Cardiac organogenesis in vitro: reestablishment of three-dimensional tissue architecture by dissociated neonatal rat ventricular cells.

The mammalian heart does not regenerate in vivo. The heart is, therefore, an excellent candidate for tissue engineering approaches and for the use of biosynthetic devices in the replacement or augmentation of defective tissue. Unfortunately, little is known about the capacity of isolated heart cells to re-establish tissue architectures in vitro. In this study, we examined the possibility that cardiac cells possess a latent organizational potential that is unrealized within the mechanically active tissue but that can be accessed in quiescent environments in culture. In the series of experiments presented here, total cell populations were isolated from neonatal rat ventricles and recombined in rotating bioreactors containing a serum-free medium and surfaces for cell attachment. The extent to which tissue-like structure and contractile function were established was assessed using a combination of morphological, physiological, and biochemical techniques. We found that mixed populations of ventricular cells formed extensive three-dimensional aggregates that were spontaneously and rhythmically contractile and that large aggregates of structurally-organized cells contracted in unison. The cells were differentially distributed in these aggregates and formed architectures that were indistinguishable from those of intact tissue. These architectures arose in the absence of three-dimensional cues from the matrix, and the formation of organotypic structures was apparently driven by the cells themselves. Our observations suggest that cardiac cells possess an innate capacity to re-establish complex, three-dimensional, cardiac organization in vitro. Understanding the basis of this capacity, and harnessing the organizational potential of heart cells, will be critical in the development of tissue homologues for use in basic research and in the engineering of biosynthetic implants for the treatment of cardiac disease.

Neonatal rat heart cells cultured in simulated microgravity.

In vitro characteristics of cardiac cells cultured in simulated microgravity are reported. Tissue culture methods performed at unit gravity constrain cells to propagate, differentiate, and interact in a two-dimensional (2D) plane. Neonatal rat cardiac cells in 2D culture organize predominantly as bundles of cardiomyocytes with the intervening areas filled by nonmyocyte cell types. Such cardiac cell cultures respond predictably to the addition of exogenous compounds, and in many ways they represent an excellent in vitro model system. The gravity-induced 2D organization of the cells, however, does not accurately reflect the distribution of cells in the intact tissue. We have begun characterizations of a three-dimensional (3D) culturing system designed to mimic microgravity. The NASA-designed High-Aspect Ratio Vessel (HARV) bioreactors provide a low shear environment that allows cells to be cultured in static suspension. HARV-3D cultures were prepared on microcarrier beads and compared to control-2D cultures using a combination of microscopic and biochemical techniques. Both systems were uniformly inoculated and medium exchanged at standard intervals. Cells in control cultures adhered to the polystyrene surface of the tissue culture dishes and exhibited typical 2D organization. Cells cultured in HARVs adhered to microcarrier beads, the beads aggregated into defined clusters containing 8 to 15 beads per cluster, and the clusters exhibited distinct 3D layers: myocytes and fibroblasts appeared attached to the surfaces of beads and were overlaid by an outer cell type. In addition, cultures prepared in HARVs using alternative support matrices also displayed morphological formations not seen in control cultures. Generally, the cells prepared in HARV and control cultures were similar; however, the dramatic alterations in 3D organization recommend the HARV as an ideal vessel for the generation of tissuelike organization of cardiac cells in vitro.

Skeletal muscle satellite cells cultured in simulated microgravity.

Satellite cells are postnatal myoblasts responsible for providing additional nuclei to growing or regenerating muscle cells. Satellite cells retain the capacity to proliferate and differentiate in vitro and, therefore, provide a useful model to study postnatal muscle development. Most culture systems used to study postnatal muscle development are limited by the two-dimensional (2-D) confines of the culture dish. Limiting proliferation and differentiation of satellite cells in 2-D could potentially limit cell-cell contacts important for developing the level of organization in skeletal muscle obtained in vivo. Culturing satellite cells on microcarrier beads suspended in the High-Aspect-Ratio-Vessel (HARV) designed by NASA provides a low shear, three-dimensional (3-D) environment to study muscle development. Primary cultures established from anterior tibialis muscles of growing rats (approximately 200 gm) were used for all studies and were composed of greater than 75% satellite cells. Different inoculation densities did not affect the proliferative potential of satellite cells in the HARV. Plating efficiency, proliferation, and glucose utilization were compared between 2-D culture and 3-D HARV culture. Plating efficiency (cells attached divided by cells plated x 100) was similar between the two culture systems. Proliferation was reduced in HARV cultures and this reduction was apparent for both satellite cells and nonsatellite cells. Furthermore, reduction in proliferation within the HARV could not be attributed to reduced substrate availability because glucose levels in medium from HARV and 2-D cell culture were similar. Morphologically, microcarrier beads within the HARV were joined together by cells into 3-D aggregates composed of greater than 10 beads/aggregate. Aggregation of beads did not occur in the absence of cells. Myotubes were often seen on individual beads or spanning the surface of two beads. In summary, proliferation and differentiation of satellite cells on microcarrier beads within the HARV bioreactor results in a 3-D level of organization that could provide a more suitable model to study postnatal muscle development than is currently available with standard culture methods.

