Development of the optic tract in the cichlid fish Haplochromis burtoni

DevelopmentalBrain Research, 26 (1986) 179-186

179

Elsevier BRD 50376

Development of the Optic Tract in the Cichlid Fish Haplochromis burtoni JOELLE C. PRESSON and RUSSELL D. FERNALD

Institute of Neuroscience, Universityof Oregon, Eugene, OR 97403 (U.S.A.) (Accepted October 15th, 1985)

Key words: visual system development - - larval teleost - - pathway guidance - - optic tract

In the teleost fish, Haplochromis burtoni, the optic tract is composed of 3 distinct components: the marginal tract, which projects to the optic tectum and is by far the largest, and the axial and medial tracts which project to diencephalic targets. In this paper we report on the normal development of these pathways in larval H. burtoni, an African cichlid fish. The earliest optic tract fibers are found in what will become the marginal optic tract. These fibers hug the wall of the diencephalon in a cohesive bundle. The first fibers in the axial tract location appear on day 5, increasing in number between days 6 and 18. Like marginal tract fibers, axial tract fibers form a cohesive bundle, It is not clear from these experiments whether the first axial tract fibers actually arrive at this location at day 5, or whether they are fibers arriving earlier that were physically displaced from the marginal tract at day 5. Medial tract fibers are not evident until day 6 of development and the number of medial tract fibers also increases as the animal gets older. Unlike fibers in the other two pathways, medial tract fibers do not travel together in a bundle. Rather, each one follows an independent trajectory to its target site. Comparison of this larval development with the adult optic tract organization which we have studied earlier suggests constraints on the mechanisms of axon guidance. INTRODUCTION The visual systems of teleost fish have often been used to study the mechanisms that guide axons from the eye to target sites in the brain. Virtually all of this work 1,7,10,19,21.26 has dealt with the fibers from the marginal optic tract ( M O T ) which project from the retina to the superficial layers of the optic tectum. There are, however, two smaller pathways, the axial

system is to describe the normal development of the pathways involved. This report describes the overall development of the optic tract in H. burtoni, including when and where the M O T , A x O T , and M d O T first appear. The findings have implications for the mechanisms of axon guidance and suggest future experiments. A preliminary report of these findings has appeared elsewhere 17.

optic tract ( A x O T ) and the medial optic tract (MdOT), that project to mid-brain sites 3,10.14.24. We

MATERIALS AND METHODS

recently described the retinotopic organization of

Experimental animals

these pathways in adult animals of the cichlid fish, Haplochromis burtoni 16. M d O T , A x O T and M O T fibers arise from adjacent retinal ganglion cells and are indistinguishable from one another in the optic nerve. However, at the optic chiasm they segregate and subsequently follow very different pathways to their brain targets. Thus, these m i n o r pathways may provide useful information about principles underlying axon guidance. The first step in the study of axon guidance in any

Larval fish from the African cichlid species Haplochromis burtoni were used in these experiments. The larval fish were obtained from special breeding colonies. Each breeding tank contained 4 females and 3 males. The tanks were maintained in conditions that mimic the natural habitat of this species 8. The water temperature was 26.5 °C, the pH at 7.8, and the light/dark cycle 12 h/12 h. The standard Tetramin diet was supplemented with brine shrimp. The breeding tanks were observed through one-way glass

Correspondence: J.C. Presson, Department of Anatomy, University of Georgetown Medical School, Washington, DC 20007, U.S.A.

180 to minimize disruption of mating behavior. The time of fertilization was determined by systematically observing the fish and noting the time of the complex, distinctive mating behavior 9. Following fertilization in H. burtoni, the eggs are carried in the female's mouth for about 2 weeks. A single female broods about 15-30 fry 9. A few fry were removed from a female's mouth at the desired time, and the rest were left with her to continue development to later stages. In this way several developmental stages could be sampled from each brood. The experiments reported here are based on a total of 25 animals from 5 broods. The fish used in this series were 4 - 1 8 days postfertilization and had all hatched from the chorion.

