Overexpression of apolipoprotein B in the heart impedes cardiac triglyceride accumulation and development of cardiac dysfunction in diabetic mice.

The heart secretes apolipoprotein B (apoB) containing lipoproteins. Herein, we examined whether the overexpression of a human apoB transgene in the heart affects triglyceride accumulation and development of cardiac dysfunction in streptozotocin-treated diabetic mice. Blood glucose, plasma free fatty acids, and plasma triglycerides were similarly affected in diabetic wild type mice and diabetic apoB transgenic mice as compared with non-diabetic mice of the same genotype. After 12 weeks, heart triglycerides were increased by 48% in diabetic wild type mice. These mice displayed an increased expression of brain natriuretic peptide and deterioration of heart function on echocardiography. In diabetic apoB transgenic mice, heart triglyceride levels were identical to those in non-diabetic wild type and apoB transgenic mice, and brain natriuretic peptide expression as well as echocardiographic indexes of heart function were only marginally affected or unaffected. The findings suggest that triglyceride accumulation in the heart is important for development of diabetic cardiomyopathy in mice, and that lipoprotein formation by cardiomyocytes plays an integrated role in cardiac lipid metabolism.

Liver and intestinal cells secrete triglyceride-rich lipoproteins. This ability is dependent on the expression of the apoB and microsomal triglyceride transfer protein (MTP) 1 genes (1,2). MTP transfers triglycerides onto the apoB polypeptide chain during its translation into the endoplasmic reticulum. ApoB serves as the principal structural protein in the resulting lipoprotein particles that are secreted from the cells. Studies of mice, which overexpress a human apoB transgene, revealed that cardiac myocytes in addition to hepatocytes and absorptive enterocytes also express the apoB and MTP genes (3) and secrete apoB containing lipoproteins (4). The apoB mRNA is not edited in cardiac myocytes (5). Consequently, the heart secretes lipoproteins containing the full-length apoB100 pro-tein rather than the truncated apoB48 protein (4). Because the formation of apoB-containing lipoproteins serves as an effective means of secreting large amounts of triglycerides from liver and intestinal cells, we hypothesized previously that the physiological role of lipoprotein formation in the heart could be the removal of triglycerides from myocytes that are not used as fuel, i.e. a reverse triglyceride transport pathway (6,7).
Recently, the idea has been put forward that triglyceride accumulation in cardiac muscle cells adversely affects cardiac function (8). Diabetes is associated with aberrations in cardiac fuel metabolism and a ϳ2-fold increase in cardiac triglyceride content (9,10). In diabetic rats, this triglyceride accumulation occurs in parallel with compromised cardiac performance (11,12). Echocardiographic studies have also revealed abnormal cardiac function in young diabetic individuals without coronary heart disease (13). It is unknown as to what extent triglyceride accumulation in cardiac myocytes affects the development of diabetes-induced cardiac dysfunction.
In this study, we examined whether the overexpression of a human apolipoprotein B transgene in the heart might attenuate cardiac lipid accumulation and consequently modulate signs of cardiac dysfunction in streptozotocin-treated diabetic mice.

