- Email: [email protected]
Identified vertebrate neurons that differ in axonal projection develop together
DEVELOPMENTAL
BIOLOGY
118,309-313
(1986)
Identified Vertebrate Neurons that Differ in Axonal Projection Develop Together BRUCE MENDELSON’
AND CHARLES B. KIMMEL
Institute of Neuroscience, University of Oregon, Eugene, Oregon 97403 Received November 27, 1985; accepted in revised form May 15, 1986 The development of identified reticulospinal neurons of the zebrafish (Brachydanio rerio) was studied in order to learn if cell specific differences in axonal projection are correlated with cell specific differences in time of neuronal development. We examined the development of individually known reticulospinal neurons that are located in close proximity in the hindbrain but that project axons to targets on opposite sides of the spinal cord. We observed that these identified neurons are generated together, and that their axons first arrive in the spinal cord together. We suggest that the selection of different axonal pathways by these neurons does not depend on the time that they develop. o 1986 Academic
Press, Inc.
INTRODUCTION
The manner in which growing axons choose specific pathways during development is not well understood. Axons may recognize and grow along “labeled pathways” that lead them to their targets (Singer et al., 1979; Goodman et ah, 1982), or the targets themselves may secrete a tropic substance toward which axons may grow (LeviMontalcini, 1982). Also, differences in time of development may be crucial; only a limited number of pathways may be available at the time an axon is growing (Gottlieb and Cowan, 1972; Horder and Martin, 1978; Rager, 1980). Although studies of the development of large populations of vertebrate neurons have suggested that time of neuronal development may not be critical with respect to the guidance of axons (Lance-Jones and Landmesser, 1980a,b; Frank and Westerfield, 1983), a rigorous test of a timing hypothesis requires examination of specific, identifiable neurons that develop at the same location in the nervous system, but that differ in axonal projection. We have examined the development of identified neurons of the zebrafish (Brachgdanio rerio) that are immediate neighbors in the brain, but whose axons project to targets on opposite sides of the spinal cord. The development of these neurons was studied by combining [3H]thymidine ([3H]TdR) autoradiography with horseradish peroxidase (HRP) histochemistry. We found that the cell types studied, independent of axonal pathway, are born at similar times, and that their axons first extend into the rostra1 spinal cord at similar times. 1To whom correspondence should be addressed: Department of Neurobiology, Anatomy and Cell Science, University of Pittsburgh Medical School, 3550 Terrace St., Pittsburgh, Pa. 15261.
We obtained no evidence that differences in time of development underlie the different axonal pathways. MATERIALS
AND METHODS
Zebrafish (B. rerio) embryos were obtained from a breeding population maintained in the laboratory, and were staged (al0 min) by morphology within 1 hour of fertilization. We use the letters “hr” to denote hours postfertilization with incubation at 28.5”C. Reticulospinal neurons were identified in larvae at 120 hr by filling them retrogradely with HRP (Metcalfe et ah, 1986). Briefly, HRP was applied to a lesion restricted to one side of the spinal cord performed at the sixth spinal segment. After survival times of l-2 hours the tissue was fixed, sectioned at 50 pm on a vibratome, and subsequently processed histochemically for HRP reaction product by either the Hanker Yates (Hanker et al., 1977) or a Diaminobenzidine (Metcalfe, 1985) procedure. The times of origin of these reticulospinal neurons were determined by combining rH]TdR autoradiography with HRP histochemistry as has been previously described (Mendelson, 1986a).Briefly, the yolk sacs of series of embryos were injected with about 5 nl of [3H]TdR ([methypH]thymidine, 77.2 Ci/mmole, New England Nuclear) at times between 5 and 32 hr. At 120 hr the reticulospinal neurons were retrogradely filled with HRP. After HRP histochemistry was performed the tissue was processed autoradiographically for the presence of rH]TdR incorporation (described in detail in Mendelson, 1986a). To determine when reticulospinal neurons grow axons into the rostra1 spinal cord of the embryo, HRP was applied to a lesion of the spinal cord that was performed
309
0012-1606/86 $3.00 Copyright All rights
0 1986 by Academic Press, Inc. of reproduction in any form reserved.
