Lipoprotein Secretion and Triglyceride Stores in the Heart*

The genes for apolipoprotein B and microsomal triglyceride transfer protein are expressed in mouse and human heart tissue. Why the heart would express these “lipoprotein assembly” genes has been unclear. Here we demonstrate that the beating mouse heart actually secretes spherical lipoproteins. Moreover, increased cardiac production of lipoproteins (e.g., in mice that express a human apolipoprotein B transgene) was associated with increased triglyceride secretion from the heart and decreased stores of triglycerides within the heart. Increased cardiac production of lipoproteins also reduced the pathological accumulation of triglycerides that occurs in the hearts of mice lacking long-chain acyl coenzyme A dehydrogenase. In contrast, blocking heart lipoprotein secretion (e.g., in heart-specific microsomal triglyceride transfer protein knockout mice) increased cardiac triglyceride stores. Thus, heart lipoprotein secretion helps regulate cardiac triglyceride stores and may protect the heart from the detrimental effects of surplus lipids.

Apolipoprotein (apo) 1 B and microsomal triglyceride transfer protein (MTP) play critical roles in the assembly and secretion of lipoproteins in hepatocytes and intestinal enterocytes (1)(2)(3). The main purpose of lipoprotein secretion by the liver and intestine is to transport triglycerides to peripheral tissues, mainly to adipose tissue for storage or to heart and skeletal muscle for use as fuel (4). If lipoprotein secretion from hepatocytes or enterocytes is abolished, for example, by null mutations in the apoB or MTP genes, there is a striking accumulation of triglycerides within those cells (2,5,6).
The concept that the function of apoB and MTP is to export lipids from the liver and intestine is well documented and accepted, but several reports have suggested that this view may be too narrow (7)(8)(9). We recently reported that apoB and MTP are also expressed in the mouse heart, and that the full-length apoB protein, apoB100, is translated by minced-up pieces of mouse hearts (7,8). Substantially more apoB100 is synthesized by the hearts of human apoB transgenic mice (HuBTg ϩ/o ) created with large fragments of human genomic DNA (7,9,10). The HuBTg ϩ/o mice have cardiac levels of the apoB mRNA that range from 1 to 8% of the hepatic level. Both the apoB and MTP genes are also expressed in the human heart, and minced-up pieces of human heart synthesize apoB100 (7). ApoB mRNA levels in human hearts are comparable with those in the hearts of HuBTg ϩ/o mice. Not surprisingly, apoB sequences have been encountered during the sequencing of clones in human heart cDNA libraries (www.ncbi.nlm.nih.gov/dbEST).
The expression of "lipoprotein assembly" genes in the heart led us to hypothesize that the heart actually synthesizes and secretes lipoproteins (7,8). The concept that the heart might export lipids was remarkable because the heart has always been viewed as a prodigious importer of lipids from the plasma lipoproteins produced in the liver and intestine (4). We further hypothesized that lipoprotein secretion by cardiac myocytes might function to unload surplus lipids from cells. A mechanism for exporting surplus lipids would make teleological sense, given that elevated levels of fatty acids and fatty acid intermediates can have deleterious effects on cardiac function (11)(12)(13) and that increased amounts of triglyceride storage can compromise contractile function and cause cell death (14,15).
In the current study, we sought to determine whether the beating heart actually secretes spherical lipoproteins containing a core of neutral lipids and, if so, to assess the effect of lipoprotein secretion on myocardial lipid homeostasis. To address the possibility that lipoproteins are synthesized by the heart, we used a Langendorff apparatus to perfuse mouse hearts and then examined the perfusion medium for secreted lipoproteins and lipids. To assess the impact of lipoprotein secretion on cardiac lipid homeostasis in vivo, we measured cardiac triglyceride stores in genetically modified mice that had increased levels of cardiac lipoprotein production and in mice with a blockade of cardiac lipoprotein production.

