A paradoxical gradient of a basal lamina-associated repellent is essential for pathfinding by the Ti1 pioneer axons in cockroach embryos.

Perturbations of in situ axon growth with proteolytic enzymes and monoclonal antibodies were used to determine the role of gradient guidance cues in the formation of the Ti1 pioneer axon trajectory in cultured cockroach embryos. Treatment with enzymes that degrade the basal lamina indicated that this substrate contains both an elastase-sensitive proximal directing cue and a collagenase-sensitive distal directing cue. The latter is shown to be a repellent of axon growth and is identical to the PROD-2 antigen that is distributed in a gradient along the proximal-distal axis of the leg with high levels in proximal regions. This means that throughout the course of their growth the axons extend in the direction of increasing levels of repellent. At a critical decision point in the trajectory the axons change both the direction of growth and the substrate to which their growth cones adhere. PROD-2 plays an essential role in both of these processes.

The in vivo regulation of pioneer axon growth by FGF-2 and heparan sulfate proteoglycans in cultured embryos of the cockroach.

Antibody perturbation experiments on cultured cockroach embryos demonstrated that a localized source of an FGF-2-like immunoreactive molecule in the head is required for the proper growth of pioneer axons in the leg. The study of axon growth in various fragments of cultured embryos and in the presence of various conditioned media showed that FGF-2 is needed to counteract the effects of an inhibitor of axon growth produced in the body trunk of the embryo. Endogenous heparan sulfate proteoglycans mediate these effects of FGF-2 on axon growth. The results of experiments with FGF-2 and/or body trunk axon growth inhibitor added to the culture medium indicate that more globally and uniformly distributed molecules may play as important a role in axon guidance as the more spatially restricted guidance cues. The results are interpreted in terms of a model that is consistent with a role for the FGF-2 receptor in axon growth.

Mesodermal guidance of pioneer axon growth.

Pioneer axons in insect legs are experimentally accessible model systems for the molecular identification and cellular localization of guidance cues regulating the path of axon growth. A detailed study of the Fe2 pioneer axons in the legs of the cockroach was performed to examine the diversity of guidance mechanisms. A detailed microscopic analysis of the axons at various points in their trajectory indicates that the Fe2 axons grow on a mesodermal substratum which contains the cues guiding their growth along a stereotyped path. An identified pair of muscle pioneer cells (MPC) are likely to play an important role in enabling the Fe2 growth cones to respond to mesodermal guidance cues. The addition of heparan sulfate, heparitinase, and phosphatidylinositol-specific phospholipase C to the medium perturbs the in situ path of growth of the Fe2 axons and the location of the MPC in cultured embryos. This indicates a role for heparan sulfate proteoglycans and glycosylphosphatidylinositol-anchored proteins in axon guidance. When these results are compared to those of similar experiments performed on the well-characterized Ti1 axons, they indicate significant differences in the mechanisms that are used for axon guidance. The Fe2 neurons are a good model for elucidating the mechanisms used to guide axon growth on nonmuscle mesodermal substrates often encountered in the periphery of vertebrate embryos.

Error correction during guidance of pioneer axons in the leg of the cockroach embryo.

Axons of the Til and Fe2 pioneer neurons in the legs of insect embryos possess separate and highly stereotyped proximal projections towards the CNS. However, quantitative analyses of deviations from the standard paths during the period of axon growth indicate that transient errors occur unexpectedly often. The distribution of legs with axons following deviant paths among the embryos analyzed is used to determine whether these errors are caused by random developmental noise or by non-random genetic or environmental factors. During the formation of the Til pathway all the errors are characterized by defasciculation of the 2 axons, occur with an average incidence of 7% and are statistically shown to be randomly caused. In comparison, during the formation of the Fe2 pathway the errors are characterized by both defasciculation and elongation in an inappropriate distal direction, occur with an incidence of 16%, and as revealed by statistical analyses, are caused by a non-random factor. Therefore, during pathfinding by these 2 pairs of axons there is a need for error-correcting mechanisms to insure the stereotypy of the final projections. These error-correcting mechanisms are suggested to have properties similar to those producing canalization as proposed by Waddington.

