|Home | About | Journals | Submit | Contact Us | Français|
Excess lipid accumulation in the heart is associated with decreased cardiac function in humans and in animal models. The reasons are unclear, but this is generally believed to result from either toxic effects of intracellular lipids or excessive fatty acid oxidation (FAO). PPARγ expression is increased in the hearts of humans with metabolic syndrome, and use of PPARγ agonists is associated with heart failure. Here, mice with dilated cardiomyopathy due to cardiomyocyte PPARγ overexpression were crossed with PPARα-deficient mice. Surprisingly, this cross led to enhanced expression of several PPAR-regulated genes that mediate fatty acid (FA) uptake/oxidation and triacylglycerol (TAG) synthesis. Although FA oxidation and TAG droplet size were increased, heart function was preserved and survival improved. There was no marked decrease in cardiac levels of triglyceride or the potentially toxic lipids diacylglycerol (DAG) and ceramide. However, long-chain FA coenzyme A (LCCoA) levels were increased, and acylcarnitine content was decreased. Activation of PKCα and PKCδ, apoptosis, ROS levels, and evidence of endoplasmic reticulum stress were also reduced. Thus, partitioning of lipid to storage and oxidation can reverse cardiolipotoxicity despite increased DAG and ceramide levels, suggesting a role for other toxic intermediates such as acylcarnitines in the toxic effects of lipid accumulation in the heart.
With the increase in the prevalence of obesity and type 2 diabetes, a series of disorders associated with ectopic deposition of fat have become more common and are termed lipotoxic diseases (1). Although the clinical presentations are in diverse tissues — leading to nonalcoholic fatty liver disease, muscle insulin resistance, and cardiac dysfunction — it is likely that they have common or overlapping pathophysiology. The heart is the most energy-demanding tissue of the body and utilizes fatty acids (FAs) as its major source of substrate for ATP generation (2). Nonetheless, excess FA oxidation (FAO) has been implicated as a cause of cardiac dysfunction in obesity and diabetes (3, 4). In humans, greater stores of cardiac lipid in obesity (5, 6) and type 2 diabetes (7) are correlated with reduced heart function.
Several genetically modified animals were created to have altered cardiac lipid content and determine how this affects heart function exclusive of systemic metabolic changes. Overexpression of fatty acyl–CoA synthetase (8), a cardiomyocyte cell surface–anchored form of lipoprotein lipase (9), or FA transport protein (10) augments heart lipid content and leads to cardiomyopathy. Overexpression of PPARα (4) and PPARγ (11) using the α–myosin heavy chain (MHC) promoter led to cardiac lipid accumulation and cardiomyopathy; the transgenes are designated MHC-Ppara and MHC-Pparg. Although this was associated with increased expression of genes controlling FAO, these hearts accumulated more triacylglycerol (TAG). Thus, the increase in lipid uptake likely outstrips the greater FAO. The dilated cardiomyopathy in these animals, especially the MHC-Ppara transgenic mice, has been compared with that occurring with diabetes (4). The pathophysiology of the cardiac dysfunction with PPARα and PPARγ overexpression is unclear but has been hypothesized to result from excess FAO or accumulation of toxic intracellular lipids (12).
PPARγ agonists cause heart failure in humans. One reason for this may be greater accumulation of salt and water (13). However, potent PPARγ agonists cause cardiomegaly in rodents (14). The ratio of PPARγ/PPARα expression in human hearts is much greater than in mice (15). Moreover, a recent report showed that PPARγ expression is markedly increased in ventricular muscle from subjects with metabolic syndrome (16). Thus, aside from serving as a model to understand the toxic effects of lipid in the heart, MHC-Pparg transgenic mice are likely to mimic pathological processes that occur with PPARγ agonist treatment of patients who have a predisposition to developing lipotoxic cardiomyopathy.
Greater intracellular toxic lipid content might lead to heart dysfunction associated with ceramide-induced apoptosis (17), increased ROS formation, mitochondrial dysfunction (18, 19), and/or ER stress (20). We hypothesized that deletion of PPARα in the heart would reduce the expression of FAO genes and FAO in the MHC-Pparg hearts and would ameliorate cardiac dysfunction. To study this, we crossed MHC-Pparg mice with PPARα-knockout (Ppara–/–) mice. Surprisingly MHC-Pparg/Ppara–/– mice had greater expression of PPAR downstream genes, including those responsible for FAO, and greater FAO. These mice, however, had improved cardiac function and survival. Many potentially toxic pathways were simultaneously improved in the MHC-Pparg/Ppara–/– mice; therefore, no single cause of the improvement can be defined. Nonetheless, our studies show that lipotoxicity is ameliorated despite greater FAO and no reduction in TAG storage.