EXPERIMENTAL PROCEDURES
Mouse Models-Hemizygous human apoB transgenic mice (HuBTg ϩ/o ) generated with an 80-kb human genomic clone (p158) have been described (16). The HuBTg ϩ/o mice, which express human apoB in the liver and the heart but not in the intestine (9,16,17), were backcrossed to C57BL/6 mice more than 14 times. Long-chain acyl coenzyme A dehydrogenase-deficient (encoded Acadl Ϫ/Ϫ ) mice have been described (18 Mice homozygous for a conditional allele of the microsomal triglyceride transfer protein gene (Mttp flox/flox ) were described previously (5). All mice except the HuBTg ϩ/o mice had a mixed genetic background. The mice were housed in a pathogen-free barrier facility with a 12-h light/ dark cycle and fed rodent chow containing 4.5% fat (Ralston Purina, St. Louis, MO). Genotypes were determined by Southern blot analysis or by polymerase chain reaction (PCR).
Heart-specific Mttp knockout mice (Mttp flox/flox ␣MHC-Cre ϩ/o ) were generated by breeding Mttp flox/flox mice with ␣MHC-Cre transgenic mice. Recombination within Mttp was assessed by Southern blot analysis of SacI-cleaved genomic DNA (5). Mttp mRNA levels in the hearts of Mttp flox/flox ␣MHC-Cre ϩ/ mice were assessed by real-time, quantitative reverse transcription (RT)-PCR. In brief, 25 ng of total heart RNA and 10 mM of oligonucleotide probes and primers were mixed, and real-time RT-PCR was carried out with the platinum quantitative RT-PCR thermoscript one-step system (Life Technologies, Inc.). The Mttp oligonucleotide primers were 5Ј-CACTCTTGGAGAAACGGTCATA-ATT-3Ј and 5Ј-CACTCTTGGAGAAACGGTCATAATT-3Ј, and the probe was 5Ј-CGTCGAGTTCTCAAGGAGATGGCTGTTC-3Ј (PE Applied Biosystems, Foster City, CA). For normalization of heart Mttp mRNA levels, a glyceraldehyde-3-phosphate dehydrogenase control RT-PCR kit was used (Taqman; PE Applied Biosystems). The RT-PCR reaction was performed for 30 min at 60°C, followed by a heat inactivation step (10 min at 95°C) and then real-time PCR (40 cycles of 95°C for 15 s and 60°C for 60 s) on an ABI Prism 7700 Sequence Detection System (PE Applied Biosystems).
Examination of Lipoprotein Particles by Electron Microscopy-Fresh plasma samples (0.5 ml) and heart perfusion medium samples (9 ml) were adjusted to d 1.10 g/ml with KBr and overlaid with 3 ml of a d 1.063 g/ml KBr solution. The samples were then ultracentrifuged in an SW40 Ti rotor (Beckman Instruments, Palo Alto, CA) at 40,000 rpm for 18 h at 12°C. The top 0.5 ml from the ultracentrifugation tubes was removed and adjusted to d 1.10 g/ml and a volume of 3 ml. The samples were again overlaid with 3 ml of a d 1.063-g/ml solution and ultracentrifuged in an SW55 Ti rotor (Beckman Instruments) for 18 h at 12°C. A carbon grid was placed on the liquid surface and then examined for lipoproteins by negative-staining electron microscopy (21,22).
Triglyceride Secretion from Isolated Mouse Hearts-To assess cardiac triglyceride secretion, hearts were perfused with continuously circulating KHB medium containing 1.2 mM [ 3 H]oleate (Amersham Pharmacia Biotech) bound to 3% albumin (specific activity, 5 mCi/liter). Samples of the perfusate were collected at baseline and at 15, 30, and 60 min. After lyophilization, lipids were extracted with 2:1 choloroform/methanol (v/v) and separated by thin layer chromatography (TLC). The fatty acid and triglyceride radioactivities were quantified in a scintillation counter. The sample activity of 3 H-triglycerides was normalized to the dry weight of the heart and to the sample activity of [ 3 H]oleate, the latter to compensate for differences in recovery between samples. The d 1.063g/ml fractions of perfusion medium samples obtained at 60 min were prepared as described above for the electron microscopy studies. At the end of the second ultracentrifugation, lipids within the top 0.5 ml of the density gradient were extracted with 3:2 hexane/isopropanol (v/v) and separated by TLC, and the triglyceride radioactivity was quantified in a scintillation counter.
