pre-axonogenesis migration of afferent pioneer cells in the grasshopper embryo

THE JOURNAL OF EXPERIMENTAL ZOOLOGY 251:217-223 (1989)

Pre-Axonogenesis Migration of Afferent Pioneer Cells in the Grasshopper Embryo

DAVID BENTLEY AND ALMA TOROIAN-RAYMOND Department of Molecular and Cell Biology, University of California, Berkeley, California 94720

ABSTRACT In insects, afferent neurons arise primarily from the ectodermal epithelium in the periphery and differentiate at the site of their precursor mitosis. Here we describe ectodermally derived cells that migrate away from their site of origin and initiate axonogenesis a t a distant location. In embryonic grasshopper limb buds, the first two pairs of afferents to differentiate are the pair of Ti1 pioneers at the limb tip and the pair of Cxl cells found at the base of the limb. While the Ti1 pioneers arise from the mitosis of a pioneer mother cell at the limb tip, the Cxl cells are shown to emerge from the epithelium at circumferential positions that are approximately 150" apart and that belong to different embryonic compartments. The cells migrate into contact with each other before initiating axonogenesis, and their axons then extend in a new direction that is orthogonal to the route of cell migration.

While independent migrations of individual cells to specific target areas are an important fea- ture of vertebrate embryogenesis, there has been little evidence that this is a significant mechanism in insect development. In the ectodermal epithelium, for example, early marking of epithelial progenitors has revealed very little mixing between clones (Campos Ortega and Harten- stein, '85). Recently, however, experiments in which specific mesodermal cells, termed muscle pioneers, are recognized early in development have shown that these cells do migrate within developing limbs and that these migrations have a major impact on subsequent development of muscles (Ho et al., '83; Ball et al., '85). We have been interested in the development of the peripheral nervous system and have encoun- tered an instance of migration of ectodermally derived cells.

With the exception of a very small number of central cells (Braunig and Hustert, 'SO), insect sensory neurons arise in the periphery from the ectodermal epithelium (Wigglesworth, '53). De- pending on cell type, they are derived by a specific sequence of cell divisions from epithelial precur- sors (Bate, '78) and extend axons to the CNS with- out migrating from their site of origin (Kamper and Murphey, '87). The routes of these afferent axons are established very early in embryogenesis by the pathfinding axons of pioneer neurons (Bate, '76; Ho and Goodman, '82; Keshi- shian and Bentley, '83). Thus the point from which these pioneer cells arise or begin to extend

0 1989 ALAN R. LISS, INC.

their axons is a critical determinant of the subsequent location and route of the peripheral nerves.

In each of the thoracic limb buds there are two pairs of pioneer cells that appear at the 30% stage of development, well before any other afferent cells. One of these pairs (Til) lies at the appendage tip and comprises the first cells to extend axons (Bate, '76). These cells are siblings and arise from an ectodermal pioneer mother cell found at the appendage tip (Keshishian, '80; Lef- cort and Bentley, '$9). The other cell pair is found at the base of the appendage and has been called the 1A cells (Ho and Goodman, '82) or Cxl cells (Keshishian and Bentley, '83; Caudy and Bentley, '86). These cells extend axon-like processes to the CNS. The relative timing of Ti1 and Cxl axonogenesis varies between limbs in an anterior- posterior gradient: in the prothoracic limb, and usually in the mesothoracic limb, the Cxl cells begin axonogenesis before the arrival of the Ti1 growth cones, whereas in the metathoracic limb, the Ti1 growth cones are the first to reach the CNS (Ho and Goodman, '82). In the work reported here, we have investigated the origin and axonogenesis of the Cxl cells.

MATERIALS AND METHODS Schistocerca americana embryos at the 30-33%

stage of development (Bentley et al., '79; Caudy and Bentley, '86) were obtained from a colony

Received January 6, 1989; revision accepted March 3, 1989.