To study the effects of PPARα deficiency in the PPARγ-mediated cardiomyopathy, we crossed the high-expressing MHC-Pparg mouse line (15) twice into the Ppara–/– background, which resulted in MHC-Pparg/Ppara–/– offspring. Overexpression of PPARγ and the absence of PPARα in the heart were confirmed in each genotype at the protein level (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI40905DS1).
PPARα deficiency is associated with increased circulating FFAs (Table (Table1).1). In MHC-Pparg mice, FFA concentrations were similar to those in controls. In MHC-Pparg/Ppara–/– mice, FFAs were equivalent to those in Ppara–/– mice. TAG levels have been reported to be variable in the Ppara–/– mice (21, 22). In our mice, they were not significantly reduced compared with those in control or MHC-Pparg mice, but this reduction reached significance when the MHC-Pparg/Ppara–/– mice were compared with either control or MHC-Pparg animals. There were no statistical differences in plasma cholesterol among these mice. PPARα deficiency is associated with decreased circulating glucose. In MHC-Pparg mice, glucose concentrations were similar to those in controls; and in MHC-Pparg/Ppara–/– mice, glucose concentrations were equivalent to those in Ppara–/– mice.
Heart weights of PPARγ mice were increased, and this increase was not reduced in MHC-Pparg/Ppara–/– mice (Figure (Figure1A).1A). However, echocardiography showed improved heart function in 3-month-old male MHC-Pparg/Ppara–/– compared with MHC-Pparg mice. As expected, MHC-Pparg mice exhibited reduced fractional shortening (FS) and increased systolic left ventricular dimensions (LVDs) compared with wild-type controls (Figure (Figure1,1, B–D). In stark contrast, LV function of the MHC-Pparg/Ppara–/– mice was improved and was not different from that of the wild-type control mice. Consistent with the echocardiography data, the survival rate increased from 58% to 92% at 130 days in MHC-Pparg/Ppara–/– mice (P < 0.05, Figure Figure1E).1E).
The finding of improved cardiac function with persistently increased cardiac size was surprising. To assess whether PPARα deficiency altered heart lipid accumulation, we stained heart tissues with oil red O. As shown in Figure Figure2A,2A, MHC-Pparg hearts had more lipid than wild-type, whereas Ppara–/– hearts had less neutral lipid. Despite the improved cardiac function, loss of PPARα did not reduce lipid accumulation in the hearts of MHC-Pparg mice. Both MHC-Pparg and MHC-Pparg/Ppara–/– mice had copious amounts of stained lipids. Heart tissue TAG data was consistent with the oil red O staining pattern, demonstrating significantly higher TAG levels in both MHC-Pparg and MHC-Pparg/Ppara–/– mice (Figure (Figure2B).2B). Heart tissue FFA levels were significantly increased in MHC-Pparg/Ppara–/– mice compared with MHC-Pparg mice (10.03 ± 0.86 vs. 7.51 ± 1.01 mol/l, P < 0.01) (Figure (Figure2C);2C); FFA levels in Ppara–/– and control hearts were similar. MHC-Pparg hearts had a significant increase in palmitic acid (C16:0) and decrease in docosahexaenoic acid (C22:6n3) (DHA) and docosapentaenoic acid (C22:5n6) (DPA) concentrations compared with control, but MHC-Pparg and MHC-Pparg/Ppara–/– hearts had a similar FA distribution (Supplemental Figure 2). These data suggested that the improved function of MHC-Pparg/Ppara–/– hearts was not due to reduction in palmitate.
Ceramide levels were increased in MHC-Pparg mice compared with control mice (Figure (Figure3A)3A) but were not altered by crossing the MHC-Pparg transgene onto the Ppara–/– background. However, C24:1 and C24 ceramides, which have been shown to be more toxic (23, 24), were slightly but significantly decreased in MHC-Pparg/Ppara–/– mice (Figure (Figure3B).3B). DAG, measured using the DAG kinase method, was also increased, but neither the level nor the distribution correlated with the improved cardiac function (Supplemental Figure 3A).
