Patients with mutations in the mitochondrial very-long-chain acyl-CoA dehydrogenase (VLCAD) gene are at risk for cardiomyopathy, myocardial dysfunction, ventricular tachycardia (VT), and sudden cardiac death. The mechanism is not known. Here we report a novel mechanism of VT in mice lacking VLCAD (VLCAD−/−). These mice exhibited polymorphic VT and increased incidence of VT after isoproterenol infusion. Polymorphic VT was induced in 10 out of 12 VLCAD−/− mice (83%) when isoproterenol was used. One out of 10 VLCAD−/− mice with polymorphic VT had VT with the typical bidirectional morphology. At the molecular level, VLCAD−/− cardiomyocytes showed increased levels of cardiac ryanodine receptor 2, phospholamban, and calsequestrin with increased [3H]ryanodine binding in heart microsomes. At the single cardiomyocyte level, VLCAD−/− cardiomyocytes showed significant increase in diastolic indo 1 and fura 2 fluorescence, with increased Ca2+ transient amplitude. These changes were associated with altered Ca2+ dynamics, to include: faster sarcomere contraction, larger time derivative of the upstroke, and shorter time-to-minimum sarcomere length compared with VLCAD+/+ control cells. The L-type Ca2+ current characteristics were not different under voltage-clamp conditions in the two VLCAD genotypes. Sarcoplasmic reticulum Ca2+ load measured as normalized integrated Na+/Ca2+ exchange current after rapid caffeine application was increased by 48% in VLCAD−/− cells. We conclude that intracellular Ca2+ handling represents a possible molecular mechanism of arrhythmias in mice and perhaps in VLCAD-deficient humans.
- inborn errors
- ryanodine receptor
- calcium ion
- ventricular tachycardia
deficiencies in mitochondrial fatty acid β-oxidation (FAO) enzymes have been implicated in cardiomyopathy, arrhythmias, sudden death, nonketotic hypoglycemia, heart and liver lipidosis, encephalopathy, and skeletal myopathy (21, 41, 46). Very-long-chain acyl-CoA dehydrogenase (VLCAD) deficiency is a FAO defect that usually presents with episodes of metabolic crisis and death in children (15) and causes ventricular tachycardia (VT) in almost 50% of the presenting cases (34). Most cases of sudden death occur in the first year of life (15). There has been little research on the mechanisms leading to cardiac dysfunction and ventricular arrhythmias in VLCAD deficiency. We developed a mouse model of VLCAD deficiency and characterized the cardiac phenotype in these mice, revealing lipid accumulation and mitochondrial proliferation in myocytes and inducible polymorphic VT, which is exacerbated by adrenergic stimulation (14).
We have hypothesized that altered intracellular Ca2+ homeostasis resulting from dysregulation of Ca2+-regulated proteins may be at the basis of the VT in VLCAD deficiency. Our hypothesis originated from partial cDNA microarray data in which we found that levels of several of the Ca2+-related proteins were quantitatively different in mouse hearts deficient in VLCAD at birth and 2 mo after birth. At a cellular level and in several metabolic pathways, Ca2+ is an important messenger that serves a vital role in cell signaling. In muscle, Ca2+ influx though sarcolemmal ion channels triggers the mobilization of Ca2+ from intracellular stores (Ca2+-induced Ca2+ release), thereby leading to the contraction of the heart (systole; see Refs. 6 and 7). Abnormalities of intracellular Ca2+ handling are hypothesized to induce diastolic depolarization that can trigger ventricular arrhythmias (29, 30, 32, 33). Abnormalities in Ca2+ handling are also found in patients with familial catecholaminergic polymorphic ventricular tachycardia (CPVT), a genetic arrhythmia syndrome characterized by polymorphic VT associated with adrenergic stimulation. Intracellular Ca2+ homeostasis is modulated by intracellular Ca2+ release channels (ryanodine receptors, RyRs), which are widely distributed in tissues: RyR1 in skeletal muscle, RyR2 in cardiac muscle, and RyR3 in a variety of tissues but in minuscule amounts. Mutations in the cardiac ryanodine receptor (RyR2) gene are known to cause both CPVT, also termed familial polymorphic ventricular tachycardia (FPVT) or bidirectional VT (28, 29, 39, 48), and arrhythmogenic right ventricular dysplasia (40, 49). These mutations cause a “gain-of-function” or increased Ca2+ leak from the sarcoplasmic reticulum (SR; see Ref. 33).
Given that VLCAD-deficient babies present with severe forms of ventricular dysfunction and arrhythmias (34), and that our mouse model of VLCAD deficiency has easily inducible ventricular arrhythmias, we investigated whether defective cardiomyocyte Ca2+ handling plays a role in the development of cardiomyopathy in the VLCAD-deficient mouse. In this study, we used our mouse model of VLCAD deficiency (VLCAD−/−; see Ref. 14) to show a novel role for mitochondrial FAO defects in the regulation of Ca2+ signaling in heart. We assessed functional changes of intracellular Ca2+ homeostasis and dynamics in hearts from VLCAD+/+ and VLCAD−/− mice. Our results provide new insights into Ca2+ handling in the settings of VLCAD deficiency in mice and perhaps in humans.