310
DEVELOPMENTAL BIOLOGY
VOLUME 118,1986
loo-130 pm caudal to the level of the somata of the neurons studied, in different groups of animals at different developmental times. The animals were fixed 20 min after the HRP application and the tissue was subsequently processed for HRP reaction product. RESULTS
The cells studied are individually known reticulospinal cells, collectively called MID neurons (Kimmel, 1982; Kimmel et aZ.,1982; Metcalfe et al, 1986). MID neurons are located in bilateral clusters of three cells each, at two axial levels of the hindbrain. The levels are separated by about 30 pm (Fig. 1). In each cluster the axons of neurons termed MiDc cells cross the midline before descending into the spinal cord in the medial longitudinal fasciculus (mlf), while the axons of MiDi cells project ipsilaterally in the mlf to descend into the spinal cord. We previously found that MiDc cells are generated at about 9 hr (Mendelson, 1986a), but the times of origin of MiDi cells were unknown. To determine the time of origin of MiDi cells, [3H]TdR autoradiography was combined with HRP histochemistry, such that MiDi and MiDc cells were both filled with HRP (as in Fig. l), and could be identified unambiguously. We found no differences in the times of origin of these neurons (Fig. 2). Neurons at both the MiD2 and MiD3 axial levels were generated at similar times. In most animals MiDi and MiDc cells were either all labeled, or all unlabeled (as in Fig. 2B), depending on the time of the [3H]TdR injection. In a few cases, however, only some of these cells were labeled, suggesting that the rH]TdR was injected during the time that the MiD
FIG. 1. Reticulospinal neurons of young zebrafish larvae (120 hr), filled with HRP from lesions restricted to the right side of the spinal cord. (A) A camera lucida drawing traced from a horizontal section;
(B and C) Nomarski micrographs of transverse sections. All of these neurons are present in the brain bilaterally, but are labeled on only the left or right sides according to axonal projection. The designation MID codes the location of the neurons studied in the Mi(ddle) D(orsa1) reticular formation of the hindbrain (Kimmel, 1982; Kimmel et al., 1982; Metcalfe et al, 1986). Clusters of neurons are present at two axial levels (2 and 3). MiD2c and MiD3c (collectively MiDc) cells have axons that cross the midline before descending in the medial longitudinal fasciculus (mlf), and are seen on the left side. MiDi cells have axons that descend ipsilaterally in the mlf, and are seen on the right, the side of the lesion. There are two MiDc neurons at each axial level. The MiDc neurons in each cluster are individually recognizable by position and/or axonal course after the decussation (Metcalfe et d, 1986). In this study the MiD2c neurons are treated collectively and only one of the two MiD3c neurons was analyzed. The other MiD3c cell (MiDlcl) was not studied. The MiDi cells are immediate neighbors of the MiDc cells, often lying medially within each cluster. Other reticulospinal cells were filled in these experiments, and serve to control for the specificity of the lesions. For example the cluster of MiV2 neurons (B), known to project axons ipsilaterally into the spinal cord (Kimmel et al., 1982) are labeled on the right. This pattern of labeling is reproducible to a considerable degree of detail in hundreds of examined preparations (Kimmel, 1982; Kimmel et al., 1982; Metcalfe et al, 1986). Scale bars = 25 pm.
311
BRIEF NOTES
A _-l MiDc
o MiDi o MiDc
l Cl
24 Hours
\
0
11 MiD2i
l l n l
Time
8 @a , ea..... 10 of khymidine
ear
/ .
20
.
.
/
’
MiD3i
.
3%
Injection
Dorsal
FIG. 2. Time of origin of the MiD neurons. (A) Relationship between E3HJTdR labeling of MiD neurons and developmental stage. All of these neurons, independent of axonal pathway, are generated together between 8 and 10 hr. The closed circles (from previous studies Mendelson, 1985, 1986a) represent MiDc neurons filled with HRP from bilateral lesions at myotome 15, below the level of termination of MiDi axons (Metcalfe et al, 1986). The squares and open circles represent new data where the MiDc and MiDi cells were identified from unilateral HRPfills at myotome 6. Each point represents the scoring of at least 10 neurons. It was shown earlier that injected [“HJTdR is incorporated into nuclei of S-phase cells within 15-30 min of the time of injection (Mendelson, 1986a). (B) The MiD3 neurons are unlabeled after an injection of [‘H]TdR at 9 hr. MiD3i and MiD3c were identified by filling them with HRP applied unilaterally at Day 5. The nuclei of many hindbrain neurons in the micrograph that are not filled with HRP, but that can be visualized with the use of Nomarski optics, have incorporated [3HJJ!dR and are labeled with silver grains. The background was less than 1 grain/250 pm* and a cell was only considered labeled if it had at least 5 grains over its nucleus, more than 30 times the level of background labeling. The micrograph shows a transverse section of 4 Frn, as has been previously described (Mendelson, 1986a). Scale bar = 25 pm.
neurons were undergoing their last round of DNA replication. In seven such animals one or both MiDc cells were found to be generated earlier than MiDi cells at the same axial level (i.e., the MiDc cells were unlabeled with r3H]TdR and the MiDi cells opposite to them in the
FIG. 3. All of the axons of MiD neurons, independent of their pathways, first appear in the rostra1 spinal cord at 22-24 hr. A, Tracing of retrogradely HRP-filled neurons in a horizontal section. All animals were fixed 20 min after HRP application and the developmental time at fixation is indicated. (B) Nomarski micrograph of MiD2c and MiD2i neurons, filled with HRP at 23 hr, in a transverse section. At this stage the somata of the MiD neurons are adjacent to the brain’s ventral surface (compare with Fig. 1B). The axons of the MiD2c (arrow) and MiD2i (arrowhead) have grown along or near this surface (Mendelson, 1986b). In these studies HRP was applied to a lesion of the spinal cord performed at the third spinal segment, 100 pm caudal to the level of the MiD3 neurons. This procedure was performed at hourly intervals between 18 and 48 hr and more than 10 good examples of HRP staining were examined for each time point (see Mendelson, 1986h for criteria). All four types of MiD neuron were consistently filled when this procedure was performed at or after 26 hr. Scale bars = 25 pm.
