STUDIES ON NERVE FIBER OUTGROWTH
IN MOTONEURON POPULATIONS
OF THE DEVELOPING
CHICK EMBRYO
BY
DENISE BETH WAYNE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988
Copyright 1988 by
Denise Beth Wayne
'Sometimes the magic works'
Thomas Berger
ACKNOWLEDGEMENTS
I would like to thank the members of my supervisory committee, Drs. Walker, West, and Ulshafer for their patience and fortitude. I would like to thank my advisor, Dr. Marieta Heaton for the many opportunities and for taking me in twice.
I would like the many other people who freely offered their time and resources particularly Dr. Mohan Raizada, Dr. Michael Young and Bill Creegan.
Last but not least, I would like to thank my family and friends who have given me limitless encouragement and support through it all.
iv
TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS ....................................... iv
ABSTRACT ............................................... vii
CHAPTERS
I. GENERAL INTRODUCTION ......................... 1
Theoretical Considerations................. 2
Mechanical Cues ......................... 3
Electrical Cues ......................... 11
Chemical Cues............................ 13
Axonal Elongation........................ 37
Cue Distribution Theories................ 41
Peripheral Nerve Patterning.............. 44
II. THE PATTERN OF EXTRAOCULAR INNERVATION BY THE OCULOMOTOR NUCLEUS OF THE CHICK........ 57
Introduction. ........................... 57
Methods .......o......................... 59
Results. ...................... .......... 61
Discussion .............................. 71
III. OCULOMOTOR DEVELOPMENT IN THE CHICK FOLLOWING EXPERIMENTAL REMOVAL OF
THE TARGET MUSCLES. .......................... 78
Introduction. ........................... 78
Methods................................. 82
Results. .................... ............ 84
Discussion. ............................. 97
IV. THE EFFECT OF SERUM AND DEFINED MEDIUM CONSTITUENTS ON NEURITE GROWTH FROM
EARLY NEURAL TUBE EXPLANTS.................. 105
Introduction ............................ 105
Methods ................................. 107
Results ................................. 109
Discussion.............................. 121
V
V. SPECIFIC RESPONSIVENESS OF CHICK TRIGEMINAL MOTOR NUCLEUS EXPLANTS
TO TARGET-CONDITIONED MEDIA................... 128
Introduction ............................ 128
Methods ................................. 130
Results ................................. 135
Discussion .............................. 140
VI. THE ONTOGENY OF SPECIFIC RETROGRADE TRANSPORT OF NERVE GROWTH FACTOR (NGF) BY MOTONEURONS
OF THE BRAINSTEM AND SPINAL CORD. ............. 145
Introduction ............................ 145
Methods ................................. 147
Results ................................. 150
Discussion .............................. 167
VII. THE RESPONSE OF CULTURED TRIGEMINAL AND
SPINAL CORD MOTONEURONS
TO NERVE GROWTH FACTOR ....................... 182
Introduction ............................ 182
Methods ................................. 185
Results ................................. 191
Discussion .............................. 201
VIII. GENERAL DISCUSSION .......................... 212
Nerve Development ....................... 212
REFERENCES ............................................. 225
BIOGRAPHICAL SKETCH .................................... 249
vi
Abstract of Dissertation Presented to the Graduate School
of the University of Florida
in Partial Fulfillment of the Requirements for
the Degree of Doctor of Philosophy
STUDIES ON NERVE FIBER OUTGROWTH
IN MOTONEURON POPULATIONS
OF THE DEVELOPING
CHICK EMBRYO
By
DENISE BETH WAYNE
April 1988
Chairman: Marieta B. Heaton Major Department: Medical Sciences (Neuroscience)
The organization of motoneurons in the mature oculomotor nucleus of the chick brainstem was examined. The retrograde label horseradish peroxidase was injected into each of the four extraocular muscles innervated by the oculomotor neurons. The results established that the segregation of motoneurons into subnuclei reflects the particular muscle which that cell group innervates.
The course of oculomotor nerve outgrowth and subnuclear organization were examined following embryonic removal of extraocular target muscles. It was observed that (1) the path of nerve outgrowth was not a function of vii
the presence of appropriate target muscles and (2) the nucleus forms appropriate morphological subnuclear groups in the absence of target muscle. The organization of the motoneurons is not initiated by contact with specific muscle but seems to be intrinsic to the motoneuron population. The disruption of the path of nerve outgrowth in the presence of muscle remnant suggests that there are multiple cues distributed along the route of nerve travel which guide it to the target.
An initial tissue culture study compared neurite growth from chick trigeminal neural tube explants in defined and serum-supplemented media. Subsequently, neurite outgrowth in the presence of medium conditioned by age-matched appropriate (jaw) and inappropriate (limb) target muscle was quantified. Medium conditioned by appropriate muscle produced significantly greater outgrowth than both medium conditioned by inappropriate muscle and unsupplemented control medium. This indicates a more specific effect of muscle produced factors on neurite outgrowth than has previously been reported.
The final two studies investigated the role of nerve growth factor (NGF) in the development of trigeminal and spinal cord motoneurons. The specific retrograde transport of NGF by motoneurons transiently during early development was demonstrated following peripheral injections of 125INGF. Additionally, early dissociates of the motoneuron viii
populations responded to NGF in vitro. The quantity of neurite and the rate of neurite initiation were enhanced in the presence of NGF in spinal cord and trigeminal motoneurons, respectively.
ix
CHAPTER I
GENERAL INTRODUCTION
Until the turn of the century the catenary theory had maintained that the nervous system was one continuous syncytium, a single sprawling web. As it became clear that the nervous system was composed of individual elements, the question which emerged was how the appropriate pattern of connections was faithfully reproduced during the course of development. Harrison's (1910, 1912) observations on isolated neurons in culture established unequivocally that the axon was a process extended from the neuronal cell body. During in situ development, the axonal process must be elaborated over long distances, following a patented course and reaching a specific target.
The cellular mechanisms which guide axons and result in the formation of specific connections have been the subject of many theoretical propositions and experimental investigations. The knowledge and insights provided by these efforts establish the foundation upon which the present work is based. The discussion which follows considers previous works which have examined: the nature and role of specific factors in axonal outgrowth, the organization of nerve pathways, the mechanisms whereby an axon elongates and, more
1
2
specifically, the development of peripheral nerves. The experimental work in the present dissertation is concerned with the organization and development of chick motoneuron populations of both the brainstem and spinal cord. The influence of target-derived factors on motoneuron process growth is of particular interest.
Theoretical Considerations
A unified axon guidance theory provides a conceptually useful model. It is, however, a simplistic reduction of a complex process. Unlike a concrete highway which is a fixed structure, the elements which contribute to form the nerve path are not immutably bonded together. Thus, the pathway for nerve growth is not an entity itself but the juncture of various factors which create a favorable path for outgrowth. The pathway is likely to be the product of multiple factors which, additionally, are likely to be changing continually along the course of the pathway. The influence of any factor may change in intensity as the axon grows and influential factors may be added or deleted from the repertoire. Thus, the pathway is quite likely a dynamic structure and this must be kept in mind when considering the significance of any single theory or factor.
Whereas the pathway provides the extra-axonal cue to growth, the axonal sensitivity is also critical to the selective growth response. The characteristics of the axonal membrane which may mediate a response will be discussed
3
later. The extra-axonal cues and the distribution of cues will be addressed first.
The types of cues which might be available to the axon in its growth can be categorized as: mechanical, electrical and chemical. Whereas these cues have been separated for clarity in the list above and the discussion below, within the realm of the developing nervous system they must certainly act in concert.
Mechanical Cues
Contact Guidance
A physical substrate for axonal elongation is an elemental requirement for the elaboration of neuronal processes. Harrison's (1910, 1912) pioneering work in tissue culture demonstrated that neurons elaborated processes only while in contact with a solid support. Neural explants that were suspended in a hanging drop of salt solution with no surfaces available for contact did not extend processes. With a solid surface available, the fibers grew along the path provided and thereby mirrored the pattern of the solid. This simple thigmotaxis of nerve fibers observed by Harrison was later expanded by Weiss. He found that fibers not only grew along the path described by a solid substratum but within a continuous solid substratum they were directed along the lines of organization within the substratum (Weiss, 1934). Thus, fibers were observed to grow along the lines of tension applied to a drying plasma
4
clot. This "contact guidance" described by Weiss has later been found to occur at the microscopic level. For example, neurite growth on a collagen substratum occurs preferentially along the axis of polymerization of the collagen fibrils (Ebendal, 1976).
Weiss' (1934) accumulated observations argued strongly against chemical or electrical trophic influences directly affecting outgrowth from neurons. He observed no directed growth in response to diffusible substances from target tissues. In addition, he argued that physical and electrical environmental influences oriented fiber growth by acting indirectly on the organization of the substratum. Weiss maintained that the mechanical ordering of the environmental substratum was the single denominator common to varied influences in directing axonal outgrowth. Invoking contact guidance to explain all forms of directed neurite outgrowth has proven to be a provincial view of the range of factors to which the growing neurite responds.
Neuronal processes have been observed to associate with non-neural cells in vitro (glia; Grainger and James, 1970; glia and fibroblast, Fallon and Raff, 1982) and in vivo (glia; Silver et al., 1982). Although the cells do provide a physical substrate for axonal growth, the specificity of the associations suggests that the neuron-cell contact is more complex than the simple provision of a solid substrate for growth. Spinal cord and retinal neurites displayed minimal
5
growth on a monolayer of fibroblast cells but extended lengthy and abundant processes on a monolayer of astrocytes (Fallon and Raff, 1982). The cell surface rather than some diffusible molecule produced by the cell has been indicated as the mediator of this growth response.
In vivo, Silver et al. (1982) examined the growth of callosal fibers which cross between the hemispheres. In embryonic mice the growing fibers are preceded in time by the appearance of a pathway of glia which form a band of cells between the hemispheres (just rostral to the lamina terminalis). This band or "sling" of glial cells is a transient structure which disappears after the pioneering fibers of the corpus callosum cross into the opposing hemisphere. In genetically acallosal mice this transient glial bridge never appears. Disruption of the bridge in normal mice by embryonic surgery results in an acallosal brain with the formation of large neuromas bilaterally. If, in addition to the interhemispheric cut of the glial bridge, a substitute inorganic bridge is implanted, the callosal fibers cross only if the glial cells first migrate across the inorganic bridge (Silver, 1983). The normal configuration of glial cells is, thus, necessary for the continued growth of the callosal fibers.
The association of growing axons and neurites with nonneural cells reported above demonstrated a definite and interesting specificity. Similarly, the fasciculation of
6
neural processes with other axons in vivo and in vitro exhibits some degree of selectivity and is discussed below.
The axonal processes of neurons may provide a path for growth of other fibers. Speidel (1933) was able to observe in vivo axonal growth in superficial regions of the tail of live tadpoles. He observed growth cones advancing along neuronal fibers that had preceded them. Ingrowing sensory axons in Drosophila wing fasciculate along the earliest growing pioneer fibers in growth to the CNS (Palka and Ghysen, 1982). Following experimental implant of barriers in chick embryos, the deflected trigeminal nerve sometimes fasciculates along the fibers of the abducens nerve (Moody and Heaton, 1983c). The examples are numerous: axons can and do grow along the tracts provided by other fibers. However, axons discriminate and do not grow along any and all fibers available. Ectopic axons of the auditory nerve in Xenopus cross perpendicular to optic tract axons and yet retain their own distinct orientation (ConstantinePaton, 1983). The fasciculation of axons is also selective in vitro (Nakajima, 1965). Neurites of retinal and sympathetic explants go to dramatic lengths to avoid contact when confronted with the other, turning, ceasing elongation or elongating on another plane (Bray et al., 1980).
Axons may be guided as much by a region which prohibits growth as by a region which favors growth. Olfactory tract axons growing rostro-caudally and optic tract axons growing
7
latero-medially both pass in close proximity to the diencephalic-telencephalic (DT) border but fail to cross it. Close inspection of the DT junction in the embryonic mouse and chick reveals the presence of a dense core of cells with little extracellular space (Silver, 1984; Silver et al., 1987). This "glial knot" region seems to prohibit growth in two respects: (1) it provides a physical barrier to growth and (2) it deprives axons of the neural-cell adhesion molecule (N-CAM) and extracellular matrix proteins present in adjacent areas where growth occurs (Silver et al., 1987). Both the physical barrier and the specific molecular characteristics of the substratum appear to be factors in the inhibition of fiber growth across this region.
In sum, the normal surface character of cells and axons appear to provide specific substrates for axonal elongation, not merely solid support.
Extracellular Space
An alternate type of mechanical guidance has been
suggested by Singer's Blueprint Hypothesis (Singer et al., 1979). Observations on the developing and regenerating spinal cord of the newt reveal the existence of extracellular channels before axonal ingrowth. These spaces are bounded by epithelial cells. It is suggested that a channel pattern arises during ontogeny and contains chemical "trace pathways" which together guide tract formation. One candidate for the role of a "trace" molecule is the neural
8
cell adhesion moleculue (N-CAM) which seems to play a role in axonal fasciculation (Rutishauser et al., 1978b) as well as neuronal adhesion to glial (Silver and Rutishauser, 1984) and muscle cells (Grumet et al., 1982; Rutishauser et al., 1983). The N-CAM has been found to increase in intensity along the endfeet of neuroepithelial cells in Xenopus spinal cord prior to the axonal ingrowth which occurs in this region (Balak et al., 1987). Components of the extracellular matrix as well as other surface associated molecules may be additional components of the "trace pathway."
A fixed blueprint of axon channels would not account for the normal growth of axons which enter the central nervous system at abnormal locations (Constantine-Paton, 1983; Giorgi and van der Loos, 1978). In these cases, information for appropriate growth seems to be widely distributed through the nervous system and not available only in rigidly select channels.
The Blueprint hypothesis may accurately represent the
existence of general pathways available for growth early in the development of long fiber tracts. The axon channels and associated "trace" molecules may be only one component of a more complex system which leads to appropriate connectivity, and the specific nature of the components may vary throughout the nervous system.
9
Spatiotemporal Sorting
Another means whereby specific patterning is
hypothesized to arise is by spatiotemporal sorting of axons. Thus, the arrangement of axons within a fiber bundle would be determined by the order of cell differentiation and axonal outgrowth. Sequentially arising axons would assume a position adjacent to those formed immediately prior to them, thus ordering axons within a nerve. The same would be true at the target with the first arriving axon occupying the nearest synaptic vacancy. The majority of data supporting this mechanism for patterning comes from studies of the developing retinotectal system. Retinal cells are generated in a central to peripheral sequence and a corresponding relationship is maintained by their axons in the retinal fiber layer. It has been suggested (Rager, 1980) that the preservation of axonal order in the optic nerve and tract does not require sorting. Contact guidance may serve to passively funnel axons through intercellular channels and "oriented glial partitions." At the target site the first arriving ganglion cell axons seem to bypass vacant tectal target areas to actively seek specific tectal sites. However, Golgi analysis reveals that the pattern of ganglion axon innervation of the tectum seems to follow a temporal gradient provided by the maturation of tectal cells (Rager and von Oeyenhauser, 1979). The evidence is consistent, then, with a spatiotemporal gradient of cell differentiation
10
producing order within this system. This evidence does not exclude the possibility, however, that there are other active cues. There might also be a chemospecific interaction of fibers with either oriented glial cells or tectal cells.
The spatiotemporal theory of nerve sorting has also been applied to peripheral nerve patterning (Horder, 1978). Some of the evidence supports a selection mechanism for axons in the periphery based on position, but there is also much data which indicates the presence of other, more specific mechanisms. Experimental inversion of the chick limb bud along the dorso-ventral axis has resulted in an incorrect pattern of innervation of limb muscles (Summerbell and Stirling, 1981; Whitelaw and Hollyday, 1983c). The nerve fibers apparently maintain their old position and follow the new pathway in a non-selective manner. The axons are unable to re-direct their growth and reach the appropriate target. Ferguson (1983) has reported appropriate innervation using the same experimental paradigm with the exception that she rotated the limb more proximally. When the rotation included the proximal nerve plexus the fibers were able to respond to the path changes. The axons shift their position within the plexus to innervate the normally appropriate target muscle. This active rearrangement of fibers is evidence against a merely passive channeling of fibers. In normal chick development too, the fibers innervating the limb muscles selectively re-organize in the plexus and nerve fascicles as
they course to the target (Lance-Jones and Landmesser, 1981a; Ferguson, 1983). The axons do not maintain a consistent position within the nerve along their proximal to distal course. This specific rearrangement of axons in both normal and experimentally altered situations contradicts what would be expected with a passive spatiotemporal sorting.
The physical channeling of non-specified axons may be an important part of normal fiber growth but this process alone does not seem sufficient to account for the specificity of nerve growth. Again, evidence in support of one guidance mechanism does not necessarily exclude the possibility that there are other mechanisms acting concurrently.
Electrical Cues
An early theory which proposed a role for endogenous electrical activity in mediating neuronal development was Bok's concept of stimulogenous fibrillation (1915). This theory suggested that as growing fiber tracts passed neuroblasts, electrical activity in the growing axons stimulated axonal outgrowth from the neuroblasts. The theory of neurobiotaxis elaborated by Ariens-Kappers (1917, 1921) was an ambitious attempt to explain much of neuronal development in terms of the influence of currents on different elements of the neuron. It was purported that axons were directed away from, while dendrites and cell bodies moved towards, the extracellular negative charge
12
(cathode) created by growing fiber tracts. This polarization of the neuron as well as the formation of multiple dendrites versus a single axon were all attributed to specific influences of current from fiber tracts on the neuron.
As early as 1920, Ingvar reported the orientation of
neurite growth in response to weak electric currents applied in culture. Weiss (1934), however, contended that the neurite orientation was due to the current's influence on the alignment of fibrils in the plasma clot substratum rather than to a direct effect on the neurites themselves. Marsh and Beams (1946) found that applied currents produced an orientation of the neurites independent of any orienting effect that the current might have on the plasma clot substratum. At sufficient current densities neurite outgrowth was enhanced toward the cathode and suppressed toward the anode. Jaffee and Poo (1979) recently replicated this experiment and suggested that the directed growth was caused by the electrophoretic redistribution of membrane glycoproteins under an applied current. The fluoresceintagged lectin, concanavalin A (Con A), has been used to label surface glycoproteins of isolated neurons (Patel and Poo, 1982) and muscle cells (Poo and Robinson, 1977) in culture. Con-A labeling has revealed a concentration of surface membrane glycoproteins on the cathodal cell face following exposure to an electric field. Additionally, the presence of Con A in the medium during neuronal exposure to
13
an electric field blocks both the directed growth of neurites and the concentration of surface Con-A receptors on the cathode directed membrane (Patel and Poo, 1982). These surface glycoproteins could be important to neurite elongation by acting in membrane-substratum adhesion or as receptors to growth factors in the medium. Endogenous electrical currents have been measured in the region of the developing limb of Xenopus (Robinson, 1983). There are also electrical fields produced by non-neural cells in vivo. These fields are present during early development of the nervous system and could possibly influence nerve development in a manner similar to the growth effects seen in vitro. An alternate interpretation is that electrical current in vitro is an artificial means of stimulating the neurite's normal growth mechanism. The appropriate stimulus in vivo may actually be binding of extracellular molecules to these Con-A binding surface glycoproteins which then results in enhanced neurite growth at the binding locus.
Chemical Cues
It is possible that molecular cues external to the
neuron may serve to guide axonal growth, thus manifesting a chemotactic phenomenon. The salient molecular cues could reasonably exist in several forms, e.g., freely diffusible molecules, substrate bound molecules, or molecules integral to a cell membrane. The neuronal response to chemotactic cues which results in a directed growth of neural processes
14
can theoretically be categorized in one of two ways. The response could be classified as a trophic or as a tropic response type. A trophic stimulus provides a nutritive or metabolically favorable climate for growth. This trophic stimulation of neural growth could produce an apparent selective path choice by increasing the amount or rate of growth in one direction thus favoring that path above others. A tropic response is characterized by a tendency to turn in response to a particular stimulus, that is, a qualitative change in orientation. It is clear that either chemotrophism or chemotropism could produce directed axonal growth along specific pathways.
A purely trophic stimulus may be thought of as affecting the quantity of neurite growth while a purely tropic stimulus produces only a qualitative change affecting the direction of neurite growth. This clarification is important in reaching an understanding of the cellular mechanism which responds to the cue. However, it is quite possible that a single molecule such as nerve growth factor (NGF) is capable of eliciting both a tropic and a trophic cellular responses. Early Theories
The role that molecular cues may play in the regulation of neural development has been discussed since the turn of the century (Ramon y Cajal, 1960). Weiss (1934) presented a spate of evidence which supported mechanical guidance of nerve outgrowth and contended that chemical and electrical
15
factors had no direct influence on nerve growth. Weiss was an influential proponent of the "resonance theory" which maintained that the establishment of connections in the nervous system was a non-selective event and that selectivity occurred later as a result of neural activity (Weiss, 1936). Sperry's studies on the recovery of function in the visual system of lower vertebrates after nerve section found a selective re-establishment of neural connections despite the functionally maladaptive behavior this yielded (Sperry, 1944, 1945; Attardi and Sperry, 1960, 1963). In these studies, the optic nerve of anurans was cut and the eye rotated prior to nerve regeneration. When vision was restored, the animals behaved as if the visual field had been rotated in precise accordance with the rotation of the eye. This inversion of the visual field was very specific as well an enduring. The result, a persistent maladaptive behavior, is inconsistent with the resonance theory which predicted that activity-induced specification would produce the appropriate functional adjustment over time.
Subsequent studies examined optic nerve regeneration in goldfish following optic nerve section combined with partial retinal ablations (Attardi and Sperry, 1960, 1963). The regeneration of fibers proved to be very specific both along the path of nerve growth and in the terminal site selected. The fibers followed the same tract to the tectum and reinnervated the appropriate tectal site, bypassing
16
inappropriate vacant sites. The chemoaffinity hypothesis elaborated by Sperry suggested that chemotactic recognition phenomena based on biochemical specificities guided nerve growth (Sperry, 1965). The highly selective nerve growth, along the appropriate pathway and to the appropriate synaptic site in the tectum, revealed by these experiments was consistent with a specific cellular chemoaffinity regulating the course of nerve growth.
While this regeneration data establishes the specificity of neural connections and is consistent with chemotactic regulation of growth, it does not directly evidence any molecular mediation of neural growth. Nerve Growth Factor (NGF)
The discovery of the nerve growth promoting factor, NGF, provided the first direct evidence of a molecular factor capable of influencing neural growth. Bueker (1948) first noted that in chick embryos with transplanted mouse tumors (sarcoma 180), the dorsal root ganglia innervated the tumor and these ganglia were enlarged. Levi-Montalcini and Hamburger (1951) pursued this observation and found that both sensory and sympathetic ganglia responded with fiber growth toward the tumor and an enlargement of the ganglion itself. In addition, it was not just those neurons in immediate contact with the tumor that showed the enhanced growth response. The growth of the tumor on the chorioallantoic membrane, separated from the embryo,
17
demonstrated unequivocally that the neurons were responding to some diffusible substance emanating from the tumor rather than to some contact mediated influence of the tumor (LeviMontalcini and Hamburger, 1953). The hypertrophy of the sensory and sympathetic ganglia was not due to the presence of an enlarged peripheral field presented by the tumor but due to the effect of a diffusible molecule produced by the tumor. Embryonic sympathetic and neural crest-derived sensory neurons respond to the presence of NGF with abundant fiber outgrowth both in vivo and in vitro (Levi-Montalcini et al., 1954). NGF provides a trophic stimulus to the neurons enhancing anabolic processes (Levi-Montalcini, 1982). In addition, Gundersen and Barrett (1979) have shown that the growth cones of sensory neurons growing in culture turn to follow a gradient of NGF provided by a nearby micropipette. The NGF, then, provides both a trophic and tropic stimulus to the growth of sensory and sympathetic fibers.
The history of NGF is a remarkable one (Levi-Montalcini, 1975). Both the existence of NGF itself and the identification of a rich source of the molecule resulted from serendipitous observations combined with keen intuition. The relatively high concentration of NGF in the male mouse submaxillary gland facilitated the characterization of the NGF protein complex which consists of three subunits referred to as a, B and T Of the three
18
subunits, the full biological activity resides in the B subunit with the role of the other two subunits still not clearly resolved. The B subunit of NGF is a Mr 26,500 dimer of two non-covalently associated peptide chains. At physiological concentrations the molecule is apparently present as a dimer and not as the dissociated monomer (Bothwell and Shooter, 1977).
The purification of mouse BNGF made possible the production of specific antibodies which blocked the biological activity of NGF in mice. The administration of the antibody to newborn mice resulted in an immunosympathectomy characterized by a virtually complete and enduring destruction of sympathetic neurons (see LeviMontalcini and Angeletti, 1968). Developmental studies of the effects of NGF deprivation have been hampered by the failure of antibodies against mouse NGF to block the activity of chick NGF in vivo. Some success in chronicling the effects of NGF deprivation during development in mammalian species has been accomplished by maternal transfer of antibodies to the embryo following auto-immunization of the mother to NGF (Johnson et al., 1980, 1983). With prenatal exposure to antibodies against NGF there is destruction of two neuronal populations, i.e., sympathetic and neural crest derived-sensory neurons. In the guinea pig, the maternal antibodies are not transferred to the embryo until the latter half of gestation (day 35 of 65-70 day
19
gestation period). The effects of NGF deprivation on early development, then, have not yet been determined. Recent efforts have been directed at finding an antibody that successfully inactivates NGF in the chick embryo which is accessible throughout development (Ebendal et al., 1986; Meier et al., 1986; Belew and Ebendal, 1986).
The NGF receptor was first characterized by binding kinetics which revealed a high affinity or slowly dissociating receptor (type I) and a low affinity or rapidly dissociating receptor (type II). Both receptor types have been found on NGF-responsive cells (sympathetic, Godfrey and Shooter, 1986; sensory, Sutter at al., 1979; PC12, Hosang and Shooter, 1985) while cell types that do not respond to NGF have only the type II receptor (mutant PC12, Green et al., 1986; Schwann cells, Zimmermann and Sutter, 1983). The type I receptor apparently mediates the biological response to NGF by internalization and subsequent transport to the cell body of the NGF-NGF receptor complex (see review, Stach and Perez-Polo, 1987). Enzymatic activity associated with the NGF receptor suggests that NGF binding may result in an active enzyme complex which may then have a role in affecting NGF action (Stach et al., 1986). There is some evidence that the two receptor types are identical in the NGF-binding moiety (Green and Greene, 1986). This is supported by the finding that there is apparently only one gene for the NGF receptor (Chao et al., 1986). The
20
differential effects produced by NGF binding could possibly result from an additional moiety associated with type I receptor (Hosang and Shooter, 1985) or differing interactions with membrane or cytoplasmic constituents (Schecter and Bothwell, 1981).
In the sympathetic nervous system, at least,
considerable evidence supports the notion that NGF acts as a retrograde messenger transported from the target organ to the innervating neurons. In vivo blockade of retrograde transport by colchicine application to the post-ganglionic nerves resulted in changes in the neurons of the superior cervical ganglion similar to the changes produced by NGF deprivation (Purves, 1976). Chemical destruction of sympathetic nerve terminals or blockade of axonal transport results in a rapid increase in the levels of NGF in sympathetic target organs with a corresponding decrease in the NGF content of the innervating sympathetic ganglion (Korsching and Thoenen, 1985). Thus, the NGF normally present in the sympathetic ganglia is derived from the target structures. Recent methodological advances have permitted the detection of low concentrations of NGF and the distinction of NGF-localization from NGF-production with probes which detect mRNANGF (Heumann et al., 1984; Whittemore et al., 1986). In adult animals, NGF production is clearly localized to sympathetic target structures and
21
not to the sympathetic ganglia (Heumann et al., 1984; Shelton and Reichardt, 1984; Bandtlow et al., 1987).
In addition to the adult structures, the production of NGF has been described in all Schwann cells of the neonatal rat, i.e. those which ensheath motor as well as sensory and sympathetic fibers (Bandtlow et al., 1987), and embryonic chick skeletal muscle (Hulst and Bennett, 1986). The apparent widespread presence of NGF in fetal rat central nervous system (Ayer-Lelievre et al., 1983) and the more recent demonstration of specific NGF-binding by chick motoneurons of the brainstem and spinal cord transiently during development (Raivich et al., 1985, 1987) suggests that NGF may be important in developmental interactions between the motoneuron and its target musculature.
The discovery and characterization of NGF has been important in providing substantiation that endogenous molecular substances could produce specific stimulation of neural growth. The recognition of NGF has stimulated the exploration for other growth factors and an interest in the role that these molecules may play in normal development. In this vein, the trophic effect that the target tissue, muscle, has on motoneuron growth has received much attention.
Tissue-derived Factors
The selectivity of a growth factor is an important aspect of its character which may determine its
22
effectiveness in mediating the formation of specific connections in vivo. Co-cultures of normal target tissues with neuronal explants has been shown to specifically enhance the neurite outgrowth from the explants (Ebendal and Jacobson, 1977; Pollack et al., 1981). The effect is specific in that target tissue type and target tissue age were important in the degree of enhancement produced. In vivo, the spatial (tissue type) and temporal (tissue age) availability of a growth factor may operate to produce specific effects during development. The diffusion of a molecule away from its source may produce a concentration gradient, and neurons in vitro have displayed the capability to direct growth to increased concentrations of the peptide NGF (Gundersen and Barrett, 1979).
Variations in the sensitivity of neural tissue to
available trophic and/or tropic factors could also produce selective growth during development. The ability of the neural explants to respond to target tissue has been shown to vary with age of the explant (Pollack and Muhlach, 1982). At least in this instance, the neural tissue seems to exhibit an intrinsic decline in the potential for growth with increasing age.
On the other side of this neuron-muscle trophic/tropic interaction is the array of factors produced by muscle. Similar to the changing growth capabilities of the neural tissue, there does seem to be a developmental change in the
23
effectiveness of muscle-produced factors in supporting neuronal survival and differentiation (Heaton and Kemperman, 1987). There is conflicting evidence regarding whether muscle is increasingly or decreasingly effective at producing a neurite promoting factor(s) with age (Nurcombe and Bennett, 1983; Eagleson and Bennett, 1986; Heaton and Paiva, 1986). The work from Bennett's laboratory suggests that motoneurons are initially dependent on an astrocyte produced factor for survival and later become dependent on a muscle derived factor (Eagleson and Bennett, 1986). The stimulation of neurite formation and enhanced survival produced by muscle conditioned medium (MCM) increases with muscle differentiation from myoblast to myotube (Nurcombe and Bennett, 1983). In contrast, other studies using explants of chick brainstem and frog spinal cord have shown that target from early stages which coincide with initial axon outgrowth in vivo produce greater enhancement of neurite outgrowth than older target tissue (Heaton and Kemperman, 1987; Pollack and Muhlach, 1981). Species (chick versus frog) and regional (spinal cord versus brainstem) differences may account for the observed differences in the neurite promoting effects of muscle-derived factors with age.
Smith and Appel (1983) found that extracts of neonatal rat skeletal muscle stimulate both neurite outgrowth and cholinergic activity in dissociated ventral spinal cord
24
neurons. The extract stimulation was specific for ventral cord neurons as demonstrated by the control levels of response from dorsal cord neurons. With muscle of increasing postnatal ages, the stimulation of neurite outgrowth by muscle extract decreased while the cholinergic activity was stimulated equally by all ages of muscle extract. There is apparently more than one specific developmental change in the trophic influence which muscle exerts on motoneurons. While the ability of muscle-derived factors to enhance neurite outgrowth appears to decline from mid-development onwards it is still apparent at birth. Several studies have found that purified populations of embryonic motoneurons respond to a factor derived from embryonic (day 11 and day 18 chick) or newborn (rat) muscle with enhanced neurite outgrowth (rat: Smith et al., 1986; chick: Calof and Reichardt, 1984; Dohrmann et al., 1986).
The selectivity of target-derived factors in enhancing neurite outgrowth has been more clearly delineated in vitro by the choice of trigeminal sensory axons for growth toward the specific region of epithelium normally innervated in vivo (Lumsden and Davies, 1986). The developmental specificity of these trophic effects strongly supports a role for these factors in developmental events in vivo. The factors which stimulate neurite growth have been characterized in vitro by the nature of the biological response and their physicochemical properties.
25
A number of analyses have revealed that conditioned medium (CM) generated by a variety of cell types is comprised of substrate-binding and soluble components (Adler et al., 1981). Collins (1978a) was the first to demonstrate that it was a substrate-bound component of heart cell conditioned medium (HCM) that promoted neurite outgrowth from ciliary ganglion neurons. This component remained bound after adsorbed to the substrate and the CM depleted of this component did not, by itself, stimulate neurite outgrowth. The neurite-promoting component which adheres to the substrate is, additionally, only effective when it is associated with the substratum (Adler and Varon, 1980). The binding of the factor to the substratum along with direct observation of an increased filopodial adherence to the substratum, led Collins to suggest that the neurite promoting factor in HCM acts by increasing the adhesivity of the substrate. Collins and Garrett (1980) then presented neurites with a choice of substrates coated with HCM or uncoated substrates and found that neurites followed the contours of the coated substrate and only rarely strayed onto the uncoated regions. This characterizes a tropic rather than trophic response because the neurites oriented to remain on the HCM-substrate rather than retracting or failing to survive on non-coated substrate.
The substrate-bound neurite promoting factor (NPF) has been isolated from both muscle conditioned medium (Collins,
26
1978a, 1978b; Collins and Garrett, 1980) and co-cultured premuscle tissue (Nurcombe and Bennett, 1983). This adsorbable component supports neurite outgrowth but not long-term neuronal survival. The non-substrate-bound component of CM, distinct from the adsorbable component, supports neuronal survival (Collins, 1978a; Adler and Varon, 1980; Nurcombe and Bennett, 1983) and promotes continued elongation of neurites once they have been initiated (Collins and Dawson, 1982).
It is possible that neural sensitivity to a trophic factor could be exogenously regulated by diffusible, substrate-bound or cell-attached molecular cues. The sensitivity of the response of sympathetic neurons to a trophic factor, NGF, can be altered by such external stimuli. Sympathetic neurons exhibit an increased sensitivity to NGF when the culture substrate has been preexposed to HCM (Edgar and Thoenen, 1982). The HCM alone has no effect on sympathetic neuron survival. Some component of HCM absorbs to the polycationic substrate and modifies the response of sympathetic neurons to NGF. This could conceivably occur by receptor induction or by selective survival of cells with previously silent NGF receptors. Weill (1983) has reported an induction of androgen receptors on dissociated spinal cord cells exposed to muscle CM in culture. It is possible that multiple trophic factors interact during development to stimulate and direct neurite
27
outgrowth. A trophic factor may act not to stimulate growth directly but to alter cellular sensitivity to another factor or stimulus.
Stimulation of neurite outgrowth from motoneurons has been produced by factors isolated from a variety of tissue sources including, but not limited to, muscle. Neurite promoting activity has been demonstrated by exposure of motoneurons to extracts of muscle tissue (Hsu et al., 1982, 1984; Smith and Appel, 1983) and medium conditioned by a variety of cell types [embryonic rat lung, fibroblast and muscle (Dribin and Barrett, 1980; 1982); embryonic chick muscle, liver, and skin (Henderson et al., 1981); bovine vascular endothelial cells, bovine vascular smooth muscle, bovine adrenal cortical cells, human skin fibroblasts, bovine corneal endothelial cells, C2-mouse skeletal muscle, PTK-1 kangaroo rat epithelium, A-431 human carcinoma, N-18 mouse neuroblastoma (Lander et al., 1982)]. These studies have demonstrated production of neurite promoting factors but not without some conflicting results. While some investigators found that chick embryo heart muscle (Collins, 1980) and rat lung, fibroblast and muscle tissue (Dribin and Barrett, 1980) all produced effective CM, Henderson et al. (1981) found that CM produced by chick embryo heart and lung did not enhance neurite outgrowth in day 41 chick neural tube dissociates. Attempts to reconcile these differences serves to underscore the difficulty in comparing results
28
derived from experiments using different species, different assays of neurite response, different substrates and different media for growth.
Characterization of Neurite-Promoting Factor
While the differing results in production of neurite promoting factors by various tissue types have yet to be reconciled, the characterization of the active factors isolated from the different sources is more consistent. The elimination of activity by exposure to trypsin (Dribin and Barrett, 1982; Henderson et al., 1981; Collins, 1978a; Smith and Appel, 1983) indicates that the factor is likely to be, or to contain, a protein. A majority of the biological activity consistently appears with a molecular weight around 50,000 daltons. More specifically, antibodies to NGF fail to block the neurite promoting activity, thus certifying that the active factor is not NGF (Dribin nd Barrett, 1980; Smith and Appel, 1983). The neurite promoting factor(s) was found to be negatively charged at neutral pH and insensitive to RNase, DNase, collagenase, and neuraminidase (sialic acid residues are not important to activity) (Dribin and Barrett, 1982; Lander et al., 1982). Lander et al. (1982) have provided a detailed characterization of the CM factor produced by corneal endothelial cells which is consistent with the results described above. The factor is identified as a heparan sulfate proteoglycan. Proteoglycans are polyanions which will bind well to positively charged
29
substrates. As in other studies, the protein portion of the molecule is vital to the activity. Enzymatic degradation of the protein or pre-exposing the substratum to only the polysaccharide portion of the proteoglycan (heparan sulfate or other purified glycosaminoglycans) yielded no neurite promoting activity. Further investigation showed that the neurite promoting activity in a variety of CM was immunoprecipitated with antibodies to the basement membrane glycoprotein laminin (Lander et al., 1985). Neurons which normally extend their processes in the peripheral nervous system exhibit enhanced neurite growth when cultured on a laminin substratum (Rogers et al., 1983). Laminin is present in vivo during development in the regions of nerve outgrowth in both the brainstem and spinal cord (Riggott and Moody, 1987; Rogers et al., 1986; see section on Molecular Path Cues). In vitro, anti-laminin antibodies inhibit neurite growth on a purified laminin substratum but fail to do so if a CM substratum is used (Edgar, 1985). This may be due to epitope differences in the purified laminin compared to the CM laminin so that antibody binds but is unable to block the region of the molecule important in the promotion of neurite growth. An antibody which is able to block the neurite promoting activity of CM factors suggests that the activity of the natural complex requires the association of laminin with heparan sulfate proteoglycan and possibly other components (Chiu et al., 1986). The nature of the laminin
30
and the associated molecules which comprise the neurite promoting activity present in CM produced by a variety of cells have yet to be fully resolved. Cell Surface Molecules
The action of diffusible tropic factors as well as the
efforts to characterize these factors have been described in preceding sections. Effective action of chemically mediated cues in neuronal development would also require that the neuronal cell membrane possess the capacity to respond appropriately to the factor. Some type of cell surface molecule must be present on the neuronal membrane to respond to the external molecular cue. The distribution of these membrane molecules would be expected to be restricted to certain developmental times, cell types, and cellular loci depending on their particular role in the development of the cell. Monoclonal antibodies and lectin labels have provided the means to examine specific cell surface characteristics.
A variety of plant lectins selectively bind to specific disaccharide residues. Binding to neuronal membranes in vivo and in vitro, lectins have been used to describe developmental (timed), topographical (cell type) and topological (cell locus) differences in cell surface glycoconjugates. The binding of several lectins to embryonic mouse retina at sequential stages has revealed a general increase in the intensity of label with development (Blanks and Johnson, 1983). Each lectin exhibited a characteristic
31
pattern of binding indicative of its affinity for a specific carbohydrate sequence (Lis and Sharon, 1977). Specific cells and retinal layers were differentially labeled with the lectins tested (Blanks and Johnson, 1983). Selective labeling of a sub-population of photoreceptors (likely cones) by peanut agglutinin and the outer but not inner segments of photoreceptors by three of the eight lectins was reported (Blanks and Johnson, 1983). This differential label evidences the distinct glycoprotein composition of these membranes.
The surface membrane features of different cell types
vary not only in the particular lectins which are bound but, also in the profile of binding by a particular lectin. Pfenninger and Maylie-Pfenninger (1981) provided quantification of lectin receptors by the binding of ferritin conjugated lectins to dissociated cells. Comparison of the binding to explants versus single dissociated cells reveals similar patterns indicating that the membrane integrity is restored to normal after the dissociation procedure. Spinal cord and superior cervical ganglion cells isolated in culture exhibited different patterns of lectin receptors which varied in several ways: the absolute receptor density varied at a given cell region, the local distribution of receptors varied (clustered versus homogeneous) and the profile of changes in receptor density between cell regions varied. Not only are there distinct
32
differences in surface glycoconjugates between different cell types but there are regional specializations of the neuronal membrane within a given cell. Evidence of this was provided by a comparison of the density of binding of each lectin among three distinct cellular regions: cell body, neurite shaft, and growth cone. The lectin receptors were equally distributed, increased or decreased in the sequence from perikaryon to growth cone. Each lectin demonstrated a distinct binding pattern on each cell type. The regional specialization of the neuronal surface membrane is confirmed as well as the distinct surface identity of different neuronal types.
Injections of radiolabeled glycoprotein precursors into a single identified Aplysia neuron has also provided evidence of regional differences in neuronal membrane glycoproteins (Ambron, 1982). Examination of the perikaryal and axonal membranes subsequent to these injections revealed different glycoproteins were specifically incorporated in the two regions. The functional role that surface glycoproteins may play in axonal outgrowth was assessed in goldfish retinal explants exposed to tunicamycin, a protein glycosylation inhibitor (Heacock, 1982). There was a concentration dependent inhibition of neurite outgrowth from the explants with exposure to tunicamycin. While protein synthesis was relatively unaffected, the neurite membrane produced was deficient in carbohydrate as evidenced by
33
lectin binding. This treatment may produce other undetected deficiencies in addition to depletion of cell surface glycoproteins. Nonetheless, the model does produce impaired neurite outgrowth and strongly suggests that cell surface glycoproteins are critical elements to this normal function.
The development and distribution of unique membrane characteristics have also been investigated with the application of monoclonal antibodies. Antibodies which recognize rat peripheral but not central neurons (Vuillamy et at., 1981) and, conversely, antibodies which recognize rat central but not peripheral neurons (Cohen and Selvandran, 1981) have both been described. The labeled antigens appear at distinct points in development yet the functional significance of this distinctive temporal sequence remains undeciphered. In the simpler nervous system of the leech, Hockfield and McKay (1983) described the distribution of axons and cell bodies labeled with several different antibodies. The axonal position in the interganglionic connectives delineated with these antibodies is consistent and predictable. Thus the axon, like the cell body, is specifically and reproducibly distributed in the nervous system. The distribution of perikarya and axons following single antibody administration suggests that the position of the cell body does not confer a similar positioning to the axon in the connectives. Quite possibly, different cell surface determinants dictate the position of
34
the cell body and the position of the axon. The three different antibodies that Hockfield and McKay (1983) applied all labeled different subsets of cells and axons. Specific labeling of subgroups of cells and axons has also been described in the mammalian CNS (cat: McKay and Hockfield, 1982). The molecular heterogeneity of cell membranes is thus a fundamental characteristic of the nervous system.
There is evidence to suggest that at least some of these cell surface antigens are important in the elaboration of processes by neurons. Henke-Fahle and Bonhoeffer (1983) have produced an antibody to chick embryo retinal cells which binds preferentially to retinal plexiform layers in vivo. When this antibody is applied in culture it inhibits neurite outgrowth from retinal explants. This effect is specific for retinal neurites as demonstrated by the lack of effect on outgrowth from dorsal root ganglion explants.
The simplicity of the insect nervous system along with the experimental removal of identified cells by laser ablation have permitted important observations to be made regarding specific cellular interactions during axonal outgrowth. The axon of an identified cell (pCC) recognizes specific axons along which the pCC axon normally fasciculates such that, in the absence of this axonal pathway, growth of the pCC cell either stops or progresses randomly (Bastiani et al.,1986; du Lac et al., 1986). An experimental paradigm which creates a temporal delay in the
35
arrival of the axonal pathway results in a cessation of the neuronal process extension until the appropriate substratum for growth arrives (Bastiani et al., 1986). The specific fasciculation of fibers along these "labeled pathways" is likely to be a fundamental feature of construction which also contributes in the elaboration of more complex nervous systems. In more complex nervous systems, however, nerve growth seems to show a hierarchy of substrate preferences rather than the absolute pathway preference that grasshopper neurons have shown. In mammals, groups of neurons may play the role represented by one cell in the grasshopper so that observations on an isolated cell population has not yet been possible in these complex systems. Several glycoproteins have recently been identified on neurons in the grasshopper and Drosophila which have a limited axonal distribution during development (grasshopper, fasciclin I and II, Bastiani et al.,1987; Drosophila, fasciclin III, Patel et al., 1987). The expression of these glycoproteins on a subset of cells and restricted to specific segments of the axonal membrane make them likely candidates for molecular mediators of specific axonal interactions which occur during development.
Neural Cell Adhesion Molecule (N-CAM)
Another cell surface molecule which has been partially characterized, neural cell adhesion molecule (N-CAM), is more general in its distribution and effects. N-CAM is a
36
cell surface glycoprotein. It is present on all neurons and cells that are precursors of striated muscle. This is a very liberal distribution both in terms of cell type and cell locus. Specificity in its effect could be achieved by differences in N-CAM density, different molecular partners for N-CAM in different places, or heterogeneity in the structure of N-CAM itself. Heterogeneity in N-CAM structure has been described both between tissue (nerve and muscle, Rutishauser et al., 1983) and between developmental ages (embryonic and adult, Edelman and Chuong, 1982). N-CAM is apparently important in cell-cell adhesion mechanisms (Rutishauser et al., 1978a; Buskirk et al., 1980), neuriteneurite adhesion mechanisms (Rutishauser et al., 1978b) and neurite-cell adhesion mechanisms (Grumet et al., 1983; Rutishauser et al., 1983). In vitro adhesion of these elements has been disrupted with antibodies to N-CAM. AntiN-CAM disrupts the normally distinct retinal cell layers in vitro (cell-cell), fasciculation of axons together in vitro (neurite-neurite) and the adhesion of spinal cord neurites to muscle cells (neurite-cell) (Edelman, 1983). N-CAM appears to act as an adhesive by covalent binding to another N-CAM molecule. This bonding is enhanced by removal of sialic acid residues which probably hinder access to the binding site. N-CAM binding is specific and does not function in the adhesion of spinal cord neurites to the collagen substratum or to fibroblast cells (Rutishauser et
37
at., 1983). While anti-N-CAM disrupts neurite fasciculation by inhibition of neurite-neurite adhesion, neurite outgrowth is not affected and growth cones appear to operate as normal. N-CAM is a cell surface molecule which mediates adhesion between cell membranes. The importance of N-CAM in binding neural membranes to muscle but not to fibroblast cells reveals a selectivity which could provide direction during normal development. N-CAM might also be a contributing force in the selectivity which operates in vivo to produce specific nerve branching patterns and specific neuron to muscle contacts. While N-CAM is apparently not critical to elongation, direction and elongation may be independently regulated events in vivo.
A variety of cell surface molecules mediating cell interactions with cells, subgroups of cells and extracellular molecules might well be imagined to act in concert to produce specific neuronal networks in vivo.