Evidence for multiple satellite cell populations and a non-myogenic cell type that is regulated differently in regenerating and growing skeletal muscle.

We have performed studies to determine if different populations of satellite cells provide nuclei to growing and regenerating skeletal muscle fibers. Satellite cells were isolated from regenerating or growing anterior tibialis muscles, and their phenotypic properties were compared in vitro. Isolates from regenerating muscle contained 31% satellite cells, and those from control muscle contained 66% satellite cells, as determined by their expression of desmin. Among the desmin-positive satellite cells present from each preparation, two distinct populations of satellite cells were evident. Approximately 28% of satellite cell colonies were composed of only large cells, contained less than 50 cells/colony, and were designated as type 1 colonies. The remainder of satellite cell colonies isolated from either regenerating or control muscles were primarily composed of small cells, contained from 60 to 150 cells/colony, and were designated as type 2 colonies. Despite dramatic differences in the ratio of myogenic to non-myogenic cell types, satellite cells from regenerating and control muscles formed myotubes and expressed myosin heavy chain at similar levels. Treatment of regenerating cultures with dexamethasone resulted in a 16% increase in the number of desmin-positive colonies and dramatically decreased the proliferation of non-myogenic cells. These results suggest that at least two distinct populations of satellite cells can be isolated from regenerating and control skeletal muscles, and that non-myogenic cells are differentially regulated in regenerating versus non-regenerating environments.

Hemin enhances differentiation and maturation of cultured regenerated skeletal myotubes.

Satellite cells, isolated from marcaine-damaged rat skeletal muscle, differentiate in culture to form contracting, cross-striated myotubes. Addition of 20 microM hemin (ferriprotoporphyrin IX chloride) to the culture medium resulted in increases in the number, size, and alignment of myotubes; in the number of myotubes that exhibited cross-striations; and in the strength and frequency of myotube contractions. Hemin increased satellite cell fusion by 27%, but decreased cell proliferative rate by 30%. Hemin increased the specific activity of creatine kinase (CK), a sensitive indicator of muscle differentiation, by 157%. Separation of CK isoenzymes by agarose gel electrophoresis showed that hemin increased only the muscle-specific CK isoenzymes (MM-CK and MB-CK). Thus, hemin seems to duplicate some of the effects of innervation on cultured myotubes by increasing contraction frequency and strength, appearance of cross-striations, and muscle-specific isoenzymes. In contrast, 3-amino-1,2,4-triazole, an inhibitor of heme biosynthesis, decreased the number of cross-striated myotubes, the strength and frequency of myotube contractions, and CK activity. These inhibitory effects were reversed by hemin. Collectively, these results demonstrate a physiologically significant role for heme in myotube maturation.

Protein metabolism during nutrient deprivation and refeeding of neonatal heart cells.

Pathological conditions or nutrient deprivation in the heart cause an imbalance between rates of protein synthesis and degradation, often resulting in a severe depletion of cardiac protein. We used cultured neonatal rat heart cells, a model system exhibiting positive nitrogen balance, to examine the effects of 10 h of starvation on myocardial glucose and protein metabolism. Cellular capacity for glucose utilization was depressed after starvation, as evidenced by lower hexokinase and other glycolytic enzyme activities and a 21% decrease in glucose usage. A 21.0% decrease in protein synthetic rate and an increase in protein degradation rate combined to yield a 29.5% decrease in total cellular protein during starvation. Degradation rates increased 29.0, 46.7, and 59.6% in 2-, 24-, and 96-h prelabeled cells, respectively, indicating that lability increased with half-life of proteins. During refeeding of starved, cultured cells, at least three proteins were synthesized at a lower rate. At the same time, proteins with approximate molecular masses of 45, 84, 92, and 174 kDa exhibited increased synthesis.

Hemin increases aerobic capacity of cultured regenerating skeletal myotubes.