312
DEVELOPMENTAL BIOLOGY
brain were labeled). In another four animals just the opposite was observed; MiDi cells were generated before MiDc cells at the same level. We conclude that the order in which these neurons are born varies within a period of l-2 hours, rather than being strictly correlated with their axonal projections. The time that the MiD neurons grow axons was determined by applying HRP to a lesion of the rostra1 spinal cord performed 100 Km caudal to the level of the MiD3 cells, in different groups of animals at different developmental times. We previously found that MiD neurons are retrogradely labeled by this procedure if and only if their axons have already grown to the level of the lesion (Mendelson, 1986b). At the earliest time that any MiD neurons were filled with HRP, between 22 and 24 hr, both MiDi and MiDc neurons were filled (Fig. 3). As in the birthday study, however, there was variability among different animals. The results from animals where the axonal pathways of all of the cell types in question could be followed unambiguously to the site of the HRP application are shown in Table 1. In these animals we observed that the HRP lesion completely transected the spinal cord and thus all of the MiD neurons that had extended axons to or past the lesion site should have been filled with HRP. In some of these animals only MiDc neurons were filled at one level, in others only MiDi cells were filled (as at the MiD3 level in Fig. 3A), and in still others MiDc neurons were filled on one side of the midline and a MiDi was filled on the other (as at the MiD2 level in Fig. 3A). We observed no consistent relationship between the earliest time that a given type of MiD neuron could be filled from the rostra1 spinal cord and either its axonal pathway or its rostrocaudal position. Rather, there appears to be variability in the times that individual types of MiD neurons first extend axons into the rostra1 spinal cord. However, by 26 hr in all animals where the MiD axonal pathways could be traced, both MiDc and MiDi neurons were filled. Thus, the sequence in which MiD neurons extend axons into the rostra1 spinal cord varies over a period of l-2
VOLUME 118, 1986
hours, the same range of variability MiD neuronal birthdays.
observed for the
DISCUSSION
We have shown that neighboring identified neurons whose axons take different pathways to the spinal cord are generated together, and have axons that first arrive in the spinal cord together. Thus we obtained no evidence for a critical role of time of development that would explain the different axonal pathways. Furthermore, in cases where we observed one of the cell types to develop ahead of the other, in some animals the axon of the earlier developing cell crossed the midline, and in other animals it descended into the spinal cord ipsilaterally. This finding further argues against timing differences being critical, and suggests that development of one of the two cell types does not have to precede that of the other. The variability between cell types was observed both for time of generation and for the time the cells could first be retrogradely filled with HRP from the rostral spinal cord. The variability was present in spite of the differences in axonal pathway lengths, of the order of 30 pm, between the neurons at different rostrocaudal levels (MiD2 vs MiD3 cells) and between those taking the crossed versus the ipsilateral pathway (MiDc vs MiDi neurons). In the spinal cord the MiD axons elongate at about 100 pm/hour (Mendelson, 1986b), and assuming the same rate of growth through the hindbrain, the longer pathways would be traversed in a few minutes, a difference we would not have detected. While we did not observe consistent timing differences, our evidence that timing is not important is circumstantial, for the retrograde labeling technique did not allow us to study the cells at the critical time when axonal outgrowth was initiated and when growth cones grew to the point, less than 20 pm away from their somata (see Fig. 3b), where the pathways diverged. For example, it is possible that one of the cell types always initiated axonogenesis first and subsequently the growth cone paused for a variable period while the later growing
TABLE 1 DEVELOPMENT OF AXONAL PROJECTIONS MiDZc” Stage* 22 hr 23 hr 24 hr 25 hr 26-38 hr
3 5 1 0 0
MiD2i
2 3 1 0 0
MiD2c and i
MiD3c
3 10 14 7 33
2 0 1 1 0
MiD3i
0 1 1 0 0
MiD3c and i
5 12 12 6 31
a Number of animals in which only the cell types indicated were filled with HRP at the specified axial level. *Time that HRP was applied to a lesion of the spinal cord performed loo-130 pm caudal to the MID somata in hours postfertilization ’ Total number of animals where the axonal pathways of the MiD cell types could be unambiguously traced to the lesion site.
Total”
10 22 20 10 45
(hr).
