Numerous investigations of pathfinding by embryonic neurons, including many leech neurons, have demonstrated that pathfinding is precise. Neurons project the correct number of growth cones which grow to their target areas by making specific choices along the way. However, one leech mechanosensory neuron, the dorsal P (PD) neuron, is unusual in the sense that it initially projects an excessive number of growth cones. One of the growth cones will form the peripheral axon while the others are eliminated. This suggests that PD is one of the earliest neurons to project a peripheral axon, i.e., it may pioneer a peripheral nerve, and that it searches with its multiple growth cones for an external cue which can guide it to its target area. Examination of the early PD axon with light and electon microscopy reveals that it indeed is the first growth cone in its nerve and that it grows in contact with a large non-neuronal (DV) cell until it reaches its target area. The DV cell has a unique morphology and location: a large cell body with thin flat processes extending from the edge of the ganglion to the target area of the PD. It is also present with its unique morphology prior to axonal outgrowth by the PD neuron. These features suggest that the DV cell may be an attractive substrate and/or axonal guidance cue for the PD peripheral axon and therefore for the entire peripheral nerve.
The analysis of how neurons make specific connections with target cells has benefited from investigations of axonogenesis by identifiable embryonic neurons (for reviews see Goodman et al. 1984; Goodman & Bastiani, 1985). These studies demonstrated that for the most part embryonic neurons do not randomly project an excessive number of axons and then eliminate the inappropriate ones. Instead, they project the appropriate number of growth cones which grow to their targets by making a series of decisions at specific choice points. Choices appear to be made by utilizing gradients of adhesivity (Nardi, 1983; Berlot & Goodman, 1984) and by specific cellular cues. Cellular cues can be non-neuronal cells (Ho, Ball & Goodman 1983) or other neurons both in the peripheral nervous system (Ho & Goodman, 1982 ; Bentley & Keshishian, 1982; Bentley & Caudy, 1983; Berlot & Goodman, 1984) and in the central nervous system (CNS) (Raper, Bastiani & Goodman 1983a,b,c; Taghert, Bastiani, Ho & Goodman 1982; Bastiani & Goodman, 1984; Raper, Bastiani & Goodman 1984; Bastiani, Raper & Goodman 1984).
An exception to the orderly projection and growth of embryonic axons is the initial stages of the development of the primary peripheral axon of the dorsal pressure (PD) neuron (Kuwada, 1982; Kuwada & Kramer, 1983), a mechanosensory neuron found in the CNS of the leech (Nicholls & Baylor, 1968; Kramer & Goldman, 1981). Early in embryogenesis the PD neuron projects numerous peripherally directed growth cones. When the peripheral growth cones are still in the vicinity of the ganglion (far from the primary target area), one of them rapidly grows directly to the proper target area while the others are eliminated. This sequence of events suggests that the growth cones are sampling the environment for a cue which could direct a growth cone to the proper target area (Kuwada, 1982; Kuwada & Kramer, 1983). This phenomenon has been investigated further by marking embryonic PD neurons and examining them with light and electron microscopic techniques. These data demonstrate that the PD neuron is the first neuron, i.e., a pioneer neuron (Bate, 1976), in its peripheral nerve and its primary peripheral axon grows along an identifiable non-neuronal cell to get to its target area.
MATERIALS AND METHODS
Embryos of the leech Haementaria ghilianii were collected from a laboratory breeding colony. Descriptions of the growth, reproduction, and care of the adult leeches can be found in Sawyer, LePont, Stuart & Kramer (1981) and of the embryos in Kuwada & Kramer (1983). The dissection of the embryos and the procedures for the physiological and light level anatomical investigations of the embryonic cells have been previously enumerated by Kuwada & Kramer (1983). In short, embryos were dissected, pinned out in a recording chamber, and mounted on a fixed-stage compound microscope outfitted for Nomarski interference contrast optics and epifluorescence illumination. Cells were visualized in living embryos with Nomarski optics and subsequently penetrated with a microelectrode to record their activity and to fill them with the fluorescent dye, Lucifer Yellow (Stewart, 1978).
For the electron microscopic analysis of early axons, neurons were injected with horseradish peroxidase (HRP) (4% dissolved in 100 mM-KCl at pH 7·4) by passing 1–2 nA of depolarizing current for 5 to 15 minutes from HRP-filled microelectrodes measuring 100 to 250 megohms in resistance. Embryos were then fixed in a solution containing 2% glutaraldehyde and 2% paraformaldehyde in 100 mM-phosphate buffer at pH 7·4 for 1 h at room temperature. Following three washes (20 min each) in the phosphate buffer, embryos were incubated for 20 min in the developing solution containing 0·5mg/ml of 3,3 diaminobenzidine, 2mg/ml of beta-D-glucose, and 0·5 mg/ml of ammonium chloride in phosphate buffer. The HRP reaction was initiated by adding 0-5 unit of glucose oxidase per ml of developing solution and allowed to develop for a to 1 h. The embryos were washed three times for 30 min each in phosphate buffer and then postfixed in 1% osmium tetroxide in phosphate buffer on ice for 1 h. Following three more washes (10 min each) in cold phosphate buffer, embryos were dehydrated, embedded in plastic, and serially sectioned for transmission electron microscopy.