Developmental changes in epitope accessibility as an indicator of multiple states of an immunoglobulin-like neural cell adhesion molecule.

Cell surface molecules with restricted spatial and temporal distributions are good candidates for mediators of the cell-cell interactions that are necessary for the development of the nervous system. A monoclonal antibody (MAb 23A7) was produced that selectively and transiently labeled a limited subset of axons in the chick embryo spinal cord. Determination of the N-terminal amino acid sequence and immunoprecipitation experiments demonstrated that the 23A7 antigen is identical to Bravo/Nr-CAM, a previously described cell adhesion molecule with immunoglobulin-like domains (E.J. de la Rosa, J.F. Kayyem, J.M. Roman, Y.-D. Stierhof, W.J. Dreyer, and U. Schwartz [1989] J. Cell Biol. 111:3087-3096; M. Grumet, V. Mauro, M.P. Goon, G.M. Edelman, and B.A. Cunningham [1991] J. Cell Biol. 113:1399-1412). The temporal distribution of the 23A7 antigen is unusual in that, immunohistochemically, MAb 23A7 binding greatly decreases after 7 days of development, whereas Western blot analysis indicates increasing levels of the antigen until 17 days of development. In contrast, an antiserum against purified Nr-CAM, which also binds only to the 23A7 antigen, labels nearly all the axons in the tissue throughout all the later stages of development. These anomalous observations are apparently not the result of differential sensitivity of the 23A7 epitope to fixation, the use of suboptimal concentrations of the MAb, or selective MAb binding to a subset of Bravo/Nr-CAM molecules produced by alternative splicing of the transcript or by posttranslational modification. These findings could indicate the existence of multiple states of Bravo/Nr-CAM, which during development, vary in the accessibility of their extracellular domains to the MAb. This suggests the existence of multiple conformation or aggregation states of this cell adhesion molecule, each of which might be performing a different function.

Molecular gradients along the proximal-distal axis of embryonic insect legs: possible guidance cues of pioneer axon growth.

It has been proposed that gradients of environmental cues direct the proximal growth of pioneer axons in embryonic insect legs. Hybridoma techniques have been used to produce 3 monoclonal antibodies (mAbs) that bind to components associated with the basal lamina/extracellular matrix that are non-uniformly distributed along the proximal-distal axis of cockroach legs at the time of pioneer axon growth. Two of these mAbs, PROD-1 and PROD-2, label the proximal parts of the leg more intensely than the distal ends. The other mAb, DIP-1, has the reverse pattern of binding with the distal parts of the leg labeled more intensely. The graded distribution of these antigens only occurs just prior to and during the growth period of the Ti1 pioneer axons. Western blot analyses and immunoprecipitations have identified the protein antigens recognized by these mAbs. The spatial and temporal distributions of these molecules in the legs and the CNS make them good candidates for environmental guidance cues of pioneer axon growth.

A multifunctional cell surface developmental stage-specific antigen in the cockroach embryo: involvement in pathfinding by CNS pioneer axons.

mAb DSS-8 binds to a 164-kD developmental stage-specific cell surface antigen in the nervous system of the cockroach, Periplaneta americana. The antigen is localized to different subsets of cells at various stages of development. The spatial and temporal distributions of DSS-8 binding were determined and are consistent with this antigen playing multiple roles in the development of the nervous system. Direct identification of some of these functions was made by perturbation experiments in which pioneer axon growth occurs in embryos that are cultured in vitro in the presence of mAb DSS-8 or its Fab fragment. Under these conditions the pioneer axons of the median fiber tract grow but follow altered pathways. In a smaller percentage of the ganglia, the immunoreagents additionally produce defasciculation of a subset of DSS-8 labeled axons. Therefore, direct roles for the DSS-8 antigen in both the guidance of pioneer axons and selective fasciculation have been demonstrated.

A role for proteoglycans in the guidance of a subset of pioneer axons in cultured embryos of the cockroach.

