THE JOURNAL OF COMPARATLVE NEUROLOGY 303329-337 (1991)
First Step of SelediveMotoneuronkonal Growth:.SelectiveOutgmmthat Discrete Sitesin the Spinal Cord HJDEAKITANAKA Department of Pharmacology, Gunma University School of Medicine, Maebashi 371, Japan
ABSTRACT Selective axonal growth at a series of choice points along pathways is essential for the establishment of precise motoneuron projections. To reveal some of the molecules responsible for this selective growth of motoneuron axons, this study investigates the phenomenon of why motoneurons extend axons outside the spinal cord, whereas interneurons do not. Axonal growth in the chick embryonic spinal cord at early stages of development was examined immunohistochemically. MAb SC1 staining of serial sections selectively revealed the entire distribution of motoneuron axons in the embryo. In the cervical segments, some motoneurons, called dorsal motoneurons, extended axons outside the cord via the dorsal root entry zone. The axons of both dorsal and ventral motoneurons were distributed only in the anterior half of the sclerotome; inside the cord motoneuron cell bodies and axons were distributed evenly along the anterior-posterior axis. Motoneurons and interneurons extended axons during the same period, and although the growth cones of both were intermixed in the same locations, only motoneuron axons grew out from the cord. The outgrowth of all st 19 motoneurons from the cord through either the dorsal or ventral exit zone strongly suggests that extension of their axons outside the cord is a selective rather than random process. Key words: motoneuron, chick embryo spinal cord, immunohistochemistry, monoclonal antibody
One of the major goals of developmental neurobiology is to elucidate the molecular mechanism of selective synapse formation. In the neuromuscular system, predetermined motoneurons selectively innervate unique target muscles. Since muscles can be widely separated, selective axonal growth of motoneurons toward their appropriate targets is the most important process involved in the establishment of precise motoneuron projections. Avian motoneurons are known to be able to innervate their target muscles even after displacement of motoneurons in the cord, or after muscles in the limb are repositioned by embryonic surgery (Lance-Jones and Landmesser, '80; Laing, '84; Stirling and Summerbell, '85). On the basis of these results, it is strongly suggested that motoneurons actively select their directions of growth toward the target muscles at the plexus in the base of the limb (Landmesser, '84) and also probably at other regions within the limb where pathway decisions must be made (Tosney and Landmesser, '85). At proximal choice points of motoneuron axonal growth in the axolotl, highly directed rather than random growth was also reported by Freeman and Davey ('86). In their study the trunk muscles were innervated by either a dorsal or a ventral ramus from the cord, but axons innervating the limb grew via the ventral ramus only. An approximately
o 1991 WILEY-LISS, INC.
10-fold increase in the ratio of ventral/dorsal ramus axons was counted, suggesting highly directed axonal growth of the limb motoneurons at this point. The first step in the series of pathway choices that must be made by motoneuron axonal growth cones is outgrowth from the cord. Only motoneurons and autonomic preganglion neurons grow out from the cord, whereas the spinal cord interneurons do not. This suggests that motoneurons selectively grow out from the cord and that cord interneurons selectively grow within the cord. Since motoneuron outgrowth occurs throughout the entire length of the spinal cord, the amount of the molecule(s) hypothesized to be responsible for the selective outgrowth would presumably be larger, and therefore more detectable than those at the other choice points of the motoneuron axonal growth pathway. The molecule(s) responsible for selective axonal growth must be localized on the cell surface of motoneurons and on the cell surface or in the extracellular matrix of cells in the pathway (Dodd and Jessell, '88). Therefore some cell Accepted September 14,1990. Address reprint requests to H. Tanaka, Department of Pharmacology, Gunma University School of Medicine, 3-39-22 Showa-machi, Maebashi 371, Japan.
H. TANAKA
330 surface components of motoneurons would be expected to differ from those of interneurons during the period of motoneuron outgrowth from the cord. In fact, such a difference has already been detected using a monoclonal antibody (MAb) SC1. MAb SC1 uniquely binds to the cell surface of motoneurons and autonomic preganglion neurons, but not to interneurons, in the spinal cord of the chick embryo at early developmental stages (Tanaka and Obata, '84). Molecular mechanisms of differences in axonal growth among motoneuron subpopulations would appear to be more difficult t o analyze at the present time than differences between motoneurons and cord interneurons, since virtually no difference in cell surface antigens between subpopulations of motoneurons has been described, with the possible exception of M7412-antigen. M7412-antigen is a membrane molecule that is heavily expressed in limb motoneurons, but only slightly expressed in axial motoneurons (Tanaka et al., '89). In order to discover some of the molecules responsible for the selective axonal growth of motoneurons, I propose to investigate why motoneurons grow out from the cord, whereas interneurons do not. Here I report the histological features of motoneuron axonal growth in avian embryos during early stages of development.
MATEXTALSANDMETHODS Immunohistochemistry
Fig. 1. Shows a 58-hour chick embryo spinal cord reproduced from Ramon y Cajal ('29) (loc. cit. Fig. 43). The cells are representative of cells in 3 successive sections. m: motoneuron, dm: dorsal motoneuron, in: interneuron, dr: dorsal root, vr: ventral root.
Chick embryos at different developmental stages (HamRetrogade labelingof motoneuronswith the burger and Hamilton, '51) were fixed with ice-cold 10% formalin in PBS for 1-6 hours, and then kept in cold (4°C) fluorescentdye, DiI 20-30% sucrose in PBS. In order to examine longitudinal Chick embryos at stage 36 were placed in a bath of sections at various anterior-posterior levels simulta- oxygenated Tyrode's solution, decapitated, eviscerated, and neously, the normally flexed embryos were pinned out in an a ventral laminectomy was performed. After pressure injecextended position on cardboard and then fixed. The fixed tion of the fluorescent dye, DiI (l,l'-dioctadecyl-3,3,3',3'embryos retained an extended configuration. Frozen sec- tetramethylindocarbocyanine perchlorate; Molecular tions 10-14 Fm thick were cut on a cryostat microtome and Probes, Inc.), through a glass pipette into the muscles and mounted on gelatin-coated slides. Some of the chick em- dorsal root ganglia of the cervical region (Honig and Hume, bryos were fixed and dehydrated in acetone at -80°C for '86; Tanaka, '87), the embryos were incubated for 5-12 several days and then incubated in xylene for 30 minutes at hours at 26-32°C in oxygenated Tyrode's solution. The room temperature, followed by embedding in paraffin solution of DiI used for injection was made by dissolving (Tanaka and Obata, '84). Serial paraffin sections 10 pm 2-10 mg of DiI in 1 ml of 100% ethanol. Since DiI is thick were mounted on gelatin-coated slides. Indirect immu- water-insoluble, any dye solution leaking during injection nohistochemistry was performed on these sections at room was easily sucked away with an aspirator and the extent of temperature. The sections were stained with hybridoma nonspecific diffusion was very low. The embryos were fixed supernatant of SC1 or 23C10 (Fujita and Obata, '84) by the with 10%formalin in PBS, cryoprotected with 20% sucrose avidin-biotin-peroxidase complex method (Vectastain ABC in PBS, embedded in OCT compound (Miles), sectioned kit, Vector Labs.). The SC1-antigen is transiently expressed frozen using a cryostat, mounted on gelatin-coated slides, on motoneurons and ventral epithelial cells among cord and then viewed in a dry condition by fluorescence microscells at early developmental stages (Tanaka and Obata, '84). The SC1-antigen purified by immunoafhity is a glycopro- COPY. tein of apparent molecular weight of approximately 100 KD by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel RESULTS electrophoresis) (unpublished). The 23C10-antigen is a Axonal growth of motoneurolls and intracellular component expressed selectively in neuronal interneurons in the chick embryospinalcord cells from very early developmental stages but has not been at early developmental stages analyzed further. Other sections were stained with fluorescein isothiocyRamon y Cajal('29), in describing early axonal growth of anate (F1TC)-conjugated goat antimouse IgG (H + L) (Cap- motoneurons and interneurons in the chick embryo spinal pel), mounted in 50% glycerol in PBS, and viewed by cord (Fig. 11, characterized two types of motoneurons. One epifluorescence. In double staining experiments with MAb type grows out from the ventral root and the other type, 23C10 and a rabbit antiserum against purified SC1- called dorsal motoneurons, grows dorsally inside the cord antigen, rhodamine isothiocyanate (R1TC)-conjugated goat and exits from the dorsal root. Interneurons located in the antirabbit IgG (H + L) (Cappel) was also used for demon- dorsal part of the cord pass the dorsal and ventral root exit strating distribution of SC1-antigen. zones and remain within the cord (Fig. 1).Thus it would
MOTONEURONOUTGROWTHFROMTHECORD
Fig. 2. Early neuronal profiles in the lumbar cord of the chick embryo at stage 18. Frozen sections of lumbar cord were stained with neuron-specific MAb 23C10 (A,B,D) and a rabbit antiserum against purified SC1-antigen (1:1,000dilution) ( C ) .The second antibodies were conjugated to horseradish peroxidase (A, B), RITC (C), and FITC (D). Since the immunoglobulin subtypes of MAbs SC1 and 23C10 are identical, rabbit antiserum was used for the double staining. A. A few motoneurons extend their axons outside the cord (arrows) and interneu-
331
rons also grow ventrally along the lateral margin of the cord. B. Slightly more anterior level than A. More neuronal profiles can be seen than A. C and D are taken from the similar level of B. SC1 antiserum stains epidermis (ep),ventral epithelial cells (ve),notochord (nc),and motoneurons (arrow). This staining pattern is identical with that of MAb SC1 (Tanaka and Obata, '84). Although only a few SC1 (+) motoneurons can be seen in C, many neuronal profiles are seen in D. Arrow in D indicates the same sites of C. Bar 50 pm for A,B, and 40 p m for C. D.