Labelling optic axons Optic fibers were labelled with cobaltous lysine. Fish older than 9 days were anesthetized by immersion in iced tank water. The cornea was pierced with a sharpened tungsten needle and the lens removed. The retina was gently disrupted with the needle and a small amount of dried cobaltous lysine was placed in the eye. The eye was sealed with a melted gelatin/glycerol (1:1:4.6 distilled water) mixture. The fish survived 5-16 h. The animals were processed according to a protocol modified from Springer and ProkoschZL The cobalt was precipitated by immersing the animals in a 2.5% ammonium sulfide solution in 0.9% saline. They were then rinsed and fixed by immersion overnight in 1.5% formaldehyde/1.5% glutaraldehyde. The entire animal was embedded in plastic (JB-4, Polysciences) and sectioned at 10/~m. The cobalt was intensified in silver nitrate and hydroquinone, suspended in gum arabic. Three intensifications of 3 min each resulted in the lowest background levels, The sections were then counterstained with toluidine blue. RESULTS

Adult optic tract To facilitate understanding the results presented here, we will first give a brief description of the optic tracts in adult H. burtonpOA 6. A cross-section through the adult brain, at a level caudal to the chiasm, is shown in Fig. 1, indicating the appearance of the MOT, M d O T and A x O T optic tracts. The

!~ ~ ~ ~i~ :~ !!!~~i~ii~~i!ci!!i~)~ ~! ~ Fig. 1. A transverse section through the optic tract of an adult H. burtoni in which all optic fibers have been labelled with cobaltous lysine (see Materials and Methods). MOTd, dorsal branch of MOT; MOTv, ventral branch of MOT. In all figures. section orientation is given by: d, dorsal; m, medial; v, ventral; I, lateral. Calibration bar, lOlt/~m.

M O T is the largest optic fiber tract and lies along the lateral wall of the forebrain. A x O T fibers segregate from M O T fibers at the level of the suprachiasmatic nucleus, travel close to the midline until the level of the rostral pretectum, and then curve laterally to rejoin the MOT. M d O T fibers lie at the medial edge of the MOT.

General description of larval fish Since a general staging scheme for teleost development is not commonly in use, a brief description of the larval fish used in these experiments will be given. The youngest animals in this series were 4 days old, just posthatching. At this stage the body lacks any pigmentation. The eye is heavily pigmented, and its ventral fissure is still visible externally. The retina contains differentiated layers in its center, and the optic nerve has formed (Fig. 2). Most of the retina is dorsal to the optic nerve head, but there is a small amount of undifferentiated retina ventrally. The retinal ganglion cell layer is ca. 5 cells thick and there are an estimated 5400 ganglion cells present. The optic tectum at this stage is primarily comprised of periventricular neurons (Fig. 2b) and the superficial layers are very thin. The oldest animals in this series were

181

Fig. 2. Transverse sections through retina and optic tecta of the youngest and oldest animals used in this study, a: retinal section from a 4-day-old animal, b: optic tectum section from a 4-day-old animal, c: retinal section from an 18-day-old animal, d: optic rectum section from an 18-day-old animal. The small arrows in (a) and (c) point to the ganglion cell layer and the arrowheads indicate the optic nerve coming from the eye. OT, optic tectum; pl, periventricular layer; sl, superficial layer. Calibration bars, 50 ~m.

182 18 days, 4 days after animals are normally free swimming II. By this time, although the eye and brain are still growing, they both have an adult appearance (Fig. 2c, d). The retinal ganglion cell layer has spread out and is only ca. 3 cells thick, and retinal ganglion cell number has increased to ca. 60,000. The superficial layers of the tectum are well developed 12.

3a). The optic fibers form a cohesive bundle and there is no evidence for the pathway diversity seen in adult animals. As the animals get older the number of fibers in this lateral pathway increases, and it becomes clear that this is the developing MOT (Fig. 3b-f). In fry of all ages the MOT carries the majority of optic fibers, just as in adults.

The development of the MOT

The development of the AxOT

At day 4, all existing optic fibers travel along the lateral wall of the developing diencephalon, taking a straight course from the eye to the optic tectum (Fig.

As described above, at 4 days of age, the AxOT is not present as a distinct pathway. By day 5, however, a few fibers are segregated from the medial edge of

Fig. 3. The development of the M O T and A O T as seen in transverse sections taken from 6 animals ranging in age from 4 to 18 days. Optic fibers have been labelled with cobaltous lysine. In all cases, the arrows point to M O T fibers and the arrowheads point to the A O T fibers: a: 4-day-old fish: b: 5-day-old; c: 6-day-old: d: 8-day-old: e: 12-day-old: f: 18-day-old fish. Calibration bars, 501~m.

183

Fig. 4. Transverse sections through the optic tracts of 8- to 18-day-old animals, showing the development point to the MdOT fibers; a: 8-day-old fish; b: lo-day-old; c: 12-day-old (orientation as in a); d: 18-day-old Calibration bar, 50pm.

of the MOT. The arrows fish (orientation as in b).