Axonal Elongation
Trophic factors have been shown to enhance neurite outgrowth. How trophic factors influence the growth mechanism is a question of considerable interest. Neurite elongation in vitro has revealed much of the normal morphology and ultrastructure which subserve the elongation mechanism. Neurites extended in culture have demonstrated that elongation occurs at the growing neurite tip, the growth cone, rather than at the proximal portion of the
38
neurite shaft close to the cell body (Bray, 1970). The growth cone is the site of an active exchange with the environment evidenced, in part, by the presence of pinocytotic vesicles (Hughes, 1933). The growth cone is a large membranous expansion observed at the tip of processes growing both in vitro and in vivo. Filopodia or microspikes and sheet-like veils (lamellipodia) rapidly protrude from and retract into the margins of the growth cone. These dynamic elements of the growth cone may act to probe the immediate environment and/or they may serve in advancing the growth cone. Letourneau (1975) has documented the importance of an adhesive substrate to neurite formation. Using a simple airblast to assay the degree of adhesion to the substrate, he found that a firm adherence to the substrate increases the probability of axon initiation, rate of elongation and degree of axonal branching. The morphology of neurons grown on highly adhesive substrates is distinct (Letourneau, 1979). These neurons display crooked neurites and growth cones that are flat and expanded with longer and more numerous microspikes. Collins (1978b) suggested that the conditioned medium (CM) factor which enhances parasympathetic neurite outgrowth does so by a specific increase in substrate adhesivity. This was based, in part, on the observation that the first visible response to CM was attachment of formerly non-adhering filopodia. Anchoring the growth cone to the substrate may play multiple roles in
39
affecting elongation. It may permit the addition or sequestration of cell membrane for surface expansion at that location. Concurrently, adhesion may be necessary for transmembrane events which trigger the assembly or stabilization of microfilament bundles. Tosney and Wessells (1983) have observed microfilament bundles only in those microspikes which are adherent to the substrate.
Trophic factors produce specific changes in neurite growth, enhancing outgrowth from particular cell types. Substrate adhesivity is clearly one important element in achieving neurite outgrowth. Whether a factor acts only to increase adhesion to the substrate or also specifically interacts with some other element of the elongation mechanism has not been determined. It is possible that adhesion may be separate from an additional and more specific means of enhancing neurite growth by trophic factors.
Studies employing dissociated cells have examined several indices of neurite growth. Neurite initiation, neurite branching and neurite length have all served to reflect the amount of neurite outgrowth produced. It is not clear that these different aspects of neurite growth are regulated equally by all influences. Rogers et al. (1983) reported that as substrate adhesivity increased, neurite initiation increased while neurite length actually decreased. The apparent inhibition of process elongation
40
which was observed may not be a direct cellular effect but, rather, it may be a counteraction to the enhanced initiation. On the highly adhesive substrata used, the cell may be operating at its maximal capacity for neurite extension. At the maximal limit to overall neurite extension an alteration in the propensity for only one of the growth modes (neurite initiation or neurite elongation) may result in a shift of neuritic materials from one form of growth to another. While this shift of materials is speculative, a recent observation (Roederer and Cohen, 1983) on regenerating neurons in the cricket CNS also suggests that a finite capacity for process extension is budgeted between the regenerating axon and numerous neuritic extensions from the cell soma.
The possibility that a declining ability to elongate on the more adhesive substrate is responsible for the observed shift in the mode of outgrowth suggests another means of directing growth in vivo. Factors which inhibit elongation of specific cellular processes could act to direct neurite growth in vivo. In this situation, growth might not be transferred from one mode of extension to another but from one path of growth to another. The cellular regulation of different modes of outgrowth (initiation versus elongation) are not yet well understood.
41
Cue Distribution Theories
The following discussions are concerned with the
distribution of cues which guide axons rather than the specific identity of the cues themselves. The determination of nerve pathway organization is common to the three discussions: genetic influences, the substrate pathway hypothesis and peripheral nerve patterning. Genetic Influences
In Drosophila and Xenopus the grouping of cells during development according to a common genetic origin has been observed. The progeny of one precursor cell all lie within a certain spatial domain named a compartment. In Drosophila, axons and dendrites do not respect compartment boundaries (Palka and Ghysen, 1982). Genetic mutations which have produced a variety of phenotypes useful for examining neural pathway regulation have revealed a genetic influence on both peripheral and central nerve paths (Palka and Ghysen, 1982; Palka et al., 1983). The peripheral paths that direct axons toward the CNS are preferred paths for axonal growth. These preferred paths are non-discriminant and provide a favored path for axons of any compartment or sensory modality. In contrast, the central paths discriminate amongst axonal types. The neuronal compartment identity which is derived from the genotype (transformed or non-transformed) determines the axonal path and pattern of branching centrally. The peripheral paths are specific in their
42
polarity, however, favoring growth directed toward the CNS while central paths are not specifically polarized and permit both rostral and caudal growth.
The genetic heritage of both the neuron and the cellular substrate for growth are likely to contribute to determining the ultimate neural patterning produced. Jacobson (1983) has described seven regions of compartmental restriction in the embryonic nervous system of Xenopus. At early cell stages separate groups of blastomeres have been identified whose progeny each constitute one of the seven compartments. Injection of horseradish peroxidase (HRP) into a blastomere contributing to the ventromedial cell group yields labeled cells in the ventral spinal cord, the ventral myotome and other ventral organs in the rostral embryo. Following such labeling with HRP, the course of the earliest growing motoneuron pioneer axons was followed through the myotome (Moody and Jacobson, 1983). Axons of labeled motoneurons make contact preferentially with labeled muscle cells, that is, cells derived from the same common progenitor cell. Thus, it is possible that pioneering axons of motoneurons are guided by following compartmentally related muscle cells. Compartmental distinctions seem to diminish as environmental complexity increases, making genetic identity less likely to serve as a cue in later development. Compartmental relationships, however, might play a significant role in establishing pathways during pioneer
43
fiber outgrowth. While these results are interesting, in order to extend these observations to other populations in more complex systems better cellular labels are needed. Antibodies which respond to selective cell groups may prove useful in extending these observations. Substrate Pathway Hypothesis
An array of cues which arise during ontogeny were proposed by Katz et al. (1980) to provide continuous, specific substrate pathways for both cell migration and axonal elongation. It was proposed that the cues which described these pathways could be derived from a pattern of molecules, cells or developmental events. Both the availability of a cue and the response of a cell to a cue would not necessarily remain static over the time course of development. This theory does not exclude previously proposed mechanisms for neural guidance. The conceptual framework it provides can accommodate a diversity of cues and a diversity of conditions within the nervous system. The model represented the neural tube on a cartographic coordinate system. The paths of the major fiber bundles early in development were mapped onto the neural tube and these served to depict proposed substrate pathways. Experimental manipulations which have produced growth of fiber tracts into the CNS at ectopic loci (Giorgi and Van der Loos, 1978; Levi-Montalcini, 1976; Katz and Lasek, 1979, 1981) evidenced growth within that sector of the neural tube
44
where the tract normally grew. Katz and Lasek (1981) suggested that the general pattern of substrate pathways within the neural tube confines motor tracts to the ventral basal plate and sensory tracts to the dorsolateral alar plate. There are already amendments to this rule of ordering as Constantine-Paton (1983) has noted the dorso-radial path taken by both normal and ectopic sensory fibers of the auditory (VIII) nerve.
The details of the substrate pathway hypothesis may
require many amendments and revisions to accurately depict the entirety of nervous system development. Taken as a whole, the theory integrates multiple guidance mechanisms operating in a dynamic environment into one coordinated and comprehensible system. This conceptualization of pathway generation is a useful one for consideration of peripheral nerve patterning.
Peripheral Nerve Patterning
Motoneurons innervating skeletal muscles reside in the basal plate of the central nervous system and their axons course through a peripheral environment to reach the appropriate target. The development of motor nerves in both the brainstem and spinal cord of the chick embryo are of specific interest to the experiments which comprise this dissertation.
The motoneuron populations in the brainstem and spinal cord are similar in many respects. Both are basal plate
45
populations which arise from a medial cell column subjacent to the proliferative cells of the ventricular zone. While some motoneurons retain this medial position (e.g., brainstem oculomotor nucleus, spinal cord medial motor column), other motoneuron populations move to a more lateral position within the basal plate (e.g., trigeminal lateral motor nucleus). Cell differentiation progresses rostrocaudally such that motoneurons of the brainstem are generated earlier than motoneurons of the spinal cord (e.g., brainstem trigeminal motoneurons generated day 2-3, Heaton and Moody, 1980; spinal cord lumbosacral lateral motor column generated day 3-4, Langman and Haden, 1970). However, motoneurons as a whole differentiate precociously with respect to the rest of the nervous system. Chick brainstem and spinal cord motoneuron populations are similar in organization, too, in that the cells which innervate a specific muscle are grouped together and assume a reliable position within the central nervous system (see Chapter II; Landmesser, 1978a).
Brainstem motoneuron populations frequently have an
associated cranial sensory ganglion which is analagous to the dorsal root ganglia (DRG) of the spinal cord. This is not always the case, however, and some cranial motor nerves have no associated sensory component (e.g., oculomotor nucleus). In the chick hindlimb, motoneuron axon outgrowth precedes sensory axon outgrowth (Honig, 1982; Tosney and
46
Landmesser, 1985c) and sensory axons apparently depend on the presence of the motor axons for appropriate formation of the muscle nerve branches (Landmesser and Honig, 1986). In the brainstem trigeminal region the motoneuron axons follow the peripheral sensory fibers to the target region (Riggott and Moody, 1987). Despite the sequence of arrival of the trigeminal fibers, it has not been resolved which population initiates the formation of the muscle nerve branches. Motoneuron populations of the brainstem and spinal cord present several different patterns of outgrowth: motoneuron axon outgrowth precedes, follows, or occurs in the absence of sensory axon outgrowth. Fundamentally these motoneuron systems share many similar features and are likely to use common mechanisms to guide outgrowth. It will be of great interest to see how these systems have adapted to suit the individual needs of outgrowth in each situation.
The segmentation imposed on peripheral nerves in the spinal cord may result in additional constraints on nerve outgrowth which are not present in the brainstem. Initial axon outgrowth in the spinal cord is restricted to the anterior half of the opposing somite (Keynes and Stern, 1984; Stern et al., 1986). There is no apparent equivalent in motor nerve outgrowth in the brainstem. Following experimental removal of the somites, the nerves exiting the spinal cord are no longer organized into discrete nerve bundles although the nerves do grow to the target region
47
(Lewis et al., 1981). The somites do not appear to provide cues necessary for the growth to the target but do impose an organization on the fibers which results in the individual nerve roots.
The motor systems of the brainstem and spinal cord are equivalent in many respects. It is likely, then, that the development of these populations is regulated by the same mechanisms although there may be important regional modifications.
Specificity
Peripheral nerve growth has proven to be specific both in vitro and in vivo. The age of the involved tissues, neural and muscle, has been shown to be important in both nerve and neurite growth. Explants of brainstem and spinal cord motoneuron populations show enhanced neuritic outgrowth in the presence of muscle tissue. The developmental stage of the muscle used to produce extract or conditioned medium has an effect on the degree of neurite outgrowth produced from motoneuron populations (Heaton and Kemperman, 1987; Heaton and Paiva, 1986; Pollack and Muhlach, 1981; Pollack et al., 1981; Nurcombe and Bennett, 1983). While there is some dispute as to whether premuscle tissue (present at initial outgrowth) or myotubes (present during muscle-nerve formation) is maximally effective, it does appear that muscle-derived factors are generally specific to periods of nerve growth. The magnitude of the neuronal response has
48
been shown to vary with the neuronal stage of development independent of the target age (Pollack and Muhlach, 1981). The time course of neural responsiveness observed is specific for different neuron types examined. Motor (spinal cord) and sensory (dorsal root ganglion) neurons in vivo send nerves to their target at different times in development. This is reflected in the differing stages at which these two types of neural explants show a maximal neurite outgrowth response to co-cultured target tissue (Pollack and Muhlach, 1982). The temporal specificity of tissue effects in vitro suggests that these are not nonspecific growth effects but may be significant for events occurring at the same stage in vivo, namely nerve outgrowth. The substrate pathway hypothesis suggests that the temporally specific expression of cues serves to establish pattern in the developing nervous system.
The growth of nerves in vivo has also been shown to be influenced by target tissue age (Swanson and Lewis, 1982). Exchanging a limb bud between an older and younger embryo prior to any limb innervation results in a peripheral nerve branching pattern characteristic of the age of the limb donor, not the host. It seems that the nerve is capable of earlier or faster outgrowth but is normally limited by some limb-specific, developmentally expressed, cue. This guidance cue could be either diffusible or pathway-bound and could act to initiate nerve outgrowth or promote elongation. The
49
paths available in the limb bud must differ with time so that the course of nerve growth in vivo is in some way regulated by the changing limb conditions.
Motoneuron pools to individual muscles are discrete and identifiable in the chick spinal cord (Landmesser, 1978a). Peripheral injections of HRP early in development result in a retrograde label which demonstrates both cells bodies and the distribution of their fibers (Lance-Jones and Landmesser, 1981a). This procedure reveals that fibers of specific cells within the spinal cord take highly specific peripheral routes. The axons begin sorting while still quite remote from their ultimate target. Within the crural and sciatic plexuses of the leg, axons selectively cross, grouping with other axons to establish specific branching patterns. The generation of a distinct pattern seems to depend on two features, selective path cues and unique cellular (axonal) identities. The fibers all respond to the path in a specific manner and the particular response is based on the individual axonal identity. Proximal Nerve Path
The organization of proximal plexuses can be affected by changes in the periphery. The presence of a supernumerary limb results in aberrant spinal nerves contributing to a limb plexus while the nerve branching pattern generally appears normal (Lance-Jones and Landmesser, 1981b). Although the gross morphology of plexus and nerve formation is
50
correct, the muscles are innervated by inappropriate motoneurons and the motoneurons do not show the usual organization within the spinal cord. The motoneuron pools are no longer discrete but are diffusely distributed through many spinal segments. The position of motoneurons in the transverse axis of the spinal cord, however, remains similar to the normal distribution pattern. The general form that the plexus takes is determined by the opposing limb, not the spinal nerves which contribute to the plexus. That is, thoracic nerves normally innervate axial muscles but contribute to the formation of an overtly normal crural, thigh, plexus when opposite a supernumerary leg. The determination of the nerve pattern (plexus formation) by the limb identity is consistent with the previously described exchange of limb buds between young and old embryos. In that case, the peripheral nerve branch pattern was a function of the limb age, not the host spinal segment age. The overall nerve form is dependent on the limb while it seems that the establishment of a specific innervation pattern is dependent on the entry of specific fibers into a particular plexus.
Under varied experimental conditions (limb shifts, spinal cord shifts, supernumerary limb; Lance-Jones and Landmesser, 1981b), when inappropriate spinal nerves contribute to a plexus the subsequent muscle innervation is by a non-select population of motoneurons in the contributing spinal cord segments. The axons do not correct
51
their course in the periphery to find their appropriate target but follow the path dictated by their point of exit from the plexus. It appears in this circumstance that there are cues for nerve growth along certain pathways but no organizational cues relevant to these incorrect axons. At the branch points in this circumstance, the axonal choice of diverging paths appears fairly random.
If fibers enter the appropriate plexus but enter from an altered position [following reversals of small segments of spinal cord (Lance-Jones and Landmesser, 1980a), or rotation of limb bud including proximal portions (Ferguson, 1983)], they are able to respond to the change. The axons re-orient within the plexus and take the appropriate path to the correct target. Within the appropriate plexus to which they are specified the axons can alter their course in a selective manner.
Distal Nerve Path
Distal to the level of fiber sorting in the plexus, the normal neural innervation of target muscle is very specific. Lance-Jones and Landmesser (1980) deleted small cord segments to leave some limb muscles without their normal innervation. The remaining nerves by-passed this uninnervated muscle and only innervated the target muscle appropriate to them. Despite the specificity, inappropriate innervation can be produced under certain experimental conditions.
52
If the periphery is altered distal to the proximal limb plexus (distal limb bud rotations: Summerbell and Stirling, 1981; Whitelaw and Hollyday, 1983c; addition of limb segments to an existing limb: Whitelaw and Hollyday, 1983b), the axons are largely incapable of responding to the altered target locus. Inappropriate innervation of target muscles results as the nerve continues along its course and seems to innervate muscle based on the position it occupies rather than the muscle identity. This position-dependent inapropriate innervation can be produced with a variety of experimental paradigms. Following a distal limb rotation, the nerves coursing along the dorsal path innervate the ventral muscles which have been rotated to a dorsal position. Supernumerary segments of limb can be experimentally added in series with normal chick limb segments (Whitelaw and Hollyday, 1983b). A resultant limb, for example, would be composed of thigh, calf, calf and foot segments, in that order. The third limb segment, the duplicated calf, was found to be innervated by nerves which normally contact the third limb segment, the foot-nerves. There was a consistent nerve pattern in the different types of duplications. The nerves innervated the first three limb segments in the usual nerve sequence. Limb segment position rather than limb segment identity determined which nerves provided innervation.
53
On some occasions appropriate muscle innervation will
occur despite an initial misdirection of the nerves. Under the experimental conditions which channeled spinal nerves into the inappropriate plexus (supernumerary limb, rostral limb shift, large cord reversal), aberrant nerves were sometimes observed to break away and take a novel path to their appropriate target (Lance-Jones and Landmesser, 1981b). The formation of aberrant nerves did not follow predictable patterns. When they did form, all the axons within an aberrant nerve were normally appropriate for the muscle that it ultimately innervated. No incorrect connections were made through aberrant nerves. The formation of aberrant paths could be imagined to occur through a shift in the balance of cues. The two routes available to an axon in this situation are the incorrect path it is traveling and the novel path to the appropriate muscle. The incorrect path of travel may in some way physically confine the axon and/or this path may have a decreasing attraction for the axons along its proximo-distal length. The incorrect path presumably has only a weak non-specific signal for those axons at the start. The formation of a novel path may rely on mechanically permissive regions in the environment and/or the presence of specific signals from the appropriate muscle. An alteration in the nerve path could occur through any of the several different combinations of changing conditions. Variability in the formation of aberrant paths
54
could arise in this way. It is very clear that specific axons course over novel pathways to reach specific muscle. This makes it seem highly likely that some cue from the muscle is important in actively guiding the nerve. Molecular Path Cues
These studies on the altered patterns of nerve growth produced by experimental manipulations in vivo led to the postulate that peripheral nerve pattern was established by two categories of cues: (1) those which establish a general concourse for fiber growth with no selective effect on the growth of specific populations thereby creating a "public highway" for growth and (2) those which provide information to direct the growth of specific subsets of axons (Tosney and Landmesser, 1985c; Tosney et al., 1986). While there are no candidates for such highly selective cues of the second category in vertebrates, several candidates have recently been identified in the simpler insect nervous system (Bastiani et al., 1987; Patel et al., 1987; see section on Cell Surface Molecules).
Despite the complexities of the vertebrate, there are several candidate molecules which may provide the "public highway" cues. The growth of peripheral neurites in vitro is enhanced on a laminin substratum (Rogers et al., 1983). In vivo, the distribution of laminin in the brainstem and spinal cord suggests that it could be providing an adhesive substratum for axonal outgrowth during early development
55
(Rogers et al., 1986; Riggott and Moody, 1987). Recently, Schwann cells have been observed to precede the growing front of motor axons in the chick forelimb (Noakes and Bennett, 1987). Experimental deletion of the Schwann cell precursors by neural crest extirpation has resulted in a disruption of axonal ingrowth into the limb muscles in both the brachial and lumbar regions of chick spinal cord (Carpenter and Hollyday, 1986). In addition to secreting the laminin glycoprotein (Cornbrooks et al., 1983; Palm and Furcht, 1983), Schwann cells secrete the proteolytic enzyme plasminogen activator (Krystock and Seeds, 1984) and nerve growth factor (Bandtlow et al., 1987; Assouline et al., 1985). These molecules may act singly or in concert to provide a path for trailblazing fibers.
Evidence suggests that multiple molecular cues mediate the interaction between motoneurons of the ciliary ganglion and embryonic myotubes in vitro (Bixby et al., 1987). Severe inhibition of neurite outgrowth was observed in this system only when the JG22 antibody to cellular extracellular matrix receptors (ECM; likely blocking cell receptors for laminin, fibronectin and collagen) was administered in combination with antibodies to two cell adhesion molecules (neural cell adhesion molecule and neural Ca2+-dependent cell adhesion molecule). Thus, multiple cell surface associated molecules appear to be significant in mediating the interaction between the processes of motoneurons and myotubes.
56
The following experiments are concerned with the
elements active in shaping early motoneuron development. The target-independent aspects of early motoneuron development will be differentiated from those aspects of growth which are target-influenced. The regionally specific nature of target-derived factors will be characterized in terms of their growth promoting effects. Finally, evidence will be provided that one of the above mentioned molecules, NGF, is important in early motoneuron development.
CHAPTER II
THE PATTERN OF EXTRAOCULAR INNERVATION
BY THE OCULOMOTOR NUCLEUS
OF THE CHICK
Introduction
The oculomotor complex lies in the rostral
mesencephalon. These motoneurons are the cell bodies of origin of the third cranial nerve, the oculomotor nerve. The oculomotor complex has been subdivided into four subnuclei on morphological grounds. Three subnuclei, dorsolateral, dorsomedial and ventromedial, innervate four of the six extrinsic muscles of the eye. A midline accumulation of cells between the ventral subnuclei, the central nucleus, is also sometimes described. The accessory subnucleus innervates intrinsic eye musculature (this innervation was not examined within this study).
The topographic grouping of motoneuron cell bodies is prevalent in the central nervous system. In the ventral spinal cord, motoneuron cell bodies are arranged in groups corresponding to the musculature of specific body regions which they innervate, such as arm, leg, and axial musculature. The segregation of motoneurons based on muscle innervated has been demonstrated within the oculomotor nucleus of mammalian (Augustine et al., 1981) and avian 57
58
(Isomura, 1973) species. The distribution of motoneurons within the chick oculomotor nucleus was reported following the use of degeneration techniques to map the projection pattern (Isomura, 1973). The results of this study were somewhat equivocal, however, because only a small percentage of the neurons were identified as degenerating.
The pattern of innervation and the origin of the
ventromedial subnuclear motoneurons have been of particular interest because of the partially contralateral innervation which this subnucleus provides. The oculomotor anlagen appears on either side of the midline in the mesencephalon between three and five days of development. Subsequently, from five to ten days, this homogeneous cell group segregates into three distinct subnuclei. During this same developmental period, a portion of the population also migrates medially and, intermingling with the contralateral migratory population, apparently crosses the midline (Puelles-Lopez et al., 1975). This migration, then, establishes the contralateral innervation. Because of the high cell packing density and the intermingling of contralateral populations, it has been difficult to determine the limits of the migratory population. Additionally, the ultimate destination of this population within the contralateral subnucleus has not been resolved. Neither the developmental observations nor degeneration mapping have definitively described the distribution of
59
motoneurons within the ventromedial subnucleus. The present study was designed to resolve the distribution of the four motoneuron populations within the three subnuclei using the horseradish peroxidase retrograde tracing technique.
Methods
Fertile White Leghorn chicken eggs were obtained from
the Poultry Science Department at the University of Florida. The eggs were set in a forced draft incubator at 370C, 70% humidity for 18 days (total incubation period = 20 days). For injection, the eggs were candled and the shell overlying the air pocket was carefully removed. The egg was secured in a wax egg holder and a drop of water placed on the intact membranes to reveal the course of the blood vessels. The chorioallantoic membrane overlying the head was slit, taking care to avoid cutting any large blood vessels. A forceps could then be used to gently pull the head out. The head was supported on dampened gauze pads. Under the dissecting microscope the right eye was anaesthetized with xylocaine. The eye was enucleated, removing the vitreous, and some of the connective tissue surrounding the eye was cut to permit some movement of the eye and access to the particular muscle to be injected. A solution of 30% HRP (Sigma, type VI) in
0.9 % saline was loaded into a five microliter Hamilton syringe fitted with a 50 Am glass cannula tip. The HRP filled syringe was mounted in a Brinkman micromanipulator and visually directed to the belly of the muscle under the
60
dissecting microscope. From 0.1-0.3 Al of HRP solution was injected into the belly of the muscle and this was visually confirmed by the spread of the dark brown HRP within the muscle. Following the injection the cannula was withdrawn and the muscle swabbed to soak up any HRP which had seeped out. An additional precaution against diffusion of leaked HRP involved packing the exposed injection area with sterile gelfoam. The egg was then returned to the incubator using dampened gauze pads to cradle the head and keep the embryo moist.
After a survival period of 18-24 hours the embryo was anaesthetized with 0.5 ml of chloral hydrate and perfused intracardially with 5 ml of saline followed by 50 ml of fixative (1.25% glutaraldehyde and 1% paraformaldehyde in
0.1M phosphate buffer, pH 7.4). The brain was then dissected out and immersed in 0.1M phosphate buffer with 10% sucrose at 40C overnight. The following day the brain was frozen on a sliding microtome and 25 Am sections were cut into the 10% sucrose phosphate buffer. Within 24 hours the sections were reacted using the tetramethyl benzidine procedure of Mesulam (1978) which demonstrates the presence of HRP with a dark brown reaction product. The sections were mounted, dried overnight and counterstained with neutral red before coverslipping.
The identity of the tissue was coded for analysis so that the injection site was unknown to the investigators.
61
Each section through the rostro-caudal extent of the nucleus was examined. Both notes and drawings were made to describe the presence, the locus and the extent of the HRP reaction product. The subnuclei are quite distinct in the rostral and mid regions of the nuclear complex. In the caudal region the distinction between the dorsal subnuclei becomes less clear as they coalesce to form a single crescent of cells overlying the dorsal and medial aspect of the medial longitudinal fasciculus. In all regions, except the most caudal portions of the dorsal subnuclei, the labeled cells can be confidently ascribed to a single subnucleus.
Results
In the adult oculomotor complex the cellular
organization is such that there are four subnuclei distinguishable with conventional histological methods (Fig. 2-1). The accessory subnucleus is located most dorsally and consists of small neurons which provide preganglionic parasympathetic innervation of the ciliary muscle and the sphincter muscle of the iris. This is analogous to the Edinger-Westphal nucleus of mammals. This innervation was not examined. The three remaining subnuclei provide innervation to four of the extraocular muscles.
The cellular organization is readily seen in the control section in Figure 2-1 taken from the mid region of the nucleus. This organization varies somewhat over the rostrocaudal extent of the complex. The rostral-most extent
62
t os
Figure 2-1. Normal organization of oculomotor nucleus.
Transverse section through the mesencephalic basal plate of a chick at 3 days posthatching. The morpologically distinct subnuclei can be readily identified.
a, accessory subnucleus; dl, dorsolateral subnucleus;
dm, dorsomedial subnucleus; vm ventromedial subnucleus;
v, ventricle. Hematoxylin and eosin stain. X 127.
63
of the complex is evidenced by the presence of the accessory, dorsolateral, and dorsomedial subnuclei. Slightly caudal to this the ventromedial subnucleus first becomes apparent. Progressing caudally the accessory subnucleus diminishes in size while the ventromedial subnucleus expands. The nerve root is first evident at mid level of the rostrocaudal extent of the complex. In the caudal reaches of the nuclear complex the accessory subnucleus disappears while the ventromedial group becomes more predominant. The dorsomedial and dorsolateral subnuclei fuse in the caudal regions to form a single extended cell group overlying the medial longitudinal fasciculus (MLF). These groups trail off caudally. The ventromedial group is the last to disappear and does so just rostral to the appearance of the trochlear motoneurons which lie lateral to the MLF. An additional central cell group has sometimes been described to lie in the midline between the two ventromedial subnuclei (Nimii et al., 1958).
Injection of the inferior rectus muscle produced a
consistent pattern of labeling. The HRP reaction product was always in cells on the ipsilateral side and virtually always within the dorsolateral subnucleus. There was an even distribution of labeled cells within this subnucleus and along its rostrocaudal extent. There was only one cell which was found to lie outside this subnucleus. This labeled cell was located along the lateral border of the dorsomedial
64
subnucleus, that is, immediately adjacent to the boundary of the dorsolateral subnucleus. The cameral lucida drawing in Figure 2-2a represents the typical pattern of labeled cells seen following injections of the right inferior rectus muscle. The photomicrograph in Figure 2-3 shows a typical profile of labeled cells following such an injection.
Injection of the medial rectus muscle produced a consistent labeling of the ipsilateral dorsomedial subnucleus. In the rostral portion of the nucleus there were many labeled cells medially. These cells were quite near the midline and distinct from other cells of the complex in their spindle shaped morphology. All other cells labeled were round to ovoid in their morphology. The labeled cells were most prominent in the mid region where there were many cells labeled in the dorsal portion of this subnucleus. Caudally, there was a more even distribution of labeled cells within the subnucleus with the labeled population extending to the lateral border. There was also an occasional labeled cell observed along the medial border of the dorsolateral subnucleus. The cameral lucida drawing in Figure 2-2b depicts the characteristic pattern of labeled cells seen after an injection of the right medial rectus muscle. A photomicrograph of a typically labeled section is shown in Figure 2-4.
Injection of the inferior oblique muscle also produced a consistent ipsilateral pattern of labeled cells. In the
65
0i0
Figure 2-2. Summary of motoneuron organization within the
oculomotor nucleus.
Camera lucida drawings of transverse sections through the mesencephalic basal plate of day 18 chick embryos following injections of the extraocular muscles with HRP. The presence of HRP reaction product within the subnuclei is represented by the dots. A. Injection of the right inferior rectus muscle typically resulted in labeled cells distributed through the right dorsolateral subnucleus. B. Injection of the right medial rectus muscle typically resulted in labeled cells distributed through the right dorsomedial subnucleus. C. Injection of the right inferior oblique muscle typically resulted in labeled cells distributed through the lateral portion of the ventromedial subnucleus. D. Injection of the right superior rectus muscle typically resulted in labeled cells distributed through the medial portion of the left, contralateral ventromedial subnucleus.
dl, dorsolateral subnucleus; dm, dorsomedial subnucleus;
vm, ventromedial subnucleus.
66
Figure 2-3. Innervation pattern: inferior rectus muscle.
Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the inferior rectus muscle. The dark HRP reaction product is localized to cells of the dorsolateral subnucleus (boxed) ipsilateral to the injection. The boxed area is shown at higher magnification in the inset. The closed arrows mark cells in which the characteristic granular HRP reaction product can be seen. At the lower magnification open arrows mark artifactual labeling which results from the endogenous peroxidase activity of blood cells. This reaction product is localized to blood vessels and can be distinguished from cellular labeling by size (very small) and shape (elongate). Abbreviations as in Figure 2-1. Neutral red stain. X 99 (inset X 233).
vv W. ~
Figure 2-3. Innervation pattern: inferior rectus muscle.
Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the inferior rectus muscle. The dark HRP reaction product is localized to cells of the dorsolateral subnucleus (boxed) ipsilateral to the injection. The boxed area is shown at higher magnification in the inset. The closed arrows mark cells in which the characteristic granular HRP reaction product can be seen. At the lower magnification open arrows mark artifactual labeling which results from the endogenous peroxidase activity of blood cells. This reaction product is localized to blood vessels and can be distinguished from cellular labeling by size (very small) and shape (elongate). Abbreviations as in Figure 2-1. Neutral red stain. X 99 (inset X 233).
67
Figure 2-4. Innervation pattern: Medial rectus muscle.
Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right medial rectus muscle. The HRP reaction product is present in cells of the dorsomedial subnucleus (boxed area) ipsilateral to the injection site. The inset of the boxed area shows many labeled cells (arrows) identifiable at higher magnification. Abbreviations as in Figure 2-1. Neutral red stain. X 90 (inset X 144).
a~ 14
it A
do 40 V
IrA AA
4W4
Figure 2-4. Innervation pattern: Medial rectus muscle.
Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right medial rectus muscle. The HRP reaction product is present in cells of the dorsomedial subnucleus (boxed area) ipsilateral to the injection site. The inset of the boxed area shows many labeled cells (arrows) identifiable at higher magnification. Abbreviations as in Figure 2-1. Neutral red stain. X 90 (inset X 144).
68
rostral regions the cells were located predominantly in the dorsal portion of the ventromedial subnucleus and more laterally than medially. Progressing to mid region the labeled cells spanned the full dorso-ventral extent of the ventromedial subnucleus but with more cells appearing dorsally. As seen rostrally, the labeled cells were still restricted to the lateral portion of the subnucleus. Caudally the number of labeled cells diminished but the same pattern continued. The labeled cells were in the lateral portion of the ventromedial subnucleus with more labeled dorsally than ventrally. Three dimensionally, this pattern constructs something of an elongated pyramid, with one short side corresponding to the dorsal border of the ventromedial subnucleus. The base of this pyramid faces rostrally with the apex projecting caudally. The camera lucida drawing in Figure 2-2c represents the characteristic labeling pattern in mid region of the complex with a photomicrograph of typically labeled cells in Figure 2-5.
Injection of the superior rectus muscle again produced a consistent pattern of labeling but, in this case, it was entirely contralateral. The pattern of labeled cells was complementary to that produced by injection of the inferior oblique muscle so that essentially the medial portion of the ventromedial subnucleus was labeled. Rostrally the labeled cells were found primarily along the medial or dorsomedial border of the contralateral ventromedial subnucleus with
69
- ,.
*, r
Figure 2-5. Innervation pattern: Inferior oblique muscle.
Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right inferior oblique muscle. Granular HRP reaction product is localized to cells of the lateral portion of the ventromedial subnucleus ipsilateral to the injection site. The boxed ventromedial subnucleus is shown at higher magnification in the inset. The arrows point to labeled cells which are located exclusively within the lateral portion of the subnucleus. Abbreviations as in Figure 2-1. Neutral red stain. X 100 (inset X 167).
70
elf.
- q
d.4
Figure 2-6. Innervation pattern: Superior rectus muscle.
Transverse section of the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right superior rectus muscle. Labeled cells are present in the ventromedial subnucleus (boxed) located contralateral to the injection site. Higher magnification of the boxed region (inset) reveals that the labeled cells (arrows) are clearly confined to the medial aspect of the contralateral ventromedial subnucleus. Abbreviations as in Figure 2-1. Neutral red stain. X 92 (inset X 138).
ilk '44
-. g
JJ
'1 4V
Figue 26. nneratin ptter: Speror rctu mucle
Transverse~~~-V sectio ofteouooo eino h
mesncphlo o a 1 dy cic ebro olowngtR
injctin f te igh sperorrecusmusle Laeld cll
are resnt n te vntrmedal sbnuleu (bxed loate
cotrVtea toteijcinst.Hihrmgiiaino
the ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ boeIein(ne)rvasta h aee el
Figure 2-1. Ineraio redtin: 92ero (inset Xu138).
71
only few cells labeled. In the mid region the labeled cells were again confined to the medial aspect but now fully spanned the dorsoventral extent of the subnucleus. Caudally this distribution persisted although ventrally the labeled cells were spread somewhat laterally. In addition, at this level, the labeled population included what appeared to be the "central nucleus" which lies in the medial area between the two ventromedial subnuclei. The camera lucida drawing in Figure 2-2d represents the pattern of labeled cells following injection of the right superior rectus muscle. The typical distribution of labeled cells is seen in the photomicrograph in Figure 2-6.
Discussion
The pattern of labeled cells following retrograde
transport of injected HRP was very consistent for all four of the extraocular muscles. The locations of each of the four motoneuron populations present within the complex were unique with very little overlap. On the few occasions where a labeled cell was found to lie outside the primary domain it was on the border of the neighboring subnucleus immediately adjacent to the parent population.
The innervation pattern reported here and summarized in Table I is as follows: the dorsolateral (DL) subnucleus innervates the ipsilateral inferior rectus muscle, the dorsomedial (DM) subnucleus innervates the ipsilateral
72
TABLE 2-1
Major projection patterns
of the oculomotor complex in the chick
Extraocular Muscles Inferior Medial Inferior Superior Oculomotor Subnucleus rectus rectus oblique rectus
Dorsolateral +, I - Dorsomedial +, I Ventromedial, lateral - +, I Ventromedial, medial - +, C Central - +, C
+, presence of HRP reaction product -, absence of HRP reaction product
I, ipsilateral projection
C, contralateral projection
73
medial rectus muscle, the lateral portion of the ventromedial (lVM) subnucleus innervates the ipsilateral inferior oblique muscle, and the medial portion of the ventromedial (mVM) subnucleus innervates the contralateral superior rectus muscle. This distribution pattern of the motoneuron populations within the oculomotor subnuclei is similar to that which has been previously described using degeneration techniques to identify the particular populations (Isomura, 1973). The reports differ, however, in the description of the segregation of the motoneurons within the ventromedial subnucleus. Isomura (1973) reported finding one population confined to the dorsal aspect (innervating the ipsilateral inferior oblique muscle) and the other to the ventral aspect of the subnucleus innervating the contralateral superior rectus muscle). Our findings do not conflict with this observation so much as refine it, since we also found more cells dorsally which innervated the ipsilateral inferior oblique muscle. However, while the preponderance of cells were dorsal there was also a wedge of labeled cells which extended ventrally. Additionally, it was clear that both dorsally and ventrally the cells were restricted to the lateral aspect of the ventromedial subnucleus. The reported differences are probably attributable to the numbers of cells labeled with the two techniques. The greater numbers of cells which were labeled with the retrograde transport technique provided a fuller
74
picture of the entire .motoneuron population and extends the observations to provide greater detail.
The division of the motoneuron populations within the
ventromedial subnucleus is of particular interest because of the unusual migration which occurs during development to produce the contralateral innervation. This migration has been described to occur in several avian species (PuellesLopez et al., 1975; Sohal, 1977; Heaton, 1981). In the chick the medial migration of cells from the oculomotor anlagen is first observable at 5 days of incubation. Until that time this medial cell column had been a single homogeneous cell group. Over the next two days the number of participating cells increases and the cells meet with the contralateral population and seem to freely intermingle. Subsequently the cells disperse and apparently continue on their migratory route to the contralateral side. Purely descriptive studies have been unable to determine the extent of the ventromedial population which migrates or the exact locus of their destination. Puelles-Lopez et al. (1975) speculated that the entire ventromedial population may participate in the migration. The present study provides rather convincing evidence that it is only those cells that comprise the medial portion of the ventromedial subnucleus which participate in its migration.
The pattern of labeling observed here was very
consistent and reliable between cases. The virtually
75
exclusive segregation of the motoneuron populations differs from the varying degrees of overlap in these populations reported in other species (cat; Gacek, 1974; frog; Matesz and Szekely, 1977; stingray; Rosiles and Leonard, 1980; baboon; Augustine et al.,1981; macaque monkey; Spencer and Porter, 1981). The technique used here permitted labeling of only a portion of the motoneuron population innervating a muscle. The volume of HRP injected and the resultant number of cells labeled was restricted by the leakage of HRP when large volumes are injected combined with the ability of embryonic muscle to incorporate the enzyme after only topical exposure. Nonetheless, the highly consistent pattern and the fact that there were no areas left unlabeled by our injections argues strongly against the existence of any significant overlap among the motoneuron populations in the chick oculomotor complex. The few instances in which a labeled cell was observed outside the primary domain for that cell group the errant cell was contiguous with the parent population but located in the neighboring subnucleus.
There are many similarities in the pattern of oculomotor innervation reported for the different species. The contralateral innervation of the superior rectus muscle has been reported in stingray (Rosiles and Leonard, 1980), frog (Matesz and Szekely, 1977), cat (Gacek, 1974), baboon (Augustine et al., 1981), and macaque monkey (Spencer and Porter, 1981). The topographical organization within the
76
oculomotor complex is particularly similar in chick, cat, and baboon. In each of these species, the motoneurons innervating the superior rectus and the inferior oblique muscles accumulate in close proximity to one another. It may be functionally important for these two motoneuron populations to be contiguous since both muscles serve to rotate the eye upwards. It is particularly interesting that they innervate muscle on opposing eyes because the inferior rectus rotates the eye up and slightly nasally while the superior rectus muscle rotates the eye upward with a temporal bias so that if activated together they would produce an upward rotation of both eyes with both eyes slightly directed to the same side. While these species demonstrate similarities in organization, in the baboon there is a greater degree of intermingling of adjacent populations. On the basis of differences in size and location of oculomotor motoneurons of the macaque monkey, Spencer and Porter (1981) suggest that there might be subpopulations of motoneurons to a particular muscle which mediates the action of that muscle in different types of eye movements. It is possible that such multiple representation of a particular muscle indicate that in this species at least, the subnuclear organization is more closely related to specific types of movement rather than to specific muscles.
77
Oculomotor organization in fish and amphibian species has been found to be somewhat different from that seen in mammalian and avian forms (e.g. bilateral innervation: stingray, Rosiles and Leonard, 1981; frog, Matesz and Szekely, 1977). This disparity may reflect the diversity of these species and their differing adaptive requirements. Their individual environments, the predators and prey likely to be encountered, and the position of the eyes in the head might all influence the types of eye movements most frequently employed.
CHAPTER III
OCULOMOTOR DEVELOPMENT IN THE CHICK
FOLLOWING EXPERIMENTAL REMOVAL OF THE TARGET MUSCLES.
Introduction
In the ventral spinal cord, motoneuron cell bodies are arranged in groups corresponding to the musculature of specific body regions which they innervate, such as arm, leg, and axial muscle. This topographic grouping of motoneurons is prevalent in the central nervous system (Landmesser, 1978a; Landmesser, 1978b). In the chick oculomotor complex, retrograde labeling has demonstrated that the motoneurons innervating four of the extraocular muscles (i.e. superior rectus, medial rectus, inferior rectus and inferior oblique) exist in segregated populations within the subnuclei (Heaton and Wayne, 1983). A similar organization of motoneuron populations is seen in the oculomotor complex of mammalian species (Augustine et al., 1981). This adult segregation of the motoneuron population develops in the chick embryo from a single, homogeneous cell group, the oculomotor anlagen (Figure 3-1). This precursor of the adult oculomotor complex is first seen in the rostral midbrain of the chick embryo at day 3 of incubation. At this time, these cells begin to send fibers out to the peripheral
78
FIGURE 3-1: Normal development of oculomotor complex
The normal development of the chick oculomotor complex is represented in these three transverse sections of the rostral mesencephalon at (a) 4, (b) 7, (c) 10 days and (d) 18 days of incubation.
(a) At 4 days the oculomotor anlagen (A), a homogeneous cell group, sits in the marginal layer of the mesencephalic neural tube (NT). The nerves (N) can be seen exiting ventrally into surrounding mesenchyme (S). X 258.
(b) At 7 days the emerging subnuclear organization is becoming apparent. Accessory (a), dorsal (d) and ventral
(v) cell groups are visible. There is a suggestion of a medial -lateral division of the dorsal cell group which will become two distinct subnuclei later. There are many cells migrating (m) between the ventral groups. The medial longitudinal fasciculus (MLF), a brainstem fiber bundle, sits next to the oculomotor cell bodies. X 337.
(c) Cells migrating between the two ventromedial subnuclei in (b) are shown here at higher magnification. The typical elongate morphology of migratory cells is apparent. The migration across the midline to the contralateral subnucleus is indicated by the presence of cells oriented in opposing directions (arrows). X 1350.
(d) At 10 days the subnuclei have separated completely. The accessory cells (a) look quite different from the other subnuclei. X 241.
(e) The adult configuration of the oculomotor complex is evident at 18 days which is just prior to hatching. The dashed line within the ventromedial subnucleus approximates the segregation of motoneurons which innervate the ipsilateral inferior oblique muscle (lateral portion) and the contralateral superior rectus muscle (medial portion). X 213.
a=accessory, dl=dorsolateral, dm=dorsomedial,
vm=ventromedial, v=ventricle, s=mesenchyme
80
4* 0
*A* -
81
musculature. By embryonic day 5 all of the extrinsic eye muscles innervated by the oculomotor complex have been reached by some fibers from that complex. In the subsequent development of the oculomotor complex, embryonic days 5-10, there is a gradual reorganization of the cell bodies so that the subnuclear distinction emerges. During this time there is also a migration of some cells to the midline where they intermingle with the corresponding cells from the opposite side. It seems that these cells then cross the midline and comprise the medial portion of the ventromedial subnucleus. The crossing of these cells thus accounts for the contralateral innervation of the superior rectus muscle (Puelles-Lopez et al., 1975; Heaton, 1981; Heaton and Wayne, 1983).
The appearance of subnuclei and the migration of a
portion of the cell population both serve to produce the muscle-specific grouping of motoneurons. The temporal contiguity between the nerve fibers first reaching the target musculature and the onset of these organizational changes in the cell bodies of the oculomotor complex suggest the possibility of a causal relationship. Thus, it is possible that some influence from the periphery either initiates or directs the cells in their rearrangements.
In order to account for specific matching between
motoneuron and muscle cell populations there seem to be three developmental possibilities. The muscle specification
82
may precede neural development and the subsequent organization of neuronal cells would be dictated by the contacts made with the peripheral target. Alternatively, neural specification may precede muscle development and the distinct neural populations may then determine muscle organization upon contact with the periphery. The third possibility is that the specificity of neural and muscle populations arise independently and that development has incorporated some means to appropriately couple specific cell populations in the periphery.
The pathway of nerve outgrowth, the subnuclear
segregation, and the migration of a distinct population of motoneurons may all be subject to some influence from the four extraocular muscles innervated. This possibility was experimentally addressed by removal of the presumptive muscle tissue prior to the appearance of the oculomotor anlagen. The subsequent development of the oculomotor nerve and nucleus was observed over the course of embryonic days 5 through 10 and is described here.
Methods
Fertile chicken eggs, Gallus gallus domesticus, were obtained from the University of Florida Poultry Science Department. The eggs were set in a forced draft incubator at 370C and 70% relative humidity for 36-40 hours to yield stage 10-11 embryos (Hamburger and Hamilton, 1951). The surgery precedes both the condensation of tissue to form the
83
extraocular muscles and nerve outgrowth from the oculomotor complex. Immediately prior to surgery the eggs were candled to reveal the position of the embryo. A window was made in the shell and after exposure, 1% neutral red stain was dropped onto the embryo to enhance contrast. The overlying vitelline membrane was carefully peeled away with fine forceps. The left optic vesicle and surrounding mesoderm, the precursor for the extraocular muscles, were removed with a microsurgical vibrating needle (Wenger, 1968). In some control operations the mesoderm was left intact and just the left optic cup was removed. The surgery is shown diagrammatically in Figure 3-2. In all cases the surgical removal of tissue was from the left side of the embryo. The egg was then resealed with parafilm and was carefully set in an incubator. The window was directed up and the egg left undisturbed for the remainder of its incubation period. The embryos were sacrificed at days 5 through 10 of incubation. They were fixed in Bouin's fixative, dehydrated, embedded in paraffin, and sectioned at 10 gm. Following mounting on slides, the sections were counterstained with hematoxylin and eosin.
Results
Muscle Removal
The operations always resulted in the presence of some residual muscle tissue. Each case was evaluated individually. The majority of cases demonstrated a dramatic
84
ov
NT S
Figure 3-2: Diagram of surgical removal of presumptive
extraocular muscle
The chick embryo at stage 10-11 of development is represented in the drawing above. The dotted line frames the tissue removed. Note that dorsally the cut followed the contours of the neural tube while the ventral cut extended to the midline (arrow).
ov= optic vesicle, s= mesenchyme, NT = neural tube
85
reduction of muscle with just a single small mass of muscle cells remaining. In one case, a severe and almost compete absence of muscle cells was seen (Figure 3-3). The volume of residual muscle present probably provides an inflated index of normalcy because it assumes that any residual muscle in the area is an appropriate target for oculomotor neurons. In some cases the only muscle available seemed to be displaced caudally suggesting that it may actually have been the remnant of another muscle group. While it is not possible to distinguish muscle cells of the oculomotor system from other muscle cells of the region, these cells may bear differences important in normal development (see Chapter V). In some cases a fair mass of muscle was present but it should be noted that this was still quite reduced compared to the normal muscle. Of even greater significance, the muscle did not separate into the distinct muscle groups but formed one consolidated mass. This consolidation was not simply a function of losing the eye as an anchoring point because control operations with eye removals only have no eye or the surrounding orbit yet still evidence the distinctive muscle groups. It seems that there is some qualitative change in the muscle so that the normal complement of distinct muscle groups is not available for innervation.