BRIEF NOTES
axon caught up. Pauses of growth cones at what could be important points along axonal pathways have been observed before, in grasshoppers (Goodman et al., 1982) and in the zebrafish, where the axons of identified motoneurons charactersitically pause at the boundary between the dorsal and ventral myotome (Eisen et al., 1986). Critical examination of this issue requires labeling neurons orthogradely at the time of axonogenesis. We estimate that, in the absence of axonal pauses, axonogenesis occurs at 20-22 hr, an hour before we first labeled neurons retrogradely. An alternative to a timing hypothesis is that the growth cones of the MiD cells specifically recognize different pathways, either near the cell bodies where they would lead to the ipsilateral versus the contralateral descending tract, or the left versus the right tracts themselves. Cell-specific selection of “labeled pathways” has been proposed as a mechanism for axonal guidance in both invertebrates (Goodman et al., 1982; Raper et al, 1984a,b; Bastiani et al, 1984) and vertebrates (e.g., Eisen et al, 1986;Lance-Jones and Landmesser, 1980a,b). Thus the mechanisms that guide axons specifically towards their targets may be conserved during evolution. Some of this work has appeared previously in the form of an abstract (Mendelson, 1984). We thank Monte Westerfield, Walter Metcalfe and Eric Frank for providing valuable comments concerning this manuscript and Reida Kimmel for her excellent technical assistance. This work was supported by NIH Grant NS 17963 and 5 T32 GM0’7257-07. REFERENCES BASTIANI, M. J., RAPER, J. A. and GOODMAN, C. S. (1984). Pathfinding in neuronal growth cones in grasshopper embryos. III. Selective affinity of the G growth cone for the P cells within the A/P fascicle. J. Neurosci. 42339-2345. EISEN, J., MYERS, P. Z., and WESTERFIELD, M. (1986). Pathway selection by growth cones of identified motoneurons in live zebra fish embryos. Nature 320,269-271. FRANK, E., and WESTERFIELD, M. (1983). Development of sensorymotor synapses in the spinal cord of the frog. J. Physiol. (London) 343,593610. GOODMAN, C. S., RAPER, J. A., Ho, R. K., and CHANG, S. (1982). Pathfinding by neuronal growth cones in grasshopper embryos. In “Developmental Order: Its origins and Regulation” (S. Subtelny and P. B. Green, eds.). pp. 275-316. Liss, New York. GO~LEIB, D. I., and COWAN, W. M. (1972). Evidence for a temporal factor in the occupation of available synaptic sites during development of the dentate gyrus. Brain Res. 41,452-456. HANKER, J. S., YATES, P. E., METZ, C. B., and RUSTIONI, A. (1977). A
313
new, specific, sensitive and non-carcinogenic reagent for the demonstration of horseradish peroxidase. H&&em 9,789-792. HORDER, T. J., and MARTIN, A. C. (1978). Morphogenetics as an alternative to chemospecificity in the formation of nerve connections. Symp. Sot. Exp. Biol 32,275-358. KIMMEL, C. B. (1982). Reticulospinal and vestibulospinal neurons in the young larva of a teleost fish, Brachydanio rerio. Prog. Brain Res. 57,1-23. KIMMEL, C. B., POWELL, S. L. and METCALFE, W. K. (1982). Brain neurons which project to the spinal cord in the young larvae of the zebrafish. J. Camp. Neural. 205,112-127. KIMMEL, C. B., SESSIONS, S. K., and KIMMEL, R. J. (1978). Radiosensitivity and time of origin of the Mauthner neurons in the zebrafish. Dev. Biol. 62,526-529. LANCE-JONES, C., and LANDMESSER, L. (1980a). Motoneurone projection patterns in embryonic chick limbs following partial deletions of spinal cord. J. Physiol (London) 302,559-580. LANCE-JONES, C., and LANDMESSER, L. (1980b). Motoneurone projection patterns in embryonic chick hind limb following partial reversals of the spinal cord. J. Physiol. (London) 302,581-602. LEVI-M• NTALCINI, R. (1982). Developmental neurobiology and the natural history of nerve growth factor. Annu. Rev. Neurosci. 5, 341362. MENDELSON, B. (1984). Is timing important in the guidance of axons of identified neurons? Sot Neurosci A&r. 10,42. MENDELSON, B. (1985). Soma position is correlated with time of development in three types of identified reticulospinal neurons. Dev. Biol 112,489-493. MENnELsON, B. (1986a). Development of reticulospinal neurons of the zebrafish: I. Time of origin. J. Camp. Neural 261, in press. MENDELSON, B. (198613). Development of reticulospinal neurons of the zebrafish: II. Early axonal outgrowth and cell body position. J. C’omp. Neural. 251, in press. METCALFE, W. K. (1985). Sensory neuron growth cones comigrate with posterior lateral line primordial cells in zebrafish. J. Comp. Neural. 238,218-224. METCALFE, W. K., MENDELSON, B., and KIMMEL, C. B. (1986). Segmental homologies among reticulospinal neurons in the hindbrain of the zebrafish larva. J. Camp. Neural. 251, in press. RAGER, G. (1980). Specificity of nerve connections anisms. Trends Neurosci. 3,43-44.
by unspecific mech-
RAPER, J. A., BASTIANI, M. J., and GOODMAN, C. S. (1983). Pathfinding by neuronal growth cones in grasshopper embryos. II. Selective fasciculation onto specific axonal pathways. J. Neurosci 3,31-41. RAPER, J. A., BASTIANI, M. J., and GOODMAN, C. S. (1984). Pathfinding by neuronal growth cones in grasshopper embryos. III. The effects of ablation of the A and P axons upon the behavior of the G growth cone. J. Neurosci. 4,2339-2345. SINGER, M., NORDLANDER, R. H., and EGAR, M. (1979). Axonal guidance during embryogenesis and regneration in the spinal cord of the newt. “The blueprint hypothesis” of neuronal pathway selection. J. Cornp. Neural. 185,1-22.