184 the MOT, at the base of the brain in the region of the developing preoptic area (Fig. 3). These few segregated fibers represent the first appearance of the AxOT. As development progresses (Fig. 3c-f), the segregation between the MOT and AxOT fibers increases. In addition, the width of the AxOT increases with the age of the animal, implying that the number of fibers in the AxOT increases. By 18 days of age (Fig. 3f), the AxOT looks very much as it does in adult animals (Fig. 1), An important feature of the AxOT fibers is that at every age they are tightly fasciculated, giving the impression that later developing fibers follow very closely alongside earlier AxOT fibers. Although the first axial tract fibers do not appear until day 5, our data do not indicate whether these fibers actually arrive at the tract on day 5. They could be fibers that arrive earlier and are then displaced from the marginal tract fibers on day 5.

Development of the MdOT MdOT fibers do not appear to be present at the earliest stages of development. At days 4 and 5 (Fig. 3a, b) there are no fibers traversing the diencephalic white matter, where one would expect MdOT fibers. Of course, these fibers could be very small and difficult to detect, but none have been seen in 4 animals of ages 4 and 5 days. The first MdOT fibers appear at day 6, but are not easily photographed until day 8. At this age (Fig. 4a) there are one or two fibers that leave the MOT and take a straight dorsal course across the diencephalic white matter. As development proceeds, the number of such fibers increases (Fig. 4b-d). By day 18, it is clear that these fibers comprise the early MdOT (Fig. 4d), but the MdOT at day 18 is not as densly populated as it is in adult animals (Fig. 1). Unlike the MOT and AxOT fibers, the MdOT fibers do not closely follow one another as they leave the MOT and approach their midbrain targets. Rather, each MdOT fiber appears to take an independent route through the developing diencephaIon. The appearance of a cohesive bundle of fibers in adults may be the result of later MdOT fibers filling in the spaces between existing MdOT fibers. DISCUSSION

Development of the optic tract In the adult teleost, H. burtoni, optic fibers take

the marginal, medial or axial pathway to the brain, depending upon their target sites. The purpose of this study was to determine when during development these 3 components of the optic tract first appear, and to examine the implications of this developmental timetable for our understanding of axon guidance mechanisms. The very earliest optic fibers to arise in the embryo do not contribute to the pathway diversity seen in adults. The first fibers to project to the brain all travel in the MOT as a cohesive bundle along the wall of the developing diencephalon. Thus something about either optic fibers or the developing brain changes between day 4 and older ages that instructs or allows optic axons to form the AxOT and the MdOT. The first AxOT fibers appear in the region of the chiasm around day 5. From the current series of experiments we cannot say whether these first AxOT fibers are those that arrived at the chiasm at day 5, or whether optic fibers that had already reached the brain are displaced by the developing white matter of the diencephalon. This question could be answered by labelling small groups of optic fibers from different retinal regions in 5-day-old animals and following their central projections. Presson et al. 16 suggested that the AxOT might be formed by the passive growth of the diencephalon in between optic fibers. From the present study, it appears that diencephalic tissue does grow in between the AxOT and MOT during development (Fig. 3). This cannot be a major mechanism in the formation of the AxOT, however. Between days 6 and 18 the number of fibers in both the AxOT and MOT increases, indicating that during this developmental period optic fibers must choose between the AxOT pathway and the MOT pathway. The mechanism of AxOT pathway choice is not known, but several possibilities can be considered. First, it is known from studies of the adult visual system 10,13A6 that AxOT fibers project to the suprachiasmatic nucleus. During the time that the AxOT is being formed, the developing preoptic area, which contains the suprachiasmatic nucleus, is very close to the position of the developing AxOT. Thus, some chemical or structural message from the suprachiasmatic nucleus might control the pathway choice of some optic fibers. Second, some structural or chemical features not related to the suprachiasmatic nucleus