Mouse Cardiac Myocyte Cultures-Cardiac myocytes were prepared from 1-5-day-old neonatal HuBTg ϩ/o and nontransgenic hearts by serial trypsinization of pieces of heart tissue as described (23). The cells were preplated for 90 min at 37°C in minimum essential medium with Hank's basic salt solution, 50 units/ml penicillin (Sigma), vitamin B12 (Sigma), and 5% bovine calf serum (HyClone Laboratories, Logan, UT). This preplating step allowed the noncardiac myocytes to adhere to the plastic dish. Cardiac myocytes were plated at a density of 400 cells/mm 2 in dishes coated with collagen I (Biocoat Collagen I Cellware; Becton Dickinson Laboratories, Bedford, MA). The medium was changed daily for 4 days. The expression of human apoB in cardiac myocytes from HuBTg ϩ/o mice was assessed by RT-PCR (Prostar RT-PCR kit; Stratagene, La Jolla, CA). The human apoB oligonucleotide primers were 5Ј-GAAGAACTTCCGGAGAGTTGCAAT-3Ј and 5Ј-CTCTTAGCCCCA-TTCAGCTCTGAC-3Ј.
Cardiac myocytes were incubated for 5 or 24 h with 1.0 ml of medium containing 10 mM [ 14 C]oleate (0.40 mCi/mmol; Amersham Pharmacia Biotech) bound to 10% albumin. The medium was then removed, and the cells were quickly washed three times with ice-cold phosphatebuffered saline containing fatty acid-free bovine serum albumin (0.2%) and then washed twice for 10 min with the same buffer. Cardiac myocytes were washed two more times with ice-cold phosphate-buffered saline. After the addition of 10 ml of an internal triglyceride standard, glycerol tri-[ 3 H]oleic acids (20 mCi/mmol; Amersham Pharmacia Biotech), lipids were extracted from the cells and the medium with 3:2 hexane/isopropanol (v/v). Triglycerides were identified by TLC, and the triglyceride radioactivity was quantified in a scintillation counter. Total cell protein was extracted by adding 0.5 ml of 0.1 M NaOH to each well for 10 min and assayed by a colorimetric assay using bovine serum albumin as a standard (Bio-Rad Laboratories, Hercules, CA). 14 C-triglycerides were normalized to the recovery of the 3 H-triglyceride standard and expressed as disintegrations per minute per milligram of cell protein.
Heart and Liver Lipid Stores-Heart and liver triglyceride stores were determined in mice that had been fasted for 18 h. All comparisons were performed on age-and sex-matched littermates. Hearts (dissected free of all vessels and pericardial fat) and ϳ100-mg liver pieces were homogenized (Ultra-Turbax T8; VWR) in the presence of a known amount of tripentadecanoic acid (Sigma) as an internal standard (24). Lipids were extracted with 3.0 ml of 2:1 chloroform/methanol. Triglycerides were identified by TLC, transesterified with methanolic HCl, and quantified by gas chromatography using the internal standard (24). Triglyceride measurements were normalized to the weight of each tissue (millimoles of triglycerides per gram of tissue).
Thiobarbituric acid-reactive substances, which have been used frequently to gauge levels of lipid peroxides (25,26), were determined in ϳ50-mg pieces of hearts from HuBTg ϩ/o Acadl Ϫ/Ϫ and Acadl Ϫ/Ϫ mice that had been fasted for 72 h. To limit the peroxidation of lipids during the procedure, hearts were homogenized in 1.15% KCl solution containing 50 mM desferrioxamine (Sigma).
Plasma Lipid and Glucose Levels-Plasma concentrations of cholesterol, free fatty acids, and triglycerides after an 18-h fast were determined by colorimetric assays (Spectrum Cholesterol Assay; Abbott Laboratories, Abbott Park, IL; free fatty acids, half micro test; Roche Diagnostics, Indianapolis, IN; and triglycerides, Roche Diagnostics). Whole-blood glucose levels were determined by a colorimetric assay on a test strip (One Touch basic blood glucose monitoring system; Lifespan, Milpitas, CA).