BIOLOGY
118,309-313
(1986)
Identified Vertebrate Neurons that Differ in Axonal Projection Develop Together BRUCE MENDELSON’
AND CHARLES B. KIMMEL
Institute of Neuroscience, University of Oregon, Eugene, Oregon 97403 Received November 27, 1985; accepted in revised form May 15, 1986 The development of identified reticulospinal neurons of the zebrafish (Brachydanio rerio) was studied in order to learn if cell specific differences in axonal projection are correlated with cell specific differences in time of neuronal development. We examined the development of individually known reticulospinal neurons that are located in close proximity in the hindbrain but that project axons to targets on opposite sides of the spinal cord. We observed that these identified neurons are generated together, and that their axons first arrive in the spinal cord together. We suggest that the selection of different axonal pathways by these neurons does not depend on the time that they develop. o 1986 Academic
Press, Inc.
INTRODUCTION
The manner in which growing axons choose specific pathways during development is not well understood. Axons may recognize and grow along “labeled pathways” that lead them to their targets (Singer et al., 1979; Goodman et ah, 1982), or the targets themselves may secrete a tropic substance toward which axons may grow (LeviMontalcini, 1982). Also, differences in time of development may be crucial; only a limited number of pathways may be available at the time an axon is growing (Gottlieb and Cowan, 1972; Horder and Martin, 1978; Rager, 1980). Although studies of the development of large populations of vertebrate neurons have suggested that time of neuronal development may not be critical with respect to the guidance of axons (Lance-Jones and Landmesser, 1980a,b; Frank and Westerfield, 1983), a rigorous test of a timing hypothesis requires examination of specific, identifiable neurons that develop at the same location in the nervous system, but that differ in axonal projection. We have examined the development of identified neurons of the zebrafish (Brachgdanio rerio) that are immediate neighbors in the brain, but whose axons project to targets on opposite sides of the spinal cord. The development of these neurons was studied by combining [3H]thymidine ([3H]TdR) autoradiography with horseradish peroxidase (HRP) histochemistry. We found that the cell types studied, independent of axonal pathway, are born at similar times, and that their axons first extend into the rostra1 spinal cord at similar times. 1To whom correspondence should be addressed: Department of Neurobiology, Anatomy and Cell Science, University of Pittsburgh Medical School, 3550 Terrace St., Pittsburgh, Pa. 15261.
We obtained no evidence that differences in time of development underlie the different axonal pathways. MATERIALS
AND METHODS
Zebrafish (B. rerio) embryos were obtained from a breeding population maintained in the laboratory, and were staged (al0 min) by morphology within 1 hour of fertilization. We use the letters “hr” to denote hours postfertilization with incubation at 28.5”C. Reticulospinal neurons were identified in larvae at 120 hr by filling them retrogradely with HRP (Metcalfe et ah, 1986). Briefly, HRP was applied to a lesion restricted to one side of the spinal cord performed at the sixth spinal segment. After survival times of l-2 hours the tissue was fixed, sectioned at 50 pm on a vibratome, and subsequently processed histochemically for HRP reaction product by either the Hanker Yates (Hanker et al., 1977) or a Diaminobenzidine (Metcalfe, 1985) procedure. The times of origin of these reticulospinal neurons were determined by combining rH]TdR autoradiography with HRP histochemistry as has been previously described (Mendelson, 1986a).Briefly, the yolk sacs of series of embryos were injected with about 5 nl of [3H]TdR ([methypH]thymidine, 77.2 Ci/mmole, New England Nuclear) at times between 5 and 32 hr. At 120 hr the reticulospinal neurons were retrogradely filled with HRP. After HRP histochemistry was performed the tissue was processed autoradiographically for the presence of rH]TdR incorporation (described in detail in Mendelson, 1986a). To determine when reticulospinal neurons grow axons into the rostra1 spinal cord of the embryo, HRP was applied to a lesion of the spinal cord that was performed
309
0012-1606/86 $3.00 Copyright All rights
0 1986 by Academic Press, Inc. of reproduction in any form reserved.