86
a b
Figure 3-3: Muscle remnants following removal
Transverse sections through the rostral mesencephalon of (a) day 6 and (b) day 7 experimental chick embryos. These cases demonstrate the range of muscle which was present in the experimental cases used for analysis, (a) representing a maximum and (b) representing a minimum. (a) The oculomotor nerve on the left, operated side can be seen coursing towards the residual muscle (rm). The experimental muscle mass (rm) always has an appearance distinct from normal muscles (m). (b) There are only a few muscle cells (rm) present at the body wall. (a) X 148; (b) X 112.
M= mesencephalon, s=mesenchyme, N= nerve
Figure 3-4: Segregation of subnuclei
Transverse sections through rostral mesencephalon of (a) 7 and (b) 9 day experimental embryos demonstrating the appearance of subnuclei. As in all cases, the left side is the operated side. (a) X 305; (b) X 433.
(a) Accessory (a), dorsal (d) and ventral (v) cell groups are visible on both sides.
(b) Accessory (a), dorsomedial (dm), dorsolateral (dl) and ventromedial (vm) cell groups are apparent on both sides. The shapes of subnuclei on experimental and control sides are disparate because they have been cut in somewhat different planes of section. The similarity in organization is clear nevertheless.
n= nerve, s=mesenchyme
88
*
(' '. .
' i -.....
a*a
b
89
Subnuclear organization and cell migration
The experimental manipulation introduced a loss of symmetry in the embryo which frequently resulted in a distortion of the neural tube. This neural tube distortion made a consistent orientation of the tissue for transverse sectioning difficult to achieve and sometimes meant that the experimental and control sides of the brain were sectioned in different planes. As a result the observed variability in the shape of the subnuclei in experimental embryos was determined largely by the plane of section of the tissue. In contrast, the degree of separation of the individual subnuclei was only slightly influenced by the plane of section. In all cases examined the subnuclei were segregated to some extent (Figure 3-4). In any individual case, the extent to which the individual subnuclei were distinguishable on the operated side reflected the organization evident on the unoperated control side of the same embryo. Therefore, the breakdown into the distinct subnuclear populations occurred regardless of the presence or absence of the normal target muscles.
Migrating cells were identified by two criteria: (1)
their position within the complex, located between the two ventral cell groups; and (2) their morphology, which consisted of an elongate cell body with a fine trailing process. In all experimental embryos migratory cells which were oriented in both directions could always be found.
90
Thus, the migrating population always contained cells from the experimental side as well as cells from the unoperated, control side (Figure 3-5). The magnitude of migration varied considerably between cases. Any precise quantification of the migrating population was made impossible by the fact that there are no distinct borders separating migrating from non-migrating cells and the migrating population itself consists of intermingled cells from both sides. In general, the variability in degree of migration did not show any relationship to the amount of muscle remaining. It is possible that the variation observed was a normal consequence of capturing cases at different developmental stages. The surgical intervention probably alters the developmental time course to some extent, thus creating a greater range in the developmental stages found amongst any age group.
Axonal pathway
In 21 of 23 cases analyzed, the oculomotor nerve
projected along a very consistent pathway to any residual muscle still present on the operated side (Figure 3-6). Among this majority which projected to the ipsilateral (normally appropriate) side, there was usually a fair amount of muscle still present in the target region. The quantity of residual muscle available did not seem to determine the path of innervation, however. A few cases showed only
etf'" 1 91
d*
"70 "No
...., ,
Figure 3-5: Migration
Transverse sections through the rostral mesencephalon of two day 7 experimental embryos demonstrating the migration of cells. The variation in numbers of migrating cells can be seen. Arrows indicate the profile of migratory cells which appear to originate from the left, operated side. (a) X 398;
(b) X 398.
|
PAGE 1
STUDIES ON NERVE FIBER OUTGROWTH IN MOTONEURON POPULATIONS OF THE DEVELOPING CHICK EMBRYO BY DENISE BETH WAYNE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1988
PAGE 2
Copyright 1988 by Denise Beth Wayne
PAGE 3
'Sometimes the magic works' Thomas Berger
PAGE 4
I ACKNOWLEDGEMENTS I would like to thank the members of my supervisory committee, Drs. Walker, West, and Ulshafer for their patience and fortitude. I would like to thank my advisor, Dr. Marieta Heaton for the many opportunities and for taking me in twice. I would like the many other people who freely offered their time and resources particularly Dr. Mohan Raizada, Dr. Michael Young and Bill Creegan. Last but not least, I would like to thank my family and friends who have given me limitless encouragement and support through it all. iv
PAGE 5
TABLE OF CONTENTS page ACKNOWLEDGEMENTS . . . . . . . . . i V ABSTRACT. .. . . . . . . . . . . Vii CHAPTERS I. GENERAL INTRODUCTION ........................ 1 Theoretical Considerations.............. 2 Mechanical Cues......................... 3 Electrical Cues ......................... 11 Chemica 1 Cues. . . . . . . 13 Axonal Elongation ....................... 37 Cue Distribution Theories ............... 41 Peripheral Nerve Patterning ............. 44 II. THE PATTERN OF EXTRAOCULAR INNERVATION BY THE OCULOMOTOR NUCLEUS OF THE CHICK ........ 57 Introduction ............................ 57 Methods . . . . . . . . 5 9 Results. . . . . . . . . 61 Discussion.............................. 71 III. OCULOMOTOR DEVELOPMENT IN THE CHICK FOLLOWING EXPERIMENTAL REMOVAL OF THE TARGET MUSCLES. . . . . . . 7 8 Introduction ............................ 78 Methods . . . . . . . . 8 2 Results. . . . . . . . . 84 Discussion. . . . . . . . 97 IV. THE EFFECT OF SERUM AND DEFINED MEDIUM CONSTITUENTS ON NEURITE GROWTH FROM EARLY NEURAL TUBE EXPLANTS ................... 105 Introduction ............................ 105 Methods. . . . . . . . . 1 O 7 Results. . . . . . . . . 109 Discussion. . . . . . . . 121 V
PAGE 6
V. SPECIFIC RESPONSIVENESS OF CHICK TRIGEMINAL MOTOR NUCLEUS EXPLANTS TO TARGET-CONDITIONED MEDIA .................. 128 Introduction ............................ 128 Methods. . . . . . . . . 13 o Results ................................. 135 Discussion. . . . . . . . 14 O VI. THE ONTOGENY OF SPECIFIC RETROGRADE TRANSPORT OF NERVE GROWTH FACTOR (NGF) BY MOTONEURONS OF THE BRAINSTEM AND SPINAL CORD ............. 145 Introduction ............................ 145 Methods. . . . . . . . . 14 7 Results ................................. 150 Discussion. . . . . . . . 167 VII. THE RESPONSE OF CULTURED TRIGEMINAL AND SPINAL CORD MOTONEURONS TO NERVE GROWTH FACTOR ....................... 182 Introduction ............................ 182 Methods..... . . . . . . . 185 Results. . . . . . . . . 191 Discussion. . . . . . . . 2 01 VIII. GENERAL DISCUSSION. . . . . . . 212 Nerve Development ....................... 212 REFERENCES. . . . . . . . . . . . 2 2 5 BIOGRAPHICAL SKETCH .................................... 249 vi
PAGE 7
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STUDIES ON NERVE FIBER OUTGROWTH IN MOTONEURON POPULATIONS OF THE DEVELOPING CHICK EMBRYO By DENISE BETH WAYNE April 1988 Chairman: Marieta B. Heaton Major Department: Medical Sciences (Neuroscience) The organization of motoneurons in the mature oculomotor nucleus of the chick brainstem was examined. The retrograde label horseradish peroxidase was injected into each of the four extraocular muscles innervated by the oculomotor neurons. The results established that the segregation of motoneurons into subnuclei reflects the particular muscle which that cell group innervates. The course of oculomotor nerve outgrowth and subnuclear organization were examined following embryonic removal of extraocular target muscles. It was observed that (1) the path of nerve outgrowth was not a function of vii
PAGE 8
the presence of appropriate target muscles and ( 2) the nucleus forms appropriate morphological subnuclear groups in the absence of target muscle. The organization of the motoneurons is not initiated by contact with specific muscle but seems to be intrinsic to the motoneuron population. The disruption of the path of nerve outgrowth in the presence of muscle remnant suggests that there are multiple cues distributed along the route of nerve travel which guide it to the target. An initial tissue culture study compared neurite growth from chick trigeminal neural tube explants in defined and serum-supplemented media. Subsequently, neurite outgrowth in the presence of medium conditioned by age-matched appropriate (jaw) and inappropriate (limb) target muscle was quantified. Medium conditioned by appropriate muscle produced significantly greater outgrowth than both medium conditioned by inappropriate muscle and unsupplemented control medium. This indicates a more specific effect of muscle produced factors on neurite outgrowth than has previously been reported. The final two studies investigated the role of nerve growth factor (NGF) in the development of trigeminal and spinal cord motoneurons. The specific retrograde transport of NGF by motoneurons transiently during early development was demonstrated following peripheral injections of 125rNGF. Additionally, early dissociates of the motoneuron viii
PAGE 9
populations responded to NGF in vitro. The quantity of neurite and the rate of neurite initiation were enhanced in the presence of NGF in spinal cord and trigeminal motoneurons, respectively. ix
PAGE 10
CHAPTER I GENERAL INTRODUCTION Until the turn of the century the catenary theory had maintained that the nervous system was one continuous syncytium, a single sprawling web. As it became clear that the nervous system was composed of individual elements, the question which emerged was how the appropriate pattern of connections was faithfully reproduced during the course of development. Harrison's (1910, 1912) observations on isolated neurons in culture established unequivocally that the axon was a process extended from the neuronal cell body. During in situ development, the axonal process must be elaborated over long distances, following a patented course and reaching a specific target. The cellular mechanisms which guide axons and result in the formation of specific connections have been the subject of many theoretical propositions and experimental investigations. The knowledge and insights provided by these efforts establish the foundation upon which the present work is based. The discussion which follows considers previous works which have examined: the nature and role of specific factors in axonal outgrowth, the organization of nerve pathways, the mechanisms whereby an axon elongates and, more 1
PAGE 11
specifically, the development of peripheral nerves. The experimental work in the present dissertation is concerned with the organization and development of chick motoneuron populations of both the brainstem and spinal cord. The influence of target-derived factors on motoneuron process growth is of particular interest. Theoretical Considerations 2 A unified axon guidance theory provides a conceptually useful model. It is, however, a simplistic reduction of a complex process. Unlike a concrete highway which is a fixed structure, the elements which contribute to form the nerve path are not immutably bonded together. Thus, the pathway for nerve growth is not an entity itself but the juncture of various factors which create a favorable path for outgrowth. The pathway is likely to be the product of multiple factors which, additionally, are likely to be changing continually along the course of the pathway. The influence of any factor may change in intensity as the axon grows and influential factors may be added or deleted from the repertoire. Thus, the pathway is quite likely a dynamic structure and this must be kept in mind when considering the significance of any single theory or factor. Whereas the pathway provides the extra-axonal cue to growth, the axonal sensitivity is also critical to the selective growth response. The characteristics of the axonal membrane which may mediate a response will be discussed
PAGE 12
later. The extra-axonal cues and the distribution of cues will be addressed first. The types of cues which might be available to the axon in its growth can be categorized as: mechanical, electrical and chemical. Whereas these cues have been separated for clarity in the list above and the discussion below, within the realm of the developing nervous system they must certainly act in concert. Mechanical Cues Contact Guidance 3 A physical substrate for axonal elongation is an elemental requirement for the elaboration of neuronal processes. Harrison's (1910, 1912) pioneering work in tissue culture demonstrated that neurons elaborated processes only while in contact with a solid support. Neural explants that were suspended in a hanging drop of salt solution with no surfaces available for contact did not extend processes. With a solid surface available, the fibers grew along the path provided and thereby mirrored the pattern of the solid. This simple thigmotaxis of nerve fibers observed by Harrison was later expanded by Weiss. He found that fibers not only grew along the path described by a solid substratum but within a continuous solid substratum they were directed along the lines of organization within the substratum (Weiss, 1934). Thus, fibers were observed to grow along the lines of tension applied to a drying plasma
PAGE 13
clot. This "contact guidance" described by Weiss has later been found to occur at the microscopic level. For example, neurite growth on a collagen substratum occurs preferentially along the axis of polymerization of the collagen fibrils (Ebendal, 1976). 4 Weiss' (1934) accumulated observations argued strongly against chemical or electrical trophic influences directly affecting outgrowth from neurons. He observed no directed growth in response to diffusible substances from target tissues. In addition, he argued that physical and electrical environmental influences oriented fiber growth by acting indirectly on the organization of the substratum. Weiss maintained that the mechanical ordering of the environmental substratum was the single denominator common to varied influences in directing axonal outgrowth. Invoking contact guidance to explain all forms of directed neurite outgrowth has proven to be a provincial view of the range of factors to which the growing neurite responds. Neuronal processes have been observed to associate with non-neural cells in vitro (glia; Grainger and James, 1970; glia and fibroblast, Fallon and Raff, 1982) and in vivo (glia; Silver et al., 1982). Although the cells do provide a physical substrate for axonal growth, the specificity of the associations suggests that the neuron-cell contact is more complex than the simple provision of a solid substrate for growth. Spinal cord and retinal neurites displayed minimal
PAGE 14
5 growth on a monolayer of fibroblast cells but extended lengthy and abundant processes on a monolayer of astrocytes (Fallon and Raff, 1982). The cell surface rather than some diffusible molecule produced by the cell has been indicated as the mediator of this growth response. In vivo, Silver et al. (1982) examined the growth of callosal fibers which cross between the hemispheres. In embryonic mice the growing fibers are preceded in time by the appearance of a pathway of glia which form a band of cells between the hemispheres (just rostral to the lamina terminalis). This band or "sling" of glial cells is a transient structure which disappears after the pioneering fibers of the corpus callosum cross into the opposing hemisphere. In genetically acallosal mice this transient glial bridge never appears. Disruption of the bridge in normal mice by embryonic surgery results in an acallosal brain with the formation of large neuromas bilaterally. If, in addition to the interhemispheric cut of the glial bridge, a substitute inorganic bridge is implanted, the callosal fibers cross only if the glial cells first migrate across the inorganic bridge (Silver, 1983). The normal configuration of glial cells is, thus, necessary for the continued growth of the callosal fibers. The association of growing axons and neurites with nonneural cells reported above demonstrated a definite and interesting specificity. Similarly, the fasciculation of
PAGE 15
neural processes with other axons in vivo and in vitro exhibits some degree of selectivity and is discussed below. 6 The axonal processes of neurons may provide a path for growth of other fibers. Speidel (1933) was able to observe in vivo axonal growth in superficial regions of the tail of live tadpoles. He observed growth cones advancing along neuronal fibers that had preceded them. Ingrowing sensory axons in Drosophila wing fasciculate along the earliest growing pioneer fibers in growth to the CNS (Palka and Ghysen, 1982). Following experimental implant of barriers in chick embryos, the deflected trigeminal nerve sometimes fasciculates along the fibers of the abducens nerve (Moody and Heaton, 1983c). The examples are numerous: axons can and do grow along the tracts provided by other fibers. However, axons discriminate and do not grow along any and all fibers available. Ectopic axons of the auditory nerve in Xenopus cross perpendicular to optic tract axons and yet retain their own distinct orientation (ConstantinePaton, 1983). The fasciculation of axons is also selective in vitro (Nakajima, 1965). Neurites of retinal and sympathetic explants go to dramatic lengths to avoid contact when confronted with the other, turning, ceasing elongation or elongating on another plane (Bray et al., 1980). Axons may be guided as much by a region which prohibits growth as by a region which favors growth. Olfactory tract axons growing rostro-caudally and optic tract axons growing
PAGE 16
7 latero-medially both pass in close proximity to the diencephalic-telencephalic (OT) border but fail to cross it. Close inspection of the OT junction in the embryonic mouse and chick reveals the presence of a dense core of cells with little extracellular space (Silver, 1984; Silver et al., 1987). This "glial knot" region seems to prohibit growth in two respects: (1) it provides a physical barrier to growth and (2) it deprives axons of the neural-cell adhesion molecule (N-CAM) and extracellular matrix proteins present in adjacent areas where growth occurs (Silver et al., 1987). Both the physical barrier and the specific molecular characteristics of the substratum appear to be factors in the inhibition of fiber growth across this region. In sum, the normal surface character of cells and axons appear to provide specific substrates for axonal elongation, not merely solid support. Extracellular Space An alternate type of mechanical guidance has been suggested by Singer's Blueprint Hypothesis (Singer et al., 1979). Observations on the developing and regenerating spinal cord of the newt reveal the existence of extracellular channels before axonal ingrowth. These spaces are bounded by epithelial cells. It is suggested that a channel pattern arises during ontogeny and contains chemical "trace pathways" which together guide tract formation. One candidate for the role of a "trace" molecule is the neural
PAGE 17
8 cell adhesion moleculue (N-CAM) which seems to play a role in axonal fasciculation (Rutishauser et al., 1978b) as well as neuronal adhesion to glial (Silver and Rutishauser, 1984) and muscle cells (Grumet et al., 1982; Rutishauser et al., 1983). The N-CAM has been found to increase in intensity along the endfeet of neuroepithelial cells in Xenopus spinal cord prior to the axonal ingrowth which occurs in this region (Balak et al., 1987). Components of the extracellular matrix as well as other surface associated molecules may be additional components of the "trace pathway." A fixed blueprint of axon channels would not account for the normal growth of axons which enter the central nervous system at abnormal locations (Constantine-Paton, 1983; Giorgi and van der Loos, 1978). In these cases, information for appropriate growth seems to be widely distributed through the nervous system and not available only in rigidly select channels. The Blueprint hypothesis may accurately represent the existence of general pathways available for growth early in the development of long fiber tracts. The axon channels and associated "trace" molecules may be only one component of a more complex system which leads to appropriate connectivity, and the specific nature of the components may vary throughout the nervous system.
PAGE 18
9 Spatiotemporal Sorting Another means whereby specific patterning is hypothesized to arise is by spatiotemporal sorting of axons. Thus, the arrangement of axons within a fiber bundle would be determined by the order of cell differentiation and axonal outgrowth. Sequentially arising axons would assume a position adjacent to those formed immediately prior to them, thus ordering axons within a nerve. The same would be true at the target with the first arriving axon occupying the nearest synaptic vacancy. The majority of data supporting this mechanism for patterning comes from studies of the developing retinotectal system. Retinal cells are generated in a central to peripheral sequence and a corresponding relationship is maintained by their axons in the retinal fiber layer. It has been suggested (Rager, 1980) that the preservation of axonal order in the optic nerve and tract does not require sorting. Contact guidance may serve to passively funnel axons through intercellular channels and "oriented glial partitions." At the target site the first arriving ganglion cell axons seem to bypass vacant tectal target areas to actively seek specific tectal sites. However, Golgi analysis reveals that the pattern of ganglion axon innervation of the tectum seems to follow a temporal gradient provided by the maturation of tectal cells (Rager and von Oeyenhauser, 1979). The evidence is consistent, then, with a spatiotemporal gradient of cell differentiation
PAGE 19
10 producing order within this system. This evidence does not exclude the possibility, however, that there are other active cues. There might also be a chemospecific interaction of fibers with either oriented glial cells or tectal cells. The spatiotemporal theory of nerve sorting has also been applied to peripheral nerve patterning (Horder, 1978). Some of the evidence supports a selection mechanism for axons in the periphery based on position, but there is also much data which indicates the presence of other, more specific mechanisms. Experimental inversion of the chick limb bud along the dorso-ventral axis has resulted in an incorrect pattern of innervation of limb muscles (Summerbell and Stirling, 1981; Whitelaw and Hollyday, 1983c). The nerve fibers apparently maintain their old position and follow the new pathway in a non-selective manner. The axons are unable to re-direct their growth and reach the appropriate target. Ferguson (1983) has reported appropriate innervation using the same experimental paradigm with the exception that she rotated the limb more proximally. When the rotation included the proximal nerve plexus the fibers were able to respond to the path changes. The axons shift their position within the plexus to innervate the normally appropriate target muscle. This active rearrangement of fibers is evidence against a merely passive channeling of fibers. In normal chick development too, the fibers innervating the limb muscles selectively re-organize in the plexus and nerve fascicles as
PAGE 20
11 they course to the target (Lance-Jones and Landmesser, 1981a; Ferguson, 1983). The axons do not maintain a consistent position within the nerve along their proximal to distal course. This specific rearrangement of axons in both normal and experimentally altered situations contradicts what would be expected with a passive spatiotemporal sorting. The physical channeling of non-specified axons may be an important part of normal fiber growth but this process alone does not seem sufficient to account for the specificity of nerve growth. Again, evidence in support of one guidance mechanism does not necessarily exclude the possibility that there are other mechanisms acting concurrently. Electrical Cues An early theory which proposed a role for endogenous electrical activity in mediating neuronal development was Bok's concept of stimulogenous fibrillation (1915). This theory suggested that as growing fiber tracts passed neuroblasts, electrical activity in the growing axons stimulated axonal outgrowth from the neuroblasts. The theory of neurobiotaxis elaborated by Ariens-Kappers (1917, 1921) was an ambitious attempt to explain much of neuronal development in terms of the influence of currents on different elements of the neuron. It was purported that axons were directed away from, while dendrites and cell bodies moved towards, the extracellular negative charge
PAGE 21
12 (cathode) created by growing fiber tracts. This polarization of the neuron as well as the formation of multiple dendrites versus a single axon were all attributed to specific influences of current from fiber tracts on the neuron. As early as 1920, Ingvar reported the orientation of neurite growth in response to weak electric currents applied in culture. Weiss (1934), however, contended that the neurite orientation was due to the current's influence on the alignment of fibrils in the plasma clot substratum rather than to a direct effect on the neurites themselves. Marsh and Beams (1946) found that applied currents produced an orientation of the neurites independent of any orienting effect that the current might have on the plasma clot substratum. At sufficient current densities neurite outgrowth was enhanced toward the cathode and suppressed toward the anode. Jaffee and Poo (1979) recently replicated this experiment and suggested that the directed growth was caused by the electrophoretic redistribution of membrane glycoproteins under an applied current. The fluoresceintagged lectin, concanavalin A (Con A), has been used to label surface glycoproteins of isolated neurons (Patel and Poo, 1982) and muscle cells (Poo and Robinson, 1977) in culture. Con-A labeling has revealed a concentration of surface membrane glycoproteins on the cathodal cell face following exposure to an electric field. Additionally, the presence of Con A in the medium during neuronal exposure to
PAGE 22
13 an electric field blocks both the directed growth of neurites and the concentration of surface Con-A receptors on the cathode directed membrane (Patel and Poo, 1982). These surface glycoproteins could be important to neurite elongation by acting in membrane-substratum adhesion or as receptors to growth factors in the medium. Endogenous electrical currents have been measured in the region of the developing limb of Xenopus (Robinson, 1983). There are also electrical fields produced by non-neural cells in vivo. These fields are present during early development of the nervous system and could possibly influence nerve development in a manner similar to the growth effects seen in vitro. An alternate interpretation is that electrical current in vitro is an artificial means of stimulating the neurite's normal growth mechanism. The appropriate stimulus in vivo may actually be binding of extracellular molecules to these Con-A binding surface glycoproteins which then results in enhanced neurite growth at the binding locus. Chemical Cues It is possible that molecular cues external to the neuron may serve to guide axonal growth, thus manifesting a chemotactic phenomenon. The salient molecular cues could reasonably exist in several forms, e.g., freely diffusible molecules, substrate bound molecules, or molecules integral to a cell membrane. The neuronal response to chemotactic cues which results in a directed growth of neural processes
PAGE 23
14 can theoretically be categorized in one of two ways. The response could be classified as a trophic or as a tropic response type. A trophic stimulus provides a nutritive or metabolically favorable climate for growth. This trophic stimulation of neural growth could produce an apparent selective path choice by increasing the amount or rate of growth in one direction thus favoring that path above others. A tropic response is characterized by a tendency to turn in response to a particular stimulus, that is, a qualitative change in orientation. It is clear that either chemotrophism or chemotropism could produce directed axonal growth along specific pathways. A purely trophic stimulus may be thought of as affecting the quantity of neurite growth while a purely tropic stimulus produces only a qualitative change affecting the direction of neurite growth. This clarification is important in reaching an understanding of the cellular mechanism which responds to the cue. However, it is quite possible that a single molecule such as nerve growth factor (NGF) is capable of eliciting both a tropic and a trophic cellular responses. Early Theories The role that molecular cues may play in the regulation of neural development has been discussed since the turn of the century (Ramon y Cajal, 1960). Weiss (1934) presented a spate of evidence which supported mechanical guidance of nerve outgrowth and contended that chemical and electrical
PAGE 24
15 factors had no direct influence on nerve growth. Weiss was an influential proponent of the "resonance theory" which maintained that the establishment of connections in the nervous system was a non-selective event and that selectivity occurred later as a result of neural activity (Weiss, 1936). Sperry's studies on the recovery of function in the visual system of lower vertebrates after nerve section found a selective re-establishment of neural connections despite the functionally maladaptive behavior this yielded (Sperry, 1944, 1945; Attardi and Sperry, 1960, 1963). In these studies, the optic nerve of anurans was cut and the eye rotated prior to nerve regeneration. When vision was restored, the animals behaved as if the visual field had been rotated in precise accordance with the rotation of the eye. This inversion of the visual field was very specific as well an enduring. The result, a persistent maladaptive behavior, is inconsistent with the resonance theory which predicted that activity-induced specification would produce the appropriate functional adjustment over time. Subsequent studies examined optic nerve regeneration in goldfish following optic nerve section combined with partial retinal ablations (Attardi and Sperry, 1960, 1963). The regeneration of fibers proved to be very specific both along the path of nerve growth and in the terminal site selected. The fibers followed the same tract to the tectum and reinnervated the appropriate tectal site, bypassing
PAGE 25
16 inappropriate vacant sites. The chemoaffinity hypothesis elaborated by Sperry suggested that chemotactic recognition phenomena based on biochemical specificities guided nerve growth (Sperry, 1965). The highly selective nerve growth, along the appropriate pathway and to the appropriate synaptic site in the tectum, revealed by these experiments was consistent with a specific cellular chemoaffinity regulating the course of nerve growth. While this regeneration data establishes the specificity of neural connections and is consistent with chemotactic regulation of growth, it does not directly evidence any molecular mediation of neural growth. Nerve Growth Factor (NGF) The discovery of the nerve growth promoting factor, NGF, provided the first direct evidence of a molecular factor capable of influencing neural growth. Bueker (1948) first noted that in chick embryos with transplanted mouse tumors (sarcoma 180), the dorsal root ganglia innervated the tumor and these ganglia were enlarged. Levi-Montalcini and Hamburger (1951) pursued this observation and found that both sensory and sympathetic ganglia responded with fiber growth toward the tumor and an enlargement of the ganglion itself. In addition, it was not just those neurons in immediate contact with the tumor that showed the enhanced growth response. The growth of the tumor on the chorioallantoic membrane, separated from the embryo,
PAGE 26
17 demonstrated unequivocally that the neurons were responding to some diffusible substance emanating from the tumor rather than to some contact mediated influence of the tumor (LeviMontalcini and Hamburger, 1953). The hypertrophy of the sensory and sympathetic ganglia was not due to the presence of an enlarged peripheral field presented by the tumor but due to the effect of a diffusible molecule produced by the tumor. Embryonic sympathetic and neural crest-derived sensory neurons respond to the presence of NGF with abundant fiber outgrowth both in vivo and in vitro (Levi-Montalcini et al., 1954). NGF provides a trophic stimulus to the neurons enhancing anabolic processes (Levi-Montalcini, 1982). In addition, Gundersen and Barrett (1979) have shown that the growth cones of sensory neurons growing in culture turn to follow a gradient of NGF provided by a nearby micropipette. The NGF, then, provides both a trophic and tropic stimulus to the growth of sensory and sympathetic fibers. The history of NGF is a remarkable one (Levi-Montalcini, 1975). Both the existence of NGF itself and the identification of a rich source of the molecule resulted from serendipitous observations combined with keen intuition. The relatively high concentration of NGF in the male mouse submaxillary gland facilitated the characterization of the NGF protein complex which consists of three subunits referred to as a, fi and 't Of the three
PAGE 27
18 subunits, the full biological activity resides in the B subunit with the role of the other two subunits still not clearly resolved. The B subunit of NGF is a Mr 26,500 dimer of two non-covalently associated peptide chains. At physiological concentrations the molecule is apparently present as a dimer and not as the dissociated monomer (Bothwell and Shooter, 1977). The purification of mouse BNGF made possible the production of specific antibodies which blocked the biological activity of NGF in mice. The administration of the antibody to newborn mice resulted in an immunosympathectomy characterized by a virtually complete and enduring destruction of sympathetic neurons (see LeviMontalcini and Angeletti, 1968). Developmental studies of the effects of NGF deprivation have been hampered by the failure of antibodies against mouse NGF to block the activity of chick NGF in vivo. Some success in chronicling the effects of NGF deprivation during development in mammalian species has been accomplished by maternal transfer of antibodies to the embryo following auto-immunization of the mother to NGF (Johnson et al., 1980, 1983). With prenatal exposure to antibodies against NGF there is destruction of two neuronal populations, i.e., sympathetic and neural crest derived-sensory neurons. In the guinea pig, the maternal antibodies are not transferred to the embryo until the latter half of gestation (day 35 of 65-70 day
PAGE 28
19 gestation period). The effects of NGF deprivation on early development, then, have not yet been determined. Recent efforts have been directed at finding an antibody that successfully inactivates NGF in the chick embryo which is accessible throughout development (Ebendal et al., 1986; Meier et al., 1986; Belew and Ebendal, 1986). The NGF receptor was first characterized by binding kinetics which revealed a high affinity or slowly dissociating receptor (type I) and a low affinity or rapidly dissociating receptor (type II). Both receptor types have been found on NGF-responsive cells (sympathetic, Godfrey and Shooter, 1986; sensory, Sutter at al., 1979; PC12, Hosang and Shooter, 1985) while cell types that do not respond to NGF have only the type II receptor (mutant PC12, Green et al., 1986; Schwann cells, Zimmermann and Sutter, 1983). The type I receptor apparently mediates the biological response to NGF by internalization and subsequent transport to the cell body of the NGF-NGF receptor complex (see review, Stach and Perez-Polo, 1987). Enzymatic activity associated with the NGF receptor suggests that NGF binding may result in an active enzyme complex which may then have a role in affecting NGF action (Stach et al., 1986). There is some evidence that the two receptor types are identical in the NGF-binding moiety (Green and Greene, 1986). This is supported by the finding that there is apparently only one gene for the NGF receptor (Chao et al., 1986). The
PAGE 29
20 differential effects produced by NGF binding could possibly result from an additional moiety associated with type I receptor (Hosang and Shooter, 1985) or differing interactions with membrane or cytoplasmic constituents (Schecter and Bothwell, 1981). In the sympathetic nervous system, at least, considerable evidence supports the notion that NGF acts as a retrograde messenger transported from the target organ to the innervating neurons. In vivo blockade of retrograde transport by colchicine application to the post-ganglionic nerves resulted in changes in the neurons of the superior cervical ganglion similar to the changes produced by NGF deprivation (Purves, 1976). Chemical destruction of sympathetic nerve terminals or blockade of axonal transport results in a rapid increase in the levels of NGF in sympathetic target organs with a corresponding decrease in the NGF content of the innervating sympathetic ganglion (Korsching and Thoenen, 1985). Thus, the NGF normally present in the sympathetic ganglia is derived from the target structures. Recent methodological advances have permitted the detection of low concentrations of NGF and the distinction of NGF-localization from NGF-production with probes which detect mRNANGF (Heumann et al., 1984; Whittemore et al., 1986). In adult animals, NGF production is clearly localized to sympathetic target structures and
PAGE 30
21 not to the sympathetic ganglia (Heumann et al., 1984; Shelton and Reichardt, 1984; Bandtlow et al., 1987). In addition to the adult structures, the production of NGF has been described in all Schwann cells of the neonatal rat, i.e. those which ensheath motor as well as sensory and sympathetic fibers (Bandtlow et al., 1987), and embryonic chick skeletal muscle (Hulst and Bennett, 1986). The apparent widespread presence of NGF in fetal rat central nervous system (Ayer-Lelievre et al., 1983) and the more recent demonstration of specific NGF-binding by chick motoneurons of the brainstem and spinal cord transiently during development (Raivich et al., 1985, 1987) suggests that NGF may be important in developmental interactions between the motoneuron and its target musculature. The discovery and characterization of NGF has been important in providing substantiation that endogenous molecular substances could produce specific stimulation of neural growth. The recognition of NGF has stimulated the exploration for other growth factors and an interest in the role that these molecules may play in normal development. In this vein, the trophic effect that the target tissue, muscle, has on motoneuron growth has received much attention. Tissue-derived Factors The selectivity of a growth factor is an important aspect of its character which may determine its
PAGE 31
22 effectiveness in mediating the formation of specific connections in vivo. Co-cultures of normal target tissues with neuronal explants has been shown to specifically enhance the neurite outgrowth from the explants (Ebendal and Jacobson, 1977; Pollack et al., 1981). The effect is specific in that target tissue type and target tissue age were important in the degree of enhancement produced. In vivo, the spatial (tissue type) and temporal (tissue age) availability of a growth factor may operate to produce specific effects during development. The diffusion of a molecule away from its source may produce a concentration gradient, and neurons in vitro have displayed the capability to direct growth to increased concentrations of the peptide NGF (Gundersen and Barrett, 1979). Variations in the sensitivity of neural tissue to available trophic and/or tropic factors could also produce selective growth during development. The ability of the neural explants to respond to target tissue has been shown to vary with age of the explant (Pollack and Muhlach, 1982). At least in this instance, the neural tissue seems to exhibit an intrinsic decline in the potential for growth with increasing age. On the other side of this neuron-muscle trophic/tropic interaction is the array of factors produced by muscle. Similar to the changing growth capabilities of the neural tissue, there does seem to be a developmental change in the
PAGE 32
23 effectiveness of muscle-produced factors in supporting neuronal survival and differentiation (Heaton and Kemperman, 1987). There is conflicting evidence regarding whether muscle is increasingly or decreasingly effective at producing a neurite promoting factor(s) with age (Nurcombe and Bennett, 1983; Eagleson and Bennett, 1986; Heaton and Paiva, 1986). The work from Bennett's laboratory suggests that motoneurons are initially dependent on an astrocyte produced factor for survival and later become dependent on a muscle derived factor (Eagleson and Bennett, 1986). The stimulation of neurite formation and enhanced survival produced by muscle conditioned medium (MCM) increases with muscle differentiation from myoblast to myotube (Nurcombe and Bennett, 1983). In contrast, other studies using explants of chick brainstem and frog spinal cord have shown that target from early stages which coincide with initial axon outgrowth in vivo produce greater enhancement of neurite outgrowth than older target tissue (Heaton and Kemperman, 1987; Pollack and Muhlach, 1981). Species (chick versus frog) and regional (spinal cord versus brainstem) differences may account for the observed differences in the neurite promoting effects of muscle-derived factors with age. Smith and Appel (1983) found that extracts of neonatal rat skeletal muscle stimulate both neurite outgrowth and cholinergic activity in dissociated ventral spinal cord
PAGE 33
24 neurons. The extract stimulation was specific for ventral cord neurons as demonstrated by the control levels of response from dorsal cord neurons. With muscle of increasing postnatal ages, the stimulation of neurite outgrowth by muscle extract decreased while the cholinergic activity was stimulated equally by all ages of muscle extract. There is apparently more than one specific developmental change in the trophic influence which muscle exerts on motoneurons. While the ability of muscle-derived factors to enhance neurite outgrowth appears to decline from mid-development onwards it is still apparent at birth. Several studies have found that purified populations of embryonic motoneurons respond to a factor derived from embryonic (day 11 and day 18 chick) or newborn (rat) muscle with enhanced neurite outgrowth (rat: Smith et al., 1986; chick: Calof and Reichardt, 1984; Dohrmann et al., 1986). The selectivity of target-derived factors in enhancing neurite outgrowth has been more clearly delineated in vitro by the choice of trigeminal sensory axons for growth toward the specific region of epithelium normally innervated in vivo (Lumsden and Davies, 1986). The developmental specificity of these trophic effects strongly supports a role for these factors in developmental events in vivo. The factors which stimulate neurite growth have been characterized in vitro by the nature of the biological response and their physicochemical properties.
PAGE 34
25 A number of analyses have revealed that conditioned medium (CM) generated by a variety of cell types is comprised of substrate-binding and soluble components (Adler et al., 1981). Collins (1978a) was the first to demonstrate that it was a substrate-bound component of heart cell conditioned medium (HCM) that promoted neurite outgrowth from ciliary ganglion neurons. This component remained bound after adsorbed to the substrate and the CM depleted of this component did not, by itself, stimulate neurite outgrowth. The neurite-promoting component which adheres to the substrate is, additionally, only effective when it is associated with the substratum (Adler and Varon, 1980). The binding of the factor to the substratum along with direct observation of an increased filopodial adherence to the substratum, led Collins to suggest that the neurite promoting factor in HCM acts by increasing the adhesivity of the substrate. Collins and Garrett (1980) then presented neurites with a choice of substrates coated with HCM or uncoated substrates and found that neurites followed the contours of the coated substrate and only rarely strayed onto the uncoated regions. This characterizes a tropic rather than trophic response because the neurites oriented to remain on the HCM-substrate rather than retracting or failing to survive on non-coated substrate. The substrate-bound neurite promoting factor (NPF) has been isolated from both muscle conditioned medium (Collins,
PAGE 35
26 1978a, 1978b; Collins and Garrett, 1980) and co-cultured premuscle tissue (Nurcombe and Bennett, 1983). This adsorbable component supports neurite outgrowth but not long-term neuronal survival. The non-substrate-bound component of CM, distinct from the adsorbable component, supports neuronal survival (Collins, 1978a; Adler and Varon, 1980; Nurcombe and Bennett, 1983) and promotes continued elongation of neurites once they have been initiated (Collins and Dawson, 1982). It is possible that neural sensitivity to a trophic factor could be exogenously regulated by diffusible, substrate-bound or cell-attached molecular cues. The sensitivity of the response of sympathetic neurons to a trophic factor, NGF, can be altered by such external stimuli. Sympathetic neurons exhibit an increased sensitivity to NGF when the culture substrate has been preexposed to HCM (Edgar and Thoenen, 1982). The HCM alone has no effect on sympathetic neuron survival. Some component of HCM absorbs to the polycationic substrate and modifies the response of sympathetic neurons to NGF. This could conceivably occur by receptor induction or by selective survival of cells with previously silent NGF receptors. Weill (1983) has reported an induction of androgen receptors on dissociated spinal cord cells exposed to muscle CM in culture. It is possible that multiple trophic factors interact during development to stimulate and direct neurite
PAGE 36
27 outgrowth. A trophic factor may act not to stimulate growth directly but to alter cellular sensitivity to another factor or stimulus. Stimulation of neurite outgrowth from motoneurons has been produced by factors isolated from a variety of tissue sources including, but not limited to, muscle. Neurite promoting activity has been demonstrated by exposure of motoneurons to extracts of muscle tissue (Hsu et al., 1982, 1984; Smith and Appel, 1983) and medium conditioned by a variety of cell types [embryonic rat lung, fibroblast and muscle (Dribin and Barrett, 1980; 1982); embryonic chick muscle, liver, and skin (Henderson et al., 1981); bovine vascular endothelial cells, bovine vascular smooth muscle, bovine adrenal cortical cells, human skin fibroblasts, bovine corneal endothelial cells, c2-mouse skeletal muscle, PTK-1 kangaroo rat epithelium, A-431 human carcinoma, N-18 mouse neuroblastoma (Lander et al., 1982)]. These studies have demonstrated production of neurite promoting factors but not without some conflicting results. While some investigators found that chick embryo heart muscle (Collins, 1980) and rat lung, fibroblast and muscle tissue (Dribin and Barrett, 1980) all produced effective CM, Henderson et al. (1981) found that CM produced by chick embryo heart and lung did not enhance neurite outgrowth in day 4 chick neural tube dissociates. Attempts to reconcile these differences serves to underscore the difficulty in comparing results
PAGE 37
28 derived from experiments using different species, different assays of neurite response, different substrates and different media for growth. Characterization of Neurite-Promoting Factor While the differing results in production of neurite promoting factors by various tissue types have yet to be reconciled, the characterization of the active factors isolated from the different sources is more consistent. The elimination of activity by exposure to trypsin (Dribin and Barrett, 1982; Henderson et al., 1981; Collins, 1978a; Smith and Appel, 1983) indicates that the factor is likely to be, or to contain, a protein. A majority of the biological activity consistently appears with a molecular weight around 50,000 daltons. More specifically, antibodies to NGF fail to block the neurite promoting activity, thus certifying that the active factor is not NGF (Dribin nd Barrett, 1980; Smith and Appel, 1983). The neurite promoting factor(s) was found to be negatively charged at neutral pH and insensitive to RNase, DNase, collagenase, and neuraminidase (sialic acid residues are not important to activity) (Dribin and Barrett, 1982; Lander et al., 1982). Lander et al. (1982) have provided a detailed characterization of the CM factor produced by corneal endothelial cells which is consistent with the results described above. The factor is identified as a heparan sulfate proteoglycan. Proteoglycans are polyanions which will bind well to positively charged
PAGE 38
29 substrates. As in other studies, the protein portion of the molecule is vital to the activity. Enzymatic degradation of the protein or pre-exposing the substratum to only the polysaccharide portion of the proteoglycan (heparan sulfate or other purified glycosaminoglycans) yielded no neurite promoting activity. Further investigation showed that the neurite promoting activity in a variety of CM was immunoprecipitated with antibodies to the basement membrane glycoprotein laminin (Lander et al., 1985). Neurons which normally extend their processes in the peripheral nervous system exhibit enhanced neurite growth when cultured on a laminin substratum (Rogers et al., 1983). Laminin is present in vivo during development in the regions of nerve outgrowth in both the brainstem and spinal cord (Riggott and Moody, 1987; Rogers et al., 1986; see section on Molecular Path Cues). In vitro, anti-laminin antibodies inhibit neurite growth on a purified laminin substratum but fail to do so if a CM substratum is used (Edgar, 1985). This may be due to epitope differences in the purified laminin compared to the CM laminin so that antibody binds but is unable to block the region of the molecule important in the promotion of neurite growth. An antibody which is able to block the neurite promoting activity of CM factors suggests that the activity of the natural complex requires the association of laminin with heparan sulfate proteoglycan and possibly other components (Chiu et al., 1986). The nature of the laminin
PAGE 39
and the associated molecules which comprise the neurite promoting activity present in CM produced by a variety of cells have yet to be fully resolved. Cell Surface Molecules 30 The action of diffusible tropic factors as well as the efforts to characterize these factors have been described in preceding sections. Effective action of chemically mediated cues in neuronal development would also require that the neuronal cell membrane possess the capacity to respond appropriately to the factor. Some type of cell surface molecule must be present on the neuronal membrane to respond to the external molecular cue. The distribution of these membrane molecules would be expected to be restricted to certain developmental times, cell types, and cellular loci depending on their particular role in the development of the cell. Monoclonal antibodies and lectin labels have provided the means to examine specific cell surface characteristics. A variety of plant lectins selectively bind to specific disaccharide residues. Binding to neuronal membranes in vivo and in vitro, lectins have been used to describe developmental (timed), topographical (cell type) and topological (cell locus) differences in cell surface glycoconjugates. The binding of several lectins to embryonic mouse retina at sequential stages has revealed a general increase in the intensity of label with development (Blanks and Johnson, 1983). Each lectin exhibited a characteristic
PAGE 40
31 pattern of binding indicative of its affinity for a specific carbohydrate sequence (Lis and Sharon, 1977). Specific cells and retinal layers were differentially labeled with the lectins tested (Blanks and Johnson, 1983). Selective labeling of a sub-population of photoreceptors (likely cones) by peanut agglutinin and the outer but not inner segments of photoreceptors by three of the eight lectins was reported (Blanks and Johnson, 1983). This differential label evidences the distinct glycoprotein composition of these membranes. The surface membrane features of different cell types vary not only in the particular lectins which are bound but, also in the profile of binding by a particular lectin. Pfenninger and Maylie-Pfenninger (1981) provided quantification of lectin receptors by the binding of ferritin conjugated lectins to dissociated cells. Comparison of the binding to explants versus single dissociated cells reveals similar patterns indicating that the membrane integrity is restored to normal after the dissociation procedure. Spinal cord and superior cervical ganglion cells isolated in culture exhibited different patterns of lectin receptors which varied in several ways: the absolute receptor density varied at a given cell region, the local distribution of receptors varied (clustered versus homogeneous) and the profile of changes in receptor density between cell regions varied. Not only are there distinct
PAGE 41
32 differences in surface glycoconjugates between different cell types but there are regional specializations of the neuronal membrane within a given cell. Evidence of this was provided by a comparison of the density of binding of each lectin among three distinct cellular regions: cell body, neurite shaft, and growth cone. The lectin receptors were equally distributed, increased or decreased in the sequence from perikaryon to growth cone. Each lectin demonstrated a distinct binding pattern on each cell type. The regional specialization of the neuronal surface membrane is confirmed as well as the distinct surface identity of different neuronal types. Injections of radiolabeled glycoprotein precursors into a single identified Aplysia neuron has also provided evidence of regional differences in neuronal membrane glycoproteins (Ambron, 1982). Examination of the perikaryal and axonal membranes subsequent to these injections revealed different glycoproteins were specifically incorporated in the two regions. The functional role that surface glycoproteins may play in axonal outgrowth was assessed in goldfish retinal explants exposed to tunicamycin, a protein glycosylation inhibitor (Heacock, 1982). There was a concentration dependent inhibition of neurite outgrowth from the explants with exposure to tunicamycin. While protein synthesis was relatively unaffected, the neurite membrane produced was deficient in carbohydrate as evidenced by
PAGE 42
33 lectin binding. This treatment may produce other undetected deficiencies in addition to depletion of cell surface glycoproteins. Nonetheless, the model does produce impaired neurite outgrowth and strongly suggests that cell surface glycoproteins are critical elements to this normal function. The development and distribution of unique membrane characteristics have also been investigated with the application of monoclonal antibodies. Antibodies which recognize rat peripheral but not central neurons (Vuillamy et at., 1981) and, conversely, antibodies which recognize rat central but not peripheral neurons (Cohen and Selvandran, 1981) have both been described. The labeled antigens appear at distinct points in development yet the functional significance of this distinctive temporal sequence remains undeciphered. In the simpler nervous system of the leech, Hockfield and McKay (1983) described the distribution of axons and cell bodies labeled with several different antibodies. The axonal position in the interganglionic connectives delineated with these antibodies is consistent and predictable. Thus the axon, like the cell body, is specifically and reproducibly distributed in the nervous system. The distribution of perikarya and axons following single antibody administration suggests that the position of the cell body does not confer a similar positioning to the axon in the connectives. Quite possibly, different cell surface determinants dictate the position of
PAGE 43
34 the cell body and the position of the axon. The three different antibodies that Hockfield and McKay (1983) applied all labeled different subsets of cells and axons. Specific labeling of subgroups of cells and axons has also been described in the mammalian CNS (cat: McKay and Hockfield, 1982). The molecular heterogeneity of cell membranes is thus a fundamental characteristic of the nervous system. There is evidence to suggest that at least some of these cell surface antigens are important in the elaboration of processes by neurons. Henke-Fahle and Bonhoeffer (1983) have produced an antibody to chick embryo retinal cells which binds preferentially to retinal plexiform layers in vivo. When this antibody is applied in culture it inhibits neurite outgrowth from retinal explants. This effect is specific for retinal neurites as demonstrated by the lack of effect on outgrowth from dorsal root ganglion explants. The simplicity of the insect nervous system along with the experimental removal of identified cells by laser ablation have permitted important observations to be made regarding specific cellular interactions during axonal outgrowth. The axon of an identified cell (pCC) recognizes specific axons along which the pee axon normally fasciculates such that, in the absence of this axonal pathway, growth of the pee cell either stops or progresses randomly (Bastiani et al.,1986; du Lac et al., 1986). An experimental paradigm which creates a temporal delay in the
PAGE 44
35 arrival of the axonal pathway results in a cessation of the neuronal process extension until the appropriate substratum for growth arrives (Bastiani et al., 1986). The specific fasciculation of fibers along these "labeled pathways" is likely to be a fundamental feature of construction which also contributes in the elaboration of more complex nervous systems. In more complex nervous systems, however, nerve growth seems to show a hierarchy of substrate preferences rather than the absolute pathway preference that grasshopper neurons have shown. In mammals, groups of neurons may play the role represented by one cell in the grasshopper so that observations on an isolated cell population has not yet been possible in these complex systems. Several glycoproteins have recently been identified on neurons in the grasshopper and Drosophila which have a limited axonal distribution during development (grasshopper, fasciclin I and II, Bastiani et al.,1987; Drosophila, fasciclin III, Patel et al., 1987). The expression of these glycoproteins on a subset of cells and restricted to specific segments of the axonal membrane make them likely candidates for molecular mediators of specific axonal interactions which occur during development. Neural Cell Adhesion Molecule (N-CAM) Another cell surface molecule which has been partially characterized, neural cell adhesion molecule (N-CAM), is more general in its distribution and effects. N-CAM is a
PAGE 45
36 cell surface glycoprotein. It is present on all neurons and cells that are precursors of striated muscle. This is a very liberal distribution both in terms of cell type and cell locus. Specificity in its effect could be achieved by differences in N-CAM density, different molecular partners for N-CAM in different places, or heterogeneity in the structure of N-CAM itself. Heterogeneity in N-CAM structure has been described both between tissue (nerve and muscle, Rutishauser et al., 1983) and between developmental ages (embryonic and adult, Edelman and Chuong, 1982). N-CAM is apparently important in cell-cell adhesion mechanisms (Rutishauser et al., 1978a; Buskirk et al., 1980), neuriteneurite adhesion mechanisms (Rutishauser et al., 1978b) and neurite-cell adhesion mechanisms (Grumet et al., 1983; Rutishauser et al., 1983). In vitro adhesion of these elements has been disrupted with antibodies to N-CAM. AntiN-CAM disrupts the normally distinct retinal cell layers in vitro (cell-cell), fasciculation of axons together in vitro (neurite-neurite) and the adhesion of spinal cord neurites to muscle cells (neurite-cell) (Edelman, 1983). N-CAM appears to act as an adhesive by covalent binding to another N-CAM molecule. This bonding is enhanced by removal of sialic acid residues which probably hinder access to the binding site. N-CAM binding is specific and does not function in the adhesion of spinal cord neurites to the collagen substratum or to fibroblast cells (Rutishauser et
PAGE 46
37 at., 1983). While anti-N-CAM disrupts neurite fasciculation by inhibition of neurite-neurite adhesion, neurite outgrowth is not affected and growth cones appear to operate as normal. N-CAM is a cell surface molecule which mediates adhesion between cell membranes. The importance of N-CAM in binding neural membranes to muscle but not to fibroblast cells reveals a selectivity which could provide direction during normal development. N-CAM might also be a contributing force in the selectivity which operates in vivo to produce specific nerve branching patterns and specific neuron to muscle contacts. While N-CAM is apparently not critical to elongation, direction and elongation may be independently regulated events in vivo. A variety of cell surface molecules mediating cell interactions with cells, subgroups of cells and extracellular molecules might well be imagined to act in concert to produce specific neuronal networks in vivo. Axonal Elongation Trophic factors have been shown to enhance neurite outgrowth. How trophic factors influence the growth mechanism is a question of considerable interest. Neurite elongation in vitro has revealed much of the normal morphology and ultrastructure which subserve the elongation mechanism. Neurites extended in culture have demonstrated that elongation occurs at the growing neurite tip, the growth cone, rather than at the proximal portion of the
PAGE 47
38 neurite shaft close to the cell body (Bray, 1970). The growth cone is the site of an active exchange with the environment evidenced, in part, by the presence of pinocytotic vesicles (Hughes, 1933). The growth cone is a large membranous expansion observed at the tip of processes growing both in vitro and in vivo. Filopodia or microspikes and sheet-like veils (lamellipodia) rapidly protrude from and retract into the margins of the growth cone. These dynamic elements of the growth cone may act to probe the immediate environment and/or they may serve in advancing the growth cone. Letourneau (1975) has documented the importance of an adhesive substrate to neurite formation. Using a simple airblast to assay the degree of adhesion to the substrate, he found that a firm adherence to the substrate increases the probability of axon initiation, rate of elongation and degree of axonal branching. The morphology of neurons grown on highly adhesive substrates is distinct (Letourneau, 1979). These neurons display crooked neurites and growth cones that are flat and expanded with longer and more numerous microspikes. Collins (1978b) suggested that the conditioned medium (CM) factor which enhances parasympathetic neurite outgrowth does so by a specific increase in substrate adhesivity. This was based, in part, on the observation that the first visible response to CM was attachment of formerly non-adhering filopodia. Anchoring the growth cone to the substrate may play multiple roles in
PAGE 48
39 affecting elongation. It may permit the addition or sequestration of cell membrane for surface expansion at that location. Concurrently, adhesion may be necessary for transmembrane events which trigger the assembly or stabilization of microfilament bundles. Tosney and Wessells (1983) have observed microfilament bundles only in those microspikes which are adherent to the substrate. Trophic factors produce specific changes in neurite growth, enhancing outgrowth from particular cell types. Substrate adhesivity is clearly one important element in achieving neurite outgrowth. Whether a factor acts only to increase adhesion to the substrate or also specifically interacts with some other element of the elongation mechanism has not been determined. It is possible that adhesion may be separate from an additional and more specific means of enhancing neurite growth by trophic factors. Studies employing dissociated cells have examined several indices of neurite growth. Neurite initiation, neurite branching and neurite length have all served to reflect the amount of neurite outgrowth produced. It is not clear that these different aspects of neurite growth are regulated equally by all influences. Rogers et al. (1983) reported that as substrate adhesivity increased, neurite initiation increased while neurite length actually decreased. The apparent inhibition of process elongation
PAGE 49
40 which was observed may not be a direct cellular effect but, rather, it may be a counteraction to the enhanced initiation. On the highly adhesive substrata used, the cell may be operating at its maximal capacity for neurite extension. At the maximal limit to overall neurite extension an alteration in the propensity for only one of the growth modes (neurite initiation or neurite elongation) may result in a shift of neuritic materials from one form of growth to another. While this shift of materials is speculative, a recent observation (Roederer and Cohen, 1983) on regenerating neurons in the cricket CNS also suggests that a finite capacity for process extension is budgeted between the regenerating axon and numerous neuritic extensions from the cell soma. The possibility that a declining ability to elongate on the more adhesive substrate is responsible for the observed shift in the mode of outgrowth suggests another means of directing growth in vivo. Factors which inhibit elongation of specific cellular processes could act to direct neurite growth in vivo. In this situation, growth might not be transferred from one mode of extension to another but from one path of growth to another. The cellular regulation of different modes of outgrowth (initiation versus elongation) are not yet well understood.