185 might dictate the pathway choice of A x O T fibers. For example, perhaps A x O T fibers have a special affinity for one another. A third possibility is that optic fibers will follow any structural feature in the environment, and that some random subset ends up in the AxOT. There is some evidence that optic fibers will follow structural features, even if they lead to an inappropriate target2.4,15. One way to refute the random choice hypothesis would be to determine that these early AxOT fibers do not arise from a random set of retinal ganglion cells. Whatever model is eventually proposed for A x O T formation, it must account for the fact that at some point during development optic fibers no longer choose the AxOT pathway. Presson et al. 16 found that central retinal fibers project via the AxOT, but peripheral retinal fibers do not. Studies of retinal growth in H. burtoni have shown that the central retinal fibers are the earliest to develop, and peripheral retinal fibers develop laterllA2, z2. Thus, the mechanism that guides fibers to the A x O T between days 5 and 18, must at some point cease to operate. The MdOT appears to be the last component of the optic tract to begin development. Prior to 6 days of .age, we have seen no fibers that take this pathway. The onset of the formation of the MdOT is quite striking. Between 6 and 8 days of age there are suddenly one or two fibers coursing through the diencephalic white matter, and these can be followed dorsally and caudally to the deep layers of the optic tectum, one of the projection sites of the M d O T 10,16. Hypotheses about how the MdOT fibers choose this route are difficult to pose, since there are no obvious features in the area of the developing M d O T that might be providing structural or chemical cues. It is possible that the MdOT target siteslO.13A6 (thalamus, d e e p tectal layers and dorsal pretectum) are providing chemical guidance, as these are present before day 6. These structures, however, are about 200 ~tm from the location where M d O T fibers leave the M O T in the youngest animals. Thanos et al. 27 have interpreted their evidence to mean that in the developing optic tectum, any chemical cues involved in optic fiber guidance can only operate over short distances. Furthermore, Presson et al. 16 have shown that MdOT fibers arise from all parts of the retina, including the more peripheral regions that are added continuously during adulthood. This makes chemical

cues from the target structures even more unlikely, since in adults the target structures are several mm from the optic chiasm, where MdOT fibers first exhibit their independent pathways. There are no obvious structural features, which the MdOT fibers might be following. It does seem clear that MdOT fibers are not following one another. Each fiber takes a slightly different path after leaving the MOT. This behavior is quite striking since fasciculation is a common feature of axonal outgrowth. It would be of interest to determine whether fibers in the MdOT differ from other optic fibers in the expression of some cell adhesion molecule, such as N-CAM 18.

The relevance of teleost optic tract development to mammalian systems An explication of the mechanisms underlying optic tract development in teleosts would be of general interest, but are the mechanisms involved likely to be applicable to the development of mammalian visual systems? One conspicuous feature of mammalian development is cell death. In mammals many retinal ganglion cells project to incorrect sites during development and then die 5. Thus, mammalian retinal ganglion cells may not rely primarily on specific guidance mechanisms, since errors do occur and are subsequently eliminated. In developing H. burtoni there is no evidence of necrotic cells in the retina 11,12suggesting that cell death is not a central mechanism of visual system development in this species. Moreover, we have no evidence that early optic fibers take inappropriate pathways. Instead, the vast majority of optic axons appear to take the correct pathway during development, and thus must rely on specific c u e s either external or internal. Early optic fibers in mammals also take appropriate pathways and thus must use some specific guidance cues. Silver has demonstrated that in the mouse the majority of early optic fibers occupy the appropriate position in the optic nerve and make the correct choice at the optic chiasm 23. Both of these phenomena appear to be the result of guidance by structural features in the optic pathway. Drager6 found that early optic fibers do reach the appropriate tectal targets in mice. Errors in projection occur at later stages in development. Thus, even if cell death is not a characteristic of the teleost visual system, the principles gleaned from studies of teleosts are likely to also prove useful for analysis of mammalian visual development.

186 ACKNOWLEDGEMENTS

00832 to R . D . F . W e w o u l d e s p e c i a l l y like to t h a n k Linda Shetton and Harry Howard for valuable help

This work

was s u p p o r t e d

by g r a n t s f r o m

the

with the photomicrographs.