Regeneration of damaged, mature muscle occurs by differentiation of satellite cells. In culture, satellite cell myoblasts proliferate, align, and fuse to form cross-striated, contracting myotubes. The biochemical changes and the factors that regulate differentiation in satellite cells have not been investigated previously. We report here that no significant differences in glucose uptake rate or glucose oxidation rate were observed between regenerating myoblasts and myotubes, whereas the aerobic oxidation of palmitic acid increased 7.3-fold between these differentiation states. Specific activities of enzymes of critical importance in aerobic metabolism or in production of ATP were increased 2- to 3.5-fold during fusion. Addition of 20 microM hemin to regenerating muscle cultures potentiated the aerobic capacity as evidenced by a 23.6% increase in palmitate oxidation rate. Hemin also increased the specific activities of all nonheme enzymes investigated with the exception of phosphofructokinase. This augmentation of aerobic metabolism together with the time frame of active muscle differentiation suggests a complex role for hemin in myogenesis.

Increased aerobic glucose oxidation by cAMP in cultured regenerated skeletal myotubes.

Previous studies of embryonic rat skeletal muscle cultures suggested that there was a correlation between intracellular adenosine 3',5'-cyclic monophosphate (cAMP) concentration and activities of enzymes of oxidative energy metabolism. We investigated the ability of agents that elevate intracellular cAMP by three different mechanisms (activation of adenylate cyclase, inhibition of phosphodiesterase, and analogues) to alter not only the activities of 11 glycolytic and mitochondrial enzymes but also the rate of flux through aerobic glucose oxidation in intact myotubes derived from regenerating rat muscle satellite cells. The enzyme activities were not consistently altered when cAMP was elevated, with the exception of the electron transport enzyme, NADH cytochrome c reductase, whose activity was elevated by exposure of the myotubes to cholera toxin (110% of control), 3-isobutyl-1-methylxanthine (112%), caffeine (119%), and 8-bromoadenosine 3',5'-cyclic monophosphate (120%). The rate of flux of aerobic glucose oxidation was elevated by all agents (116-157% of control) except cholera toxin. This study allowed us to compare the metabolic characteristics of myotube cultures derived from satellite cells with those from embryonic muscle, from a previous study. Despite differences between these two models, together, the data present strong evidence that an increase in intracellular cAMP can cause an increase in oxidative capacity.

Preferential distribution of non-esterified fatty acids to phosphatidylcholine in the neonatal mammalian myocardium.

Non-esterified fatty acids are used to a limited extent as an energy source in the newborn-mammalian heart. Therefore additional roles for palmitic and oleic acids during this early period of growth and development were investigated in the cultured neonatal-rat heart cell model system. Our results indicate significant differences in nonesterified-fatty-acid metabolism exist in this system in comparison with the adult rat or embryonic chick heart. Initial rates of depletion of palmitate and oleate from serum-free growth medium by heart cells obtained from 2-day-old rats and maintained in culture for 10 or 11 days were 111 +/- 2 and 115 +/- 3 pmol/min per mg of protein respectively. In serum-containing medium, the initial depletion rates were 103 +/- 3 and 122 +/- 4 pmol/min per mg of protein respectively, when endogenous serum nonesterified-fatty-acid concentrations were included in rate calculations. Less than 1% of the intracellularly incorporated fatty acids were found in aqueous products at any time. After 25 h, 15.5% of the initial palmitate was deposited intracellularly in the phosphatidylcholine lipid fraction, 4.2% in the triacylglycerol + fatty-acid-ester fraction and 3.1% in the sphingomyelin fraction. These results contradict the classical view, based on findings with the lipid-dependent adult heart, that exogenous nonesterified fatty acids are directed intracellularly primarily to pathways of oxidation or to storage as triacylglycerol. More importantly, it underscores the significance of exogenous non-esterified fatty acids in membrane biosynthesis of the developing mammalian heart. Included here is a new method for one-dimensional t.l.c. separation of metabolically important polar lipids.

Control of enzyme activity levels by serum and hydrocortisone in neonatal rat heart cells cultured in serum-free medium.