Criteria utilized to stage the embryos and descriptions of the embryos at various stages can be found in Kuwada & Kramer (1983). The stage of an embryo was designated by the stage number followed by, in parentheses, the ratio of the number of days elapsed at 27 °C since the beginning of the stage to the total number of days in the stage. A stage 9(3/4) embryo has completed the first 3 days of stage 9.
The leech is segmented and includes 21 essentially identical midbody segments. Each segment contains a ganglion which is linked to the periphery by a set of nerves; successive ganglia are linked by connectives (Muller, Nicholls & Stent 1981). The skin of the leech is innervated by a number of mechanosensory neurons whose cell bodies reside in the ganglia (Nicholls & Baylor, 1968; Kramer & Goldman, 1981). The primary peripheral axon of the dorsal pressure-sensitive neuron (PD) mainly innervates half the dorsal skin of a segment ipsilateral to the PD cell body via the dorsal posterior (DP) nerve in Haementaria ghilianii (Kramer & Kuwada, 1983).
The primary peripheral axon of the PD is one of the earliest peripheral axons to develop (Kuwada & Kramer, 1983). Initially (stage 9(2/4)) the PD neuron projects up to five peripherally directed growth cones which can radiate anteriorly, laterally, and posteriorly but rarely towards the ventral midline and they extend up to 50 μm from the cell body (Fig. 1A). However, approximately 1 day later a single primary peripheral axon, presumably developed from one of the initial neurites, is quite evident with its growth cone growing directly to the dorsal germinal plate which will later develop into the dorsal body wall and skin (Fig. IB). Later the other initial peripheral neurites are eliminated (Kuwada & Kramer, 1983; Fig. 3A).
The pathway of the primary growth cone of all PD neurons is stereotyped from segment to segment and from animal to animal. This, in conjunction with the early behaviour of the numerous initial peripheral growth cones, suggests that the PD neuron establishes a primary peripheral axon by utilizing a specific substrate or guidance cue: one of the early growth cones finds an attractive substrate or cue which will lead it to the appropriate target area of the germinal plate (nascent body wall and skin) while the other growth cones are eliminated (Kuwada, 1982; Kuwada & Kramer, 1983). The pathway of the primary peripheral growth cone entails a number of cells and structures which may serve as an appropriate substrate and/or pathfinding cue (Kuwada, 1982). One of the most salient, the dorsoventral (DV) cell, has a relatively large, ovoid cell body (5 μm wide and 10 μm long) located approximately half way between the ganglion and the future lateral edge (the boundary between the ventral and dorsal halves of the leech) (Fig. 3). The DV cell has a complex network of flat, thin processes approximately 30 μm wide and 100–150 μm long, which appear to insert into the ventral germinal plate near the ganglion, emerge onto the inner surface of the ventral germinal plate approximately 20–30 μgm from the ganglion, and insert into the dorsal germinal plate near the future lateral edge. Thus, the DV cell spans the territory which the primary peripheral growth cone of the PD neuron must cross to get to its target area (Fig. 3B). Not only does the PD axon run right by (Fig. 2C) and just under the DV cell body, but it can be seen to run along the entire length of the DV cell’s flat processes (Fig. 3). The morphology of the DV cell is so unique that it is easily identifiable in sections viewed with the electron microscope (Fig. 4). In addition to its prominent cell body, it has a network of processes 50 nm to 1 μm in thickness which cover and demarcate a column of cells in the area where the flattener muscles will develop later. In the adult leech the dorsal posterior (DP) nerve, which contains the primary axon of the PD neuron, is found embedded in the flattener muscles (Kuwada & Kramer, 1983). Additionally, the DV cell can be found in its normal position with its unique morphology at stage 9(1-5/4) when the PD neuron has not yet projected any peripheral growth cones (Fig. 7B).
In three cases the PD neuron from a midbody ganglion was filled with HRP in stage 9(2·5–3/4) embryos, when the primary peripheral axon is first evident, for examination with the electron microscope. In each case the growth cone of the PD neuron and distal portions of the primary axon were more distal than any other growth cones or axons in the DP pathway. The PD growth cone is, therefore, the earliest growth cone to traverse this pathway (Fig. 5A,6A). Other DP axons fasciculated with the PD axon more proximally (Fig. 6C). Axons were easy to distinguish from other cellular profiles, some of which belonged to the DV cell, because of the regular array of microtubules which characterized the axons. In those same sections, the HRP filled PD growth cone and axon were associated and in contact with the processes and cell body of the DV cell (Fig. 5,6). The DV cell processes were identified by following them from the DV cell body in serially sectioned preparations (n=2).