218 D. BENTLEY AND A. TOROIAN-RAYMOND

maintained at Berkeley. To examine normal development, embryos were dissected, fixed in 4% formaldehyde, labeled with serum antibodies against horseradish peroxidase (anti-HRP; protocol in Caudy and Bentley, '86), whole mounted, viewed, and photographed in a Zeiss Universal epifluorescence microscope (n > 100). These antibodies selectively label grasshopper neurons (Jan and Jan, '82; Snow et al., '87).

To delay the progress of Cxl cell migration with respect to the acquisition of anti-HRP binding sites, embryos were dissected under sterile condi- tions and cultured for 30-36 hr at pH 7.0, 31 ? 1°C in a COz incubator (5% COB) in RPMI- supplemented saline solution (Lefcort and Bent- ley, '87) (n > 100). To disrupt microfilaments, saline contained 0.1 pgiml dihydrocytochalasin B (n > 251, 2 kg/ml cytochalasin B (n > 251, or 0.05 kg/ml cytochalasin D (Bentley and Toroian- Raymond, 1986) (n > 25); to disrupt or to stabilize microtubules, saline contained 0.1 pg/ml colchicine (n > 25), or 7 x lo-' M taxol (n > 251, respectively.

To determine the circumferential location of Cxl neurons, embryos were cultured 24 hr at 30 k 1°C in RPMI saline with 0.05 kg/ml cytochalasin D, fixed in 4% formaldehyde, and labeled with rabbit anti-HRP primary antibody (as above), biotin goat anti-rabbit secondary antibody (Vec- tastain A-B reagent), and peroxidase-conjugated avidin, and then treated with diaminobenzidine. The tissue was dehydrated, mounted in soft Eponi Araldite, serially sectioned (10 pm sections) per- pendicular to the limb axis, and viewed and photographed with Nomarski optics on a Zeiss microscope (n = 10).

To examine the location of the posterior limb compartment cells, we used a monoclonal antibody (gift of N. Patel and C.S. Goodman) against an epitope in the homeodomain of engrailed and inuected protein (labeling protocol in Patel et al., in preparation; see also Kornberg et al., '85). Staining was intensified by including 8% NiC12/ 10 ml diaminodenzidine (n = 9). For double- labeling, anti engrailedlinuected and anti-HRP primary antibodies were processed in the same solution (n = 1). Preparations were imaged on a BioRad laser scanning confocal microscope.

RESULTS In 30-31% stage prothoracic limb buds, anti-

HPR antibodies, which bind a neural-specific carbohydrate moiety (Jan and Jan, '82; Snow et al., ,871, label two pairs of afferent cells, the Ti1

pioneers at the limb tip and the Cxl cells at the base (Fig. 1D). When first evident, the Cxl cells are usually in close proximity and, while they often have an apron of lamellipodia and filopodia, have not begun axonogenesis. In occasional embryos, the Cxl cells are separated in the anterior- posterior axis, extend processes toward each other, and have a "dumb-bell" configuration (Fig. 1C). This suggests that the cells may have sepa- rate origins and may migrate into contact.

The frequency with which this configuration is observed, and the distance between the cells, can be increased by removing embryos from the egg at the 28-29% stage and culturing them in a supplemented RPMI medium (Fig. 1C). This treatment appears to delay cell migration with respect to acquisition of anti-HRP binding sites. In extreme cases, two cells are seen to lie at the opposite sides of the limb bud, to be circumferentially elongated, and to be extending processes along the limb circumference toward each other (Fig. 1B).

Migration of the cells can be further delayed, or blocked, by culturing in the microfilament- disrupting agents cytochalasin B, dihydrocytochalasin B, or cytochalasin D. Under these con- ditions, two cells are seen to label within the ectodermal epithelium, one at the anterior face of the limb and one at the posterior face of the limb (Fig. 1A). The posterior Cxl cell labeling in Fig- ure 1A has its circular apical endfoot still inserted in the external face of the epithelium. Thus both of these cells appear to be derived from the epithelium.

The circumferential locus of the cells can be determined more precisely by examining limb cross sections containing migration-delayed cells (Fig. 2C). The cells arise from sites that are approximately 150" apart on the limb circumference (Fig. 2C). The anterior cell emerges from the epithelium before the posterior cell (Fig. ZC), ac- quires anti-HRP binding before the posterior cell, and often appears to migrate slightly farther around the limb circumference than the posterior cell. The cells encounter each other near the ventral midpoint of the limb (Fig. 2D).