MATERIALS AND METHODS
All animals in this study were cared for according to the Institutional Animal Care and Use Committee at Vanderbilt University. Mice were generated and genotyped as previously described (13, 14). Mice of 2–3 mo of age were included in these studies. Intracardiac electrophysiological studies were performed in 12 VLCAD−/− and 16 VLCAD+/+ male mice. Subcellular fractions were performed in 24 VLCAD−/− mice (16 males, 8 females) and 12 VLCAD+/+ mice (8 males, 4 females). Upon death or at autopsy, whole heart ventricles and skeletal muscle (slow and fast twitch) were harvested for total protein analysis.
Intracardiac electrophysiology studies.
Mice were anesthetized with pentobarbital sodium (0.070 mg/g ip). A surface electrocardiogram (ECG; lead I) was obtained by placement of subcutaneous 27-gauge needles in each foreleg. ECG channels were amplified (0.1 mV/cm) and filtered between 0.05 and 400 Hz. Under an operating microscope, a cut down of the right internal jugular vein was performed, and an octapolar 2-Fr electrode catheter (CIBer cath; NuMED) was placed in the right atrium and right ventricle, guided by electrogram tracing to verify placement.
Bipolar electrogram recordings were obtained from the right atrium, right ventricle, and His positions. Signals were amplified and filtered between 40 and 400 Hz. Bipolar pacing was performed using a programmable stimulator (Medtronic 2356) modified by the manufacturer to deliver coupling intervals as short as 10 ms. Pacing threshold (in mA) was determined for each pacing site, and stimulation was performed for 1.0- to 2.0-ms pulse width at two times the diastolic capture threshold. Electrophysiological intervals (RR, PR, QRS, QT, AH, HV, and AV) were measured in standard fashion. Standard clinical electrophysiological pacing protocols were used to determine all basic electrophysiological parameters. The AV-His-Purkinje conduction properties were assessed through rapid atrial pacing at rates up to 1,000 beats/min. The Wenckebach cycle length, defined as the longest atrial paced cycle length that failed to conduct 1:1 to the ventricle, was determined. Programmed right atrial stimulation was performed to determine AV node effective refractory period, defined as the longest premature stimulus that failed to conduct to the ventricle. Ventricular effective refractory period was determined at three drive cycle lengths. Single, double, and triple extra stimuli were delivered at three drive cycle lengths to determine inducibility of VT. The duration and cycle length of induced tachycardias were recorded. After baseline measurements were completed, isoproterenol was administered (100 μg ip), and the protocols were repeated to assess the effects on conduction and refractoriness.
Preparation and characterization of subcellular fractions from mouse hearts.
Subcellular fractionation was performed by the method described by Chamberlain and Fleischer (10). Cardiac microsomal fractions enriched in SR were analyzed by SDS-PAGE (6% resolving gel), and the protein profile was visualized with Coomassie blue staining. For Western blot analysis of RyR2 after SDS-PAGE, proteins were transferred to an Immobilon-P membrane in blot transfer buffer (in mM: 48 Tris, 39 glycine, and 1.3 SDS, pH 9.2) using a Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell (Bio-Rad). The vacant binding sites on the membrane were blocked by incubating the membrane in wash buffer (10 mM Tris·Cl, pH 8.0, 0.55 M NaCl, and 0.05% Tween 20) containing 5% nonfat dry milk protein for 1 h. The Immobilon-P membrane was probed with a RyR2-specific antibody (1:1,000) in blocking buffer for 1 h, washed three times with wash buffer, and then incubated with secondary antibody goat anti-rabbit IgG conjugated to peroxidase (Sigma) in blocking buffer. The membrane was again washed three times with wash buffer, and immune complexes were detected by the enhanced chemiluminescence Western blotting analysis system (Amersham Pharmacia Biotech). RyR1, -2, and -3 specific antibodies were obtained from the Fleischer laboratory, Vanderbilt University (26). With the use of SDS-PAGE (10% resolving gels), Western blots were performed for VLCAD (antibody from Strauss laboratory Vanderbilt University; see Ref. 13), calsequestrin, SR Ca2+-ATPase (SERCA) 2a, and phospholamban (antibodies from Santa Cruz Biotechnology, Santa Cruz, CA). Quantification of the different blots was done by densitometry (NIH software image).
[3H]ryanodine binding assay.