Metabolic Labeling of Pieces of Heart Tissue-HuBTg ϩ/o and HuBTg ϩ/o Acadl Ϫ/Ϫ mice were fasted for 18 h, and the hearts were excised and labeled with [ 35 S]Promix (Amersham Pharmacia Biotech; Ref. 27). Briefly, three mouse hearts (ϳ300 mg) were minced with a razor blade into ϳ0.5-mm 3 pieces, placed in a 1.6-ml Eppendorf tube, and washed twice with 1.0 ml of methionine-and cysteine-free Dulbecco's modified Eagle's medium (D-0422; Sigma) supplemented with 7% fetal calf serum, 1.6 mM glutamate, and 1.6 mM sodium pyruvate. The heart tissue was then incubated with 0.715 mCi of [ 35 S]Promix and incubated at 37°C for 3 h. After addition of protease inhibitors, the medium from three tubes was pooled, and any debris was pelleted by centrifugation (12,000 ϫ g for 1 min). The [ 35 S]Promix labeling medium was subjected to discontinuous sucrose gradient ultracentrifugation (7,27). The gradients were ultracentrifuged in an SW55 Ti rotor at 35,000 rpm for 65 h at 12°C; the bottom of the tube was pierced, and the gradient was unloaded into five fractions. Human apoB was immunoprecipitated from each of the fractions and analyzed by polyacrylamide gel electrophoresis and autoradiography. The densities of the fractions were determined by refractometry.

RESULTS
Lipoprotein Secretion by the Heart-Hearts from HuBTg ϩ/o mice and nontransgenic littermate controls were perfused in a Langendorff apparatus, and the perfusate was examined for lipoproteins by electron microscopy. Each mouse heart was first perfused exhaustively in a noncirculating fashion with KHB perfusion medium (retrograde coronary perfusion of 5 ml/min for 60 min) to remove any plasma lipoproteins. The hearts were then perfused with KHB perfusion medium containing 1.2 mM oleate, and the buffer was circulated. After 5 min, virtually no lipoproteins were detectable in the perfusion medium (Fig. 1, A and B), indicating that most plasma lipoproteins had been washed away. After 60 min, the perfusion medium from both HuBTg ϩ/o and nontransgenic mice contained spherical lipoproteins (Fig. 1, C and D). Most of these lipoproteins in the perfusates from HuBTg ϩ/o mouse hearts were in the low density lipoprotein size range (18.3 Ϯ 5.7 nm in diameter (mean Ϯ S.D.); n ϭ 200 particles). The same was true for the lipoproteins in the perfusates of the nontransgenic hearts (18.5 Ϯ 6.9 nm; n ϭ 200). The lipoproteins in the plasma of the nontransgenic mice were larger and more variable in size (22.9 Ϯ 17 nm; range, 12-106 nm; n ϭ 200; Fig. 1E). The lipoproteins in the plasma of the HuBTg ϩ/o mice were skewed toward smaller particles (17.5 Ϯ 4.3 nm; n ϭ 200), reflecting the larger numbers of low density lipoprotein particles in the plasma of those mice (Ref. 16 and Fig. 1F).
The heart perfusion experiments suggested that the beating heart secretes lipoproteins. The fact that the lipoproteins were spherical suggested that the heart must secrete core lipids (e.g., triglycerides). To address the issue of triglyceride secretion by the heart, we perfused mouse hearts for 60 min with continuously circulating medium containing 1.2 mM [ 3 H]oleate. 3 Htriglycerides were detected in the perfusates of nontransgenic hearts, and the amounts increased with time ( Fig. 2A). Significantly more 3 H-triglycerides were present in the perfusates of the hearts of the HuBTg ϩ/o mice at 60 min ( Fig. 2A). Adding an MTP inhibitor to the medium reduced the amount of 3 H-triglycerides secreted from the HuBTg ϩ/o mouse hearts (Fig. 2A).
The difference in the amounts of triglycerides in the perfusates of HuBTg ϩ/o and nontransgenic hearts reflected, in large part, the different amounts of 3 H-triglycerides in the d Ͻ 1.063-g/ml fraction. At 60 min, the amount of 3 H-triglycerides in the d Ͻ 1.063-g/ml fraction was significantly greater for the HuBTg ϩ/o mice than for the nontransgenic mice (Fig. 2B).