Long-chain FA coenzyme A (LCCoA) and acylcarnitine are intermediates in lipid oxidation. PPARα deficiency significantly increased total intracellular LCCoA content in MHC-Pparg hearts (86.5 ± 10.4 nmol/g vs. 61.9 ± 9.7 nmol/g; P < 0.01) (Figure (Figure3C3C and Supplemental Figure 3B), consistent with reduced FAO. LCCoAs are converted to acylcarnitines prior to mitochondrial FA β-oxidation; LCCoA can also be utilized for peroxisome β-oxidation and TAG synthesis (25, 26). Despite increased LCCoA content, acylcarnitine content was reduced in MHC-Pparg/Ppara–/– compared with MHC-Pparg mice (1,189.7 ± 163.5 nmol/g vs. 1,611.7 ± 98.6 nmol/g; P < 0.01) (Figure (Figure3D),3D), especially associated with a reduction in medium- and long-chain acylcarnitines (Supplemental Figure 3C). Because acylcarnitine is a toxic catabolite (27–29), reduction of acylcarnitine content and increased TAG storage might have, at least partially, contributed to amelioration of cardiac dysfunction in MHC-Pparg/Ppara–/– mice.
Electron microscopy showed distorted mitochondrial contours and more lipid droplets within the sarcoplasm of cardiomyocytes in the MHC-Pparg and MHC-Pparg/Ppara–/– mice. MHC-Pparg/Ppara–/– hearts showed a striking redistribution of lipid storage into large droplets (Figure (Figure4A).4A). In some areas of MHC-Pparg/Ppara–/– hearts, lipid droplets were surrounded by mitochondria (Figure (Figure4A).4A). In addition, as shown at higher power, disrupted cristae were seen in MHC-Pparg but not in MHC-Pparg/Ppara–/– hearts (Figure (Figure4B).4B). Lipid droplet size was significantly increased in MHC-Pparg/Ppara–/– mice compared with MHC-Pparg mice (1.93 ± 0.81 vs. 0.77 ± 0.32 μm, P < 0.001) (Figure (Figure4C).4C). MHC-Pparg mice had greater cardiac uptake of VLDL-TAG than wild-type or Ppara–/– mice (Figure (Figure4D),4D), and this was not reduced in MHC-Pparg/Ppara–/– mice. Cardiac 2-deoxyglucose uptake was increased in MHC-Pparg and Ppara–/– mice but was not further increased in MHC-Pparg/Ppara–/– mice (Figure (Figure4E).4E).
To further investigate the metabolism of these hearts, we determined palmitate oxidation and myocardial oxygen consumption (MVO2) in isolated working hearts (30). Rates of palmitate oxidation were not increased in MHC-Pparg mice (Figure (Figure5A).5A). However, palmitate oxidation rates were increased by 27.3% and 32.5% in MHC-Pparg/Ppara–/– mice versus controls and MHC-Pparg hearts, respectively (P < 0.05). MVO2 was reduced by 31% in Ppara–/– hearts relative to control hearts (P < 0.001, Figure Figure5B)5B) and was unchanged in MHC-Pparg hearts. However, PPARα deficiency increased MVO2 by 18% in MHC-Pparg hearts compared with controls (P < 0.001). Cardiac efficiency, which reflects work performed per unit of oxygen consumed, was reduced in MHC-Pparg hearts because cardiac work was reduced despite “normal” levels of MVO2 (Figure (Figure5C).5C). In MHC-Pparg/Ppara–/– hearts, although cardiac efficiency was similar to that in MHC-Pparg hearts, cardiac power was increased by 49% relative to MHC-Pparg (P < 0.001, Figure Figure5D).5D). Thus, MHC-Pparg/Ppara–/– hearts utilized more oxygen due to greater FAO but had improved function. And, at least under nonischemic conditions, MHC-Pparg/Ppara–/– mice had improved cardiac function associated with greater FAO. This was associated with increased mitochondrial mass as estimated by the ratio of the mitochondrially encoded gene ATPase6 to β-actin and mitochondrial transcription factor A (mtTFA) to 18s (Figure (Figure5,5, E and F). There was no change in mitochondrial function detected in permeabilized cardiac fibers (Supplemental Figure 4, A–C).