PAGE 50
41 Cue Distribution Theories The following discussions are concerned with the distribution of cues which guide axons rather than the specific identity of the cues themselves. The determination of nerve pathway organization is common to the three discussions: genetic influences, the substrate pathway hypothesis and peripheral nerve patterning. Genetic Influences In Drosophila and Xenopus the grouping of cells during development according to a common genetic origin has been observed. The progeny of one precursor cell all lie within a certain spatial domain named a compartment. In Drosophila, axons and dendrites do not respect compartment boundaries (Palka and Ghysen, 1982). Genetic mutations which have produced a variety of phenotypes useful for examining neural pathway regulation have revealed a genetic influence on both peripheral and central nerve paths (Palka and Ghysen, 1982; Palka et al., 1983). The peripheral paths that direct axons toward the CNS are preferred paths for axonal growth. These preferred paths are non-discriminant and provide a favored path for axons of any compartment or sensory modality. In contrast, the central paths discriminate amongst axonal types. The neuronal compartment identity which is derived from the genotype (transformed or non-transformed) determines the axonal path and pattern of branching centrally. The peripheral paths are specific in their
PAGE 51
polarity, however, favoring growth directed toward the CNS while central paths are not specifically polarized and permit both rostral and caudal growth. 42 The genetic heritage of both the neuron and the cellular substrate for growth are likely to contribute to determining the ultimate neural patterning produced. Jacobson (1983) has described seven regions of compartmental restriction in the embryonic nervous system of Xenopus. At early cell stages separate groups of blastomeres have been identified whose progeny each constitute one of the seven compartments. Injection of horseradish peroxidase (HRP) into a blastomere contributing to the ventromedial cell group yields labeled cells in the ventral spinal cord, the ventral myotome and other ventral organs in the rostral embryo. Following such labeling with HRP, the course of the earliest growing motoneuron pioneer axons was followed through the myotome (Moody and Jacobson, 1983). Axons of labeled motoneurons make contact preferentially with labeled muscle cells, that is, cells derived from the same common progenitor cell. Thus, it is possible that pioneering axons of motoneurons are guided by following compartmentally related muscle cells. Compartmental distinctions seem to diminish as environmental complexity increases, making genetic identity less likely to serve as a cue in later development. Compartmental relationships, however, might play a significant role in establishing pathways during pioneer
PAGE 52
43 fiber outgrowth. While these results are interesting, in order to extend these observations to other populations in more complex systems better cellular labels are needed. Antibodies which respond to selective cell groups may prove useful in extending these observations. Substrate Pathway Hypothesis An array of cues which arise during ontogeny were proposed by Katz et al. (1980) to provide continuous, specific substrate pathways for both cell migration and axonal elongation. It was proposed that the cues which described these pathways could be derived from a pattern of molecules, cells or developmental events. Both the availability of a cue and the response of a cell to a cue would not necessarily remain static over the time course of development. This theory does not exclude previously proposed mechanisms for neural guidance. The conceptual framework it provides can accommodate a diversity of cues and a diversity of conditions within the nervous system. The model represented the neural tube on a cartographic coordinate system. The paths of the major fiber bundles early in development were mapped onto the neural tube and these served to depict proposed substrate pathways. Experimental manipulations which have produced growth of fiber tracts into the CNS at ectopic loci (Giorgi and Van der Loos, 1978; Levi-Montalcini, 1976; Katz and Lasek, 1979, 1981) evidenced growth within that sector of the neural tube
PAGE 53
44 where the tract normally grew. Katz and Lasek (1981) suggested that the general pattern of substrate pathways within the neural tube confines motor tracts to the ventral basal plate and sensory tracts to the dorsolateral alar plate. There are already amendments to this rule of ordering as Constantine-Paton (1983) has noted the dorso-radial path taken by both normal and ectopic sensory fibers of the auditory (VIII) nerve. The details of the substrate pathway hypothesis may require many amendments and revisions to accurately depict the entirety of nervous system development. Taken as a whole, the theory integrates multiple guidance mechanisms operating in a dynamic environment into one coordinated and comprehensible system. This conceptualization of pathway generation is a useful one for consideration of peripheral nerve patterning. Peripheral Nerve Patterning Motoneurons innervating skeletal muscles reside in the basal plate of the central nervous system and their axons course through a peripheral environment to reach the appropriate target. The development of motor nerves in both the brainstem and spinal cord of the chick embryo are of specific interest to the experiments which comprise this dissertation. The motoneuron populations in the brainstem and spinal cord are similar in many respects. Both are basal plate
PAGE 54
45 populations which arise from a medial cell column subjacent to the proliferative cells of the ventricular zone. While some motoneurons retain this medial position (e.g., brainstem oculomotor nucleus, spinal cord medial motor column), other motoneuron populations move to a more lateral position within the basal plate (e.g., trigeminal lateral motor nucleus). Cell differentiation progresses rostrocaudally such that motoneurons of the brainstem are generated earlier than motoneurons of the spinal cord (e.g., brainstem trigeminal motoneurons generated day 2-3, Heaton and Moody, 1980; spinal cord lumbosacral lateral motor column generated day 3-4, Langman and Haden, 1970). However, motoneurons as a whole differentiate precociously with respect to the rest of the nervous system. Chick brainstem and spinal cord motoneuron populations are similar in organization, too, in that the cells which innervate a specific muscle are grouped together and assume a reliable position within the central nervous system (see Chapter II; Landmesser, 1978a). Brainstem motoneuron populations frequently have an associated cranial sensory ganglion which is analagous to the dorsal root ganglia (DRG) of the spinal cord. This is not always the case, however, and some cranial motor nerves have no associated sensory component (e.g., oculomotor nucleus). In the chick hindlimb, motoneuron axon outgrowth precedes sensory axon outgrowth (Honig, 1982; Tosney and
PAGE 55
46 Landmesser, 1985c) and sensory axons apparently depend on the presence of the motor axons for appropriate formation of the muscle nerve branches (Landmesser and Honig, 1986). In the brainstem trigeminal region the motoneuron axons follow the peripheral sensory fibers to the target region (Riggott and Moody, 1987). Despite the sequence of arrival of the trigeminal fibers, it has not been resolved which population initiates the formation of the muscle nerve branches. Motoneuron populations of the brainstem and spinal cord present several different patterns of outgrowth: motoneuron axon outgrowth precedes, follows, or occurs in the absence of sensory axon outgrowth. Fundamentally these motoneuron systems share many similar features and are likely to use common mechanisms to guide outgrowth. It will be of great interest to see how these systems have adapted to suit the individual needs of outgrowth in each situation. The segmentation imposed on peripheral nerves in the spinal cord may result in additional constraints on nerve outgrowth which are not present in the brainstem. Initial axon outgrowth in the spinal cord is restricted to the anterior half of the opposing somite (Keynes and Stern, 1984; Stern et al., 1986). There is no apparent equivalent in motor nerve outgrowth in the brainstem. Following experimental removal of the somites, the nerves exiting the spinal cord are no longer organized into discrete nerve bundles although the nerves do grow to the target region
PAGE 56
47 (Lewis et al., 1981). The somites do not appear to provide cues necessary for the growth to the target but do impose an organization on the fibers which results in the individual nerve roots. The motor systems of the brainstem and spinal cord are equivalent in many respects. It is likely, then, that the development of these populations is regulated by the same mechanisms although there may be important regional modifications. Specificity Peripheral nerve growth has proven to be specific both in vitro and in vivo. The age of the involved tissues, neural and muscle, has been shown to be important in both nerve and neurite growth. Explants of brainstem and spinal cord motoneuron populations show enhanced neuritic outgrowth in the presence of muscle tissue. The developmental stage of the muscle used to produce extract or conditioned medium has an effect on the degree of neurite outgrowth produced from motoneuron populations (Heaton and Kemperman, 1987; Heaton and Paiva, 1986; Pollack and Muhlach, 1981; Pollack et al., 1981; Nurcombe and Bennett, 1983). While there is some dispute as to whether premuscle tissue (present at initial outgrowth) or myotubes (present during muscle-nerve formation) is maximally effective, it does appear that muscle-derived factors are generally specific to periods of nerve growth. The magnitude of the neuronal response has
PAGE 57
48 been shown to vary with the neuronal stage of development independent of the target age (Pollack and Muhlach, 1981). The time course of neural responsiveness observed is specific for different neuron types examined. Motor (spinal cord) and sensory (dorsal root ganglion) neurons in vivo send nerves to their target at different times in development. This is reflected in the differing stages at which these two types of neural explants show a maximal neurite outgrowth response to co-cultured target tissue (Pollack and Muhlach, 1982). The temporal specificity of tissue effects in vitro suggests that these are not nonspecific growth effects but may be significant for events occurring at the same stage in vivo, namely nerve outgrowth. The substrate pathway hypothesis suggests that the temporally specific expression of cues serves to establish pattern in the developing nervous system. The growth of nerves in vivo has also been shown to be influenced by target tissue age (Swanson and Lewis, 1982). Exchanging a limb bud between an older and younger embryo prior to any limb innervation results in a peripheral nerve branching pattern characteristic of the age of the limb donor, not the host. It seems that the nerve is capable of earlier or faster outgrowth but is normally limited by some limb-specific, developmentally expressed, cue. This guidance cue could be either diffusible or pathway-bound and could act to initiate nerve outgrowth or promote elongation. The
PAGE 58
49 paths available in the limb bud must differ with time so that the course of nerve growth in vivo is in some way regulated by the changing limb conditions. Motoneuron pools to individual muscles are discrete and identifiable in the chick spinal cord (Landmesser, 1978a). Peripheral injections of HRP early in development result in a retrograde label which demonstrates both cells bodies and the distribution of their fibers (Lance-Jones and Landmesser, 1981a). This procedure reveals that fibers of specific cells within the spinal cord take highly specific peripheral routes. The axons begin sorting while still quite remote from their ultimate target. Within the crural and sciatic plexuses of the leg, axons selectively cross, grouping with other axons to establish specific branching patterns. The generation of a distinct pattern seems to depend on two features, selective path cues and unique cellular (axonal) identities. The fibers all respond to the path in a specific manner and the particular response is based on the individual axonal identity. Proximal Nerve Path The organization of proximal plexuses can be affected by changes in the periphery. The presence of a supernumerary limb results in aberrant spinal nerves contributing to a limb plexus while the nerve branching pattern generally appears normal (Lance-Jones and Landmesser, 1981b). Although the gross morphology of plexus and nerve formation is
PAGE 59
50 correct, the muscles are innervated by inappropriate motoneurons and the motoneurons do not show the usual organization within the spinal cord. The motoneuron pools are no longer discrete but are diffusely distributed through many spinal segments. The position of motoneurons in the transverse axis of the spinal cord, however, remains similar to the normal distribution pattern. The general form that the plexus takes is determined by the opposing limb, not the spinal nerves which contribute to the plexus. That is, thoracic nerves normally innervate axial muscles but contribute to the formation of an overtly normal crural, thigh, plexus when opposite a supernumerary leg. The determination of the nerve pattern (plexus formation) by the limb identity is consistent with the previously described exchange of limb buds between young and old embryos. In that case, the peripheral nerve branch pattern was a function of the limb age, not the host spinal segment age. The overall nerve form is dependent on the limb while it seems that the establishment of a specific innervation pattern is dependent on the entry of specific fibers into a particular plexus. Under varied experimental conditions (limb shifts, spinal cord shifts, supernumerary limb; Lance-Jones and Landmesser, 1981b), when inappropriate spinal nerves contribute to a plexus the subsequent muscle innervation is by a non-select population of motoneurons in the contributing spinal cord segments. The axons do not correct
PAGE 60
51 their course in the periphery to find their appropriate target but follow the path dictated by their point of exit from the plexus. It appears in this circumstance that there are cues for nerve growth along certain pathways but no organizational cues relevant to these incorrect axons. At the branch points in this circumstance, the axonal choice of diverging paths appears fairly random. If fibers enter the appropriate plexus but enter from an altered position [following reversals of small segments of spinal cord (Lance-Jones and Landmesser, 1980a), or rotation of limb bud including proximal portions (Ferguson, 1983)], they are able to respond to the change. The axons re-orient within the plexus and take the appropriate path to the correct target. Within the appropriate plexus to which they are specified the axons can alter their course in a selective manner. Distal Nerve Path Distal to the level of fiber sorting in the plexus, the normal neural innervation of target muscle is very specific. Lance-Jones and Landmesser (1980) deleted small cord segments to leave some limb muscles without their normal innervation. The remaining nerves by-passed this uninnervated muscle and only innervated the target muscle appropriate to them. Despite the specificity, inappropriate innervation can be produced under certain experimental conditions.
PAGE 61
52 If the periphery is altered distal to the proximal limb plexus (distal limb bud rotations: Summerbell and Stirling, 1981; Whitelaw and Hollyday, 1983c; addition of limb segments to an existing limb: Whitelaw and Hollyday, 1983b), the axons are largely incapable of responding to the altered target locus. Inappropriate innervation of target muscles results as the nerve continues along its course and seems to innervate muscle based on the position it occupies rather than the muscle identity. This position-dependent inapropriate innervation can be produced with a variety of experimental paradigms. Following a distal limb rotation, the nerves coursing along the dorsal path innervate the ventral muscles which have been rotated to a dorsal position. Supernumerary segments of limb can be experimentally added in series with normal chick limb segments (Whitelaw and Hollyday, 1983b). A resultant limb, for example, would be composed of thigh, calf, calf and foot segments, in that order. The third limb segment, the duplicated calf, was found to be innervated by nerves which normally contact the third limb segment, the foot-nerves. There was a consistent nerve pattern in the different types of duplications. The nerves innervated the first three limb segments in the usual nerve sequence. Limb segment position rather than limb segment identity determined which nerves provided innervation.
PAGE 62
53 On some occasions appropriate muscle innervation will occur despite an initial misdirection of the nerves. Under the experimental conditions which channeled spinal nerves into the inappropriate plexus (supernumerary limb, rostral limb shift, large cord reversal), aberrant nerves were sometimes observed to break away and take a novel path to their appropriate target (Lance-Jones and Landmesser, 1981b). The formation of aberrant nerves did not follow predictable patterns. When they did form, all the axons within an aberrant nerve were normally appropriate for the muscle that it ultimately innervated. No incorrect connections were made through aberrant nerves. The formation of aberrant paths could be imagined to occur through a shift in the balance of cues. The two routes available to an axon in this situation are the incorrect path it is traveling and the novel path to the appropriate muscle. The incorrect path of travel may in some way phygically confine the axon and/or this path may have a decreasing attraction for the axons along its proximo-distal length. The incorrect path presumably has only a weak non-specific signal for those axons at the start. The formation of a novel path may rely on mechanically permissive regions in the environment and/or the presence of specific signals from the appropriate muscle. An alteration in the nerve path could occur through any of the several different combinations of changing conditions. Variability in the formation of aberrant paths
PAGE 63
could arise in this way. It is very clear that specific axons course over novel pathways to reach specific muscle. This makes it seem highly likely that some cue from the muscle is important in actively guiding the nerve. Molecular Path Cues 54 These studies on the altered patterns of nerve growth produced by experimental manipulations in vivo led to the postulate that peripheral nerve pattern was established by two categories of cues: (1) those which establish a general concourse for fiber growth with no selective effect on the growth of specific populations thereby creating a "public highway" for growth and (2) those which provide information to direct the growth of specific subsets of axons (Tosney and Landmesser, 1985c; Tosney et al., 1986). While there are no candidates for such highly selective cues of the second category in vertebrates, several candidates have recently been identified in the simpler insect nervous system (Bastiani et al., 1987; Patel et al., 1987; see section on Cell Surface Molecules). Despite the complexities of the vertebrate, there are several candidate molecules which may provide the "public highway" cues. The growth of peripheral neurites in vitro is enhanced on a laminin substratum (Rogers et al., 1983). In vivo, the distribution of laminin in the brainstem and spinal cord suggests that it could be providing an adhesive substratum for axonal outgrowth during early development
PAGE 64
55 (Rogers et al., 1986; Riggott and Moody, 1987). Recently, Schwann cells have been observed to precede the growing front of motor axons in the chick forelimb (Noakes and Bennett, 1987). Experimental deletion of the Schwann cell precursors by neural crest extirpation has resulted in a disruption of axonal ingrowth into the limb muscles in both the brachial and lumbar regions of chick spinal cord (Carpenter and Hollyday, 1986). In addition to secreting the laminin glycoprotein (Cornbrooks et al., 1983; Palm and Furcht, 1983), Schwann cells secrete the proteolytic enzyme plasminogen activator (Krystock and Seeds, 1984) and nerve growth factor (Bandtlow et al., 1987; Assouline et al., 1985). These molecules may act singly or in concert to provide a path for trailblazing fibers. Evidence suggests that multiple molecular cues mediate the interaction between motoneurons of the ciliary ganglion and embryonic myotubes in vitro (Bixby et al., 1987). Severe inhibition of neurite outgrowth was obs~rved in this system only when the JG22 antibody to cellular extracellular matrix receptors (ECM; likely blocking cell receptors for laminin, fibronectin and collagen) was administered in combination with antibodies to two cell adhesion molecules (neural cell adhesion molecule and neural ca2+_dependent cell adhesion molecule). Thus, multiple cell surface associated molecules appear to be significant in mediating the interaction between the processes of motoneurons and myotubes.
PAGE 65
56 The following experiments are concerned with the elements active in shaping early motoneuron development. The target-independent aspects of early motoneuron development will be differentiated from those aspects of growth which are target-influenced. The regionally specific nature of target-derived factors will be characterized in terms of their growth promoting effects. Finally, evidence will be provided that one of the above mentioned molecules, NGF, is important in early motoneuron development.
PAGE 66
CHAPTER II THE PATTERN OF EXTRAOCULAR INNERVATION BY THE OCULOMOTOR NUCLEUS OF THE CHICK Introduction The oculomotor complex lies in the rostral mesencephalon. These motoneurons are the cell bodies of origin of the third cranial nerve, the oculomotor nerve. The oculomotor complex has been subdivided into four subnuclei on morphological grounds. Three subnuclei, dorsolateral, dorsomedial and ventromedial, innervate four of the six extrinsic muscles of the eye. A midline accumulation of cells between the ventral subnuclei, the central nucleus, is also sometimes described. The accessory subnucleus innervates intrinsic eye musculature (this innervation was not examined within this study). The topographic grouping of motoneuron cell bodies is prevalent in the central nervous system. In the ventral spinal cord, motoneuron cell bodies are arranged in groups corresponding to the musculature of specific body regions which they innervate, such as arm, leg, and axial musculature. The segregation of motoneurons based on muscle innervated has been demonstrated within the oculomotor nucleus of mammalian (Augustine et al., 1981) and avian 57
PAGE 67
58 (Isomura, 1973) species. The distribution of motoneurons within the chick oculomotor nucleus was reported following the use of degeneration techniques to map the projection pattern (Isomura, 1973). The results of this study were somewhat equivocal, however, because only a small percentage of the neurons were identified as degenerating. The pattern of innervation and the origin of the ventromedial subnuclear motoneurons have been of particular interest because of the partially contralateral innervation which this subnucleus provides. The ocutomotor anlagen appears on either side of the midline in the mesencephalon between three and five days of development. Subsequently, from five to ten days, this homogeneous cell group segregates into three distinct subnuclei. During this same developmental period, a portion of the population also migrates medially and, intermingling with the contralateral migratory population, apparently crosses the midline (Puelles-Lopez et al., 1975). This migration, then, establishes the contralateral innervation. Because of the high cell packing density and the intermingling of contralateral populations, it has been difficult to determine the limits of the migratory population. Additionally, the ultimate destination of this population within the contralateral subnucleus has not been resolved. Neither the developmental observations nor degeneration mapping have definitively described the distribution of
PAGE 68
59 motoneurons within the ventromedial subnucleus. The present study was designed to resolve the distribution of the four motoneuron populations within the three subnuclei using the horseradish peroxidase retrograde tracing technique. Methods Fertile White Leghorn chicken eggs were obtained from the Poultry Science Department at the University of Florida. The eggs were set in a forced draft incubator at 37c, 70% humidity for 18 days (total incubation period= 20 days). For injection, the eggs were candled and the shell overlying the air pocket was carefully removed. The egg was secured in a wax egg holder and a drop of water placed on the intact membranes to reveal the course of the blood vessels. The chorioallantoic membrane overlying the head was slit, taking care to avoid cutting any large blood vessels. A forceps could then be used to gently pull the head out. The head was supported on dampened gauze pads. Under the dissecting microscope the right eye was anaesthetized with xylocaine. The eye was enucleated, removing the vitreous, and some of the connective tissue surrounding the eye was cut to permit some movement of the eye and access to the particular muscle to be injected. A solution of 30% HRP (Sigma, type VI) in 0.9 % saline was loaded into a five microliter Hamilton syringe fitted with a 50 m glass cannula tip. The HRP filled syringe was mounted in a Brinkman micromanipulator and visually directed to the belly of the muscle under the
PAGE 69
60 dissecting microscope. From 0.1-0.3 l of HRP solution was injected into the belly of the muscle and this was visually confirmed by the spread of the dark brown HRP within the muscle. Following the injection the cannula was withdrawn and the muscle swabbed to soak up any HRP which had seeped out. An additional precaution against diffusion of leaked HRP involved packing the exposed injection area with sterile gelfoam. The egg was then returned to the incubator using dampened gauze pads to cradle the head and keep the embryo moist. After a survival period of 18-24 hours the embryo was anaesthetized with 0.5 ml of chloral hydrate and perfused intracardially with 5 ml of saline followed by 50 ml of fixative (1.25% glutaraldehyde and 1% paraformaldehyde in O.lM phosphate buffer, pH 7.4). The brain was then dissected out and immersed in O.lM phosphate buffer with 10% sucrose at 4c overnight. The following day the brain was frozen on a sliding microtome and 25 m sections were cut into the 10% sucrose phosphate buffer. Within 24 hours the sections were reacted using the tetramethyl benzidine procedure of Mesulam (1978) which demonstrates the presence of HRP with a dark brown reaction product. The sections were mounted, dried overnight and counterstained with neutral red before coverslipping. The identity of the tissue was coded for analysis so that the injection site was unknown to the investigators.
PAGE 70
61 Each section through the rostro-caudal extent of the nucleus was examined. Both notes and drawings were made to describe the presence, the locus and the extent of the HRP reaction product. The subnuclei are quite distinct in the rostral and mid regions of the nuclear complex. In the caudal region the distinction between the dorsal subnuclei becomes less clear as they coalesce to form a single crescent of cells overlying the dorsal and medial aspect of the medial longitudinal fasciculus. In all regions, except the most caudal portions of the dorsal subnuclei, the labeled cells can be confidently ascribed to a single subnucleus. Results In the adult oculomotor complex the cellular organization is such that there are four subnuclei distinguishable with conventional histological methods (Fig. 2-1). The accessory subnucleus is located most dorsally and consists of small neurons which provide preganglionic parasympathetic innervation of the ciliary muscle and the sphincter muscle of the iris. This is analogous to the Edinger-Westphal nucleus of mammals. This innervation was not examined. The three remaining subnuclei provide innervation to four of the extraocular muscles. The cellular organization is readily seen in the control section in Figure 2-1 taken from the mid region of the nucleus. This organization varies somewhat over the rostrocaudal extent of the complex. The rostral-most extent
PAGE 71
62 Figure 2-1. Normal organization of oculomotor nucleus. Transverse section through the mesencephalic basal plate of a chick at 3 days posthatching. The morpologically distinct subnuclei can be readily identified. a, accessory subnucleus; dl, dorsolateral subnucleus; dm, dorsomedial subnucleus; vm ventromedial subnucleus; v, ventricle. Hematoxylin and eosin stain. X 127.
PAGE 72
63 of the complex is evidenced by the presence of the accessory, dorsolateral, and dorsomedial subnuclei. Slightly caudal to this the ventromedial subnucleus first becomes apparent. Progressing caudally the accessory subnucleus diminishes in size while the ventromedial subnucleus expands. The nerve root is first evident at mid level of the rostrocaudal extent of the complex. In the caudal reaches of the nuclear complex the accessory subnucleus disappears while the ventromedial group becomes more predominant. The dorsomedial and dorsolateral subnuclei fuse in the caudal regions to form a single extended cell group overlying the medial longitudinal fasciculus (MLF). These groups trail off caudally. The ventromedial group is the last to disappear and does so just rostral to the appearance of the trochlear motoneurons which lie lateral to the MLF. An additional central cell group has sometimes been described to lie in the midline between the two ventromedial subnuclei (Nimii et al., 1958). Injection of the inferior rectus muscle produced a consistent pattern of labeling. The HRP reaction product was always in cells on the ipsilateral side and virtually always within the dorsolateral subnucleus. There was an even distribution of labeled cells within this subnucleus and along its rostrocaudal extent. There was only one cell which was found to lie outside this subnucleus. This labeled cell was located along the lateral border of the dorsomedial
PAGE 73
64 subnucleus, that is, immediately adjacent to the boundary of the dorsolateral subnucleus. The cameral lucida drawing in Figure 2-2a represents the typical pattern of labeled cells seen following injections of the right inferior rectus muscle. The photomicrograph in Figure 2-3 shows a typical profile of labeled cells following such an injection. Injection of the medial rectus muscle produced a consistent labeling of the ipsilateral dorsomedial subnucleus. In the rostral portion of the nucleus there were many labeled cells medially. These cells were quite near the midline and distinct from other cells of the complex in their spindle shaped morphology. All other cells labeled were round to ovoid in their morphology. The labeled cells were most prominent in the mid region where there were many cells labeled in the dorsal portion of this subnucleus. Caudally, there was a more even distribution of labeled cells within the subnucleus with the labeled population extending to the lateral border. There was also an occasional labeled c~ll observed along the medial border of the dorsolateral subnucleus. The cameral lucida drawing in Figure 2-2b depicts the characteristic pattern of labeled cells seen after an injection of the right medial rectus muscle. A photomicrograph of a typically labeled section is shown in Figure 2-4. Injection of the inferior oblique muscle also produced a consistent ipsilateral pattern of labeled cells. In the
PAGE 74
65 0 8E)?JQ GO Figure 2-2. Summary of motoneuron organization w ithin the oculomotor nucleus. Camera lucida drawings of transverse sections through the mesencephalic basal plate of day 18 chick embryos following injections of the extraocular muscles with HRP. T h e presence of HRP reaction product within the subnuclei is represented by the dots. A. Injection of the right inferior rectus muscle typically resulted in labeled cells distributed through the right dorsolateral subnucleus B. Injection of the right medial rectus muscle typically resulted in labeled cells distributed through the right dorsomedial subnucleus. c. Injection of the right inferior oblique muscle typically resulted in labeled cells distributed through the lateral portion of the ventromedial subnucleus. D Injection of the right superior rectus muscle typically r esulted i n labeled cells distributed through the medial portion of the left, contralateral ventromedial subnucleus. dl, dorsolateral subnucleus; dm, dorsomedial subnucleus; vm, ventromedial subnucleus.
PAGE 75
66 ',I~ r "~t, f \. ( . ..., .. . a } :! ""' -. v. ~ :- ;-. ; V \J I 1 ~ A f ,!'"., I,. ; .' ::, .,i:~ I ,. ', l,._ ,. .. ,. ., ...... r; .. -.. .... I "'.z t. ,. ~ r _.ji. r ,: i.': ; :.---i/i:.;;, ;:,;,. .:\ : .... ; ~,~ :!', ;: \ ifV';.# ~ :. . : :,, \ ~"-!. ,.,;,."'-,+.._-f. ._ -;,."> -... "~ T (; 4 d> r > C ,I ... .,.JIil -f, ... )1; ... -~ ( .. :....)! < (J;" .. J ,.' ',1. : .._ ._ ., ~ -<"- ,"'II. C ::If ,.. :.>k .. . .. ;., .. f ~ .. '.i ~ : i :
PAGE 76
I' .:. ,. ., ,. ... "' .. ., (' ~ ; / .8 -.. .. ., a ( ,(_ t .. ,-' ,.... ~ ,, ry ... l 67 f I f, ., ., ,, i I i {'Figure 2-4. Innervation pattern: Medial rectus muscle. Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right medial rectus muscle. The HRP reaction product is present in cells of the dorsomedial subnucleus (boxed area) ipsilateral to the injection site. The inset of the boxed area shows many labeled cells (arrows) identifiable at higher magnification. Abbreviations as in Figure 2-1. Neutral red stain. X 90 (inset X 144).
PAGE 77
68 rostral regions the cells were located predominantly in the dorsal portion of the ventromedial subnucleus and more laterally than medially. Progressing to mid region the labeled cells spanned the full dorso-ventral extent of the ventromedial subnucleus but with more cells appearing dorsally. As seen rostrally, the labeled cells were still restricted to the lateral portion of the subnucleus. Caudally the number of labeled cells diminished but the same pattern continued. The labeled cells were in the lateral portion of the ventromedial subnucleus with more labeled dorsally than ventrally. Three dimensionally, this pattern constructs something of an elongated pyramid, with one short side corresponding to the dorsal border of the ventromedial subnucleus. The base of this pyramid faces rostrally with the apex projecting caudally. The camera lucida drawing in Figure 2-2c represents the characteristic labeling pattern in mid region of the complex with a photomicrograph of typically labeled cells in Figure 2-5. Injection of the superior rectus muscle again produced a consistent pattern of labeling but, in this case, it was entirely contralateral. The pattern of labeled cells was complementary to that produced by injection of the inferior oblique muscle so that essentially the medial portion of the ventromedial subnucleus was labeled. Rostrally the labeled cells were found primarily along the medial or dorsomedial border of the contralateral ventromedial subnucleus with
PAGE 78
69 Figure 2-5. Innervation pattern: Inferior oblique muscle. Transverse section through the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right inferior oblique muscle. Granular HRP reaction product is localized to cells of the lateral portion of the ventromedial subnucleus ipsilateral to the injection site. The boxed ventromedial subnucleus is shown at higher magnification in the inset. The arrow s point to labeled cells which are located exclusively within the lateral portion of the subnucleus. Abbreviations as in Figure 2-1. Neutral red stain. X 100 (inset X 167).
PAGE 79
C. .. ( V .. ... r ,... "' .,. < : . ........ -' ~ -~ 'lff-t b ... r ......_ If'! .. .. ) ;.' : 70 \ Figure 2-6. Innervation pattern: Superior rectus muscle. Transverse section of the oculomotor region of the mesencephalon of an 18 day chick embryo following HRP injection of the right superior rectus muscle. Labeled cells are present in the ventromedial subnucleus (boxed) located contralateral to the injection site. Higher magnification of the boxed region (inset) reveals that the labeled cells (arrows) are clearly confined to the medial aspect of the contralateral ventromedial subnucleus. Abbreviations as in Figure 2-1. Neutral red stain. X 92 (inset X 138).
PAGE 80
71 only few cells labeled. In the mid region the labeled cells were again confined to the medial aspect but now fully spanned the dorsoventral extent of the subnucleus. Caudally this distribution persisted although ventrally the labeled cells were spread somewhat laterally. In addition, at this level, the labeled population included what appeared to be the "central nucleus" which lies in the medial area between the two ventromedial subnuclei. The camera lucida drawing in Figure 2-2d represents the pattern of labeled cells following injection of the right superior rectus muscle. The typical distribution of labeled cells is seen in the photomicrograph in Figure 2-6. Discussion The pattern of labeled cells following retrograde transport of injected HRP was very consistent for all four of the extraocular muscles. The locations of each of the four motoneuron populations present within the complex were unique with very little overlap. On the few occasions where a labeled cell was found to lie outside the primary domain it was on the border of the neighboring subnucleus immediately adjacent to the parent population. The innervation pattern reported here and summarized in Table I is as follows: the dorsolateral (DL) subnucleus innervates the ipsilateral inferior rectus muscle, the dorsomedial (DM) subnucleus innervates the ipsilateral
PAGE 81
TABLE 2-1 Major projection patterns of the oculomotor complex in the chick Extraocular Muscles Oculomotor Subnucleus Inferior Medial Inferior rectus rectus oblique Dorsolateral Dorsomedial Ventromedial, lateral Ventromedial, medial Central + I I +, I +, presence of HRP reaction product absence of HRP reaction product I, ipsilateral projection C, contralateral projection + I I 72 Superior rectus +, C + C
PAGE 82
73 medial rectus muscle, the lateral portion of the ventromedial (lVM) subnucleus innervates the ipsilateral inferior oblique muscle, and the medial portion of the ventromedial (mVM) subnucleus innervates the contralateral superior rectus muscle. This distribution pattern of the motoneuron populations within the oculomotor subnuclei is similar to that which has been previously described using degeneration techniques to identify the particular populations (Isomura, 1973). The reports differ, however, in the description of the segregation of the motoneurons within the ventromedial subnucleus. Isomura (1973) reported finding one population confined to the dorsal aspect (innervating the ipsilateral inferior oblique muscle) and the other to the ventral aspect of the subnucleus innervating the contralateral superior rectus muscle). Our findings do not conflict with this observation so much as refine it, since we also found more cells dorsally which innervated the ipsilateral inferior oblique muscle. However, while the preponderance of cells were dorsal there was also a wedge of labeled cells which extended ventrally. Additionally, it was clear that both dorsally and ventrally the cells were restricted to the lateral aspect of the ventromedial subnucleus. The reported differences are probably attributable to the numbers of cells labeled with the two techniques. The greater numbers of cells which were labeled with the retrograde transport technique provided a fuller
PAGE 83
74 picture of the entire .motoneuron population and extends the observations to provide greater detail. The division of the motoneuron populations within the ventromedial subnucleus is of particular interest because of the unusual migration which occurs during development to produce the contralateral innervation. This migration has been described to occur in several avian species (PuellesLopez et al., 1975; Sohal, 1977; Heaton, 1981). In the chick the medial migration of cells from the oculomotor anlagen is first observable at 5 days of incubation. Until that time this medial cell column had been a single homogeneous cell group. Over the next two days the number of participating cells increases and the cells meet with the contralateral population and seem to freely intermingle. Subsequently the cells disperse and apparently continue on their migratory route to the contralateral side. Purely descriptive studies have been unable to determine the extent of the ventromedial population which migrates or the exact locus of their destination. Puelles-Lopez et al. (1975) speculated that the entire ventromedial population may participate in the migration. The present study provides rather convincing evidence that it is only those cells that comprise the medial portion of the ventromedial subnucleus which participate in its migration. The pattern of labeling observed here was very consistent and reliable between cases. The virtually
PAGE 84
75 exclusive segregation of the motoneuron populations differs from the varying degrees of overlap in these populations reported in other species (cat; Gacek, 1974; frog; Matesz and Szekely, 1977; stingray; Rosiles and Leonard, 1980; baboon; Augustine et al.,1981; macaque monkey; Spencer and Porter, 1981). The technique used here permitted labeling of only a portion of the motoneuron population innervating a muscle. The volume of HRP injected and the resultant number of cells labeled was restricted by the leakage of HRP when large volumes are injected combined with the ability of embryonic muscle to incorporate the enzyme after only topical exposure. Nonetheless, the highly consistent pattern and the fact that there were no areas left unlabeled by our injections argues strongly against the existence of any significant overlap among the motoneuron populations in the chick oculomotor complex. The few instances in which a labeled cell was observed outside the primary domain for that cell group the errant cell was contiguous with the parent population but located in the neighboring subnucleus. There are many similarities in the pattern of oculomotor innervation reported for the different species. The contralateral innervation of the superior rectus muscle has been reported in stingray (Rosiles and Leonard, 1980), frog (Matesz and Szekely, 1977), cat (Gacek, 1974), baboon (Augustine et al., 1981), and macaque monkey (Spencer and Porter, 1981). The topographical organization within the
PAGE 85
76 oculomotor complex is particularly similar in chick, cat, and baboon. In each of these species, the motoneurons innervating the superior rectus and the inferior oblique muscles accumulate in close proximity to one another. It may be functionally important for these two motoneuron populations to be contiguous since both muscles serve to rotate the eye upwards. It is particularly interesting that they innervate muscle on opposing eyes because the inferior rectus rotates the eye up and slightly nasally while the superior rectus muscle rotates the eye upward with a temporal bias so that if activated together they would produce an upward rotation of both eyes with both eyes slightly directed to the same side. While these species demonstrate similarities in organization, in the baboon there is a greater degree of intermingling of adjacent populations. On the basis of differences in size and location of oculomotor motoneurons of the macaque monkey, Spencer and Porter (1981) suggest that there might be subpopulations of motoneurons to a particular muscle which mediates the action of that muscle in different types of eye movements. It is possible that such multiple representation of a particular muscle indicate that in this species at least, the subnuclear organization is more closely related to specific types of movement rather than to specific muscles.
PAGE 86
77 Oculomotor organization in fish and amphibian species has been found to be somewhat different from that seen in mammalian and avian forms (e.g. bilateral innervation: stingray, Rosiles and Leonard, 1981; frog, Matesz and Szekely, 1977). This disparity may reflect the diversity of these species and their differing adaptive requirements. Their individual environments, the predators and prey likely to be encountered, and the position of the eyes in the head might all influence the types of eye movements most frequently employed.
PAGE 87
CHAPTER III OCULOMOTOR DEVELOPMENT IN THE CHICK FOLLOWING EXPERIMENTAL REMOVAL OF THE TARGET MUSCLES. Introduction In the ventral spinal cord, motoneuron cell bodies are arranged in groups corresponding to the musculature of specific body regions which they innervate, such as arm, leg, and axial muscle. This topographic grouping of motoneurons is prevalent in the central nervous system (Landmesser, 1978a; Landmesser, 1978b). In the chick oculomotor complex, retrograde labeling has demonstrated that the motoneurons innervating four of the extraocular muscles (i.e. superior rectus, medial rectus, inferior rectus and inferior oblique) exist in segregated populations within the subnuclei (Heaton and Wayne, 1983). A similar organization of motoneuron populations is seen in the oculomotor complex of mammal ian species (Augustine et al., 1981). This adult segregation of the motoneuron population develops in the chick embryo from a single, homogeneous cell group, the oculomotor anlagen (Figure 3-1). This precursor of the adult oculomotor complex is first seen in the rostral midbrain of the chick embryo at day 3 of incubation. At this time, these cells begin to send fibers out to the peripheral 78
PAGE 88
FIGURE 3-1: Normal development of oculomotor complex The normal development of the chick oculomotor complex is represented in these three transverse sections of the rostral mesencephalon at (a) 4, (b) 7, (c) 10 days and (d) 18 days of incubation. (a) At 4 days the oculomotor anlagen (A), a homogeneous cell group, sits in the marginal layer of the mesencephalic neural tube (NT). The nerves (N) can be seen exiting ventrally into surrounding mesenchyme (S). X 258. (b) At 7 days the emerging subnuclear organization is becoming apparent. Accessory (a), dorsal (d) and ventral (v) cell groups are visible. There is a suggestion of a medial -lateral division of the dorsal cell group which will become two distinct subnuclei later. There are many cells migrating (m) between the ventral groups. The medial longitudinal fasciculus (MLF), a brainstem fiber bundle, sits next to the oculomotor cell bodies. X 337. (c) Cells migrating between the two ventromedial subnuclei in (b) are shown here at higher magnification. The typical elongate morphology of migratory cells is apparent. The migration across the midline to the contralateral subnucleus is indicated by the presence of cells oriented in opposing directions (arrows). X 1350. (d) At 10 days the subnuclei have separated completely. The accessory cells (a) look quite different from the other subnuclei. X 241. (e) The adult configuration of the oculomotor complex is evident at 18 days which is just prior to hatching. The dashed line within the ventromedial subnucleus approximates the segregation of motoneurons which innervate the ipsilateral inferior oblique muscle (lateral portion) and the contralateral superior rectus muscle (medial portion). X 213. a=accessory, dl=dorsolateral, dm=dorsomedial, vm=ventromedial, v=ventricle, s=mesenchyme
PAGE 89
s a 1 .. .,;, I .. ,., : I .., -,. .... ~,,_ .. "5~ .. ,.: ....... .. 80 -~ ...... ..; ...... ... ., _. 4 ... ..