310
DEVELOPMENTAL BIOLOGY
VOLUME 118,1986
loo-130 pm caudal to the level of the somata of the neurons studied, in different groups of animals at different developmental times. The animals were fixed 20 min after the HRP application and the tissue was subsequently processed for HRP reaction product. RESULTS
The cells studied are individually known reticulospinal cells, collectively called MID neurons (Kimmel, 1982; Kimmel et aZ.,1982; Metcalfe et al, 1986). MID neurons are located in bilateral clusters of three cells each, at two axial levels of the hindbrain. The levels are separated by about 30 pm (Fig. 1). In each cluster the axons of neurons termed MiDc cells cross the midline before descending into the spinal cord in the medial longitudinal fasciculus (mlf), while the axons of MiDi cells project ipsilaterally in the mlf to descend into the spinal cord. We previously found that MiDc cells are generated at about 9 hr (Mendelson, 1986a), but the times of origin of MiDi cells were unknown. To determine the time of origin of MiDi cells, [3H]TdR autoradiography was combined with HRP histochemistry, such that MiDi and MiDc cells were both filled with HRP (as in Fig. l), and could be identified unambiguously. We found no differences in the times of origin of these neurons (Fig. 2). Neurons at both the MiD2 and MiD3 axial levels were generated at similar times. In most animals MiDi and MiDc cells were either all labeled, or all unlabeled (as in Fig. 2B), depending on the time of the [3H]TdR injection. In a few cases, however, only some of these cells were labeled, suggesting that the rH]TdR was injected during the time that the MiD
FIG. 1. Reticulospinal neurons of young zebrafish larvae (120 hr), filled with HRP from lesions restricted to the right side of the spinal cord. (A) A camera lucida drawing traced from a horizontal section;
(B and C) Nomarski micrographs of transverse sections. All of these neurons are present in the brain bilaterally, but are labeled on only the left or right sides according to axonal projection. The designation MID codes the location of the neurons studied in the Mi(ddle) D(orsa1) reticular formation of the hindbrain (Kimmel, 1982; Kimmel et al., 1982; Metcalfe et al, 1986). Clusters of neurons are present at two axial levels (2 and 3). MiD2c and MiD3c (collectively MiDc) cells have axons that cross the midline before descending in the medial longitudinal fasciculus (mlf), and are seen on the left side. MiDi cells have axons that descend ipsilaterally in the mlf, and are seen on the right, the side of the lesion. There are two MiDc neurons at each axial level. The MiDc neurons in each cluster are individually recognizable by position and/or axonal course after the decussation (Metcalfe et d, 1986). In this study the MiD2c neurons are treated collectively and only one of the two MiD3c neurons was analyzed. The other MiD3c cell (MiDlcl) was not studied. The MiDi cells are immediate neighbors of the MiDc cells, often lying medially within each cluster. Other reticulospinal cells were filled in these experiments, and serve to control for the specificity of the lesions. For example the cluster of MiV2 neurons (B), known to project axons ipsilaterally into the spinal cord (Kimmel et al., 1982) are labeled on the right. This pattern of labeling is reproducible to a considerable degree of detail in hundreds of examined preparations (Kimmel, 1982; Kimmel et al., 1982; Metcalfe et al, 1986). Scale bars = 25 pm.
311
BRIEF NOTES
A _-l MiDc
o MiDi o MiDc
l Cl
24 Hours
\
0
11 MiD2i
l l n l
Time
8 @a , ea..... 10 of khymidine
ear
/ .
20
.
.
/
’
MiD3i
.
3%
Injection
Dorsal
FIG. 2. Time of origin of the MiD neurons. (A) Relationship between E3HJTdR labeling of MiD neurons and developmental stage. All of these neurons, independent of axonal pathway, are generated together between 8 and 10 hr. The closed circles (from previous studies Mendelson, 1985, 1986a) represent MiDc neurons filled with HRP from bilateral lesions at myotome 15, below the level of termination of MiDi axons (Metcalfe et al, 1986). The squares and open circles represent new data where the MiDc and MiDi cells were identified from unilateral HRPfills at myotome 6. Each point represents the scoring of at least 10 neurons. It was shown earlier that injected [“HJTdR is incorporated into nuclei of S-phase cells within 15-30 min of the time of injection (Mendelson, 1986a). (B) The MiD3 neurons are unlabeled after an injection of [‘H]TdR at 9 hr. MiD3i and MiD3c were identified by filling them with HRP applied unilaterally at Day 5. The nuclei of many hindbrain neurons in the micrograph that are not filled with HRP, but that can be visualized with the use of Nomarski optics, have incorporated [3HJJ!dR and are labeled with silver grains. The background was less than 1 grain/250 pm* and a cell was only considered labeled if it had at least 5 grains over its nucleus, more than 30 times the level of background labeling. The micrograph shows a transverse section of 4 Frn, as has been previously described (Mendelson, 1986a). Scale bar = 25 pm.
neurons were undergoing their last round of DNA replication. In seven such animals one or both MiDc cells were found to be generated earlier than MiDi cells at the same axial level (i.e., the MiDc cells were unlabeled with r3H]TdR and the MiDi cells opposite to them in the
FIG. 3. All of the axons of MiD neurons, independent of their pathways, first appear in the rostra1 spinal cord at 22-24 hr. A, Tracing of retrogradely HRP-filled neurons in a horizontal section. All animals were fixed 20 min after HRP application and the developmental time at fixation is indicated. (B) Nomarski micrograph of MiD2c and MiD2i neurons, filled with HRP at 23 hr, in a transverse section. At this stage the somata of the MiD neurons are adjacent to the brain’s ventral surface (compare with Fig. 1B). The axons of the MiD2c (arrow) and MiD2i (arrowhead) have grown along or near this surface (Mendelson, 1986b). In these studies HRP was applied to a lesion of the spinal cord performed at the third spinal segment, 100 pm caudal to the level of the MiD3 neurons. This procedure was performed at hourly intervals between 18 and 48 hr and more than 10 good examples of HRP staining were examined for each time point (see Mendelson, 1986h for criteria). All four types of MiD neuron were consistently filled when this procedure was performed at or after 26 hr. Scale bars = 25 pm.