We determined whether reduction of PPARα expression would reduce PPARα target gene expression in the MHC-Pparg mice. We unexpectedly found an increase in PPAR target gene expression in MHC-Pparg/Ppara–/– mice (Figure (Figure66 and Supplemental Table 1); mRNAs were increased for genes mediating lipid uptake, synthesis, oxidation, lipolysis, and storage, such as fatty acid translocase (Cd36), fatty acid synthase (Fasn), acyl-CoA oxidase (AOX), adipose TAG lipase (Atgl), adipose differentiation related protein (Adrp), and diacylglycerol acyltransferase1 (Dgat1). The increase in LCCoA content was associated with increased mRNA expression of long-chain fatty acyl-CoA synthetase (Acsl4), an enzyme expressed in the heart peroxisome (31). However mitochondrial acyl-CoA thioesterase 2 (Acate2), an enzyme that catalyzes the hydrolysis of acyl-CoAs to the FFA and coenzyme A (CoASH), was also increased in MHC-Pparg/Ppara–/– mice (Supplemental Table 1). Expression of PPARγ coactivator 1α (Pgc-1α) and oxidative phosphorylation–related (OXPHOS) genes were unchanged by crossing with Ppara–/– mice (Supplemental Table 1). Additionally, PPARα deficiency dramatically increased cellular retinol-binding protein III (CRBPIII) expression in MHC-Pparg mice (5.91-fold, P < 0.05). In contrast, there was no change in genes regulating glucose metabolism. Thus, the absence of PPARα appeared to increase the efficacy of the MHC-Pparg transgene in driving expression of PPAR target genes in the cardiomyocyte. PPARα deficiency did not alter endogenous PPARδ and PPARγ expression (Supplemental Figure 5).
Because loss of PPARα was, curiously, associated with greater expression of several PPARγ downstream genes, we tested the hypothesis that PPARα blocked the actions of MHC-Pparg by treating MHC-Pparg mice with the potent PPARα agonist WY-14,643. Ten days of WY-14,643 treatment decreased plasma TAG without altering FFA or glucose levels (Figure (Figure7A).7A). Heart tissue oil red O staining showed reduced heart neutral lipid in WY-14,643–treated MHC-Pparg mice (Figure (Figure7B). 7B). Expression of PPARγ downstream targets Cd36, Fasn, and CRBPIII (32) was reduced compared with that in nontreated MHC-Pparg mice. FAO and mitochondrial biogenesis genes, such as Lcad, Ucp3, and Pgc-1α, were also reduced with WY-14,643 treatment. However, expression of Cpt1, AOX, and Dgat1 was not altered by WY-14,643 treatment in MHC-Pparg mice (Figure (Figure7C). 7C). These data suggested that PPARγ activity may be regulated by PPARα.
The pattern of intracellular lipid storage relates to the metabolic characteristics of cells; larger droplets have reduced total surface area and are likely to reflect more inert lipid storage (33). We then studied the reasons for the redistribution of cellular lipids. Adrp, perilipin 4, and cell death–inducing DFFA-like effector c (Fsp27) are associated with the surface of intracellular lipid droplets and regulate their formation and mobilization (33–35) and are regulated by PPARγ (36). mRNA expression of Adrp, perilipin 4, and Fsp27 was upregulated in MHC-Pparg mice by crossing with Ppara–/– mice (Figure (Figure6,6, Supplemental Table 1, and Supplemental Figure 6).
Several intracellular lipids are potent signaling molecules and are thought to function as activators of classical PKCs (37, 38). To determine whether lipid redistribution led to altered intracellular lipid activation, we measured heart PKC activation. MHC-Pparg increased membrane PKCα and PKCδ. This increase was ameliorated in the MHC-Pparg/Ppara–/– mice (Figure (Figure8A).8A). Thus, the altered lipid distribution changed lipid activation of this potentially toxic pathway.
Greater FAO is associated with more ROS production (39) and in some situations cellular apoptosis (40). However, TUNEL-positive myocytes were reduced in MHC-Pparg/Ppara–/– compared with MHC-Pparg mice, and expression of the apoptosis-related protein Bax and p-JNK was decreased in the heart of MHC-Pparg/Ppara–/– mice (Figure (Figure8,8, B–D). Heart tissue intracellular O2– levels measured using dihydroethidium and fluorescence staining were increased in MHC-Pparg mice, but this was markedly attenuated in MHC-Pparg/Ppara–/– mice (Figure (Figure9A).9A). Mitochondrial and ER stress markers prohibitin and protein disulfide isomerase (PDI) (41, 42) were also decreased in the MHC-Pparg/Ppara–/– hearts (Figure (Figure9,9, B and C). This was associated with increased expression of Sod2 but no differences in expression of Gpx1, catalase, Sod1, and Ucp2 between MHC-Pparg and MHC-Pparg/Ppara–/– mice. Expression of cardiac hypertrophic markers and apoptosis-related genes, such as natriuretic peptide precursor type B (BNP), serine hydrolase-like (Serh1), DNA damage–inducible transcript 3 (chop), and caspase-6, was significantly decreased in MHC-Pparg/Ppara–/– mouse hearts (Figure (Figure6B,6B, Figure Figure9D,9D, and Supplemental Table 1).
