The normal and abnormal postnatal development of retinogeniculate projections in golden hamsters: An anterograde horseradish peroxidase tracing study

Developmental BrainResearch, 12 (1984) 191-205

191

Elsevier

The Normal and Abnormal Postnatal Development of Retinogeniculate Projections in Golden Hamsters: An Anterograde Horseradish Peroxidase Tracing Study K.-F. SO, H. H. WOO and L. S. JEN

Department of Anatomy, Facultyof Medicine, Universityof Hong Kong (Hong Kong) (Accepted September 13th, 1983)

Key words: development of retinal pathways - - lateral geniculate nucleus - - hamster

In adult hamsters, the dorsal lateral geniculate nucleus (LGd), which lacks a noticeable pattern of cellular lamination, receives fibers predominantly from the contralateral eye except for a medial segment which receives fibers from the ipsilateral eye. Using the method of anterograde transport of horseradish peroxidase (HRP), it is shown in this study that the contralateral and ipsilateral retinogeniculate fibers innervate the LGd by day 0 with the development of the contralateral fibers slightly ahead of the ipsilateral ones. The entire contralateral LGd is filled with retinal fibers by day 1. The ipsilateral LGd is almost completely covered with retinal fibers on day 2 but with the fiber density much higher in the dorsal half of the nucleus. Thus, fibers from both eyes overlap with each other completely beginning on day 2 in the LGd, The segregation of these fibers becomes obvious on day 6 as indicated by a decrease in the density of ipsilateral fibers in the ventral portion of the LGd while the ipsilateral projection continues to concentrate in the dorsal half of the nucleus. A low density area in the dorsomedial part of the contralateral LGd is observed on day 7. By day 8, the segregation of the contralateral and ipsilateral projections has achieved an adult-like pattern. Thus, there seems to be two phases for the normal development of the retinogeniculate fibers. In the first phase, axons from both eyes grow in and occupy the entire LGd. In the second phase, those axons occupying inappropriate areas of the LGd are eliminated to form the adult pattern. The effect of unilateral eye removal at birth on the development of the retinogeniculate projection from the remaining eye was also studied with the anterograde HRP method. The ipsilateral fibers in the experimental animals are distributed in the lateral portion of the nucleus in the first two postnatal days. The entire LGd is not filled with ipsilateral fibers until day 4. From day 6 onwards, the ipsilateral fibers are more extensive than those of the normal animals. In addition to a dense projection to the dorsal two-thirds of the LGd, moderate amount of ipsilateral axons can be detected in the remaining ventral portion of the nucleus in day 6 and older experimental animals. The development of the contralateral retinal fibers in the experimental animals is similar to that of the normal from day 1 to day 6, i.e. the entire LGd is densely filled with crossed optic axons. However, on day 7, there is no decrease in the density of the contralateral fibers in the medial sector of the LGd. The lack of restriction of retinal fibers is observed in all subsequent stages of development. These results suggest that axo-axonal interaction is important for the process of segregation of retinogeniculate projections observed in normal developing hamsters. INTRODUCTION

tially o v e r l a p p i n g inputs was also d i s c o v e r e d in the d e v e l o p m e n t of o r d e r l y c o n n e c t i o n s in central n e u r a l

A n e m e r g i n g g e n e r a l principle in the d e v e l o p m e n t

structures including o l f a c t o r y cortex26, h i p p o c a m -

of c o n n e c t i o n s in the m a m m a l i a n p e r i p h e r a l n e r v o u s

pus 9 and c e r e b e l l u m 6. T h e s e results increase the possibility that this p h e n o m e n o n m i g h t be a g e n e r a l prin-

system is that each skeletal muscle fiber is initially p o l y n e u r o n a l l y i n n e r v a t e d , and all but o n e of t h e s e c o n n e c t i o n s s u b s e q u e n t l y d i s a p p e a r 2,24,29. A similar

ciple for the d e v e l o p m e n t of o r d e r l y n e u r a l c o n n e c -

p h e n o m e n o n was also o b s e r v e d in the d e v e l o p m e n t

extensively in the visual system.

of c o n n e c t i o n s in the m a m m a l i a n a u t o n o m i c n e r v o u s system. In the rat s u b m a n d i b u l a r g a n g l i o n , n e u r o n s

T h a t the retinal fibers f r o m the two eyes originally o v e r l a p and s u b s e q u e n t l y s e g r e g a t e was r e p o r t e d in the dorsal nucleus of the lateral g e n i c u l a t e b o d y

i n n e r v a t e d by single p r e g a n g l i o n i c axons in adulth o o d are also p o l y i n n e r v a t e d at n e o n a t a l stageslL R e c e n t l y , this p h e n o m e n o n of s e g r e g a t i o n of ini-

tions. This p h e n o m e n o n has b e e n i n v e s t i g a t e d m o s t

( L G d ) and s u p e r i o r colliculus (SC) in the m o n k e y 27, h a m s t e r n,36,37, o p o s s u m 4, rat 17.21 , cat 33,39, grey squir-

Correspondence: K.-F. So, Department of Anatomy, Faculty of Medicine, University of Hong Kong, Hong Kong, 0165-3806/84/$03.00 © 1984 Elsevier Science Publishers B.V.