Grasshopper limb buds are divided into anterior and posterior epithelial compartments that can be delineated by an antibody against an epitope in the homeodomain of the engrailed and inuected protein (Patel et al., in preparation; see also Kornberg et al., '85). Labeling early limb buds with this antibody (Fig. 2A,B), or double- labeling with anti-HRP (Fig, ZD), indicates that the site from which the anterior cell arises is in

Fig. 1. Migration and axonogenesis of Cxl cells. Anti-HRP antibody labeled, whole-mounted prothoracic limb buds. A: Pre- migration: the plane of focus is at the apical surface of the posterior limb epithelium, where the posterior Cxl cell still has its apical endfoot inserted (solid arrow). The anterior Cxl cell (open arrow) is out of focus at the anterior surface of the limb. The Ti1 pioneer neurons can be seen at the limb tip (asterisk). Embryo cultured in 0.1 pgiml dihydrocytochalasin B for 36 hr at 32°C. B: Initial stage of migration: the anterior (open arrow) and posterior (solid arrow) Cxl cells are migrating toward each other around the limb circumference. A prominent process leads the anterior cell, which has several filopodia inserted within the epithelium. The Ti1 pioneer neurons (asterisk) at the limb tip have not begun axonogenesis. Embryo cultured for 6 hr in supplemented RPMI medium. C: Completion of migration: major processes of the anterior and posterior Cxl cells are in contact, although the cell bodies axe still about two cell diameters apart. Embryo cultured for 6 hr in supplemented RPMI medium. D: Cell-cell apposition: the Cxl cells have completed migration and have established a broad zone of apposition but have not initiated axonogenesis. Embryo fixed immediately after dissection. E: Initiation of axonogenesis: the Cxl cells are now oriented along the limb axis, rather than circumference, and are extending prominent filopodia (large arrowhead) toward the CNS. The Ti1 pioneer neurons are at about the 33-34% stage of axonogenesis, and filopodia (small arrowhead) from one Ti1 growth cone are in contact with the Cxl cells. Embryo fixed immediately after dissection. F Axonogenesis: major processes (large arrowhead) extend from the Cxl cells to the border of the CNS, where they encounter emerging efferent growth cones. The Ti1 pioneer neurons are at the 34% stage, and their growth cones (small arrowhead) are in contact with the Cxl cells. Embryo fixed at dissection. Size calibrations: 50 Fm,


Fig. 2. Circumferential and compartmental location of Cxl cells. A Dorsal aspect of whole-mounted prothoracic (left) and mesothoracic (right) limbs buds labeled with antibody against engruiled and inuected protein and imaged in a confocal laser scanning microscope. The optical section is equatorial in the prothoracic leg and tangential to the dorsal epithelium in the mesothoracic leg. The antibody labels the posterior compartment (arrowheads). B: Ten micrometer sections through the prothoracic (left), mesothoracic (middle), and metathoracic (right) limb buds of a 32% stage embryo labeled with antibody against engruiled and invected protein. The meso- and metathoracic sections include a portion of ventral body wall. In the limb and body wall, the posterior compartment is labeled (arrowheads). C: Ten micrometer cross sections through the prothoracic (left) and mesothoracic (right) limb buds of an approximately 30% stage embryo a t the level of the Cxl cells, labeled with anti- HPR antibody. In the prothoracic limb, an anterior Cxl cell has almost completed emergence from the epithelium (open arrow) and is migrating circumferentially toward the ventral midline. The posterior Cxl cell (solid arrow) is still fully within the epithelium with apical and basal endfeet. In the mesothoracic limb, the anterior Cxl cell (arrowhead) is labeled faintly within the epithelium. Embryo cultured in 0.05 pg/ml cytochalasin D for 24 hr at 30°C. Dorsal, up; anterior, t o left. D: Ten micrometer cross sections through the prothoracic limb bud (right) and the next anterior appendage bud (left) in an approximately 31% stage embryo double-labeled with antibody against the engruiled and inuected protein product and with anti-HRP. In the prothoracic limb, one (or both) Cxl cell (open arrow) lies at a midventral position just anterior to the compartment boundary (arrowheads). In the anterior appendage, the posterior cell (solid arrow) can be seen emerging from the epithelium of the posterior compartment (arrowheads). Dorsal, up: anterior, to left. Sections in B and D viewed with transmission fiber optics on a confocal microscope. Size calibrations: 50 km.

