Adaptive Mechanisms That Preserve Cardiac Function in Mice Without Myoglobin

Homozygous myoglobin-deficient (Mb^–/–) mice were generated using gene disruption technology as previously described.⁸ A complete absence of myoglobin mRNA and protein was previously documented.⁸ Offspring of heterozygote intercrosses (Mb^+/– × Mb^+/–) from timed pregnancies were harvested at 3 periods during embryogenesis (gestational days 8 to 9 [E8 to E9], E9.5 to E10.5, and E11 to E12.5) and genotyped by Southern blot assays. Viable progeny from the heterozygote intercrosses were genotyped and analyzed as described below.

All embryos and embryonic tissues were staged, harvested, fixed in 4% paraformaldehyde overnight, either paraffin-processed or quick-frozen in liquid nitrogen, and cryosectioned as previously described.⁹ The staging of embryos was performed by counting the presence of the vaginal plug as day 0.5 after conception and by the number of somites. Tissue for genotyping the embryos was obtained from the yolk sac. Immunohistochemistry was performed as previously described.^{2 9} The whole-mount immunohistochemical staining protocol included overnight fixation of the embryonic hearts in 4% paraformaldehyde and overnight incubation with both the primary (rabbit anti-CD31 or platelet-endothelial cell adhesion molecule 1 [PECAM-1]; 1:50, Pharmingen) and secondary antiserum (horseradish peroxidase–conjugated goat anti-rat serum; 1:100, Vector Labs). The embryonic hearts were then incubated with dimethylaminoazobenzene/NiCl₂ solution, postfixed overnight in 4% paraformaldehyde/0.1% glutaraldehyde, and photographed using an Olympus SZH stereomicroscope.

In addition to whole-mount immunohistochemical staining with PECAM-1, we used a second technique to identify the endothelial cells of blood vessels in the developing and adult wild-type and mutant hearts. Embryonic (E12.5 to E16.5) or adult male Mb^+/+ and Mb^–/– hearts were isolated, immersion-fixed overnight in methyl-Carnoy’s fixative, embedded in paraffin, sectioned, and stained using the biotinylated lectin Bandiera simplicifolia lectin B4 (10 μg/mL; Vector labs).¹⁰ Substrate was developed with dimethylaminoazobenzene, and slides were either cover-slipped or lightly counterstained with hematoxylin and analyzed microscopically (magnification of 20× to 100×). For each embryo (n=3 for each genotype), the entire section was analyzed at 3 separate levels. For each adult animal, 30 fields were examined at 3 separate levels (n=3 for each genotype). The vasculature was quantified by 2 blinded investigators. Statistical analysis was performed with a Student’s paired t test.

Total RNA was isolated from individually dissected embryonic or adult hearts or whole embryos using the Tripure isolation kit (Boehringer Mannheim). Four micrograms of total RNA were used in each reverse transcription (RT) reaction (Retro-script, Ambion). Complementary DNA (2 μL) was then used as a template for the polymerase chain reaction (PCR) in a 20-μL reaction volume including 100 ng of each primer, 2 mmol/L MgCl₂, Taq buffer, and 1 U of Taq polymerase (GIBCO/BRL). Fifteen microliters of each PCR reaction were loaded on a 2% agarose gel, as previously described.⁹ For myoglobin, a single forward primer was positioned to include exon 1, and amplification by PCR (25 cycles) was done using reverse primers complementary to sequences from exon 2. Semiquantitative RT-PCR using RNA isolated from individually dissected embryonic (E12.5) or adult (3-month-old) hearts was performed as previously described.⁹ For each primer pair, PCR conditions were used to establish linearity of the output signal relative to the input template DNA.⁹ PCR primer pairs used for this study are included in the online data supplement (located at http://www.circresaha.org).

Stained sections were examined with a Leitz Laborlux-S microscope equipped with an Optronics VI-470 CCD camera and Scion Image 1.62 analysis software. Image processing was completed using Adobe Photoshop 5.0 and printed using a Kodak XLS 8600 PS printer.

Mice (25 to 30 g) were euthanized, and their thoracic contents were excised and placed in a 4°C perfusate. The heart and aorta were dissected free and perfused in the Langendorff mode at 37.5°C with a (retrograde) perfusion pressure of 100 cm H₂O. The preparation was then converted to a working mode with a left atrial pressure of 15 cm H₂O and an anterograde perfusion pressure of 80 cm H₂O^{11 12} (online data supplement).