192 relS and brushtailed possum32. A similar segregation phenomenon was demonstrated for the normal and the abnormal recrossed retinocollicular projections in hamsters with unilateral SC removal on the day of birth 34. Although much evidence is accumulating which illustrates this phenomenon, studies are just beginning to investigate the underlying mechanisms. In the hamsters the whole process of overlap and segregation of retinal axons occurs during the postnatal period, thus making them particularly suitable for experimental study. There are at least two possible mechanisms that should be examined37: (1) axon-substrate interactions (mechanisms by which the fibers coming from the contralateral and ipsilateral eye become segregated and restricted to their final terminal zones without any direct interaction between the two axon populations); and (2) axo-axonal interaction (segregation due to some form of interaction or competition between the fibers from the contralateral and ipsilateral eyes). In trying to distinguish between these two mechanisms, one eye was removed in hamsters at birth and the retinal projection from the remaining eye was studied when the animals were 6 weeks old 37. A similar experiment has been carried out in the monkey. One eye was removed at prenatal day 64 and the animal was 3 months old postnatally when the projection from the other eye was studied 2~. In both cases, the retinogeniculate axons did not form the characteristic adult laminar pattern. Instead, the retinal fibers occupied the entire contralateral geniculate nucleus. This finding seems to be consistent with the axoaxonal interaction hypothesis, i.e. failure of segregation is due to the lack of fiber interaction. However, it is possible that fibers from the remaining eye might have been segregated during early development and later sprouted and refilled the ipsilateral area before the animals reached the adult stage, The following experiment was therefore performed to investigate this possibility. One of the two populations of retinal fibers was removed by means of unilateral enucleation in hamsters at birth and the development of the retinal fibers from the remaining eye was studied on various closely spaced postnatal days using the anterograde horseradish peroxidase (HRP) method. This is important for a critical examination of the above possibility. To serve as controls, the distribu-

tion pattern of the developing retinogemculatc proiections in a series of normal hamsters of comparable ages was also studied with the same technique. MATERIALS AND METHODS Golden hamsters (Mesocricetus auratus) were bred in a room with a light-dark cycle of 14 h light, 10 h dark. The time of impregnation was obtained by observation of mating behavior. A golden hamster gives birth 16 days after mating. The first 24 h period following birth is referred to as day 0, the next 24 h as day 1 etc. When a phenomenon regarding the developing retinogeniculate projections is said to occur on a certain day, it is meant that this phenomenon is observed in animals injected on that day; the animals were killed late in the same 24 h period. Survival times longer than 24 h were not used for the developing animal in order to avoid difficulty in interpreting the result. The normal development of the retinogeniculate projections was studied in 64 hamsters with the anterograde H R P method. At various postnatal days (days 0-42, Table I), the right eye of each animal was injected with 0.2 to 3 mg of H R P (Sigma type V1) dissolved in 0.5 to 3/xl of Tris buffer at pH 7.5. The amount of H R P solution injected was judged by the volume of the vitreous body for each animal. Injection of too much H R P solutions was avoided because it would result in leakage and might contaminate the brain. Details of the injection procedure have al-

TABLE I Animals used for studying the normal development of the retinogeniculate projections Days after birth 0 1

2 3 4 6 7 8 10 14 17 42 or adult

Number of animals

Survival time (h)

HRP (mg) dosage

7

9-19 21 18-24 15-21 14-24 19-24 18.5 19-24 22-24 18-24 19-24 17-105

0.2-0.3 0.3 0.4 1t.4 0.4-0.6 0.4-0.6 11.6-0.7 0.6-11.8 0.6-0.8 1.0 1.0 2.11-3.0

3

3 4 8 6 1 8 6 7 3 8

193 ready been published 34. H y p o t h e r m i a was sufficient anesthetic for animals younger than 4 days. O l d e r animals were anesthetized by inhalation of ether. Animals younger than 17 days were allowed to survive from 9 to 24 h after eye injection; a longer survival time was used for most older animals (Table I). The animals were reanesthetized and perfused with 0.9% saline followed by a fixative consisting of 1% paraformaldehyde and 1.25% glutaraldehyde in 0.1 M phosphate buffer. The brains were r e m o v e d , postfixed in the same fixative for 3-5 h and put into phosphate buffer solution containing 10% sucrose until they sank. Frontal sections were cut frozen at 50/~m thickness. Four 1 in 5 series of sections from each brain were reacted for the d e m o n s t r a t i o n of H R P with tetramethyl benzidine (TMB) 22 as the chromogen and two of the series were counterstained with Neutral red. The remaining unreacted sections were stained with cresyl violet. The a b n o r m a l d e v e l o p m e n t of the retinogeniculate projections was studied in 71 hamsters which had the left eye r e m o v e d under h y p o t h e r m i a anesthesia a few hours after birth. A t various postnatal days (days 1-42), the remaining right eye was injected with H R P solution and after an a p p r o p r i a t e survival time (Table II), the animals were perfused and the brain was processed for the H R P reaction as described above. All the sections were e x a m i n e d with bright- and dark-field illumination. Photomicrographs illustrat-

