ABSTRACT
A strain of axolotl, Ambystoma mexicanum, that carries the cardiac lethal or c gene presents an excellent model system in which to study inductive interactions during heart development. Embryos homozygous for gene c contain hearts that fail to beat and do not form sarcomeric myofibrils even though muscle proteins are present. Although they can survive for approximately three weeks, mutant embryos inevitably die due to lack of circulation. Embryonic axolotl hearts can be maintained easily in organ culture using only Holtfreter’s solution as a culture medium. Mutant hearts can be induced to differentiate in vitro into functional cardiac muscle containing sarcomeric myofibrils by coculturing the mutant heart tube with anterior endoderm from a normal embryo. The induction of muscle differentiation can also be mediated through organ culture of mutant heart tubes in medium ‘conditioned’ by normal anterior endoderm. Ribonuclease was shown to abolish the ability of endoderm-conditioned medium to induce cardiac muscle differentiation. The addition of RNA extracted from normal early embryonic anterior endoderm to organ cultures of mutant hearts stimulated the differentiation of these tissues into contractile cardiac muscle containing well- organized sarcomeric myofibrils, while RNA extracted from early embryonic liver or neural tube did not induce either muscular contraction or myofibrillogenesis. Thus, RNA from anterior endoderm of normal embryos induces myofibrillogenesis and the development of contractile activity in mutant hearts, thereby correcting the genetic defect.
Introduction
It has been established in chicks (Orts-Llorca & Gil, 1965) and salamanders (Jacobson & Duncan, 1968) that anterior endoderm is the most potent inducer of heart differentiation. However, as in most inductive interactions that have been studied, the mechanism by which the inductive stimulus is passed from one tissue to another remains unknown.
The strain of Ambystoma mexicanum that carries the cardiac lethal or c gene provides an excellent model system in which to study inductive interactions that occur during development of the heart. Embryos that are homozygous for the c gene do not develop beating hearts and inevitably die from the resulting lack of circulation. Experiments using parabiotically joined normal and mutant embryos (Humphrey, 1972) have shown that the characteristic morphology of mutant embryos including small gills, microcephaly and ascites, can be prevented by providing the mutant embryo with a source of circulation. Thus, the only tissue that appears to be adversely affected by gene c is the heart. Lemanski and workers (Lemanski, Mooseker, Peachy & Iyengar, 1976; Lemanski, Fuldner & Paulson, 1980) and Starr, Diaz & Lemanski (1985) have demonstrated that the mutant hearts contain the contractile proteins actin, α-actinin and myosin. However, the mutant heart is distinctive in its lack of organized sarcomeric myofibrils.
Two sets of experiments provide evidence that there is a lack of inductive stimulus for heart differentiation in c/c embryos. First, in transplantations performed by Humphrey (1972) mutant heart tissue placed in a normal host differentiates into functional contractile cardiac muscle. On the other hand, normal presumptive heart tissue placed into a mutant embryo fails to contract. The second piece of evidence comes from experiments in which hearts were removed from mutant embryos and maintained in organ culture. Mutant hearts cultured in this manner do not beat and do not develop sarcomeric myofibrils (Hill & Lemanski, 1979). However, if anterior endoderm is cocultured with mutant heart tubes, the mutant heart tissue differentiates into functional contractile cardiac muscle containing normally organized myofibrils (Lemanski, Paulson & Hill, 1979).
This paper describes our efforts to elucidate the mechanisms by which anterior endoderm is able to influence the differentiation of mutant heart tissue. First, by using medium conditioned with normal anterior endoderm to culture mutant hearts we were able to show that a diffusible factor(s) must be involved in this interaction. Second, experiments were performed to identify and characterize the active factor(s) present in conditioned medium. Enzyme-inactivation experiments showed that the activity of conditioned medium was destroyed by ribonuclease. Mutant hearts were then organ cultured in Holtfreter’s solution containing RNA extracted from normal embryonic anterior endoderm. The addition of this RNA to the culture medium stimulated differentiation of mutant heart tissue into functional cardiac muscle composed of cells containing well- organized sarcomeric myofibrils. RNA derived from other embryonic axolotl tissues was not able to induce differentiation of mutant myocardium. Thus, RNA from normal anterior endoderm is able to induce myofibrillogenesis in the mutant heart.