PAGE 90
81 musculature. By embryonic day 5 all of the extrinsic eye muscles innervated by the oculomotor complex have been reached by some fibers from that complex. In the subsequent development of the oculomotor complex, embryonic days 5-10, there is a gradual reorganization of the cell bodies so that the subnuclear distinction emerges. During this time there is also a migration of some cells to the midline where they intermingle with the corresponding cells from the opposite side. It seems that these cells then cross the midline and comprise the medial portion of the ventromedial subnucleus. The crossing of these cells thus accounts for the contralateral innervation of the superior rectus muscle (Puelles-Lopez et al., 1975; Heaton, 1981; Heaton and Wayne, 1983) The appearance of subnuclei and the migration of a portion of the cell population both serve to produce the muscle-specific grouping of motoneurons. The temporal contiguity between the nerve fibers first reaching the target musculature and the onset of these organizational changes in the cell bodies of the oculomotor complex suggest the possibility of a causal relationship. Thus, it is possible that some influence from the periphery either initiates or directs the cells in their rearrangements. In order to account for specific matching between motoneuron and muscle cell populations there seem to be three developmental possibilities. The muscle specification
PAGE 91
82 may precede neural development and the subsequent organization of neuronal cells would be dictated by the contacts made with the peripheral target. Alternatively, neural specification may precede muscle development and the distinct neural populations may then determine muscle organization upon contact with the periphery. The third possibility is that the specificity of neural and muscle populations arise independently and that development has incorporated some means to appropriately couple specific cell populations in the periphery. The pathway of nerve outgrowth, the subnuclear segregation, and the migration of a distinct population of motoneurons may all be subject to some influence from the four extraocular muscles innervated. This possibility was experimentally addressed by removal of the presumptive muscle t issue prior to the appearance of the oculomotor anlagen. The subsequent development of the oculomotor nerve and nucleus was observed over the course of embryonic days 5 through 1 0 and is described here. Methods Fertile chicken eggs, Gallus gallus domesticus, were obtained from the University of Florida Poultry Science Department. The eggs were set in a forced draft incubator at 370c and 70% relative humidity for 36-40 hours to yield stage 10-11 embryos (Hamburger and Hamilton, 1951). The surgery precedes both the condensation of tissue to form the
PAGE 92
83 extraocular muscles and nerve outgrowth from the oculomotor complex. Immediately prior to surgery the eggs were candled to reveal the position of the embryo. A window was made in the shell and after exposure, 1% neutral red stain was dropped onto the embryo to enhance contrast. The overlying vitelline membrane was carefully peeled away with fine forceps. The left optic vesicle and surrounding mesoderm, the precursor for the extraocular muscles, were removed with a microsurgical vibrating needle (Wenger, 1968). In some control operations the mesoderm was left intact and just the left optic cup was removed. The surgery is shown diagrammatically in Figure 3-2. In all cases the surgical removal of tissue was from the left side of the embryo. The egg was then resealed with parafilm and was carefully set in an incubator. The window was directed up and the egg left undisturbed for the remainder of its incubation period. The embryos were sacrificed at days 5 through 10 of incubation. They were fixed in Bouin's fixative, dehydrated, embedded in paraffin, and sectioned at 10 m. Following mounting on slides, the sections were counterstained with hematoxylin and eosin. Results Muscle Removal The operations always resulted in the presence of some residual muscle tissue. Each case was evaluated individually. The majority of cases demonstrated a dramatic
PAGE 93
8 4 r-----, I Figure 3-2: Diagram of surgical removal of presumpt ive extraocular muscle The chick embryo at stage 10-11 represented in the drawing above. The the tissue removed. Note that dorsally of development is dotted line frame s the cut followed the contours of the neural tube while the ventral cut extended to the midline (arrow). ov= optic vesicle, s= mesenchyme, NT= neural tube
PAGE 94
85 reduction of muscle with just a single small mass of muscle cells remaining. In one case, a severe and almost compete absence of muscle cells was seen (Figure 3-3). The volume of residual muscle present probably provides an inflated index of normalcy because it assumes that any residual muscle in the area is an appropriate target for oculomotor neurons. In some cases the only muscle available seemed to be displaced caudally suggesting that it may actually have been the remnant of another muscle group. While it is not possible to distinguish muscle cells of the oculomotor system from other muscle cells of the region, these cells may bear differences important in normal development (see Chapter V). In some cases a fair mass of muscle was present but it should be noted that this was still quite reduced compared to the normal muscle. Of even greater significance, the muscle did not separate into the distinct muscle groups but formed one consolidated mass. This consolidation was not simply a function of losing the eye as an anchoring point because control operations with eye removals only have no eye or the surrounding orbit yet still evidence the distinctive muscle groups. It seems that there is some qualitative change in the muscle so that the normal complement of distinct muscle groups is not available for innervation.
PAGE 95
\ a ) i 1 ,. ... s l ,. ..: : ., ... .. ~ ., im ... : : = : -\>~i~~-.:, : :,.' f : ~;.,. r i ,." } ... ~ ~ f .. .'j .. "' ., ,,. \ b / ., /. 'i I I I .. r-. Figure 3-3: Muscle remnants following removal 86 M .. s ,...~~' ., ~ i ~ ~ .. Transverse sections through the rostral mesencephalon of (a) day 6 and (b) day 7 experimental chick embryos. These cases demonstrate the range of muscle which was present in the experimental cases used for analysis, (a) representing a maximum and (b) representing a minimum. (a) The oculomotor nerve on the left, operated side can be seen coursing towards the residual muscle (rm). The experimental muscle mass (rm) always has an appearance distinct from normal muscles (m). (b) There are only a few muscle cells (rm) present at the body wall. (a) X 148; (b) X 112. M= mesencephalon, s=mesenchyme, N= nerve
PAGE 96
Figure 3-4: Segregation of subnuclei Transverse sections through rostral mesencephalon of (a) 7 and (b) 9 day experimental embryos demonstrating the appearance of subnuclei. As in all cases, the left side is the operated side. (a) X 305; (b) X 433. (a) Accessory (a), dorsal (d) and ventral (v) cell groups are visible on both sides. (b) Accessory (a), dorsomedial (dm), dorsolateral (dl) and ventromedial (vm) cell groups are apparent on both sides. The shapes of subnuclei on experimental and control sides are disparate because they have been cut in somewhat different planes of section. The similarity in organization is clear nevertheless. n= nerve, s=mesenchyme
PAGE 97
88 a ,. . s
PAGE 98
89 Subnuclear organization and cell migration The experimental manipulation introduced a loss of symmetry in the embryo which frequently resulted in a distortion of the neural tube. This neural tube distortion made a consistent orientation of the tissue for transverse sectioning difficult to achieve and sometimes meant that the experimental and control sides of the brain were sectioned in different planes. As a result the observed variability in the shape of the subnuclei in experimental embryos was determined largely by the plane of section of the tissue. In contrast, the degree of separation of the individual subnuclei was only slightly influenced by the plane of section. In all cases examined the subnuclei were iegregated to some extent (Figure 3-4). In any individual case, the extent to which the individual subnuclei were distinguishable on the operated side reflected the organization evident on the unoperated control side of the same embryo. Therefore, the breakdown into the distinct subnuclear populations occurred regardless of the presence or absence of the normal target muscles. Migrating cells were identified by two criteria: (1) their position within the complex, located between the two ventral cell groups; and (2) their morphology, which consisted of an elongate cell body with a fine trailing process. In all experimental embryos migratory cells which were oriented in both directions could always be found.
PAGE 99
90 Thus, the migrating population always contained cells from the experimental side as well as cells from the unoperated, control side (Figure 3-5). The magnitude of migration varied considerably between cases. Any precise quantification of the migrating population was made impossible by the fact that there are no distinct borders separating migrating from non-migrating cells and the migrating population itself consists of intermingled cells from both sides. In general, the variability in degree of migration did not show any relationship to the amount of muscle remaining. It is possible that the variation observed was a normal consequence of capturing cases at different developmental stages. The surgical intervention probably alters the developmental time course to some extent, thus creating a greater range in the developmental stages found amongst any age group. Axonal pathway In 21 of 23 cases analyzed, the oculomotor nerve projected along a very consistent pathway to any residual muscle still present on the operated side (Figure 3-6). Among this majority which projected to the ipsilateral (normally appropriate) side, there was usually a fair amount of muscle still present in the target region. The quantity of residual muscle available did not seem to determine the path of innervation, however. A few cases showed only
PAGE 100
91 .-1) ., \ .. .,,, .., .. Figure 3-5: Migration Transverse sections through the rostral mesencephalon of two day 7 experimental embryos demonstrating the migration of cells. The variation in numbers of migrating cells can be seen. Arrows indicate the profile of migratory cells which appear to originate from the left, operated side. (a) X 398; (b) X 398.
PAGE 101
92 minimal muscle still present (Figure 3-3). Even with only a few muscle cells remaining, the nerve followed a normal path to the ipsilateral side. It also seemed that in some cases in which innervation was ipsilateral, the contralateral muscle was not too remote for possible innervation. In some cases, the neural tube had rotated so that the ventral aspect of the tube now faced the remaining eye. Despite the proximity of the contralateral muscle, in these cases, the nerve traveled a course of greater distance to reach any residual muscle on the ipsilateral side. In the majority of cases, the nerve persisted in following the normally appropriate path to the ipsilateral side despite the apparent availability of the full complement of normal muscles on the control side of the embryo. The contralateral muscles were not somehow inherently unsuitable for innervation by these neurons because in two cases the nerve did cross to innervate the contralateral side (Figure 3-7). In both of these cases, there was still a fairly sizeable remnant of muscle present on the operated side. Once again, then, the muscle remnant did not determine the innervation path of the operated nerve. In all these cases, it appears that pathway cues acted independent of target muscle to guide the nerve to the target region.
PAGE 102
Figure 3-6: Most common axon path Transverse sections of rostral mesencephalon (M) of (a) day 7 and (b) day 6 embryonic chick brains. These cases demonstrate the most frequently seen path of nerve (n) growth, i.e., an ipsilateral course to the remnant of muscle (rm) in the periphery. (The few muscle cells which remained in case a are shown in Fig. 3-3b) Normal muscles (m) were available in all cases on the contralateral side and are present in (b) but not visible at the level shown in (a). (a) X 160; (b) X 134. c= ciliary ganglion
PAGE 103
.. \ \ ? / 1 / / r \ ,.:. ~': -:. \ r !, ;.. '!.", > b l ~ .) -I i \ I .:_ '\ ~.'1 t '.' \ k\, 'A-~ \ -\ '-i : .: ,,.. ~., 1 :,;, s 't 94 s ... ,,,,,.. ... ,, ., ,;, ~ I ... I I 6 (; :... ..,.. 'ii.
PAGE 104
Figure 3-7: Crossed axon path The crossed axonal pathway observed in two cases is depicted here. (a and b) Day 6 embryo -The course of the two nerves (n) to the control muscles (nm) is clearly seen. In (a) the muscle remnant (rm) which persisted in the periphery on the experimental side is shown from a section at another level. (c and d) Day 9 embryo -The two nerves (n) exit the mesencephalon and appear headed toward the control side. Following the nerves through successive sections the experimental nerve passes under a midline blood vessel to travel to the contralateral side. (d) The nerves are still distinct as they approach the control ciliary ganglion (c). Remnants of muscle (rm) can be seem in the periphery on the experimental side. The bar marks the midline of the neural tube ( a) X 119 ; ( b) X 113 ; ( c) X 113 ; ( d) X 2 2 1. M= mesencephalon, r= retina
PAGE 105
a f I I L J ( ; \J : ...:, f I 96 ..... ) .. /_,.. ~ -... : .. : 1 \ I, ..... 'i
PAGE 106
97 Discussion The present experiment has shown that neither the subnuclear organization nor the initial course of axon outgrowth of the oculomotor neurons is dependent on the presence of the target extraocular muscles. The peripheral course of nerve outgrowth, however, did seem to be influenced by the presence of muscle in that once in the periphery the nerve path was directed to any muscle cells present. The presence of the motoneuron population does not appear to be sufficient to organize residual muscle normally. The experimental removal of the lateral mesoderm resulted in the quantitative and qualitative alteration of the extraocular target muscles. The quantitative reduction of muscle was evident on gross inspection. Along with the reduction in the mass of muscle, the residual muscle also appeared morphologically altered. The muscle cells which were identified by position and histological staining properties were present as single condensed balls of muscle cells rather than the distinct elongate morphology that the individual muscles assume in the normal embryo. Normal development of the different muscles of the wing has recently been described to occur from independent groups of myoblasts rather than from a large muscle precursor which later cleaves to form the individual muscles (Noakes et al., 1986). If the same is true of the development of the
PAGE 107
98 extraocular muscles, then the accumulation of the remaining muscle cells into a single mass may indicate a breakdown in the normal muscle identity. Control operations in which only the eye and not the surrounding mesoderm was removed demonstrated that even in the absence of the eye the individual muscles could be distinguished (elongate morphology). This observation supports the concept that the muscle remnant is qualitatively altered as well as being reduced in mass. While the quantitative depletion of muscle mass was severe in only a few cases, the morphology and position of residual muscle cells suggest that the muscle was substantially altered in the majority of the cases. In the normal embryo the motoneuron populations are grouped within the nucleus based on the particular muscle innervated. Since the normal arrangement of the nucleus into subnuclear groups occurs immediately after contact is made with the muscle we decided to investigate whether the contact with the appropriate muscle was essential for this organization to occur. We found that regardless of the degree of depletion of the target muscle the nuclear organization on the experimental side of the embryo matched that seen on the control side. It is important to address whether this gross morphological organization into subnuclei reflects a selective grouping of motoneuron cell bodies or merely a
PAGE 108
99 random distribution of the cells. There is no means to identify the population of motoneurons independent of the muscles which they innervate. However, the fact that subnuclear organization is not a passive process but is, in part, accomplished by an active migration of a restricted subpopulation of cells suggests that subnuclear organization itself is a specific and active process (Puelles-Lopez et al., 1975). The observation that the distinction of subnuclei occurs independent of the presence of the target muscle suggests that the specificity of the neuronal populations is not derived from contact with the muscles. The proximal pathway of nerve outgrowth, i.e., that immediately adjacent to the exit from the neural tube, appeared in the present study to be wholly independent of any influence of the muscle. This observation is consistent with the results of experimental manipulations of the motoneuron pathway in the chick limb (Summerbell and Stirling, 1981; Whitelaw and Hollyday, 1983c). In these studies, the proximal nerve path seems to be determined by the organization of the limb plexus and not by the position of the limb muscles. In a few of our cases where the depletion of extraocular muscle on the experimental side was nearly complete, the nerve persisted in following the normal ipsilateral course of growth. Of greater import, in two cases the nerve crossed the midline and followed a path to the contralateral target muscle despite the presence of
PAGE 109
100 residual muscle on the ipsilateral side. It is probably significant that these two cases were part of surgeries in which an effort was made to reduce the residual target muscles by removing more mesenchyme including that ventral to, but not crossing the midline of, the neural tube. This removal of ventral mesenchyme may have served to remove or disrupt cues significant to the nerve path followed immediately after exiting the neural tube and thus account for the move across the midline. The initial course of outgrowth reflected no responsiveness to the presence or the absence of muscle in the ipsilateral periphery. The factor or factors which subserve the initial pathway selection by these brainstem motoneurons is of great interest. Some of the mechanical pathway constraints provided by the plexus region in the limb do not seem to be present in the brainstem. Specific pathways are followed in both regions, however, and the determination of these routes remains unresolved. Some recent studies have suggested that Schwann cells which migrate out from the neural crest may be an important component of the pathway for motor axon outgrowth (Carpenter and Hollyday, 1986; Noakes and Bennett, 1987). From the preceding, it appears that target muscle does not dictate the organization of the motoneuron cell bodies and it does not determine the initial course of nerve outgrowth. How then does muscle influence nerve growth in
PAGE 110
101 vivo? In vitro studies have repeatedly demonstrated that both muscle-conditioned media (Dribin and Barrett, 1980, 1982; Henderson et al., 1981) and extracts of muscle cells (Hsu et al., 1982; Smith and Appel, 1983, Smith et al., 1986; Dohrmann et al., 1986) will enhance neurite outgrowth from motoneuron populations. In the present study we observed that once the nerve reached the periphery, it always grew to any available muscle cells which remained. The nerve frequently reached muscle which was considerably displaced from the normal target position, suggesting that the presence of muscle is somehow recognized by the growing nerve. In the normal development of the chick limb, axonal outgrowth is highly patterned and consistent (Lance-Jones and Landmesser, 1981a, 1981b; Tosney and Landmesser, 1985a, 1985b). Specific motor axons seem to make unique decisions in the course of growth and branching within the muscle nerve (Tosney and Landmesser, 1985a). Muscle derived factors may act locally to affect nerve growth with diminishing capacity to influence growth with distance. In vivo the muscle-produced factors probably influence nerve growth in concert with other environmental cues so the efficacy of action must in some measure be balanced against other, perhaps competitive, influences on the nerve. That the local influence of muscle on nerve growth is specific is suggested by both in vitro and in vivo studies. Experiments in culture have indicated that the activity of
PAGE 111
102 muscle-produced factors capable of influencing neurite growth is dependent on the age and regional origin of the muscle tissue (Pollack and Muhlach, 1981; Chapter V). In the chick limb bud, nerves experimentally induced to grow along an incorrect pathway occasionally break off and forge a wholly new path to their specific muscle (Lance-Jones and Landmesser, 1981b). Only neurons normally appropriate to innervate a particular muscle are observed to reach this muscle by novel pathways, suggesting that neurons can respond to cues from muscle over a short distance and that there is a great deal of selectivity in these effects. The specific differences in populations of muscle or motoneurons which could provide a substrate for individual recognition to occur still await identification. The early developmental organization of the muscle and the motoneuron populations often appear to occur independent of the other, supporting the third possibility mentioned above. A recent report, for example, indicates that individual chick wing muscles develop independently and prior to any contact from the muscle nerve (Noakes et al., 1986). It has also been demonstrated that in aneural chick wing the initial differentiation of muscle fibers is unaffected by the absence of neural input (Phillips and Bennett, 1984). Conversely, our results indicate that the presence of the normal motoneuron population alone is not sufficient to organize the residual muscle appropriately. If
PAGE 112
103 the motoneurons were capable of organizing the muscle, the manipulation in the present study should have resulted in merely a quantitative reduction of muscle and no other distortion of muscle morphology. Conversely, early neural development appears to progress independent of the presence of the muscle. Thus, in the current instance, subnuclear organization emerged normally in the relative absence of target musculature (i.e., in all experimental cases, individual extraocular muscles could not be identified). The initial path of axon outgrowth was also established independent of muscle. The path which the oculomotor nerve followed in this experiment, either ipsilateral or contralateral, was not determined by the muscle remnant. It would seem, then, that both the neurons and the muscle cells are capable of organizing independent of the other. If this is the case, then specificity of the contacts made between the neuronal and muscle populations must be achieved during development by some set of signals which leads to the correct coupling of specific subpopulations in the periphery. It seems likely that muscle and motor axons interact quite specifically in the periphery. There is a matching of motoneuron and muscle based on the segmental origin of each (Laskowski and Sanes, 1987) and their physiological characteristics (i.e., firing pattern and fiber type; Vogel and Landmesser, 1987). Since it is apparent that neuron and
PAGE 113
104 muscle are uniquely identified independent of contact with the other, it is of some interest how this specific peripheral matching arises. Cell surface related molecules and released factors are equally likely to play important roles in the recognition which occurs. In summary, the present study has shown that the peripheral target musculature is not necessary for the determination of the subnuclear organization and the proximal nerve path. Conversely, the presence of the normal oculomotor neuronal population is not sufficient to organize residual muscle into the morphologically distinct elongate muscle groups normally present. As previous studies have indicated, target muscle does seem to act locally to influence nerve growth once the nerve fibers have reached their peripheral target region.
PAGE 114
CHAPTER IV THE EFFECTS OF SERUM AND DEFINED MEDIUM ON NEURITE GROWTH FROM EARLY NEURAL TUBE EXPLANTS Introduction Embryonic chick neural tissue is frequently used in in vitro studies of neural development. Most commonly the tissue is excised at a developmental stage when the desired cell population is fully established and a maximum number of cells is available. Under these conditions, however, most of the neurons of interest have initiated processes and at least some of these cells have already contacted their target. In order to use the in vitro paradigm to investigate an early event of neuronal differentiation such as the initial growth of processes, the tissue must be isolated and cultured at earlier developmental times. Thus, it is of interest to establish the response characteristics of relatively undifferentiated neurons of the central nervous system to the conditions of culture. Neuronal growth and survival in culture is dependent on some type of supplementation of the basic nutrient medium. Both serum and chick embryo extract have proven to be effective in supporting cells in culture. Serum, in particular, has been widely used in the culture of neurons. This practice may create problems since serum itself has 105
PAGE 115
106 been reported to affect neuronal cell differentiation (Wolinsky and Patterson, 1985) and neurite outgrowth (Skaper and Varon, 1985). Recently, in order to circumvent the potential problems posed by serum, chemically defined constituents have been successfully used in lieu of serum to supplement growth media for a variety of neuronal cell types (rat cerebral cortex [Borg et al., 1985); rat cerebellar cortex [Gallo et al., 1986); fetal rat brain [Honegger et al., 1979); chick dorsal root ganglia [Bottenstein et al., 1980)). In the present study, we compared initial neurite outgrowth from early neural tube explants cultured in medium supplemented with fetal bovine serum, and in medium supplemented with chemically defined components. The influence of serum on the outgrowth induced by the defined medium was subsequently examined by comparing growth in defined medium to growth in medium with both serum and the defined supplements combined. For this determination the early trigeminal motor system of the chick embryo was used as a model. In vivo, the chick trigeminal motoneuron population appears immediately adjacent to the midline between 48 and 84 hours of incubation, with the peak time of cell production occurring from 50-56 hours (Heaton and Moody, 1980). These neurons are the first to undergo differentiation in the rostral metencephalon. By thirty-six hours of incubation the metencephalic region of the neural
PAGE 116
107 tube is well circumscribed by newly formed constrictions. At 40 hours, the time of explantation, the trigeminal region is well defined but the motoneuron population has yet to be established. Methods Fertile White Leghorn chick embryos were incubated in a forced draft incubator at 37c and 60-70% humidity. At 36-40 hours of incubation (st. 10-12; Hamburger and Hamilton, 1951) the rostral metencephalic region of the neural tube was excised. This region is well defined by the constrictions which have developed in the neural tube by these stages. The excised tissue was plated onto a substratum coated with both poly-DL-ornithine (PORN, 0.04 mg/ml) and collagen (0.5 mg/ml). The medium consisted of Ham's F-12, 1.6% glutamine, 1% penicillin/ streptomycin, and 0.7% fungizone which was then supplemented with either 11% fetal bovine serum (FBS) or a defined set of supplements. Chick neural tube explanted at this stage of development requires some type of supplementation of the basic nutrient medium (Ham's F-12) in order to survive in culture. The defined supplements consisted of 5 g/ml insulin, 10 g/ml transferrin, 30 nM sodium selenite, 100 M putrescine, and 20 nM progesterone with the following additions: the potassium concentration was increased to 20 mM and bovine serum albumin (0.5 mg/ml) was added. The transferrin concentration (10 g/ml) is the only component which differs
PAGE 117
108 from the previously described defined media Nl ( 5 g/ml; Skaper et al., 1979) and N2 (100 g/ml; Bottenstein and Sato, 1979). The explants were given 24 hours to adhere in the presence of 0.5 ml of medium and then an additional 1.5 ml of medium was added. The medium was changed every third day thereafter. The day of explantation was considered to be day o in vitro. Neurite outgrowth was assessed on days 4, 8 and 12 of culture using a Nikon Diaphot inverted phase contrast microscope with an attached air curtain incubator. The cultures were photographed at 40X magnification and enlarged 7X upon printing for a total magnification of 280X. A template of concentric circles spaced at 1 cm intervals was placed over each photomicrograph. Beginning at the innermost circle (diameter= 4mm), the number of neurites which intersected with that line was counted. This was repeated at subsequent circles until the limits of outgrowth were reached. Each neurite could be counted at multiple intersection points depending on the branching frequency and the length of each neuritic process. The number of neurite intersection points were then totaled to yield the neurite count for that culture at that time. The neurites were scored blindly so that the scorer did not know under which condition the explant had been cultured. Only groups of explants which had been cultured at the same time were compared in each of the two experimental sets
PAGE 118
109 (i.e. serum compared to defined supplementation, and defined to combined supplementation). This was done to avoid the introduction of variability due to time dependent differences in viability of the embryos or the culture components. A growth profile was plotted for those cultures for which the neurite outgrowth had been assessed at all three sampling times. The maximal growth that a given culture demonstrated was considered to represent 100%. The quantity of neurite elaborated during each four day interval (0-4, 4-8, and 8-12) was then used to determine what percent of the overall growth occurred during that time span. Results Defined versus serum-supplemented medium Neurite outgrowth in defined medium exceeded that in serum-supplemented medium at all three time points examined. The difference in neurite outgrowth is evident in typical explants after culture for 8 days in serum (Fig. 4-1, 4-3a) or the defined supplementation (Fig. 4-2, 4-3b). The medians of the neurite outgrowth scores listed in table 4-1 confirm this outgrowth difference. However, the outgrowth in defined medium was significantly greater than that in serum only at the 4 and 8 day time points ( p=.002, Mann-Whitney two tail u-test). The explants grown in defined medium demonstrated abundant processes early in culture. At later times the beaded appearance of some processes signaled the presence of
PAGE 119
110 degenerating neurites. The profile of neurite growth clearly demonstrates a decline in neurite number subsequent to day eight (Fig. 4-4). This decline in neurites probably reflects a reduction in neurite extension as well as the failure to support neurites already established. In contrast, the serum-supplemented explants maintained a slowed, but steady expansion of their neurite number during the 8 to 12 day time period. The size of the explants grown in serum-supplemented medium was consistently greater than those grown in the defined medium. The central core of the explant was surrounded by a cellular mat of non-neuronal and possibly some neuronal cells which have migrated away from the explant. The neurite growth was largely confined to this cellular mat. However, neurites did grow onto the exposed substratum. This occurred very rarely with explants grown in the serum-supplemented medium and in these cases the neurites do not extend very far on this surface. Neurites extend onto the exposed substratum with much greater frequency when explants were grown in the defined medium and these neurites were seen to extend some distance on the acellular surface. Defined compared to combined medium At all three time points examined, outgrowth in the combined medium exceeded that in the defined medium. The difference in explant size and neurite outgrowth is evident
PAGE 120
111 ~ _,. Figure 4-1: Explant cultured for 8 days in serum-supplemented medium Phase contrast photomicrograph of typical explant of trigeminal region of the metencephalon cultured for 8 days in serum-supplemented medium. The explant appears healthy although few neurites have been extended. ( X 140).
PAGE 121
112 Figure 4-2: Explant cultured for 8 days in defined medium Phase contrast photomicrograph of typical trigeminal metencephalic explant cultured for 8 days in the defined medium supplements. The explant appears healthy and has extended many neuritic processes. Both the central core of the explant and the layer of cells surrounding this core are confined to a smaller area than that circumscribed by the serum-supplemented explant shown in Fig. 4-1. X 158.
PAGE 122
Figure 4-3: Cellular morphology in the different growth media. Phase contrast photomicrographs of neurites from explants cultured for 8 days in serum-supplemented (a) or defined (b) media shown here at higher magnification. The greater neurite density and complexity characteristic of culture in the defined medium is evident. Explants cultured for 12 days in the combined medium exhibit abundant and elaborate neurite outgrowth characterized by the lengthy extension of processes (c, double arrow) onto the exposed substratum with no non-neuronal cells present (c, arrows). In addition, small dark spindle-shaped cells appeared at the perimeter of the explant in many cases (d, arrows). (a), (b) X 434; (c), (d) X 173.
PAGE 123
114
PAGE 124
115 Table 4-1: Median Neurite Outgrowth in Defined Compared to Serum-Supplemented Medium (11% FBS) Defined Medium 11% FBS p value Mann-Whitney two-tail U test Day 4 129.5 (n=20) 18.5 (n=l6) .002 Median Neurite Outgrowth at: Day 8 Day 12 431 (n=25) 89 (n=20) .002 380.5 (n=22) 159 (n=l8) not significant Outgrowth in defined medium was significantly greater than outgrowth in serum-supplemented medium after 4 and 8 days in culture but not after 12 days in culture.
PAGE 125
1 1 6 Serum-Supplemented Medium Defined Medium 80 80 60 60 40 Cl) 40 Cl) Cl Cl as as 20 c 20 c Cl) Cl) 0 0 ... ... 0 Cl) 0 Cl) Q. Q. -20 -20 40 .,--40 0-4 4-8 8 1 2 0 4 4 8 8-1 2 Days in Culture Days i n Culture F igure 4-4: Profile of neurite outgrowth in serum-supplemented and defined media. Each explant was sampled at three time points. The change in the neuri t e index which occurred over each time interva l was expresse d as a percentage of the maximal outgrowth achieved by that particular explant. The values plotted above represent med ian values for the percent o f total neurite outgrowth achieved over time. The resultant growth profiles graphically display the temporal distribution o f growth in the different media. I n defined medium there is more outgrowth at ear l ier times. Subsequent t o day 8 there is a decrease in neurite number of most explants in d efined medium (10 of 17). In seru m supplemented medium neuri te number is still increasing in most explants ( 11 o f 14) during the 8-12 day interval but not a s greatly as it had increased during the preceding .interval. S ome of these explants (3 of 14) show a decrease in neurite number over this interval.
PAGE 126
117 in typical explants after culture for 8 days in the defined supplements (Fig. 4-2) or the combined medium supplements (Fig. 4-5). The observable difference in neurite outgrowth is supported by the medians of neurite outgrowth listed in table 4-2. However, the differences in outgrowth are statistically significant only at the 8 and 12 day time points (p<.002, Mann-Whitney two tail U-test). Thus, the abundant elaboration of processes which occurred early in culture in defined medium was not inhibited by the presence of serum. The additional presence of serum increased outgrowth during the first days in culture but this growth was not significantly greater than that seen in the defined medium alone. The combined medium did produce significantly greater outgrowth than the defined medium after 8 and 12 days in culture. The growth profiles for these groups indicated that while the explants grown in defined medium demonstrated a loss of neurites during the 8-12 day time period, the explants grown in the combined medium showed a continued increase in neurites (Fig. 4-6). The explants were consistently much larger than those grown either in defined or serum-supplemented media. This was due to the larger size of the explant core, to a more extensive flat (non-neuronal) cell layer surrounding the explant core, and to the abundant neurite outgrowth (Fig. 4-5). The cellular composition was noticeably different from explants in the other conditions in that small dark cells appeared at the periphery of the
PAGE 127
Figure 4-5: Explant cultured for 8 days in the combined medium 118 Phase contrast photomicrograph of typical trigeminal metencephalic explant cultured for 8 days in the combined medium. Abundant neurite outgrowth emanates from all aspects of the explant. ( X 115).
PAGE 128
Defined Medium Combined Medium p value Table 4-2: Median Neurite Outgrowth in Defined Compared to Combined Medium Median Neurite Outgrowth Day 4 Day 8 53 465 (n=ll) (n=ll) 126 1421 (n=ll) (n=ll) Mann-Whitney Two-tail not U test significant <.002 119 at: Day 12 78 (n=ll) 2105 (n=ll) <.002 At all three times sampled the amount of outgrowth in the combined medium (defined constituents plus 11% FBS) exceeded the outgrowth in the defined medium alone. However, the differences were statistically significant only at the 8 and 12 day time points, not at day 4.
PAGE 129
120 D e fin ed M edium Co mbin ed Med ium 90 9 0 7 0 70 50 50 Q) 30 O l Q) 30 Ol (1) (1) c 10 c 10 Q) Q) 0 cii -10 0 ... -10 Q) C. C. -30 -30 5 0 -50 -70 -7 0 0-4 4 8 8 1 2 D a y s in C ultu re 0-4 4 8 8 -12 D ays In Cultu r e F igure 4-6: Profile of neurite outgrowth in the defined and t h e combined medium supplements The med ian v alue s for the percent of total n eurite outgrowth a chieved over each of the three time intervals are plotted above. The profile of o utgrowth in the defined medium is basically the same a s when previously compared to the serumsupplemented medium. T h e explants cultured i n defined med ium show a greater percentage of their overal l g r o wth during the 0 4 day interval t h a n the explants -cultured in c ombine d medium. N o t e however, that the absolute quantity of neurite outgrowth sco r e d i s greater i n the combined medium ( see Tabl e 4-2). Most of the explants cultured in defined med ium (10 of 1 1 ) show a r eduction in neurite n umber over the 8-12 day interval. Similar to the p rofile of growth in serum suppleme n ted medium, n e urite number in combined medium is stil l increasi n g during the 8-12 day interval but nor as g reatly as it had increased durin g the p receding interval. None o f the c omb i ned med ium explants ( 0 o f 11) have shown a reduction in neurite n u mber ove r the 8-12 day interval.
PAGE 130
121 explant (Fig. 4-3d). These dark spindle-shaped cells were not characterized further. The neurites frequently extended very long distances onto the acellular substratum (Fig. 4-3c) Discussion In previous studies, explants of spinal cord and peripheral ganglia have demonstrated neuronal survival and healthy neurite outgrowth in the presence of defined medium constituents (Skaper et al., 1979; Bottenstein et al., 1980). These neuronal populations had been fully established in vivo, however. Thus, they have already undergone their terminal cell division and have extended processes. The age of a neuronal population clearly has bearing on the response demonstrated when isolated in culture (e.g., Pollack and Muhlach, 1981; 1982). In the present study, the survival and initial differentiation of very early explants of chick metencephalic neuroblasts was compared in different culture media. At the time of explantation, the rostral metencephalon contains only neuroblasts, i.e., dividing cells. A major projection from this region is provided by the trigeminal (V) motor nucleus. In vivo this population develops precociously: process outgrowth begins at about day 2.5, shortly following the onset of production of this population from the precursor neuroblasts (Heaton and Moody, 1980; Moody and Heaton, 1981). At early times in culture the
PAGE 131
122 trigeminal motoneuron population would certainly be the predominant neuronal population in the excised tissue. At later observation times interneurons probably contribute to the neurite output of the explants. At all time points examined, however, the V motoneuron population is likely to provide an important contribution to the neuritic output of the explant since it is the predominant neuronal output of this region in vivo. While there are multiple indices of neuronal differentiation, the characteristic assessed in this study was length and complexity of neurite outgrowth. The neurite growth index approximates the quantity of outgrowth produced. In order to actually measure neurite outgrowth rather than estimate outgrowth it is necessary to use dissociated cells. The explant preparation, however, offers some advantages. While growing in an in vitro environment, a cellular context is maintained in the cell's immediate surroundings. This cell contact is critical for neurons of the central nervous system which are unable to differentiate in dissociated culture. In order to assess early growth and differentiation in culture currently, then, it is necessary to use the explant system. The estimate of neurite outgrowth used in this study counted the number of neurite intersections with an overlying template of concentric circles. The semiquantitative approach of this method seemed a better means to estimate relative neurite outgrowth than methods which
PAGE 132
123 rely on subjective rankings of explants. The assessment used in this study does not make any compensation in scoring for neurite fasciculation. The introduction of such a weighted ranking in the scoring of neurite intersections was considered at the onset of the study but was not invoked. The additional element of scoring seemed unnecessary because fasciculation did not occur to a greater extent in one culture condition but increased with increasing neurite outgrowth in any condition. The metencephalic explants adhered to the substratum well when cultured in either defined or serum supplemented medium. However, the amount of neurite outgrowth differed greatly in the two conditions. The absolute quantity of neurite seen was much greater when explants were grown in defined medium. This is especially true at the earlier time points following explantation. The differences in neurite numbers could be accomplished via several different avenues of cellular growth. The greater neurite quantity demonstrated could be due to an increase in numbers of neurons surviving, enhanced survival of a select neuronal subpopulation or to an increase in the process outgrowth from individual neurons. These possibilities are not easily distinguished with the explant preparation. The observed differences in neurite outgrowth could be attributed to suppressed neurite growth in the presence of serum as well as to an enhancement of growth by the defined
PAGE 133
124 medium constituents. Fetal bovine serum is composed of an unknown assortment of factors which influence cellular growth processes. It is possible that along with stimulators of cell growth there are growth inhibitors as well. In the presence of a basic nutrient medium (e.g., F-12) with no additional supplements (i.e., serum or defined), neuronal cells do not survive. It is possible that serum could support neuronal survival but prevent subsequent neurite outgrowth. However, when fetal bovine serum is added to defined medium, subsequent outgrowth exceeds that in defined medium alone indicating that, under these conditions, FBS does not have a net inhibitory effect. The outgrowth produced in combined medium was far greater than the additive effects of serum and defined media. It is therefore unlikely that serum and the defined components are acting via the same mechanism. The temporal profile of outgrowth suggests that serum and defined medium components may provide complementary support for different aspects of the outgrowth process. The defined medium constituents appear to support some early aspect of neurite outgrowth but are unable to maintain neurites at later culture times. Conversely, serum appears to support neurite outgrowth for prolonged culture periods but it is unable to initiate neurites in the quantity seen in defined medium. As noted previously, the explants are a mixed cell population so that the temporal response sequence
PAGE 134
125 may reflect either an effect on different stages of growth of one population or effects on different cell populations. These possibilities can not be readily differentiated with explant cultures. As neurite outgrowth is observed to increase in the different conditions there is an increasingly greater frequency of growth of processes onto the exposed (acellular) substratum. Neurite growth has been shown to be quite substrate specific (Collins, 1978a, 1978b; Fallon, 1985; Letourneau, 1975, 1979). This extended outgrowth on the exposed substratum suggests that under these conditons either the properties of the neurites themselves are different or that the properties of the substratum have been altered. While neurite outgrowth is obviously affected, the support of non-neuronal cells also differs in the tested culture conditions. This was evidenced by the density and spread of the non-neuronal cell carpet which surrounds the central explant core. As has been previously reported (Skaper et al., 1979; Bottenstein et al., 1980), nonneuronal cells are not well supported by defined medium conditions compared to the serum supplemented medium. It is possible that the greater number of non-neuronal cells present in serum supplemented medium interferes with neurite outgrowth rather than the serum itself producing a direct effect on the neuronal populations. There is recent evidence
PAGE 135
126 to suggest that non-neuronal cells cause some inhibition of neurite outgrowth in culture (Sisken et al., 1985). However, the results in combined medium suggest that an increase in mere number of non-neuronal cells is not sufficient to inhibit neurite outgrowth. Both the extent of non-neuronal spread and the quantity of neurites elaborated were greater in the combined medium than in defined medium alone. The demonstrated effect on neurite outgrowth may still be mediated by non-neuronal cells but it must be due to the alteration of some specific aspect of the nonneuronal cells rather than just an increase in their numbers. In summary, both defined and serum-supplemented media support survival and differentiation of chick metencephalic explants. However, there are qualitative differences in the explant growth. The explants grown in defined medium demonstrate a greater quantity of neurite, a faster rate of neurite appearance and a lesser spread of non-neuronal cells which surround the explant. However, the neurites are not supported well at the longer culture times (beyond 8 days). The explants grown in the serum-supplemented medium demonstrated much less neurite outgrowth than explants in defined medium but the neurites showed a steady increase in number at all culture times examined. And finally, the explants grown in the defined and serum supplements combined indicated that there was no inhibition of outgrowth by some
PAGE 136
127 serum component. These explants demonstrated the rapid and abundant elaboration of processes evident in the defined medium alone in addition to the continued support and elaboration of neurites at later times which the serum produced.
PAGE 137
CHAPTER V SPECIFIC RESPONSIVENESS OF CHICK TRIGEMINAL MOTOR NUCLEUS EXPLANTS TO TARGET-CONDITIONED MEDIA Introduction The formation of peripheral nerves during development is thought to be guided by extra-axonal cues which promote axonal elongation, i.e. trophic factors, and/or cues which provide directional information to the growing axon, i.e. tropic factors. This theoretical distinction between stimulation of axonal elongation as opposed to directional guidance of growing axons may not always be clearly isolated in vivo. Thus, the selective distribution of trophic, growth stimulating, cues may provide for directional growth of an axonal population. The most likely candidates for the sources of these trophic or tropic influences have been either the populations of cells distributed along the path of nerve outgrowth or the target of nerve outgrowth, muscle. In vitro, both explants and dissociates of motoneurons have demonstrated enhanced neurite outgrowth in the presence of medium conditioned by dissociated muscle tissue (muscleconditioned medium, MCM) (Dribin and Barrett 1980, 1982; Henderson et al., 1981) or muscle extract (Hsu et al., 1984; Smith and Appel, 1983). These observations support the 128
PAGE 138
129 concept that muscle or some factor released by muscle cells may play a role in nerve development in vivo. While there is some in vivo evidence suggesting that specific muscle groups can attract the ingrowth of specific axonal populations (Lance-Jones and Landmesser, 1981b), the degree of specificity of such influences has not been closely examined. In addition, most previous in vitro studies have examined neurite growth from motoneuron populations which had already extended processes in vivo. The present study examined the specific influence of presumptive muscle tissue from different body regions on the initial neurite outgrowth from explants of a brainstem motoneuron population. For this determination, chick metencephalic neural tubes containing the trigeminal (V) motor nucleus were explanted at Hamburger and Hamiltion (1951) stage 11 (36-40 hours) and cultured in the presence of a standard unsupplemented culture medium or in the presence of one of two types of conditioned media (CM). The CM were derived from dissociates of (1) the mandibular process of the first visceral arch at stage 22 (3!-4 days of incubation) (this tissue comprises the normal target region for the trigeminal motoneuron population) or (2) the rostral limb bud, also at stage 22. The visceral arch and the limb bud were at equivalent developmental stages in vivo, both having just begun to be innervated.
PAGE 139
Methods Explant preparation 130 Fertile white leghorn chicken eggs were set in a forced draft incubator at 37c for 36-40 hours (st. 11). The eggs were candled, opened and the embryo lightly counterstained with a 1% solution of neutral red. The trigeminal (V) region of the neural tube, the rostral metencephalon, was excised using the vibrating needle of Wenger (1968). The rostral metencephalic region is just rostral to the otocyst. It is clearly identified at this stage by the accumulation of migrating neural crest cells and the newly formed neural tube constrictions. This stage was chosen for explantation because it is just prior to the initial neuronal generation within the V motoneuron population (stage 12). It is also prior to the neural crest condensation to form the trigeminal ganglion and the subsequent penetration of the brainstem by ganglionic afferents (st. 13, Heaton and Moody, 1980). The explants were cultured for two days before the experimental media were added. At this time, cytosine arabinoside (Ara-c, 2 g/ml) was added. Ara-c is toxic to dividing cells and thereby inhibits the proliferation of any fibroblasts present and limits the neuronal populations to those which have been established at this time (Burry, 1983). In vivo, the proliferation of the trigeminal motoneurons is completed by 72-84 hours of incubation (Heaton and Moody, 1980). Thus,
PAGE 140
131 addition of Ara-c to the cultures at the time equivalent to 86 hours in vivo effectively restricts the neuronal population to those already generated, that is, primarily the trigeminal motoneurons. The size of the neural tube segment explanted was fairly consistent. The area of a number of explants was measured using a Bioquant II digitizing tablet. The initial explant size was relatively uniform with a mean surface area of 0.57 mm2 (S.E.M.= 0.02). The neural tube explants were randomly assigned to one of the three experimental groups and plated on collagen/ polyornithine coated 35 mm dishes. The explants were positioned in the center of the dish and approximately 0.3 ml of the standard medium was added for an overnight incubation to allow adherence to the substratum to occur. The following morning this medium was removed and another 2 ml of the standard medium was added. Twenty-four hours later the medium was replaced with medium containing Ara-c. Media used were: (1) 1.5 ml of standard F-12, (2) 1.0 ml of standard F-12 plus 0.5 ml of jaw muscle consitioned medium (JMCM), or (3) 1.0 ml of standard F-12 plus 0.5 ml of limb muscle conditioned medium (LMCM). Preparation of conditioned media Fertile white leghorn chick eggs were obtained from the Poultry Science Department at the University of Florida and set at 37oc in a forced draft incubator. Both the
PAGE 141
132 presumptive jaw and limb muscle were dissected from day 3\-4 (st. 22, Hamburger and Hamilton, 1951) chick embryos. The presumptive jaw muscle, the mandibular process of the first visceral arch, and the limb bud are both regions clearly identified at this time and easily excised with fine forceps. Once removed, the tissue was held at 37c in Hank's BSS until dissociation. To dissociate, the tissue was incubated at 37c in 0.1% trypsin (Gibco) in 1 M phosphate buffered saline (pH 7.2) for 20 minutes. A ratio of approximately 0.16 mg of tissue to each 1 ml of solution was maintained. The incubation medium was then removed and the tissue rinsed twice in the standard F-12 medium without fetal bovine serum. The cells were then dissociated by gentle trituration with a flame narrowed glass pipette. The cells were plated in 2 ml aliquots into 35mm plastic culture dishes. The omission of serum in culturing the target cells was necessary in order to permit a meaningful assay of the protein content of the medium after conditioning by the cells. The protein content of serum greatly exceeds that generated in the conditioning. The conditioned medium was collected after 3 days of culture. The medium was removed and centrifuged for 90 seconds. The supernatant was withdrawn and assayed for the protein content, using the dye-binding assay of Bradford (1976). The protein content was determined (BSA standard) and the level in each batch of CM adjusted to 5 g/ml by
PAGE 142
133 successive dilutions with the standard medium (minus FBS). This adjustment was to insure that responses to CM were not simply a function of general nutritive differences resulting from disparities in protein content. Typically, however, the conditioned media showed equal protein levels so that no adjustment was necessary. The CM was made fresh immediately prior to addition to explants. The explants were cultured in 1.5 ml of medium. This consisted of 1.5 ml of the standard medium (controls) or 1.0 ml of the standard medium plus 0.5 ml of either mandibular process (jaw) conditioned medium (JMCM) or limb bud-conditioned medium (LMCM). The explants were cultured for 2 days before CM was added to the medium. The media were replaced with fresh batches of the respective media after 3 days. The neurite outgrowth of the explants was assessed after 6 days in culture. Culture Conditions The standard culture medium consisted of Ham's F-12, 10% fetal bovine serum (FBS; Gibco and Hyclone), 1.6% glutamine, 1 % penicillin/ streptomycin (Gibco), and 0.7% fungizone (Gibco). This medium was used alone or supplemented with one of the two CM to culture the neural tube explants. To generate the conditioned medium the dissociated target tissues were cultured in this medium without FBS. The explants were cultured in collagen /polyornithine coated 35 mm plastic dishes. For coating, the dish surface was covered
PAGE 143
134 with a sterile solution of 0.5 mg/ml collagen in 0.1% acetic acid and 0.04 mg/ml poly-DL-ornithine. All but approximately 0.2 ml of the solution was removed and the dishes were allowed to dry overnight. The dishes were stored in a desiccator at 4c until use. Analyses Ten groups of explants were cultured,and each group initially consisted of 12-18 explants. All three of the experimental conditions were equally represented within each group so that inter-group variability in conditions could not contribute to the results. General viability of the explants was evaluated daily using a Nikon inverted phase contrast photomicroscope with an attached air curtain incubator. Note was taken of intactness of the explant mass, presence and appearance of neuritic growth, and continued adherence of the explant and its processes to the substratum. After 6 days in culture, the explants were photographed and drawn using a Nikon drawing tube. The micrographs were used for data collection with the drawings providing an alternative if the micrographs were unsatisfactory. Explant area was measured using the Bioquant II digitizing tablet and the digitizing morphometry program. Neuritic outgrowth was assessed using a procedure similar to Pollack et al. (1981). A template of nine concentric circles (center diameter= 20 mm, subsequent intervals 10 mm) was centered over the explant and the
PAGE 144
135 numbers of neurites which intersected each circle were counted until the limits of explant outgrowth had been reached. This means of measure provides an index of neurite length and complexity. The micrographs were magnified 80 X, so the center represented an actual explant diameter of 0.25 mm with each interval representing a 0.125 mm increase in distance from the explant. The "back-up" drawings were magnified only 40 X so the template used when scoring these had a center of 10 mm diameter and intervals between subsequent rings of 5 mm. Results Neurite outgrowth was quantified after 6 days in culture but daily observations were made on the overall condition of the explants. There were no differences between the groups in their general health as indicated by the continued adhesion of the explant and the neuritic processes to the substratum. In addition, none of the groups demonstrated the beaded processes indicative of neuritic deterioration (Pollack, 1980) and most explants showed some degree of neurite outgrowth. There was no difference between groups in the numbers of explants surviving 1 week in culture. Overall, 82% of the explants plated (106 of 130) survived. This is comparable to the survival after 1 week of control explants (86 % or 30 of 35 plated), the LMCM explants (80 % or 35 of 44 plated), and the JMCM explants (82 % or 41 of 50 plated).