ing the distribution pattern of the retinogeniculate fibers were all taken at the mid-geniculate or slightly rostral level as illustrated in Fig. l B . RESULTS Dorsal nucleus o f the lateral geniculate body As a reference for the description of the development of the retinogeniculate projections, the various sectors and borders of the dorsal lateral geniculate nucleus ( L G d ) will be described briefly. The L G d in hamster does not show a clear p a t t e r n of cellular lamination as observed in other non-rodent species such as cat and monkey. H o w e v e r , based on retinal projection pattern, the L G d can be divided into a large a c and small ai and fl sectors 11.37 (Fig. 1). The ctc sector receives a heavy p r o j e c t i o n from the contralateral eye. It occupies the whole rostrocaudal extent of the nucleus and it consists of neurons with diameters ranging from 9 to 16~m40. However, the cells just immediately adjacent to the optic

--~

B

A

TABLE II Animals used for studying the development of the retinogeniculate projections after unilateral eye removal at birth Days after birth

Number of animals

Survival time (h)

HRP (mg) dosage

1 2 3 4 5 6 7 8 9 10 14 17 42 or adult

7 10 6 9 4 5 4 4 1 5 7 4 5

12-22 18-24 17-24 18-24 21 20-24 16-24 15-24 24 24 16-24 19-24 67-72

0.2-0.3 O.3 0.4 0.4-0.6 0.4-0.6 0.4-0.6 0.6-0.8 0.6-0.8 0.6-0.8 0.6-0.8 1.0 1.0 2.0-3.0

Fig. 1. Three different rostrocaudal levels of the dorsal lateral geniculate nucleus are illustrated in line drawings of coronal soctions through the diencephalon of an adult hamster. Only the right half of the section is shown. The rostral, middle and caudal thirds of the LGd are represented in A, B and C, respectively. Different sectors of the dorsal lateral geniculate nucleus are represented by ac, ai and fl (for details see text). Abbreviations: LGd and LGv, dorsal and ventral lateral geniculate nucleus respectively; LP, lateral posterior nucleus; PT, pretecturn; RT, thalamic reticular nucleus; ZI, zona incerta; nat, mammillothalamic tract; f, fornix; or, optic tract; st, stria terminalis; cp, cerebral peduncle. Scale = 1 mm.

194 tract are smaller, with diameters of 5-8/~m 4°. The czi sector receives a projection from the ipsilateral eye. It is located in the dorsomedial portion of the rostral

es retinal projection from both eyes. The borders of the LGd can be defined in the following manner. The lateral border is formed by the

half of the LGd, occupying about 8% of the total vol-

optic tract. Medially, the (z sectors of the nucleus are

ume of the nucleus. The neurons in the czi sector are slightly more densely packed than in the ac sector,

bounded by the external medullary lamina of the thalamus. The ventral and dorsal borders of the LGd

and their diameters range from 8 to 12/~m ~0. The

can be defined using cell packing density as a crite-

third part of the L G d is the fl sector which is located

rion. However, the dorsal borders in animals young-

just medial to the ventral part of the ac sector. This fl sector, similar to that in the rabbit 30 and opossumZ5 is

er than day 2 are difficult to delineate.

found approximately in the rostral third of the k G d . The neurons in this sector are smaller than in the rest of the nucleus, with diameter 6-10/~m ~. In the most rostral sections of the LGd, the fl sector is e m b e d d e d in the external medullary lamina but more caudally, most of the fibers of the external medullary lamina split into two bundles separating the fl sector medially from the a sector and laterally from the ventroba-

Normal development of the retinogeniculate projection Day O. In animals with an eye injection of H R P

sal nucleus. The fl sector, as indicated below, receiv-

shortly after birth and perfused within 11 h, the contralateral optic tract (OT) was heavily filled with H R P reaction product. Axons were observed to enter the contralateral L G d from the superficial optic tract, and some of the fibers in the mid-portion of nucleus had reached the medial edge of the nucleus

Figs. 2-8. Dark-field photographs of coronal sections illustrating the retinal projections to the contralateral (left-hand column) and ipsilateral (right-hand column) lateral geniculate body in the hamster. Each pair of photographs was taken from one brain at the same rostrocaudal level (mid-geniculate or slightly rostral level as illustrated in Fig. 1B). The time of unilateral eye injection with HRP is illustrated for each pair of photographs. Stars define the dorsal and ventral borders of the LGd. The broken white lines define the borders of the ipsilateral LGd. The magnification is the same for all the photographs. Scale = 100 urn. Figs. 2-5. Dark-field photographs illustrating the normal development of the retinogeniculate projections. Fig. 2. A and B: contralateral and ipsilateral LGd, respectively, of a normal day 0 hamster; 11 h survival. Note that the retinal fibers have penetrated more deeply in the mid-portion of the contralateral LGd. No fibers were observed inside the ipsilateral LGd.

195

Fig. 3. A and B: contralateral and ipsilateral L G d , respectively, of a normal day 0 hamster; 19 h survival. A few fibers are observed inside the ipsilateral L G d (big star). The continuous long whitelines are artifactual. C and D: contralateral and ipsilateral LGd, respectively, of a normal day 1 hamster; 21 h survival. P h o t o m o n t a g e was needed for C and D in order to have the entire L G d in focus. E and F: contralateral and ipsilateral LGd, respectively, of a normal day 2 hamster; 18 h survival. The open arrows point to the retinal fibers located in the fl sector. The solid arrows point to the retinal fibers located inside the ventrobasal nucleus.

