Carnitine palmitoyl transferase-I inhibition prevents ventricular remodeling and delays decompensation in pacing-induced heart failure

Abstract

Objective: Experimental evidence suggests that modulation of myocardial substrate metabolism can markedly affect the progression of chronic heart failure (HF). We tested whether the inhibition of carnitine palmitoyl transferase-I (CPT-I), the enzyme regulating mitochondrial fatty acid oxidation, slows left ventricular remodeling and deterioration of function in pacing-induced HF.

Methods: Normal dogs (n=9) were compared to untreated dogs with pacing-induced HF (n=9) and HF dogs treated with 65 mg/kg/day of oxfenicine (HF+Oxf, n=9), a CPT-I inhibitor.

Results: HF+Oxf reached terminal failure (LV end-diastolic pressure=25 mm Hg) 6 days later than untreated HF (P<0.05). At 28 days of pacing, hemodynamic alterations and LV dilation were significantly attenuated and the 25% decrease in LV wall thickness was completely prevented in HF+Oxf vs. untreated HF, as was the activation of matrix metalloproteinase-2 and -9, markers of tissue remodeling. Oxfenicine also prevented HF-induced transcriptional down-regulation of CPT-I, medium chain acyl-CoA dehydrogenase, GAPDH and citrate synthase, key enzymes of cardiac energy metabolism. In addition, mRNA, but not protein levels of the nuclear receptor peroxisome proliferator-activated receptor-α were reduced in untreated HF, while they did not change significantly in HF+Oxf, as compared to control.

Conclusions: CPT-I inhibition early in the development of HF prevented LV wall thinning and delayed the time to end-stage failure. While these results are limited to an experimental model of disease, they nevertheless suggest that CPT-I inhibition might be effective for slowing the progression of clinical HF.

This article is referred to in the Editorial by Rupp et al. (pages 423–426) in this issue.

1. Introduction

Clinical and animal studies show that in the early stages of heart failure (HF) myocardial substrate selection is relatively normal, whereas in advanced HF there is reduced myocardial oxidation of fatty acids, the main fuel for normal cardiac muscle [1]. Reduced fatty acid oxidation (FAO) in advanced HF has been attributed to down-regulation FAO enzymes, as seen in a decreased activity and expression of both carnitine-palmitoyl transferase-I (CPT-I), the enzyme that regulates long chain fatty acyl uptake by mitochondria, and medium chain acyl-CoA dehydrogenase (MCAD), a key enzyme of FAO [2,3]. These findings are consistent with reduced expression/activation of the peroxisome proliferator-activated receptor-α (PPARα) in HF, which regulates the transcription of FAO genes and is activated by long chain fatty acid ligands [4,5]. It has been proposed that pharmacological inhibition of CPT-I might slow the progression of HF by improving cardiac energy metabolism due to reduced FAO and increasing the oxidation of glucose, which is a slightly more oxygen efficient fuel [1,6]. Administration of a CPT-I inhibitor to rats with left ventricular (LV) hypertrophy secondary to aortic banding partially prevented deterioration in cardiac function [7]. Rupp et al. suggested that a secondary effect of CPT-I inhibition might be the activation of the expression of PPARα-regulated genes (and perhaps other genes) that lead to improved cardiac function and prevention of LV remodeling [8]. This approach differs from a pure reactivation of the PPARα gene regulatory pathway, which was shown by Young et al. to exacerbate cardiac function in overloaded hearts [5]. In fact, there would be suppression of FAO due to CPT-I inhibition, and, at the same time, extramitochondrial accumulation of lipid ligands that would stimulate the activity of the PPARα pathway and the expression of FAO enzymes [8].

Although the inhibition of CPT-I has been proposed as a therapy for HF, this approach has not been evaluated in a large animal model of HF. In addition, it is not known whether CPT-I inhibition leads to greater expression of PPARα-regulated genes. Thus, the aim of the present study was to test if inhibition of CPT-I slows LV remodeling and functional deterioration in pacing-induced HF. Pacing-induced HF is a model of dilated cardiomyopathy that follows a very predictable time course [9,10], therefore, the effects of CPT-I inhibition on the transition to decompensation could be precisely determined. Left ventricular contractile function and remodeling were assessed, including activation of matrix metalloproteinases (MMPs), markers of interstitial matrix remodeling [11], as well as expression and activity of key enzymes of substrate metabolism.

