Search

The now classical idea that programmed cell death (apoptosis) contributes to a plethora of developmental processes still has lost nothing of its impact. It is, therefore, important to establish effective three-dimensional (3D) reconstruction as well as simulation techniques to decipher the exact patterns and functions of such apoptotic events. The present study focuses on the question whether and how apoptosis promotes neurulation-associated processes in the spinal cord of Tupaia belangeri (Tupaiidae, Scandentia, Mammalia). Our 3D reconstructions demonstrate that at least two craniocaudal waves of apoptosis consecutively pass through the dorsal spinal cord. The first wave appears to be involved in neural fold fusion and/or in selection processes among premigratory neural crest cells. The second one seems to assist in establishing the dorsal signaling center known as the roof plate. In the hindbrain, in contrast, apoptosis among premigratory neural crest cells progresses craniocaudally but discontinuously, in a segment-specific manner. Unlike apoptosis in the spinal cord, these segment-specific apoptotic events, however, precede later ones that seemingly support neural fold fusion and/or postfusion remodeling. Arguing with Whitehead that biological patterns and rhythms differ in that biological rhythms depend "upon the differences involved in each exhibition of the pattern" (Whitehead in An enquiry concerning the principles of natural knowledge. Cambridge University Press, London, 1919, p. 198) we show that 3D reconstruction and simulation techniques can contribute to distinguish between (static) patterns and (dynamic) rhythms of apoptosis. By deciphering novel patterns and rhythms of developmental apoptosis, our reconstructions help to reconcile seemingly inconsistent earlier findings in chick and mouse embryos, and to create rules for computer simulations.

The neurogenic trigeminal placode develops from the crescent-shaped panplacodal primordium which delineates the neural plate anteriorly. We show that, in Tupaia belangeri, the trigeminal placode is represented by a field of focal ectodermal thickenings which over time changes positions from as far rostral as the level of the forebrain to as far caudal as opposite rhombomere 3. Delamination proceeds rostrocaudally from the ectoderm adjacent to the rostral midbrain, and contributes neurons to the trigeminal ganglion as well as to the ciliary ganglion/oculomotor complex. Proliferative events are centered on the field prior to the peak of delamination. They are preceded, paralleled and, finally, outnumbered by apoptotic events which proceed rostrocaudally from non-delaminating to delaminating parts of the field. Apoptosis persists upon regression of the placode, thereby exhibiting a massive "wedge" of apoptotic cells which includes the postulated position of the "ventrolateral postoptic placode" (Lee et al. in Dev Biol 263:176-190, 2003), merges with groups of lens-associated apoptotic cells, and disappears upon lens detachment. In conjunction with earlier work (Washausen et al. in Dev Biol 278:86-102, 2005) our findings suggest that apoptosis contributes repeatedly to the disintegration of the panplacodal primordium, to the elimination of subsets of premigratory placodal neuroblasts, and to the regression of placodes.

"Starburst" cholinergic amacrines specify the response of direction-selective ganglion cells to image motion. Here, development of cholinergic amacrines was studied in the tree shrew Tupaia belangeri (Scandentia) by immunohistochemistry with antibodies against choline acetyltransferase (ChAT) and neurofilament proteins. Starburst amacrines expressed ChAT much earlier than previously thought. From embryonic day 34 (E34) onward, orthotopic and displaced subpopulations segregated from a single cluster of immunoreactive precursor cells. Orthotopic starburst amacrines rapidly took up positions in the inner nuclear layer. Displaced starburst amacrines were first arranged in a monocellular row in the inner plexiform layer, and, with a delay of 1 week, they descended to the ganglion cell layer. Conversely, dendritic stratification of displaced amacrines slightly preceded that of orthotopic ones. Starburst amacrines expressed the medium-molecular-weight neurofilament protein (NF-M) from E34 to postnatal day 11 (P11) and coexpressed alpha-internexin from E36.5 to P11. Consequently, neurofilaments composed of alpha-internexin and NF-M may stabilize developing dendrites of starburst amacrines. During the first 2 postnatal weeks, subpopulations of anti-NF-M-labeled ganglion cells costratified with the preexisting dendritic strata of starburst amacrines in the ON sublamina, OFF sublamina, or both. Hence, anti-NF-M-labeled ganglion cells may include direction-selective ones. Thereafter, NF-M and alpha-internexin proteins disappeared from starburst amacrines, and NF-M immunoreactivity was lost in the dendrites of ganglion cells. Our findings suggest that NF-M and alpha-internexin are important for starburst amacrines and ganglion cells to recognize each other and, thus, contribute to the formation of early developing retinal circuits in the inner plexiform layer., ((c) 2007 Wiley-Liss, Inc.)