EXPERIMENTAL PROCEDURES
Animals-5-7-week-old wild type mice (C57BL/6 mice, 20 female and 20 male mice) and human apoB transgenic mice (C57BL/6NTac-TgN mice, 20 female and 20 male mice) were obtained from M&B (Ry, Denmark) housed at the Panum Institute, University of Copenhagen (Copenhagen, Denmark) and fed standard laboratory chow (Altromin number 1314, Rugaarden, Denmark). The apoB transgenic mice were generated with an ϳ87-kilobase pair human genomic fragment from a genomic P1 bacteriophage clone, p158. This transgene directs human apoB overexpression in the liver and in the heart but not in the intestine (7,14). The transgenic mice used in this study had been backcrossed to a C57Bl6 background Ͼ14 times. The study was approved by the Danish Government body supervising animal experiments (Dyreforsoegstilsynet).
Diabetes-Streptozotocin (STZ) (40 g/g/day) (Sigma) was injected intraperitoneally once a day for five consecutive days followed by a two-day recovery and injections for an additional 2 days. STZ (4 g/l) was dissolved in citrate buffer (0.05 M, pH 4.5, 4°C), kept on ice, and injected within 20 min. Control mice received the citrate buffer vehicle (200 l/mouse/day) following the same treatment regimen.
Tissue and Blood Samples-Blood samples for plasma insulin and lipid analyses were drawn into tubes containing Na 2 -EDTA and centrifuged at ϳ4000 ϫ g for 10 min at 4°C. Plasma was stored at Ϫ80°C until analyses. After echocardiographic examination, the heart and liver were removed and briefly rinsed in 0.9% NaCl (4°C). The heart was carefully cleaned of pericardial fat. Cross-sectional slices of the ventricular portion of heart including right and left ventricular tissue and liver biopsies were snap-frozen in liquid nitrogen and stored at Ϫ141°C for subsequent lipid and RNA extraction. A ϳ3-mm slice from the apex of the heart was fixed in 3% paraformaldehyde.
Echocardiography-The mice were anesthetized with a subcutaneous injection of 1:1 hypnorm (Fentanyl, Fluanisone (10 mg/ml), Janssen):diazepam (6 ml/kg) before transthoracic echocardiography, and Doppler flow analyses were performed with the Vivid Five instrument (GE Ultrasound, Copenhagen, Denmark) and a 10-MHz transducer head (15). Echocardiographic investigations were repeated in each mouse after an intraperitoneal injection of dobutamine (1.0 -1.5 g/g body weight) (16). Echocardiography and data analyses were performed by E. Bollano without any knowledge of treatment and mouse genotype.
Plasma Analyses-The glucose concentration in tail blood was determined with a Humacheck Plus glucose meter (Human, Wiesbaden, Germany). Plasma insulin was measured with ELISA by Dr. B. Rolin (Novo Nordisk A/S). Plasma lipid concentrations were determined with enzymatic kits: free fatty acids (Wako NEFA C kit, TriChem Aps, Frederikssund, Denmark), triglycerides (GPO-TRINDER, Sigma; this kit also measures free glycerol), total cholesterol (CHOD-PAP, Roche Molecular Biochemicals), and high density lipoprotein (HDL) cholesterol (HDL-C, Roche Molecular Biochemicals).
RNA Purification-Frozen heart and liver biopsies (40 -50 mg) were homogenized with a Polytron PT1200CL (Buch & Holm, Herlev, Denmark) in TRIzol reagent (Invitrogen). Total RNA was extracted according to the manual by the manufacturer and suspended in RNase-free H 2 O. The RNA concentration was calculated from the absorbance at 260 nm (A 260 ). The RNA integrity was assured always by electrophoresis on a 1% agarose gel. Synthesis and Amplification of cDNA-First strand cDNA was synthesized from 1 g of total RNA with Moloney murine leukemia virus reverse transcriptase (40 units, Roche A/S, Avedore, Denmark) and random hexamer primers in 10-l reactions. The primers for amplification of mouse apoB, MTP, glyceraldehyde-3-phosphate dehydrogenase, and ␤-actin are described elsewhere (18). The primers for brain natriuretic peptide (BNP) (5Ј-CTGAAGGTGCTGTCCCAGAT-3Ј and 5Ј-GTTCTTTTGTGAGGCCTTGG-3Ј) and human apoB (5Ј-GGAGCTGCT-GGACATTGCTA-3Ј and 5Ј-ATGGCAGCTTTCTGGATCAT-3Ј) were obtained from Sigma-Genosys (Pampisford, United Kingdom). The human apoB transcript could not be amplified from wild type mouse liver cDNA, and the mouse apoB transcript could not be amplified from human HepG2 cell cDNA. The specificity of each PCR was further confirmed by DNA sequencing of upper and lower strands of PCR transcripts (18). We never observed any amplification of genomic DNA from the cDNA preparations.
mRNA Quantification-Quantitative real-time PCR analysis of mRNA expression was done with a LightCycler and the DNAmaster SYBR Green kit (Roche A/S). The PCR reactions (20 l) contained 2 l of SYBR Green I mixture, 2-3 mM MgCl 2 , 10 pmol of each primer, cDNA synthesized from 20 ng of total RNA, and PCR grade H 2 O. For each mRNA transcript in each tissue biopsy, the time point of the log-linear increase in amplified DNA during the PCR was determined with the fit-point option of the LightCycler software. The relationship between that time point and the relative concentration of an mRNA transcript was determined by analyzing in each run the dilution series of cDNA from wild type mouse livers or HepG2 cells (cDNA synthesized from 100, 20, 2, and 0.2 ng of total RNA). All differences among groups were analyzed with Student's t test.