[3H]ryanodine binding was determined at 60 nM ryanodine as described previously (26). Briefly, cardiac microsomes (50 μg) were incubated in 50 μl of buffer containing 10 mM potassium-HEPES, pH 7.4, 0.15 M KCl, 25 μM CaCl2, and 60 nM [3H]ryanodine [∼15,000 counts·min−1 (cpm)·pmol−1, obtained from Amersham] for 1 h at room temperature. Nonspecific binding was measured in the presence of 20 μM cold ryanodine (Sigma). Free ligand was separated from the bound ligand by sedimenting the microsomes in a Beckman TL-100.1 rotor at 95,000 revolutions/min for 15 min at 4°C. The supernatants were removed by aspiration, the pellets were rinsed two times and resuspended in 200 μl water, and the radioactivity was measured in 5 ml of Cytoscint (ICN, Cleveland, OH) in a Beckman LS 5000TD scintillation counter. Upon completion of the experiment, radioactivity was measured and expressed as picomoles per milligram of protein. Binding was calculated based on the formula (total cpm − nonspecific cpm)/(cpm/pmol × mg protein) (26).
Preparation of ventricular myocytes.
Mice were anesthetized by intraperitoneal injection of Avertin solution (5 mg Avertin/10 g body wt, T48402; Sigma-Aldrich) containing heparin (3 mg/10 ml, H9399; Sigma-Aldrich). The heart was rapidly excised and placed in ice-cold Ca2+-free and glucose-free HEPES-buffered Tyrode solution. The Tyrode solution contained (in mM) 140 Na+, 4.5 K+, 0.5 Mg2+, 150 Cl−, 0.4 H2PO, 10 HCO3−, and 10 HEPES. The pH of all solutions was adjusted to 7.4 using NaOH. The aorta was cannulated, and the heart was perfused with Tyrode solution at room temperature for 10 min to stop contractions. The perfusion was switched to Tyrode solution containing 10 μM Ca2+, 178 U/ml collagenase (CLS2; Worthington Biochemical), and 0.64 U/ml protease (P5147; Sigma-Aldrich) for 12 min at 37°C. Tissues from the atria and aorta were discarded. Remaining ventricular tissue was coarsely minced and placed in Tyrode solution containing the C-16 fatty acid energy substrate palmitate (10 μM, palmitic acid P0500; Sigma-Aldrich) and 0.5 mM Ca2+. Myocytes were dissociated by gentle agitation and used within 3 h after isolation.
Measurement of intracellular Ca2+ transients.
Dissociated cardiomyocytes were loaded for 15 min with the acetoxymethyl ester forms of the Ca2+-sensitive fluorescence indicators indo 1 (20 μM I1203; Invitrogen) or fura 2 (5 μM, F1221; Invitrogen). Cells were then centrifuged for 10 min at 27 g and resuspended in Tyrode solution containing 1.0 mM Ca2+. Ca2+ transients were measured in Tyrode solution containing 1.0 mM Ca2+ and 10 μM palmitate at 37°C, supplemented with an insulin-transferrin-selenium mixture (100× medium supplement, 41400, consisting of 10 μg/ml insulin, 5.5 μg/ml transferrin, 6.7 ng/ml sodium selenite; GIBCO-Invitrogen) to improve the performance of the isolated cells at the high stimulation frequencies (42). Myocytes were field stimulated at 600 and 250 beats/min via two parallel platinum wires. Fluorescence measurements were carried out using an inverted microscope (Axiovert 200; Carl Zeiss) equipped with a ×63, 1.4 NA oil immersion lens (Plan Apochromat; Carl Zeiss). For the fura 1 measurements, the excitation wavelength was rapidly switched between 340 ± 10 and 380 ± 10 nm at a rate of ∼200 ratios/s by using a monochromator (Optoscan; Cairn Research). For the indo 1 measurements, the excitation wavelength was fixed at 365 ± 15 nm. The excitation light was reflected on the cells by a dichroic mirror (415DCLP for fura 1 and 390DRLP for indo 1; Omega Optical). Fura 2 fluorescence emission was recorded at a wavelength of 510 nm.
With the use of a bandpass emission filter (510WB40; Omega Optical), indo 1 fluorescence was split into two components by a second dichroic mirror (450DCLP; Omega Optical), and each component passed through a 495-nm (495DF20; Omega Optical) or a 405-nm bandpass filter (405DF43; Omega Optical). The light was collected by 1-mm-diameter optical fibers mounted behind the emission filters imaging a ∼16-μm-diameter spot on the cell. Fluorescence emission was registered by photomultiplier modules (H6780; Hamamatsu) and amplified by a custom-built low-noise direct current-coupled amplifier. The signals were digitized at a sampling rate of 20 kHz by an analog/digital (A/D) converter board (PCI-6071E; National Instruments) in a conventional personal computer (PC).