Materials and methods
Preparation of conditioned medium
Ambystoma mexicanum embryos were removed from their jelly coats at various times between stages 29 and 33 (Schreckenberg & Jacobson, 1975). Embryos from both +/+ × +/+ matings and +/+ × +/c matings were dissected; however, the spawnings were used separately. The embryos were placed in sterile modified Holtfreter’s solution (3·5 g NaCl, 100mg CaCl2, 50mg KC1, 204 mgMgSO4 and 200 mg NaHCO3/litre) containing 1% antibiotic/ antimycotic (Gibco). Anterior endoderm was dissected from ten embryos and placed in 150 μ of sterile Holtfreter’s solution in a single well of a Falcon 24-well tissue culture plate. The plate was covered and placed in an 18°C incubator in air for 48 h. Following this culture period the supernatant was drawn off and centrifuged at 15000 g for 10 min to remove any cellular debris. At this point, the conditioned medium was divided into equal samples and used for organ cultures, biochemical analysis or frozen in liquid N2 for later use.
Organ culture of embryonic hearts
Stage-34 embryos were removed from their jelly coats in sterile modified Holtfreter’s solution containing 1 % antibiotic/antimycotic. The embryos were classified as normal or mutant by whether or not they possessed a beating heart. As expected, there were approximately 25 % mutant embryos in +/c × +/c spawnings. The normal and mutant hearts were removed using glass needles to minimize damage during dissection and placed in 10 μl of Holtfreter’s solution, conditioned medium or Holtfreter’s solution containing RNA on a small piece of sterile dental wax in a Falcon tight-sealing culture dish. The cultures were maintained at 18 °C for 24–48 h and their activities were monitored periodically under a dissecting microscope during the culture period.
Enzyme inactivation of conditioned medium
Insoluble enzymes attached to agarose (Sigma) were used to treat the conditioned medium. The enzyme/agarose complexes were washed three times with Holtfreter’s solution prior to use. 200 μl of bovine pancreatic ribonuclease A (0·65 units), 200/4 of bovine pancreatic trypsin (2units) or 200/4 of type VI-A neuraminidase from Clostridium perfringens (2·44 units) were added to 200/4 of conditioned medium. The mixture was agitated at 18°C for 1h. Enzymes attached to agarose were removed by centrifugation and the enzyme-treated conditioned medium was then used for organ cultures of normal and mutant hearts. As a control in these experiments, 200/4 conditioned medium was diluted with 200 μl of Holtfreter’s solution and treated in the same manner as conditioned medium containing enzymes. Conditioned medium was placed in a boiling water bath for 3 min to denature proteins. The boiled conditioned medium was then used to culture mutant hearts. The activity of ribonuclease attached to agarose was tested under the conditions used to inactivate conditioned medium. Polycytidylic acid was incubated with 200μl of insoluble ribonuclease for time periods of up to 1 h at 18 °C. Degradation of polycytidylic acid was measured spectro-photometrically at 280 μm (Zimmerman & Sandeen, 1965).