Effects of STZ on Blood Glucose and Plasma Lipids-Injec-
tions of the pancreatic ␤-cell toxin STZ conferred a pronounced and sustained increase in blood glucose levels and a corresponding decrease in plasma insulin levels in male mice (Table  I). However, in female mice, the diabetic response after STZ treatment was marginal ( Table I). The blood glucose concen-  trations were similarly increased by STZ treatment in wild type and apoB transgenic mice (Table I). The plasma concentrations of triglycerides and free fatty acids were not affected significantly by STZ treatment in either apoB transgenic or wild type mice (Table I). As expected, apoB transgenic mice had higher total and VLDLϩLDL plasma cholesterol concentrations than wild type mice (Table I).
MTP and ApoB mRNA Expression in Heart and Liver-Realtime PCR analyses were used to assess whether STZ-induced diabetes might affect heart or liver expression of the MTP or apoB genes. The cardiac MTP mRNA levels of male mice were increased significantly by STZ in both wild type and apoB transgenic mice (Fig. 1A). In contrast, there was no effect of STZ-induced diabetes on MTP gene expression in the liver (Fig.  1A). The expression of MTP mRNA in the heart was ϳ2 and ϳ1% of the expression in the liver in STZ-treated and control mice, respectively. STZ treatment did not affect heart or liver expression of either the endogenous mouse apoB gene or the human apoB transgene (Fig. 1, B-C). Human apoB mRNA level in the heart of the transgenic mice was ϳ4% of that in the liver. This finding is in close accordance with a previous estimate by RNase protection assays (3). The heart expression of endogenous mouse apoB mRNA was ϳ0.1% of that in the liver (Fig.  1B). The overexpression of human apoB mRNA had no effect on mouse apoB or MTP mRNA expression. Cardiac and hepatic human apoB, mouse apoB, and MTP mRNA levels were similar in STZ-treated and control female mice (data not shown).

Effect of Diabetes and ApoB Overexpression on Cardiac Lipid
Accumulation-In male wild type mice, STZ treatment increased the mean cardiac triglyceride content by 48% (p Ͻ 0.05) ( Fig. 2A). In contrast, STZ treatment did not affect the cardiac triglyceride content of male apoB transgenic mice ( Fig. 2A). In female mice, the cardiac triglyceride content was similar in STZ-treated and control wild type mice (4.3 Ϯ 0.4 versus 4.1 Ϯ 0.3 nmol/mg wet weight). The cardiac triglyceride content of non-diabetic control mice was unaffected by apoB overexpression ( Fig. 2A). The enzymatic method for measuring triglycerides in cardiac biopsies also measured glycerol. In male mice, the heart glycerol content was increased in STZ-treated wild type mice compared with their controls, whereas there was no difference in the glycerol content between STZ-treated apoB transgenic mice and their controls (Fig. 2B). STZ did not significantly affect free glycerol levels in either wild type or apoB transgenic female mice (data not shown). The levels of free glycerol were higher in apoB transgenic control mice than in wild type control mice (Fig. 2B).  To assess whether the overexpression of the apoB transgene in the heart affected cardiac stores of other lipid classes than triglycerides, we measured a series of cardiac lipids in addition to triglycerides (Table II). There was no effect of STZ treatment or apoB overexpression on cardiac contents of cholesterol or phospholipids in male mice (Table II) or in female mice (data not shown). The TLC analysis confirmed the normalization of cardiac triglycerides in the STZ-treated apoB transgenic male mice (Table II).
Liver triglyceride and total cholesterol stores were not affected by STZ treatment in wild type male mice. The hepatic triglyceride content was lower in STZ-treated apoB transgenic male mice compared with control apoB transgenic male mice (Table II). A reduction of hepatic triglyceride stores in STZtreated diabetic mice has been described previously (19).

Effects of Diabetes and ApoB Overexpression on Cardiac Function-Echocardiography and Doppler flow analyses
showed that STZ-induced diabetes affects cardiac function of wild type male mice (Fig. 3, A-C, and Table III). In accordance with previous findings in rats (11,15), the heart rate was slightly decreased in STZ-treated mice, although this effect was not statistically significant. The indexes of systolic function (e.g. circumferential shortening (Fig. 3B) and diameter of the left ventricle at the end of the systole (Table III)) and diastolic function (E-wave deceleration time (Fig. 3C)) were affected negatively by STZ at the base-line recordings. For each of these parameters, the effect of STZ treatment was less pronounced and/or absent in apoB transgenic mice. In apoB transgenic male mice, the heart rate tended to be increased, whereas circumferential shortening, diameter of the left ventricle at the end of the systole, and E-wave deceleration time at base line were not significantly affected after STZ treatment (Fig. 3, B and C, and Table III).
Increased BNP expression is a sensitive indicator of cardiac dysfunction (20). STZ-treated diabetic rats have increased BNP gene expression in the heart (21). In this study, STZ-treated wild type male mice had increased cardiac BNP mRNA levels compared with vehicle-treated wild type mice (Fig. 3D). In contrast, BNP gene expression was not affected by STZ treatment in the apoB transgenic male mice (Fig. 3D). Diabetes has also been associated with apoptosis of cardiac myocytes (22). Terminal deoxynucleotidyltransferase dUTP nick-end (TUNEL) staining of histological sections from the apical section of each mouse heart in this study revealed very few apoptotic cells with similar occurrence in STZ-treated and vehicle-treated mice (data not shown), and STZ treatment did not affect heart weight.