H. TANAKA
332
Fig. 3. Three examples of MAb SC1 staining of the transverse paraffin sections of cervical levels at stage 19. MAb SC1 stains motoneurons (m) and ventral epithelial cells (ve) in the cord (sc) and dorsal root ganglion cells (d). In the cervical segments there are two motoneuron groups. One grows out from the cord via the ventral root
(vr) and the other via the dorsal root (dr). Arrow in A shows the dorsal motoneuron of which the axonal growth trajectory can be followed, Note that no SC1-positive structures can be seen beyond the dorsal exit and inside the cord in A, as well as in other examples (B and C ) . nc: notochord. Bar for A, B, C, 50 Fm.
seem that although motoneurons and interneurons both extend axons during roughly the same period, motoneurons selectively grow out from the cord, whereas interneurons remain inside. However, since the cells shown in Figure 1 were representative of three sections, it was not clear whether the growth cones of both motoneurons and interneurons grew during the same period and were therefore confronted with the same choice. In order to confirm the results in Figure 1and to examine axonal growth at slightly earlier developmental stages than those shown, the chick lumbar cord a t stage 18 was stained with neuron-specific MAb 23C10 and a rabbit polyclonal antibody against purified SC1-antigen (Fig. 2 ) . In the chick embryo at early developmental stages, the antigen recognized by MAb SC1 is expressed on only motoneurons and ventral epithelial cells among the various cord cells, whereas interneurons do not express the antigen (Tanaka and Qbata, '84). Figure 2 shows that only a few motoneurons have extended axons outside the cord a t this time and that many SC1(-) axons, which are presumably interneurons, have started to grow toward the ventral cord. These suggest that the growth cones of both motoneurons and interneurons temporally and spatially coexist at the dorsal or ventral exit points of motoneurons from the cord. Similar early axonal growth of both motoneurons and interneurons has been shown in the mouse embryo by Golgi analysis (Wentworth, '84a,b). In the chick embryo, Yaginuma et al. ('90) recently showed, by using a neuron-specific monoclonal antibody, that many interneurons extended axons along a primitive longitudinal pathway and a circumferential pathway during the same period of motoneuron outgrowth.
Dorsal motOneuPOllS Immunohistochemistry with MAb SC1 on serial sections of young chick embryo reveals the entire distribution of motoneuron axons, in contrast to earlier methods involving the Golgi or other histological techniques. Like those described by Ramon y Cajal ('29) and others (Windle and Orr, '34; Barron, '431, dorsal motoneurons were also revealed by MAb SC1 (Fig. 3). Although MAb SC1 also stained dorsal root ganglion cells and their axons, the earliest dorsal root ganglion cells had just started to extend axons at stage 19 and few of them had reached the cord. The S C l ( + ) fibers at the dorsal root entry zone were considered to be the axons of dorsal motoneurons. These dorsal motoneurons only occurred in the cervical segments (Figs. 3 , 4 ) and were more numerous in anterior levels (Fig. 4). In the transverse plane of the cord, motoneurons were observed to grow out from the cord at very narrow zones of ventral root exit and dorsal root entry from the earliest time. The ventral root exit and dorsal root entry zones are referred to as the ventral and dorsal exit zones, respectively, in this work. Motoneurons virtually never extended axons outside the cord in the regions between these ventral and dorsal exit zones. Furthermore, no motoneurons were observed to grow beyond the dorsal exit zone or to wander within the cord when examined at stage 19 (Fig. 3). All motoneurons grew out from the cord without exception. These histological observations therefore suggest that motoneurons actively select to grow out from the cord at either the ventral or dorsal exit zones or that they are constrained elsewhere. Dorsal motoneurons usually did not grow out from the cord between the ventral and dorsal exit zones, as shown in
MOTONEURON OUTGROWTH FROM THE CORD
333
limb
Fig. 4. Distribution of motoneuron axons in the transverse plane of a chick embryo at stage 19 along the anterior-posterior axis from the upper cervical cord to the lumbar cord. Motoneuron axons were stained with MAb SC1 and traced from camera lucida drawings of serial paraffin sections. A: anterior; P: posterior; D: dorsal; V: ventral directions.
Figure 3. However, a few rare exceptions were observed in chick embryos at stage 19 (Fig. 5). In these cases, the axons of the dorsal motoneurons exited from the cord on the way to dorsal exit zone, even though other axons in the same section grew normally. The existence of these rare exceptions indicates that the sclerotome cells adjacent to the cord between the ventral and dorsal exits have the ability to support motoneuron axonal growth, but that usually motoneurons fail to grow out at such regions. MAb SC1 staining of motoneurons was inadequate to determine if dorsal motoneurons were transient during development or if they still existed in older embryos, since it is transient and disappears at older stages. Therefore, dorsal motoneurons were labeled retrogradely with the fluorescent dye DiI a t stage 36, the time by which almost all naturally occurring motoneuron cell death has ended in the lumbar lateral motor columns (Hamburger, '75). Since DiI was injected into muscles and dorsal root ganglia, the dorsal funiculus was densely labeled. However, sensory fibers projecting into the motor column, which ran ventrally from the medial part of the dorsal funiculus toward the motor column inside the gray matter (Kudo and Yamada, '87; Davis et al., '89), were not labeled. Therefore, the labeled fibers along the boundary between white and gray matters were presumably axons of dorsal motoneurons. In fact, dorsal motoneurons were visible at stage 36 and could be
divided into two groups (Fig. 6). One group had cell bodies mixed among the other motoneurons in the ventrolateral motor column, which extended their axons into the ventral root. The other group was located in the intermediolateral portion of the cord, probably corresponding to the Lenhossek column cells. Such cells located in this portion of the cord and extending their axons dorsally were not detected by MAb SC1 staining of serial sections of chick embryos at stage 19. These cells might then have migrated farther from the ventrolateral motor column, as shown for autonomic preganglion cells (Levi-Montalcini, '50), or might differentiate later than stage 19, or might not express the SC1-antigen. The distribution of these cells along the anterior-posterior axis was not examined. Dorsal motoneuron axons could be followed in a single section (Figs. 3, 6), and appeared to grow directly toward the dorsal side and to maintain their configuration during development.