Fluorescence emission ratios were converted to intracellular Ca2+ concentrations ([Ca2+]i) according to the equation (20) [Ca2+]i = Kd × β[R − Rmin]/(Rmax − R), using published in vivo dissociation constants for fura 1 and indo 1 of 371 (23) and 844 (3) nM, respectively. R values in the previous equation were calculated as ratios of 405/495 nm fluorescence emission (indo 1) and 340/380 nm excitation at 510 nm emission (fura 2). β Values were calculated as ratios of Ca2+-free/saturated indicator at excitation/emission wavelengths of 365/495 nm (indo 1) and 380/510 nm (fura 2). Rmin and Rmax were determined in vivo. Cells were loaded with Ca2+ dye and exposed to 10 mM caffeine two times to empty the SR. For Rmin determination, cells were superfused with Ca2+-free Tyrode solution containing 5 mM EGTA and 10 μM of the nonfluorescent ionophore bromo-A-23187 (B7272; Sigma-Aldrich). Measurements were taken after the fluorescence reached stable values at both wavelengths. For Rmax determination, extracellular Ca2+ concentration was gradually increased to 10 mM in the presence of the metabolic inhibitor cyanide p-(trifluoromethoxy)-phenylhydrazone (3 μM, C2920; Sigma-Aldrich) to avoid hypercontracture. The Ca2+ ionophore was present throughout the calibration procedure.
Sarcomere contraction measurements.
L-type Ca2+ current (ICa) was recorded in the whole cell mode voltage-clamp configuration according to previously published methods (22). Briefly, pipettes (2–3 MΩ) contained (in mM) 120 Cs+, 3 Ca2+, 126 Cl−, 1 MgATP, 1 NaGTP, 5 phosphocreatine, 10 HEPES, and 10 EGTA (1). The pH was adjusted to 7.2 with CsOH. The bath solution contained Tyrode solution as described above with 10 μM palmitic acid and 1.8 mM Ca2+. The holding potential was −90 mV. ICa was measured during 500-ms test pulses at potentials ranging from −40 to +40 mV following a 50-ms pulse at −50 mV to inactivate Na+ current (12, 31).
SR Ca2+ content.
SR Ca2+ content was estimated by integrating current through the Na+/Ca2+ exchanger (NCX) as described by Diaz et al. (11). Briefly, Na+ and K+ currents were eliminated in the bath solution by adding Cs+ and tetraethylammonium chloride. Pipettes contained (in mM) 120 Cs+, 10 HEPES, 10 TEA, 5 phosphocreatine, 1 MgATP, and 1 NaGTP; pH was adjusted to 7.2 with CsOH. The extracellular bath solution contained (in mM) 137 N-methyl-d-glucamine (NMDG), 25 Cs+, 10 HEPES, 10 glucose, 1.8 Ca2+, and 0.5 Mg2+; pH was adjusted to 7.4 with HCl. Cells were rapidly perfused with a spritz of modified bath solution containing 20 mM caffeine and equimolar substitution of NaCl for NMDG. The resulting inward current was integrated using PCLAMP 9.2 (Molecular Devices) and normalized for total membrane capacitance vs. cell surface area.
Data were expressed as means ± SE. Statistical analysis for the Western blots was performed using the Mann-Whitney Test. Data analysis for the Ca2+ fluorescence was accomplished with Matlab (R14; The MathWorks). Ca2+ transient upstroke and decay were fit to mathematical functions for calculating numerical time derivatives (51). Because derivatives of high-bandwidth fluorescence recordings are extremely susceptible to noise, we used these empirical models to obtain smooth derivatives of the signal. Ca2+ transient upstroke was best fit by [Ca2+]i(t) = [Ca2+]max × [P(t)/1 − P(t)]n, where t is time, [Ca2+]max is the amplitude of Ca2+ transient, with P(t) = 0.5(1 − e−t/τ)m. The parameters τ, n, and m were used as fit parameters and were not further analyzed. Transient decay was best fit by a sum of two exponentials, [Ca2+]i = Ae−t/τ1 + Be−t/τ2 with fit parameters A, B, τ1, and τ2 that were not further analyzed. The null hypothesis was rejected for P < 0.05.
Polymorphic and bidirectional VT in the VLCAD−/− mice.
VLCAD−/− mice, although viable, demonstrated easily inducible VT in the absence of physiological stress. There was no difference in heart rate, PR interval, Wenckebach cycle length (defined as the longest atrial paced cycle length with failure of 1:1 conduction to the ventricle), or ventricular effective refractory period (defined as the longest coupling interval that fails to capture the ventricle) between VLCAD+/+ and VLCAD−/− mice. With the use of programmed ventricular stimulation, VT could be induced in 6/12 (50%) of VLCAD−/− mice compared with 2/16 (12%) wild-type mice. VT in the VLCAD−/− mice was consistently polymorphic. VT inducibility was increased in the VLCAD−/− mice to 10/12 (83%) when isoproterenol was used. VT with the typical bidirectional morphology with isoproterenol infusion was observed in 1 out of 10 mice (Fig. 1). Isoproterenol did not increase arrhythmia inducibility in wild-type mice.
Increased expression of RyR2 isoform in mouse hearts and augmented ryanodine binding in microsomes of VLCAD−/− mice.