The DV cell has a resting potential of approximately −35 mV, input resistance of 2 megohms, and linear membrane characteristics (Fig. 8). The lack of electrogenic properties in addition to its unusual morphology suggests that the DV cell is a non-neuronal cell. Furthermore, it is dye coupled (Lucifer Yellow) to many of the cells found along the DP pathway which are covered by the DV processes (Fig. 7A), but was never found to be dye coupled to the PD neuron nor any other neuron.
Pathfinding by the primary peripheral growth cone of the PD neuron
Analyses of axonal growth by identifiable embryonic neurons have established that neurons make specific connections with each other by utilizing various external cues (Raper et al. 1983a,b,c; Ho et al. 1983; Bentley & Caudy, 1983; Berlot & Goodman, 1984; Raper et al. 1984; Bastiani et al. 1984). Likewise the manner in which the primary peripheral axon of the PD neuron is established suggests that axonal guidance cues may be involved: an initial projection of an excessive number of peripheral growth cones followed by the rapid growth of one of the growth cones directly to its target area and the elimination of the others (Kuwada, 1982; Kuwada & Kramer, 1983). Consistent with this hypothesis are the two main findings of this investigation. 1) The primary growth cone of the PD neuron is the first growth cone to traverse the DP pathway and, therefore, establishes the DP nerve. 2) Its growth cone and axon grow in contact with the processes of the large, non-neuronal DV cell. Furthermore, the DV cell possesses a unique morphology which makes it especially suitable as a substrate and/or guidance cue and is present with its unique morphology in its characteristic position before the PD neuron has projected any peripheral growth cones. These characteristics suggest that the DV cell may be involved in the guidance of the pioneer PD growth cone.
In the embryonic grasshopper nervous system neurons project the correct number of growth cones and they seem to make pathfinding decisions on the basis of selective filopodial contact (Raper et al. 1983c; Taghert et al. 1982; Bastiani & Goodman, 1984; Bastiani et al. 1984). Many neurons in the embryonic leech also project the correct number of growth cones which proceed to grow in a stereotyped and regular fashion (Kuwada & Kramer, 1983; Kramer & Kuwada, 1983; Kuwada, 1984). The establishment of the primary peripheral axon of the PD neuron is unusual because an excessive number of growth cones are initially projected. Furthermore, the other peripheral axons of the PD neuron, which are projected later, and the central axon of the PD neuron, which is projected before any peripheral growth cones, are all established by extension of a single growth cone (Kuwada & Kramer, 1983; Kramer & Kuwada, 1983; Kuwada, 1984). The reason for the difference between the manner in which the PD primary peripheral axon and many other embryonic axons are established is not known. One possibility is that the guidance cue is beyond reasonable filopodial reach. Therefore, the projection of multiple growth cones may be necessary in order for filopodial contact with the appropriate cue. This point of view appears to conflict with the notion that the DV cell is the guidance cue since its processes are found in the vicinity of the ganglion prior to extension of peripheral growth cones by the PD neuron. However, these processes appear to be inserted into the germinal plate or developing body wall near the ganglion and emerge onto the inner surface of the nascent body wall approximately 20–30 μm lateral to the ganglion. Thus, the closest DV processes may not be readily accessible to the filopodia of the PD neuron. The DV processes more lateral are accessible but may be far enough from the ganglion to require projection of multiple growth cones in order to insure filopodial contact.
Is the DV cell a muscle pioneer?
The embryonic grasshopper contains large mesodermal cells known as muscle pioneers with morphologies similar to the DV cell (Ho et al. 1983). Muscle pioneers apparently function as a scaffold for organizing specific muscles and as axonal guidance cues for outgrowing motor neurons which normally innervate those muscles. It is possible that the DV cell may function as a muscle pioneer in the embryonic leech for a number of reasons. (1) The DV cell is probably of mesodermal origin since the Lan 3–14 monoclonal antibody, which recognizes leech muscles (Zipser & McKay, 1981; Stuart, Thompson, Weisblat & Kramer, 1982), recognizes the DV cell (unpublished results). (2) It occupies the location where the dorsoventral flattener muscles will develop and covers and demarcates a column of cells some of which may develop into the flattener muscles. (3) It may serve as an axonal guidance cue for the pioneering PD growth cone and the DP nerve. Growth cone guidance by mesodermal derivatives both in the CNS and periphery has been previously suggested in the embryonic leech (Kuwada, 1982; Kuwada & Kramer, 1983; Kramer & Kuwada, 1983). This suggestion awaits experimental confirmation, and the large and conspicuous nature of the DV cell makes it attractive for future manipulations.
This research was supported by NSF (SPI80–09177) and NIH (NS 06551) postdoctoral fellowships to the author and a NSF research grant (BNS 79–23459) to W. B. Kristan, Jr. I thank M. Bastiani and W. B. Kristan, Jr. for critically reading the manuscript and F. Thomas for help with some of the EM micrographs. I am especially indebted to M. Bastiani for his expert help with the electron microscopy.