2. Methods

2.1. Surgical instrumentation

Studies were performed on male mongrel dogs (n=26; aged 12 to 18 months; weight 23–27 kg). The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). The animals were sedated with acepromazine maleate (1 mg/kg, i.m.), anesthetized with sodium pentobarbital (25 mg/kg, i.v.), ventilated and instrumented as previously described [3,9]. Briefly, a thoracotomy was performed in the left fifth intercostal space. One catheter was placed in the descending thoracic aorta and a solid-state pressure gauge (P6.5, Konigsberg Instruments) was inserted in the left ventricle through the apex. A Doppler flow transducer (Craig Hartley) was placed around the left circumflex coronary artery. A human, screw-type, unipolar myocardial pacing lead was placed on the left ventricular wall. Seven out of the nine HF dogs that received oxfenicine had a catheter permanently implanted into the coronary sinus. Wires and catheters were run subcutaneously to the intra scapular region, the chest was closed in layers and the pneumothorax was reduced. Antibiotics were given after surgery and the dogs were allowed to fully recover. After 7–10 days of recovery, dogs were trained to lie quietly on the laboratory table.

2.2. Experimental protocol

HF was induced in 18 dogs by pacing the LV at 210 bpm for 3 weeks. The pacing rate was increased to 240 bpm thereafter. Dogs were checked twice daily for well being, and care was taken to insure that the dogs did not suffer undue discomfort. Nine of the paced dogs received 65 mg/kg/day per os of oxfenicine (HF+Oxf), a selective inhibitor of CPT-I [12,13] from the first day of pacing until end-stage failure. Oxfenicine doses of 17 to 65 mg/kg partially inhibit myocardial fatty acid oxidation in dogs and in pigs [12,13] and are well below the doses that trigger the moderate cardiac hypertrophy shown with long-term administration in normal dogs [14]. In pilot experiments in conscious dogs, we found that 5–7 days of treatment with 65 mg/kg oxfenicine per os caused a 50% reduction in cardiac free fatty acids uptake (from 18 to 9 μmoL/min). In order to obtain normal myocardium as control, a group of 8 similarly instrumented healthy dogs was used. Experiments were conducted in conscious dogs, at spontaneous heart rate, with the pacemaker turned off, after an overnight fasting. Hemodynamic variables were recorded and echocardiographic measurements were performed at baseline and every week until the end of the pacing protocol. Dogs were considered in end-stage HF when LV end-diastolic pressure reached 25 mm Hg and clinical signs of severe decompensation were observed [3,9]. At this point, the dogs were anesthetized with 30 mg/kg i.v of sodium pentobarbital to harvest and snap-freeze LV tissue samples as previously described [3,15]. The heart was then removed, dissected and weighed.

2.3. Hemodynamics, echocardiographic recordings and calculated parameters

The aortic catheter was attached to a P23ID strain-gauge transducer for measurement of aortic pressure. LV pressure was measured using the solid-state pressure gauge. The first derivative of LV pressure, LV dP/dt, was obtained using an operational amplifier (National Semiconductor LM 324). Coronary blood flow was measured with a pulsed Doppler flowmeter (Model 100, Triton Technology). All signals were recorded on an eight-channel direct-writing oscillograph (Gould RS 3800). The analog signals were also stored in computer memory through an analog-digital interface (National Instruments), at a sampling rate of 250 Hz [3,9,15]. Two-dimensional and M-mode echocardiography was also performed. Images were obtained from a right parasternal approach at the mid-papillary muscle level. Measurements were made according to the criteria of the American Society of Echocardiography [16].

2.4. Gene expression analysis

RNA was extracted using standard methods and analyzed using reverse transcription followed by real-time quantitative polymerase chain reaction (PCR) for the transcripts of interest, as described previously [15,17]. Analysis was focused on the traditional “housekeeping” genes β-actin and ribosomal 18S, as well as on glyceraldehyde-3-phosphate dehydrogenase (GAPDH), on the PPARα-regulated genes CPT-I (muscle and liver isoforms), MCAD, pyruvate dehydrogenase kinase-4 (PDK4), and uncoupling protein 3 (UCP3), and on citrate synthase (CS), RXRα and PPARα. Specific quantitative assays were designed from dog sequences available in GenBank, except in the cases of CS, MCAD, the muscle isoform of CPT-I (m-CPT-I), and RXRα, for which the dog sequences were not available. In contrast, these genes have been cloned for the human, mouse and rat. Multiple assays were then designed using the human sequence for these genes, in a region that was highly conserved between the three species. These human-specific assays were subsequently tested for compatibility in the dog. Primers and probes were designed to be isoform specific, spanning sites where two exon join (splice sites) when such sites are known (preventing recognition of the assay to any potential contaminating genomic DNA). Sequences of these primers and probes are given in Table 1. Standard RNA was made for all assays by the T7 polymerase method (Ambion), using total RNA isolated from the dog heart. Transcript levels are expressed as the number of molecules of mRNA per nanogram of total RNA (as measured by UV spectrophotometry).