Epibranchial placodes and rhombencephalic neural crest provide precursor cells for the geniculate, petrosal, and nodose ganglia. In chick embryos and in Tupaia belangeri, apoptosis in rhombomeres 3 and 5 helps to select premigratory precursor cells and to segregate crest cell streams derived from the even-numbered rhombomeres. Much less is known about the patterns and functions of apoptosis in epibranchial placodes. We found that, in Tupaia belangeri, combined anlagen of the otic placode and epibranchial placode 1 transiently share a primordial low grade thickening with post-otic epibranchial placodes. Three-dimensional reconstructions reveal complementary, spatially, and temporally regulated apoptotic and proliferative events that demarcate the otic placode and epibranchial placode 1, and help to individualize three pairs of epibranchial placodes in a rostrocaudal sequence. Later, rostrocaudal waves of proliferation and apoptosis extend from dorsal to ventral parts of the placodes, paralleled by the dorsoventral progression of precursor cell delamination. These findings suggest a role for apoptosis during the process of neuroblast generation in the epibranchial placodes. Finally, apoptosis eliminates remnants of the placodes in the presence of late invading macrophages.

Whether rhombomere-specific patterns of apoptosis exist in the developing hindbrain of vertebrates is under debate. We have investigated the sequence of apoptotic events in three-dimensionally reconstructed hindbrains of Tupaia belangeri (8- to 19-somite embryos). Apoptotic cells were identified by structural criteria and by applying an in situ tailing technique to visualize DNA fragmentation. Seven rhombomeres originated from three pro-rhombomeres. Among pre-migratory neural crest cells in the dorsal thirds of the neural folds, the earliest apoptotic concentrations appeared in the developing third rhombomere (r3). Dorsal apoptotic maxima then persisted in r3, extended from r3 to r2, and also arose in r5. Transverse apoptotic bands increased the total amount of apoptotic cells in odd-numbered rhombomeres first in r3 and, with a delay, also in r5. This sequence of apoptotic events was paralleled by an approximate rostrocaudal sequence of neural crest cell delamination from the even-numbered rhombomeres. Thus, large-scale apoptosis in r3 and r5 helped to establish crest-free zones that segregated streams of migrating neural crest cells adjacent to r2, r4, and r6. The sequence of apoptotic events observed in the dorsal thirds of rhombomeres matches that reported for the chick embryo. Other shared features are apoptotic peaks in the position of a circumscribed ventricular protrusion of fusing parts of the neural folds in r1 and r2, and Y-shaped apoptotic patterns composed of apoptotic maxima in the dorsal and lateral thirds of r1, r2, and r3. These rhombomere-specific patterns of apoptosis may therefore represent a conserved character, at least in amniotes.

The patterns of cell death and of macrophages were investigated in the forebrain and eyes of the tree shrew Tupaia belangeri during five phases of optic cup formation. Seventeen embryos were studied. Three- dimensional reconstructions were made from one embryo of each phase. In phase 1 (V-shaped optic evagination) a midline band of cell death passes through the closing anterior neuroporus. From phases 2 (optic vesicle) to 5 (far-advanced invagination) the midline band of cell death extends in the dorsal wall of the forebrain to its rostral pole and, further, into its ventral wall. At the approximate future position of the optic chiasm this ventral pycnotic area, predicted but so far unidentified by others, is connected to a previously described second band of cell death passing through the optic anlagen. Recently, evidence has been presented that chicken embryos develop holoprosencephaly and cyclopia when ventral forebrain structures are lost secondary to experimentally induced apoptosis. Our findings in Tupaia suggest that, in cases of spontaneous malformations of this kind, such an atypical pycnotic area in the ventral telencephalon might result from the defective regulation of cell death processes during optic cup formation. In the forebrain and eyes of Tupaia, the occurrence of bands of cell death precedes the appearance of the earliest intraepithelial macrophages. From phase 3 (onset of invagination) onwards almost all of them are concentrated along the band of cell death.