PAGE 145
136 The surface area covered by the explants after 1 week in culture did not vary significantly between the groups, although the explants cultured in JMCM did cover a slightly greater area. The average surface extent of explants after 1 week were 0.79 mm2 for controls (S.E.M.= 0.24), 0.80 mm2 for LMCM (S.E.M.= 0.18), and 0.88 mm2 for JMCM (S.E.M.= 0.20). After 6 days in culture the neurite outgrowth was quantified as described above. The resultant neurite growth index is a sum of the numbers of neurite intersections and reflects the process number and complexity. The mean neurite growth indices for the three groups are compared in Figure 5-1. The mean growth index for the control group was 92.6 (S.E.M.= 15.1). This did not differ significantly (P > .20, Mann-Whitney u-test) from the growth index for the LMCM explants which was 109.5 (S.E.M.=10.5). The explants grown in JMCM had a mean growth index of 170.9 (S.E.M.= 17.2) and this was significantly greater than both the control group ( P < .002) and the LMCM group ( P < .004). The general appearance of the neurites in all groups was the same. Typical explants are depicted in Figures 5-2 (control) and 5-3 (JMCM). The outgrowth in the JMCM group was denser and more complex than that seen in the other groups. The neurites in the JMCM group extended greater distances from the explant core, frequently extending 800-1200 m from the explant.
PAGE 146
180 150 130 -0 C ..... 90 <.!) -~ 60 z 30 0 T l T J.. T l. Control Limb Muscle Jaw Muscle Conditioned Media Conditioned Media Figure 5-1: Mean neurite growth index 137 The neurite growth index is an estimator of the quantity of neurite elaborated by an explant. The mean neurite growth indices for trigeminal motor nucleus explants cultured in control medium (N=30), limb-muscle conditioned medium (N=35), and jaw-muscle conditioned medium (N=41) are compared above. Neurite outgowth is significantly enhanced in the jaw-muscle conditioned medium group compared to the limb-muscle conditioned medium group (P<.004 ) and control group (P<.002). The vertical bars represent the standard error of the mean.
PAGE 147
138 NTE Figure 5-2: Control explant Typical trigeminal motor nucleus explant following culture in the control medium for 6 days. Darkfield photomicrograph reveals the sparse neuritic outgrowth from the neural tube explant (NTE). X 66.
PAGE 148
139 Figure 5-3: Jaw-muscle conditioned medium explant. Typical trigeminal motor nucleus explant cultured for 6 days in jaw-muscle conditioned medium. In contrast to the control explant, this darkfield photomicrograph reveals abundant and complex neurite outgrowth from all aspects of the neural tube explant (NTE). X 48.
PAGE 149
140 Discussion The results presented here demonstrate that presumptive muscle tissues from different body regions are not equivalent in their ability to stimulate growth from cultured motor neuron explants. Medium conditioned by muscle from the appropriate target region, the mandibular process, produces greater neurite outgrowth from trigeminal motoneuron explants than medium conditioned by muscle from a normally remote body region, limb bud. While it has previously been demonstrated that muscle-conditioned medium (Dribin and Barrett, 1980, 1982; Henderson et al., 1981) or muscle extract (Hsu et al., 1983; Smith and Appel, 1983) is able to enhance outgrowth from motoneuron populations, the specificity of the effect demonstrated here could have great functional importance in vivo. Enhanced outgrowth from metencephalic explants in the presence of medium conditioned by muscle from the mandibular process is most likely the result of specific responsiveness by the trigeminal motoneurons. There was little variability in the initial size of the explants so that the observed differences in outgrowth are not likely to be attributable to initial differences between the groups. The explants primarily consisted of trigeminal motoneurons as a result of the inclusion of Ara-c in the culture medium. Ara-C serves both to minimize the presence of non-neuronal cells and to limit the neuronal populations to those that have already
PAGE 150
undergone their terminal cell division at the time of its addition. 141 The quantity of neurite outgrowth produced by an explant could result from alterations in the number or specific population of cells within an explant or alterations in the amount of neurite elaborated by a fixed population of cells. It did not appear here that cell survival contributed to the outcome based on the similarities in terminal explant sizes seen in the different groups. While it is possible that increased cell survival may have made a small contribution it does not seem likely that this was the major venue for increased production of neurites. It seems most likely that the alteration in neurite quantity was produced through some combination of (1) an increase in the numbers of cells stimulated to initiate processes, (2) an increase in the numbers of neurites initiated from those cells which formed processes, (3) an increase in the degree of branching which occurs on initiated processes, or (4) a change in the rate or length of processes elaborated. Although Ara-c greatly reduced the numbers of non-neuronal cells present it is possible that some indirect effect of conditioned medium on the non-neuronal cells may have contributed something to the observed differences in outgrowth. In vivo studies have suggested that there is a great deal of specificity which occurs in the pairing of neuron and muscle in the periphery (Laskowski and Sanes, 1987;
PAGE 151
142 Vogel and Landmesser, 1987). The isolation of elements in tissue culture has permitted demonstration of muscle-derived factors which are able to enhance survival and neurite outgrowth from purified populations of motoneurons (Dohrmann et al., 1986; Smith et al., 1986). Dohrmann et al. (1986) precipitated by ammonium sulfate two separate factors from embryonic muscle extract which potentiated either survival of, or neurite outgrowth from purified embryonic chick motoneurons in culture. The neurite promoting factor was found to contain a laminin-like molecule. Similarly, Smith et al. (1986) have isolated several fractions from newborn rat muscle extract which produce specific effects on the development of identified motoneurons in vitro. In particular, two factors were isolated which stimulated neurite outgrowth, a 55 kDa neutral glycoprotein which acts only on motoneurons and a 35 kDa acidic glycoprotein which stimulated general outgrowth from ventral spinal cord neurons. Thus, muscle appears to contain several distinct factors which subserve different trophic functions such as cell survival, process outgrowth, and enzyme synthesis. The different factors are quite specific and selective both in the cellular process affected and in the cell type affected. In addition, several studies have documented the developmental change in the ability of muscle to stimulate neurite outgrowth from spinal cord explants (Nurcombe and Bennett, 1983; Pollack and Muhlach, 1981).
PAGE 152
143 The possible developmental significance of the specific responsiveness observed in the present study is suggested by the normal processes correspondent to the in vitro observations. The explants were observed in culture during the period during which nerve outgrowth occurs in vivo. The effective conditioned medium was generated by muscle tissue during the time period at which it normally would be first receiving neural innervation in vivo. The specific stimulation of neurite formation during the period which corresponds to nerve outgrowth in vivo suggests that muscle may play some important developmental role in influencing process outgrowth from the appropriate motoneurons. The specificity of the matching between nerve and muscle is evidenced in varied ways. Motor axons from several spinal cord levels which are combined within the phrenic nerve preferentially innervate diaphragmatic muscle from similar segmental levels (Laskowski and Sanes, 1987). Motor nerves routed to an incorrect muscle prior to initial innervation apparently find the appropriate muscle fiber type to innervate within the foreign muscle (Vogel and Landmesser, 1987). Motoneurons are distinctly identified prior to contact with muscle based on their position within the central nervous system and their physiological characteristics. While other elements of the embryonic milieu may contribute in guiding nerve growth to the appropriate target, much evidence indicates that muscle
PAGE 153
144 directly affects motoneuron growth. The recognition which occurs between neuron and muscle in the periphery could well be subserved by target-released or membrane-associated molecules. The degree of specificity in the interactions between neuron and muscle and the molecular basis of the interactions remain important questions in motoneuron development.
PAGE 154
CHAPTER VI THE ONTOGENY OF SPECIFIC RETROGRADE TRANSPORT OF NERVE GROWTH FACTOR (NGF) BY MOTONEURONS OF THE BRAINSTEM AND SPINAL CORD. Introduction The neurotrophic Nerve Growth Factor (NGF) has been demonstrated to be physiologically significant for the development of sympathetic and neural crest-derived sensory neurons (see review in Levi-Montalcini and Angeletti, 1968). It was long believed that NGF effectiveness was limited to these two neuronal populations. Recently, however, evidence has emerged indicating that central cholinergic populations are also (1) responsive to NGF (Gnahn et al., 1983; Kromer, 1987), (2) exhibit specific uptake and retrograde transport of injected l25I-NGF (Seiler and Schwab, 1984), and (3) bear specific NGF receptors (Hefti et al., 1986; Richardson et al., 1986; Taniuchi et al., 1986). The specific binding of radiolabeled NGF, a consistent hallmark of receptors for NGF, has been described in the adult rat brain (Richardson et al., 1986) and in the developing brain of the chick embryo (Raivich et al., 1985, 1987). Somewhat surprisingly, in the embryonic chick, motoneurons of the spinal cord and brainstem exhibit 145
PAGE 155
146 specific NGF binding. This NGF-binding is only evident transiently during an early developmental period. Some evidence also exists that motoneurons are responsive to NGF during this time: Explants of 40 hour chick brainstem basal plate respond to NGF with increased neuritic outgrowth (Heaton, 1987). Retrograde transport of NGF has been shown to occur selectively in NGF-responsive populations (e.g., Hendry et al., 1974). The interruption of retrograde transport in sympathetic neurons by ligation or by colchicine (Purves, 1976), results in a neuronal response similar to that observed following antibody-induced NGF deprivation. While the role of retrogradely transported NGF in directing the cellular response has not been established, it is clear that NGF-responsive cells selectively transport both exogenous and endogenous NGF (Stockel et al., 1975a, 1975b; Korsching and Thoenen, 1983). The physiological significance of NGF to the early development of motoneuron populations remains to be clarified. An initial step in this direction is the determination that in addition to specific binding, the motoneurons are capable of uptake and retrograde NGF transport. Previously, we have described the specific retrograde transport of labeled NGF by embryonic chick spinal cord motoneurons (Wayne and Heaton, 1988). The present study was an attempt to define the temporal limits
PAGE 156
147 of this transport capacity. In addition, we were interested in extending the observation of NGF transport to motoneurons of the developing trigeminal (V) motor nucleus, which also bind labeled NGF (Raivich et al., 1987), and which appear to be responsive to NGF early in their development (Heaton, 1987) Methods The ~-NGF (generously provided by Dr. Michael Young and Dr. Eugene Johnson) was iodinated by the chloramine-T method (Young et al., 1979). Briefly, 100 l of lM and 200 l of O.lM potassium phosphate buffers (pH=7.0), and 100 l ~-NGF (82.2 g/ml) were combined with 2 mci of Na125r in the vbottom shipping vial and mixed thoroughly by micropipetting. The reaction was begun with the addition of 10 l chloramine-T (1 mg/ml), continually mixed by micropipette, and allowed to proceed exactly 30 seconds before 10 l of mercaptoethanol (diluted 1:400 in distilled water) was added to halt the reaction. The resultant solution was then placed on a plastic column (1 X 1 cm) of Bio-Rad AG-1X8 resin (200-400 mesh) which had been equilibrated with 0.1% bovine serum albumin in O.lM potassium phosphate buffer (pH=7.0). The fractions of about 200 leach were collected in plastic tubes typically yielding three peak fractions beginning with the third or fourth fraction collected. The specific activity was 10-40 Ci/g NGF and the 125r-NGF was used
PAGE 157
148 within five days of labeling. Cytochrome-C (Sigma) was iodinated by the same method for control injections. Fertile White Leghorn chicken. eggs were obtained from the Poultry Science Department at the University of Florida and were set in a forced draft incubator at 37.5c. Chick embryos at 3 4 5, 6, 8, 10, 11 and 14 days of incubation were injected with 125r-NGF in the mandibular process of the first visceral arch (the precursor to jaw musculature innervated by motor V) or the jaw muscles proper (pseudotemporalis profundus, pseudotemporalis superficialis and adductor mandibulae). Similarly, injections into the caudal limb bud or leg muscle were made at 4, 5, 6, 7, 8, 10, and 13 days of incubation. The lumbar region was always the target of this inquiry except in one case in which the wing muscle was injected in a chick embryo at 14 days of incubation and the brachial region subsequently examined. The early injection times were chosen to correspond to the times of initial axonal ingrowth into the target structures (mandibular arch day 3-3! ; Heaton et al., 1978 ; limb bud, day 4 !-5; Tosney and Landmesser, 1985c). The final injection times were chosen to correspond to the times at which the motoneurons are likely to cease transporting NGF as predicted by the observed NGF binding to motoneuron populations of the brainstem and spinal cord (Raivich et al., 1985, 1987).
PAGE 158
149 Eggs were candled at the appropriate ages and a lateral window made in the shell overlying the embryo. A small hole was cut in the chorioallantoic membrane taking care to avoid prominent blood vessels, and, for caudal limb bud or leg injections, the amnion was retracted to expose the injection area. Either the limb or jaw region was immobilized using a hair loop or a blunt glass probe. A Hamilton 5 l syringe with an affixed glass micropipette tip was mounted in a micromanipulator for the injections. The glass needle was guided to the injection site under a dissecting microscope. The insertion of the needle and delivery of the injection were visually verified under the dissecting microscope while a second investigator controlled the syringe cannula and the volume injected. The injections ranged from 0.1 -1.2 l volumes with 0.1 l volume approximating from 30,000-86,790 cpm (1.750-1.366 ng fi-NGF). The animals were allowed 3\-6 hours survival time before being sacrificed by decapitation. The tissue was fixed by immersion in 1% paraformaldehyde/ 2.5% glutaraldehyde for 1-2 hours at room temperature. The tissue was then thoroughly rinsed in O.lM potassium phosphate buffer (pH=7.4) overnight. The tissue was dehydrated in a graded series of alcohols and xylene and embedded in paraffin. Serial 6-10 micron sections were cut and mounted on poly-L-lysine (5 mg/ml) coated slides. The slides were de-paraffinized and air dried before being dipped in Kodak NTB2 autoradiography
PAGE 159
150 emulsion diluted 1:1 with distilled water. After being dipped, the slides were stored in light tight boxes at 4c for 1-3 months. The slides were developed in Kodak D-19, lightly counterstained with hematoxylin, and coverslips were affixed with Permount. The slides were examined with both darkfield and brightfield optics on a Zeiss photomicroscope. Results Jaw musculature injections The jaw muscle or presumptive muscle region was injected with 125r-NGF in chick embryos from 3 !-14 days of incubation. Following injections at 3 4 5, 6 and 8 days of incubation labeled embryos displayed an accumulation of grains predominantly ipsilateral to the injection localized to some or all of the following: (1) the peripheral injection site, (2) along the peripheral nerve, (3) overlying cell bodies of the trigeminal (V) ganglion, (4) overlying the brainstem entry zone region and (5) overlying cell bodies of the lateral trigeminal motor nucleus. Following injections at 10, 11, and 14 days of incubation the label was localized to three of the five areas seen labeled at the earlier stages: (1) the peripheral injection site, (2) along the peripheral nerve and (3) overlying cell bodies of the trigeminal (V) ganglion. Table 6-1 summarizes the labeling observed in the trigeminal region following peripheral injections of 125I-NGF. Figures 6-1, 6-2, and 6-3
PAGE 160
AGE 3.5 4.5 5 6 8 10 1 1 14 n 5 4 5 5 3 2 3 1 INJ SITE NERVE ++ +++++ +++ +++ +++ +++++ ++++ +++++ +++ +++ ++ ++ +++ +++ + + TABLE 6-1 : NGF TRANSPORT IN TRIGEMINAL SYSTEM V GANG +++(+) ++++ +++++ +++++ +++ ++ +++ + V ROOT ENTRY ZONE +++++ ++++ +++++ +++++ + V MN ++(++) ++++ +++++ ++(+) + MARGINAL ZONE +++++ ++++ ++++(+) +(+) PATH EZ TO TECTUM ++++ ++++ ++++ ++++ NOT E Each + denotes one case which had grains present in that structure. The indicates that no cases showed label in that structure. The parentheses (+) indicates that the labeling was very slight. Abbreviations: V, trigeminal; E Z, entry zone; GANG, ganglion; MN, motoneuron; INJ, injection I-' Ul I-'
PAGE 161
Figure 6-1: NGF transport by trigeminal (V) motoneurons at days 3 and 5 of incubation Cross section of the metencephalic neural tube following peripheral injections of 125r-NGF at day 3 (6-la, b, c, d) and day 5 (6-le, f, g, h). Labeling is bilateral at both ages although always more intense on the left, injected, side. The continuity of the target region across the midline makes it impossible to confine injections to just one side. At day 3 bright field (6-la, c) and dark field (6-lb, d) photomicrographs reveal the presence of transported NGF in the accumulation of autoradiographic silver grains. This label is most prominent in the entry zone (Z) region where trigeminal sensory afferents enter the brainstem. The trigeminal motor root (MR) which courses along the medial aspect of the trigeminal ganglion (G) is not heavily labeled. The trigeminal motor population (M) which lies within the metencephalon just medial to the afferent fiber entry zone is clearly labeled (6-lb, d). While there is a diffuse increase in grain density over the trigeminal ganglion there is not the discrete labeling of cell bodies which is seen at later stages (see Fig. 6-2). The mandibular process (P) which is the target region for the trigeminal motor fibers can be partially seen at the lower magnification (6-lb). (a), (b) X 234; (c), (d) X 469. At day 5 bright field (6-le, g) and dark field (6-lf, h) photomicrographs reveal the accumulation of silver grains over the entry zone (Z) which is the region where the trigeminal sensory fibers enter the brainstem and over the trigeminal lateral motor nucleus (M). Label which extends dorsally from the entry zone region (arrow, 6-lf) may represent fibers of the mesencephalic sensory nucleus. (G, ganglion) (e), (f) X 234; (g), (h) X 410.
PAGE 162
153
PAGE 163
154 Figure 6-2: NGF transport by trigeminal motoneurons at day 8 of incubation Cross section of the rostral metencephalon (trigeminal reqion) of the day 8 chick embryo following injection of 125I-NGF into the jaw muscle. At low power brightfield (a) and darkfield (b) photomicrographs reveal the accumulation of silver grains over the trigeminal motor nerve (motor root, MR), over the trigeminal ganglia (G), and over the trigeminal lateral motor nucleus (M, arrow). At higher magnification (c, d) the label is clearly localized to cell bodies in the lateral motor nucleus (arrow) and trigeminal ganglion (g). In darkfield (d) a slight increase in grain density is evident in the brainstem at the point of sensory fiber entry. (a) X 138; (b) X 133; (c), (d) X 492.
PAGE 164
155 .., / cl .. .I.' Figure 6-3: Trigeminal motoneurons cease transport of NGF by day 11 of incubation Cross section of the brainstem trigeminal region of the day 11 chick embryo following 125I-NGF injection of the jaw muscle. At both low (a, b) and high (c, d) magnification brightfield (a, c) and darkfield (b, d) photomicrographs clearly evidence the labeling of cells within the trigeminal ganglia (g) and the absence of label in the trigeminal lateral motor nuclei (m). The motor root (mr) which courses along the medial aspect of the ganglion and is visible on both sides at this level (a, b) is not labeled. (a), (b) X 112; (c), (d) X 245.
PAGE 165
156 show the typical labeling pattern in the trigeminal region of the brainstem at various ages. The point at which the V ganglionic fibers accumulate to enter the brainstem, the equivalent of the spinal cord dorsal root entry zone, was labeled at the younger ages but was not labeled with injections of embryos at day 10 or older. The labeling of this entry zone was more intense at the younger ages. The disappearance of the labeling of the entry zone region corresponded with cessation of labeling by the motoneurons. In embryos receiving injections at the earlier times, particularly day 5, in addition to the areas previously cited label extended dorsally along the lateral borders of the brainstem. This corresponds to the course which the peripheral fibers of the mesencephalic sensory nucleus of the trigeminal system take within the central nervous system. The V motor neurons were observed to be labeled following injections at days 3 } 4 } 5, 6 and 8 (Table 6-1). There was never any label evident in the dorsal motor V nucleus which is formed by a secondary dorsal migration of cells from the lateral nucleus between days 5 and 6 of incubation (Heaton and Moody, 1980). Additionally, at the younger ages, particularly days 3 } and 4 } there was a diffuse increase in grain density in the marginal zone of the brainstem basal plate region (Figure 6-3). The marginal zone labeling extended medially and appeared to include the medial cell column which lies just lateral to the midline
PAGE 166
Figure 6-4: Embryonic trigeminal motoneurons transport NGF specifically Cytochrome-C is a molecule which is similar to NGF in its molecular weight and isoelectric point. In the darkfield photomicrograph (6-4a), the mandibular process of the first visceral arch which is the target region for trigeminal motoneurons has been labeled in the day 6 chick embryo by injection of 1251-cytochrome-C. The dashed line outlines the limits of the mandibular process. Although the process is continuous across the midline (L), the label here appears on only one side. Transverse section of the metencephalic trigeminal region of the day 6 chick embryo following injection of the target region with radiolabeled cytochrome-C. Although the trigeminal ganglion is not present on the injected side at this level the sensory fibers can be seen entering the brainstem (F) in the brightfield image (6-4b). The lateral motor nucleus (m) lies just medial to the entering sensory fibers. The darkfield photomicrograph (6-4c) of this section reveals the lack of any radiolabel in either the lateral motor nucleus or the sensory fibers. (L, midline). (a) X 362; (b), (c) X 332.
PAGE 167
158 L ; 1' .. Jr-. .... (, : :'> i # : ,,.., -_). > ? ~ ... J ._;:.> V
PAGE 168
159 below the ventricular cell layer. Lateral to the medial cell column is the area traversed by V motor neurons in their medial to lateral migration, from their site of origin in the medial cell column to their termination site as the lateral motor V nucleus. Also contributing to the marginal zone are the processes of trigeminal motoneurons which are already extended to the peripheral target region during the course of the medial to lateral cell migration in the brainstem (Heaton and Moody, 1980). Injections of the jaw muscle at days 10, 11 and 14 of incubation resulted in the accumulation of grains over the injection site, the peripheral nerve, and the trigeminal ganglion, but not in the V motor nucleus or the entry zone region of the brainstem. Control embryos which were injected with 125r-cytochrome-C displayed an accumulation of grains only at the peripheral injection site (Figure 6-5). Limb bud injections The leg muscle or limb bud was injected with 125r-NGF in chick embryos at 4, 4!, 5, 6, 7, 8, 10 and 13 days of incubation. The wing muscle was injected in one embryo at 14 days of incubation. Table 6-2 summarizes the distribution of label following these injections. Those cases injected prior to day 13 demonstrated an accumulation of grains predominantly ipsilateral to the injection site and localized: (1) to the peripheral injection site, (2) along the peripheral nerve and ventral roots, (3) overlying the
PAGE 169
160 cell bodies of the dorsal root ganglion, (4) overlying the dorsal root entry zone, and (5) overlying the motoneuron cell bodies in the lateral motor column. Injections at 13 and 14 days resulted in label localized: (1) to the injection site, (2) along the peripheral nerve and (3) overlying cell bodies of the dorsal root ganglion. The typical pattern of labeling after limb bud or leg injections of NGF at various ages are depicted in Figures 6-5, 6-6, 6-7 and 6-8. The accumulation of grains over the dorsal root entry zone was particularly evident in those cases injected at 4 and 5 days of incubation. The intensity of labeling in the entry zone region diminished markedly with age. It was barely detectable at 10 days of incubation and was not evident at all at the older injection times. Labeling of the lateral motor column was observed following 125I-NGF injections at days 4, 4!, 5, 6, 7, 8 and 10 of incubation (Table 6-2). The label of lumbar motoneurons following injections at day 4 was barely detectable while that following injections at day 4 was definite. In a few cases in which very large injections were made, specific labeling was seen bilaterally. In all cases of bilateral labeling the label was much heavier on the side ipsilateral to the injection. Injections after day 10 (days 13 and 14) did not produce specific label of the lateral motor column while the peripheral nerve and the dorsal root
PAGE 170
TABLE 6-2: NGF TRANSPORT IN SPINAL CORD DORSAL LMC & INJ ROOT VENTRAL SITE NERVE DRG FIBERS DREZ ROOTS AGE n 4 2 ++ ++ ++ ++ 4 5 1 + + + + + 5 4 ++++ ++++ ++++ ++ ++++ ++++ 6 2 ++ ++ ++ ++ ++ ++ 8 4 ++++ ++++ ++++ ++++ +++ ( +) ++++ 1 0 7 ++++ ++++ ++++ (+) (+++) ++++ + ++ + + + ++ + ( +) 13 1 + + + 14 1 + + + NOTE Each+ denotes one case which had grains present in that structure. The -indicates that no label was evident in that structure. The parentheses (+) indicates that the labeling was very slight. Abbreviations: INJ, injection; DRG, dorsal root ganglion; DREZ, dorsal root entry zone; LMC, lateral motor column.
PAGE 171
Figure 6-5: Spinal cord motoneurons transport 125r-NGF at day 6 of incubation. (a) Transverse section of the lumbar spinal cord following limb injection with radiolabeled NGF at day 6 of incubation. In addition to the presence of label within the main nerve trunk (n), label is localized to muscle nerve branches (mn) and cutaneous nerves (arrowheads). Ipsilateral to the injection, silver grains are accumulated over the dorsal root ganglion (DRG) and the lateral motor column (LMC). (b) At higher magnification, the label in the dorsal root ganglion (DRG) and the lateral motor column (LMC) is clearly localized to neuronal cell bodies (short arrow). The motor root appears to have been cut superficially leaving a trace of label (curved arrow). (a) X 227; (b) X 910.
PAGE 172
163
PAGE 173
a C ./ Y .. I j r 164 Figure 6-6: Spinal cord motoneurons transport l25r-NGF at day 10 of incubation. Transverse section of the lumbar spinal cord following injection of the target limb with radiolabeled NGF at day 10 of incubation. Low power brightfield (6-6a) and darkfield (6-6b) photomicrographs reveal labeled cells in the dorsal root ganglion (DRG) and in the lateral motor column (LMC) ipsilateral to the injection site. The labeling of a discrete cluster of motoneurons which is readily seen at higher magnification (6-6c) suggests transport from a single muscle. The path of the motor root is marked by the arrow in 6-6c. Label of the motor root fibers (MR) is apparent when the section is viewed by darkfield microscopy (6-6d). The profile of labeled motoneuron cell bodies evident in Figure 6-6d (arrow) results from the very high grain density accumulated over the cells which paradoxically appear black by darkfield microscopy. (a), (b) X 275; (c), (d) X 551.
PAGE 174
165 Figure 6-7: Spinal cord motoneurons do not transport 125r-NGF at day 13 of incubation. Transverse section of the brachial spinal cord following injection of wing muscle with radiolabeled NGF at day 13 of incubation. The injection was ipsilateral to the spinal cord region shown by brightfield (6-7a) and darkfield (6-7b) microscopy. Radiolabel was localized exclusively to cell bodies of the ipsilateral dorsal root ganglion (DRG). Neither the motor root (MR) nor the motoneurons of the lateral motor column (LMC) were labeled. Autoradiographic artifact which appears below the spinal cord is due to emulsion deposition in the dense cartilage. (a), (b) X 373.
PAGE 175
166 a b .... J \ .r Figure 6-8: Embryonic spinal cord motoneurons transport NGF specifically. Transverse section of the lumbar spinal cord at day 5 of incubation following target limb injection with either (a) 1251-NGF or (b) 1251-cytochrome-C. Following the injection of radiolabeled NGF, specific label is apparent in the peripheral nerve (N), cells of the dorsal root ganglion (DRG) and lateral motor column (LMC), and at the dorsal root entry zone (arrowhead). A similar injection of radiolabeled cytochrome-C results in the presence of label at the peripheral injection site (In) with no label present in either the dorsal root ganglion (DRG) or the lateral motor column (LMC). (a) X 285; (b) X 185.
PAGE 176
167 ganglia still demonstrated an accumulation of grains (Fig 6-7). The specificity of the label was tested with comparable injections of 125I-cytochrome-C. In none of these cases was there an accumulation of grains overiying neural structures, but only that present at the injection site (Fig. 6-8). Discussion In this study we have described the retrograde transport of radiolabeled NGF by embryonic motoneurons of the chick brainstem and spinal cord. In the lumbar spinal cord, peripheral injections of 125I-NGF on embryonic day 10 but not day 13 resulted in labeling localized to limb structures, peripheral nerves, cell bodies in the dorsal root ganglia and lateral motor column, and the dorsal root entry zone. Similarly, in the brainstem trigeminal system, peripheral injections of 125I-NGF on embryonic day 8 but not day 10 resulted in labeling localized to the jaw, peripheral nerves, cell bodies in the trigeminal ganglion and trigeminal lateral motor nucleus, the brainstem entry zone region, and, possibly, the fibers of the mesencephalic nucleus of Vas they course from the brainstem entry zone region dorsally to the tectum. Central transport of NGF At the later injection times (day 10 trigeminal region, day 13 spinal cord), transported NGF did not appear in any structures of the central nervous system (CNS). The observed transport of 125I-NGF into the CNS was a characteristic of
PAGE 177
168 transport at the early embryonic stages. The central transport of NGF in early chick embryos appeared to occur in several neuronal populations. Most notably, the retrograde transport of NGF by early embryonic motoneurons of the brainstem and spinal cord is not seen at later stages. The neural crest-derived neurons of peripheral sensory ganglia retrogradely transport NGF during development and in the adult (Brunso-Bechtold and Hamburger, 1979; Stockel et al., 1974). The observed retrograde transport of NGF by neural crest-derived sensory neurons of both the trigeminal and dorsal root ganglia is not a novel finding but simply extends this previously observed transport to earlier developmental times. In the present study, however, the sensory neurons, during early embryonic stages, appeared to transport NGF beyond the cell body and into the CNS. This was evidenced by the presence of silver grains localized to the CNS region where peripheral sensory fibers enter the CNS. The accumulation of grains over the entry zone region was clearly seen until day 8 of incubation in both the spinal cord and brainstem following peripheral injections. Specific to the very early embryonic stages, then, the ganglionic sensory neurons appear to anterogradely transport NGF from their cell bodies into the central nervous system. Also, exclusive to the early injections in the trigeminal system, it appeared that fibers of the trigeminal mesencephalic nucleus, a neural crest-derived sensory
PAGE 178
169 nucleus located in the CNS, may specifically transport NGF. The path that the fibers of the mesencephalic nucleus of V follow in the brainstem was labeled at the early embryonic stages (days 3 4 5 and 6). The transport of NGF into the CNS of the embryonic chick is consistent with the widespread presence of NGF-like immunoreactivity previously described in the fetal rat brain (Ayer-Lelievre et al., 1983). Retrograde transport of NGF Specific binding of NGF can be mediated by either of two distinct NGF receptors. The two types of NGF receptors, slow or high affinity receptors (type I) and fast or low affinity receptors (type II), have been identified in a variety of cell types (sensory, Sutter et al., 1979; sympathetic, Godfrey and Shooter, 1986; PC12, Hosang and Shooter, 1985). While the roles of the two receptor populations are still somewhat unresolved, it has been determined that binding to the high affinity receptor is sufficient for internalization (Hosang and Shooter, 1987) and the initiation of neurite outgrowth to occur (Gunning et al., 1981; Sutter at al., 1979). In clonal cell lines with receptors for NGF, NGF is internalized only via the high affinity receptor (Bernd and Greene, 1984). Additionally, mutant cell lines that possess only the low affinity receptor neither internalize NGF nor demonstrate a response to it (Green et al., 1986). Thus, it appears that only the high affinity receptor is involved in
PAGE 179
internalization, retrograde transport, and the subsequent initiation of a biological response. 170 The motoneuron receptor has yet to be isolated and characterized. The NGF receptors in the brainstem of the embryonic and adult rat, however, have been found to be identical to those isolated from the peripheral nervous system (Yan and Johnson, 1987; Taniuchi et al., 1986). It is possible that motoneurons may have some previously uncharacterized form of NGF receptor uniquely present in the embryonic nervous system (Bernd, 1986). However, the binding and transport of NGF displayed by motoneurons is consistent with that reported for other NGF responsive populations and suggests the presence of at least the high affinity (type I) receptor. The retrograde transport of NGF is thought to be significant in mediating its cellular effects. The signal which evokes this cellular response is not known but both NGF and the NGF receptor are likely candidates. Previous research has revealed that biologically responsive neuronal populations transport both exogenous (Stockel et el., 1975b) and endogenous NGF (Korsching and Thoenen, 1983; Palmatier et al., 1984). Although the NGF molecule is intact until reaching the cell body (Johnson et al., 1978; Dumas et al., 1979), it alone is not the signal. If either NGF or antibodies to NGF are introduced directly into the cytoplasm of PC12 pheochromocytoma cells by microinjection or cell
PAGE 180
171 fusion, there is no effect on neurite formation or the induction of choline acetyltransferase (Heumann et al., 1981, 1984; Seeley et al., 1983). In addition to the transport of NGF itself, Johnson et al. (1987) have documented the in vivo retrograde transport of the receptor for NGF. A current postulate is that the NGF-NGF receptor complex is the cellular signal which induces the biological response (Compito et al., 1986). Motoneuron NGF-transport The retrograde transport of NGF by motoneurons was specific in two respects: temporally and molecularly. The ability of motoneurons to transport NGF was found to be specific to an early developmental period. Motoneurons of the brainstem and spinal cord no longer retrogradely transport radiolabeled NGF at embryonic day 10 in the brainstem trigeminal region and embryonic day 13 in the lumbar spinal cord. At all embryonic ages injected, sensory neurons of the trigeminal and dorsal root ganglia persisted in the uptake and retrograde transport of injected NGF. NGF transport was not a function of some non-selective ability of embryonic motoneurons to retrogradely transport molecules from the periphery: comparable injections of radiolabeled cytochrome-C, a molecule similar to NGF in its molecular weight and isoelectric point, never resulted in labeling of neurons of the brainstem or spinal cord. Spatially, the label was not always unilateral although the proportion of
PAGE 181
172 label was always greater ipsilateral to the injection site. This observation is consistent with previous reports of specific transport following intraocular injections of NGF in which leakage into the peripheral circulation was thought to contribute to the contralateral labeling (Stockel et al., 1974). Additionally, in the trigeminal system, the continuity of the jaw region across the midline makes it difficult to inject one side completely exclusive of the other. The time of cessation of NGF transport by embryonic motoneurons of the brainstem and spinal cord approximates that which would be predicted by the in vitro binding studies of Raivich et al. (1985, 1987). The diminution of radiolabeled NGF binding to motoneuron populations followed a rostro-caudal progression, decreasing in the metencephalon around day 8 of incubation followed by a decrease in the lateral motor column around day 12 of incubation. In the present study, trigeminal motoneurons transported NGF at day 8 but not at day 10 of incubation while lumbar motoneurons transported NGF at day 10 but no longer did so at day 13 of incubation. The retrograde transport of 125I-NGF by day 10 embryonic lumbar motoneurons which was consistently seen in the present study differs from the previously published report of Brunso-Bechtold and Hamburger (1979). These investigators found no retrograde transport of radiolabeled NGF by spinal cord motoneurons after subcutaneous implant of
PAGE 182
173 an NGF infiltrated polyacrylamide gel pellet in the leg of day 10 chick embryos, although NGF was transported by the dorsal root ganglia. There are several differences between the delivery of NGF in this earlier study compared to that in the present study which could contribute to the differing results. Brunso-Bechtold and Hamburger (1979) used relatively small quantities of labeled NGF (0.26 ng 0.03 ng) compared to the quantities injected in the current study (from 3 to 18.5 ng NGF depending on age injected). In addition, in the earlier study, the access of the NGF to limb structures depended on the rate of diffusion from the pellet and the distance of the pellet from the innervated structures (i.e., muscle and skin). The direct injection of a solution of NGF into the limb, as in the present study, is likely to have provided greater availability of NGF to limb muscles than the subcutaneously implanted pellet. Recently, the use of labeled cDNA or cRNA probes for mRNA have permitted the identification of NGF-producing cells. Schwann cells of the neonatal but not adult rat and embryonic chick hindlimb muscles (day 17) express mRNANGF (Bandtlow et al., 1987; Hulst and Bennett, 1986). It seems, therefore, that during development, peripheral nerves may have more than one source of NGF available to them. The retrograde transport of NGF by motoneurons does not provide any evidence for a biological response by motoneurons to NGF. It strongly suggests, however, that such
PAGE 183
174 a response may exist. To date, specific retrograde transport of NGF has only been described in neurons known to be responsive to it. The presence of the high affinity NGF receptor, internalization of NGF after binding, and a biological response to NGF are highly correlated in a variety of cell types. In all cases to date, internalization of NGF is integral to a subsequent cellular response. The biological significance of NGF transport to the development of motoneurons is not known. A recent evocative report describes an increase in the size and basophylia of spinal cord motoneurons in Xenopus laevis tadpoles following seven daily NGF injections (Levi-Montalcini and Aloe, 1985). In vitro, Xenopus embryonic neural tube dissociates respond to NGF with an increased rate of elongation and increased branching of neurites (Katz, 1986) while chick metencephalic basal plate explants respond to NGF with increased neuritic outgrowth (Heaton, 1987). These reports demonstrate that NGF can influence the growth of centrally derived neuronal populations and lend support to the suggestion that NGF may be significant to normal motoneuron development. It is clear that motoneuron populations transport NGF during a very dynamic period of their development. The spinal cord motoneurons transport NGF as early as day 4-4! of incubation and no longer do so at day 13 of incubation. In the brainstem, the trigeminal motoneurons were observed
PAGE 184
175 to retrogradely transport NGF at day 3 of incubation and no longer evidenced such transport by day 10. At the time of initial advance into the target limb (stage 24!-25 or 4 !-5 days; Tosney and Landmesser, 1985c), the spinal cord motoneuron axons are able to bind and transport NGF. The trigeminal motoneuron axons are at or near their target site by day 3-3! of incubation (Heaton et al., 1978). Like motoneurons of the spinal cord, then, trigeminal motoneurons retrogradely transport NGF during the very early stages of axonal ingrowth into the target structure. Along with the axonal outgrowth that occurs during this period of NGF sensitivity, there are also other significant developmental events occurring. In the limb, there is formation of the individual muscle nerves and the establishment of specific connectivities with the target muscles (Tosney and Landmesser, 1985c). Within the central nervous system, motoneurons of the trigeminal system translocate from their medial origin to their adult lateral position in the brainstem. This lateral migration of trigeminal motoneurons occurs from 50 hours to day 6 of incubation (Heaton and Moody, 1980) and appears to be dependent on the ingrowth of the ganglionic fibers into the brainstem (Moody and Heaton, 1983b, 1983c). It is possible that the responsiveness of trigeminal motoneurons to NGF begins before day 3 and that NGF plays a role in the developmental interaction between the V ganglion and the V
PAGE 185
176 motoneuron populations. During the latter part of the period of NGF sensitivity, the motoneuron populations are depleted by the process of natural cell death. In addition to the morphological changes occurring during this period the motoneurons are also differentiating biochemically. One indicator of this maturation process at the molecular level is the appearance of calcitonin gene-related peptide in spinal cord motoneurons of day 9 chick embryos (Fontaine et al., 1986). Further investigation is needed to determine the role that NGF may play in mediating these various aspects of motoneuron development: axonal outgrowth, specific connectivity with target muscle, cell migration, naturally occurring cell death and biochemical differentiation. Mesencephalic y nucleus In the brainstem, the trigeminal motoneuron population may not be the only central neuronal population of the trigeminal system to retrogradely transport injected NGF. The mesencephalic sensory population arises from the cranial neural crest and subsequently migrates into the central nervous system (CNS). The fibers of the mesencephalic V nucleus carry proprioceptive information from the jaw musculature into the CNS via the trigeminal root. The cell bodies lie dorsally in the tectum rostral to the level of the trochlear nucleus. Retrograde transport of NGF by the mesencephalic sensory nucleus of V was suggested by the presence of a path of silver grains extending from the entry
PAGE 186
177 zone region dorsally to the overlying tectum. This path of silver grains follows the course of the fibers of the mesencephalic V neurons within the central nervous system. Unfortunately, our tissue sections did not extend far enough rostrally to include the cell bodies of the mesencephalic nucleus of V. Accordingly, it is not possible to definitively state that mesencephalic V neurons have transported NGF in this experiment. Cell culture of mesencephalic sensory neurons isolated from day 10 chick embryos has indicated that, while the cells have receptors for NGF, varied NGF concentrations produce no observable differences in cell survival or process formati~n (Davies et al., 1987). As indicated by the temporal pattern of cranial NGF tranport presented in this study, neurons from the mesencephalon of day 10 chick embryos are likely to be beyond their period of NGF sensitivity. In vivo, Straznicky and Rush (1985a) found that NGF administered daily in chick embryos from day 6 through days 11, 12 and 14 did not prevent normal cell death in the mesencephalic V nucleus but rather seemed to delay it by approximately 48 hours. These investigators also reported that NGF had no effect on the nuclear size of the mesencephalic V neurons at day 11. It seems important to closely examine the influence of NGF on the mesencephalic V nucleus prior to day 10 before sensitivity to NGF is excluded. The suggested transport of NGF by the
PAGE 187
178 mesencephalic nucleus of Vin chick embryos before day 6 of incubation deserves further investigation. The possible role of NGF in the early development of the mesencephalic V neurons, specifically, and in developing sensorimotor systems, more generally, needs to be clarified. Sensory afferents In both the spinal cord and the brainstem, transport of 125I-NGF into the central processes of ganglionic afferents was evidenced by the accumulation of silver grains over the respective fiber entry zones. Similarly, Aloe and LeviMontalcini (1985) have reported the accumulation of silver grains over the tracts of the peripheral sensory fibers within the brainstem and spinal cord following radiolabeled NGF injections into the periorbital region of Xenopus laevis tadpoles. In the brainstem and spinal cord, the entry zone region is composed of fibers derived from (1) entering ganglionic fibers and/or (2) rostro-caudally oriented sensory fibers which have entered at another CNS level. In the adult rat there is no evidence of NGF in either the dorsal roots or the spinal cord (Korsching and Thoenen, 1985). The presence of labeling in the entry zone region seen in the chick in the present study and also in Xenopus laevis tadpoles (Levi-Montalcini and Aloe, 1985) suggest that NGF may be present in both the dorsal roots and the CNS early in development.
PAGE 188
179 The eniry zone labeling observed in the present study diminished with the age of the embryos injected. While the quantities of NGF injected were small (3-18.5 ng NGF depending on the embryonic age), these quantities are still likely to be considerably more than the embryonic system is normally exposed to during those developmental stages. Thus, it is possible that the accumulation of label in central ganglionic processes was an anomaly due to overloading the cellular capacity to handle internalized NGF. This seems unlikely, however, considering that the uptake of NGF is apparently mediated by a specific and finite receptor population. The developmental significance of the central transport of a peripherally derived factor is of particular interest. Such central transport of NGF could conceivably be important in the development of the central processes of the peripheral sensory cells or, alternatively, the growth factor could be carried centrally and transferred to another neuronal population. In the lumbar spinal cord there is some evidence to suggest that ganglionic afferents provide trophic support important to motoneuron survival during development (Okada and Oppenheim, 1984). However, exogenous supplementation of NGF during in vivo development of the chick has produced no change in the number of lumbar motoneurons surviving either the natural period of cell
PAGE 189
180 death (Oppenheim et al., 1982) or cell death induced by limb removal during this same period (Bueker, 1948). In the brainstem, the central processes of the V ganglion do not make direct synaptic contact with the V motoneurons as the dorsal root afferents do in the spinal cord (Dubbledam and Karten, 1978). In the trigeminal system, then, the direct synaptic transfer of NGF from the central processes of afferent fibers to motoneurons is not possible. However, the possibility that the NGF present in the central sensory processes has a trophic effect on V motoneuron development can not be excluded. In vivo removal of the trigeminal ganglion precursor cells suggests that the V ganglion provides some trophic stimulus critical to V motoneuron development (Moody and Heaton, 1983a, 1983b). Along with the depletion of the trigeminal ganglion population produced by the surgical excision there is a proportional decrease in the trigeminal motor population (Moody and Heaton, 1983a). Both the mediator of the ganglionic influence on motoneuron development and the role of centrally transported NGF in the development of the trigeminal system have yet to be determined. NGF in early development In summary, during the early stages of development of the embryonic chick, there is specific transport of NGF which is not observable at later times. Embryonic motoneurons of both the brainstem and the spinal cord
PAGE 190
181 retrogradely transport NGF in a specific manner. The functional significance of NGF transport to motoneuron development is not clear but the period of sensitivity to NGF is one which encompasses many events in the process of motoneuron differentiation. In addition to the transport by embryonic motoneurons, central sensory neurons of the mesencephalic nucleus of V may retrogradely transport NGF. Peripheral sensory neurons of the trigeminal and dorsal root ganglia retrogradely transport NGF throughout development and anterogradely transport NGF into their central processes transiently. The specific and transient nature of the transport of NGF into the central nervous system suggests that NGF may have a critical role in the ongoing early differentiative events. Thus, the role of NGF in development is expanding in two dimensions: (1) temporally it seems to be important at earlier developmental times than had been traditionally believed, and (2) spatially it seems to be have some role in the development of the central as well as the peripheral nervous system.