Hearts were perfused in the working mode as described above. After stabilization with a standard Krebs-Henseleit buffer, the buffer was changed to the following mixture of physiological substrates: 3-¹³C lactate (1.2 mmol/L), 1,3-¹³C acetoacetic acid (0.17 mmol/L), 3-¹³C pyruvate (0.12 mmol/L), BSA (0.75%), U-¹³C labeled fatty acids (0.5 mmol/L; consisting of 50% palmitic acid, 23% oleic acid, 15% linoleic acid, 11% palmitoleic acid, and 1% stearic acid), and unlabeled glucose (8 mmol/L).¹³ After 30 minutes of perfusion, hearts were freeze-clamped and extracted in perchloric acid, neutralized with KOH, reconstituted in D₂O, and analyzed by nuclear magnetic resonance (NMR) spectroscopy.¹⁴ Proton-decoupled ¹³C NMR spectra were obtained at 150.87 MHz using a Varian Inova spectrometer with a broadband probe, a 45-degree observe pulse, an interpulse delay of 1.5 s, 43 488 data points, and 4000 scans.

Ventricular hypertrophy was induced in age- and sex-matched adult Mb^+/+ and Mb^–/– mice by continuous administration of isoproterenol (0.028 g/mL at a rate of 1.0 μL/h), as described previously.^{15 16} After 7 days of continuous infusion, the mice were weighed and euthanized. The hearts were excised, weighed, and either fixed in 4% paraformaldehyde or frozen for isolation of total RNA.² Cardiac hypertrophy was determined by the comparison of heart weight to body weight, histological analysis, and gene expression.

After continuous anesthesia with 2.0% isoflurane, age- and sex-matched adult Mb^+/+ and Mb^–/– mice underwent a left thoracotomy/pericardiotomy to expose the left ventricle, and the left anterior descending coronary artery was ligated with 8-0 prolene sutures. The residual pneumothorax was evacuated to restore negative pressure, and buprenorphine (0.05 mg/kg) was administered for postoperative pain control. Two months after ischemic injury, transthoracic echocardiography was performed under isoflurane anesthesia. The mice were weighed and euthanized, and their hearts were excised, weighed, fixed in 4% paraformaldehyde, and processed for histological analysis.

An expanded Materials and Methods section can be found in an online data supplement available at http://www.circresaha.org.

Analysis of a large group of embryos from 16 male and 16 female heterozygote intercrosses at sequential stages of embryonic development demonstrated a Mendelian distribution of genotypes (1:2:1 ratio) only in embryos at E9.0 days or younger (online data supplement). Between days E9.5 and E10.5, the majority of embryos homozygous for the mutated myoglobin allele (Mb^–/–) exhibited a number of severe defects. These included congestive heart failure manifested by prominent pericardial effusion and vascular insufficiency manifested by diffuse hemorrhages, a generalized developmental delay, and reduced size (Figure 1). Similar defects were observed more rarely in animals bearing a single copy of the myoglobin-null allele (Mb^+/–). Of the myoglobin-null embryos that developed beyond this stage (E11.0 days or older), there was no evidence of further fetal loss or structural abnormalities (online data supplement). These findings were noted in 2 strains of mice (C57BL/6 and Sv 129).

To further define the morphological defects associated with reduced expression of myoglobin, histological sections of Mb^+/+, Mb^+/– (data not shown), and Mb^–/– embryos were compared. Histological analysis of runted mutant embryos revealed evidence of myocardial thinning, congestive heart failure, and vascular insufficiency with peripheral hemorrhage (Figure 1). Other than a generalized developmental delay, morphological defects in Mb^–/– embryos were limited to the cardiovascular system, and other major developmental events were not noted until the time of death. Among the 12 failing Mb^–/– embryos at E9.5 to E10.5, 4 had a severe pericardial effusion consistent with congestive heart failure and 8 had extensive hemorrhages. Ultrastructural analysis of phenotypically abnormal but viable Mb^–/– hearts at E10.5 revealed evidence of edema in cells of both the endocardial and myocardial layers, increased cytoplasmic vacuoles, and a less well-developed sarcomeric myosin actin contractile apparatus compared with normal wild-type littermates (data not shown). The TdT-mediated dUTP nick-end labeling assay showed no evidence of an apoptotic process associated with the hearts of failing Mb^–/– embryos (data not shown), which is consistent with an oncotic process associated with the demise of the mutant embryos.

Hearts were dissected free from Mb^–/– and Mb^+/+ embryos that survived to E12.5. Immunostaining to detect PECAM-1, a marker for endothelial cells, demonstrated hypervascularity in the viable Mb^–/– heart (n=3) compared with wild-type (Mb^+/+) hearts (n=3) (Figure 2). These results were confirmed using a specific lectin that identifies the endothelial cells. Further, we observed that the formation of myocardial vascularization was established earlier during cardiac development in the absence of myoglobin (Figure 2). There was a 48% increase in capillary density in the ventricles of Mb^–/– mice at E13.5 (Figure 2).