the interior of the anterior limb compartment and that the site from which the posterior cell arises is in the interior of the posterior compartment. The two cells meet close to the ventral compartment boundary.

When the circumferentially migrating cells encounter each other, they establish an extensive zone of apposition (Fig. 1D) and extend a halo of radial filopodia (Fig. 1E). Subsequently, filopodia that are longer and straighter than other filopodia are seen extending proximally along the ventral midline toward the CNS (see also Caudy and Bentley, '86). Such filopodia are not seen extending distally along the midline. The cell bodies also reorient from elongation along the limb cir-

cumference (Fig. 1B) to slight elongation along the limb axis (Fig. lE,F). Axon-like processes then emerge from the cells and extend in a straight axial path to the CNS (Fig. 1F).

The cessation of Cxl cell migration and initiation of axonogenesis might be caused by a variety of possible mechanisms, including contact between the two Cxl cells or contact with some fea- ture of the ventral axial pathway to the CNS. We investigated these alternatives by delaying or blocking Cxl migration with cytoskelton-altering agents. When migration is delayed by disrupting microtubules with colchicine (Fig. 3A) or stabiliz- ing microtubules with taxol (Fig. 3B), the Cxl cells can be prevented from arriving at their nor-

MIGRATION OF PIONEER CELLS 221

Fig. 3. Reaction of Cxl cell processes to the circumferential and limb axis pathways. Anti-HRP antibody labeled, whole- mounted embryos. A: Migration of the anterior (open arrow) and posterior (solid arrow) Cxl cells has been delayed by culture in colchicine-containing medium (0.1 Fgiml; 30 hr; 32°C). Although the cells are not closely apposed, both cells have extended filopodia (arrowheads) on the axial pathway to the CNS. B: Migration of the Cxl cells has been delayed by culture in taxol(7 x 10:'; 30 hr; 32°C). The cells have not reached close apposition or their normal location for axonogenesis, and they have not been contacted by filopodia of the Ti1 pioneer neurons. However, both cells are extending long filopodia (arrowheads) along the axial axonogenesis pathway to the CNS. C: Migration of both Cxl cells has been blocked by culture with cytochalasin B (2 )*g/ml; 36 hr). The posterior cell (solid arrow) has extended a long process with a growth-cone-like enlargement at its end along the circumferential migration route. D: Migration of both C x l cells has been blocked by culture in cytochalasin medium (as in C). The anterior cell (open arrow) has extended a long process, tipped by as growth-cone-like enlargement, which has reached the position of the axial axon pathway. At this point, the growth cone is bifurcating and extending a lamellipodium proximally (arrowhead) along the axon route. Size calibrations: 50 pm.

ma1 midventral axonogenesis site and also from establishing close cell-cell apposition. In this situation, long filopodia emerge from both cells and extend toward (Fig. 3A) or to (Fig. 3B) the CNS along the midventral axial route. These filopodia never extend straight proximally from the two separated cell bodies, but only extend along the normal path of axonogenesis. This suggests the presence of a specific, local pathway that is recognized by filopodia, even in the absence of close Cxl cell apposition.

In the presence of cytochalasin D, migration of both Cxl cells can be prevented (Fig. 3C,D). In this situation, either the posterior or the anterior cell may extend an axon-like process tipped by a

growth-cone-like enlargement. These processes extend circumferentially along the route on which the cell would normally migrate. If such a process reaches the point of the midventral axial pathway, the growth cone can bifurcate and extend a lamellipodium along this route (projec- tion of filopodia from these cells is blocked in the presence of cytochalasin; Bentley and Toroian- Raymond, '86), even in the absence of contact with the other Cxl cell.