2.5. Western immunoblot analysis

Protein was extracted from frozen tissue as previously described [3,15] and its concentration measured by spectrophotometrically with a colorimetric assay (Bio-Rad DC reagent). Fifty microgram of total protein was separated by electrophoresis, transferred onto a PVDF membrane, and incubated with specific antibody against MCAD (1:2000 dilution, Cayman Chemical) and PPARα, RXRα, [3] and the peroxisome proliferator-activated receptor gamma coactivator-1 (PGC-1, dilution 1:100, Chemicon). After conjugation with the secondary antibody, the membranes were developed in a chemiluminescence substrate solution (Pierce SuperSignal Chemiluminescens Substrate), and re-probed for calsequestrin [15] to verify the uniformity of protein loading. Bands were visualized by autoradiography and quantified using commercially available software. Results are expressed as percentage of the density of a standard sample loaded on all membranes in triplicate.

2.6. Metabolite and enzymatic measurements

FFA in plasma was measured by an enzymatic spectrophotometric assay (NEFA C, Wako Chemicals) and cardiac uptake was calculated by multiplying coronary blood flow by the difference in FFA concentration between paired blood samples from aorta and coronary sinus [9]. The content of the triglycerides in the heart were measured using an enzymatic spectrophotometric assay (Triglyceride E kit, Wako Chemicals) as previously described [3,15]. The increase in LV end-diastolic volume and thinning of the LV free wall in HF has been shown to be dependent upon an increase in the activity of MMP-2 and MMP-9, which correspond to loosening of the collagen matrix surrounding cardiomyocytes [11,18]. The activities of cardiac MMP-2 and MMP-9 were measured by loading 10 μg of tissue extract on 10% acrylamide gels impregnated with 1 mg/ml gelatin [19]. Bands resulting from gelatin digestion were quantified in arbitrary units using NIH image software. Malondialdehyde and 4-hydroxyalkenals, end products of peroxidation of polyunsaturated fatty acids with related esters, were measured in cardiac tissue extract with a colorimetric method by employing a commercially available assay kit (Calbiochem) [20]. The activity of CPT-I was assayed radiochemically, and the activities of MCAD and CS were measured spectrophotometrically, as previously described [3,15,21].

2.7. Statistical analysis

Data are presented as mean ± standard error of the mean (SEM). Differences between control, untreated HF and HF+Oxf were tested by a one-way or two-way ANOVA, as appropriate, followed by Tukey post-hoc test. For all the statistical analyses, significance was accepted at P<0.05.

3. Results

3.1. Hemodynamic alterations

We have previously shown that our protocol of pacing induces a compensated HF during the first 3 weeks (210 beats/min), while decompensation occurs late during the fourth week (240 beats/min) [9]. Therefore, in the present study, we compared hemodynamic changes found at 3 weeks of pacing and later, i.e., during the critical transition from compensated to decompensated failure. Fig. 1, panel A, shows that a chronic CPT-I inhibition significantly delayed the onset of decompensation. We consider a LV end-diastolic pressure of 25 mm Hg as the hallmark of end-stage failure. This pressure was reached after 34 ± 1 days of pacing in HF+Oxf, which was 6 days later than in untreated HF (P<0.05). At 28 ± 1 days pacing, corresponding to end-stage failure in untreated HF, LV and aortic pressure (Fig. 1B and C) were significantly higher in HF+Oxf compared to untreated HF, whereas the progressive fall in dP/dt_max (Fig. 1D) was not affected by the CPT-I inhibition. Control blood flow prior to the initiation of pacing in the left circumflex coronary artery was 35.2 mL/min in untreated HF and 36.6 mL/min in HF+Oxf (N.S.) and did not change during the progression of failure. Spontaneous heart rate increased significantly within each group from 91.5 ± 4.9 to 128.0 ± 5.5 beats/min in untreated HF and from 95.0 ± 6.1 to 120 ± 7.0 in HF+Oxf, with no difference between groups.