RNA extraction
Anterior endoderm, presumptive liver and neural tube tissues were dissected from 50 stage-29 +/ + normal embryos. The tissues were homogenized in 6 M-guanidinium isothiocyanate, 5mM-sodium citrate (pH 7·0), 0·lM-β-mercaptoethanol and 0·5 % N-lauroylsarcosine. Following homogenization, 1g of caesium chloride was added to the mixtures. The homogenates were then layered onto a 1 · 2 ml cushion of 5–7M-CSC1 in 0-IM-EDTA (pH7-5) in a polyallomer tube and centrifuged at 31000 revs min−1 for 12 h at 20°C in a Beckman SW60 Ti rotor. Following centrifugation, the supernatants were discarded and the pellets of RNA were dissolved in 10mM-Tris-HCl (pH7·4), 5HIM-EDTA and 1 % SDS. The RNAs were then extracted with a 4:1 mixture of chloroform and 1-butanol, centrifuged at 2000g for 10 min at 4°C and the aqueous layer recovered. The organic phases were re-extracted with the Tris buffer and the two aqueous phases combined. RNAs were precipitated by adding 0-1 (v/v) of 3M-sodium acetate (pH5·2) and 2·2 (v/v) of 95% ethanol (Glisin, Crkvenjakov & Byus, 1974; Foster, Rich, Karr & Przybyla, 1982). Prior to use in organ cultures, the RNAs were reprecipitated from RNase-free distilled water and dissolved in modified Holt-freter’s solution. The concentration and purity of RNA in each sample was determined by spectroscopy at 260 nm and 280 nm.
Electron microscopy
At the end of the culture period, hearts were fixed in a glutaraldehyde-paraformaldehyde-picric acid mixture (Ito & Kamovsky, 1968), postfixed in OsO4, dehydrated and embedded in Epon. Thin sections were stained with lead citrate and uranyl acetate.
Results
Hearts removed from normal embryos and maintained in organ culture beat in a strong and rhythmic manner. Ultrastructurally they showed the same myofibrillar organization as normal hearts in vivo (Figs 1, 2). In contrast to the normal hearts, mutant hearts both in vivo and in organ culture in modified Holtfreter’s solution did not beat or contain organized myofibrils, although amorphous areas and scattered filaments were present (Figs 3,4).
Myocardium of a normal stage-39 axolotl embryo. Note the presence of sarcomeric myofibrils. A, A-band; I, I-band; L, lipid; Z, Z-line. Bar, 0·5μm (X38600).
Myocardial cells of normal embryonic heart explanted into Holtfreter’s solution organ culture at stage 34 and maintained for 48 h. These hearts are comparable to stage-39 in vivo and also contain sarcomeric myofibrils. A, A-band; I, I-band; y, yolk; Z, Z-line. Bar, 0·5 μm (x38600).
Myocardium of a mutant stage-39 axolotl embryo. No organized myofibrils are present; however, scattered 15 nm filaments (arrows) and amorphous areas (am) are often seen. Bar, 0·5μm (x38600).
Myocardial cells of a mutant axolotl embryo heart explanted into Holtfreter’s solution organ culture at stage 34 and maintained for 48 h. These hearts are comparable to mutant stage-39 hearts in vivo. No sarcomeric myofibrils are present. Instead the tissue contains patches of amorphous material (am) and randomly arranged 15 nm filaments (arrows). L, lipid; y, yolk. Bar, 0·5μm (X38600).
Myocardial cells of a mutant axolotl embryo heart explanted into Holtfreter’s solution organ culture at stage 34 and maintained for 48 h. These hearts are comparable to mutant stage-39 hearts in vivo. No sarcomeric myofibrils are present. Instead the tissue contains patches of amorphous material (am) and randomly arranged 15 nm filaments (arrows). L, lipid; y, yolk. Bar, 0·5μm (X38600).
Experiments using medium conditioned by normal anterior endoderm to organ culture mutant hearts showed that the endoderm tissue itself was not necessary for differentiation of the presumptive heart tissue to occur (Table 1). The active factor(s) produced by anterior endoderm was released into the culture medium and retained its ability to stimulate myofibrillogenesis in mutant hearts as indicated macroscopically by the development of propagated rhythmic contractions in the heart tubes and ultrastructurally by the presence of organized sarcomeric myofibrils in the myocardium (Fig. 5).
Myocardium of a mutant stage-34 axolotl embryo that has been cultured in conditioned medium prepared using normal anterior endoderm. This ‘corrected’ mutant heart was beating strongly at the end of the culture period and ultrastructurally resembles the normal hearts seen in Figs 1 and 2. Sarcomeric myofibrils are prominent throughout the myocardium. A, A-band; 1, I-band; L, lipid; y, yolk; Z, Z-line. Bar, 0·5 μm (×38600).