Several molecules involved in the development of the nervous system have specific binding sites for the glycosaminoglycan (GAG) side chains of proteoglycans. Exogenous GAGs should bind to these sites, competitively inhibit interactions with proteoglycans, and perturb development. GAGs added to the culture medium perturb the in situ growth of pioneer axons in cultured cockroach embryos by producing axon defasciculation and growth in incorrect directions. The specificity of this phenomenon is evident from the following observations: Of all the GAGs tested only heparin and heparan sulfate produced perturbation; of the six axon tracts being pioneered during the culture period only two of them are perturbed by the GAGs; and similar perturbations are produced when embryos are cultured in the presence of heparinase II and heparitinase.

A morphological correlate of target recognition by regenerating motor axons in the cockroach.

A specific cell recognition process during regeneration of severed axons of identified cockroach motor neurons eventually leads to the reformation of the original innervation pattern of target muscles in the leg. This occurs even though, at early times after nerve crush, the multiple branches of each regenerating axon grow into both appropriate and inappropriate muscles. In this study, we sought to examine whether there are any structural differences between regenerating axon branches in appropriate and inappropriate muscles that could lead to an understanding of why only those in inappropriate muscles are eliminated. A neuron subset-specific monoclonal antibody, NSS-2A, which labels the inhibitory motor neurons, was used to make their axon branches visible at various times after nerve crush. In inappropriate muscles, these axons grow primarily parallel to the muscle fibers and are later eliminated. In the appropriate muscles, these axon branches initially also grow parallel to the muscle fibers, but subsequently grow many interstitial collaterals. The formation of the collateral branches is a morphological correlate of the specific interaction of a neuron with its appropriate muscle. The simultaneous occurrence of axonal elimination and collateral sprouting supports the idea that the two processes are causally related, as suggested by the sibling neurite bias hypothesis.

The reappearance of a developmental stage-specific antigen in adult regenerating neurons of the cockroach.

A monoclonal antibody has previously been described that binds to all neurons in the 15 d (50% development) cockroach embryo but to only a small subset of neurons in the adult (Denburg et al., 1989). Experiments were performed in order to determine whether the developmental stage-specific antigen recognized by this antibody would reappear in adult neurons that were induced to undergo axonal regeneration by axotomy. It is demonstrated here that after nerve crush motor, sensory and interneurons undergo axonal regeneration and regain their ability to bind this antibody. This indicates that the developing and regenerating states of these neurons selectively use the same molecules to perform apparently similar cellular functions. The increase and subsequent decrease of antibody binding as a function of time after nerve crush was determined for each of these adult neurons. Correlations between the temporal distribution of the antigen and cellular events occurring during axonal regeneration are consistent with a role for this molecule in axon growth and the elimination of inappropriate synaptic connections. The antigen was localized to the external surface of the plasma membrane, and preliminary biochemical characterization has led to the tentative identification of the antigen as a glycolipid. These characteristics distinguish this growth-associated antigen from other previously described molecules whose temporal distribution has implicated a role for them in axon growth.

An axon growth associated antigen is also an early marker of neuronal determination.

We have previously described the generation of a monoclonal antibody (DSS-3) that binds to all neurons in cockroach embryos at 50% development and to only a small subset of interneurons in the adult nervous system. This developmental stage-specific antigen was observed to reappear in all axotomized adult neurons that were undergoing axonal regeneration. In the present study the time course of the appearance of this growth-associated antigen during embryonic development was determined. Unexpectedly, the antigen was observed to be present in embryonic neurons long before axon growth. In addition, all cells in the CNS neuronal lineage (neuroblasts, ganglion mother cells, and neurons) bind the antibody as soon as they can be morphologically identified. However, the antigen is also transiently present in all neuroepithelial cells at a stage prior to the morphological differentiation of some of them to neuroblasts. Analogous patterns of DSS-3 binding to cells involved in the development of sensory neurons and leg pioneer neurons are observed. The DSS-3 antigen is therefore a very early marker for the capacity of ectodermal epithelial cells to develop along a neuronal lineage.

Developmental stage-specific antigens in the nervous system of the cockroach.