We suspected that lipoproteins are produced by cardiac myocytes. This conclusion was bolstered by experiments with cardiac myocytes from newborn HuBTg ϩ/o mice. HuBTg ϩ/o myocytes expressed the apoB gene, as judged by RT-PCR (Fig. 3A). After incubating the myocytes with [ 14 C]oleate, HuBTg ϩ/o myocytes secreted ϳ2-fold more 14 C-triglycerides than the nontransgenic cardiac myocytes (2.9-, 1.6-, 1.3-, 2.0-, and 2.1-fold more in five experiments). An example of one such experiment is shown in Fig. 3B.
Heart Lipoprotein Secretion and Cardiac Triglyceride Stores-Gene defects that block lipoprotein secretion are associated with increased intracellular lipid stores in the liver and intestine (2,6,24). We hypothesized that interfering with lipid and lipoprotein secretion in the heart might also increase triglyceride stores in the heart. We compared heart triglyceride stores in heart-specific MTP knockout mice (Mttp flox/flox mice carrying an MHC-Cre transgene) and littermate controls (nontransgenic Mttp flox/flox mice; n ϭ 20 in each group). Heartspecific recombination occurred as predicted in mice carrying the ␣MHC-Cre transgene (Fig. 4A), and Mttp mRNA levels were reduced by ϳ90% (Fig. 4B). Although there was no difference in liver triglyceride stores in the two groups of mice ( Fig. 4C; p ϭ 0.77), the triglyceride stores in hearts of the Mttp flox/flox /␣MHC-Cre mice were almost 2-fold higher than in controls (p Ͻ 0.0005; Fig. 4C). The increase in cardiac triglyceride stores was almost certainly attributable to the Mttp gene excision and not simply to the expression of the ␣MHC-Cre transgene. In a control experiment, there was no difference in cardiac triglyceride stores in ␣MHC-Cre transgenic mice and littermate control mice that were homozygous for a wild-type Mttp allele (2.40 Ϯ 0.40 mol/g in ␣MHC-Cre transgenic mice (n ϭ 4) versus 2.75 Ϯ 0.80 mol/g in the nontransgenic mice (n ϭ 5); p ϭ 0.45).
To investigate whether increased heart lipoprotein production reduces cardiac triglyceride stores, we assessed heart triglyceride stores in C57BL/6 HuBTg ϩ/o mice (n ϭ 30 females) and nontransgenic littermate controls (n ϭ 29 females). The plasma triglycerides and cholesterol levels were higher in the HuBTg ϩ/o mice than in nontransgenic controls (Fig. 5A), but there were no differences in plasma glucose or free fatty acid levels (Fig. 5B). Cardiac triglyceride stores were reduced by ϳ75% in the HuBTg ϩ/o mice (p Ͻ 0.0001; Fig. 5C). There was no difference in liver triglyceride stores (Fig. 5C).
Long-chain acyl coenzyme A dehydrogenase (Acadl) is required for the oxidation of long-chain fatty acids within mitochondria. Homozygous Acadl Ϫ/Ϫ mice have elevated plasma fatty acid levels and accumulate pathological levels of triglycerides in both heart and liver in response to fasting (18). To determine whether increased heart lipoprotein production from the human apoB transgene affects heart triglyceride stores in Acadl Ϫ/Ϫ mice, we measured heart triglyceride stores in Acadldeficient mice that carried the human apoB transgene (Acadl Ϫ/ ϪHuBTg ϩ/o ) and in nontransgenic Acadl Ϫ/Ϫ littermate control mice. As before, the plasma cholesterol and triglyceride levels were higher in the Acadl Ϫ/Ϫ HuBTg ϩ/o than in the Acadl Ϫ/Ϫ mice (Fig. 6A), but there were no differences in glucose or fatty acid levels (Fig. 6B). Heart triglyceride stores were significantly lower in Acadl Ϫ/Ϫ HuBTg ϩ/o mice than in the Acadl Ϫ/Ϫ controls (p Ͻ 0.01; Fig. 6C). Once again, however, the human apoB transgene did not affect triglyceride stores in the liver (Fig. 6C).