Myocardium of a mutant stage-34 axolotl embryo that has been cultured in conditioned medium prepared using normal anterior endoderm. This ‘corrected’ mutant heart was beating strongly at the end of the culture period and ultrastructurally resembles the normal hearts seen in Figs 1 and 2. Sarcomeric myofibrils are prominent throughout the myocardium. A, A-band; 1, I-band; L, lipid; y, yolk; Z, Z-line. Bar, 0·5 μm (×38600).
Conditioned media from different sources were not uniformly effective in inducing myofibrillogenesis in mutant hearts. We found that conditioned medium produced by anterior endoderm removed from embryos of a homozygous (+/+ × +/+) spawning was significantly more effective than that produced by anterior endoderm from heterozygous (+/+×+/c) embryos (Table 1). Another set of experiments suggested that anterior endoderm from younger embryos was more effective in promoting differentiation of mutant heart tissue than from older embryos (Table 1). In fact by stage 33 it appears that the anterior endoderm is no longer able to provide the active factor(s) for differentiation of mutant hearts in culture.
Enzyme-inactivation studies were done in an effort to discover the identity of the active factor(s) in the conditioned medium. The treatment of conditioned medium with neuraminidase had no effect on the ability of conditioned medium to correct mutant hearts. Boiling of conditioned medium did not affect its activity. Treatment with trypsin resulted in a reduction of the activity of conditioned medium, although it retained the ability to induce rhythmic contractions in 38 % of the cultured mutant hearts. The only enzyme that completely abolished the activity of conditioned medium was ribonuclease (Table 2). An increase in the acid-soluble absorbance of polycytidylic acid incubated with the ribonuclease at times up to 1 h indicated that the enzyme was able to degrade RNA under the experimental conditions used. In further studies, RNA was extracted from normal embryonic anterior endoderm and added to organ cultures of mutant hearts. Mutant hearts organ cultured in modified Holtfreter’s solution plus RNA were indistinguishable from those cultured in conditioned medium. The hearts beat rhythmically (Table 3) and ultrastructural examination revealed the presence of sarcomeric myofibrils (Fig. 6). In contrast to the effects of endoderm RNA on mutant heart tissue, RNAs extracted from embryonic fiver and neural tube were not able to induce differentiation of functional cardiac muscle from mutant heart tubes. Mutant hearts cultured in modified Holtfreter’s solution containing either liver RNA or neural tube RNA did not beat (Table 3) and did not contain sarcomeric myofibrils. Ultrastructurally, these hearts looked like mutant hearts in vivo (Fig. 3) or mutant hearts cultured only in Holtfreter’s solution (Fig-4).
Myocardium of a mutant stage-34 axolotl embryo organ cultured in modified Holtfreter’s solution containing RNA extracted from normal anterior endoderm. This mutant heart is also considered to be ‘corrected’ as it was beating rhythmically at the end of the culture period and contains well- organized myofibrils. A, A-band; /, I-band; L, lipid; y, yolk; Z, Z-line. Bar, 0·5μm (×38600).
Myocardium of a mutant stage-34 axolotl embryo organ cultured in modified Holtfreter’s solution containing RNA extracted from normal anterior endoderm. This mutant heart is also considered to be ‘corrected’ as it was beating rhythmically at the end of the culture period and contains well- organized myofibrils. A, A-band; /, I-band; L, lipid; y, yolk; Z, Z-line. Bar, 0·5μm (×38600).