Lipotoxicity-induced organ dysfunction is becoming an increasingly common cause of human disease. In the heart, an organ that primarily uses lipids as a source of energy, lipid accumulation assessed as TAG content correlates with reduced function (4, 5, 43, 44). This relationship is found with obesity, metabolic syndrome, and diabetes (6). The pathological pathways linking cellular lipids to cardiac dysfunction are obscure. Our objective was to determine whether excess FAO, one putative cause of lipotoxicity, was corrected in the MHC-Pparg hearts by reduced PPARα expression and whether this would also correct heart dysfunction. By performing a cross of MHC-Pparg onto the Ppara–/– background, we rescued the heart dysfunction and uncovered a number of fundamental issues related to the causes of heart lipotoxicity. We clearly show that several factors previously implicated as causes of lipotoxicity were not corrected in this model. These include FAO and TAG storage, neither of which were decreased in the MHC-Pparg/Ppara–/– hearts. In addition, there were no major changes in DAG or ceramide levels. Therefore, although loss of PPARα improved heart function and normalized 5-month survival in the MHC-Pparg mice, it did not do this by the expected mechanisms.
Why did the hearts improve? What was altered was the distribution and size of the lipids: the droplets were significantly larger. Macro- versus microsteatosis has been associated with less dysfunction in the liver (45). Membrane-associated (active) PKCα and PKCδ were both reduced by the PPARγ/Ppara–/– cross, an indication that the redistribution of stored lipid was associated with reduced lipid signaling. MHC-Pparg/Ppara–/– hearts had reduced levels of acylcarnitines. High cardiac tissue levels of acylcarnitine can disrupt biomembranes through nonspecific detergent actions, induce electrophysiological alterations, and inhibit several critical enzyme systems (46–48). Because LCCoA was increased in the rescued mice, this lipid pattern suggests that the rescued mice more efficiently oxidized acylcarnitines and did not convert unused FAs to this potentially harmful lipid. mRNA levels of Atgl and Dgat1 were increased in the rescued MHC-Pparg/Ppara–/– hearts. In other situations, Dgat1 (49, 50) and Atgl (51, 52) expression correlates with increased FAO, and transgenic expression of either of these enzymes relieves lipotoxicity in the heart (53). Thus, we speculate that it is preferable to shunt FAs to the lipid droplet and then allow TAG hydrolysis to oxidative substrates, rather than directly convert internalized FAs into LCCoA and acylcarnitines.
Cardiac muscle is the most energy-requiring tissue in the body and primarily uses FAs and to a great extent lipoprotein-derived FAs (54). It has been suggested that reduction of FAO might be beneficial in that glucose oxidation is a less oxygen-requiring event. Theoretically, such a switch in substrate use would be especially important in the setting of ischemia and has led to the development of FA inhibitors as a treatment for angina (55). In models of lipotoxicity and hypertrophy, it has also been argued that excess FAO is harmful (4, 56). In contrast, high-fat diets are not harmful in rat models of cardiac hypertrophy (57) and in human heart failure (58). Greater FAO in this and one other model (53) of lipotoxicity is clearly not toxic.
Although cardiac toxicity especially in diabetic models is associated with defective mitochondrial function due to mitochondrial uncoupling (59), we found no evidence for changes in mitochondrial function. It should be noted that FFA levels were increased in the PPARα knockout. While this could have supplied more FA to the hearts, the isolated perfused heart studies and gene expression studies clearly show that MHC-Pparg/Ppara–/– hearts were primed to oxidize more FAO.
Prohibitin is a mitochondrial marker protein localized in the inner membrane of mitochondria. It responds to mitochondrial stress and is induced by metabolic stress caused by an imbalance in the synthesis of mitochondrial- and nuclear-encoded mitochondrial proteins (41). MHC-Pparg heart sections had much more intense staining for prohibitin compared with those from control mice. PPARα deficiency markedly reduced the staining in MHC-Pparg mice. These data indicate that, in MHC-Pparg/Ppara–/– mice, the improved cardiac function was associated with reduced mitochondrial stress.
MHC-Pparg/Ppara–/– hearts had less ROS, reduced ER stress–related proteins, and less apoptosis. PDI, a very abundant protein in the ER, is an essential folding catalyst and chaperone of the ER (60, 61). There was a large increase in PDI staining in hearts from MHC-Pparg compared with control mice, but this was reduced with PPARα deficiency. These data suggested that MHC-Pparg mice had more PDI protein to catalyze disulfide formation, which protects against protein misfolding in the ER.