196 (Fig. 2A). There were also a few deep optic axons going through the LGd on their way to other visual cell groups in the mesencephalon. In day 0 animals which were allowed to survive for a longer period of time, more fibers were found in the contralateral nucleus and a few axons had penetrated into the ,/3 sector (Fig. 3A, arrow). Very sparse retinal fibers were observed in the ipsilateral OT. No fibers, however, could be detected inside the ipsilateral LGd in the short survival day 0 animals (Fig. 2B). But in cases with a longer survival time, the optic fibers had penetrated into the LGd, reaching about half way between the OT and the medial border of the L G d (Fig. 3B, big star). Thus, axons from the ipsilateral eye have not innervated as much of the LGd as fibers from the contralateral eye. Day 1. The a sector of the contralateral LGd was filled with H R P reaction products on day 1. The deep optic axons passing through the nucleus seemed to be more numerous on this day compared to the day 0 cases (compare Figs. 2A, 3A and C). In these animals, retinal fibers were observed inside the [;~ sector of the nucleus (Fig. 3C, open arrow). A few axons had penetrated beyond the fl sector and the external medullary lamina and were found inside the ventrobasal nucleus of the thalamus (Fig. 3C, solid arrow). The population of ipsilateral optic fibers penetrated the LGd in the form of a cone. The base of the cone abutted the O T and the tip of the cone, consisting of sparse HRP-filled fibers, had reached the medial edge of the nucleus (Fig. 3D). A few fibers could be observed innervating the fl sector of the ipsilateral LGd (Fig. 3D, arrow) in 2 of the 3 cases. Day 2. In day 2 animals, the contralateral retinogeniculate fibers were densely and evenly distributed in the a sector of the LGd including the dorsomedial region (Fig. 3E). Some retinal fibers were found in the fl sector (open arrow, Fig. 3E). As in the day 1 animals, a few fibers had penetrated beyond the caudal part of the fl sector and the thalamic external medullary lamina to the lateral margin of the ventrobasal nucleus (Fig. 3E, solid arrow). This phenomenon was not found in later stages. The ipsilateral L G d was almost completely filled with optic axons on day 2. The axons were densely distributed in the dorsal half of the nucleus, but sparse H R P reaction products could be observed in the ventral portion (Fig. 3F) and the /J sector (arrow,

Fig. 3F) of the LGd. Thus, on day 2. tile contr~thncral and ipsilateral retinal fibers occupied 111c entire g G d and formed an overlapping pro,jection with c:lch other. Da)' 3. In day 3 animals, fibcrs from both c\'cs were still overlapping m the LGd. The oxerall distribution pattern of the contralateral retina/ fibers was about the same as that of the day-2 animals except no fibers were detected in the ventrobasal nucleus. The ipsilateral fibers were concentrated in the dorsal half of the nucleus. However. the quantity of H R P reaction products in the ventral half of the nucleus appeared to be slightly higher than that observed in day 2 animals. Day 4. In day 4 animals, the contralateral fibers were heavily and evenly distributed throughout the (z sector of the LGd (Fig. 4A). Sparse retinal fibers could be found in the contralateral and ipsilateral F~ sectors. Similar fibers were found in older animals and further description of their development will not be presented. The ipsilateral fibers of day 4 animals could still he observed covering the entire LGd (Fig. 4B). Howcver, the dorsomedial portion of the nucleus was innervated more densely than in day 3 animals. Day O. The entire a sector of the nucleus of the normal day-6 animal was densely filled with contralateral retinal fibers (Fig. 4C). But in some animals, the retinal fibers in the dorsomedial region were slightly sparser, suggesting that some fibers might have already started to be eliminated from that region. In the ipsilateral LGd, the density of fibers was greatly reduced in the ventral region (Fig. 4D). Retinal axons were found to concentrate further at the dorsomedial portion of the rostral half of the LGd. Thus, by day 6, fibers from both eyes, while still overlapping, began to segregate from each other and concentrate in the areas corresponding to their adult definitive territory. Day 7. In day 7 animals, the reduction of the fiber density in the dorsomedial region of the ~x sector of the ipsilateral LGd to form a medial 'hole' became obvious (Fig. 4E, arrow). On the other hand, retinal fibers in the ipsilateral LGd became more concentrated in the dorsomedial area. Very few ipsilateral fibers could be found in the remaining area of the LGd (Fig. 4F).

197

Fig. 4. A and B: contralateral and ipsilateral LGd, respectively, of a normal day 4 hamster; 18 h survival. Note that the whole ipsilateral nucleus is filled with retinal fibers. C and D: contralateral and ipsilateral LGd, respectively, of a normal day 6 hamster; 19 h survival. Note the decrease in density of retinal fibers in the ventral half of the ipsilateral LGd. E and F: contralateral and ipsilateral LGd respectively of a normal day 7 hamster; 18 h survival. The medial 'hole' is quite obvious on this day (arrow, 3E).