The morphogenesis of the megamitochondria in the retinal cones of prenatal, young postnatal and adult tree shrews (Tupaia belangeri) was studied by transmission electron microscopy and three-dimensional reconstruction techniques. The initial assembly of the supranuclear cone mitochondria and their subsequent migration towards the developing inner segment conform to the morphogenetic pattern known from other mammals. Within the first postnatal week, however, a marked increase in both the number of the cristae and the matrix density occurs in the inner segment mitochondria of Tupaia. These mitochondria then grow, initially exhibiting a basal-to-apical size-gradient. In the 17-day-old Tupaia, this gradient is superseded by a radial size-gradient that, in addition to the single apical megamitochondrion, is characteristically found in the adult Tupaia. The number of megamitochondria remains almost constant from day 12 of postnatal ontogenesis to the adult stage. Each megamitochondrion consists of an apically located body from which several long processes project towards the base of the inner segment. In the older stages, the number of small mitochondria that most probably have budded off from the megamitochondrial processes clearly increases. We consider that megamitochondria in the cone inner segments of Tupaia arise by the growth of a single mitochondrion and not by the fusion of smaller mitochondria.

The morphogenesis of the manubrium sterni was studied in a series of dated embryos of Tupaia belangeri. In addition to the sternal bands, the "paired suprasternal Anlage" takes part in the shaping of the manubrium sterni as reported by Klima (1968) for other mammals. It forms skeletal elements that mediate between the clavicle and the manubrium: the sternocalvicular ligament and the paired prominence on the dorsal surface of the manubrium, which underlies the clavicles. The paired prominence corresponds to the praeclavium present in some therians. Very probably, the discus articularis of the sternoclavicular articulation of some primates can be attributed to the suprasternal Anlage. There was, however, no indication that the ossa suprasternalia of primates develop from the suprasternal Anlagen: In Tupaia these Anlagen do not form the cranial part of the manubrium. Klima's "unpaired Anlage" develops differently in Tupaia than in other therians. It consists of connective tissue and is not integrated into the manubrium. It presents an insertion surface for the M. pectoralis major, which shifts its origin onto the manubrium, after the sternal bands have fused. The homology of the "unpaired Anlage" and the "pars chondralis interclaviculae" is doubtful.

The ontogeny of the mandibular joint and the middle ear region was studied in Tupaia javanica, Microcebus murinus, Nycticebus coucang and Tarsius bancanus. During development, a passage connection was found between the mandibular condyle and Meckel's cartilage that is produced by the primordium of the lateral pterygoid muscle. The articular disk is formed separately, and it appears later in development. A theory is presented on the interpretation of these findings.

The therian scapula was until now thought to show very primitive features during early morphogenesis, as are found in the scapula of adult monotremes (elevated position of the scapula, lack of a spina and a fossa supraspinata, laterally directed cavitas glenoidalis). A morphogenetic study of the scapula of Tupaia belangeri has proved some of these assumptions to be wrong. The scapula undergoes a tilting which shifts its angulus articularis cranially, but no descent of the scapula could be found. The supraspinous fossa, which was supposed to develop very late in ontogeny from the anterior border of the scapula (Lewis 1902, Cheng 1955), is present in Tapaiai from the start. Part of it ossifies in membrane. The scapular spine does not develop as a cartilaginous outgrowth from the anterior border, but is formed mainly as an appositional bone along the lateral surface of the scapula. The glenoid cavity and the humerus are initially directed laterally. They attain their definitive form after the heart has migrated downward and the arms have been adducted. This represents a true plesiomorphous character state in therian ontogeny.

Studied the morphogenesis of the Fenestra rotunda and of the Aquaeductus cochleae in a series of 23 dated embryos and postnatal stages of Tupaia belangeri. The ontogeny of the Fenestra rotunda is the result of the caudal growth of the Processus recessus (DE BEER 1937). The Processus arises from the caudal ridge of the floor of the cochlear part of the otic capsule. On the 28th d of ontogeny (the gestation period of Tupaia belangeri is 43 d), it is fused with the lateral edge of the parachordal plate. On the 40th d, the Processus recessus joins the ventral surface of the canalicular part of the otic capsule, which develops a small cartilaginous process to meet it. In Tupaia, the Processus recessus is a large cartilaginous plate in a nearly horizontal position. It does not reach the plane of the Foramen perilymphaticum. The Processus recessus can be regarded as a part of the parachordal plate that was shifted laterally together with the Recessus scalae tympani by the enlargement of the cochlear part of the otic capsule in the ancestors of living mammals. The Processus forms the floor of the Aquaeductus cochleae, by which the laterally shifted Recessus scalae tympani of mammals remains connected with the cranial cavity. The Aquaeductus cochleae contains the Ductus perilymphaticus connecting the Cavum perilymphaticum of the inner ear with the Cavum leptomeningeum. The Fenestra rotunda of mammals is homologous with the lateral aperture of the Recessus scalae tympani of reptiles. In some mammals (e.g. Micropotamogale), the Membrana tympani secundaria spans the lateral aperture of the Recessus scalae tympani, as in many reptiles. Both the Membrana tympani secundaria of reptiles and that of mammals are homologous. Secondarily, in a large number of therian mammals (e.g. Myotis [Frick 1952]), the tympanic cavity extends into the Recessus scalae tympani displacing the Membrana tympani secundaria medially from the lateral aperture of the Recessus scalae tympani (= Fenestra rotunda of mammals) and even into the plane of the Foramen perilymphaticum. Thereby the Fossula fenestrae rotundae is formed, which in bounded medially by the Membrana tympani secundaria.