3.2. LV ejection fraction and dimensions

At 28 days of pacing, LV ejection fraction was significantly greater in HF+Oxf compared to untreated HF, however at end-stage failure, it was similar in both groups and significantly reduced from the pre-pacing control values (Fig. 2A). The increase in LV end-diastolic diameter at 4-week pacing was significantly less in HF+Oxf compared to untreated HF, and became significantly different from pre-pacing baseline values only in end-stage failure (Fig. 2B). On the other hand, end-diastolic LV free wall and septum thickness decreased by 25% and 30%, respectively, in untreated HF at 4 weeks, whereas they did not change in HF+Oxf (Fig. 2C and D). The heart weight/body weight ratio was 8.5 ± 0.2 g/kg in normal dogs and increased significantly in HF+Oxf (10.9 ± 0.5 g/kg, P<0.05), but not in untreated HF (9.2 ± 0.4 g/kg).

3.3. Metabolite and enzymatic measurements

The catheter remained permanently patent in coronary sinus only in a small number of animals. In particular, it remained patent until the third week of pacing in three HF+Oxf dogs, and until end-stage failure in two HF+Oxf dogs. After 1 and 2 weeks of pacing and oxfenicine administration, cardiac FFA uptake was reduced from 19.0 ± 3.3 μmoL/min (baseline) to, respectively, 6.2 ± 3.1 and 7.1 ± 1.3 μmoL/min (P<0.05). At 3 weeks it was 10.1 ± 4.4 μmoL/min (n=3) and 2.2 μmoL/min in end-stage failure (n=2). FFA uptake measurements were not performed in untreated HF dogs. The myocardial triglyceride content was not significantly different between the normal group (12.2 ± 3.5 μmoL/g wet weight), the untreated HF group (13.2 ± 2.1 μmoL/g wet weight), and the HF+Oxf group (15.8 ± 2.0 μmoL/g wet weight). In end-stage failure, the activities of MMP-2 and MMP-9 were increased by 146% and 295%, respectively, in untreated HF compared to normal hearts, but not in the HF-Oxf group (Fig. 3). Similarly, levels of lipoperoxidation end products (malondialdehyde+4-hydroxyalkenals) were higher in untreated HF, compared to normal hearts (7.9 ± 0.97 vs. 5.3 ± 0.67 nmoL/mg wet weight, P<0.05), but not in HF+Oxf (4.3 ± 0.4 nmoL/mg wet weight, N.S.).

Compared to the normal group, HF caused significant down-regulation of PPARα, RXRα, GAPDH, CS, m-CPT-I, PDK4, and UCP3, a modest but significant increase in the ribosomal RNA 18S, with no significant changes in β-actin, l-CPT-I or MCAD (Fig. 4). This down-regulation was prevented by CPT-I inhibition. In HF-Oxf, we found even an increase in m-CPT-I, PDK4, and UCP3 gene expression, compared to control, that however, due to the large variability among dogs, did not result statistically significant. The effects of oxfenicine on gene expressions were observed whether mRNA values were expressed relative to total RNA content (Fig. 4) or relative to the housekeeping ribosomal RNA 18S (data not shown).

The enzymatic activities of CPT-I, MCAD and citrate synthase were decreased in the untreated HF group compared to control. Oxfenicine caused a complete recovery to control values of both MCAD and citrate synthase activities and in part of CPT-I activity (Table 2). The protein expression of MCAD was also reduced by 40% in the untreated HF group, but not in the HF+Oxf group (Table 2). To investigate whether this was due to altered protein expression of nuclear receptors regulating the expression of enzymes of FAO, the protein expression of PPARα and RXRα, as well as of the co-activator PGC-1, was determined. HF without treatment resulted in a decrease in RXRα protein expression, as previously reported by us [3], while PPARα and PGC-1 were not significantly different from control hearts (Table 2). Treatment with oxfenicine had no effect on the protein expression of RXRα, PPARα or PGC-1.

4. Discussion

Current medical therapies for HF can improve clinical symptoms and slow the progression of contractile dysfunction and expansion of LV chamber volume, nevertheless there is still progression, and the prognosis for even the optimally treated patients remains poor. Thus, there is a need for novel therapies for HF, independent of the neurohormonal axis, that can improve cardiac performance and prevent or reverse LV dysfunction and remodeling. In the present investigation, we demonstrate that an early and sustained treatment with the CPT-I inhibitor oxfenicine prevents LV chamber dilation, wall thinning and delays by almost 1 week the onset of decompensation in pacing-induced HF. While our results are limited to an experimental model of disease, they nevertheless support the concept that CPT-I inhibition might be effective for slowing the development and progression of clinical HF [1,6–8].