Given our microarray results suggesting changes in Ca2+-related genes and other reports showing a possible link between polymorphic/bidirectional VT and excessive RyR2 activity (32, 33, 39), we first tested whether RyRs were differently expressed in the VLCAD−/− mice. Western blot analyses using isoform-specific antibodies showed a 3.04 ± 0.8-fold increase in RyR2 expression in VLCAD−/− mice compared with VLCAD +/+ hearts (Fig. 2, A–C). We also found that in VLCAD−/− mice RyR1 levels were increased 3.3 ± 0.1-fold in skeletal muscle, and RyR3 levels were increased 2.4 ± 0.8-fold in brain homogenates (data not shown). We subsequently performed [3H]ryanodine binding assays with purified microsomal fractions from the VLCAD+/+ and VLCAD−/− mice. This was done because [3H]ryanodine has been demonstrated to bind with high affinity to the open conformational state of the RyR. [3H]ryanodine has been used as an index of RyR channel gating (30, 35, 50). Ryanodine binding was 1.7 ± 0.3 pmol/mg protein in the VLCAD+/+ microsomes and 3.5 ± 0.2 pmol/mg protein in the VLCAD−/− microsomes (Fig. 2D).
Enhanced diastolic Ca2+ concentration, Ca2+ transients, and SR Ca2+ load in VLCAD−/− cardiomyocytes.
We tested whether the observed RyR2 upregulation and increased [3H]ryanodine binding were associated with enhanced Ca2+ transients in intact cardiac myocytes from VLCAD−/− hearts. We isolated ventricular cardiac myocytes from VLCAD+/+ and VLCAD−/− mice and measured Ca2+ transients using Ca2+-sensitive ratiometric dyes. We found that both diastolic fluorescence and amplitude of Ca2+ transient were increased in VLCAD−/− cardiac myocytes as shown in Fig. 3. We confirmed our results in two series of independent experiments using two different fluorescence dyes with different Ca2+-binding affinities, and at two different pacing rates. Because the physiological heart rate in a mouse is on the order of 600–700 beats/min, we initially measured Ca2+ transients at 10 Hz (Fig. 3A). Under these conditions, Ca2+ transients were significantly larger in VLCAD−/− cardiac myocytes compared with VLCAD+/+ control cells. However, rapid pacing itself increased diastolic Ca2+ concentration in all cells (Fig 3C). The increased binding of Ca2+ to intracellular ligands that occurs when diastolic Ca2+ concentration is high may by itself increase free Ca2+ concentration and the amplitude of the Ca2+ transient. Therefore, we performed measurements at 4 Hz to achieve a lower baseline Ca2+ concentration and to reduce mitochondrial Ca2+ loading. Although diastolic fluorescence was significantly lower in VLCAD+/+ control myocytes at 4 Hz, i.e., 159 ± 40 nM (Fig. 3B), it was still threefold higher in VLCAD−/− cardiac myocytes (Fig. 3, B and C). These data were consistent with the fluorescence data at 10 Hz. The amplitude of Ca2+ transient was significantly increased in VLCAD−/− cells at 4 Hz (Fig. 3, B and D), and this increase was even more pronounced at the lower pacing rate (Fig. 3D).
To further characterize Ca2+ transients in VLCAD−/− cardiac myocytes, we subsequently calculated the maximum time derivatives of Ca2+ transient upstroke and decay. We found that maximum slopes of the transients were significantly increased in VLCAD−/− myocytes (Fig. 4A), although times of fluorescence rise and fall were about the same (Fig. 4B). One of our hypotheses was that the increase in Ca2+ transient amplitude was the result of an increase in SR Ca2+ content. We tested this hypothesis by measuring the integrated NCX current after fast application of caffeine (Fig. 4C). The integrated normalized NCX current was 3.38 ± 0.69 pC/pF in the VLCAD−/− cardiomyocytes compared with 2.28 ± 0.55 pC/pF in wild-type control cells. This result indicated an increased SR Ca2+ load in VLCAD−/− cardiomyocytes compared with WT control cells.
Enhanced sarcomere contraction velocity in VLCAD−/− cardiomyocytes.
We also measured sarcomere contraction and relaxation to test whether changes in time course and amplitude of Ca2+ transients had any impact on cardiomyocyte contractility. We found that sarcomere contraction was significantly hastened in VLCAD−/− cardiomyocytes compared with VLCAD+/+ control cells (Fig. 5A). Duration of contraction transient (as measured at 50% of the maximum contraction) was shortened by ∼30% in VLCAD−/− cells (Fig. 5B), which was consistent with an increase in contraction velocity. Maximum contraction velocity was increased by 28% (Fig. 5D), and time-to-maximum was decreased by 25% (Fig. 5E) in VLCAD−/− cardiomyocytes. There was, however, no difference in sarcomere relaxation (Fig. 5, D and E). Interestingly, amplitude of sarcomere contraction was the same in VLCAD−/− (85 ± 15 nm) and VLCAD+/+ (96 ± 10 nm) cardiac myocytes (Fig. 5C). There was no difference in diastolic sarcomere lengths in VLCAD−/− (1.749 ± 0.003 μm) compared with VLCAD+/+ (1.744 ± 0.007 μm) cells (Fig. 5A).
Altered Ca2+ release in VLCAD−/− mice was not the result of changes in the function of the L-type Ca2+ channel.