Motoneuronoutgrowth in the longitudinal Plane Although motoneurons differentiate uniformly along the anterior-posterior axis in each segment, they extended their axons only into the anterior half of the sclerotome (Fig. 7A), as previously shown by Keynes and Stern ('84). The axons of motoneurons opposite the posterior half of the sclero-
H. TANAKA
334
Fig. 5. Rare exceptional outgrowth of dorsal motoneurons. Usually motoneuron outgrowth from the cord is restricted to either the dorsal or the ventral exit. However, a few exceptional outgrowths were observed among serial sections of several chick embryos at stage 19. In two examples of serial sections (A and B or C and D) dorsal motoneu-
ions grow out from the cord at the midpoint between the dorsal and ventral exits (indicated by arrowheads), and these exceptional growth cones join with the other axons via the normal exit. vr: ventral root, nc: notochord. Bar, 100 pm.
tome initially grew out from the cord, but then turned abruptly toward the anterior half of the sclerotome of either the same or the next posterior segment (Fig. 7B, arrowheads). The same trajectory has also been observed in neural crest migration (Teillet et al., '87). Many of the more anterior dorsal motoneurons extend axons into the ascending root of the spinal accessory nerve, whereas the more posterior dorsal motoneurons extend axons into the dorsal roots (Windle and Orr, '34). Such posteriorly located dorsal motoneurons also grow out from the cord into the anterior half of the sclerotome (Fig. 8A,B) as did motoneurons extending into the ventral roots (Figs. 7, BE), whereas inside the cord, motoneuron cell bodies (Fig. 8D,E) and axons of dorsal motoneurons (Fig. 8B-D) were distributed uniformly along the anterior-posterior axis. Dorsal root ganglia were also formed in the anterior half of the sclerotome (Fig. 8B-D).
observed to wander inside the cord or to grow beyond the dorsal exit zone. These results indicate that motoneurons extend their axons outside the cord selectively rather than randomly, although they extend dendrites inside the cord in the later stages of development. Furthermore, the outgrowth of motoneurons seemed to involve active selection, because interneurons extended their axons within the cord during the same period as motoneurons but did not grow out from the cord. However, since interneurons were not examined with interneuron-specific markers, a possibility remains that some interneurons grew out from the cord. Both ventral and dorsal motoneurons grow out from the cord within very narrow zones in the transverse plane. What is the signal that indicates the exit point? The only known structure to account for this are groups of nucleated cells that breach the basal lamina of the neural tube exclusively in the region of the ventral root, preceding axon outgrowth (Lunn et al., '87). The basal lamina of the neural tube probably plays a role in restricting the site where motoneurons can grow out, because usually motoneuron outgrowth does not occur in the region between the ventral and dorsal exits even though the sclerotome cells adjacent to such a region can support motoneuron growth, as shown in Figure 5 . It is unlikely that the similar cells observed at the ventral exits breached the basal lamina in the exceptional cases shown in Figure 5 . Such rare exceptional outgrowth might occur through disruptions in the basal lamina of the cord. However, breaching of the basal lamina is not sufficient to explain motoneuron outgrowth. It seems more likely that some diffusible signals for outgrowth
DISCUSSION This study has described features of motoneuron outgrowth from the cord using specific monoclonal antibodies SC1 and 23C10 and will serve as background for further investigations of why motoneurons grow out from the cord, whereas interneurons do not. In contrast to Golgi- and silver-impregnation techniques, MAb SC1 staining of serial sections selectively revealed the entire distribution of motoneuron axons and cell bodies in the chick embryo at early developmental stages. All motoneurons observed at st 19 were in the process of growing out from the cord. None were
MOTONEURONOUTGROWTHFROMTHECORD
Fig. 6. Retrograde labeling of dorsal motoneurons with fluorescent dye DiI. The dye intensely stains both ventral and dorsal motoneurons (m), the ventral root (vr) and dorsal funiculus (do. The labeled axons of dorsal motoneurons (arrowheads) can be seen along the boundary between cord cells and spinal tracts. Labeled dorsal motoneurons are divided into two populations (arrows in A). Bar for A, B, 100 pm.
(Tessier-Lavigne et al., '88) originate from the sclerotome or that fixed signals (Harris, '89) are localized around the exits. Since motoneurons grow out from the cord at sites where somites have been removed (Lewis et al., '81; Tosney, '88), the signals for outgrowth might be localized around the exits rather than emitted from somites, and only motoneurons and autonomic preganglion cells might respond to such signals. The axonal trajectory of the dorsal motoneurons, as shown by the arrow in Figure 3A, can be divided into three steps. In the first step dorsal motoneurons must ignore or fail to respond to whatever signal is causing ventral motoneurons to exit via the ventral root. Second, they must grow dorsally along the lateral margin of the spinal cord. At the third step they grow out from the cord via the dorsal exit zone. This suggests that dorsal motoneurons are capable of distinguishing between the signals present at dorsal and ventral exit points. Along the lateral margin of the cord, dorsal motoneurons grow dorsally whereas commissural interneurons grow ventrally, as shown in Figure 1. Using monoclonal antibodies, we have detected distribution gradients of antigens localized at the basal lamina of the neural tube along the ventral-dorsal axis (unpublished). Such antigens might therefore determine the growth directions of dorsal motoneurons and commissural interneurons. In transplantation experiments the brachial cord forms two plexuses at the base of the hindlimb (Narayanan and Hamburger, '71, Straznicky, '83). Brachial motoneurons
335
Fig. 7. Distribution of motoneuron axons and cell bodies in the longitudinal plane of the trunk level. Frozen sections of chick embryo trunk at stages 18 (A)and 17 (B) were stained with MAb SC1 and FITC-conjugated second antibody. A. Motoneuron axons are selectively localized in the anterior half of the sclerotome (st), between the dermomyotome (dm) and spinal cord (sc), in each segment. Since the plane of section was not parallel to the anterior-posterior axis of the embryo, ventral epithelial cells (ve) are seen on the left. B. Axons of the motoneurons opposite the posterior half of the sclerotome (arrowheads) can be seen to turn toward the anterior half of the sclerotome of the next posterior segment. a: anterior, p: posterior. Bar, 100 ym for A, and 70 pm for B.
normally form one plexus at the wing. Furthermore, wing buds transplanted onto the face are innervated by cranial nerves in patterns macroscopically identical with those in the control (Swanson and Lewis, '82). These findings of transplantation experiments indicate that in the limb there are regions where any motoneurons can grow well regardless of their identity. These regions form nonspecific highways for motoneuron axonal growth. However, motoneuron axons show a certain hierarchy for these nonspecific highways. Motoneurons and autonomic preganglion cells alone among the spinal cord neurons extend axons outside the cord. Motoneurons can be divided into two large subpopulations, one consisting of motoneurons that innervate axial muscles and the other consisting of motoneurons that innervate limb muscles. Axial motoneurons are known to have innervation specificity, and they do not grow very well into the limb when it is placed opposite them by embryonic surgery (Lance-Jones and Landmesser, '8 1).Limb motoneurons can be further subdivided into lateral and medial groups based on the cell positions in the lateral motor
H. TANAKA
336
A B
C
D E
I I1
0
Fig. 8. Distribution of motoneuron axons and cell bodies in the longitudinal plane of the cervical spinal cord at stage 19. Frozen sections, of which the planes of section are illustrated at the top right were stained with MAb SC1 by the Vectastain AEK method. Axons of dorsal motoneurons (A and B), ventral roots (El, and dorsal root ganglia (B-E) are selectively localized in the anterior half of the
sclerotome. Inside the cord, axons of dorsal motoneurons (arrows in B-D) and motoneuron cell bodies (D and E) are distributed uniformly along the anterior-posterior axis. The epidermis is also positive for MAb SC1 staining. a: anterior, p: posterior, sc: spinal cord, d: dorsal root ganglion, vr: ventral root, ep: epidermis, m: motoneuron cell bodies. Bar: 100 wm.
column. Lateral and medial motoneurons innervate muscles derived from dorsal and ventral muscle masses, respectively (Landmesser, '78; Hollyday, '80). This innervation specificity is maintained even in abnormal situations such as the supernumerary limb (Hollyday, '81; Ferguson, '83). Furthermore, motoneurons are subdivided into individual motoneuron pools that are predetermined to innervate specific target muscles. At each choice point along nonspecific highways, these subpopulations of motoneurons described above appear actively to select their correct directions. For this selection, specific chemical cues acting as signals for growth direction
along the nonspecific highway are considered to arise. Since motoneuron outgrowth from the cord is the first level of hierarchy in motoneuron axonal growth, I attempt to analyze it in this study. The purpose of this analysis is also to clarify why motoneurons and autonomic preganglion cells alone can enter the nonspecific highway. Molecules located in the anterior half of the sclerotome, such as cytotactin/tenascin (Tan et al., '87; Mackie et al., '88), butyrylcholinesterase (Layer et al., '88) and M7412-antigen (Tanaka et al., '89), are among the candidate molecules that form the nonspecific highway. Since SC1-antigen is located on the cell surface of motoneurons and autonomic pregan-
MOTONETJRON OUTGROWTH FROM THE CORD glion cells during the period of outgrowth from the cord, it can be regarded as a candidate molecule for recognition of the nonspecific highway. However, the present work has also revealed the need for additional differences between dorsal and ventral motoneurons for these to achieve their divergent patterns of exit from the cord.