312
DEVELOPMENTAL BIOLOGY
brain were labeled). In another four animals just the opposite was observed; MiDi cells were generated before MiDc cells at the same level. We conclude that the order in which these neurons are born varies within a period of l-2 hours, rather than being strictly correlated with their axonal projections. The time that the MiD neurons grow axons was determined by applying HRP to a lesion of the rostra1 spinal cord performed 100 Km caudal to the level of the MiD3 cells, in different groups of animals at different developmental times. We previously found that MiD neurons are retrogradely labeled by this procedure if and only if their axons have already grown to the level of the lesion (Mendelson, 1986b). At the earliest time that any MiD neurons were filled with HRP, between 22 and 24 hr, both MiDi and MiDc neurons were filled (Fig. 3). As in the birthday study, however, there was variability among different animals. The results from animals where the axonal pathways of all of the cell types in question could be followed unambiguously to the site of the HRP application are shown in Table 1. In these animals we observed that the HRP lesion completely transected the spinal cord and thus all of the MiD neurons that had extended axons to or past the lesion site should have been filled with HRP. In some of these animals only MiDc neurons were filled at one level, in others only MiDi cells were filled (as at the MiD3 level in Fig. 3A), and in still others MiDc neurons were filled on one side of the midline and a MiDi was filled on the other (as at the MiD2 level in Fig. 3A). We observed no consistent relationship between the earliest time that a given type of MiD neuron could be filled from the rostra1 spinal cord and either its axonal pathway or its rostrocaudal position. Rather, there appears to be variability in the times that individual types of MiD neurons first extend axons into the rostra1 spinal cord. However, by 26 hr in all animals where the MiD axonal pathways could be traced, both MiDc and MiDi neurons were filled. Thus, the sequence in which MiD neurons extend axons into the rostra1 spinal cord varies over a period of l-2
VOLUME 118, 1986
hours, the same range of variability MiD neuronal birthdays.
observed for the
DISCUSSION
We have shown that neighboring identified neurons whose axons take different pathways to the spinal cord are generated together, and have axons that first arrive in the spinal cord together. Thus we obtained no evidence for a critical role of time of development that would explain the different axonal pathways. Furthermore, in cases where we observed one of the cell types to develop ahead of the other, in some animals the axon of the earlier developing cell crossed the midline, and in other animals it descended into the spinal cord ipsilaterally. This finding further argues against timing differences being critical, and suggests that development of one of the two cell types does not have to precede that of the other. The variability between cell types was observed both for time of generation and for the time the cells could first be retrogradely filled with HRP from the rostral spinal cord. The variability was present in spite of the differences in axonal pathway lengths, of the order of 30 pm, between the neurons at different rostrocaudal levels (MiD2 vs MiD3 cells) and between those taking the crossed versus the ipsilateral pathway (MiDc vs MiDi neurons). In the spinal cord the MiD axons elongate at about 100 pm/hour (Mendelson, 1986b), and assuming the same rate of growth through the hindbrain, the longer pathways would be traversed in a few minutes, a difference we would not have detected. While we did not observe consistent timing differences, our evidence that timing is not important is circumstantial, for the retrograde labeling technique did not allow us to study the cells at the critical time when axonal outgrowth was initiated and when growth cones grew to the point, less than 20 pm away from their somata (see Fig. 3b), where the pathways diverged. For example, it is possible that one of the cell types always initiated axonogenesis first and subsequently the growth cone paused for a variable period while the later growing
TABLE 1 DEVELOPMENT OF AXONAL PROJECTIONS MiDZc” Stage* 22 hr 23 hr 24 hr 25 hr 26-38 hr
3 5 1 0 0
MiD2i
2 3 1 0 0
MiD2c and i
MiD3c
3 10 14 7 33
2 0 1 1 0
MiD3i
0 1 1 0 0
MiD3c and i
5 12 12 6 31
a Number of animals in which only the cell types indicated were filled with HRP at the specified axial level. *Time that HRP was applied to a lesion of the spinal cord performed loo-130 pm caudal to the MID somata in hours postfertilization ’ Total number of animals where the axonal pathways of the MiD cell types could be unambiguously traced to the lesion site.
Total”
10 22 20 10 45
(hr).