We suspected that the increased triglyceride stores within the hearts of Acadl-deficient mice might be associated with the Hearts were perfused in a Langendorff perfusion apparatus with noncirculating Krebs-Henseleit medium for 1 h (ϳ300 ml/heart) and then with circulating medium containing 1.2 mM oleate. Perfusion medium samples (from the circulating medium) were obtained after 5 and 60 min, and the d Ͻ 1.063-g/ml fractions were examined by electron microscopy. A, nontransgenic heart, 5 min; B, HuBTg ϩ/o heart, 5 min; C, nontransgenic heart, 60 min; D, HuBTg ϩ/o heart, 60 min; E, d Ͻ 1.063-g/ml fraction of the plasma from a nontransgenic mouse; F, d Ͻ 1.063-g/ml fraction of the plasma from an HuBTg ϩ/o mouse. secretion of more buoyant lipoproteins from the heart. However, this did not appear to be the case. The buoyant density of the apoB100-containing lipoproteins produced by minced-up pieces of hearts from Acadl Ϫ/Ϫ HuBTg ϩ/o mice was the same as that of lipoproteins produced by pieces of HuBTg ϩ/o hearts (Fig. 6D).
The lower cardiac levels of lipids in Acadl Ϫ/Ϫ HuBTg ϩ/o mice than in Acadl Ϫ/Ϫ mice raised the possibility that the hearts of those two groups of mice might contain different amounts of lipid peroxides. Interestingly, the cardiac levels of thiobarbituric acid-reactive substances were lower in Acadl Ϫ/Ϫ HuBTg ϩ/o mice than in Acadl Ϫ/Ϫ mice (p Ͻ 0.001; Fig. 6E), suggesting that increased levels of lipoprotein production in Acadl Ϫ/Ϫ HuBTg ϩ/o mice also reduce lipid peroxide levels. DISCUSSION This study shows that the beating mouse heart actually synthesizes and secretes spherical lipoproteins and also secretes triglycerides. The fact that the secretion of spherical lipoproteins would be accompanied by triglyceride secretion was not particularly surprising, given that the principal function of apoB in other tissues is to package neutral lipids, particularly triglycerides, for secretion. The amount of triglyceride secretion from the heart increased with increased levels of lipoprotein synthesis and secretion (e.g. in HuBTg ϩ/o mice) and decreased with reduced levels of lipoprotein secretion (e.g. in the setting of an MTP inhibitor). Importantly, our studies suggest that the expression of lipoprotein assembly genes influences cardiac lipid homeostasis. Blocking lipoprotein secretion by knocking out Mttp in the heart significantly increased cardiac triglyceride stores within the heart after an 18-h fast. Increasing lipoprotein secretion with a human apoB transgene also affected cardiac triglyceride stores. Cardiac triglyceride stores in HuBTg ϩ/o mice were ϳ75% lower than in nontransgenic littermates. The human apoB transgene also reduced the accumulation of triglycerides and thiobarbituric acid-reactive substances in the hearts of mice lacking Acadl.
Why would the heart, a prodigious importer and consumer of lipids, secrete lipoproteins? One obvious possibility is that lipoprotein secretion is involved in reverse triglyceride transport, exporting surplus and potentially toxic lipids away from the heart. That cardiac triglyceride stores were reduced in the HuBTg ϩ/o mice and increased in the heart-specific Mttp knockout mice is consistent with this possibility. Lipoprotein secretion from the heart could be particularly relevant during a prolonged fast or during exercise, when both the uptake and the oxidation of fatty acids increase (28). Interestingly, an increased rate of triglyceride depletion from working rat hearts cannot be fully explained by increases in ␤-oxidation, prompting speculation that triglycerides can be removed by other mechanisms (29,30). Our studies raise the possibility that some of this unexplained triglyceride turnover could be accounted for by lipoprotein secretion by the heart.
The fact that perturbations in cardiac lipoprotein production were associated with reciprocal changes in cardiac triglyceride stores supports a role for lipoprotein secretion in exporting lipids away from the heart and into the general circulation. However, lipoprotein secretion might very well play a role in redistribution of lipids between myocytes, just as apoE secretion by astrocytes may play a role in redistributing cholesterol within the central nervous system (31).