W h i t e h a l l F o u n d a t i o n , N I H E Y 05051 a n d F06 T W

REFERENCES 1 Bodick, N. and Levinthal, C., Growing optic nerve fibers follow neighbors during embryogenesis, Proc. Natl. Acad. Sci. U.S.A., 77 (1980) 4374-4378. 2 Bohn, R.C. and Stelzner, D.J., The abberant retino-retinal projection during optic nerve regeneration in the frog. I. Time course of formation and cells of origin, J. Comp. Neurol., 196 (1981) 605-620. 3 Braford M.R. and Northcutt, R.G., Organization of the diencephalon and pretectum of the ray-finned fishes. In R.E. Davis and R.G. Northcutt (Eds.), Fish Neurobiology, Vol. 2, University of Michigan Press, Ann Arbor, 1983, pp. 117-163. 4 Bunt, S.M. and Lund, R.D., Development of a transient retino-retinal pathway in hooded and albino rats, Brain Res., 211 (198l)399-404. 5 Cowan. W.M., Fawcett, J.W., O'Leary, D.M.O. and Stanfield, B.B., Regressive events in neurogenesis, Science, 225 (1984) 1258-1265. 6 Drager, U.C., The time of origin of ganglion cells giving rise to crossed and uncrossed projections in the mouse retina, Soc. Neurosci. Abstr., 10 (1984) 141. 7 Easter, S.S., Jr. and Stuermer, C.A.O., An evaluation of the hypothesis of shifting terminals in the goldfish optic tecturn, J. Neurosci., 4 (1984) 1052-11163. 8 Fernald, R.D. and Hirata, N.R., Field study of Haplochromis burtoni: habitats and cohabitants, Environ. Biol. Fishes, 2 (1977) 299-308. 9 Fernald, R.D. and Hirata, N.R., Field study of Haplochromis burtoni: Quantitative behavioral observations, Anita. Behav., 25 (1977) 964-975. 10 Fernald, R.D., Retinal projections in the African cichlid fish, Haplochromis burtoni, J. Comp. Neurol., 206 (19821 379-389. 11 Fernald, R.D., Neural basis of visual pattern recognition. In J.P. Ewert, R.R. Capranica and D.J. Engle (Eds.), Advances in Vertebrate Neuroethology, Plenum, New York, 1983, pp. 569-580. 12 Fernald, R.D., Growth of the teleost eye: novel solutions to complex constraints, Env. Biol. Fish., 13 11985) 113-123. 13 Fernald, R.D. and Shelton, L.C., The organization of the diencephalon and pretectum in the cichlid fish, Haplochromis burtoni, J. Comp. Neurol., 238 (1985) 2(/2-217.

14 Fraley, S.M. and Sharma, S.C., Topography of retinal axons in the diencephalon of goldfish. Cell Tiss. Res.. 238 (1984) 529-538. 15 Harris, W.A., Holt, C.E., Smith, T.A. and Gallenson, N., Growth cones of developing retinal cells in vivo, on culture surfaces, and in collagen matrices, J. Neurosci. Res., 13 (1985) 1/11-122. 16 Presson, J., Fernald, R.D. and Max. M., The organization of retinal projections to the diencephalon and pretectum in the cichlid fish, Haplochromis burtoni, J. Comp. Neurol., 235 (1985) 360-374. 17 Presson, J.. Shelton, L.C. and Fernald, R.D., Development of the optic tract in a cichlid fish, Haplochromis burtoni, Soc. Neurosci. Abstr., 10 (1984) 465. 18 Rutishauser, U., Developmental biology of a neural cell adhesion molecule, Nature (London), 310 (1984) 594-597. 19 Rusoff, A.C., Paths of axons in the visual system of perciform fish and implications of these paths for rules governing axonal growth, J. Neurosci., 4 (1984) 1414-1428. 211 Scholes, J.H., Nerve fiber topography in the retinal projection to the rectum, Nature (London), 278 (1979) 620-624, 21 Scholes, J.H., Ribbon optic nerves and axonal growth patterns in the retinal projection to the tectum, Br. Soc. Dev. Biol. Syst., 5 (1981) 181-214. 22 Shelton, L.C. and Fernald, R.D., What is the relationship between embryonic and adult growth patterns in the teleost visual system? Soc. Neurosci. Abstr., 9 11983) 101. 23 Silver, J., Studies on the factors that govern directionality in the embryonic optic nerve and at the chiasm of mice, J. Comp. Neurol., 223 119841 238-251. 24 Springer, A D . and Mednick, A.S., Retinofugal and retinopetal projections in the cichlid fish Astronotus ocellatus, J. Comp. Neurol., 236 (19851 179-196. 25 Springer, A.D. and Prokosch. J., Surgical and intensitifcation procedures for defining visual pathways with cobaltous lysine, J. Histochern. Cvtochem., 30 (19821 1234-1242. 26 Stuermer. C.A.O., Rules for retinotectal terminal arborizations in the goldfish optic tectum: a whole mount study, J. Comp. Neurol., 229 (1984) 214-232. 27 Thanos, S., Bonhoeffer, F. and Rutishauser, U., Fiber fiber interactions and tectal cues influence the development of the chicken retinotectal projection, Proc. Natl. Acad. Sci. U.S.A.. 81 (1984) 1906-19111.