The chain of molecular events that led to delayed onset of decompensation in HF+Oxf is undoubtedly too complex to dissect using an in vivo model. We therefore focused the present study on the final effects of CPT-I inhibition on a group of myocardial molecular alterations selected from those that characterize HF. First, we tested whether the preserved LV wall thickness was associated with a reduced activation of MMPs, whose critical role in determining cardiac remodeling in HF has been well documented [11,18]. Intriguingly, the marked activation of MMP-2 and -9 found in untreated HF was completely prevented in HF+Oxf. It is difficult to establish whether the lack of MMP activation was the primary cause or rather the consequence of the preserved LV wall thickness. The generation of lipid peroxidation products, a second characteristic marker of myocardial injury [22], was also attenuated by oxfenicine treatment.

Consistent with previous reports [2,3], gene expression of PPARα, RXRα and of PPARα/RXRα-regulated enzymes was reduced in failing hearts. A novel finding of the present study is that CPT-I inhibition prevented the transcriptional down-regulation of PPARα and RXRα, as well as the HF-induced down-regulation of PPARα-activated genes, namely CPT-I, MCAD, PDK4 and UPC3. Consistent with gene expression, protein expression and activity of CPT-I and MCAD was also preserved by sustained oxfenicine administration. Although we could not directly measure PPARα activation in vivo, a possible explanation of these findings is that the inhibition of CPT-I results is accumulation of lipid ligands outside the mitochondria that activate the PPARα/RXRα and preserve the expression of mRNA expression of key FAO enzymes even in end-stage failure. The hypothesis that CPT-I inhibitors can indirectly stimulate PPARα has been formulated also by other authors [8]. These results raise the question of whether the beneficial effects of CPT-I inhibition can be attributed to the combination of PPARα activation and reduced fatty acid transport into the mitochondria. Additional work is needed before the role of pathophysiological PPARα activity and fatty acid oxidation in cardiac hypertrophy and HF are understood. Oxfenicine prevented also the decrease in gene expression and activity of the Krebs cycle enzyme citrate synthase. This effect was not related to changes in protein expression of PGC-1, a critical cofactor involved in mitochondrial biogenesis [4,23].

In the present study, we found that CPT-I inhibition delayed the onset of decompensation by almost 1 week, which corresponds to an approximate 25% increase in the duration of pacing. This is a remarkable improvement if one considers that the progression of failure becomes increasingly rapid during the fourth week, when the pacing rate is increased to 240 beats/min and the heart is subjected to extreme pacing-induced stress [9]. To our knowledge, few interventions have shown similar effects with respect to slowing the progression of pacing-induced HF [24]. The present findings are supported by animal studies in hypertrophic cardiomyopathy: 6-week administration of etomoxir, another CPT-I blocker, improved LV function and remodeling in rats subjected to aortic banding [7].

Several limitations of our study should be considered. First, cardiac free fatty acid oxidation was not determined. Second, we cannot exclude the possibility that part of the beneficial effects of oxfenicine was due to mechanisms independent of CPT-I blockade. To our knowledge, however, such mechanisms have not been reported in literature. Third, the effects of one-month administration of oxfenicine was not assessed in normal dogs, however toxicology studies in dogs found oxfenicine increased heart weight to body weight ratio only after 3–12 months at dose of 500–750 mg/kg/day, while doses in the range used in the present study did not cause cardiac pathology [14]. Third, the expression of mRNA for PPARα target genes (e.g., mCPT-1, MCAD, PDK4 and UCP3) does not necessarily reflect the activation state of the PPARα/RXRα heterodimer. Finally, one cannot rule out the possibility that more long-term CPI-I inhibition could have deleterious effects on the myocardium related to altered expression of enzymes involved in FFA and carbohydrate metabolism.

In conclusion, the results of the present investigation show that treatment with oxfenicine early in the development of HF prevented LV wall thinning and delayed the time to end-stage failure. Treatment with oxfenicine also increased the mRNA levels of genes encoding key enzymes of cardiac substrate metabolism, particularly those regulated by PPARα. These results, while limited to an experimental model of disease, suggest that CPT-I inhibition has very novel effects on the myocardium, and support the concept that this approach may potentially be an effective therapy for HF.

Acknowledgements

This study was supported by the National Heart, Lung and Blood Institute grants R01 HL-62573 (F.A. Recchia), and by P01 HL-74237 (W.C. Stanley and F.A. Recchia) and P01 HL-43023 (T.H. Hintze). It was also supported in part by the American Heart Association Texas Affiliate grant 0365028Y (M.E. Young). Dr. Linke received a post-doctoral fellowship (Li 946/1-1) from the German Research Foundation.

The authors thank Joseph P. Sterk, Tracy McElfresh and Hazel Huang for their assistance with the biochemical analysis, as well as Melissa A. Stavinoha for her assistance with the RNA analysis.

References

Author notes