We tested whether the observed increase in Ca2+ transient amplitude could be the result of an increase in transmembrane L-type ICa. An increase in Ca2+ transient amplitude could have been the result of an increase in Ca2+ influx through sarcolemmal ion channels. Ca2+ release from the SR is solely graded via Ca2+ influx through the L-type Ca2+ channel, given that contribution from NCX reverse mode is believed to be small (∼7%) in mouse and rat ventricles (5, 6). We found that both amplitude and time course of ICa were not changed under voltage-clamp conditions in VLCAD−/− cardiac myocytes using a test potential of 0 mV (Fig. 6A). The slow and the fast time constants of ICa inactivation were the same, with 92 ± 6 and 12.0 ± 3 ms in VLCAD−/− and 92 ± 5 and 12 ± 3 ms in VLCAD+/+ cardiomyocytes as indicated by the recovery period of the current traces in Fig. 6A. We also measured the current-voltage relationship of mean peak ICa and found no differences between VLCAD−/− and VLCAD+/+ cells (Fig. 6B), supporting that the increase in Ca2+ transient amplitude in VLCAD−/− cells was not because of an increase in transmembrane ICa.
Ablation of the VLCAD gene in mice leads to alteration in other Ca2+-related proteins in heart.
Given the changes in Ca2+ homeostasis and the observed increase in amplitude of Ca2+ transient in VLCAD−/− cardiac myocytes, we tested whether the expression of other Ca2+-related proteins was also changed in our model. Western blot analyses showed a sevenfold increase in calsequestrin levels (Fig. 7, A and B). Protein levels of SERCA2a were not changed (Fig. 7, A and D). The monomeric form of phospholamban was unchanged although the pentameric form of phospholamban was increased 6.0 ± 0.9-fold in the VLCAD−/− cells compared with control (Fig. 7D).
VLCAD deficiency in mice leads to polymorphic and bidirectional VT.
Children with mitochondrial VLCAD deficiency present with cardiomyopathy, ventricular arrhythmias, and sudden unexpected death (34, 46). Deficiencies in several of the mitochondrial FAO enzymes are also associated with arrhythmias and sudden unexpected death in humans, but the mechanism is unclear. In this article, we demonstrated that VLCAD deficiency in mice leads to catecholamine-sensitive polymorphic VT and in one case to bidirectional VT. Our results show that VLCAD deficiency in mice leads to alterations in intracellular Ca2+ homeostasis, which might mimic electrophysiological aspects of RyR2 and/or calsequestrin mutations in humans (17). RyR2 and/ or calsequestrin mutations in humans lead to FPVT/CPVT, although many patients with FPVT/CPVT do not have identifiable mutations in RyR2 or calsequestrin (38). A mutation in ankyrin B was found in a patient with a phenotype similar to catecholaminergic VT (37). These observations suggest that genes other than RyR2 and calsequestrin are implicated in the genesis of these arrhythmias. Our results in VLCAD−/− mice suggest that deficiencies in fatty acid metabolic enzymes could provide an alternative mechanism for FPVT/CPVT.
VLCAD deficiency in mice leads to an increase in diastolic and systolic Ca2+ transient amplitude and SR Ca2+ load in isolated cardiomyocytes.
One major finding in this paper is that systolic and diastolic Ca2+ levels were significantly increased in VLCAD−/− compared with VLCAD+/+ cardiomyocytes. Field stimulation at the physiological rate of 10 Hz led to diastolic Ca2+ concentration of >400 nM in indo 1-loaded wild-type control cells. Under the same pacing conditions, diastolic [Ca2+]i was further increased to ∼700 nM in the VLCAD−/− cells (Fig. 3C). These high diastolic Ca2+ levels were most likely from SR inability to sequester sufficient Ca2+ at such high stimulation rate as suggested by Field et al. (16). High diastolic Ca2+ levels also alter cytosolic Ca2+ buffering efficiency, thereby complicating the interpretation of Ca2+ transient amplitudes (7). Approximately 99% of the total Ca2+ that enters the cytosol during a Ca2+ transient is rapidly complexed with a variety of Ca2+ buffers (4–6). These buffers are Ca2+-binding sites that are unoccupied in diastole, but they bind part of the Ca2+ released during systole, thus attenuating the rise in [Ca2+]i. Conversely, in the setting of our high diastolic [Ca2+]i, there would be fewer binding sites available to buffer the rise in [Ca2+]i following a release. As a consequence, we tested whether the increase in transient amplitude was the result of less systolic Ca2+ buffering in the VLCAD−/− cardiomyocytes. We estimated the increase in free Ca2+ concentration (i.e., the increase in Ca2+ fluorescence) in VLCAD−/− that could be expected from the increase in diastolic Ca2+ concentration alone and compared it with our measured values. Based on the findings of Berlin et al. (4), we used the traditional Michaelis-Menten relationship to estimate total [Ca2+]i using a lumped Kd of 0.63 μM for all cytosolic Ca2+ buffers and a total buffer capacity of the cytosol equal to 162 μM. Under these assumptions, we found that the increase in diastolic [Ca2+]i in VLCAD−/− cardiac myocytes could account for nearly 80% of the increase in Ca2+ transients in these cells when stimulated at 10 Hz (Fig. 3A).