Discussion
How one tissue influences the developmental pathway of another is an old but still unanswered question. For example, it has been established that the anterior endoderm induces differentiation of the overlying mesoderm tissue to form the heart at a very early stage in development (Orts-Llorca & Gil, 1965; Jacobson & Duncan, 1968). However, it is not clear how the inductive process or transfer of information takes place. The cardiac lethal mutation in the axolotl, Ambystoma mexicanum, provides a unique opportunity to explore the induction phenomenon. In homozygous c/c mutant embryos there is apparently a lack of appropriate inductive stimulus to the presumptive heart (Hill & Lemanski, 1979). The heart does not develop beyond stage 34 and remains a quiescent structure (Humphrey, 1972). Although muscle proteins are present (Lemanski et al. 1976, 1980; Starr et al. 1985) they are not organized into myofibrillar structures. While other investigators have provided evidence of coordinated muscular contractions propagated throughout the lengths of mutant heart tubes in culture (Kulikowski & Manasek, 1978), we only occasionally find contractions in the conus region of the heart. We have never observed contractions propagated throughout the lengths of mutant heart tubes either in vivo or when cultured in Holtfreter’s solution alone. Neither Kuli-kowski & Manasek (1978) nor our laboratory has been able to demonstrate by electron microscopy the presence of sarcomeric myofibrils in cardiac tissue of mutant embryos (Lemanski, 1973; Hill & Lemanski, 1979). Thus, we conclude that either in vivo or in organ culture, presumptive heart tissue from cardiac lethal mutant embryos is unable to differentiate fully due to the lack of a final inductive stimulus. What makes this mutation so useful in studying heart development is that the mutant heart can be ‘rescued’ in vitro by culturing it with normal anterior endoderm (Lemanski et al. 1979). Thus, in this experimental system we have a tissue that failed to receive the proper input for differentiation at the usual time in development. However, we can still influence the fate of this tissue by providing it with an exogenous inductive stimulus.
Since previous work in this laboratory had shown that mutant heart tissue could be induced to differentiate by coculturing with normal anterior endoderm (Lemanski et al. 1979) our first question was whether cell-to-cell contact with the endoderm tissue was necessary for this induction to occur or whether a diffusible substance might be involved. We found that the inductive stimulus produced by anterior endoderm was indeed able to be transmitted via the conditioned medium to mutant heart tissue without the simultaneous presence of endoderm in the culture. Clearly, the endoderm must be producing a diffusible factor(s)which presumptive mutant heart tissue responds to. Endoderm obtained from embryos produced by a homozygous normal (+/+) mating was more effective in inducing cardiac muscle differentiation than that taken from embryos of a heterozygous spawning (+/+ × +/c). Not surprisingly, since Jacobson & Duncan (1968) have shown that the inductive action of endoderm in vivo is greatest at early stages of development in both the California newt, Taricha torosa, and the salamander,Ambystoma tigrinum, we found that conditioned medium produced from stage-29 endoderm was more potent than that produced from stage-33 endoderm.
However, even using stage-29 homozygous embryos to provide the endoderm with which to condition the medium we were only able to correct approximately 70 % of the mutant hearts. There are several possible reasons why the other 30 % do not develop the ability to contract. First, even though the hearts are removed very carefully for organ culture, it is possible that some damage occurs to the structure and thus inhibits its further differentiation. Another explanation might be that the heart normally receives its inductive signals much earlier in development than we are able to reproduce in vitro, since it is not possible at present to distinguish a mutant embryo from its normal siblings until stage 34, when the normal heart begins to beat. Thus, by the time the mutant hearts are exposed to conditioned medium in organ culture they are very possibly near the end of the time they are competent to respond to an inductive stimulus.
Enzyme inactivation was chosen as the method to identify the class of compounds to which the active factor belonged. To overcome the problem of having to add both enzyme and enzyme inhibitor to the conditioned medium we used enzymes attached to agarose, which then could be removed by centrifugation following the incubation period. Of the three enzymes added to conditioned medium, i.e. trypsin to degrade proteins, neuraminidase to cleave sialic acid residues from glycoproteins and ribonuclease to inactivate RNAs, only treatment with ribonuclease completely abolished the activity of the conditioned medium. The corollary of this experiment, namely adding RNA from normal anterior endoderm, was then done. In two separate experiments, using mutant embryos from two different spawnings, endoderm RNA added to the Holtfreter’s solution used to organ culture the mutant hearts resulted in the development of contractile cardiac muscle. To corroborate the macroscopic evidence of differentiation, namely development of contractile activity, organ cultured hearts were examined ultrastructurally. Mutant hearts cultured in the presence of endoderm RNA resembled hearts from normal embryos, containing organized sarcomeric myofibrils (Figs 1,6). Mutant hearts cultured with liver or neural tube RNA did not contain these structures. Thus, RNA derived from normal anterior endoderm is able to substitute for the endoderm tissue itself in inducing myofibrillogenesis in mutant heart tissue.