198

Fig. 5. A and B: contralateral and ipsilateral LGd, respectively, of a normal day 8 hamster:22, h survival. The distribution of the retinogeniculate projection observed on day 8 closely resembles the adult pattern. C and D: contralateral and ipsilater~d LGd of a normal day 14 hamster; 18 h survival. Note the presence of ipsilateral fibers in the [3 sector (arrox~, D).

199

Day 8 and older. In the day 8 animals, the ipsilater-

from both eyes at the margins of the ai and ac sectors

al and contralateral retinal fibers were almost com-

of the LGd. In addition, a few passing fibers on their

pletely segregated from each other (Fig. 5A and B).

way to other visual centers in the mesencephalon

There might be, however, a slight overlap of fibers

could be found inside the medial 'bole' in the contra-

Figs. 6-8. Dark-field photographs of LGd from animals with unilateral eye enucleation on day 0, and injection of the remaining eye with HRP on various postnatal days. Fig. 6. A and B: contralateral and ipsilateral LGd, respectively, of an experimental day 1 hamster; 24 h survival. C and D: contralateral and ipsilateral LGd, respectively, of an experimental day 2 hamster; 22 h survival. Note that the ipsilateral fibers in both day 1 and 2 experimental animals do not penetrate as deeply into the LGd as in the normal animals of the same ages. The long continuous white lines in all the figures are artifactual.

Fig. 7. A and B: contralateral and ipsilateral LGd. respectively, of an experimental day 3 hamster; 21 h survival. C and D: contralateral and ipsilateral LGd, respectively. of an experimental day 6 hamster: 20 h survival. Note that the ipsilateral projection is enlarged. E and F: contralateral and ipsilateral LGd. respectively. of an experlmental day 7 hamster: 21 h survival. Note that no medial ‘hole’ can be detected in the contralateral LGd.

201 lateral LGd in day 8 and older animals (Fig. 5A and C). On the ipsilateral side, the retinal fibers had grouped together as a concentrated patch in the dorsomedial part of the nucleus forming the ai sector (Fig. 5B and D). Therefore, the two fiber populations had attained their adult-like distribution pattern by day 8 and the pattern remained similar in all the older animals that were examined.

Abnormal development of the retinogeniculate projections in animals with unilateral eye enucleation on the day of birth Since the unilateral eye removal was carried out on day 0, the development of the fibers from the remaining eye in animals of the same age group was not studied. The distribution of the contralateral fibers in the 1- to 6-day-old experimental animals (Fig. 6A, C) is similar to that of the normal hamsters and therefore no further description will be presented. Day 1. There was a slight suggestion that the distribution pattern of the ipsilateral fibers in the experimental day 1 animal (Fig. 6B) were different from that of the normal day 1 animals. The projections were concentrated in a region close to the optic tract, although a few axons could be detected in the ipsilateral/3 sector. Day 2. The differences between the normal and abnormal ipsilateral retinogeniculate projections were quite obvious on day 2. Unlike the normal animals, there seemed to be fewer retinal fibers and they distributed in a more restricted area in the ipsilateral nucleus. Optic axons were concentrated in the dorsolateral region of the LGd close to the optic tract (Fig. 6D). Only a few axons were found in the medial half of the a sector or in the/3 sector of the ipsilateral LGd. Thus, it seemed that the quantity of ipsilateral fibers in the experimental animals was less than that in the intact animals. In addition, most of these fibers did not penetrate as deeply into the LGd as those in the intact animals of the same age. Day 3. In the ipsilateral LGd of the day 3 experimental animals, dense retinal fibers have penetrated deeper into the nucleus occupying the lateral threefourths of the a sector of the LGd (Fig. 7B). A few axons were found inside the ipsilateral/3 sector. Day 4. The atrophy of the ipsilateral LGd was quite obvious in the day 4 experimental animals. Un-

crossed retinal fibers were found in the entire ct sector of the nucleus in day 4 and older animals. As in the normal day 4 animals, the uncrossed fibers innervated more densely the dorsal than the ventral portion of the LGd. Day 6. In the experimental day 6 animals, the ipsilateral fibers began to concentrate in the dorsomedial area of the LGd. However, the area occupied by these dense fibers was larger than normal (Fig. 7D). In addition, unlike in the normal animals a moderate amount of HRP reaction products could be found in the ventral part of the LGd. Day 7. The abnormal distribution pattern of the contralateral fibers was detected for the first time in day 7 experimental animals. Unlike the normal day 7 animal, the entire a sector of the contralateral LGd was filled with retinal fibers and there was no observable reduction in fiber density in the dorsomedial portion of the nucleus (Fig. 7E). The distribution pattern of the ipsilateral projection was similar to that of the previous day (Fig. 7F). Day 8 and older. In animals with one eye removed at birth and the remaining eye injected on day 8 or later, the distribution pattern of the retinal fibers in the ipsilateral and contralateral LGd was similar in appearance to those of the day 7 animals. There was no decrease of retinal fibers in the dorsomedial region in the a sector of the contralateral LGd (Fig. 8A and C), and therefore a medial ~hole" which was observed in the normal animals was not detected in the experimental animals. In addition, the area in the dorsal portion of the LGd occupied by dense ipsilateral retinal fibers remained anomalously enlarged, and the ventral portion of the nucleus was occupied by a moderate amount of uncrossed retinal axons (Fig. 8B and D). DISCUSSION