Although our focus was primarily on cardiac function, we also uncovered evidence that at least in the heart, PPARγ and PPARα are not equivalent and their activation is not always synchronous. Unexpectedly, MHC-Pparg/Ppara–/– hearts had greater expression of several genes regulating FA metabolism. This was surprising, and we then questioned whether in the heart PPARγ was a more potent activator of these genes and whether its actions in the MHC-Pparg mice were actually blocked by the presence of PPARα. We tested this by treating MHC-Pparg mice with the potent agonist WY-14,643. Expression of two PPARγ downstream genes, Cd36 and CRBPIII, was reduced by this agonist; a cardiac PPARα target, Cpt1, was induced. Therefore, it appears that not all PPAR activation is of similar potency and PPARγ actions were diminished by the actions of PPARα. The molecular events for this phenomenon — competition for promoter binding sites, coactivators or repressors, or lipid agonists — are under study.
Because all three members of the PPAR gene family have been overexpressed in the heart, their functional and genetic effects can be compared. PPARα, which increases FAO, leads to greater lipid uptake in the heart, and this lipid uptake is out of proportion to the induction of oxidation (4). PPARα is most highly expressed in the liver, a tissue in which lipid uptake that exceeds oxidation can be compensated by TAG secretion in lipoproteins. In contrast, hearts from MHC-Ppard mice have increased FAO but no cardiac dysfunction or accumulation of lipid; oxidation and lipid uptake are balanced (62). Thus, this gene expression program is most suitable for muscles in which greater oxidation allows for more energy production but without leading to toxicity, e.g., with exercise (63). Like PPARα, PPARγ increases lipid uptake in excess of that needed for oxidation. However, the MHC-Pparg transgene does not increase Pdk4 and is not associated with reduced glucose uptake (15). These changes in gene expression are appropriate to increase lipid stores in adipose but can lead to toxicity in muscles. The overlapping and competing actions of PPAR transcription factors may be central to metabolic regulation.
Although PPARγ is expressed in the heart at much lower levels than in adipose (64, 65), use of agents that lead to marked induction of this transcription factor are associated with cardiac toxicity (66). In humans, several PPARγ agonists increase the incidence of clinical heart failure (67, 68). PPARγ agonists have a number of actions, including salt and water retention and vasodilatation (69), and these actions or direct cardiac toxicity might be responsible for the adverse effects in humans. Hearts of patients with metabolic syndrome have greater lipid accumulation; these hearts also have increased expression of PPARγ (70). Thus, cardiac PPARγ expression is either responsible for or coincidental with cardiac lipid content in humans.
Reduced cardiac function with increased lipid is found in patients with obesity (71) and diabetes (72). Aside from lipid, heart dysfunction with diabetes is postulated to involve microvascular disease, glycosylation of intracellular proteins, and greater FAO (73). Models in which lipid accumulation is increased solely in the heart are likely to be more flagrant representations of human cardiac lipotoxicity and provide the opportunity to dissect the roles of lipids and FAO in cardiac dysfunction exclusive of generalized metabolic derangements. Our data show that processes thought to cause lipotoxicity can be illustrated and eliminated by genetic methods. Excess FAO was not corrected in the MHC-Pparg/Ppara–/– mice. Thus, this is not a likely cause of toxicity. Rather, further induction of FAO appears to be protective. Stores of TAG, although a marker of the disease, are also not a cause of toxicity. Improved cardiac function correlated with a redistribution of intracellular lipids and reduced activation of PKCs, despite unchanged levels of total DAG and ceramide. Because this cross-corrected many processes associated with lipotoxicity — ROS, apoptosis, and PKC signaling were reduced — a single specific cause and toxic pathway cannot be discerned. Reduced lipid uptake into the heart corrects cardiac toxicity in MHC-Ppara mice (74, 75) presumably by reducing accumulated lipid. Our study underscores another option for treatment of lipotoxic cardiomyopathy and perhaps other lipotoxic diseases, namely increased cellular lipid oxidation and/or greater storage of relatively “nontoxic” lipid moieties.
The high-expressing MHC-Pparg mouse line (15) in a mixed C57BL/6 × CBA/J background was backcrossed with C57BL/6 wild-type mice (The Jackson Laboratory) for 10 generations, resulting in greater than 99% C57BL/6 purity. MHC-Pparg mice were then crossed twice into the Ppara–/– (The Jackson Laboratory) background, resulting in MHC-Pparg/Ppara–/– offspring. Studies of gene expression, histology, lipid analysis, and cardiac function by echocardiography were performed using mice fed chow diets with overnight fasting. All animal experiments were approved by the Institutional Animal Care and Use Committee of the Columbia University Medical Center.