The idea that RNA can be used to direct specific differentiation pathways is not new. Early work, by Niu (1958) using embryonic ectoderm from Ambystoma tigrinum, and by Sanyal & Niu (1966), Butros (1965) and Niu & Deshpande (1973) using postnodal pieces of chick blastoderm, described a variety of differentiation patterns when these tissues were cultured with the addition of RNA from several different sources. When axolotl ectoderm treated with thymus RNA is implanted into host flank tissue, it is capable of differentiating into small bodies resembling thymus (Niu, 1958). When postnodal chick blastoderm is exposed to brain RNA, neural tissue results (Sanyal & Niu, 1966). The application of liver RNA leads to more variable results with some neural tissue developing (Sanyal & Niu, 1966) and in other cases myoblasts, mesonephric tubules and endoderm tissue being present (Butros, 1965). Kidney RNA, on the other hand, does not seem to induce neural differentiation but instead tubular structures are formed (Sanyal & Niu, 1966). The RNA most extensively studied using chick postnodal blastoderm is that from embryonic heart. Postnodal chick blastoderm cultured with RNA derived from embryonic chick heart is able to differentiate into pulsating structures containing actin and myosin. These proteins are arranged into organized sarcomeric myofibrils similar to those seen in the heart in vivo (Deshpande & Siddiqui, 1977). The inducer RNA has been characterized as a 7S poly(A)RNA which hybridizes to repetitive chick DNA and shares homology with noncoding segments of myosin genes (Khandekar, Saidapet, Krauskopf, Zarraga, Lin, Mendola & Siddiqui, 1984). There is evidence in chick blastoderm (Siddiqui, 1983) and in primary cultures of chick myoblasts (Mroczkowski, Dym, Siegel & Heywood, 1980), that muscle cells can take up and use exogenously supplied RNA. The addition of RNA extracted from adult chicken heart to cultures of embryonic chick myocardial cells results in the maturation of cell membranes (McLean, Renaud, Niu & Sperelakis, 1977). Following an incubation period of 6-14 days, myocardial cells cultured with this RNA developed adult-type electrophysiological properties such as fast sodium channels which were sensitive to tetrodotoxin. This effect is abolished by the administration of cyclohexamide.
In the present studies, we have demonstrated that RNA isolated from anterior endoderm induces myo-fibrillogenesis in mutant axolotl heart tissue. Nevertheless, we have no firm evidence yet that its effect on mutant heart tissue in vitro is the same as during in vivo embryonic development. The precise mechanism) by which anterior endoderm RNA acts on mutant hearts remains unknown. At the present time we do not know whether the active RNA is functioning as messenger RNA or if it is acting in some other capacity, such as regulating transcription or translation (Mroczkowski, McCarthy, Zezza, Bragg & Heywood, 1984; Siddiqui, Khandekar, Krauskopf, Mendola, Zarraga & Saidapet, 1984), or affecting the cell membranes of the developing myocardium (McLean et al. 1977). Studies are currently in progress to characterize the RNA that is able to overcome the lack of myofibrillogenesis in cardiac lethal mutant embryos.
ACKNOWLEDGEMENTS
We would like to thank Linda Riles and José Diaz for their excellent technical assistance. Special thanks are due also to Chris Starr for technical support and advice in RNA extraction procedures. The axolotl colony at Indiana University kindly supplied many of the embryos for this study. This work was supported by a Muscular Dystrophy Postdoctoral fellowship to L.A.D. and NIH Grants HL37702, HL32184 and a Grant from the American Heart Association to L.F.L.