Normal development of the retinogeniculate projection The use of HRP as an anterograde tracer has provided a very sensitive method for the study of retinofugal projections in the developing hamster. With the aid of this technique, ipsilateral optic axons were observed to innervate the LGd on the day of birth; these were not detected in previous autoradiographic or degeneration studies 11,37. The results, however,

202

Fig. 8. A and B: contralateral and ipsilateral L(Sd, respectively, of an experimental day 8 hamstcr. C and D: contralateral and ipsilateral LGd, respectively, of an experimental day 14 hamster. Note that the whole contralateral LGd at both ages is filled with retinal fibers. The area of fibers in the ipsilateral LGd is enlarged.

203 support previous observations that the development of the contralateral fibers is slightly ahead of the ipsilateral ones. This study also confirmed the earlier finding that there are two phases in the normal development of retinogeniculate projections 1k36,37. In the first place, the LGd receives progressively denser and overlapping projections from both eyes. In the second phase, the inappropriate projections are eliminated, beginning on day 6, and the fibers from the two eyes eventually segregate to form the adultlike pattern by postnatal day 8.

Lack of normal restriction of the retinogeniculate projection in the experimental animals In hamsters with unilateral eye removal at birth, the distribution pattern of the retinogeniculate projections from the remaining eye was studied carefully on various closely spaced days until postnatal day 17. The second phase of development was not observed as a progressive restriction of retinogeniculate fibers to form the adult pattern as in normal animals did not take place at any stage of development. Thus, we have not observed the possibility that crossed retinal fibers from the remaining eye might have become restricted during early development but later sprouted and refilled the ipsilateral area. Similarly, the uncrossed fibers did not become restricted during development but remained to be broadly distributed in the LGd. It is unlikely that the restriction might be delayed later than day 17 because major rearrangements of retinal axons in hamsters are usually completed by day 14 as indicated in another study35. Thus, we can conclude that fibers from the remaining eye are not able to form the adult pattern by interaction with the local substrate in the dorsal lateral geniculate nucleus. In order for the retinogeniculate fibers to form the adult-like pattern it is important to have the fibers from both eyes present during early postnatal development. These results support the earlier proposal that axo-axonal interaction is important in the segregation phenomenon 37.

Possible factors underlying the axo-axonal interaction hypothesis The increase in density of the terminal arborization of the axons may enhance the competitiveness of the axons to form or maintain their mature distribution pattern since the segregation phenomenon oc-

curs as the two populations of retinogeniculate projections reach a certain density. The possible importance of the terminal density factor has also been illustrated for the retinocollicular fibers in another experimental situation 34. The actual interaction between the retinogeniculate axons, however, could be carried out via physical and/or chemical 1 means. The fibers could push and shuffle each other into their appropriate location by their reactions to direct contact, or the retinogeniculate axons may inhibit or accelerate each other's growth by releasing a substance in the local environment. Both of these may be influenced by neural activity 23,24 which may be different for different populations of retinal ganglion cells. If electrical activity is involved, it does not seem to require patterned sensory input because the adult distribution of the retinogeniculate fibers is established prior to the time of eye opening around day 15. In addition, diffuse light stimulation through the eyelid does not seem to be important because retinogeniculate projections in hamsters dark-reared since birth still segregate to form the adult pattern on day 838. Whether spontaneous activity of the retinal ganglion cells might be important in the segregation phenomenon remains to be investigated.

Delay in growth of the ipsilateral retinogeniculate projection in normal hamsters and its further delay in experimental animals In this study, we have confirmed the previous observation that in the hamster the ipsilateral fibers appear later than the contralateral fibers 1k37. Similar observations have been made in the opossum 4 and ferret3. This raises the possibility that the pioneering axons of the contralateral eye might provide some form of guidance for the later-arriving fibers from the ipsilateral side. This hypothesis was tested in the present series of experiments by surgical removal of the contralateral fiber population by means of unilateral enucleation at birth. In these animals, there seem to be less ipsilateral retinogeniculate fibers, and most of them do not penetrate as deeply into the LGd in the first two postnatal days when compared to the normal animals. One interpretation is that once the crossed afferent fibers were removed, the uncrossed fibers lost their guidance, and as a result it took them a longer period of time to reach the target neurons. However, one cannot rule out the factor of

204 increase of vacated terminal space as a result of removal of the contralateral projections as underlying this delayed innervation phenomenon. In addition, the formation of axonal debris as a result of enucleation might also hinder the growth of the ipsilateral fibers.