Blood from fasted (16 hours) mice was collected from the retro-orbital plexus for the measurement of plasma total cholesterol (TC), TAG, and FFAs. For measurement of tissue lipids, hearts were perfused with PBS and homogenized at 4°C in 1 M NaCl buffer containing lipase inhibitors to prevent TAG hydrolysis. Lipids were extracted from heart tissues (50 mg) according to methods modified from that of Folch et al. (76). The dried lipids were solubilized in PBS containing 2% Triton X-100. Heart and plasma TC, TAG, and FFAs were measured enzymatically using an Infinity kit (Thermo Electron Corp.) and a NEFA C kit (Wako). Heart tissue FA analysis was performed using gas chromatographic methods as described previously (77).
Heart tissues from 3-month-old mice were prepared, and whole-cell proteins were isolated using RIPA lysis buffer (sc-24948; Santa Cruz Biotechnology Inc.). Thirty micrograms of nuclear proteins were subjected to Western blot analysis with the following antibodies: PPARγ (sc-7196; 1:100 dilution) and PPARα (sc-9000; 1:200 dilution) from Santa Cruz Biotechnology Inc.; Bax (ab54481; 1:500 dilution), PKCα (ab4142; 1:1,000 dilution), PKCδ (ab4143; 1:1,000 dilution), and Gapdh (ab9485; 1:5,000 dilution) from Abcam Inc. Immunoreactive bands were visualized using SuperSignal West Pico Chemiluminescent Substrate (Pierce). The bands were quantified by densitometry using Molecular Analysis Software (Bio-Rad).
Total RNA was prepared using a Pure Link Micro-to-Midi Total Purification System kit (Invitrogen). One microgram of RNA was initially treated with DNase I (Invitrogen) for 15 minutes. The RNA samples were then reverse transcribed using the ThermoScript RT-PCR Kit (Invitrogen). Quantitative real-time RT-PCR (qRT-PCR) was performed using an ABI 7700 (Applied Biosystems). Amplification was performed using SYBR Green PCR Master Mix (Applied Biosystems). Primers used for PCR amplification are listed in Supplemental Table 2. Analysis was performed using Sequence Detection Software (Applied Biosciences). Standard curves were generated using undiluted and diluted (1:10, 1:100, and 1:1,000) cDNA samples from heart tissue. Correlation coefficients were 0.98 or greater. Data were normalized with 18s rRNA.
Mouse heart tissue samples in RNA later were sent to Ocean Ridge Biosciences (ORB) for analysis using mouse exonic evidence-based oligonucleotide (MEEBO) microarrays (lot 20788). MEEBO microarrays were printed by Microarrays Inc. and contained 38,083 70-mer oligonucleotides probes complementary to constitutive exons of most mouse genes, as well as alternatively spliced exons, and control sequences. For statistical analysis, samples were binned in two treatment groups (MHC-Pparg and MHC-Pparg/Ppara–/–). The log2-transformed and normalized spot intensities for the 19,420 detectable probes were examined for differences between the treatment groups by 1-way ANOVA using National Institute on Aging (NIA) Array Analysis software. This ANOVA was conducted using the Bayesian error model and 10 degrees of freedom. Statistical significance was determined using the false discovery rate (FDR) method. It is the proportion of false positives among all probes with P values lower or equal to the P value of the probes that we considered significant. FDR is an intermediate method between the P value and Bonferroni correction (multiplying P value by the total number of probes). The equation is:
where r is the rank of a probe ordered by increasing P values, pi is the P value for probe with rank i, and N is the total number of probes tested. FDR value increases monotonously with increasing P value.
Mitochondrial DNA was quantified by calculating the ratio of mitochondrial gene copy number (ATPase6) to nuclear gene copy number (β-actin) (78, 79). Heart DNA was extracted from frozen tissue with a DNeasy tissue kit (QIAGEN).
Neutral lipids were assessed in hearts taken from 16-hour-fasted mice perfused with PBS. The hearts were embedded in Tissue-Tek Optimal Cutting Temperature compound (Sakura). Midventricular sections of myocardium (6 μM in thickness) were stained with oil red O and H&E and counterstained with hematoxylin. We used dihydroethidium (D23107; Invitrogen) to examine the superoxide in the frozen heart tissue sections. Immunohistochemical staining techniques were used on frozen sections to examine the mitochondria-associated protein prohibitin (ab1836; Abcam). An Endoplasmic Reticulum Labeling Kit was used for detecting ER-associated PDI with a primary antibody directed against PDI antibody and an Alexa Fluor 488 dye–labeled secondary antibody (S34252; Invitrogen). Apoptotic cardiomyocytes were stained by a TUNEL protocol according to the manufacturer’s specifications (R&D Systems). The data were quantified by counting TUNEL-positive myocytes per square millimeter cellular area.