ACKNOWIXDGMENTS 1 thank Dr. Barbara Fredette and the referee for critical reading and improvement of the manuscript, Dr. S.C. Fujita for supplying MAb 23C10, Ms. A. Agata €or histology, Ms. Y. Roppongi for the photography, and Ms. Y . Aoki for typing the manuscript.
IxlEFL4mcm Barron, D.H. (1943) The early development of the motor cells and columns in the spinal cord of the sheep. J. Comp. Neurol. 78:l-27. Davis, B.M., E. Frank, F.A. Johnson, and S.A. Scott (1989) Development of central projections of lumbosacral sensory neurons in the chick. J. Comp. Neurol. 279556566. Dodd, J., and T.M. Jessell, (1988) Axon guidance and the patterning of neuronal projections in vertebrates. Science 242692499. Ferguson, B.A. (1983) Development of motor innervation of the chick following dorsal-ventral limb bud rotations. J. Neurosci. 3:1760-1772. Freeman, J.M., and D.F. Davey (1986) The precision of pathway selection by developing peripheral axons in the axolotl. J. Embryol. Exp. Morph. 91 :117-134. Fujita, S.C., and K. Obata (1984) Monoclonal antibodies demonstrate regional specificity in the spinal funiculi of the chick embryo. Neurosci. Res. 1:131-148. Hamburger, V. (1975) Cell death in the development of the lateral motor column of the chick embryo. J. Comp. Neurol. 160.335346. Hamburger, V., and H.L. Hamilton (1951) A series of normal stages in the development of the chick embryo. J. Morphol. 88:49-82. Harris, W.A. (1989) Local positional cues in the neuroepithelium guide retinal axons in embryonic Xenopus brain. Nature 339218-221. Hollyday, M. (1980) Organization of motor pools in the chick lumbar lateral motor column. J. Comp. Neurol. 194:143-170. Hollyday, M. (1981) Rules of motor innervation in chick embryos with supernumerary limbs. J. Comp. Neurol. 202:439%465. Honig, M.G., and R.I. Hume (1986) Fluorescent carbocyanine dyes allow living neurons of identified origin to be studied in long-term cultures. J. Cell Biol. 103:171-187. Keynes, R.J., and C.D. Stern (1984) Segmentation in the vertebrate nervous system. Nature 310:786-789. Kudo, N., and T. Yamada (1987) Morphological and physiological studies of development of the monosynaptic reflex pathway in the rat lumbar spinal cord. J. Physiol. 389:441459. Laing, N.G. (1984) Motor innervation of proximally rotated chick embryo wings. J. Embryol. Exp. Morph. 83.213-223. Lance-Jones, C., and L. Landmesser (1980) Motoneurone projection patterns in the chick hind limb following early partial reversals of the spinal cord. J. Physiol. 302581-602. Lance-Jones, C., and L. Landmesser (1981).Pathway selection by embryonic chick motoneurons in an experimentally altered environment. Proc. R. SOC.Lond. B214t19-52. Landmesser, L. (1978) The distribution of motoneurons supplying chick hind limb muscles. J. Physiol. 284:371-389. Landmesser, L. (1984) The development of specific motor pathways in the chick embryo. Trends in Neurosci. 7r336-339. Layer, P.G., R. Alber, and F.G. Rathjen (1988) Sequential activation of
337 butyrylcholinesterase in rostra1 half somites and acetylcholinesterase in motoneurones and myotomes precedinggrowth of motor axons. Development 102:387-396. Levi-Montalcini, R. (1950) The origin and development of the visceral system in the spinal cord of the chick embryo. J. Morphol. 86253-284. Lewis, J . , A. Chevallier, M. Kieny, and L. Wolpert (1981) Muscle nerve branches do not develop in chick wings devoid of muscle. J. Embryol. Exp. Morph. 64911-232. Lunn, E.R., J. Scourfield, R.J. Keynes, and C.D. Stern (1987) The neural tube origin of ventral root sheath cells in the chick embryo. Development 101247-254. Mackie, E.J., R.P. Tucker, W. Halfter, R. Chiquet-Ehrismann, and H.H. Epperlein (1988) The distribution of tenascin coincides with pathways of neural crest cell migration. Development 102:237-250. Narayanan, C.H., and V. Hamburger (1971) Motility in chick embryos with substitution of lumbosacral by brachial and brachial by lumbosacral spinal cord segments. J. Exp. Zool.178:415-432. Ramon y Cajal, S. (1929) Studies on vertebrate neurogenesis. Translated by Guth, L., 1960. Springfield, IL: Charles C. Thomas. Stirling, R.V., and D. Summerbell (1985) The behaviour of growing s o n s invading developing chick wing buds with dorsoventral or anteroposterior axis reversed. J. Embryol. Exp. Morph. 85.251-269. Straznicky, C. (1983) The patterns of innervation and movements of ectopic hindlimb supplied by brachial spinal cord segments in the chick. Anat. Embryol. 167:247-262. Swanson, G.J., and J. Lewis (1982). The timetable of innervation and its controlin the chick wingbud. 3. Embryol. Exp. Morph. 71:121-137. Tan, S.-S., K.L. Crossin, S. Hoffman, and G.M. Edelman (1987)Asymmetric expression in somites of cytotactin and its proteoglycan ligand is correlated with neural crest cell distribution. Proc. Natl. Acad. Sci. USA 84:7977-7981. Tanaka, H. (1987) Chronic application of curare does not increase the level of motoneuron survival-promoting activity in limb muscle extracts during the naturally occurringmotoneuron cell death period. Dev. Biol. 124:347357. Tanaka, H., and K. Obata (1984) Developmental changes in unique cell surface antigens of chick embryo spinal motoneurons and ganglion cells. Dev. Biol. IO6:26-37. Tanaka, H., A. Agata, and K. Obata (1989) A new membrane antigen revealed by monoclonal antibodies is associated with motoneuron axonal pathways. Dev. Biol. 132:419435. Teillet, M.-A,, C. Kalcheim, and N.M. Le Douarin (1987) Formation of the dorsal root ganglia in the avian embryo: Segmental origin and migratory behavior of neural crest progenitor cells. Dev. Biol. 120:329-347. Tessier-Lavigne, M., M. Placzek, A.G.S. Lumsden, J. Dodd, and T.M. Jessell (1988) Chemotropic guidance of developing axons in the mammalian central nervous system. Nature 336:775-778. Tosney, K. (1988) Proximal tissues and patterned neurite oi.tgrowth at the lumbosacral level of the chick embryo: Partial and complete deletion of the somite. Dev. Biol. 127266-286. Tosney, K.W., and L.T. Landmesser (1985) Growth cone morphology and trajectory in the lumbosacral region of the chick embryo. J. Neurosci. 523454358, Wentworth, L.E. (1984a) The development of the cervical spinal cord of the mouse embryo. I. A golgi analysis of ventral root neuron differentiation. J. Comp. Neurol. 222.231-95. Wentworth, L.E. (1984b) The development of the cervical spinal cord of the mouse embryo. 11. A golgi analysis of sensory, commissural, and association cell differentiation. J. Comp. Neurol. 222:96-115. Windle, W.F., and D.W. Orr (1934) The development of behavior in chick embryos: Spinal cord structure correlated with early somatic motility. J. Comp. Neurol. 60:287-307. Yaginuma, H., T. Shiga, S. Hommma, R. Ishihara, and R.W. Oppenheim (1990) Identification of early developing axon projections from spinal interneurons in the chick embryo with a neuron specific P-tubulin antibody: evidence for a new 'pioneer' pathway in the spinal cord. Development 108:705-7 16.