Inherited defects in the mitochondrial ␤-oxidation pathway can cause a striking accumulation of triglycerides in the heart and are associated with heart failure as well as sudden cardiac death (13). In mice lacking Acadl, triglycerides accumulate during fasting (18). We found that the human apoB transgene significantly reduced cardiac triglyceride stores in Acadl-deficient mice but not to the low levels that we observed in HuBTg ϩ/o mice with wild-type levels of Acadl expression. We suspect that the capacity of heart lipoprotein secretion to export lipids is relatively limited and fixed. Two pieces of evidence tend to support this concept. First, the absolute magnitude of the transgene-associated reduction in triglyceride stores was quite similar in the presence and absence of Acadl deficiency. Second, even though the Acadl-deficient mice had increased triglyceride stores in the heart, the lipoproteins secreted by the hearts of HuBTg ϩ/o Acadl Ϫ/Ϫ and HuBTg ϩ/o mice had similar buoyant densities.
Hepatic triglyceride stores were not affected by the human apoB transgene, either in the presence or in the absence of Acadl deficiency, despite high-level transgene expression in the liver. Why would the effects of the transgene on triglyceride stores be different in the heart and the liver? One potential explanation is that the liver serves a unique role as a clearinghouse for triglycerides and that any transgene-induced increase in triglyceride secretion in that organ is simply balanced by an increase in lipid uptake, thereby preventing net changes in triglyceride stores. Alternatively, transgenic overexpression of apoB in the liver may not be associated with increased triglyceride secretion. In the liver, lipoprotein particles containing very small amounts of triglycerides and cholesterol esters are thought to be assembled in the rough endoplasmic reticulum as apoB is translated (32). However, the bulk of triglycerides in nascent very low density lipoprotein particles are thought to be added later in a specialized compartment of the smooth endoplasmic reticulum (32)(33)(34). In this so-called second step of lipoprotein assembly, small "apoB particles" are thought to fuse with large apoB-free triglyceride particles, generating nascent very low density lipoprotein for secretion. Because most lipids in the liver are added independently of apoB production, overexpression of apoB might have a negligible impact on triglyceride secretion rates (and thus little impact on triglyceride stores). In line with this view, hepatic triglyceride secretion rates in HuBTg ϩ/o mice and controls are similar. 2 In contrast to the liver, the heart seemingly lacks the capacity for this second step (we found that most of the lipoproteins in the perfusates of mouse hearts were small, much smaller 2 L.-S. Huang, personal communication. than very low density lipoprotein in normal mouse plasma; Ref. 35). Thus, unlike the situation in the liver, the amount of apoB synthesis and secretion in the heart may regulate triglyceride secretion rates. Of note, certain hepatic cancer cell lines have lost the capacity for the second step of lipoprotein assembly and thus secrete low density lipoprotein-sized particles (similar to those that we observed in the heart perfusates). In those cell lines, apoB overexpression increases lipid secretion (36,37).
The heart-specific Mttp knockout mice had increased cardiac triglyceride stores. Whether this finding is relevant to the human lipoprotein deficiency syndrome abetalipoproteinemia is not clear. There have been occasional cases of cardiomyopathy in abetalipoproteinemia (38), but as far as we are aware, no one has assessed cardiac lipid stores in this condition. Even if such measurements were feasible, they might be difficult to interpret. Patients with abetalipoproteinemia have extremely low plasma triglyceride levels (typically ϳ5 mg/dl; Ref. 2), which might protect them from increased triglyceride storage in the heart. Our results might, however, be relevant to the treatment of human hyperlipidemic patients with MTP inhibitors (39,40). Those drugs might influence cardiac triglyceride stores, at least under some conditions.
Our studies support the notion that cardiac lipoprotein secretion is involved in reverse triglyceride transport. High levels of triglycerides in the heart cause diminished contractile function, hypertrophy, and myocyte death (14,15,41). Therefore, lipoprotein secretion might be physiologically important, serving to ward off the deleterious effects of surplus lipids. We would not be surprised if future studies uncovered roles for lipoprotein secretion in other tissues. Recently, we were intrigued to learn that both apoB and MTP are expressed in the ␤ cells of pancreatic islets, 3 where the accumulation of lipids is thought to be toxic (42).