A specific effort was made to obtain monoclonal antibodies that bind to macromolecules that play a role in the development of the nervous system. It was considered that good candidates for such molecules were those that were only transiently present in the embryonic nervous system. Hybridomas were prepared from spleen cells taken from mice that had been immunized with nerve cords from cockroach embryos at the 43-50% stage of development. The hybridoma supernatants were screened for antibody binding to frozen sections of both embryonic and adult thoracic ganglia. Cell lines that produced monoclonal antibodies that transiently bound to the embryonic nervous system were saved and cloned. These developmental stage-specific monoclonal antibodies either did not bind to the adult nervous system or bound to it with a pattern very different from that in the embryonic nervous system. The developmental stage-specific antigens detected by these monoclonal antibodies were organized into four categories based on the part of the embryonic nervous system in which they were transiently localized. These include binding to the cell bodies of all neurons, cell bodies of subsets of neurons or neuroblasts, subsets of axons, and the neuropile. Preliminary biochemical characterization of the antigens showed that many of these antibodies were recognizing carbohydrate epitopes. Functions for these antigens, most of which are components of the cell surface, are tentatively proposed.

Absence of competitive interactions among axon terminals of regenerating motor neurons.

Competition among axon terminals is usually considered to contribute to the formation of patterned synaptic connections. During axonal regeneration of motor neurons in the cockroach, leg muscles initially become innervated by appropriate and inappropriate motor neurons. All axon terminals from inappropriate neurons eventually are eliminated, resulting in the reformation of the original innervation pattern. Destruction of an identified motor neuron by the intracellular injection of pronase did not prevent the elimination of inappropriate axon terminals in the muscle normally innervated by that motor neuron. Therefore, competition does not play a role in the reinnervation of the leg muscles. This indicates a major role for specific cell-cell recognition.

Differences in surface molecules of motor axon terminals correlated with cell-cell recognition.

The cell-cell interactions leading to the formation of synaptic connections among cells in the nervous system may be mediated by cell surface macromolecules. In the cockroach the specific reformation of the original innervation pattern of a set of leg muscles during axonal regeneration indicates a significant contribution from cell-cell recognition. Macromolecules mediating such a process would be expected to be distributed differentially among the axon terminals of the various motor neurons. Monoclonal antibodies have been isolated that selectively bind to the surfaces of axon terminals of some motor neurons and not others. Preliminary biochemical characterization indicates that these antigens are glycoproteins and are good candidates for consideration as recognition macromolecules.

Monoclonal antibodies to the cockroach nervous system.

In the cockroach nervous system individual motor neurons may be identified with respect to their position in the thoracic ganglia and to the muscles they innervate. When their axons are cut they have the ability to regrow such that when regeneration is completed they have specifically reinnervated their normal target muscles. This suggests the existence of a specific intercellular recognition process between motor neurons and muscles, and that neurons innervating different muscles are biochemically distinct from one another. The goal of this study was to use hybridoma techniques to obtain monoclonal antibodies that bind to some motor neurons and not others. Mice were injected with whole nerve cord and hybridoma supernatants were screened immunohistochemically on sections of ganglion and leg muscle. The monoclonal antibodies were categorized according to four types of specificity: tissue, regional, cell-type, and neuron-subset specificities. Antibodies expressing neuron-subset specificity were obtained very rarely. The probability of their occurrence could be increased by treating the mice with immunosuppressant drugs after initial administration of immunogen or by fixing the immunogen with paraformaldehyde in a manner similar to that of the tissue sections used in the screening process. Two of the neuron-subset specific monoclonal antibodies (MAbs) are of particular interest with respect to the goals of this study. They bind to axon terminals in the muscles of some neurons and not others. They do not bind to neuronal cell bodies in the ganglion, which makes identification of the neurons difficult. However, from the known innervation pattern of the coxal depressor muscles it appears that one of these MAbs selectively binds to axon terminals from either the inhibitory motor neurons or the dorsal unpaired median cells. Other antibodies of interest bind selectively to the synapse-rich neuropile in the ganglia or to peripheral parts of the nervous system like the nerve roots.

Plasticity in the cockroach neuromuscular system.