First Step of SelediveMotoneuronkonal Growth:.SelectiveOutgmmthat Discrete Sitesin the Spinal Cord HJDEAKITANAKA Department of Pharmacology, Gunma University School of Medicine, Maebashi 371, Japan
ABSTRACT Selective axonal growth at a series of choice points along pathways is essential for the establishment of precise motoneuron projections. To reveal some of the molecules responsible for this selective growth of motoneuron axons, this study investigates the phenomenon of why motoneurons extend axons outside the spinal cord, whereas interneurons do not. Axonal growth in the chick embryonic spinal cord at early stages of development was examined immunohistochemically. MAb SC1 staining of serial sections selectively revealed the entire distribution of motoneuron axons in the embryo. In the cervical segments, some motoneurons, called dorsal motoneurons, extended axons outside the cord via the dorsal root entry zone. The axons of both dorsal and ventral motoneurons were distributed only in the anterior half of the sclerotome; inside the cord motoneuron cell bodies and axons were distributed evenly along the anterior-posterior axis. Motoneurons and interneurons extended axons during the same period, and although the growth cones of both were intermixed in the same locations, only motoneuron axons grew out from the cord. The outgrowth of all st 19 motoneurons from the cord through either the dorsal or ventral exit zone strongly suggests that extension of their axons outside the cord is a selective rather than random process. Key words: motoneuron, chick embryo spinal cord, immunohistochemistry, monoclonal antibody
One of the major goals of developmental neurobiology is to elucidate the molecular mechanism of selective synapse formation. In the neuromuscular system, predetermined motoneurons selectively innervate unique target muscles. Since muscles can be widely separated, selective axonal growth of motoneurons toward their appropriate targets is the most important process involved in the establishment of precise motoneuron projections. Avian motoneurons are known to be able to innervate their target muscles even after displacement of motoneurons in the cord, or after muscles in the limb are repositioned by embryonic surgery (Lance-Jones and Landmesser, '80; Laing, '84; Stirling and Summerbell, '85). On the basis of these results, it is strongly suggested that motoneurons actively select their directions of growth toward the target muscles at the plexus in the base of the limb (Landmesser, '84) and also probably at other regions within the limb where pathway decisions must be made (Tosney and Landmesser, '85). At proximal choice points of motoneuron axonal growth in the axolotl, highly directed rather than random growth was also reported by Freeman and Davey ('86). In their study the trunk muscles were innervated by either a dorsal or a ventral ramus from the cord, but axons innervating the limb grew via the ventral ramus only. An approximately
o 1991 WILEY-LISS, INC.
10-fold increase in the ratio of ventral/dorsal ramus axons was counted, suggesting highly directed axonal growth of the limb motoneurons at this point. The first step in the series of pathway choices that must be made by motoneuron axonal growth cones is outgrowth from the cord. Only motoneurons and autonomic preganglion neurons grow out from the cord, whereas the spinal cord interneurons do not. This suggests that motoneurons selectively grow out from the cord and that cord interneurons selectively grow within the cord. Since motoneuron outgrowth occurs throughout the entire length of the spinal cord, the amount of the molecule(s) hypothesized to be responsible for the selective outgrowth would presumably be larger, and therefore more detectable than those at the other choice points of the motoneuron axonal growth pathway. The molecule(s) responsible for selective axonal growth must be localized on the cell surface of motoneurons and on the cell surface or in the extracellular matrix of cells in the pathway (Dodd and Jessell, '88). Therefore some cell Accepted September 14,1990. Address reprint requests to H. Tanaka, Department of Pharmacology, Gunma University School of Medicine, 3-39-22 Showa-machi, Maebashi 371, Japan.
H. TANAKA
330 surface components of motoneurons would be expected to differ from those of interneurons during the period of motoneuron outgrowth from the cord. In fact, such a difference has already been detected using a monoclonal antibody (MAb) SC1. MAb SC1 uniquely binds to the cell surface of motoneurons and autonomic preganglion neurons, but not to interneurons, in the spinal cord of the chick embryo at early developmental stages (Tanaka and Obata, '84). Molecular mechanisms of differences in axonal growth among motoneuron subpopulations would appear to be more difficult t o analyze at the present time than differences between motoneurons and cord interneurons, since virtually no difference in cell surface antigens between subpopulations of motoneurons has been described, with the possible exception of M7412-antigen. M7412-antigen is a membrane molecule that is heavily expressed in limb motoneurons, but only slightly expressed in axial motoneurons (Tanaka et al., '89). In order to discover some of the molecules responsible for the selective axonal growth of motoneurons, I propose to investigate why motoneurons grow out from the cord, whereas interneurons do not. Here I report the histological features of motoneuron axonal growth in avian embryos during early stages of development.
MATEXTALSANDMETHODS Immunohistochemistry
Fig. 1. Shows a 58-hour chick embryo spinal cord reproduced from Ramon y Cajal ('29) (loc. cit. Fig. 43). The cells are representative of cells in 3 successive sections. m: motoneuron, dm: dorsal motoneuron, in: interneuron, dr: dorsal root, vr: ventral root.
Chick embryos at different developmental stages (HamRetrogade labelingof motoneuronswith the burger and Hamilton, '51) were fixed with ice-cold 10% formalin in PBS for 1-6 hours, and then kept in cold (4°C) fluorescentdye, DiI 20-30% sucrose in PBS. In order to examine longitudinal Chick embryos at stage 36 were placed in a bath of sections at various anterior-posterior levels simulta- oxygenated Tyrode's solution, decapitated, eviscerated, and neously, the normally flexed embryos were pinned out in an a ventral laminectomy was performed. After pressure injecextended position on cardboard and then fixed. The fixed tion of the fluorescent dye, DiI (l,l'-dioctadecyl-3,3,3',3'embryos retained an extended configuration. Frozen sec- tetramethylindocarbocyanine perchlorate; Molecular tions 10-14 Fm thick were cut on a cryostat microtome and Probes, Inc.), through a glass pipette into the muscles and mounted on gelatin-coated slides. Some of the chick em- dorsal root ganglia of the cervical region (Honig and Hume, bryos were fixed and dehydrated in acetone at -80°C for '86; Tanaka, '87), the embryos were incubated for 5-12 several days and then incubated in xylene for 30 minutes at hours at 26-32°C in oxygenated Tyrode's solution. The room temperature, followed by embedding in paraffin solution of DiI used for injection was made by dissolving (Tanaka and Obata, '84). Serial paraffin sections 10 pm 2-10 mg of DiI in 1 ml of 100% ethanol. Since DiI is thick were mounted on gelatin-coated slides. Indirect immu- water-insoluble, any dye solution leaking during injection nohistochemistry was performed on these sections at room was easily sucked away with an aspirator and the extent of temperature. The sections were stained with hybridoma nonspecific diffusion was very low. The embryos were fixed supernatant of SC1 or 23C10 (Fujita and Obata, '84) by the with 10%formalin in PBS, cryoprotected with 20% sucrose avidin-biotin-peroxidase complex method (Vectastain ABC in PBS, embedded in OCT compound (Miles), sectioned kit, Vector Labs.). The SC1-antigen is transiently expressed frozen using a cryostat, mounted on gelatin-coated slides, on motoneurons and ventral epithelial cells among cord and then viewed in a dry condition by fluorescence microscells at early developmental stages (Tanaka and Obata, '84). The SC1-antigen purified by immunoafhity is a glycopro- COPY. tein of apparent molecular weight of approximately 100 KD by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel RESULTS electrophoresis) (unpublished). The 23C10-antigen is a Axonal growth of motoneurolls and intracellular component expressed selectively in neuronal interneurons in the chick embryospinalcord cells from very early developmental stages but has not been at early developmental stages analyzed further. Other sections were stained with fluorescein isothiocyRamon y Cajal('29), in describing early axonal growth of anate (F1TC)-conjugated goat antimouse IgG (H + L) (Cap- motoneurons and interneurons in the chick embryo spinal pel), mounted in 50% glycerol in PBS, and viewed by cord (Fig. 11, characterized two types of motoneurons. One epifluorescence. In double staining experiments with MAb type grows out from the ventral root and the other type, 23C10 and a rabbit antiserum against purified SC1- called dorsal motoneurons, grows dorsally inside the cord antigen, rhodamine isothiocyanate (R1TC)-conjugated goat and exits from the dorsal root. Interneurons located in the antirabbit IgG (H + L) (Cappel) was also used for demon- dorsal part of the cord pass the dorsal and ventral root exit strating distribution of SC1-antigen. zones and remain within the cord (Fig. 1).Thus it would
MOTONEURONOUTGROWTHFROMTHECORD
Fig. 2. Early neuronal profiles in the lumbar cord of the chick embryo at stage 18. Frozen sections of lumbar cord were stained with neuron-specific MAb 23C10 (A,B,D) and a rabbit antiserum against purified SC1-antigen (1:1,000dilution) ( C ) .The second antibodies were conjugated to horseradish peroxidase (A, B), RITC (C), and FITC (D). Since the immunoglobulin subtypes of MAbs SC1 and 23C10 are identical, rabbit antiserum was used for the double staining. A. A few motoneurons extend their axons outside the cord (arrows) and interneu-
331
rons also grow ventrally along the lateral margin of the cord. B. Slightly more anterior level than A. More neuronal profiles can be seen than A. C and D are taken from the similar level of B. SC1 antiserum stains epidermis (ep),ventral epithelial cells (ve),notochord (nc),and motoneurons (arrow). This staining pattern is identical with that of MAb SC1 (Tanaka and Obata, '84). Although only a few SC1 (+) motoneurons can be seen in C, many neuronal profiles are seen in D. Arrow in D indicates the same sites of C. Bar 50 pm for A,B, and 40 p m for C. D.
