predetermined. We suggest that, in addition to effects in carbohydrate metabolism, insulin resistance in liver results in the inability of first phase insulin to suppress hepatic TRL production which results in hypertriglyceridemia leading to high levels of plasma FA which accentuate insulin resistance in other target organs. As recently reviewed [17,254] the role of insulin as a stimulator of hepatic lipogenesis and TRL production has been long established. Several lines of evidence support that insulin is stimulatory to the production of hepatic TRL in vivo. First, population based studies support a positive relationship between plasma insulin and total TG and VLDL [253]. Second, there is a strong association between chronic hyperinsulinemia and VLDL overproduction [309].(ABSTRACT TRUNCATED AT 400 WORDS)

Opposite in vitro and in vivo regulation of hepatic apolipoprotein A-I gene expression by retinoic acid. Absence of effects on apolipoprotein A-II gene expression

We studied the pharmacological potential of retinoids to modulate apolipoprotein (apo) A-I and apoA-II gene expression and production in vitro in the human cell line HepG2 as well as in primary cultures of adult rat hepatocytes and in vivo in the rat. In HepG2 cells, addition of all-trans retinoic acid (RA) doubled apoA-I mRNA within 24 hours and protein secreted in the culture medium after 48 hours. The induction of apoA-I mRNA by RA was completely blocked by actinomycin D, suggesting that RA acts at the transcriptional level in HepG2 cells. In primary cultures of rat hepatocytes, addition of RA increased apoA-I mRNA in a dose- and time-dependent manner as well as the secretion of apoA-I protein. Similar changes in apoA-I mRNA were observed with 9-cis RA. However, in vivo, hepatic apoA-I mRNA levels decreased after a single administration of RA at 10 mg/kg and remained low after prolonged treatment or at a higher dose, and serum apoA-I concentrations did not change. Furthermore, RA treatment did not substantially affect apoA-II mRNA levels or protein secretion either in vitro or in vivo. As a control, RA receptor-beta mRNA levels increased after RA both in vitro and in vivo. In conclusion, RA treatment selectively induces apoA-I and not apoA-II expression in vitro but not in vivo. These results therefore show additional regulatory effects of RA on apoA-I gene expression in vivo and raise questions about the usefulness of RA in the treatment of atherosclerosis.