PAGE 191
CHAPTER VII THE RESPONSE OF CULTURED TRIGEMINAL AND SPINAL CORD MOTONEURONS TO NERVE GROWTH FACTOR Introduction The ability of nerve growth factor (NGF) to stimulate process outgrowth from peripherally-derived sympathetic and developing sensory neurons was recognized very early in its history (see Levi-Montalcini, 1982). Later, it was found that these NGF-responsive neurons have specific NGF receptors which mediate its internalization and retrograde axonal transport (Banerjee et al., 1973; Hendry et al., 1974). More recently it has been demonstrated that neuronal populations in the central nervous system specifically bind and retrogradely transport NGF (Raivich et al., 1985; Raivich et al., 1987; Seiler and Schwab, 1984; and see preceding chapter). The presence of specific binding does not necessarily mean that the cells are biologically responsive to NGF. Two types of NGF receptors, type I (high affinity) and type II (low affinity), have consistently been identified among the different cells which bind NGF (Sutter et al., 1979; Olender and Stach, 1980; Schechter and Bothwell, 1981). The type II receptor can be found alone on certain clonal cell lines and 182
PAGE 192
183 non-neuronal cell types which specifically bind NGF but do not subsequently internalize it or issue a biological response to it (Green et al., 1986; Zimmermann and Sutter, 1983). In contrast, all neuronal cells which have been examined exhibit binding by both of the receptor types (Godfrey and Shooter, 1986). While both receptors bind NGF, it is the type I receptor which has been implicated in mediating NGF internalization, transport and the subsequent cellular response (Bernd and Greene, 1984). The internalization and retrograde transport of NGF seem to be necessary parts of a sequence which results in a biological response to the growth factor (Purves, 1976; Eveleth and Bradshaw, 1987). Sensory neurons retrogradely transport NGF both during development and in the mature animal (Brunso-Bechtold and Hamburger, 1979; Stoeckel et al., 1975). The expressed response to NGF, however, varies. Neural crest-derived sensory neurons of the dorsal root ganglion (DRG) respond to NGF with enhanced survival and process outgrowth during development while DRG cells of the postnatal animal respond with increased substance P levels (Davies and Lindsay, 1984; Kessler and Black, 1981). The nature of the cellular response to NGF is apparently agedependent in a specific cell type. The exact form of the cellular response to NGF is likely to be related to ongoing developmental events of critical importance to the cell.
PAGE 193
184 Motoneurons of the brainstem and spinal cord have previously been demonstrated to bind NGF (Raivich et al., 1985; Raivich et al. 1987) and to specifically transport it during a relatively brief period in early development (spinal cord motoneurons transport NGF from day 4 to day 10; trigeminal motoneurons transport NGF from day 3! to day 8; see preceding chapter). While the motoneuron NGF receptor has yet to be isolated and characterized, the ability of motoneurons to retrogradely transport NGF suggests that there is some biological response to it by these cells. The nature of the response of motoneurons to NGF has yet to be identified. As in other neuronal populations, the particular response to NGF is likely to be related to cell functions important during the period of NGF transport. Motoneurons transport NGF in vivo during the developmental period of axonal outgrowth and elongation and the formation of specific contact with muscle (Chapter VI; Tosney and Landmesser, 1985a, 1985c). The preceding retrograde transport study showed that in the chick, lumbar spinal cord motoneurons and brainstem trigeminal motoneurons transport NGF at the time of initial axonal advance into their respective targets, i.e. the limb (day 4!-5; Tosney and Landmesser, 1985c) or the mandibular process of the first visceral arch (day 3-3!; Heaton et al., 1978). Motoneuron process formation is a highly significant event during this developmental period. Additionally, the ability
PAGE 194
185 of motoneurons to retrogradely transport NGF persists through the period of natural cell death which follows process outgrowth. The possibility that NGF may affect motoneuron cell survival and process outgrowth during early development was examined in the present study. The survival and process outgrowth of dissociates of day 4 trigeminal basal plate and day 5 ventral spinal cord cultured in the presence of NGF was evaluated and is reported here. Methods Cell Preparation Fertile White Leghorn chicken eggs were obtained from the Poultry Science Department of the University of Florida. The eggs were incubated at 37c, 60-70% relative humidity for either 4 days for brainstem tissue or 5 days for spinal cord tissue. At these ages the motoneuron populations are well established and retrogradely transport NGF, but natural cell death has not yet commenced. Using sterile techniques the basal plate in the trigeminal region of the brainstem was dissected free from day 4 chick embryos. The tissue was held in sterile BSS in a tissue culture incubator (37c, 97% air 3% CO2 ) until dissociation. The ventral spinal cord in the lumbar and brachial regions was dissected free from day 5 chick embryos. For dissociation the tissue was incubated in 0.1% trypsin at 37c for 8 minutes then finely minced before ending the trypsinization by transfer to the control
PAGE 195
186 culture medium containing 5% FBS and 0.05% BSA. The tissue was rinsed twice in culture medium then dissociated by gentle trituration. The dissociated cells were plated in 0.5 ml of the appropriate culture medium onto glass coverslips with a prepared substratum of collagen, polyornithine and laminin (see below). The coverslips were affixed with paraffin to the bottom of a 35 mm plastic tissue culture dish which had a 18 mm diameter hole drilled in the bottom. This created a small well within the dish and decreased the surface area available to the cells during the initial plating. The plated cells were allowed to adhere for 2 hr before an additional 1.5 ml of medium was added to the 35 mm culture dish. Cytosine arabinoside (Ara-c; 2 g/ml final concentration) was also added at this time. Ara-c is toxic to rapidly dividing cells (Burry, 1983) and its addition serves two purposes. It acts to virtually eliminate nonneuronal cells and, since motoneurons are generated precociously (trigeminal motoneurons generated by 72-84 hours, Heaton and Moody, 1980; spinal cord motoneurons generated by 96 hours, Hamburger, 1948), it increases the percentage of motoneurons relative to other neuronal populations. The control medium consisted of Ham's F-12 (Gibco), penicillin (100 U/ml), streptomycin (100 g/ml), fungizone (1.75 g/ml), 0.015M glutamine, 0.6% glucose, 0.05% bovine serum albumin (BSA), 5 g/ml insulin, 10 g/ml
PAGE 196
187 transferrin, 30 nM selenium and KCl to bring the potassium ion concentration to 20 mM. The experimental medium was supplemented with 30 ng/ml 2.5S murine nerve growth factor (generously provided by Dr. E.J.Johnson Jr.). Glass coverslips were sonicated in 95% alcohol, then rinsed in distilled water and sterilized by ultraviolet light exposure before the substratum was applied. The coverslips were first coated by exposure to a combined solution of 0.5 mg/ml rat tail collagen (Sigma) and 0.04 mg/ml poly-DL-ornithine (Sigma). They were then dried overnight and stored desiccated in a refrigerator until use. On the day of use the coverslips were briefly exposed to UV light then 100 l of a laminin solution (100 g/ml in distilled water) (Collaborative Research, Inc.) was micropipetted onto the center of the coverslip. The coverslips were incubated at 37c for 1.5 to 2 hours then rinsed 3X with BSS immediately prior to cell plating. Data Collection Each group consisted of four culture dishes: two experimental dishes (spinal cord or brainstem) and a control dish for each tissue type. Each dish received the cell dissociates of either two brainstem basal plates or the ventral spinal cord from the brachia! and lumbar regions of one embryo. The dishes were given an identifying number which was recorded at the time of dissociation. Following the plating period this identifying number was covered with
PAGE 197
188 tape bearing a random number. The dishes were identified by these random numbers during the subsequent observations. After completion of data collection at 48 hours the tape was removed and the code recorded. The dissociates were viewed on a Nikon diaphot inverted phase contrast microscope with an attached air curtain incubator for prolonged observations. Phase-bright cells, characteristic of neurons, were counted immediately following the two hour plating period and then again after 24 and 48 hr in culture. Circumscribed areas for cell counts were demarcated by one of two means: either 50 mesh copper EM grids (Polysciences) were affixed to the bottom of the coverslip or coverslips were used which contained an etched grid pattern (Bellco). Areas were chosen for observation at the initial time point and observations were repeated from these areas at the later time points. The criterion for using an area was that the cells be sufficiently dense with minimal clustering of cell bodies. In each group the density of cells in the two dishes containing the same tissue type always proved to be fairly similar so that differences in the initial cell density could not contribute to the results. The initial cell count was as close to 2000 cells as possible. This could usually be achieved by counting approximately 20 areas of 0.6 mm2 but varied somewhat depending on the cell density. At 24 hours 2 to 6 circumscribed fields used for the cell counts were chosen for assessing neurite outgrowth. All
PAGE 198
189 of the process bearing neurons within these areas were traced using a drawing tube attached to the Nikon inverted microscope. The neurite outgrowth was later quantified using a Bioquant digitizing tablet in concert with an Apple IIe computer. The percentage of cells with processes within each of these areas was calculated in conjunction with the cell counts for these areas. Cell Labeling Two types of cell labeling techniques were applied. Motoneurons were identified in culture by injection of their target musculature with a fluorescent dye which is incorporated by the motoneurons. The fluorescent carbocyanine dye l,l'-dioctadecyl-3,3,3131 -tetramethylindocarbocyanine perchlorate (diI) was loaded into a Hamilton microliter syringe with an attached glass micropipette tip. This is a lipid soluble dye that is incorporated into the cell membrane. It has been demonstrated to be incorporated by motoneuron axons with peripheral injections and is not transferred to neighboring cells or across synapses with the time course used (Honig and Hume, 1986). From 0.05 to 0.10 microliters of diI were injected into the limb buds or the presumptive jaw muscle at day 4! or day 3! respectively. Care was taken to disrupt the embryo as little as possible. Following the injection the eggs were resealed with parafilm and the eggs returned to the incubator. Twelve to sixteen hours later the embryo was
PAGE 199
190 removed from the egg, decapitated immediately and the relevant area of the nervous system dissected free. The tissue was prepared for culture as described above. Cultured cells were incubated with 125r-NGF to assess possible NGF binding. After being in culture for either 24 or 48 hours the medium was removed and the cells were rinsed three times over 2 hours in medium free of both NGF and serum. The cells were then incubated in serum-free medium with 125r-NGF (10 ng/ml, specific activity 21 Ci/g) at 37c for 60 minutes. Non-saturable NGF binding was established in control cultures which contained an excess of unlabeled NGF (5 g/ml) along with the labeled NGF. Following the incubation the cells were rapidly but gently rinsed six times with ice cold O.lM phosphate buffer (pH 7.4) with 0.05% BSA (PBS/BSA). The cells were then fixed in 1% paraformaldehyde/ 2.5% glutaraldehyde in PBS at room temperature for 45 minutes. The cells were then rinsed sequentially in PBS, PBS/BSA, PBS and water. Once dry the coverslips were mounted with Permount, cell side up, on glass slides. Following drying the slides were again rinsed several times in distilled water and allowed to dry completely before dipping. The slides were then dipped in Kodak NTB-2 autoradiography emulsion diluted 1:1 with distilled water with 0.1% liquid detergent added to reduce the formation of bubbles. The dipped slides, secured from exposure to light, were dried on ice for two hours. The
PAGE 200
slides were then boxed with desiccant, sealed, and stored refrigerated from 1-4 weeks before developing. The slides were developed in Kodak D-19. The coverslips were then removed from the slides and mounted cell side down on a clean slide with glycerol. Results Neurite Quantity 191 The quantity of neurite elaborated after 24 and 48 hours in culture is represented by the neurite index. The neurite index is the total quantity of neurite in a given field relative to the number of cells bearing processes in that field. The mean neurite index for each group is presented in Table 7-1. After 24 hours in culture in the presence of NGF, the mean neurite index of ventral spinal cord dissociates was significantly greater than that of the controls (p< 0.0005, t-test, 1-tailed). The mean neurite indices of the trigeminal basal plate dissociates cultured in NGF and control medium, however, did not differ significantly at this time point. After 48 hours in culture both the ventral spinal cord and the trigeminal basal plate dissociates showed slightly greater outgrowth in the NGF than in control conditions but only in the case of the spinal cord was this difference statistically significant (p<0.05). Neurite Initiation It is difficult to assess neurite initiation wholly independent of cell survival. In an effort to reduce the
PAGE 201
(a) NGF CON SIGN (b) NGF CON SIGN Table 7-1: NGF EFFECT ON PROCESS FORMATION Neurite Index TRIGEMINAL SPINAL 24 HOUR 48 HOUR 24 HOUR 50.4 71. 3 76.3 n=32 n=32 n=18 54.2 65.1 63.2 n=32 n=32 n=18 NS NS t .0005 Note: Values are in microns % Neurite Initiation 192 CORD 48 HOUR 96 n=18 82.9 n=l8 t .05 TRIGEMINAL SPINAL CORD 24 HOUR 44.8 n=18 43.7 n=18 NS 48 HOUR 82.9 n=20 67.9 n=20 t.025 24 HOUR 58.4 n=17 64.8 n=18 NS 48 HOUR 89.8 n=l8 83.5 n=18 NS
PAGE 202
193 impact that cell survival would have, neurite initiation was determined by two methods. The number of cells with processes relative to the number of cells surviving in a given field was recorded after 24 and 4 8 hours in culture. The percentage of neurite initiation which results is a ratio which can be altered by a change in either the numerator (the number of cells with processes) or the denominator (the number of cells surviving). The second method used to determine neurite initiation was not dependent on overall cell survival. The number of cells with processes at 48 hours relative to the number of cells with processes at 24 hours in a given field were compared to yield a percentage increase in neurite initiation. Table 7-1 summarizes the mean percentages for neurite initiation as determined by the first method described. After 24 hours in culture neither the trigeminal basal plate nor the ventral spinal cord dissociates evidenced any significant difference in percentage of neurite initiation in NGF compared to control cultures. Similarly, after 48 hours in culture the ventral spinal cord dissociates cultured in NGF and control medium showed no significant differences in neurite initiation as determined by both methods. In contrast, the trigeminal basal plate dissociates showed a significant increase in neurite initiation after 48 hours cultured in the presence of NGF (p<0.025,). Both methods of determining
PAGE 203
neurite initiation resulted in a significant increase for trigeminal basal plate dissociates at 48 hours. Cell Survival 194 Cell survival is assessed by a comparison of the number of cells surviving at a given point in time relative to the number of cells surviving at the time of the culture initiation, i.e. 2 hours. Cell survival, therefore, is expressed as a percentage. Cells counts were recorded at the time of plating and at two subsequent points in time, thus producing data for percent cells surviving at 24 and 48 hours. The mean percentage of cell survival for each group is presented in Table 7-2. The presence of NGF does not have any effect on the survival of either the trigeminal or spinal cord dissociates at either time sampled. Cell Labeling Dissociated cells of the trigeminal basal plate or the ventral spinal cord were incubated with 125r-NGF after 24 or 48 hours in culture. Subsequent processing for autoradiography demonstrated that both the cell body and processes bind NGF at both times (Figure 7-1, 7-2). The binding appeared heavier at 48 hours than at 24 hours, that is, greater binding to both cell bodies and processes was seen. No attempt was made to quantify this difference. Inclusion of unlabeled NGF in the incubation medium along with radiolabeled NGF resulted in no labeling of cells indicating that the binding observed was saturable (Fig.
PAGE 204
195 TABLE 7-2: NGF EFFECT ON CELL SURVIVAL % Cell Survival TRIGEMINAL NGF 24 HOUR 80.6 n=23 CON 80.3 n=23 SIGN NS LEVEL 48 HOUR 43.6 n=18 40.8 n=18 NS SPINAL CORD 24 HOUR 64.8 n=19 62.4 n=19 NS 48 HOUR 29.6 n=16 35.2 n=18 NS
PAGE 205
Figure 7-1: 125I-NGF binding by trigeminal neurons Dissociates of trigeminal basal plate were cultured for 24 (a, b, c, and d) or 48 (e, f) hours in the presence of 30 ng/ml NGF. The presence of specific binding sites for NGF was evidenced by incubation with 125I-NGF. Non-specific NGF binding was evaluated following incubation with l25I-NGF plus an excess of unlabeled NGF. NGF binds specifically to both cell bodies and processes at both culture times although more cells appear to bind NGF after 48 hours in culture. Differential interference contrast (DIC) (a) and dark field (b) photomicrographs of NGF binding to trigeminal neurons grown for 24 hours in culture. DIC (c) and dark field (d) photomicrograph of non-specific NGF binding to trigeminal neurons grown for 24 hours in culture. DIC (e) and dark field (f) photomicrographs of NGF binding to trigeminal neurons grown for 48 hours in culture. (a), (b) X 600; (c), (d) X 650; (e), (f) X 550.
PAGE 206
a ':~ ) :~.? .. ,. ,,.,,,. i ,...,, .. (' .,,.. C ...-<' I ,. ,,. ,." ~ .. ~ ./ ,. ( ,. / I ,. t ( / .,. ,. --' I \ /. ~ _,,.; ;, r,J' I' / r r 197
PAGE 207
198 Figure 7-2: 125I-NGF binding by spinal neurons Cells were cultured for 2 4 hours in the presence of 30 ng/ml NGF then rinsed before incubation with either 125I-NGF (a and b) or 125I-NGF plus an e xces s of unlabeled NGF to display non-specific binding (c and d). DIC (a) and darkfield (b) photomicrographs of specific NGFbinding to ventral spinal cord dissociates. Binding of NGF to both cell bodies and processes is evident (arrows). DIC (c) and darkfield (d) photomicrographs of non-specific biding of NGF to ventral spinal cord dissociates. Arrow points to artifactually labeled debris evident in both fields. (a), (b), (c), (d) X 540.
PAGE 208
Figure 7-3: Identification of motoneurons in culture by retrograde labeling The presence of motoneurons in culture was confirmed by injection of the target musculature with the fluorescent dye diI. Motoneurons incorporate the dye and are thereby fluorescently tagged. The punctate appearance of rhodamine fluorescence in labeled cells is evident in neurons at the center of the field in both trigeminal (b) and spinal (d) dissociates (arrows). Phase contrast (a) and rhodamine fluorescence (b) photomicrographs of metencephalic basal plate dissociates cultured for 48 hours in the presence of 30 ng/ml NGF. Trigeminal motoneurons were labeled at 3 !-4 day of incubation and isolated for culture 12-16 hours later. DIC (c) and rhodamine fluorescence (d) photomicrographs of ventral spinal cord dissociates cultured for 48 hours in control medium. Spinal motoneurons were labeled at 4 !-5 days of incubation and isolated for culture 12-16 hours later. (a), (b), (c), (d) X 686.
PAGE 209
200 a -
PAGE 210
201 7-2, 7-1). Cells cultured either in the presence or absence of NGF prior to the incubation with 125r-NGF both exhibit specific NGF binding. Twelve to sixteen hours following peripheral jaw or limb injections of the fluorescent dye diI, dissociation of the basal plate (trigeminal or spinal cord regions) resulted in the presence of labeled cells in culture, verifying the presence of motoneurons (Figure 7-3). The percentage of cells labeled with diI was low. While this was not quantified, it appeared to be consistent with the labeling which is expected with this methodology (Honig and Hume, 1986). Discussion Trigeminal basal plate and ventral spinal cord dissociates both displayed some enhancement of process formation in the presence of NGF while cell survival was not affected at any time point observed. The direct effect of NGF on neurite growth from both populations of motoneurons during early development adds support to the contention that NGF may play some role in the normal development of motoneurons. Prior to a discussion of the biological significance of the results, however, the nature of the information about cell growth which can be derived from the growth equations should be considered. The neurite index, a measure of neurite quantity, was significantly elevated in spinal cord dissociates after 24
PAGE 211
202 and 48 hours in culture. Similarly, the trigeminal basal plate dissociates showed a trend in this direction after 48 hours in culture although it did not reach levels of statistical significance. The neurite index is a straightforward measurement of the quantity of process elaborated. At the cellular level increases in the neurite index could be subserved by either an increase in (1) the rate of process extension or (2) the number of processes formed. Overall process number is a function of both primary processes initiated from the cell body and secondary processes branching from established neurites. Process formation appeared to occur at a low rate in both the experimental and the control conditions. Process number was not quantified because, at the cell densities used, crossing of adjacent processes obscured an accurate assessment of branching. The trigeminal basal plate dissociates did not respond to NGF with a significant change in the neurite index but did show a significant enhancement of the percentage of cells initiating processes after 48 hours in culture. In order to quantify neurite initiation some assumption must be made about cell survival. Neurite initiation was calculated by two different methods, each relying on a different assumption. The first method divided the number of cells with processes in a given area by the number of surviving cells in that area at that time. This equation assumes that
PAGE 212
203 any experimental influence on cell survival is equal on both sides of the equation, that is, equally affecting cells with processes and those without processes. If this is true then an increase in this ratio can be attributed to the addition of cells initiating a process. This may not be a fair assumption, however. There may be selective support of the population of cells which initiate processes. A second method was therefore introduced to assess neurite initiation at 48 hours. The number of cells with processes at 48 hours was divided by the number of cells with processes at 24 hours in that same area. This method makes an alternate assumption about cell survival, i.e., that any influence of NGF on cell survival would be of constant proportions over the 48 hour time sampled. This is perhaps a more conservative assumption. Both methods of determination of neurite initiation revealed that the presence of NGF resulted in a significant increase in the percentage of trigeminal basal plate dissociates initiating processes after 48 hours in culture. Since the two methods of assessment operate on different assumptions it seems quite likely that selective neuronal survival is not responsible for the observed increase in neurite initiation, although it might play a contributory role. However, without continually monitoring cells over these intervals, it is not possible to separate selective survival of cells with
PAGE 213
processes from an accretion of this population by new additions. 204 The initiation of processes by neurons is a morphological reflection of the cell differentiation process. It is not possible to distinguish whether NGF affects initiation by inducing morphological differentiation or whether it merely acts to potentiate the process of neurite initiation. In other differentiating neuronal populations, NGF is thought to act by a quantitative enhancement of the level of gene expression rather than invoking qualitative changes in expression which alter the neuronal phenotype (Thoenen and Edgar, 1982). The differential influence of NGF on the trigeminal and spinal cord populations is a bit puzzling. These two neuronal populations, while isolated at slightly different incubation times (day 4, trigeminal; day 5, spinal cord), should be approximately equivalent in terms of their developmental stage. Although there was a tendency for NGF to increase the neurite index in both populations, this increase was statistically significant only in the spinal cord cultures. The role of NGF in the enhancement of neurite initiation is less certain. NGF did not have a significant effect on neurite initiation by spinal cord dissociates but did significantly enhance such initiation by trigeminal dissociates after 48 hours in culture. Inherent differences in the individual motoneuron populations may account for the
PAGE 214
205 seemingly discrepant responses to NGF. The control values for both the neurite index and neurite initiation are greater for the spinal cord dissociates than for the comparable trigerninal dissociates. Thus, differences in the endogenous growth rates of the two rnotoneuron populations may affect the sensitivity or the time course of the population response to NGF. Retrograde labeling with the fluorescent dye diI positively confirmed the presence of rnotoneurons in both populations of dissociated cells. While not quantified, only a small proportion of the isolated cells were fluorescently labeled. This is consistent with the rate of labeling previously reported for day 6 chick rnotoneurons following peripheral diI injections (Honig and Hurne, 1986). The isolation and culturing procedures used in the present study, however, should produce a population consisting almost entirely of rnotoneurons. At the time of isolation, the rnotoneuron populations have reached a maximum. Motoneuron differentiation is largely completed while the diminution of the populations by natural cell death has not yet begun. After the cells were plated, Ara-c was added to eliminate those cells still dividing. Since rnotoneurons differentiate precociously, Ara-C effectively excludes most other neurons and non-neuronal cells from the culture. The dissociates consisted almost entirely of rounded phase bright cells which are characteristically neuronal with the
PAGE 215
206 flat non-neuronal cell profile appearing very rarely. The low rate of fluorescent labeling detected probably results from two factors: (1) the limited quantities of dye which were injected likely labeled only a small proportion of motoneurons initially and (2) only those cells which were highly fluorescent would be detectable with the unaided eye. The intention was not to quantify the motoneurons in culture but merely to confirm that motoneurons were present. The cultured cells retained the ability to specifically bind 125r-NGF as revealed by autoradiography. Both trigeminal basal plate and ventral spinal cord dissociates cultured in either NGF or control medium displayed NGF binding. The cells bound NGF on both the cell body and along processes. The relative density of binding has previously been described to be 2-to 5-times greater on the cell body of cultured neurons (sympathetic, Claude et al., 1982; sensory, Carbonetto and Stach, 1982). The density of binding appeared to be somewhat greater at 48 than at 24 hours but this difference was not quantified. The NGF-binding did not differentiate between high and low affinity receptors but the presence of NGF transport in vivo (Chapter VI) suggests that both receptors are probably present. The observed enhancement of process formation independent of any effect on cell survival is a common observation. Embryonic motoneuron dissociates of both chick and rat respond to distinct tissue-derived factors
PAGE 216
207 selectively with either cell survival or enhanced neurite outgrowth (chick: Calaf and Reichardt, 1984; Dohrmann et al., 1986; Eagleson and Bennett, 1986; rat: Smith et al., 1986). Similarly, in chick parasympathetic ciliary ganglion neurons and in dissociated embryonic Xenopus neural tube cells, NGF has been shown to stimulate process formation with no effect on cell survival (Collins and Dawson, 1983; Katz, 1986). These observations are quite consistent with those in the present study. There is considerable accumulated evidence suggesting that NGF may be physiologically significant to motoneuron growth during early development. Motoneurons of the brainstem and spinal cord both bind and retrogradely transport NGF during a transient period early in development (Raivich et al., 1985, 1987; Chapter VI). In one case, motoneurons have demonstrated a response to exogenous NGF: Levi-Montalcini and Aloe (1985) noted the hypertrophy of spinal cord motoneuron cell bodies following daily NGF injections into the periorbital area of Xenopus tadpoles. In vitro, embryonic Xenopus neural tube dissociates respond to NGF with increased neurite initiation and branching and an increase in the rate of neurite elongation as indicated by an increase in the number of elongation steps taken by an axon (Katz, 1986). An entirely different approach has raised the possibility that a protein specifically expressed in motoneurons may be induced by exposure to NGF: In this
PAGE 217
208 study, Leonard et al. (1986) labeled adult rat brain with anti-sense probes against mRNA which was induced in PC12 pheochromocytoma cells in response to NGF. One of these probes bound specifically to motoneuron populations of the brainstem and spinal cord of the adult rat. The dorect effect of NGF on process outgrowth from early chick motoneurons observed in the current study contributes to the evidence suggesting that NGF affects motoneuron growth during early development. Previous studies which failed to note any effect of NGF on process outgrowth from dissociates of embryonic spinal cord differed from the present study in a variety of ways. Doherty et al. (1984), for example, used an antibody to neurofilament protein to quantitatively evaluate the differentiation of a mixed population of cells, dissociates of embryonic chick spinal cord (day 7), in the presence of NGF. The cell population included whole spinal cord dissociates with no mitotic inhibitor added to exclude other neuronal and non-neuronal populations. Trophic interactions between cell populations could have made detection of an NGF effect more difficult. In addition, the method used for quantifying the neuronal response in this study did not separate an effect on process formation from an effect on cell survival. Thus, variability in cell survival, the response of the different cell populations present, and the interactions between the cell populations may all have
PAGE 218
209 served to obscure a difference in a single aspect of growth occurring in only one cell type. A more specific study of cell growth using purified populations of embryonic chick motoneurons also failed to find a growth effect produced by NGF. In this investigation, Gundersen and Park (1984) found that NGF did not affect the elongation rate, directed growth and substrate adherence of purified populations of day 6 embryonic chick spinal cord motoneurons. The lack of effect on directed growth and substrate adherence reported by Gundersen and Park (1984) are not necessarily inconsistent with the reported enhancement of neurite quantity in the presence of NGF shown in the present investigation. It is quite possible that an effect on the growth rate was not evident with the sampling interval used (Katz, 1986). In addition, unlike the present study, laminin was not a component of the substratum used. A laminin substratum might be necessary for NGF to affect outgrowth from avian motoneurons. In vivo, laminin is distributed along the pathway of motoneuron axon outgrowth in both the brainstem and spinal cord (Riggott and Moody, 1987; Rogers et al., 1986). Laminin has been shown to potentiate the NGF-induced survival of embryonic sympathetic neurons (Edgar, 1985). While laminin itself has been shown to stimulate neurite outgrowth from both peripheral and central neuronal populations (Rogers et al., 1983), it may
PAGE 219
210 also play a role in mediating neurite outgrowth in response to NGF. NGF does not appear to affect the adhesivity of motoneuron neurites to the culture substratum. In contrast to dorsal root ganglion cells, dissociates of embryonic chick motoneurons cultured on a collagen and poly-L-lysine substratum showed no change in growth cone adhesivity in the presence of NGF (Gundersen and Park, 1984). Axonal growth can be mathematically fractionated into distinct components which are (1) the average rate of growth steps, (2) the average step size, and (3) the probability of taking a successful step (Katz et al., 1984). The last of these components, the probability of taking a successful step, is likely dependent on adhesion to the substratum. Embryonic Xenopus neural tube dissociates cultured in the presence of NGF showed no change in the probability of making a successful step but did show an increase in the average rate of growth steps (Katz, 1986). The rate at which the growth cone takes growth steps is thought to be dependent on the rate of movement of the core cytoskeleton (Katz et al., 1984). The type I receptor for NGF which has been consistently linked to the biological response to the growth factor appears to be directly connected to the cytoskeletal core of the neurite (Schecter and Bothwell, 1981; Vale and Shooter, 1982). NGF binding to the type I receptor may be responsible for enhancing process
PAGE 220
formation in developing motoneurons through these direct cytoskeletal attachments. 211 In summary: motoneurons of the trigeminal region of the brainstem basal plate and ventral spinal cord isolated early in development respond to NGF with enhanced process formation, but no effect on cell survival is seen. The dissociation of an influence on neurite outgrowth from an influence on cell survival is characteristic of previous observations on tissue-derived trophic factors and NGF alike (Calof and Reichardt, 1984; Dohrmann et al., 1986; Collins and Dawson, 1983; Katz, 1986). That NGF may influence motoneuron development has been suggested by the presence of specific binding of NGF and the specific retrograde transport of NGF by motoneurons transiently during development (Raivich et al., 1985, 1987; Chapter VI). The present study provides the first evidence of a direct effect of NGF on process formation in embryonic chick motoneurons.
PAGE 221
CHAPTER VIII GENERAL DISCUSSION Nerve Development The neuron is unique in that its axonal process extends a considerable distance from the cell body to make specific contact with a target cell. During development the growing process follows a highly consistent path to reach its target. Two broad spectrum questions immediately arise. What are the factors which produce the consistently patterned path of nerve growth? What are the embryonic features which insure that specific contacts are established between the neuron and its target? Past and current research are working both to focus the questions and to provide specific answers at the cellular level. The research in this dissertation has employed both in vivo and in vitro techniques to address some aspects of the above questions in motoneuron systems of the brainstem and spinal cord. Motoneuron Systems The fundamental organization of the somatic motor systems of the brainstem and spinal cord are the same. Throughout the nervous system motoneurons arise precociously during development. While the cell bodies reside in the basal plate of the central nervous system (CNS), the nerve 212
PAGE 222
213 exits the CNS immediately and grows through a peripheral environment to innervate somatic musculature. Motoneuron systems are particularly well suited for studies of pattern and specificity in nerve formation. The early differentiation of motoneuron populations means that nerve outgrowth occurs in a relatively simple environment. In addition, both the neuron and its target musculature exist in clearly circumscribed populations. Thus, this system is especially useful for investigating the factors which establish specific neural pathways. Organization in the Central Nervous System (CNS) Motoneuron cell bodies within the CNS are grouped in populations which can be identified by the specific muscle innervated. Small injections of individual muscles in the chick with the retrograde tracer horseradish peroxidase (HRP) have resulted in the labeling of discrete groups of motoneurons in the lumbosacral spinal cord and brainstem (Landmesser, 1978a; Chapter II). The distribution of the motoneurons in the brainstem oculomotor nucleus was fully characterized in the first experiment (Chapter II). In the oculomotor complex, the muscle-specific grouping of the motoneuron cell bodies is reflected in three morphologically distinct subnuclei. Additionally, a circumscribed subpopulation of cells within the ventral subnucleus migrates across the midline.
PAGE 223
214 The arrangement of cells into the distinct subnuclei and the migration of a specific subpopulation across the midline occurs subsequent to nerve outgrowth and the initial neural contact with muscle. Is it possible, then, that the contact with muscle triggers or directs the subsequent segregation of the motoneuron cell bodies which occurs? This proposition is not too dissimilar from the resonance theory which suggested that the pattern of central connections was specified by peripheral stimuli, i.e. the pattern of peripheral connections (Weiss, 1936). The experimental test of this question required the removal of the presumptive target muscles before neural contact was made. In the oculomotor system, the cellular organization as represented by the morphologically distinct subnuclei and the migration of cells across the midline arose normally in the absence of the appropriate target musculature. Other studies have shown that in the chick limb the axonal projection pattern is initially correct so that the specificity is not achieved by a selective removal of inappropriate connections (Landmesser, 1978b). Together, these results suggest that the neural population which innervates a specific muscle has some identifying characteristics prior to contact with that muscle. The surface features which identify cells to specific groups in the CNS may also identify fibers to the appropriate path in the periphery. Cell surface molecules
PAGE 224
215 are likely candidates as specific recognition signals between cells. The adhesion molecule N-CAM which exhibits homophilic binding produces selective adhesion between cells displaying N-CAM on their surface (Rutishauser et al., 1978a). Some evidence suggests that specific adhesive interactions between cells of the dorsal root ganglion and spinal cord may be mediated by the specific expression of cell surface glycoconjugates and complimentary carbohydratebinding proteins (lectins) (Dodd and Jessell, 1986; Regan et al., 1986). While immunological techniques have made it possible to label specific cell surface molecules and describe selective distributions it is difficult to unequivocally demonstrate a specific role for these molecules in developmental events. The complexity of the system and the possibility that multiple molecular markers subserve a given interaction may be confounding efforts to identify the causative agent with certainty. Proximal Nerve Path The initial path of nerve growth represented by the course of the nerve as it first exits the brainstem is determined independent of the target muscles in the oculomotor system. This was particularly evident in two cases in which, irrespective of the presence of residual target muscle on the operated side, the nerve crossed to innervate muscle on the contralateral side. This was a rare occurrence in which cues that determine the initial course
PAGE 225
216 of nerve outgrowth were apparently disrupted independent of the target disruption. As an interesting note, when the nerve crossed to the contralateral side it did not fuse with the opposing nerve but followed a separate but convergent path to the target. The pathway does not appear to be immutably fixed to one place but each nerve, it seems, creates its own path. This is consistent with a recent observation on nerve development in the chick forelimb (Noakes and Bennett, 1987). Schwann cell precursors immediately precede the growing front of motor fibers along the entire path of nerve growth. The presumptive Schwann cells associated with each nerve seem to advance slightly in front of the nerve and may provide the substratum for nerve growth. Experimental intervention by removal of the Schwann cell precursors disrupts nerve growth into the limbs in the chick embryo (Carpenter and Hollyday, 1986). Schwann cells have been shown to produce several molecules which could be important in creating a pathway for growth, namely, laminin (Cornbrooks et al., 1983; Palm and Furcht, 1983), plasminogen activator (Krystosek and Seeds, 1984; AlvarezBuylla and Valinsky, 1985) and NGF (Assouline et al., 1985; Bandtlow et al., 1987). A laminin substratum has been shown to enhance process formation by both central and peripheral neurons in culture (Rogers et al., 1983). In the chick embryo spinal cord and brainstem the timing and pattern of the appearance of
PAGE 226
217 laminin in the periphery is consistent with a role as a substratum for nerve growth during development (Rogers et al., 1986; Riggott and Moody, 1987). The release of the enzyme plasminogen activator is likely to result in a focal production of the proteolytic enzyme plasmin. It is speculated that the plasmin generating system plays a role in morphogenetic events in development perhaps by degrading components of the extracellular matrix or by selective effects on matrix-or membrane-associated molecules (Monard, 1985; Hawkins and Seeds, 1986). In addition, during development the Schwann cell may be an important source of NGF with the relevance of this fact to motoneuron development only recently becoming apparent. Not only do embryonic chick motoneurons specifically bind NGF as shown by Raivich et al. (1985, 1987) but we have demonstrated that motoneurons of the brainstem and spinal cord retrogradely transport NGF during early development (Chapter VI). Retrograde transport of NGF has been consistently linked to a subsequent biological response. It is possible, in fact, that the NGF-NGF receptor complex acts as the intracellular message which initiates the cellular response (Stach and Perez-Polo, 1987). Thus, this specific retrograde transport of NGF provides strong support for the premise that NGF binding to motoneurons is functionally significant for the cell.
PAGE 227
218 The motoneurons were observed to transport NGF very early in development. Axons of trigeminal motoneurons have invaded their target region, the mandibular arch, by day 3! of incubation and were labeled following 125I-NGF injections at this time. In the chick spinal cord lumbosacral motoneurons were faintly labeled after injections at day 4 while there was heavy labeling in the lateral motor column after injections at day 4!. It is not clear whether this constitutes a developmental change in the ability of the motor axons to transport NGF or a difference in the accessability of the label to the motor axons at the two times. At day 4 motoneuron axons have gathered in the plexus region but have not yet entered the limb bud. Labeling at day 4 may depend on a few early entering axons present in the limb bud or, alternatively, the diffusion of injected NGF to the waiting axons at the plexus region. It is interesting that although the first axons arrive at the plexus region at day 3!, there is a waiting period of almost 24 hours before the first growth cones leave the plexus region and enter the limb proper (Tosney and Landmesser, 1985c). It is possible that a developmental change in the sensitivity to NGF triggers the re-initiation of growth and the entry of the motor fibers into the limb. The onset of NGF binding and transport by motoneurons may still need some clarification. Interactions at target muscle
PAGE 228
219 The muscle-independent nature of nerve growth proximal to the neural tube is contrasted by the muscle-influenced course of nerve growth in the target region. Following removal of the presumptive extra-ocular muscles, the distal nerve path was always directed toward any residual muscle cells in the periphery. These results suggest that, in vivo, the trophic influence of muscle on nerve is exerted at the local level. In cases where little muscle remained, once in the periphery, the nerve followed a path some distance caudally to reach any available muscle cells. The limitations of the muscle-derived influence may not be imposed by distance alone but by regional differences in competitive or restrictive factors in the peripheral environment. In vivo muscle does seem to be able to attract the ingrowth of nerve as demonstrated by the nerve path following removal of the oculomotor target muscles. Other observations have suggested that this influence can be quite specific. In the chick limb, axons which had been experimentally misdirected into an inappropriate nerve occasionally broke away from the tract of nerve growth and created a new path to reach the normally appropriate target muscle (Lance-Jones and Landmesser, 1981b). One aspect of the specificity in the effects of muscle on nerve growth was investigated in vitro using the brainstem trigeminal system. The presumptive target muscle in this system, the mandibular process of the first visceral
PAGE 229
220 arch, is easily identified and isolated thereby being well suited for in vitro study. Explants of the trigeminal motoneurons isolated prior to nerve outgrowth in vivo, exhibited enhanced neurite growth when cultured in medium conditi0ned by appropriate target tissue (presumptive jaw muscle) versus inappropriate target tissue (presumptive limb muscle). Muscle-produced factors are apparently regionally specific having maximal effect on neural tissue which normally innervates it in vivo. It is possible that this specificity is imparted to the target tissue by differential origin during development. In the cranial region muscle tissue is derived from lateral plate mesoderm while trunk muscle arises from somitic myoepithelial cells. Regionally specific growth promoting effects which may be quite important in normal development have only begun to receive adequate consideration experimentally. The neurite promoting activity present in conditioned media (CM) produced by a variety of cell types has been precipitated.by antibodies to laminin (Lander et al., 1985). Although laminin-stimulated neurite outgrowth is effectively blocked by these antibodies, the CM derived neurite promoting activity is not affected by antibodies to laminin. It is probable that differences in the laminin molecule or its configuration in these tissue derived growth promoting complexes protect the activity-dependent site while the epitope for antibody binding has remained available. The
PAGE 230
221 molecular basis of the regionally-specific neurite-promoting effect of muscle has yet to be characterized. Evidence has been presented, however, that a well characterized trophic molecule, NGF, may play a role in the developmental interaction between motoneurons and the target muscle. We have demonstrated that embryonic motoneurons of the brainstem and spinal cord retrogradely transported NGF injected at the target site and that this ability to transport NGF was restricted to an early period of development. In addition, previous work has shown that target muscle apparently provides a source of NGF during development as evidenced by the presence of mRNA sequences homologous to cDNA for NGF in embryonic chick muscle (Hulst and Bennett, 1983). The availability of NGF and its specific transport by motoneurons support the premise that NGF may play a functionally significant role in motoneuron development. Direct evidence of a cellular effect was obtained by isolating the elements in vitro. Dissociates of trigeminal and spinal motoneurons have revealed that NGF can affect process outgrowth. The initiation of neurites from trigeminal motoneuron dissociates was enhanced after culture for 48 hours in the presence of NGF. Process outgrowth from spinal cord motoneurons was enhanced in a somewhat different manner. The rate of initiation did not differ from the control but the quantity of process formed was greater after
PAGE 231
222 both 24 and 48 hours in culture. The difference in the nature of the response of the two motoneuron populations to NGF may be attributable to differences in the endogenous growth rates or the in vivo growth requirements of the two systems, i.e. enhanced fasciculation versus enhanced elongation. It is worth noting that in the culture experiments which demonstrated an influence of NGF on motoneurons, a laminin substratum was used. This may be an important factor in the response. Laminin exists in the periphery at the time of nerve outgrowth in vivo (Rogers et al., 1986; Riggott and Moody, 1987). Schwann cells which have been found to produce both laminin and NGF (Cornbrooks et al., 1983; Bandtlow et al., 1987) apparently precede the front of the growing nerve in the chick forelimb (Noakes and Bennett, 1987). A substratum of laminin and NGF provided to the growing motor axons is quite possibly critical for normal outgrowth. NGF may, additionally, be an important component of the substratum which supports peripheral nerve regeneration. Evidence has shown that during peripheral nerve regeneration in the rat, Schwann cells exhibit a dramatic increase in the low affinity NGF receptor which binds NGF specifically but does not internalize it (Taniuchi et al., 1986). Summary Within the central nervous system oculomotor complex, neuronal cell bodies assumed a distinct organization which
PAGE 232
223 was independent of contact with the target muscle. It is likely that cell surface molecules confer the cellular identity which enables cell bodies to selectively aggregate and enables the axons of these cells to contact the appropriate muscle groups. Muscle did not affect the initial path of nerve growth but did influence nerve growth locally in vivo. Skeletal muscle-produced factors were regionally-specific as evidenced by their differential abilities to enhance trigeminal motoneuron neurite growth in vitro. The specific effects of jaw muscle conditioned medium on trigeminal basal plate explants is the first evidence that there is such selectivity in the ability of muscle to promote motoneuron neurite growth. The active molecular complex, while not yet isolated, is likely to include laminin in association with other components which may be the determinants of the cellular specificity. Quite strikingly, motoneurons of the brainstem and spinal cord were shown to specifically transport NGF during a transient period of early development. Tissue culture provided further evidence of the significance of NGF to motoneuron development by demonstrating a specific enhancement of motoneuron neurite formation in the presence of NGF. In vivo, there seem to be multiple sources of NGF available to the growing axon. Both the Schwann cell, which precedes axonal ingrowth, and the target muscle are likely
PAGE 233
224 to be providing NGF at this time (Noakes and Bennett, 1987; Bandtlow et el., 1987; Hulst and Bennett, 1987). These results provide an important contribution in furthering the understanding of the factors which shape peripheral nerve development.
PAGE 234
REFERENCES Adler, R., M. Manthorpe, S.D. Skaper and S. Varon (1981) Poly-ornithine-attached neurite-promoting factors (PNPFs). Culture sources and responsive neurons. Brain Res. 206: 129-144. Adler, R. ands. Varon (1980) Cholinergic neuronotrophic factors V. Segregation of survival-and neuritepromoting activities in heart conditioned medium. Brain Res. 188: 437-448. Alvarez-Buylla, A. and J.E. Valinsky (1985) Production of plasminogen activator in cultures of superior cervical ganglia and isolated Schwann cells. Proc. Natl. Acad. Sci. USA 82: 3519-3523. Ambron, R.T. (1982) Differences in the distribution of specific glycoproteins among the regions of a single identified neuron. Brain Res. 239: 489-505. Ariens-Kappers, C.U. (1917) Further contributions on neurobiotaxis. IX. An attempt to compare the phenomena of neurobiotaxis with other phenomena of taxis and tropism. The dynamic polarization of the neurone.~ Comp. Neurol. 27: 261-298. Ariens-Kappers, C.U. (1921) On structural laws in the nervous system: The principle of neurobiotaxis. Brain 44: 125-149. Assouline, J., E.P. Bosch, R. Lim, R. Jensen and N.J. Pantazis (1985) Detection of a neurite promoting factor with some simialrity to nerve growth factor in conditioned media from rat primary Schwann cells and astroglia. Soc. Neurosci. Abst. 11: 933. Attardi, D.G. and R.W. Sperry (1960) Central routes taken by regenerating optic fibers. Physiologist 3: 12. Attardi, D.G. and R.W. Sperry (1963) Preferential selection of central pathways by regenerating optic fibers. Exp. Neurol. 7: 46-64. 225
PAGE 235
226 Augustine, J.R., E.G. Deschamps and J.G. Ferguson, Jr. (1981) Functional organization of the oculomotor nucleus in the baboon. Am.~ Anat. 161: 393-403. Ayer-LeLievre, c.s., T. Ebendal, L. Olson and A. Sieger (1983) Localization of nerve growth factor-like immunoreactivity in rat nervous tissue. Med. Biol. 61, 296-304. Balak, K., M. Jacobson, J. Sunshine and U. Rutishauser (1987) Neural cell adhesion molecule expression in Xenopus embryos. Dev. Biol. 119: 540-550. Bandtlow, C.E., R. Heumann, M.E. Schwab and H. Thoenen (1987) Cellular localization of nerve growth factor synthesis by in situ hybridization. EMBO ~. 6: 891-899. Banerjee, S.P., S.H.Snyder, P. Cuatrecasas and L.A. Greene (1973) Binding of nerve growth factor receptor in sympathetic ganglia. Proc. Natl. Acad. Sci. USA 70: 2519-2523. Bastiani, M.J., S. du Lac, and C.S. Goodman (1986) Guidance of neuronal growth cones in the grasshopper embryo. I. Recognition of a specific axonal pathway by the pee neuron.~Neurosci. 6: 3518-3531. Bastiani, M.J., A.L. Harrelson, P.M. Snow, and c.s. Goodman (1987) Expression of fasciclin I and II glycoproteins on subsets of axon pahtways during neuronal development in the grasshopper. Cell 48: 745-755. Belew, M. and T. Ebendal (1986) Chick embryo nerve growth factor. Fractionation and biological activity. Exp. Cell Res. 167: 550-558. Bernd, P. and L.A. Greene (1983) Electron microscopic radioautographic localization of iodinated nerve growth factor bound to and internalized by PC12 cells.~. Neurosci. 3: 631-643. Bernd, P. and L.A. Greene (1984) Association of 125I-nerve growth factor with PC12 pheochromocytoma cells. Evidence for internalization via high-affinity receptors only and for long-term regulation by nerve growth factor of both high-and low-affinity receptors.~Biol. Chem. 259: 15509-15516.