BRIEF NOTES
axon caught up. Pauses of growth cones at what could be important points along axonal pathways have been observed before, in grasshoppers (Goodman et al., 1982) and in the zebrafish, where the axons of identified motoneurons charactersitically pause at the boundary between the dorsal and ventral myotome (Eisen et al., 1986). Critical examination of this issue requires labeling neurons orthogradely at the time of axonogenesis. We estimate that, in the absence of axonal pauses, axonogenesis occurs at 20-22 hr, an hour before we first labeled neurons retrogradely. An alternative to a timing hypothesis is that the growth cones of the MiD cells specifically recognize different pathways, either near the cell bodies where they would lead to the ipsilateral versus the contralateral descending tract, or the left versus the right tracts themselves. Cell-specific selection of “labeled pathways” has been proposed as a mechanism for axonal guidance in both invertebrates (Goodman et al., 1982; Raper et al, 1984a,b; Bastiani et al, 1984) and vertebrates (e.g., Eisen et al, 1986;Lance-Jones and Landmesser, 1980a,b). Thus the mechanisms that guide axons specifically towards their targets may be conserved during evolution. Some of this work has appeared previously in the form of an abstract (Mendelson, 1984). We thank Monte Westerfield, Walter Metcalfe and Eric Frank for providing valuable comments concerning this manuscript and Reida Kimmel for her excellent technical assistance. This work was supported by NIH Grant NS 17963 and 5 T32 GM0’7257-07. REFERENCES BASTIANI, M. J., RAPER, J. A. and GOODMAN, C. S. (1984). Pathfinding in neuronal growth cones in grasshopper embryos. III. Selective affinity of the G growth cone for the P cells within the A/P fascicle. J. Neurosci. 42339-2345. EISEN, J., MYERS, P. Z., and WESTERFIELD, M. (1986). Pathway selection by growth cones of identified motoneurons in live zebra fish embryos. Nature 320,269-271. FRANK, E., and WESTERFIELD, M. (1983). Development of sensorymotor synapses in the spinal cord of the frog. J. Physiol. (London) 343,593610. GOODMAN, C. S., RAPER, J. A., Ho, R. K., and CHANG, S. (1982). Pathfinding by neuronal growth cones in grasshopper embryos. In “Developmental Order: Its origins and Regulation” (S. Subtelny and P. B. Green, eds.). pp. 275-316. Liss, New York. GO~LEIB, D. I., and COWAN, W. M. (1972). Evidence for a temporal factor in the occupation of available synaptic sites during development of the dentate gyrus. Brain Res. 41,452-456. HANKER, J. S., YATES, P. E., METZ, C. B., and RUSTIONI, A. (1977). A
313
new, specific, sensitive and non-carcinogenic reagent for the demonstration of horseradish peroxidase. H&&em 9,789-792. HORDER, T. J., and MARTIN, A. C. (1978). Morphogenetics as an alternative to chemospecificity in the formation of nerve connections. Symp. Sot. Exp. Biol 32,275-358. KIMMEL, C. B. (1982). Reticulospinal and vestibulospinal neurons in the young larva of a teleost fish, Brachydanio rerio. Prog. Brain Res. 57,1-23. KIMMEL, C. B., POWELL, S. L. and METCALFE, W. K. (1982). Brain neurons which project to the spinal cord in the young larvae of the zebrafish. J. Camp. Neural. 205,112-127. KIMMEL, C. B., SESSIONS, S. K., and KIMMEL, R. J. (1978). Radiosensitivity and time of origin of the Mauthner neurons in the zebrafish. Dev. Biol. 62,526-529. LANCE-JONES, C., and LANDMESSER, L. (1980a). Motoneurone projection patterns in embryonic chick limbs following partial deletions of spinal cord. J. Physiol (London) 302,559-580. LANCE-JONES, C., and LANDMESSER, L. (1980b). Motoneurone projection patterns in embryonic chick hind limb following partial reversals of the spinal cord. J. Physiol. (London) 302,581-602. LEVI-M• NTALCINI, R. (1982). Developmental neurobiology and the natural history of nerve growth factor. Annu. Rev. Neurosci. 5, 341362. MENDELSON, B. (1984). Is timing important in the guidance of axons of identified neurons? Sot Neurosci A&r. 10,42. MENDELSON, B. (1985). Soma position is correlated with time of development in three types of identified reticulospinal neurons. Dev. Biol 112,489-493. MENnELsON, B. (1986a). Development of reticulospinal neurons of the zebrafish: I. Time of origin. J. Camp. Neural 261, in press. MENDELSON, B. (198613). Development of reticulospinal neurons of the zebrafish: II. Early axonal outgrowth and cell body position. J. C’omp. Neural. 251, in press. METCALFE, W. K. (1985). Sensory neuron growth cones comigrate with posterior lateral line primordial cells in zebrafish. J. Comp. Neural. 238,218-224. METCALFE, W. K., MENDELSON, B., and KIMMEL, C. B. (1986). Segmental homologies among reticulospinal neurons in the hindbrain of the zebrafish larva. J. Camp. Neural. 251, in press. RAGER, G. (1980). Specificity of nerve connections anisms. Trends Neurosci. 3,43-44.
by unspecific mech-
RAPER, J. A., BASTIANI, M. J., and GOODMAN, C. S. (1983). Pathfinding by neuronal growth cones in grasshopper embryos. II. Selective fasciculation onto specific axonal pathways. J. Neurosci 3,31-41. RAPER, J. A., BASTIANI, M. J., and GOODMAN, C. S. (1984). Pathfinding by neuronal growth cones in grasshopper embryos. III. The effects of ablation of the A and P axons upon the behavior of the G growth cone. J. Neurosci. 4,2339-2345. SINGER, M., NORDLANDER, R. H., and EGAR, M. (1979). Axonal guidance during embryogenesis and regneration in the spinal cord of the newt. “The blueprint hypothesis” of neuronal pathway selection. J. Cornp. Neural. 185,1-22.