Transthoracic echocardiography was performed in conscious mice using a high-resolution imaging system with a 30-MHz imaging transducer (Vevo 770; VisualSonics). Each parameter was measured using M-mode view. Percent fractional shortening (%FS) was calculated as follows: %FS = (LVDd – LVDs)/LVDd × 100, where LVDd is left ventricular diastolic dimension and LVDs is left ventricular systolic dimension (80).
Left ventricles from 3-month-old mice were fixed with 2.5% glutaraldehyde in 0.1 M Sorensen’s buffer (0.2 M monobasic phosphate/0.2 M dibasic phosphate, 1:4 vol/vol; pH 7.2), postfixed in osmium tetroxide, and embedded in EPON 812 (Electron Microscopy Sciences). Ultrathin sections were stained with uranyl acetate and lead citrate and examined under a JEM-1200ExII electron microscope (JEOL).
Human VLDL was isolated from normal subjects by sequential ultracentrifugation and was labeled with [carboxyl-14C]triolein (PerkinElmer) via cholesterol ester transfer protein as previously described (81). Sixteen-hour-fasted mice were first injected intravenously with 1 × 106 decays per minute (DPM) of 2-deoxy-d-[3H]glucose (PerkinElmer). Fifty-five minutes after 2-deoxy-d-[3H]glucose injection, 1 × 106 DPM of [14C-TAG]VLDL was injected. Five minutes after VLDL injection, blood was collected, and the vasculature was thoroughly perfused with 10 ml PBS via cardiac puncture. Tissues were then excised, and accumulated radioactivity for [3H]glucose and [14C-TAG]VLDL was measured. Amounts of glucose and VLDL injected were adjusted by plasma radioactivity counts at 30 seconds after each injection and were compared with plasma counts at the end of the experiments.
Cardiac metabolism was measured in isolated working mouse hearts as described previously (30).
Mitochondrial oxygen consumption and ATP production were measured in permeabilized cardiac fibers isolated from 12-week-old wild-type, MHC-Pparg, Ppara–/–, and MHC-Pparg/Ppara–/– mice using techniques described previously (59, 82).
Cardiac ceramide and diacylglycerol (DAG) levels were determined using the DAG kinase method as described previously (83). Individual species of DAG and ceramides were measured by liquid chromatography–mass spectrometry as described previously (84).
Heart tissues (100 mg) from 3-month-old mice were homogenized and extracted in cold 20 mM Tris-HCl (pH 8.0), 2 mM EDTA, 2 mM EGTA, 6 mM 2-mercaptoethanol, 0.1 mM sodium vanadate, 50 mM NaF, and complete inhibitor cocktail (Roche). The homogenate was solubilized and centrifuged at 4°C for 1 hour at 100,000 g. Separation of cytosol and membrane fraction was done as described previously (85). Equivalent amounts of the two fractions were subjected to SDS-PAGE. Proteins were electroblotted onto nitrocellulose membranes, which were probed with rabbit anti-peptide antibodies specific for PKCα and PKCδ isozymes (Santa Cruz Biotechnology Inc.). PKC isoenzymes were then visualized by incubation of the membrane with enhanced chemiluminescence reagents and exposure to X-ray film. Densitometry of PKC bands was carried out using a Medical Dynamics Personal Densitometer and analyzed using QuantityOne software (Bio-Rad).
PPARα agonist (WY-14,643) was purchased from Sigma-Aldrich and given as 50 mg/kg/d in aqueous hydroxypropyl methylcellulose once daily for 10 days by oral gavage (86).
We analyzed data using the Prism software package (GraphPad Software). Comparisons between two groups were performed using unpaired 2-tailed Student’s t tests. All values are presented as mean ± SD. Differences between groups were considered statistically significant at P < 0.05.
We thank K.G. Bharadwaj, H. Jiang, and L. Liu for help with the kinetic studies, PCR, and measurement of ceramide and DAG; and Jamie Soto for help with mitochondrial analyses. These studies were supported by grants HL45095, HL73029, P50 HL077113, and U01HL087947 from the National Heart, Lung, and Blood Institute.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2010;120(10):3443–3454. doi:10.1172/JCI40905.