DISCUSSION In grasshopper embryo limb buds, afferent

nerve routes are established by the axons of a set of pioneer neurons (Bate, '76; Ho and Goodman,


'82; Bentley and Keshishian, '83). The earliest pioneer cells are two pairs, the Ti1 cells located at the limb tip and the Cxl cells located at the base of the limb. As is usual with insect peripheral neurons, the Ti1 cells are siblings and arise at the site of axonogenesis by the division of an epithelial mother cell (Keshishian, '80; Lefcort and Bentley, '89). Surprisingly, the Cxl cell pair does not arise in this conventional fashion. Rather, the cells emerge separately from the epithelium at two sites separated by approximately 150" of circumferential arc and then migrate around the limb circumference to a new site near the ventral midline where they initiate axonogenesis.

Cell migration and growth cone extension are closely related processes (Bray and White, 'SS), but the signals that terminate locomotion of mi- gratory neurons and initiate axonogenesis are not well understood. In this system, three events that might act as signals for the time or place of axonogenesis occur almost simultaneously: 1) arrival of each migrating Cxl cell at the ventral midline, 2) contact between the Cxl cells, and 3) contact of the Cxl cells by the arriving Ti1 growth cones. Observation of normal embryogenesis eliminates the last possibility, as the Cxl cells can begin axonogenesis before arrival of the Ti1 growth cones (Ho and Goodman, '82; Caudy and Bentley, '86). The site of Cxl axonogenesis might be determined by the circumferential location at which the migrating cells collide. By preventing cel1:cell collision with taxol or colchicine, our results suggest that this probably is not the case. Rather, the route along which filopodia or lamellipodia are preferentially extended appears to be independent of the location of the C x l cells and to be established by other local features near the ventral midline. Unidentified anti-HRP labeling cells occasionally seen at this location may be in- volved in this process (Caudy and Bentley, '86). Our results are most consistent with the hy- pothesis that the change from migration to axonogenesis is at least partially caused by a local change in extracellular signals at the junction of the circumferential migration route and the midventral axonogenesis route.

The placement of the Cxl cells at their postmi- gratory location is important for the further development of the peripheral nervous system. The growth cones of the Ti1 pioneer neurons that are migrating proximally in the limb are blocked by the differentiating coxa-trochanter limb segment boundary (Caudy and Bentley, '87). The growth

cones extend processes circumferentially along this boundary and are able to cross the boundary only when their filopodia contact the C x l cells (Bentley and Caudy, '83; Caudy and Bentley, '86). Thus, the final location of the Cxl cells appears critical for the circumferential placement of the major peripheral nerve, 5B1, pioneered by the Ti1 neurons.

Cxl cells are found in each limb bud, and simi- lar cells are found in the mouthpart appendages and in the antennae. They form a class of cells derived from the ectodermal epithelium, which appear to undertake precisely targeted individual migrations of significant distance through the embryo. Although few in number, these migrating cells have a disproportionate impact on embryogenesis, because, at least in the case of Cxl cells, their positioning is a major determinant of the routing of peripheral nerves. In the mesoderm, muscle pioneers comprise another small class of cells whose individual migrations have a major impact upon subsequent organization of the embryo (Ball et al., '85). These observa- tions suggest that targeted, individual cell migration should be viewed as a significant mechanism in insect development.

ACKNOWLEDGMENTS We thank C.S. Goodman and N. Pate1 for the

gift of anti engrailedlinvected antibody and assis- tance in its use. Support was provided by NIH Jacob Javits award (NS09074-19) and the March of Dimes Birth Defects Foundation (1-1089).

LITERATURE CITED Ball, E.E., R.K. Ho, and C.S. Goodman (1985) Muscle develop-

ment in the grasshopper embryo. I. Muscles, nerves and apodemes in the metathoracic leg. Dev. Biol., 111;383-398.

Bate, C.M. (1976) Pioneer neurones in an insect embryo. Na- ture, 260.5446.