A serum-free, hormone-supplemented medium (SFHM) for maintaining neonatal rat heart cells in culture has been developed in this laboratory (Mohamed et al., 1983). Morphological assessment of heart cells grown in SFHM show it to be similar to commonly used serum-supplemented media. To quantitatively compare cell behavior in SFHM with serum-supplemented media, the activities of ten regulatory enzymes which represent four metabolic pathways were studied in heart cells cultured in SFHM. The enzyme activities which were measured included hexokinase, glucose-6-phosphate dehydrogenase, 6-phosphogluconate dehydrogenase, phosphofructokinase, pyruvate kinase, NAD+-linked sn-glycerol-3-phosphate dehydrogenase, malate dehydrogenase, NAD+-linked isocitrate dehydrogenase, NADH-cytochrome c reductase, and succinic cytochrome c reductase. Rat heart cells maintained in culture on SFHM are not only qualitatively and quantitatively similar to those maintained in serum-supplemented medium but also provide a more suitable model system for metabolic studies of neonatal cardiac tissue for several reasons: 1) many enzyme activities that may represent dedifferentiation are elevated by serum; 2) NAD-linked glycerol-3-phosphate dehydrogenase activity in cells maintained on SFHM is similar to the in vivo activity; 3) cells beat at or near the in vivo frequency and can be maintained 3 months on SFHM; 4) the SFHM is chemically defined and thus can be completely manipulated by the investigator. The effects of three concentrations of hydrocortisone (HC) (5,000 ng/ml, 50 micrograms/ml, 0 ng/ml) on heart cells cultured in SFHM supported our previous conclusion that function (beating) and growth (protein accumulation) are inversely related in cultured neonatal rat heart cells.

Triiodothyronine depresses the NAD-linked glycerol-3-phosphate dehydrogenase activity of cultured neonatal rat heart cells.

The activity of NAD-linked alpha-glycerol-3-phosphate dehydrogenase (NAD-G3PDH; EC 1.1.1.8) was depressed by 35% when the thyroid hormone 3,3',5-triiodo-L-thyronine (20 micrograms/liter) was added to the serum-free, hormonally supplemented medium of cultured neonatal rat heart cells. The degree of depression was greater (65%) when the medium contained normal serum levels of hydrocortisone and insulin. There is a dramatic inverse dose-response relationship between triiodothyronine levels and NAD-G3PDH activity. The classic elevation by thyroid hormones of the FAD-linked alpha-glycerol-3-phosphate dehydrogenase (FAD-G3PD; EC 1.1.99.5) was observed concurrently. The medium-glucose depletion rate in triiodothyronine-free cells was depressed 32% through 11 days-in-culture, indicating reduced glycolytic activity. The activities of nine other metabolically important enzymes which were measured during this study, including hexokinase, glucose-6-phosphate dehydrogenase, 6-phosphogluconate dehydrogenase, phosphofructokinase, pyruvate kinase, malate dehydrogenase, NAD-isocitrate dehydrogenase, NADH cytochrome c reductase, and succinic cytochrome c reductase, did not respond to varying triiodothyronine concentrations.

Myocytes and fibroblasts exhibit functional synergism in mixed cultures of neonatal rat heart cells.

Cardiac cells obtained from neonatal rat heart contain a mixed population of cell types that can be enriched in culture in either myocytes or fibroblast-like cells. A metabolic comparison of mixed heart cell cultures with enriched cultures of the same age-in-culture and initial cell density showed that mixed cultures used glucose more rapidly than either enriched myocytes or fibroblasts. Mixed cultures were shown to respond to deprivation of insulin and of serum with decreases in the rate of glucose usage and decreases in the protein content of cells, whereas enriched cultures did not respond in the expected manner to insulin deprivation. Mixed, 11-day-old cells also exhibited greater increases in cellular protein and greater resistance to the stress of starvation than enriched cultures. Palmitate usage, however, was similar in all cultures examined. We conclude that mixed cultures may serve as a better model system to study cardiac metabolism and to monitor the effects of drugs and hormones on the neonatal myocardium. In addition, it is clear from our results that myocytes and fibroblastic-like cells coexist in a metabolically functional synergism.

Glucose metabolism, insulin effects, and developmental age of cultured neonatal rat heart cells.

Cultured heart cells from 2-3 day old and 5-6 day old neonatal rats have been used as a model system for the characterization of carbohydrate metabolism in developing cardiac tissue. The rate of depletion of glucose from the growth medium was dependent on 1) the age of the animals from which the cultured cells were obtained, and 2) the presence and absence of serum and/or insulin in the growth medium. The glucose depletion rate in insulin and serum-containing medium was 9.63 +/- 0.96 nmol/min/mg protein for heart cell cultures from 2 day old rats and 3.51 +/- 0.68 nmol/min/mg protein in heart cell cultures from 5 day old rats. Appearance of lactate in the medium during these experiments occurred at the rates of 18.6 +/- 7.9 nmol/min/mg and 6.4 +/- 1.2 nmol/min/mg, respectively. In the absence of serum and insulin, the medium glucose depletion rates were 5.7 +/- 1.6 and 2.2 +/- 0.5 nmol/min/mg for cells derived from 2-day-old and 5-day-old rats, respectively. It is apparent from these data that immature cardiac cells depend upon glucose as a primary source of energy for muscle contraction and cellular growth, and that less-efficient energy-yielding metabolic pathways are used to obtain ATP.