Adult Mb^–/– mice also exhibited an increased vascular supply in the atria (46% increase) and ventricles (28% increase) compared with wild-type controls (Figure 3). These findings are in agreement with those of Godecke et al,¹⁷ who used ultrastructural morphometric techniques and reported a 31% increase in capillaries in the myoglobin mutant ventricle. There were no apparent differences in the ultrastructural morphology or the histochemical staining of succinate dehydrogenase activity, a marker of mitochondrial content (online data supplement).^{1 18} We also observed no apparent differences in the profile of serum electrolytes (data not shown), hemoglobin concentration, or hematocrit between age- and sex-matched adult Mb^+/+ or Mb^–/– mice (online data supplement).

Using northern analysis and semiquantitative RT-PCR analysis, we observed enhanced expression in Mb^–/– (and Mb^+/–) embryonic and adult ventricles of a number of genes that are known to be induced in response to hypoxic conditions. Hypoxia-inducible factor-1 (HIF-1), HIF-2 (ePAS), stress proteins (heat shock proteins), and vascular endothelial growth factor are all markedly induced in the myoglobin-deficient ventricles of both embryos (Figure 4) and adults (Figure 5 and online data supplement). These changes in gene expression plausibly drive cellular adaptations, such as increased vascularization, that preserve cardiac function when myoglobin is absent.

Using the working heart preparation, we previously reported no significant differences in cardiac performance between wild-type and myoglobin mutant hearts, even under hypoxic conditions.⁸ Analysis of hemodynamics in working hearts perfused with a physiological mixture of substrates under normoxic conditions revealed no significant differences in coronary flow, cardiac output, or oxygen consumption in Mb^–/– and Mb^+/+ hearts (online data supplement and the Table above). The results obtained in the present study agree closely with the hemodynamic measurements obtained by Grupp et al¹² using the working heart preparation in wild-type mice.

Myocardial oxygen consumption has a number of determinants, including heart rate, contractility, systolic wall tension, and substrate utilization.^{1 13 14} Under physiological conditions, the adult heart generates energy or ATP primarily from the oxidative phosphorylation of fatty acids as the preferred substrate. To define the preferred myocardial substrate for oxidative phosphorylation in the absence of myoglobin, working heart preparations were perfused with ¹³C-labeled substrates and then analyzed using NMR spectroscopy.^{13 14} We found that fatty acids were the preferred substrate in both Mb^–/– and wild-type hearts. The relative use of fatty acids and acetoacetate for the production of acetyl-coenzyme A units feeding into the Krebs cycle was similar in the 2 groups, but there was a modestly greater use of ¹³C-labeled lactate in the hearts from Mb^–/– animals (Table). The use of these ¹³C-labeled substrates in the present study agrees with previously published results using this technique in the rat heart.¹⁴

Working heart preparations were then challenged by the administration of the β-adrenergic agonist isoproterenol. Despite a 20% increase in heart rate, no differences were observed between Mb^+/+ and Mb^–/– hearts with respect to myocardial oxygen consumption in response to isoproterenol stimulation (Figure 6). These results support the conclusion that cardiac function in response to short-term adrenergic stimulation is preserved in the absence of myoglobin.

To assess the response to long-term adrenergic stimulation in this knockout model, ventricular hypertrophy was induced by a 7-day continuous administration of isoproterenol.^{15 16} Using this well-characterized model of cardiac hypertrophy, we determined the functional and structural adaptations in response to adrenergic stimulation in mice that lacked myoglobin. No significant differences in heart weight were noted between age- and sex-matched wild-type and myoglobin mutant mice. Knockout and wild-type mice that were subjected to long-term isoproterenol infusion had a 34% (P<0.005) and 22% (P<0.05) increase, respectively, in the heart weight to body weight ratio (online data supplement) and an induction of atrial natriuretic factor expression in the ventricle compared with their respective controls (Figure 6). There was no evidence of congestive heart failure in either Mb^+/+ or Mb^–/– mice in response to this stimulus. Histological analysis (Figure 6) after long-term isoproterenol infusion revealed the presence of focal regions of myocardial fibrosis, but the frequency of such abnormalities did not seem to differ between Mb^+/+ and Mb^–/– animals.

As a further challenge, ischemic injury was induced surgically in wild-type and myoglobin mutant mice. We observed no differences in survival during the perioperative period or during a 2-month survival period between wild-type or knockout mice. Furthermore, 2 months after the ischemic insult, wild-type and Mb^–/– mice had comparable increases in heart size and decreases in left ventricular function (Figure 7). Using transthoracic echocardiography, fractional shortening was depressed in wild-type (wild-type control [n=5], 0.60±0.02; wild-type with coronary ligation [n=3], 0.34±0.05; P<0.001) and myoglobin mutant (Mb^–/– control [n=5], 0.59±0.03; Mb^–/– with coronary ligation [n=3], 0.34±0.05; P<0.001] mice after coronary artery ligation. Although myoglobin-deficient mice survived this ischemic challenge, these data do not exclude the possibility that myoglobin may function in oxygen metabolism during the early phase(s) of ischemic injury.