Bate, C.M. (1978) Development of sensory systems in arthropods. In: Handbook of Sensory Physiology. Volume IX: Development of Sensory Systems. M. Jacobson, ed. Springer Verlag, Berlin, pp. 1-53.

Bentley, D., H. Keshishian, M. Shankland, and A. Toroian- Raymond (1979) Quantitative staging of embryonic development of the grasshopper, Schistocercu nitens. J. Embryol. Exp. Morphol., M:47-74.

Bentley, D., and M. Caudy (1983) Pioneer axons lose directed growth after selective killing of guidepost cells. Nature, 304:62-65.

Bentley, D., and A. Toroian-Raymond (1986) Disoriented pathfinding by pioneer neurone growth cones deprived of filopodia by cytochalasin treatment. Nature, 323:712-715.

Braunig, P., and R. Hustert (1980) Proprioceptors with central cell bodies in insects. Nature, 283:768-770.

MIGRATION OF PIONEER CELLS 223

Bray, D., and J.G. White (1988) Cortical flow in animal cells. Science, 239:883-887.

Campos-Ortega, J.A., and V. Hartenstein (1985) The Embry- onic Development of Drosophila melanogaster. Springer, Berlin.

Caudy, M., and D. Bentley (1986) Pioneer growth cone steer- ing along a series of neuronal and non-neuronal cues of different affinities. J. Neurosci., 6:1781-1795.

Caudy, M., and D. Bentley (1987) Pioneer growth cone behav- ior at a differentiating limb segment boundary in the grasshopper embryo. Dev. Biol., 119;454-465.

Ho, R.K., and C.S. Goodman (1982) Peripheral pathways are pioneered by an array of central and peripheral neurones in grasshopper embryos. Nature, 297:404-406.

Ho, R.K., E.E. Ball, and C.S. Goodman (1983) Muscle pioneers: Large mesodermal cells that erect a scaffold for developing muscles and motoneurones in grasshopper embryos. Nature, 301:66-69.

Jan, L.Y., and Y.N. J an (1982) Antibodies to horseradish peroxidase as specific neuronal markers in Drosophila and grasshopper embryos. Proc. Natl. Acad. Sci. U.S.A.,

Kamper, G., and R.K. Murphey (1987) Synapse formation by sensory neurons after cross-species transplantation in crickets: The role of positional information. Dev. Biol., 122: 492-502.

Keshishian, H. (1980) The origin and morphogenesis of

79:2700-2704.

pioneer neurons in the grasshopper metathoracic leg. Dev. Biol., 80;388-397.

Keshishian, H., and D. Bentley (1983) Embryogenesis of peripheral nerve pathways in grasshopper legs. Dev. Biol.,

Kornberg, T., I. Siden, P. O’Farrell, and P. Simon (1985) The engrailed locus of Drosophila: In situ localization of tran- scripts reveals compartment-specific expression. Cell, 40: 45-53.

Lefcort, F., and D. Bentley (1987) Pathfinding by pioneer neurons in isolated, opened and mesoderm-free limb buds of embryonic grasshoppers. Dev. Biol., 119:466-480.

Lefcort, F., and D. Bentley (1989) Organization of cyto- skeletal elements and organelles preceding growth cone emergence from an identified neuron in situ. J. Cell Biol.,

Patel, N., K. Coleman, S. Poole, T. Kornberg, and C.S. Good- man Expression of engrailed-related proteins in arthropods, chordates, and annelids (submitted).

Snow, P.M., N.H. Patel, A.L. Harrelson, and C.S. Goodman (1987) Neural-specific carbohydrate moiety shared by many surface glycoproteins in Drosophila and grasshopper embryos. J. Neurosci. 7:4137-4144.

Wigglesworth, V.B. (1953) The origin of sensory neurones in an insect, Rhodnius prolixus (Hemiptera). Q. J. Microsc. Sci., 94:93-112.

96:98- 124.

108:1737-1749.

pre-axonogenesis migration of afferent pioneer cells in the grasshopper embryo

Documents