Enlargement of the ipsilateral retinogeniculate projections after unilateral eye enucleation at birth Previous studies have shown that there was a dramatic increase of the uncrossed retinal axon termination area in the LGd after neonatal unilateral enucleation in the rat 19 and mouse 12. However, a similar surgical operation in the cat produced a much less marked effect 13. In both rabbit s and hamster *0, no expansion of the uncrossed pathway in the L G d was observed. In view of this, we have re-examined this issue in animals with one eye removed at birth using the more sensitive anterograde H R P technique. The results in the present study indicate that the ipsilateral projection in the enucleated animals was anomalously enlarged from day 6 onwards when compared to the normal animals. Instead of concentrating in the dorsomedial portion of the nucleus as in normal animals, the dense expanded projection covers almost the dorsal two-thirds of the LGd, and a moderate projection is also found anomalously located in the remaining ventral portion of the nucleus. This increase cannot be explained by the shrinkage of the nucleus because the absolute size of the increased anomalous retinogeniculate projection is larger than that of the normal ipsilateral projection. This is in contrast to what was previously reported ~0. The difference might be because H R P is a much more sensitive anterograde tracing techniqueS5, 36, and is therefore able to demonstrate projections that cannot be detected by other anatomical methods.

REFERENCES 1 Aguilar, C. E., Bisby, M. A., Cooper, E. and Diamond, J., Evidence that axoplasmic transport of trophic factors is involved in the regulation of peripheral nerve field in salamanders, J. Physiol. (Lond.), 234 (1973) 449-464.

2 Brown, M. C., Jansen, J. K. S. and Van Essen, D., Polyneuronal innervation of skeletal muscle in new-born rats and its elimination during maturation, J. Physiol. (Lond.), 261 (1976) 38%422. 3 Card Linden, D., Guillery, R. W. and Cucchiaro. J., The

However, the dense portion of the expanded ipsilateral projection is limited only to the dorsal t~othirds of the nucleus. This suggests that the ipsilatcral fibers might have a higher affinity for cells ill the dot-sal portion of the LGd. Alternatively, the ipsilateral fibers may not be as competitive as other fibers innervating the ventral portion of the LGd. Such other competitive fibers could arise from neural structures including the visual cortex or the superior colliculus. While the enlarged retinogeniculate projections induced by unilateral enucleations can be interpreted as a result of a failure of the elimination of transient a x o n s : or a reduction in the loss of retinal ganglion cells 31, recent studies in the rat 7,~4-~<:° have suggested axonal sprouting may also contribute to the enlarged projection observed. This is consistent with our present results that the enlarged retinogeniculate projections observed in the mature animals with one eye removed at birth is at least as dense, if not denser, and more extensive than the ipsilateral projection in the normal animals at any stage of development. Thus, we are led to suggest that both lack of elimination of transient fibers and sprouting of these and other optic axons seem to take part in generating the abnormally enlarged retinogeniculate projections in early unilaterally enucleated hamsters. ACKNOWLEDGEMENTS

The authors would like to express their gratitude to Dr. D. Tay for his comments on the manuscript, Miss L. C. Li and Mr. K. C. Lau for technical assistance and Miss P. S. Kwan for secretarial help. This study is supported by research grants from University of Hong Kong, Wing Lung Bank Medical Research Fund and Pauline Chan Medical Research Fund.

dorsal lateral geniculate nucleus of the normal ferret and its postnatal development, J. comp. Neurol., 203 (1981) 189-211. Cavalcante, L. A. and Rocha-Miranda, C. E.. Postnatal development of retinogeniculatc, retinopretectal and retinotectal projections in the opossum, Brain Res. l-t6 (1978) 231-248. Chow, K. L., Mathcrs, L. H. and Spear, P. D.. Spreading of uncrossed retinal projection in superior colliculus of neonatally enucleated rabbits, J. comp. Neurol,, 151 (1973) 307-322..