The interpretation of our high baseline Ca2+ fluorescence levels in the VLCAD−/− cardiomyocytes needs to also take into account the contribution of mitochondrial Ca2+ loading. Mitochondrial Ca2+ has been shown to be significant when resting Ca2+ concentration exceeds 300–500 nM, as demonstrated by Zhou et al. (53). Mitochondrial dye loading cannot be avoided using the ester form of indo 1. This is why we subsequently performed Ca2+ fluorescence measurements at a lower stimulation frequency of 4 Hz using fura 2 (Fig. 3B). The ester form of fura 2 accumulates less in intracellular compartments (9). The lower pacing rate using fura 2 significantly reduced diastolic Ca2+ concentration in wild-type control cardiac myocytes to a value of <200 nM (Fig. 3, B and C), which compares well with literature data (8, 18). However, under the slow pacing condition, amplitude of Ca2+ transients was higher in all myocytes, which may be a result of an increased uptake of Ca2+ in the SR (negative staircase effect; see Refs. 16 and 47).
More importantly, our fura 2 experiments continued to show a marked increase in baseline fluorescence in VLCAD−/− cardiac myocytes, which translated to a more than threefold increase in diastolic Ca2+ concentration compared with VLCAD+/+ control cells (Fig. 3, B and C). Because of reduced diastolic Ca2+ concentrations at the lower pacing rate, Ca2+ buffering in VLCAD−/− myocytes is expected to be higher at the slow compared with the fast pacing rate because more intracellular Ca2+ buffer binding sides should be available when diastolic Ca2+ concentration is lower. We estimated that, at the slower pacing rate of 4 Hz, the change in Ca2+ buffering may account for ∼60% of the increase in transient amplitude. The observed increase in transient amplitude was even higher at the lower pacing rate. Likely, at these lower diastolic Ca2+ levels, the effect of VLCAD deficiency on the amplitude of Ca2+ transient was more pronounced because of the larger Ca2+ concentration gradient between the SR and the cytosol (negative staircase). Interestingly, the normalized integrated NCX current was increased by 48% after rapid caffeine application in VLCAD−/− cells (Fig. 4C). The increased caffeine-induced Ca2+ release together with the increased expression of SR Ca2+-binding protein calsequestrin (Fig. 7, A and B) indicate that the SR Ca2+ content was increased in VLCAD−/− cardiac myocytes.
VLCAD deficiency in mice accelerates sarcomere shortening.
Biochemical and physiological changes in Ca2+ homeostasis led to changes in sarcomere shortening velocity in the VLCAD−/− cardiomyocytes. Sarcomere contraction was significantly hastened in VLCAD−/− myocytes compared with control cells (Fig. 5, A and D), which was consistent with an enhanced Ca2+ transient in these cells (Figs. 3 and 4). Faster contraction was evident in a shorter contraction transient (Fig. 5B), a larger time derivative of the upstroke (Fig. 5D), and a shorter time-to-minimum sarcomere length (Fig. 5E). Surprisingly, the mean contraction amplitude was not changed in the VLCAD−/− cardiomyocytes (Fig. 5C; P = 0.53). The observed increase in diastolic Ca2+ in VLCAD−/− cardiac myocytes did not translate into a decreased diastolic sarcomere length (Fig. 5A). It might be speculated that myofilament response to cytosolic Ca2+ was altered in VLCAD−/− cardiac myocytes, but a direct measurement of myofilament Ca2+ sensitivity in the VLCAD cardiomyocytes would be required to test this hypothesis.
VLCAD deficiency alters expression of Ca2+ regulatory proteins.
We have previously shown that VLCAD−/− mice have a number of biochemical changes suggestive of complex alteration in lipid metabolism and lipid transport that are present in the heart at birth, whereas ultrastructural abnormalities develop postnatally (14). Although these changes may be part of the compensatory molecular events that occur in the absence of VLCAD, it is also conceivable that the biochemical changes that we presently report are directly related to the disease phenotype. Here, we report changes in the expression of several Ca2+ regulatory proteins in heart. These findings have not been previously reported in VLCAD deficiency or in other models of mitochondrial FAO defects. We found that all RyR isoforms (i.e., RyR1 in skeletal muscle, RyR2 in heart, and RyR3 in brain) were upregulated. [3H]ryanodine binding was also increased in VLCAD−/− microsomes compared with microsomes from control mouse hearts. In our model, we also found that the pentameric form of phospholamban was upregulated in VLCAD−/− myocytes (Fig. 7D). Although several studies show that phospholamban asserts its inhibitory function by binding to SERCA in its monomeric form (2, 24, 25) there is some evidence that the pentameric form of phospholamban may directly interact with SERCA2a, causing a decrease in the affinity of the pump for Ca2+ (45, 52). However, the mechanism of this interaction is not fully understood. Pentameric phospholamban has been proposed to form an ion pore in SR membranes (1a, 27, 43, 44) that is selectively permeable to Ca2+ with spontaneous opening and closing properties (27). The exact role of RyR upregulation and elevated levels of calsequestrin and other Ca2+-related proteins in VLCAD deficiency is not entirely clear. Future studies are needed to assess their biological and clinical implications for other models of mitochondrial fatty acid oxidation defects in mice and perhaps in humans.