H. TANAKA
332
Fig. 3. Three examples of MAb SC1 staining of the transverse paraffin sections of cervical levels at stage 19. MAb SC1 stains motoneurons (m) and ventral epithelial cells (ve) in the cord (sc) and dorsal root ganglion cells (d). In the cervical segments there are two motoneuron groups. One grows out from the cord via the ventral root
(vr) and the other via the dorsal root (dr). Arrow in A shows the dorsal motoneuron of which the axonal growth trajectory can be followed, Note that no SC1-positive structures can be seen beyond the dorsal exit and inside the cord in A, as well as in other examples (B and C ) . nc: notochord. Bar for A, B, C, 50 Fm.
seem that although motoneurons and interneurons both extend axons during roughly the same period, motoneurons selectively grow out from the cord, whereas interneurons remain inside. However, since the cells shown in Figure 1 were representative of three sections, it was not clear whether the growth cones of both motoneurons and interneurons grew during the same period and were therefore confronted with the same choice. In order to confirm the results in Figure 1and to examine axonal growth at slightly earlier developmental stages than those shown, the chick lumbar cord a t stage 18 was stained with neuron-specific MAb 23C10 and a rabbit polyclonal antibody against purified SC1-antigen (Fig. 2 ) . In the chick embryo at early developmental stages, the antigen recognized by MAb SC1 is expressed on only motoneurons and ventral epithelial cells among the various cord cells, whereas interneurons do not express the antigen (Tanaka and Qbata, '84). Figure 2 shows that only a few motoneurons have extended axons outside the cord a t this time and that many SC1(-) axons, which are presumably interneurons, have started to grow toward the ventral cord. These suggest that the growth cones of both motoneurons and interneurons temporally and spatially coexist at the dorsal or ventral exit points of motoneurons from the cord. Similar early axonal growth of both motoneurons and interneurons has been shown in the mouse embryo by Golgi analysis (Wentworth, '84a,b). In the chick embryo, Yaginuma et al. ('90) recently showed, by using a neuron-specific monoclonal antibody, that many interneurons extended axons along a primitive longitudinal pathway and a circumferential pathway during the same period of motoneuron outgrowth.
Dorsal motOneuPOllS Immunohistochemistry with MAb SC1 on serial sections of young chick embryo reveals the entire distribution of motoneuron axons, in contrast to earlier methods involving the Golgi or other histological techniques. Like those described by Ramon y Cajal ('29) and others (Windle and Orr, '34; Barron, '431, dorsal motoneurons were also revealed by MAb SC1 (Fig. 3). Although MAb SC1 also stained dorsal root ganglion cells and their axons, the earliest dorsal root ganglion cells had just started to extend axons at stage 19 and few of them had reached the cord. The S C l ( + ) fibers at the dorsal root entry zone were considered to be the axons of dorsal motoneurons. These dorsal motoneurons only occurred in the cervical segments (Figs. 3 , 4 ) and were more numerous in anterior levels (Fig. 4). In the transverse plane of the cord, motoneurons were observed to grow out from the cord at very narrow zones of ventral root exit and dorsal root entry from the earliest time. The ventral root exit and dorsal root entry zones are referred to as the ventral and dorsal exit zones, respectively, in this work. Motoneurons virtually never extended axons outside the cord in the regions between these ventral and dorsal exit zones. Furthermore, no motoneurons were observed to grow beyond the dorsal exit zone or to wander within the cord when examined at stage 19 (Fig. 3). All motoneurons grew out from the cord without exception. These histological observations therefore suggest that motoneurons actively select to grow out from the cord at either the ventral or dorsal exit zones or that they are constrained elsewhere. Dorsal motoneurons usually did not grow out from the cord between the ventral and dorsal exit zones, as shown in
MOTONEURON OUTGROWTH FROM THE CORD
333
limb
Fig. 4. Distribution of motoneuron axons in the transverse plane of a chick embryo at stage 19 along the anterior-posterior axis from the upper cervical cord to the lumbar cord. Motoneuron axons were stained with MAb SC1 and traced from camera lucida drawings of serial paraffin sections. A: anterior; P: posterior; D: dorsal; V: ventral directions.
Figure 3. However, a few rare exceptions were observed in chick embryos at stage 19 (Fig. 5). In these cases, the axons of the dorsal motoneurons exited from the cord on the way to dorsal exit zone, even though other axons in the same section grew normally. The existence of these rare exceptions indicates that the sclerotome cells adjacent to the cord between the ventral and dorsal exits have the ability to support motoneuron axonal growth, but that usually motoneurons fail to grow out at such regions. MAb SC1 staining of motoneurons was inadequate to determine if dorsal motoneurons were transient during development or if they still existed in older embryos, since it is transient and disappears at older stages. Therefore, dorsal motoneurons were labeled retrogradely with the fluorescent dye DiI a t stage 36, the time by which almost all naturally occurring motoneuron cell death has ended in the lumbar lateral motor columns (Hamburger, '75). Since DiI was injected into muscles and dorsal root ganglia, the dorsal funiculus was densely labeled. However, sensory fibers projecting into the motor column, which ran ventrally from the medial part of the dorsal funiculus toward the motor column inside the gray matter (Kudo and Yamada, '87; Davis et al., '89), were not labeled. Therefore, the labeled fibers along the boundary between white and gray matters were presumably axons of dorsal motoneurons. In fact, dorsal motoneurons were visible at stage 36 and could be
divided into two groups (Fig. 6). One group had cell bodies mixed among the other motoneurons in the ventrolateral motor column, which extended their axons into the ventral root. The other group was located in the intermediolateral portion of the cord, probably corresponding to the Lenhossek column cells. Such cells located in this portion of the cord and extending their axons dorsally were not detected by MAb SC1 staining of serial sections of chick embryos at stage 19. These cells might then have migrated farther from the ventrolateral motor column, as shown for autonomic preganglion cells (Levi-Montalcini, '50), or might differentiate later than stage 19, or might not express the SC1-antigen. The distribution of these cells along the anterior-posterior axis was not examined. Dorsal motoneuron axons could be followed in a single section (Figs. 3, 6), and appeared to grow directly toward the dorsal side and to maintain their configuration during development.