205 6 Crepel, F., Mariani, J. and Delhaye-Bouchaud, N., Evidence for a multiple innervation of Purkinje cells by climbing fibers in the immature rat cerebellum, J. Neurobiol., 7 (1976) 567-578. 7 Cunningham, T. J., Early eye removal produces excessive bilateral branching in the rat: application of cobalt filling method, Science, 194 (1976) 857-859. 8 Cusick, C. G. and Kaas, J. H., Retinal projections in adult and newborn grey squirrels, Develop. Brain Res., 4 (1982) 275-284. 9 Errington, M. L., Stifling, R. V. and Bliss, T. V. P., The development of afferent pathways converging on CA 3 pyramidal cells of the rat hippocampus, Neurosci. Lett., Suppl. 1 (1978) 29. 10 Frost, D. O. and Schneider, G. E., Normal and abnormal uncrossed retinal projections in Syrian hamsters as demonstrated by Fink-Heimer and autoradiographic techniques, Neurosci. Abstr., 2 (1976) 812. 11 Frost, D. O., So, K.-F. and Schneider, G. E., Postnatal development of retinal projections in Syrian hamsters: a study using autoradiographic and anterograde degeneration techniques, Neuroscience, 4 (1979) 1649-1677. 12 Godement, P., Saillour, P. and Imbert, M., The ipsilateral optic pathway to the dorsal lateral geniculate nucleus and superior colliculus in mice with prenatal or postnatal loss of one eye, J. comp. Neurol., 190 (1980) 611-626. 13 Guillery, R. W., Experiments to determine whether retinogeniculate axons can form translaminar collateral sprouts in the dorsal lateral geniculate nucleus of the cat, J. comp. Neurol., 146 (1972) 407-420. 14 Jen, L. S. and Lund, R. D., Experimentally induced enlargement of the uncrossed retinotectal pathway in rats, Brain Res., 211 (1981) 37-57. 15 Jen, L. S., So, K.-F. and Chang, A.-B., An anterograde HRP study of the retinotectal pathways in albino and pigmented guinea pigs, Brain Res., 263 (1983) 331-335. 16 Lam, K., Sefton, A. J. and Bennett, M. R., Loss of axons from optic nerve of the rat during early postnatal development, Develop. Brain Res., 3 (1982) 487-491. 17 Land, P. W. and Lund, R. D., Development of the rat's uncrossed retinotectal pathway and its relation to plasticity studies, Science, 205 (1979) 698-700. 18 Lichtman, J. W. and Purves, D., The elimination of redundant preganglionic innervation to hamster sympathetic ganglion cells in early postnatal life, J. Physiol. (Lond.), 301 (1980) 213-228. 19 Lund, R. D., Cunningham, T. J. and Lund, J. S., Modified optic projections after unilateral eye removal in young rats, Brain Behav. Evol., 8 (1973) 51-72. 20 Lund, R. D., Land, P. W. and Boles, J., Normal and abnormal uncrossed retinotectal pathways in rats: an HRP study in adults, J. comp. Neurol., 189 (1980) 711-720. 21 Maxwell, P. E. and Land, P. W., Development of retinogeniculate projections in albino and pigmented rats, Anat. Rec., 199 (1981) 165A. 22 Mesulam, M.-M., Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: a noncarcinogenic blue reaction-product with superior sensitivity for visualizing neural afferents and efferents, J. Histochem. Cytochem., 26 (1978) 106-117. 23 Meyer, R. L., Tetrodotoxin blocks the formation of ocular

dominance columns in goldfish, Science, 218 (1982) 589-591. 24 O'Brien, R. A. D., Ostberg, A. J. C. and Vrbov~, G., Observations on the elimination of polyneuronal innervation in developing mammalian skeletal muscle, J. Physiol. (Lond.), 282 (1978) 571-582. 25 Oswaldo-Cruz, E. and Rocha-Miranda, C. E., The Brain of the Opossum (Didelphis marsupialis), Instituto de Biofisica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, 1968. 26 Price, J. L., Moxley, G. F. and Schwob, J. E., Development and plasticity of complementary afferent fiber systems to the olfactory cortex, Exp. Brain Res., Suppl. 1 (1976) 148-154. 27 Rakic, P., Prenatal genesis of connections subserving ocular dominance in the rhesus monkey, Nature (Lond.), 261 (1976) 467-471. 28 Rakic, P., Genetic and epigenetic determinants of local neuronal circuits in the mammalian central nervous system. In F. O. Schmitt and F. G. Worden (Eds.), The Neurosciences, Fourth Study Program, The MIT Press, Cambridge, MA, 1979, pp. 109-127. 29 Redfern, P. A., Neuromuscular transmission m newborn rats, J. Physiol. (Lond.), 209 (1970) 701-709. 30 Rose, J. E. and Malis, L. I., Geniculo-striate connections in the rabbit. II. Cytoarchitectonic structure of the striate region and of the dorsal lateral geniculate body; organization of the geniculo-striate projection, J. comp. Neurol., 125 (1965) 121-140. 31 Sengelaub, D. R. and Finlay, B. L., Early removal of one eye reduces normally occurring cell death in the remaining eye, Science, 213 (1981) 573-574. 32 Sanderson, K. J., Dixon, P. G. and Pearson, L. J., Postnatal development of retinal projections in the brushtailed possum, Trichosurus vulpecula, Develop. Brain Res., 5 (1982) 161-180. 33 Shatz, C. J. and DiBerardino, A. C., Prenatal development of the retinogeniculate pathway in the cat, Neurosci. Abstr., 6 (1980) 485. 34 So, K.-F., Development of abnormal recrossing retinotectal projections after superior colliculus lesions in newborn Syrian hamsters, J. comp. Neurol., 186 (1979) 241-258. 35 So, K.-F. and Schneider, G. E., Abnormal recrossing retinotectal projections after early lesions in Syrian hamsters: age-related effects, Brain Res., 147 (1978) 277-295. 36 So, K.-F. and Schneider, G. E., Postnatal development of retinogeniculate projections in Syrian hamsters: an anterograde HRP study, Neurosci. Abstr., 4 (1978) 127. 37 So, K.-F., Schneider, G. E. and Frost, D. O., Postnatal development of retinal projections to the lateral geniculate body in Syrian hamsters, Brain Res., 142 (1978) 343-352. 38 So, K.-F., Woo, H. H. and Jen, L. S., Effect on segregation of retinogeniculate projections in early unilateral enucleated or dark-reared hamsters, Neurosci. Abstr., 8 (1982) 816. 39 Williams, R. W. and Chalupa, L. M., Prenatal development of retinocollicular projections in the cat: an anterograde tracer transport study, J. Neurosci., 2 (1982) 604-622. 40 Yang, S. Z., So, K.-F. and Jen, L. S., Connection between the superior colliculus and lateral geniculate body in the golden hamster, in preparation.