Altogether, the present work supports the concept of a previously unrecognized connection between a clinically important metabolic defect, upregulation of SR Ca2+ stores and release proteins, enhanced sarcomere contraction, and increased susceptibility to adrenergically mediated arrhythmias. Our current hypothesis is that changes in the regulation of [Ca2+]i are likely directly related to the cardiac phenotype of VLCAD deficiency and may serve as an added substrate for VT in the VLCAD−/− mouse. Enhanced SR Ca2+ load may increase spontaneous Ca2+ release, which is a hypothesized molecular mechanism of polymorphic and bidirectional VT. Whether changes in RyR, phospholamban, or calsequestrin were true contributors to the observed ventricular arrhythmias is not known. Our findings, however, support a mechanism in which VLCAD deficiency, perhaps through decreased ATP production, leads to alteration in intracellular Ca2+ homeostasis and elevated SR Ca2+ load. These changes have the potential to alter excitation-contraction coupling, leading to cardiac rhythm abnormalities. In the future, we will need to assess whether there are inherent or preexisting changes in RyR2, phospholamban, and/or calsequestrin that lead to abnormal Ca2+ handling or whether the accumulation of toxic metabolites resulting from VLCAD deficiency causes the observed changes in Ca2+ homeostasis. Given that these Ca2+ changes were consistently observed in the VLCAD−/− mice, abnormalities in intracellular Ca2+ handling may represent a plausible cellular mechanism for ventricular arrhythmia in VLCAD deficiency and/or several of the FAO enzyme defects.
There are several limitations to our study. Ca2+ fluorescence amplitude may differ from SR Ca2+ release because sarcolemmal ion currents were not blocked in our experiments. Although in mouse ventricles most of the Ca2+ that enters the cytosol during systole is released from the SR (∼92%, compared with ∼70% in humans), other ion currents (such as the reverse-mode NCX) might have partly contributed to the increase in systolic free Ca2+ concentration in VLCAD−/− myocytes (5, 6). These currents were not individually measured. Furthermore, mitochondrial Ca2+ loading can be expected to occur in VLCAD−/− myocytes even when diastolic Ca2+ concentration is low (19, 53). Because both indo 1 and fura 2 are known to accumulate in noncytoplasmic compartments (9), fluorescence originating from Ca2+ trapped in the mitochondria may superimpose the cytosolic component of Ca2+ transient. Therefore, it is important to note that quantitative changes in maximum time derivatives of total Ca2+ fluorescence are only suggestive of changes in SR Ca2+ release and uptake fluxes. Additional experiments would be required to determine the exact contribution of Ca2+ fluxes resulting from noncytoplasmic compartments, the sarcolemma, and the mitochondria. Notwithstanding these limitations, our findings remain true: Ca2+ homeostasis is altered in the VLCAD−/− mouse heart, which is a possible underlying mechanism for polymorphic VT in mice and perhaps in humans.
In summary, in this research report, we have provided evidence that VLCAD deficiency in mice is associated with polymorphic and bidirectional VT, elevated levels of critical Ca2+ regulatory proteins with elevated diastolic and systolic Ca2+, and elevated SR Ca2+ stores. Findings in this mouse model raise the following critical questions. 1) Does VLCAD deficiency in humans contribute to a subset of cardiac arrhythmias that resembles defects in Ca2+ release channels of the heart? 2) Do our findings represent a molecular mechanism for arrhythmias in other models of FAO deficiency? Long-term follow-up of patients with VLCAD deficiency has been limited by the fact that most of the cases present in childhood with sudden death as presenting sign. Patients who survive into adulthood may have no symptoms and structurally normal hearts (without cardiac hypertrophy or dilation). Our data, however, raise the possibility that these patients may still be at risk for VT specifically related to changes in Ca2+ handling in the absence of cardiac hypertrophy or dilation. Our results also suggest that VLCAD deficiency in humans may contribute to a subset of cases with FPVT/CPVT and point to a molecular mechanism for arrhythmias in other cases of FAO deficiency.
This work was supported by the Robert Wood Johnson Foundation, the Vanderbilt Physician-Scientist Program, and a Grant in AID from the American Heart Association-South East Affiliate (V. J. Exil), the Vanderbilt Institute for Integrative Biosystems Research and Education (A. A. Werdich and F. Baudenbacher), and the National Heart, Lung, and Blood Institute (HL-070250, HL-62494, and HL-046681). M. E. Anderson is an Established Investigator of the American Heart Association.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society