Motoneuronoutgrowth in the longitudinal Plane Although motoneurons differentiate uniformly along the anterior-posterior axis in each segment, they extended their axons only into the anterior half of the sclerotome (Fig. 7A), as previously shown by Keynes and Stern ('84). The axons of motoneurons opposite the posterior half of the sclero-
H. TANAKA
334
Fig. 5. Rare exceptional outgrowth of dorsal motoneurons. Usually motoneuron outgrowth from the cord is restricted to either the dorsal or the ventral exit. However, a few exceptional outgrowths were observed among serial sections of several chick embryos at stage 19. In two examples of serial sections (A and B or C and D) dorsal motoneu-
ions grow out from the cord at the midpoint between the dorsal and ventral exits (indicated by arrowheads), and these exceptional growth cones join with the other axons via the normal exit. vr: ventral root, nc: notochord. Bar, 100 pm.
tome initially grew out from the cord, but then turned abruptly toward the anterior half of the sclerotome of either the same or the next posterior segment (Fig. 7B, arrowheads). The same trajectory has also been observed in neural crest migration (Teillet et al., '87). Many of the more anterior dorsal motoneurons extend axons into the ascending root of the spinal accessory nerve, whereas the more posterior dorsal motoneurons extend axons into the dorsal roots (Windle and Orr, '34). Such posteriorly located dorsal motoneurons also grow out from the cord into the anterior half of the sclerotome (Fig. 8A,B) as did motoneurons extending into the ventral roots (Figs. 7, BE), whereas inside the cord, motoneuron cell bodies (Fig. 8D,E) and axons of dorsal motoneurons (Fig. 8B-D) were distributed uniformly along the anterior-posterior axis. Dorsal root ganglia were also formed in the anterior half of the sclerotome (Fig. 8B-D).
observed to wander inside the cord or to grow beyond the dorsal exit zone. These results indicate that motoneurons extend their axons outside the cord selectively rather than randomly, although they extend dendrites inside the cord in the later stages of development. Furthermore, the outgrowth of motoneurons seemed to involve active selection, because interneurons extended their axons within the cord during the same period as motoneurons but did not grow out from the cord. However, since interneurons were not examined with interneuron-specific markers, a possibility remains that some interneurons grew out from the cord. Both ventral and dorsal motoneurons grow out from the cord within very narrow zones in the transverse plane. What is the signal that indicates the exit point? The only known structure to account for this are groups of nucleated cells that breach the basal lamina of the neural tube exclusively in the region of the ventral root, preceding axon outgrowth (Lunn et al., '87). The basal lamina of the neural tube probably plays a role in restricting the site where motoneurons can grow out, because usually motoneuron outgrowth does not occur in the region between the ventral and dorsal exits even though the sclerotome cells adjacent to such a region can support motoneuron growth, as shown in Figure 5 . It is unlikely that the similar cells observed at the ventral exits breached the basal lamina in the exceptional cases shown in Figure 5 . Such rare exceptional outgrowth might occur through disruptions in the basal lamina of the cord. However, breaching of the basal lamina is not sufficient to explain motoneuron outgrowth. It seems more likely that some diffusible signals for outgrowth
DISCUSSION This study has described features of motoneuron outgrowth from the cord using specific monoclonal antibodies SC1 and 23C10 and will serve as background for further investigations of why motoneurons grow out from the cord, whereas interneurons do not. In contrast to Golgi- and silver-impregnation techniques, MAb SC1 staining of serial sections selectively revealed the entire distribution of motoneuron axons and cell bodies in the chick embryo at early developmental stages. All motoneurons observed at st 19 were in the process of growing out from the cord. None were
MOTONEURONOUTGROWTHFROMTHECORD
Fig. 6. Retrograde labeling of dorsal motoneurons with fluorescent dye DiI. The dye intensely stains both ventral and dorsal motoneurons (m), the ventral root (vr) and dorsal funiculus (do. The labeled axons of dorsal motoneurons (arrowheads) can be seen along the boundary between cord cells and spinal tracts. Labeled dorsal motoneurons are divided into two populations (arrows in A). Bar for A, B, 100 pm.
(Tessier-Lavigne et al., '88) originate from the sclerotome or that fixed signals (Harris, '89) are localized around the exits. Since motoneurons grow out from the cord at sites where somites have been removed (Lewis et al., '81; Tosney, '88), the signals for outgrowth might be localized around the exits rather than emitted from somites, and only motoneurons and autonomic preganglion cells might respond to such signals. The axonal trajectory of the dorsal motoneurons, as shown by the arrow in Figure 3A, can be divided into three steps. In the first step dorsal motoneurons must ignore or fail to respond to whatever signal is causing ventral motoneurons to exit via the ventral root. Second, they must grow dorsally along the lateral margin of the spinal cord. At the third step they grow out from the cord via the dorsal exit zone. This suggests that dorsal motoneurons are capable of distinguishing between the signals present at dorsal and ventral exit points. Along the lateral margin of the cord, dorsal motoneurons grow dorsally whereas commissural interneurons grow ventrally, as shown in Figure 1. Using monoclonal antibodies, we have detected distribution gradients of antigens localized at the basal lamina of the neural tube along the ventral-dorsal axis (unpublished). Such antigens might therefore determine the growth directions of dorsal motoneurons and commissural interneurons. In transplantation experiments the brachial cord forms two plexuses at the base of the hindlimb (Narayanan and Hamburger, '71, Straznicky, '83). Brachial motoneurons
335
Fig. 7. Distribution of motoneuron axons and cell bodies in the longitudinal plane of the trunk level. Frozen sections of chick embryo trunk at stages 18 (A)and 17 (B) were stained with MAb SC1 and FITC-conjugated second antibody. A. Motoneuron axons are selectively localized in the anterior half of the sclerotome (st), between the dermomyotome (dm) and spinal cord (sc), in each segment. Since the plane of section was not parallel to the anterior-posterior axis of the embryo, ventral epithelial cells (ve) are seen on the left. B. Axons of the motoneurons opposite the posterior half of the sclerotome (arrowheads) can be seen to turn toward the anterior half of the sclerotome of the next posterior segment. a: anterior, p: posterior. Bar, 100 ym for A, and 70 pm for B.
normally form one plexus at the wing. Furthermore, wing buds transplanted onto the face are innervated by cranial nerves in patterns macroscopically identical with those in the control (Swanson and Lewis, '82). These findings of transplantation experiments indicate that in the limb there are regions where any motoneurons can grow well regardless of their identity. These regions form nonspecific highways for motoneuron axonal growth. However, motoneuron axons show a certain hierarchy for these nonspecific highways. Motoneurons and autonomic preganglion cells alone among the spinal cord neurons extend axons outside the cord. Motoneurons can be divided into two large subpopulations, one consisting of motoneurons that innervate axial muscles and the other consisting of motoneurons that innervate limb muscles. Axial motoneurons are known to have innervation specificity, and they do not grow very well into the limb when it is placed opposite them by embryonic surgery (Lance-Jones and Landmesser, '8 1).Limb motoneurons can be further subdivided into lateral and medial groups based on the cell positions in the lateral motor
H. TANAKA
336
A B
C
D E
I I1
0
Fig. 8. Distribution of motoneuron axons and cell bodies in the longitudinal plane of the cervical spinal cord at stage 19. Frozen sections, of which the planes of section are illustrated at the top right were stained with MAb SC1 by the Vectastain AEK method. Axons of dorsal motoneurons (A and B), ventral roots (El, and dorsal root ganglia (B-E) are selectively localized in the anterior half of the
sclerotome. Inside the cord, axons of dorsal motoneurons (arrows in B-D) and motoneuron cell bodies (D and E) are distributed uniformly along the anterior-posterior axis. The epidermis is also positive for MAb SC1 staining. a: anterior, p: posterior, sc: spinal cord, d: dorsal root ganglion, vr: ventral root, ep: epidermis, m: motoneuron cell bodies. Bar: 100 wm.
column. Lateral and medial motoneurons innervate muscles derived from dorsal and ventral muscle masses, respectively (Landmesser, '78; Hollyday, '80). This innervation specificity is maintained even in abnormal situations such as the supernumerary limb (Hollyday, '81; Ferguson, '83). Furthermore, motoneurons are subdivided into individual motoneuron pools that are predetermined to innervate specific target muscles. At each choice point along nonspecific highways, these subpopulations of motoneurons described above appear actively to select their correct directions. For this selection, specific chemical cues acting as signals for growth direction
along the nonspecific highway are considered to arise. Since motoneuron outgrowth from the cord is the first level of hierarchy in motoneuron axonal growth, I attempt to analyze it in this study. The purpose of this analysis is also to clarify why motoneurons and autonomic preganglion cells alone can enter the nonspecific highway. Molecules located in the anterior half of the sclerotome, such as cytotactin/tenascin (Tan et al., '87; Mackie et al., '88), butyrylcholinesterase (Layer et al., '88) and M7412-antigen (Tanaka et al., '89), are among the candidate molecules that form the nonspecific highway. Since SC1-antigen is located on the cell surface of motoneurons and autonomic pregan-
MOTONETJRON OUTGROWTH FROM THE CORD glion cells during the period of outgrowth from the cord, it can be regarded as a candidate molecule for recognition of the nonspecific highway. However, the present work has also revealed the need for additional differences between dorsal and ventral motoneurons for these to achieve their divergent patterns of exit from the cord.