Development of the epicardium was studied in embryos of Tupaia belangeri from the 13th to 15th day of ontogeny. The greater part of the epithelium of the epicardium does not differentiate locally from the myoepicardium (cardiac splanchnopleure, splanchnic mesoderm), but rather from the coelomic epithelium of the septum transversum. The myoepicardium of the future atria and ventricles differentiates into myocardial cells only. On ontogenetic day 13, bulbar protrusions (the "villi" of Kurkiewicz 1909) are formed on the surface of the septum transversum and extend into the pericardial cavity, primarily between the sinoatrial and the ventricular regions of the embryonic heart. These protrusions are covered by flattened interdigitating cells, and they are filled with intercellular fluid of the mesenchyme of the septum transversum. Many mitoses are found among the cells. From these protrusions free vesicles are formed which are discharged into the pericardial cavity. The vesicles attach to the surface of the myoepicardium, i.e. to the developing myocardial cells. The vesicles open, and their cells spread out onto the surface of the heart to form the primary epicardium. This process begins on the dorsal surface of the heart, close to the protrusions of the septum transversum, there are, however, further isolated patches of primary epicardium in other regions of the surface of the heart. After the epicardial cells have settled onto the myocardium, mitoses become rare among them. On day 15, most of the myocardium is coated by the primary epicardium and the protrusions on the septum transversum disappear. A "bare" myocardium, as found on ontogenetic days 12 and 13 in Tupaia, might be a primitive (plesiomorphic) condition among chordates. In adult Branchiostoma, the coelomic epithelium which coats the contractile blood vessels had been found to differentiate into muscle cells that remain uncoated on the side facing the coelomic cavity (Franz 1933; Joseph 1914, 1928).

The development of the heart of Tupaia belangeri from the first endothelial-lined lumina to the cardiac loop is described in 20 embryos with 2 to 14 somites, from ontogenetic days 11 and 12. Bilateral endocardial tubes transporting blood are found in the 8-somite embryo; in the middle cardiac plate, angioblasts and angiocysts are located between them. In the 9-somite embryo, formation of the cardiac loop has started, the endocardial tubes approach each other closely, most of the angiocysts have been incorporated by the expanding endocardial tubes, and fusion of the endocardial lumina has started in the cono-truncal area. Apparently, much of the endocardial cardiac loop found in the 9-somite embryo has been produced by the disproportionate lengthening of a segment of the endocardial tubes, which is very short in the 8-somite embryo. In the 13-somite embryo the endocardial tubes have largely fused, but tube-like strands of endothelia, remnants of the original endothelial walls separating them, form a "palisade" and mark the original boundary between them. Myoepicardial differentiations of the splanchnopleure begin separately on both sides of the embryo and gradually spread craniad until they coalesce in the midline, in front of the anterior intestinal portal. The caudal portions of the endocardial tubes with initial myoepicardial and cardiac jelly differentiations do not contribute to the definitive heart. The anterior intestinal portal is very broad in Tupaia. Contradictions in the literature as to the bilaterality of cardiac primordia of eutherian mammals are discussed. The hypothesis is developed that bilateral endocardial tubes and bilateral myoepicardial differentiations of the splanchnopleure develop in species with a large yolk-sac, relatively late closure of the foregut, and a broad anterior intestinal portal (e.g., Tupaia, ferret, and cat, etc.). This is probably the primitive condition in eutherian mammals. In species with a small yolk-sac and/or reversal of germ layers (man, rodents), the foregut and anterior intestinal portal are formed earlier, the heart primordium reaches its median position ventral to the foregut in the angiocyst-stage, and the first endocardial lumina appear close to the midline. In these species, the primordium of the endocardium seems to be plexiform and without clear evidence for bilaterality.