Oxidative titrations of reduced cytochrome aa3: correlation of midpoint potentials and extinction coefficients observed at three major absorption bands.

Anaerobic oxidative titrations of purified cytochrome aa3 were monitored at three wavelengths (444, 604, and 820 nm), in both the absence and the presence of carbon monoxide. Computer simulation of each titration curve was utilized to ascertain the midpoint potentials of the four oxidation-reduction centers of the enzyme. For experiments performed under nitrogen, two components were found to titrate with low potential (heme aL = 220 mV, CuL = 240 mV) and two with high potential (heme ath, cuH = 340 mV), consistent with results obtained previously in reductive titrations. Unequal heme extinction coefficients were observed at 444 nm. Oxidation by either potassium ferricyanide or 1,1'-bis(hydroxymethyl)ferricinium ion showed that the low potential heme component contributed 75% of the absorbance change at 444 nm. At 820 nm, the entire absorbance change could be attributed to a single, low potential copper component. Midpoint potentials calculated for the carbon monoxide complexed enzyme agreed with previously reported values. The copper components retained the values observed under nitrogen, while the titratable heme group gave an apparent midpoint potential of 260 mV. These results enable us to assign absorbance changes at various wavelengths to specific redox components of cytochrome aa3.

Oxidative titrations of reduced cytochrome aa3: influence of cytochrome c and carbon monoxide on the midpoint potential values.

Oxidative titrations were performed on the electrostatic complex formed between cytochrome c and cytochrome aa3 at low ionic strength. Midpoint potentials of the redox centers in the proteins in 1:1 and 2:1 complexes were compared with those in mixtures of the cytochromes at high ionic strength. Computer simulations of all titrations yielded midpoint potentials for the components of cytochrome aa3 which were consistent with literature values for isolated cytochrome aa3 or mixture of cytochromes c and aa3. However, the unequal heme extinction coefficients observed previously (Schroedl, N.A., and Hartzell, C.R. (1977), Biochemistry 16, 1327) during oxidative titrations of cytochrome aa3 became equal in magnitude under these experimental conditions. The binding of cytochrome c to cytochrome aa3 changed the midpoint potentials of cytochrome aa3 by 15-20 mV, while the midpoint potentials for cytochrome c were altered by 50-60 mV. Careful analysis of these titrations including computer simulation revealed that cytochrome c was able to bind to cytochrome aa3 only after cytochrome aL2+ had become oxidized. When bound to cytochrome aa3, the midpoint potential of cytochrome c was 210 7V. Titrations performed under a carbon monoxide atmosphere revealed cytochrome aa3 midpoint potentials unchanged from reported values. Cytochrome c again exhibited a midpoint potential of 210 mV after binding to cytochrome aa3.

Oxidative titrations of reduced cytochrome aa3: anisotropic extinction behavior observed in the heme alpha-band region.

Direct chemical titrations of reduced, purified cytochrome aa3 were monitored at 604 nm. Anaerobic oxidation by potassium ferricyanide or 1,1'-bis(hydroxymethyl)-ferricinium ion was compared with the natural substrate, molecular oxygen. The four electrons from reduced cytochrome aa3 were donated to one oxygen molecule, resulting in a linear titration with only fully oxidized or fully reduced enzyme molecules present. Unlike the linear or sigmoidal titrations which resulted from various reductive titrations reported previously, pronounced hyperbolic curves were obtained with the chemical oxidants. The midpoint potential values exhibited by the four metal centers of cytochrome aa3 were determined by computer simulation of the titration curves. A high potential pair of components (heme a = 340 mV, Cu = 340 mV) and a low potential pair (heme a = 220 mV, Cu = 240 mV) were observed, consistent with literature values. Unique to these equilibrium studies was a split in the heme a extinction coefficients at 604 nm. The low potential heme a component contributed 80-85% to the total absorbance change. A polarographic assay was used to verify that the integrity of cytochrome aa3 remained intact during equilibrium titrations spanning several hours. These experimental results indicate that simple chemical reversibility does not exist for cytochrome aa3 under anaerobic conditions.