The retrograde transport of wheat germ agglutinin-conjugated horseradish peroxidase extracellularly injected into a leg muscle was used to identify the regenerating cockroach motor neurons that have grown an axonal branch into that muscle. At least 66% of the animals with crushed nerve roots eventually reform the original innervation pattern of this muscle with no mistakes. In spite of this apparent specificity the cockroach neuromuscular system can express plasticity as evidenced by the correction of mistakes made at early stages of regeneration. These mistakes are corrected through elimination during the time interval between 40 and 60 days after nerve crush. In addition, when the distal segments of the leg are removed, thus depriving some motor neurons of their normal target muscles, many of them form stable inappropriate axonal branches in denervated as well as fully innervated muscles. These observations are discussed in terms of possible mechanisms responsible for the specificity of the cellular interactions and in terms of their relevance to understanding the development of vertebrate neuromuscular systems.

Lectin receptors in the cockroach neuromuscular system. I. Distribution among the set of coxal depressor muscles.

Fluorescent derivatives of plant lectins have been used to determine if there are biochemical differences among the cell surfaces of 6 muscles in the leg of the cockroach innervated in a fixed pattern by two identified motor neurons, Df or Ds. Histochemical analysis of the binding of fluorescein isothiocyanate conjugates of the lectins to frozen sections of muscles has demonstrated that the surfaces of muscles innervated by motor neuron Ds have more alpha-N-acetylgalactosamine and/or D-galactose than do those of muscles innervated by motor neuron Df. Biochemical analysis of the glycoprotein lectin receptors by sodium dodecylsulphate polyacrylamide gel electrophoresis has shown that approximately 10% of them are distributed among the various muscles in a manner that correlates with innervation by Df or Ds. Some of these macromolecules may be responsible for the biochemical differences in the muscle cell surfaces that could be specifically recognized by motor neurons. This intercellular recognition could mediate the reformation of the original innervation pattern during axonal regeneration.

Lectin receptors in the cockroach neuromuscular system. II. Effects of denervation on muscle lectin receptors.

The binding of fluorescent derivatives of plant lectins to the 6 coxal depressor muscles in the leg of the cockroach was examined at 1, 2, 3 and 5 weeks after denervation. Histochemical analysis of the binding of the lectins to frozen sections of the muscles demonstrated that all detectable binding was to the surfaces of the muscle fibers. In addition, within one week of denervation those muscles which in the innervated state have different amounts of D-galactose and/or alpha-N-acetylgalactosamine on their surfaces, now all have identical amounts of these carbohydrates. Biochemical analysis of concanavalin A and wheat germ agglutinin glycoprotein receptors by sodium dodecylsulphate polyacrylamide gel electrophoresis indicated that approximately 17% of all the receptors detected altered their relative levels in denervated muscles. These changes were observed in those receptors that were present in equal amounts in each of the muscles as well as in those receptors that were distributed among the 6 muscles in a manner that correlated with innervation by an identified motor neuron. However, in spite of these changes that reduced the biochemical differences in the denervated CDMs it was still possible to distinguish among the 6 muscles by the nature of the Con A and WGA glycoprotein receptors. Denervated muscles still present biochemically different cell surfaces to regenerating motor neurons.

Lectin receptors in the cockroach neuromuscular system. III. Distribution and identification within the thoracic ganglia of the nervous system.

A fluorescence microscopic study of the binding of an array of 10 fluorescein isothiocyanate conjugated lectins to frozen sections of the cockroach thoracic ganglia was performed. Although a region of the neuropile receiving direct input from sensory neurons was observed to have distinctive lectin binding properties, no difference was detected between the lectin binding properties of Df and Ds, the two identified motor neurons that innervate the coxal depressor muscles in the leg. In addition, the three types of neurons identified in the ganglia, excitatory motor neurons, inhibitory motor neurons and dorsal unpaired median cells all had identical lectin binding properties. Therefore, in order to demonstrate the existence of macromolecules responsible for giving a biochemical identity to the various neurons it will be necessary to perform biochemical analyses of single cells or apply immunological techniques. A biochemical analysis of the ganglionic Con A and WGA receptors detected after fractionation by SDS polyacrylamide gel electrophoresis revealed the presence of some receptors not found in extracts of muscles and which may be specific for ganglia. In addition, a set of ganglionic lectin receptors were detected which were of similar solubility and lectin binding properties to a set of receptors found in muscle extracts and which disappeared from the muscles within one week of their denervation. It is suggested that such lectin receptors are found in all motor neurons and are transported to axon terminals within the muscles.