PAGE 236
227 Bixby, J.L., R.S. Pratt, F. Lilien, and L.F. Reichardt (1987) Neurite outgrowth in muscle cell surfaces involves extracellular matrix receptors as well as ca2+_ dependent and -independent cell adhesion molecules. Proc. Natl. Acad. Sci. USA 84: 2555-2559. Blanks, J.C. and L.V. Johnson {1983) Selective lectin binding of the developing mouse retina.~Comp. Neurol. 221: 31-41. Bok, S.T. {1915) Stimulogenous fibrillation as the cause of the structure of the nervous system. Psychiatr. Neurol. Bl. Arnst. 19: 393-408. Borg, J., B. Spitz, G. Hamel, and J. Mark (1985) Selective culture of neurons from rat cerebral cortex: Morphological characterization, glutamate uptake and related enzymes during maturation in various culture media. Devtl. Brain Res. 18: 37-49. Bothwell, M.A. and E.M. Shooter (1977) Dissociation equilibrium constant of B nerve growth factor.~Biol. Chem. 252: 8532-8536. Bottenstein, J.E. and G.H. Sato (1979) Growth of a rat neuroblastoma cell line in serum-free supplemented medium. Proc. Natl. Acad. Sci. USA 76, 514-517. Bottenstein, J.E., S.D. Skaper, S.S. Varon, and G.H. Sato {1980) Selective survival of neurons from chick embryo sensory ganglionic dissociates utilizing serum-free supplemented medium. Exp. Cell Res. 125: 183-190. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248-254. Bray, D. {1970) Surface movements during the growth of single explanted neurons. Proc. Natl. Acad. Sci. USA 65: 905-910. Bray, D., P. Wood, and R.P. Bunge (1980) Selective fasciculation of nerve fibers in culture. Exp. Cell Res. 130: 241-250. Brunso-Bechtold, J.K. and V. Hamburger {1979) Retrograde transport of nerve growth factor in chicken embryo. Proc. Natl. Acad. Sci. USA 76: 1494-1496.
PAGE 237
228 Bueker, E.D. (1948) Implantation of tumors in the hind limb field of the embryonic chick and developmental response of the lumbosacral nervous system. Anat. Rec. 102: 369-390. Burry, R.W. (1983) Antimitotic drugs that enhance neuronal survival in olfactory bulb cell cultures. Brain Res. 261: 261-275. Buskirk, D.R., J.-P. Thiery, U. Rutishauser, and G.M. Edelman (1980) Antibodies to a neural cell adhesion molecule disrupts histogenesis in cultured chick retinae. Nature 285: 486-489. Calof, A.L. and L.F. Reichardt (1984) Motoneurons purified by cell sorting respond to two distinct activities in myotube-conditioned medium. Dev. Biol. 106: 194-210. Carbonetto, s. and R.W. Stach (1982) Localization of nerve growth factor bound to neurons growing nerve fibers in culture. Dev. Brain Res. 3: 463-473. Carney, P.R. and J. Silver (1983) Studies on cell migration and axon guidance in the developing distal auditory system of the mouse.~Comp. Neurol. 215: 359-369. Carpenter, E.M. and M. Hollyday (1986) Defective innervation of chick limbs in the absence of presumptive Schwann cells. Soc. Neurosci. Abst. 12: 1210. Chao, M.V., M.A. Bothwell, A.H. Ross, H. Koprowski, A.A. Lanahan, C.R. Buck and A. Sehgal (1986) Gene transfer and molecular cloning of the human NGF receptor. Science 232: 518-521. Chiu, A.Y., W.D. Matthew and P.H. Patterson (1986) A monoclonal antibody that blocks the activity of a neurite regeneration-promoting gactor: Studies on the binding site and its localization in vivo.~Cell Biol. 103: 1383-1398. Claude, P., E. Hawrot, D.A. Dunis and R.B. Campenot (1982) Binding, internalization, and retrograde transport of 125r-nerve growth factor in cultured rat sympathetic neurons.~Neurosci. 2: 431-442. Cohen, J. and S.Y. Selvendran (1981) A neuronal cell-surface antigen is found in the central nervous system but not in peripheral neurones. Nature 291: 421-423.
PAGE 238
229 Collins, F. (1978a) Induction of neurite outgrowth by a conditioned-medium factor bound to the culture substratum. Proc. Natl. Acad. Sci. USA 75: 5210-5213. Collins, F. (1978b) Axon initiation by ciliary neurons in culture. Dev. Biol. 65: 50-57. Collins, F. (1980) Neurite outgrowth induced by the substrate associated material from nonneuronal cells. Dev. Biol. 79: 247-252. Collins, F. and A. Dawson (1982) Conditioned medium increases the rate of neurite elongation: separation of this activity from the substratum-bound inducer of neurite outgrowth.~Neurosci. 2: 1005-1010. Collins, F. and A. Dawson (1983) An effect of nerve growth factor on parasympathetic neurite outgrowth. Proc. Natl. Acad. Sci. USA 80: 2091-2094. Collins, F. and J.E. Garrett, Jr. (1980) Elongating nerve fibers are guided by a pathway of material released from embryonic non-neuronal cells. Proc. Natl. Acad. Sci. USA 77: 6226-6228. Compito, G.A., B.J. Wagner, and R.W. Stach (1986) Disulfide bong formation between nerve growth factor and the nerve growth factor receptor from embryonic sensory neurons. ~Neurochem. 47: 554-562. Constantine-Paton~ M. (1983) Trajectories of axons in ectopic VIIItn nerves. Dev. Biol. 97: 239-244. Cornbrooks, C.J., D.J. Carey, J.A. McDonald, R. Timpl and R.P. Bunge (1983) In vivo and in vitro observations on laminin production by Schwann cells. Proc. Natl. Acad. Sci., USA 80: 3850-3854. Davies, A.M., C. Bandtlow, R. Heumann, S. Korsching, H. Rohrer, and H. Thoenen (1987) Timing and site of nerve growth factor synthesis in developing skin in relation to innervation and expression of the receptor. Nature 326: 353-358. Davies, A.M. and R.M. Lindsay (1984) Neural crest-derived spinal and cranial sensory neurones are equally sensitive to NGF but differ in their response to tissue extracts. Devlt. Brain Res. 14: 121-127.
PAGE 239
Dodd, J. and T.M. Jessell (1986) Cell surface glycoconjugates and carbohydrate-binding proteins: Possible recognition signals in sensory neurone development.~. Exp. Biol. 124: 225-238. Doherty, P., J.G. Dickson, T.P. Flanigan, and F.S. Walsh (1984) The effect of nerve growth factor and its antibodies on neurofilament protein expression in primary cultures of sensory and spinal neurons. Neurosci. Lett. 51: 55-60. 230 Dohrmann, U., D. Edgar, M. Sendtner and H. Thoenen (1986) Muscle-derived factors that support survival and promote fiber outgrowth from embryonic chick spinal motor neurons in culture. Dev. Biol. 118: 209-221. Dribin, L. and J.N. Barrett (1980) Conditioned medium enhances neuritic outgrowth from rat spinal cord explants. Dev. Biol. 74: 184-195. Dribin, L. and J.N. Barrett (1982) Characterization of neuritic outgrowth-promoting activity of conditioned medium on spinal cord explants. Dev. Brain Res. 4: 435-442. Dubbledam, J.L. and H.J. Karten (1978) The trigeminal system in the pigeon (Columbia livia) I. Projection of the Gasserian ganglion.~Comp. Neural. 180: 661-678. du Lac, S., M.J. Bastiani, and C.S. Goodman (1986) Guidance of neuronal growth cones in the grasshopper embryo. II. Recognition of a specific axonal pathway by the ace neuron.~Neurosci. 6: 3532-3541. Dumas, M., M.E. Schwab, and H. Thoenen (1979) Retrograde axonal transport of specific macromoleucles as a tool for characterizing nerve terminal membranes.~ Neurobiol. 10: 179-197. Eagleson, K.L. and M.R. Bennett (1986) Motoneurone survival requirements during development: The change from immature astrocyte dependence to myotube dependence. Dev. Brain Res. 29: 161-172. Ebendal, T. (1976) The relative roles of contact inhibition and contact guidance in orientation of axons extending on aligned collagen fibrils in vitro. Exp. Cell Res. 98: 159-169. Ebendal, T. and c.-0. Jacobson (1977) Tissue explants affecting extension and orientation of axons in cultured chick embryo ganglia. Exp. Cell Res. 105: 379-387.
PAGE 240
231 Ebendal, T., D. Larhammar, and H. Persson (1986) Structure and expression of the chicken B nerve growth factor gene. EMBO ~5: 1483-1487. Edelman, G.M. (1983) Cell adhesion molecules. Science 219: 450-457. Edelman, G.M. and c.-M. Chuong (1982) Embryonic to adult conversion of neural cell adhesion moleucles in normal and staggerer mice. Proc. Natl. Acad. Sci. USA 79: 7036-7040. Edgar, H. (1985) Nerve growth factors and molecules of the extracellular matrix in neuronal development.~Cell Sci. Suppl. 3: 107-113. Edgar, D. and H. Thoenen {1982) Modulation of NGF-induced survival of chick sympathetic neurons by contact with a conditioned medium factor bound to the culture substrate. Dev. Brain Res. 5: 89-92. Fallon, J. (1985) Preferential outgrowth of central nervous system neurites on astrocytes and Schwann cells as compared with nonglial cells in vitro.~ Cell Biol. 100: 198-207. Fallon, J.R. and M.C. Raff (1982) Specificity and character of neurite outgrowth on defined cell types. Soc. Neurosci. Abst. 8: 302. Ferguson, B.A. {1983) Development of motor innervation of the chick following dorsal-ventral limb bud rotations. ~Neurosci. 3(9): 1760-1772. Fontaine, B., A. Klarsfeld, T. Hockfelt and J.-P. Changeux (1986) Calcitonin gene-related peptide, a peptide present in spinal cord motoneurons, increases the number of acetylcholine receptors in primary cultures of chick embryo myotubes. Neurosci. Lett. 71: 59-65. Gacek, R.R. {1974) Localization of neurons supplying the extraocular muscles in the kitten using horseradish peroxidase.~ Neural. 44: 381-403. Gallo, V., M.T. Ciotti, F. Aloisi, and G. Levi (1986) Developmental features of rat cerebellar neural cells cultured in a chemically defined medium.~Neurosci. Res. 15: 289-301.
PAGE 241
232 Gasser, R.F. and A.G. Hendrickx (1968) The development of the trigeminal nerve in Baboon embryos (Papio sp.). ~ Comp. Neural. 136: 159-182. Gaze, R.M. and M.J. Keating (1972) The visual system and 'neuronal specificity'. Nature 237: 375-378. Giorgi, P.P. and H. van der Loos (1978) Axons from eyes grafted in Xenopus can grow into the spinal cord and reach the optic tectum. Nature 275: 746-748. Gnahn, H., F. Hefti, R. Heumann, M.E. Schwab, and H. Thoenen (1983) NGF-mediated increase of choline acetyltransferase (ChAT) in the neonatal rat forebrain: Evidence for a physiological role of NGF in the brain? Devtl. Brain Res. 9: 45-52. Godfrey, E.W. and E.M. Shooter (1986) Nerve growth factor receptors on chick embryo sympathetic ganglion cells: Binding characteristics and development.~Neurosci. 6: 2543-2550. Grainger, F. and D.W. James (1970) Association of glial cells with the terminal parts of neurite bundles extending from chich spinal cord in vitro. z. Zellforsch. 108: 93-105. Green~ S.H. and L.A. Greene (1986) A single Mr 103,000 1~5I-B-nerve growth factor-affinity-labeled species represents both the low and high affinity forms of the nerve growth factor receptor.~Biol. Chem. 261: 15316-15326. Green, S.H., R.E. Rydel, J.L. Connolly and L.A. Greene (1986) PC12 cell mutants that possess low-but not highaffinity nerve growth factor receptors neither respond to nor internalize nerve growth factor.~Cell Biol. 102: 830-843. Greene, L.A., D.E. Burstein and M.M. Black (1980) The priming model for the mechanism of action of nerve growth factor: Evidence derived from clonal PC12 pheochromocytoma cells. In: Tissue Culture in Neurobiology. E. Giacobini, A. Vernadakis and A. Shahar eds., pp. 313-320, Raven Press, New York. Grumet, M., U. Rutishauser, and G.M. Edelman (1982) Neural cell adhesion molecule is on embryonic muscle cells and mediates adhesion to nerve cells in vitro. Nature 295: 693-695.
PAGE 242
233 Grumet, M., U. Rutishauser and G.M. Edelman (1983) Neuronglia adhesion is inhibited by antibodies to neural determinants. Science 222: 60-62. Gundersen, R.W. and J.N. Barrett (1979) Neuronal chemotaxis: Chick dorsal root axons turn toward high concentrations of Nerve Growth Factor. Science 206: 1079-1080. Gundersen, R.W. and K.H.C. Park (1984) The effects of conditioned media on spinal neurites: Substrateassociated changes in neurite direction and adherence. Dev. Biol. 104: 18-27. Ham, R.G. (1965) Clonal growth of cells in a chemically defined, synthetic medium. Proc. Natl. Acad. Sci. USA 53: 288-293. Hamburger, V. (1948) The mitotic patterns in the spinal cord of the chick embryo and their relation to histogenetic processes.~Comp. Neurol. 88: 221-283. Hamburger, V. (1961) Experimental analysis of the dual origin of the trigeminal ganglion in the chick embryo. ~Exp. Zool. 148: 91-124. Hamburger, V. and H.L. Hamilton (1951) A series of normal stages in the development of the chick embryo.~Morph. 88: 49-92. Harrison, R.G. (1910) The outgrowth of the nerve fiber as a mode of protoplasmic movement.~Exptl. Zool. 9: 787-846. Harrison, R.G. (1912) The cultivation of tissues in extraneous media as a method of morphogenetic study. Anat. Rec. 6: 181. Hawkins, R.L. and N.W. Seeds (1986) Effect of proteases and their inhibitors on neurite outgrowth from neonatal mouse sensory ganglia in culture. Brain Res. 398: 63-70. Heacock, A.M. (1982) Glycoprotein requirement for neurite outgrowth in goldfish retina explants: Effects of tunicamycin. Brain Res. 241: 307-315. Heaton, M.B. (1981) The development of the oculomotor nuclear complex in the Japanese quail embryo. !d..!.. Comp. Neurol. 198: 633-648. Heaton, M.B. (1987) Influence of nerve growth factor on chick trigeminal motor nucleus explants. ~Comp. Neurol. 265: 362-366.
PAGE 243
234 Heaton, M.B. and H. Kemperman (1987) The influence of muscle-conditioned media on chick embryo brainstem neurons in culture.~Neurosci. Res. 17: 384-390. Heaton, M.B. and S.A. Moody (1980) Early development and migration of the trigeminal motor nucleus in the chick embryo.~ Comp. Neurol. 189: 61-99. Heaton, M.B., S.A. Moody and M.E. Kosier (1978) Peripheral innervation by migrating neuroblasts in the chick embryo. Neurosci. Lett. 20: 55-59. Heaton, M.B. and M. Paiva (1986) The influence of target tissue age on neurite outgrowth from chick embryo trigeminal motor nucleus explants. Dev. Biol. 116: 314-318. Hefti, F., J. Hartikka, A. Salvatierra, W.J. Wiener, and D.C. Mash (1986) Localization of nerve growth factor receptors in cholinergic neurons of the human basal forebrain. Neurosci. Lett. 69: 37-41. Henderson, C.E., M. Huchet and J.-P. Changeux (1981) Neurite outgrowth from embryonic chick spinal neurons is promoted by media conditioned by muscle cells. Proc. Natl. Acad. Sci. USA 78: 2625-2629. Hendry, I.A., K. Stockel, H. Thoenen, and L.L. Iversen (1974) The retrograde axonal transport of nerve growth factor. Brain Res. 68: 103-121. Henke-Fahle, S. and F. Bonhoeffer (1983) Inhibition of axonal growth by a monoclonal antibody. Nature 303: 65-67. Heumann, R., S. Korsching, J. Scott and H. Thoenen (1984) Relationship between levels of nerve growth factor (NGF) and its messenger RNA in sympathetic ganglia and peripheral target tissues. EMBO ~. 3: 3183-3189. Heumann, R., M. Schwab, and H. Thoenen (1981) A second messenger required for nerve growth factor biological activity? Nature 292: 838-840. Hockfield, s. and R. McKay (1983) Monoclonal antibodies demonstrate the organization of axons in the leech.~ Neurosci. 3: 369-375. Honig, M.G. (1982) The development of sensory projection patterns in embryonic chick hind limb.~Physiol. 330: 175-202.
PAGE 244
235 Honig, M.G. and R.I. Hurne (1986) Fluorescent carbocyanine dyes allow living neurons of identified origin to be studied in long-term cultures.~Cell Biol. 103: 171-187. Honneger, P., D. Lenoir, and P. Favrod (1979) Growth and differentiation of aggregating fetal brain cells in a serum-free defined medium. Nature 282: 305-308. Herder, T.J. (1978) Functional adaptability and rnorphogenetic opportunism, the only rules for limb development? Zoon 6: 181-192. Hosang, M. and E.M. Shooter (1985) Molecular characteristics of nerve growth factor receptors on PC12 cells.~Biol. Chern. 260: 655-662. Hsu, L., D. Natyzak, and G.L. Trupin (1982) Neuronotrophic effects of skeletal muscle fractions on spinal cord differentiation.~Ernbryol. Exp. Morph. 71: 83-95. Hsu, L., D. Natyzak and G.L. Trupin (1984) Neuronotrophic effects of skeletal muscle fractions on neurite development. Muscle Nerve 7: 211-217. Hughes, A. (1933) The growth of embryonic neurites. A study on cultures of chick neural tissues.~Anat. 87: 150-164. Hulst, J.R. and M. Bennett (1986) Motoneurone survival factor produced by muscle increases in parallel with messenger RNA sequences homologous to BNGF-cDNA. Dev. Brain Res. 25: 153-156. Ingvar, s. (1920) Reaction of cells to galvanic current in tissue cultures. Proc. Soc. Exp. Biol. Med. 17:198-199. Isornura, G. (1973) On the localization of nerve centers of the extrinsic ocular muscles in the fowl. Anat. Anz. Bd. 133: 33-50. Jacobson, M. (1983) Clonal organization of the central nervous system of the frog.~Neurosci. 3: 1019-1038. Jaffe, L. and M.-M. Poo (1979) Neurites grow faster towards the cathode than the anode in a steady field.~Exp. Zool. 209: 115-128. Johnson, E.M., R.Y. Andres, and R.A. Bradshaw (1978) Characterization of the retrograde transport of nerve growth factor (NGF) using high specific activity [125I]NGF. Brain Res. 150: 319-331.
PAGE 245
236 Johnson, E.M., P.D. Gorin, L.D. Brandeis and J. Pearson (1980) Dorsal root ganglion neurons are destroyed by exposure in utero to maternal antibody to nerve growth factor. Science 210: 916-918. Johnson, E.M., P.A. Osborne, R.E. Rydel, R.E. Schmidt and J. Pearson (1983) Characterization of the effects of autoimmune nerve growth factor deprivation in the developing guinea-pig. Neurosci. 8: 631-642. Johnson, E.M., M. Taniuchi, H.B. Clark, J.E. Springer, s. Koh, M.W. Tayrien, and R. Loy (1987) Demonstration of the retrograde transport of nerve growth factor receptor in the peripheral and central nervous system.~ Neurosci. 7: 923-929. Katz, M.J. (1986) Quantitative effects of NGF on the growth of embryonic frog axons. Brain Res. 366: 211-216. Katz, M.J., E.B. George, and L.J. Gilbert (1984) Axonal elongation as a stochastic walk. Cell Motil. 4: 351-370. Katz, M.J. and R.J. Lasek (1979) Substrate pathways which guide growing axons in Xenopus embryos.~. Comp. Neurol. 183: 817-832. Katz, M.J. and R.J. Lasek (1981) Substrate pathways demonstrated by transplanted Mauthner axons.~. Comp. Neurol. 195: 627-641. Katz, M.J., R.J. Lasek and H.J.W. Nauta (1980) Ontogeny of substrate pathways and origin of the neural curcuit pattern. Neurosci. 5: 821-833. Kessler, J.A. and I.B. Black (1981) Similarities in development of substance P and somatostatin in peripheral sensory neurons: Effects of capsaicin and nerve growth factor. Proc. Natl. Acad. Sci. USA 78: 4644-4647. Keynes, R.J. and C.D. Stern (1984) Segmentation in the vertebrate nervous system. Nature 310: 786-789. Korsching, s. and H. Thoenen (1983) Quantitative demonstration of the retrograde axonal transport of endogenous nerve growth factor. Neurosci. Lett. 39: 1-4. Korsching, s. and H. Thoenen (1985) Nerve growth factor supply for sensory neurons: Site of origin and competition with the sympathetic nervous system. Neurosci. Lett. 54: 201-205.
PAGE 246
237 Kromer, L.F. (1987) Nerve growth factor treatment after brain injury prevents neuronal death. Science 235: 214-216. Krystosek, A. and N.W. Seeds (1984) Peripheral neurons and Schwann cells secrete plasminogen activator.~Cell Biol. 98: 773-776. Lance-Jones, C. and L. Landmesser (1980) Motoneurone projection patterns in the chick hind limb following early partial reversals of the spinal cord.~Physiol., 302: 581-602. Lance-Jones, C. and L. Landmesser (1981a) Pathway selection by chick lumbosacral motoneurons during normal development. Proc. Roy. Soc. Ser.~, 214: 1-19. Lance-Jones, C. and L. Landmesser (1981b) Pathway selection by embryonic chick motoneurons in an experimentally altered environment. Proc. Roy. Soc. Ser.~, 214: 19-52. Lander, A.D., D.K. Fujii, D. Gospodarowicz and L.F. Reichardt (1982) Characterization of a factor that pormotes neurite outgrowth --evidence linking activity to a heparan sulfate proteoglycan. ~Cell Biol. 94: 574-585. Lander, A.D., D.K. Fujii and L.F. Reichardt (1985) Laminin is associated with the "neurite outgrowth-promoting factors" found in conditioned media. Proc. Natl. Acad. Sci. USA 82: 2183-2187. Landmesser, L. (1978a) The distribution of motoneurons supplying chick hind limb muscles.~ Physiol. 284: 371-389. Landmesser, L. (1978b) The development of motor projection patterns in the chick hind limb.~ Physiol. 284: 391-414. Landmesser, L. and M. Honig (1982) The effect of motoneuron removal on sensory neuron outgrowth in chick hindlimb. Soc. Neurosci. Abst. 8: 929. Langman, J. and c.c. Haden (1970) Formation and migration of neuroblasts in the spinal cord of the chick embryo.~ Comp. Neural. 138: 419-432. Laskowski, M.B. and J.R.Sanes (1987) Topographic mapping of motor pools onto skeletal muscles.~ Neurosci. 7: 252-260.
PAGE 247
238 Leonard, D.G.B., L.A. Greene, and E.B. Ziff (1986) Characterization of nerve growth factor-regulated mRNAs. Soc. Neurosci. Abst. 12: 215. Letourneau, P.C. (1975) Possible roles for cell-tosubstratum adhesion in neuronal morhpogenesis. Dev. Biol. 44: 77-91. Letourneau, P.C. (1979) Cell-substratum adhesion of neurite growth cones, and its role in neurite elongation. Exp. Cell Res. 124: 127-138. Levi-Montalcini, R. (1976) The nerve growth factor: its role in growth, differentiation and function of the sympathetic adrenergic neuron. Prog. Brain Res. 45: 235-258. Levi-Montalcini, R. (1982) Developmental neurobiology and the natural history of nerve growth factor. Ann. Rev. Neurosci. 5: 341-362. Levi-Montalcini, R. and L. Aloe (1985) Differentiating effects of murine nerve growth factor in the peripheral and central nervous systems of Xenopus laevis tadpoles. Proc. Natl. Acad. Sci. USA 82: 7111-7115. Levi-Montalcini, R. and P.U. Angeletti (1968) Nerve growth factor. Physiol. Rev. 48: 534-565. Levi-Montalcini, R. and V. Hamburger (1951) Selective growth-stimulating effects of mouse sarcoma on the sensory and sympathetic nervous system of the chick embryo.~. Exp. Zool. 116: 321-362. Levi-Montalcini, R. and V. Hamburger (1953) A diffusible agent of mouse sarcoma producing hyperplasia and hyperneurotization of the chick embryo.~. Exp. Zool. 123: 233-388. Levi-Montalcini, R., H. Meyer and V. Hamburger (1954) Invitro experiments on the effects of mouse sarcoma 180 and 37 on the spinal and sympathetic ganglia of the chick embryo. Cancer Res. 14: 49-57. Lewis, J., A. Chevallier, M. Kieny, and L. Wolpert (1981) Muscle nerve branches do not develop in chick wings devoid of muscle.~. Embryol. Exp. Morph. 64: 211-232. Lis, H. and H. Sharon (1977) Lectins: Their chemistry and application to immunology. In The Antigens, M. Sela, ed., pp.429-529, Academic Press, New York.
PAGE 248
239 Lumsden, A.G.S. and A.M. Davies (1986) Chemotropic effect of specific target epithelium in the developing mammalian nervous sytem. Nature 323: 538-539. Marsh, G. and H.W. Beams (1946) In vitro control of growing chick nerve by applied electric currents.~Cell Comp. Physiol. 27: 139-157. Matesz, C. and G. Szekely (1977) The dorsomedial nuclear group of cranial nerves in the frog. Acta. Biol. Acad. Sci. Hung. 28: 461-474. McKay, R.D.G. and S.J. Hockfield (1982) Monoclonal antibodies distinguish antigenically discrete neuronal types in the vertebrate central nervuos system. Proc. Natl. Acad. Sci. USA 79: 6747-6751. Meier, R., M. Becker-Andre, R. Getz, R. Heumann, A. Shaw, and H. Thoenen (1986) Molecular cloning of bovine chick nerve growth factor (NGF): delineation of conserved and unconserved domains and their relationship to the biological activity and antigenicity of NGF. EMBO ~5: 1489-1493. Menesini-Chen, M.G., J.S. Chen and R. Levi-Montalcini (1978) Sympathetic nerve fibers ingrowth in the central nervous system of neonatal rodent upon intracerebral NGF injection. Arch. Ital. Biol. 116: 53-84. Mesulam, M. M. (1978) Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: A non-carcinogenic blue reaction product with superior sensitivity for visualizing neural afferents and efferents.~ Histochem. Cytochem. 26: 106-117. Monard, D. (1985) Neuronal cell behaviour: Modulation by protease inhibitors derived from non-neuronal cells. Cell Biol. Interntl. Rep. 9: 297-305. Moody, S.A. and M.B. Heaton (1981) Morphology of migrating trigeminal motor neuroblasts as revealed by horseradish peroxidase retrograde labeling techniques. Neurosci. 6: 1707-1723. Moody, S.A. and M.B. Heaton (1983a) Developmental relationships between trigeminal ganglia and trigeminal motoneurons in chick embryos. I. Ganglion development is necessary for motoneuron migration.~Comp. Neural. 213: 327-343.
PAGE 249
240 Moody, S.A. and M.B. Heaton (1983b) Developmental relationships between trigerninal ganglia and trigerninal rnotoneurons in chick embryos. II. Ganglion axon ingrowth guides rnotoneuron migration.~Comp. Neurol. 213: 344-349. Moody, S.A. and M.B. Heaton (1983c) Developmental relationships between trigerninal ganglia and trigerninal rnotoneurons in chick embryos. III. Ganglion perikarya direct motor axon growth in the periphery.~Comp. Neurol. 213: 350-364. Moody, S.A. and M. Jacobson (1983) Compartmental relationships between anuran primary spinal rnotoneurons and semitic muscle fibers that they first innervate.~ Neurosci. 3: 1670-1682. Nakajima, s. (1965) Selectivity in fasciculation of nerve fibers in vitro.~Comp. Neurol. 125: 193-204. Nirnii, K., T. Sakai, and J. Takasu (1958) The ontogenetic development of the oculornotor nucleus in the chick. Takushirna ~~Med. 5: 311-325. Noakes, P.G. and M.R. Bennett (1987) Growth of axons into developing muscles of the chick forelimb is preceded by cells that stain with Schwann cell antibodies.~ Comp. Neurol. 259: 330-347. Noakes, P.G., A.W. Everett, and M.R. Bennett (1986) The growth of muscle nerves in relation to formation of primary rnyotubes in the developing chick forelimb.~ Comp. Neurol. 248: 245-256. Noden, D.M. (1975) Analysis of the migratory behavior of avian cephalic nerual crest cells. Dev. Biol. 42: 106-130. Nurcornbe, v. and M.R. Bennett (1983) The growth of neurites from explants of brachial spinal cord exposed to different components of wing bud rnesenchyrne. ~Comp. Neurol. 219: 133-142. Okado, N. and R.W. Oppenheim (1984) Cell death of rnotoneurons in the chick embryo spinal cord. IX. The loss of rnotoneurons following removal of afferent inputs.~Neurosci. 4: 1639-1652. Olender, E.J. and R.W. Stach (1980) Sequestration of 125Ilabeled B nerve growth factor by sympathetic neurons.~ Biol. Chern. 255: 9338-9343.
PAGE 250
Palka, J. and A. Ghysen (1982) Segments, compartments and axon paths in Drosophila. TINS 5: 382-385 241 Palka, J., M. Schubiger and R.L. Ellison (1983) The polarity of axon growth in the wings of Drosophila melanogaster. Dev. Biol. 98: 481-492. Palm, S.L. and L.T. Furcht (1983) Production of laminin and fibronectin by Schwannoma cells: Cell-protein interactions in vitro and protein localization in peripheral nerve in vivo.~. Cell Biol. 96: 1218-1226. Palmatier, M.A., B.K. Hartman, and E.M. Johnson (1984) Demonstration of retrogradely transported endogenous nerve growth factor in axons of sympathetic neurons.~ Neurosci. 4: 751-756. Patel, N. and M.-M. Poo (1982) Orientation of neurite growth by extracellular electric fields.~Neurosci. 2: 483-496. Patel, N.H., P.M. Snow, and C.S. Goodman (1987) Characterization and cloning of fasciclin III: A glycoprotein expressed on a subset of neurons and axon pathways in Drosophila. Cell 48: 975-988. Pfenninger, K.H. and M.-F. Maylie-Pfenninger (1981) Lectin labeling of sprouting neurons I. Regional distribution of surface glycoconjugates. ~Cell Biol. 89: 536-546. Phillips, W.D. and M.R. Bennett (1984) Differentiation of fiber types in wing muscles during embryonic development: Effect of neural tube removal. Dev. Biol. 106: 457-468. Pollack, E.D. (1980) Target-dependent survival of tadpole spinal cord neurites in tissue culture. Neurosci. Lett. 16: 269-274. Pollack, E.D. and W.L. Muhlach (1981) Stage dependency in eliciting target-dependent enhanced neurite outgrowth from spinal cord explants in vitro. Dev. Biol. 86: 259-263. Pollack, E.D. and W.L. Muhlach (1982) Target control of neuronal development during formation of the spinal reflex arc: An operant model.~. Neurosci. Res. 8: 343-355.
PAGE 251
242 Pollack, E.D., W.L. Muhlach, and V. Liebig (1981) Neurotropic influence of mesenchymal limb target tissue on spinal cord neurite growth in vitro.~ Comp. Neurol. 200: 393-405. Poo, M.-M. and K.R. Robinson (1977) Electrophoresis of concanavalin A receptors along embryonic muscle cell membrane. Nature 265: 602-605. Puelles-Lopez, L., F. Malagon-Cobos, and J.M. Genis-Galvez (1975) The migration of oculomotor neuroblasts across the midline in the chick embryo.~ Neurol. 47: 459-469. Purves, D. (1976) Functional and sturctural changes in mammalian sympathetic neurones following colchicine application to post-ganglionic nerves.~Physiol. 259, 159-175. Rager, G. (1980) Specificity of nerve connections by unspecific mechanisms. TINS 3: 43-44. Rager, G. and B. von Oeyenhauser (1979) Ingrowth and ramificaton of retinal fibers in the developing optic tectum of the chick embryo~ Brain Res. 35: 213-227. Raivich, G., A. Zimmermann and A. Sutter (1985) The spatial and temporal pattern of BNGF receptor expression in the developing chick embryo. EMBO ~. 4: 637-644. Raivich, G., A. Zimmermann and A. Sutter (1987) Nerve growth factor (NGF) receptor expression in chicken cranial development. ~. Comp. Neurol. 256: 229-245. Ramon y Cajal, s. (1960) Studies on Vertebrate Neurogenesis. Revised and translated by L. Guth. Springfield, Illinois: Thomas. Regan, L.J., J. Dodd, S.H. Barondes and T.M. Jessell (1986) Selective expression of endigenous lactose-binding lectins and lactoseries glycoconjugates in subsets of rat sensory neurons. Proc. Natl. Acad. Sci. USA 83: 2248-2252. Richardson, P.M., V.M.K. Verge Issa, and R.J. Riopelle (1986) Distribution of neuronal receptors for nerve growth factor in the rat.~. Neurosci. 6: 2312-2321. Riggott, M.J. and S.A. Moody (1987) Distribution of laminin and fibronectin along peripheral trigeminal axon pathways in the developing chick.~. Comp. Neurol. 258: 580-596.
PAGE 252
243 Robinson, K.R. (1983) Endogenous electrical current leaves the limb and prelimb region of the Xenopus embryo. Dev. Biol. 97: 203-211. Roederer, E. and M.J. Cohen (1983) Regeneration of an identified central neuron in the cricket. I. Control of sprouting from soma, dendrites, and axon.~Neurosci. 3: 1835-1847. Rogers, S.L., K.J. Edson, P.C. Letourneau and s.c. McLoon (1986) Distribution of laminin in the developing peripheral nervous system of the chick. Dev. Biol. 113: 429-435. Rogers, S.L., P.C. Letourneau, S.L. Palm, J. McCarthy and L.T. Furcht (1983) Neurite extension by peripheral and central nervous system neurons in response to substratum-bound fibronectin and laminin. Dev. Biol. 98: 212-220. Rosiles, J.R. and R.B. Leonard (1980) The organization of the extraocular motor nuclei in the Atlantic Stingray, Dasyatis sabina. Comp. Neurol. 193: 677-687. Rutishauser, U., W.E. Gall and G.M. Edelman (1978b) Adhesion among neural cells of the chick embryo IV. Role of the cell surface molecule CAM in the formation of neurite bundles in cultures of spinal ganglia.~Cell Biol. 79: 382-393. Rutishauser, U., M. Grumet and G.M. Edelman (1983) Neural CAM mediates initial interactions between spinal cord neurons and muscle cells in culture.~Cell Biol. 97: 145-152. Rutishauser, U., J.-P. Thiery, R. Brackenbury and G.M. Edelman (1978a) Adhesion among neural cells of the chick embryo III. Relationship of the surface molecule CAM to cell adhesion and development of histotypic patterns.~ Cell Biol. 79: 371-381. Schechter, A.L. and M.A. Bothwell (1981) Nerve growth factor receptors on PC12 cells: Evidence for two receptor classes with differing cytoskeletal association. Cell 24: 867-874. Seeley, P.J., C.H. Keith, M.L. Shelanski, and L.A. Greene (1983) Pressure microinjection of nerve growth factor and anti-nerve growth factor into the nucleus and cytoplasm: Lack of effects on neurite outgrowth from pheochromocytoma cells.~Neurosci. 3: 1488-1494.
PAGE 253
244 Seiler, M. and N.E. Schwab (1984) Specific retrograde transport of nerve growth factor (NGF) from neocortex to nucleus basalis in the rat. Brain Res. 300: 33-39. Shelton, D.L. and L.F Reichardt (1984) Expression of the B nerve growth factor gene correlates with the density of sympathetic innervation in effector organs. Proc. Natl. Acad. Sci. USA 81: 7951-7955. Silver, J. (1983) Symposium on cell recognition in development of the nervous system. Soc. Neurosci. Abst. 9: 151. Silver, J. (1984) Studies on the factors that govern directionality of axonal growth in the embryonic optic nerve and at the chiasm of mice.~. Comp. Neural. 223: 238-251. Silver, J., S.E. Lorenz, D. Wahlsten and J. Coughlin (1982) Axonal guidance during development of the great cerebral commissures: Descriptive and experimental studies, in vivo, on the role of preformed glial pathways.~. Comp. Neural. 210: 10-29. Silver, J., M. Poston, and U. Rutishauser (1987) Axon pathway boundaries in the developing brain. I. Cellular and moleuclar determinants that separate the optic and olfactory projections.~Neurosci. 7: 2264-2272. Silver, J. and U. Rutishauser (1984) Guidance of optic axons in vivo by a preformed adhesive pathway on neuroepithelial endfeet. Dev. Biol. 106: 485-499. Singer, M., R.H. Nordlander and M. Egar (1979) Axonal guidance during embryogenesis and regeneration in the spinal cord of the newt: The Blueprint Hypothesis of neuronal pathway patterning.~Comp. Neural. 185: 1-21. Sisken, B.F., E. Roberts, and I. Goetz (1985) Triethanolamine, Tris, Hepes, and cytosine arabinoside show neuritogenic activity in cultured chick embryo ganglia. Exp Neural 88: 27-43. Skaper, S.D., R. Adler ands. Varon (1979) A procedure for purifying neuron-like cells in cultures from central nervous tissue with a defined medium. Dev. Neurosci. 2: 233-237. Smith, R.G. and S.H. Appel (1983) Extracts of skeletal muscle increase neurite outgrowth and cholinergic activity of fetal rat spinal motor neurons. Science 219: 1079-1081.
PAGE 254
245 Smith, R.G., K. Vaca, J. McManaman, and S.H. Appel (1986) Selective effects of skeletal muscle extract fractions on motneuron development in vitro.~Neurosci. 6: 439-447. Sohal, G.S. (1977) Development of the oculomotor nucleus, with special reference to the time of cell origin and cell death. Brain Res. 138: 217-228. Speidel, c.c. (1933) Studies of living nerves II. Activities of amoeboid growth cones, sheath cells and myelin segments as revealed by prolonged observation of individual nerve fibers in frog tadpoles. Am.~Anat. 52: 1-44. Spencer, R.F. and J.D. Porter (1981) Innervation and structure of extraocular muscles in the monkey in comparison to those of the cat.~ Comp. Neurol. 198: 649-665. Sperry, R.W. (1944) Optic nerve regeneration with return of vision in anurans. ~Neurophysiol. 7: 57-69. Sperry, R.W. (1945) Restoration of vision after crossing of optic nerves and after contralateral transplantation of eye.~Neurophysiol. 8: 15-28. Sperry, R.W. (1963) Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. USA 50: 703-710. Stach, R.W. and J.R. Perez-Polo (1987) Binding of nerve growth factor to its receptor.~Neurosci. Res. 17: 1-10. Stach, R.W., B.M. Stach and D.J. Ennulat (1986) Phenylmethylsulfonyl fluoride (PMSF) inhibits nerve growth factor binding to the high affinity (type I) nerve growth factor receptor. Biochem. Biophys. Res. Comm. 134: 1000-1005. Stern, C.D, S.M. Sisodiya, and R.J. Keynes (1986) Interactions between neurites and somite cells: Inhibition and stimulation of nerve growth in the chick embryo.~Embryol. Exp. Morph. 91: 209-226. Stockel, K., u. Paravicini, and H. Thoenen (1974) Specificity of the retrograde axonal transport of nerve growth factor. Brain Res. 76: 413-421.
PAGE 255
246 Stockel, K., M. Schwab, and H. Thoenen {1975b) Comparison between the retrograde axonal transport of nerve growth factor and tetanus toxin in motor, sensory and adrenergic neurons. Brain Res. 99: 1-16. Stoeckel, K., M. Schwab, and H. Thoenen {1975a) Specificity of retrograde transport of nerve growth factor {NGF) in sensory neurons: A biochemical and morphological study. Brain Res. 89: 1-14. Straznicky, C. and R.A. Rush {1985a) The effect of nerve growth factor on developing primary sensory neurons of the trigeminal nerve in chick embryos. Anat. Embryol. 171: 91-95. Straznicky, C. and R.A. Rush {1985b) Nerve growth factor treatment does not prevent dorsal root ganglion cell death induced by target removal in chick embryos. Anat. Embryol. 171: 357-363. Summerbell, D. and V. Stirling (1981) The innervation of dorsoventrally reversed chick wings: evidence that motor axons do not actively seek out their appropriate targets.~. Embryol. Exp. Morph. 61: 233-247. Sutter, A., R.J. Riopelle, R.M. Harris-Warwick and E.M. Shooter {1979) Nerve growth factor receptors. Characterization of two distinct classes of binding sites on chick embryo sensory ganglia cells.~. Biol. Chem. 254: 5972-5982. Swanson, G.J. and J. Lewis {1982) The timetable of innervation and its control in the chick wing bud.~ Embryol. Exp. Morph. 71: 121-137. Taniuchi, M., H.B. Clark, and E.M. Johnson {1986) Induction of nerve growth factor receptor in Schwann cells after axotomy. Proc. Natl. Acad. Sci. USA 83: 4094-4098. Thoenen, H. and D. Edgar {1982) The regulation of neuronal gene expression. TINS 5: 311-313. Tosney, K.W. and L.T. Landmesser {1985a) Specificity of early motoneuron growth cone outgrowth in the chick embryo.~ Neurosci. 5: 2336-2344. Tosney, K.W. and L.T. Landmesser {1985b) Growth cone morphology and trajectory in the lumbosacral region of the chick embryo.~ Neurosci. 5: 2345-2358.
PAGE 256
247 Tosney, K.W. and L.T. Landmesser (1985c) Development of the major pathways for neurite outgrowth in the chick hindlimb. Dev. Biol. 109: 193-214. Tosney, K.W., M. Watanabe, L. Landmesser, and U. Rutishauser (1986) The distribution of NCAM in the chick hindlimb during axon outgrowth and synaptogenesis. Dev. Biol. 114: 437-452. Tosney, K.W. and N.K. Wessells (1983) Neuronal motility: The ultrasturcture of veils and microspikes correlates with their motile activities.~Cell Sci. 61: 389-411. Vale, R.D. and E.M. Shooter (1982) Alteration of binding properties and cytoskeletal attachment of nerve growth factor receptors in PC12 cells by wheat germ agglutinin. ~Cell Biol. 94: 710-717. Vogel, M. and L. Landmesser (1987) Distribution of fiber types in embryonic chick limb muscles innervated by foreign motoneurons. Dev. Biol. 119: 481-495. Vuillamy, T., s. Rattray and R. Mirsky (1981) Cell-surface antigen distinguishes sensory and autonomic peripheral neurones from central neurones. Nature 291: 418-420. Wakade, A.R., D. Edgar and H. Thoenen (1982) Substrate requirement and media supplements for the long-term survival of chick sympathetic and sensory neurons cultured without serum. Exp. Cell Res. 140: 63-70. Wayne, D.B. and M.B. Heaton (1983) Development of the chick oculomotor complex following removal of the target muscles. Anat. Rec. 205: 211a. Wayne, D.B. and M.B. Heaton (1988) Retrograde transport of NGF by early chick embryo spinal cord motoneurons. Dev. Biol. in press. Weill, C.L. (1983) Induction of spinal cord androgen receptors by muscle factors. Soc. Neurosci. Abst. 9: 843. Weiss, P. (1934) In vitro experiments on the factors determining the course of the outgrowing nerve fiber.~ exp. Zool. 68: 393-448. Weiss, P. (1936) Selectivity controlling the centralperipheral relations in the nervous system. Biol. Rev. 11: 494-531.
PAGE 257
248 Wenger, B.S. (1968) Construction and use of the vibrating needle for embryonic operations. Bioscience 18: 226-228. Whitelaw, V. and M. Hollyday (1983a) Thigh and calf discrimination in the motor innervation ofthe chick hindlimb following deletions of limb segments~ Neurosci. 3 (6): 1199-1215. Whitelaw, V. and M. Hollyday (1983b) Position-dependent motor innervation of the chick hindlimb following serial and parallel duplications of limb segments.~. Neurosci. 3: 1216-1225. Whitelaw, V. and M. Hollyday (1983c) Neural pathway constraints in the motor innervation of the chick hindlimb following dorsoventral rotations of distal limb segments.~. Neurosci. 3: 1226-1233. Whittemore, S.R., T. Ebendal, L. Larkfors, L. Olson, A. Sieger, I. Stromberg and H. Persson (1986) Developmental and regional expression of B nerve growth factor messenger RNA and protein in the rat central nervous system. Proc. Natl. Acad. Sci. USA 83: 817-821. Windle, W.F. (1933) Neurofibrillar development in the central nervous system of cat embryos between 8 and 12 mm long.~ Comp. Neural. 58: 643-723. Wolinsky, E.J. and P.H. Patterson (1985) Rat serum contains a developmentally regulated cholinergic inducing activity.~Neurosci. 5: 1509-1512. Yan, Q. and E.M. Johnson (1987) A quantitative study of the developmental expresssion of nerve growth factor (NGF) receptor in rats. Dev. Biol. 121: 139-148. Young, M., M.H. Blanchard and J. Saide (1979) Nerve growth factor: Radioimmunoassay and bacteriophage immunoassay. In Methods of Hormone Radioimmunoassay, B.M. Jaffee and h d nd d d H.R. Berman, es., 2 E ., pp. 946-949. Aca em1c Press, New York. Zimmermann, A. and A. Sutter (1983) B-Nerve growth factor (BNGF) receptors on glial cells. Cell-cell interaction between neurones and Schwann cells in cultures of chick sensory ganglia. EMBO ~. 2: 879-885.
PAGE 258
BIOGRAPHICAL SKETCH I was born in Newark, N.J., in 1953. Following a childhood spent playing in the cornfields of New Jersey, I graduated from Livingston High School in 1971. I then migrated north to Worcester, Massachusettt, and Clark University. I spent my undergraduate years making the rounds of biology classes and finally found something that was more than just interesting, the nervous system. What truly sustained me at Clark, however, were my afternoons rowing crew on Lake Quinsigamond. I received my B.A. in 1976 and left Worcester with my diploma and the certainty that I was finished with my formal schooling. I moved on to Boston where my first two years were spent as a research assistant processing tissue for electron microscopy in the Pathology Department at Harvard Medical School. While I enjoyed the work and found the results esthetically pleasing, I realized that renal glomeruli did not compare to the intrigue that the nervous system held for me. I spent my third year in Boston working at Eastern Mountain Sports. Boston offered me a richness of family and friends that made leaving particularly difficult. Nevertheless, the time had arrived, so I bought a new pair 249
PAGE 259
of cross country skis and headed for graduate school in Florida. 250 I came to the Department of Neuroscience at the University of Fl6rida in July of 1979. I arrived with an interest in the nervous system and an affinity for anatomy. I was happily introduced to the mysteries of neural development in the laboratory of Dr. Marieta Heaton. Upon finishing my dissertation, I will be moving to St. Louis (and the cortex) to take a post-doctoral position in the laboratory of Dr. Alan Pearlman to continue exploring the mechanisms involved in early neural development. I am also hoping to finally put my skis to use.
PAGE 260
I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ~-\;~ Marieta B. Heaton, Chairperson Associate Professor of Neuroscience I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosop~ [<'./~ Don E. Walker Professor of Neuroscience I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ~ M11e~tJ: Associate Professor of Anatomy and Cell Biology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. &;~~ Associate Professor of Neuroscience This dissertation was submitted to the Graduate Faculty of the College of Medicine and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. April 1988 William B. Deal Dean, College of Medicine fx-a.d.d_ Madely'1?r1.ockhart Dean, Graduate School
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EBB2ODCPF_FLJ4GR INGEST_TIME 2017-06-05T16:54:24Z PACKAGE AA00055797_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
|