ACKNOWIXDGMENTS 1 thank Dr. Barbara Fredette and the referee for critical reading and improvement of the manuscript, Dr. S.C. Fujita for supplying MAb 23C10, Ms. A. Agata €or histology, Ms. Y. Roppongi for the photography, and Ms. Y . Aoki for typing the manuscript.
IxlEFL4mcm Barron, D.H. (1943) The early development of the motor cells and columns in the spinal cord of the sheep. J. Comp. Neurol. 78:l-27. Davis, B.M., E. Frank, F.A. Johnson, and S.A. Scott (1989) Development of central projections of lumbosacral sensory neurons in the chick. J. Comp. Neurol. 279556566. Dodd, J., and T.M. Jessell, (1988) Axon guidance and the patterning of neuronal projections in vertebrates. Science 242692499. Ferguson, B.A. (1983) Development of motor innervation of the chick following dorsal-ventral limb bud rotations. J. Neurosci. 3:1760-1772. Freeman, J.M., and D.F. Davey (1986) The precision of pathway selection by developing peripheral axons in the axolotl. J. Embryol. Exp. Morph. 91 :117-134. Fujita, S.C., and K. Obata (1984) Monoclonal antibodies demonstrate regional specificity in the spinal funiculi of the chick embryo. Neurosci. Res. 1:131-148. Hamburger, V. (1975) Cell death in the development of the lateral motor column of the chick embryo. J. Comp. Neurol. 160.335346. Hamburger, V., and H.L. Hamilton (1951) A series of normal stages in the development of the chick embryo. J. Morphol. 88:49-82. Harris, W.A. (1989) Local positional cues in the neuroepithelium guide retinal axons in embryonic Xenopus brain. Nature 339218-221. Hollyday, M. (1980) Organization of motor pools in the chick lumbar lateral motor column. J. Comp. Neurol. 194:143-170. Hollyday, M. (1981) Rules of motor innervation in chick embryos with supernumerary limbs. J. Comp. Neurol. 202:439%465. Honig, M.G., and R.I. Hume (1986) Fluorescent carbocyanine dyes allow living neurons of identified origin to be studied in long-term cultures. J. Cell Biol. 103:171-187. Keynes, R.J., and C.D. Stern (1984) Segmentation in the vertebrate nervous system. Nature 310:786-789. Kudo, N., and T. Yamada (1987) Morphological and physiological studies of development of the monosynaptic reflex pathway in the rat lumbar spinal cord. J. Physiol. 389:441459. Laing, N.G. (1984) Motor innervation of proximally rotated chick embryo wings. J. Embryol. Exp. Morph. 83.213-223. Lance-Jones, C., and L. Landmesser (1980) Motoneurone projection patterns in the chick hind limb following early partial reversals of the spinal cord. J. Physiol. 302581-602. Lance-Jones, C., and L. Landmesser (1981).Pathway selection by embryonic chick motoneurons in an experimentally altered environment. Proc. R. SOC.Lond. B214t19-52. Landmesser, L. (1978) The distribution of motoneurons supplying chick hind limb muscles. J. Physiol. 284:371-389. Landmesser, L. (1984) The development of specific motor pathways in the chick embryo. Trends in Neurosci. 7r336-339. Layer, P.G., R. Alber, and F.G. Rathjen (1988) Sequential activation of
337 butyrylcholinesterase in rostra1 half somites and acetylcholinesterase in motoneurones and myotomes precedinggrowth of motor axons. Development 102:387-396. Levi-Montalcini, R. (1950) The origin and development of the visceral system in the spinal cord of the chick embryo. J. Morphol. 86253-284. Lewis, J . , A. Chevallier, M. Kieny, and L. Wolpert (1981) Muscle nerve branches do not develop in chick wings devoid of muscle. J. Embryol. Exp. Morph. 64911-232. Lunn, E.R., J. Scourfield, R.J. Keynes, and C.D. Stern (1987) The neural tube origin of ventral root sheath cells in the chick embryo. Development 101247-254. Mackie, E.J., R.P. Tucker, W. Halfter, R. Chiquet-Ehrismann, and H.H. Epperlein (1988) The distribution of tenascin coincides with pathways of neural crest cell migration. Development 102:237-250. Narayanan, C.H., and V. Hamburger (1971) Motility in chick embryos with substitution of lumbosacral by brachial and brachial by lumbosacral spinal cord segments. J. Exp. Zool.178:415-432. Ramon y Cajal, S. (1929) Studies on vertebrate neurogenesis. Translated by Guth, L., 1960. Springfield, IL: Charles C. Thomas. Stirling, R.V., and D. Summerbell (1985) The behaviour of growing s o n s invading developing chick wing buds with dorsoventral or anteroposterior axis reversed. J. Embryol. Exp. Morph. 85.251-269. Straznicky, C. (1983) The patterns of innervation and movements of ectopic hindlimb supplied by brachial spinal cord segments in the chick. Anat. Embryol. 167:247-262. Swanson, G.J., and J. Lewis (1982). The timetable of innervation and its controlin the chick wingbud. 3. Embryol. Exp. Morph. 71:121-137. Tan, S.-S., K.L. Crossin, S. Hoffman, and G.M. Edelman (1987)Asymmetric expression in somites of cytotactin and its proteoglycan ligand is correlated with neural crest cell distribution. Proc. Natl. Acad. Sci. USA 84:7977-7981. Tanaka, H. (1987) Chronic application of curare does not increase the level of motoneuron survival-promoting activity in limb muscle extracts during the naturally occurringmotoneuron cell death period. Dev. Biol. 124:347357. Tanaka, H., and K. Obata (1984) Developmental changes in unique cell surface antigens of chick embryo spinal motoneurons and ganglion cells. Dev. Biol. IO6:26-37. Tanaka, H., A. Agata, and K. Obata (1989) A new membrane antigen revealed by monoclonal antibodies is associated with motoneuron axonal pathways. Dev. Biol. 132:419435. Teillet, M.-A,, C. Kalcheim, and N.M. Le Douarin (1987) Formation of the dorsal root ganglia in the avian embryo: Segmental origin and migratory behavior of neural crest progenitor cells. Dev. Biol. 120:329-347. Tessier-Lavigne, M., M. Placzek, A.G.S. Lumsden, J. Dodd, and T.M. Jessell (1988) Chemotropic guidance of developing axons in the mammalian central nervous system. Nature 336:775-778. Tosney, K. (1988) Proximal tissues and patterned neurite oi.tgrowth at the lumbosacral level of the chick embryo: Partial and complete deletion of the somite. Dev. Biol. 127266-286. Tosney, K.W., and L.T. Landmesser (1985) Growth cone morphology and trajectory in the lumbosacral region of the chick embryo. J. Neurosci. 523454358, Wentworth, L.E. (1984a) The development of the cervical spinal cord of the mouse embryo. I. A golgi analysis of ventral root neuron differentiation. J. Comp. Neurol. 222.231-95. Wentworth, L.E. (1984b) The development of the cervical spinal cord of the mouse embryo. 11. A golgi analysis of sensory, commissural, and association cell differentiation. J. Comp. Neurol. 222:96-115. Windle, W.F., and D.W. Orr (1934) The development of behavior in chick embryos: Spinal cord structure correlated with early somatic motility. J. Comp. Neurol. 60:287-307. Yaginuma, H., T. Shiga, S. Hommma, R. Ishihara, and R.W. Oppenheim (1990) Identification of early developing axon projections from spinal interneurons in the chick embryo with a neuron specific P-tubulin antibody: evidence for a new 'pioneer' pathway in the spinal cord. Development 108:705-7 16.