DISCUSSION
The metabolism of fat is essential for the function of the heart. The present results suggest that accelerated lipoprotein formation by cardiac myocytes can remove excess triglycerides from the diabetic heart. This idea is in accordance with a recent seminal study (23), which demonstrated that hearts from apoB transgenic mice secrete more triglycerides in apoB-containing lipoproteins than control mice. Also, the overexpression of apoB decreased heart triglyceride stores in mice with a genetic defect in the metabolism of long chain fatty acids that results in cardiac triglyceride accumulation during prolonged fasting (23).
MTP is rate-limiting for the production and secretion of apoB-containing lipoproteins from the liver (24 -26). Previous studies (26 -29) have shown a tight correlation among MTP mRNA, protein, and activity levels in vitro and in vivo. Because the MTP mRNA expression levels were increased in the heart of diabetic mice, this study is compatible with the idea that lipoprotein secretion from the heart might be regulated. This notion has also gained support from studies of heart biopsies from patients undergoing cardiac surgery. 2 Those studies suggest that the MTP mRNA expression is increased in hypoxic human left ventricle compared with normoxic human left ventricle, and that the MTP mRNA level is negatively associated with cardiac triglyceride storage in patients with ischemic heart disease. An increase of lipoprotein secretion rates from cardiac myocytes during increased fat load (such as in diabetes with increased flux of free fatty acids into the heart and cardiac ischemia with decreased ␤-oxidation of fatty acids) could contribute to maintain a constant (low) level of triglycerides in cardiac myocytes. Remarkably, in this study, we only observed increased MTP gene expression in the heart and not in the livers of STZ-treated diabetic mice. The latter observation is in accordance with a more elaborate study of MTP in diabetic mouse livers (18). Further studies are needed to determine whether it might be possible to increase lipoprotein secretion from the heart without increasing the secretion from the liver.
We were able to test the functional consequences of cardiac triglyceride accumulation in the diabetic mouse heart, because cardiac triglyceride stores in diabetic mice were increased in wild type mice but were normal in apoB transgenic mice. Importantly, the blood levels of major cardiac fuel substrates (i.e. glucose, free fatty acids, and triglycerides) were similarly affected by STZ treatment in wild type and in apoB transgenic mice. Moreover, in the heart, the effect of apoB overexpression on cardiac lipids was confined to triglycerides, whereas there was no effect of apoB overexpression on cardiac cholesterol or phospholipid content. The assessments of cardiac BNP gene expression as well as biomechanical cardiac function suggested that a diabetes-specific affection of the heart in STZ-treated male mice is attenuated by a normalization of cardiac triglyceride stores. In apoB transgenic mice, several echocardiographic measures reflecting both systolic and diastolic function were either unaffected by STZ-induced diabetes or less affected than in wild type mice. This finding implies that lipoprotein secretion from the heart plays an integrated physiological role in cardiac function. Still, some parameters of systolic function during dobutamine stress were decreased by STZ treatment in both wild type and apoB transgenic mice. This indicates that the decline in systolic function of STZ-treated mice was not entirely related to cardiac triglyceride accumulation.
A causal link of cardiac triglyceride accumulation and cardiac dysfunction has been proposed recently (8). The normalization of cardiac triglycerides in genetically obese fa/fa rats with a peroxisome proliferator-activated receptor ␥ agonist attenuates the development of cardiac dysfunction (30). Cardiac overexpression of peroxisome proliferator-activated receptor ␣ (31) or long chain acyl-CoA synthetase (32) in mouse hearts both confer an increase in triglyceride stores and a concomitant decrease of cardiac function. This study supports the idea that a normalization of cardiac triglyceride stores improves cardiac function in diabetic mice. This conclusion may add evidence to an emerging concept (8,32,33) that the accumulation of neutral lipids in cardiac myocytes causes "lipotoxic" heart disease. It is also conceivable that the accelerated removal of triglycerides by increased lipoprotein secretion affects intracellular fluxes of free fatty acids or perhaps cardiac fuel substrate utilization. An effect of apoB overexpression on intracellular intermediate metabolites in cardiomyocytes is compatible with the finding that cardiac free glycerol levels were increased in control apoB transgenic mice versus control wild type.
In conclusion, this study supports the idea that lipoprotein formation by the heart plays an integrated role in cardiac lipid homeostasis and affects cardiac physiology in diabetic mice.