0306-4522/90 63.00 + 0.00 Pergamon Press plc C 1990 IBRO

Neuroscience Vol. 37, No. I, pp. 101-I 14. 1990 Printed in Great Britain

EFFECTS OF VISUAL DEPRIVATION ON THE POSTNATAL DEVELOPMENT OF THE GENICULOCORTICAL PROJECTION IN KITTENS N. KATO Institute

for Brain

Research,

Faculty

of Medicine,

Kyoto

University,

606 Kyoto,

Japan

Abstract-In kittens reared with either monocular, binocular or reverse suture, beginning before the physiological eyelid opening (around one week of age) and lasting until after one month, the cortical laminar distribution of geniculocortical afferents to area I7 was examined by using orthograde transport of wheat germ agglutinin conjugated with horseradish peroxidase, and compared with that in normal kittens. In normal kittens, at birth, the afferents were distributed most densely in layer I and, to a lesser extent, widely from the upper part of layer II to layers V or VI. After one month, the alferents were found mainly in and around layer IV and very sparsely in layer I. Neither binocular nor monocular suture affected this normal development. In contrast, when the present procedure of monocular suture had been followed by opening the sutured lid and suturing the other lid (reverse suture) for one week, the distribution was altered. The density of the alTerents in layer I was increased while the labelled terminals in deeper layers were as segregated in and around layer IV as observed in normal kittens. Such increase in density of the afferents resulted only when the injected tracer covered the medial or intermediate part of the C complex of the lateral geniculate nucleus. To confirm these findings, geniculate neurons retrogradely labelled by horesradish peroxidase injections into layer I of area 17 were examined in normal and reverse-sutured kittens. In both kinds of kittens, the labelled neurons were dense in the C complex, and absent or sparse in the A laminae. But, the number was higher in reverse-sutured kittens. These results suggest an involvement of geniculocortical layer I projections in reorganization of neuronal circuits in the visual cortex.

tangential and laminar distributions of the kitten’s geniculocortical afferents markedly change during the first few postnatal months. As for the laminar distribution, the afferents at birth are very dense in layer I and widely distributed in deeper layers whereas, after one month, they become largely segregated in and around layer IV.zo.22In the tangential direction. the distribution of the afferents devoted to one eye in layer IV, initially continuous, become

opened eye become much wider than those devoted to the formerly open, newly deprived eye.4~28*29 On the other hand, little is known about the effects of eyelid suture on the change in the laminar distribution. In the present study, the orthograde and retrograde horseradish peroxidase (HRP) methods were used to elucidate the effects of eyelid suture on the postnatal change in the laminar distribution of

The

columnar, forming ocular dominance interdigitate evenly with the columns

columns devoted

geniculocortical afferents. A particular emphasis was put on the analysis of tracer injection sites. Some of the present work has appeared in brief fotm”*”

which to the

other eye.24*25,3s it is well known that this change in the tangential direction is not affected much by binocular eyelid suture, but severely altered by monocular suture.3.34~35.‘8After dominance columns

monocular suture, the ocular of the undeprived eye become much wider and richer in innervation than those for the deprived eye. This change induced by monocular

suture is reversible during the first few months. When the sutured eyelid is opened and the other lid closed (reverse suture), the columns devoted to the newly

EXPERIMENTAL PROCEDURES

preparation Sixteen kittens were used for the orthograde study and

Animal

seven others for the retrograde study (Table 1). Six served as normal controls. The other I7 kittens were deprived of normal vision on the days indicated in Table 1. Before their eyelids opened naturally, three kittens underwent bilateral eyelid suture while the other 14 had only the right eyelid sutured. Among these I4 kittens, I I underwent reverse suture about one month later, having the right eyelid opened and the left eyelid sutured. Suture was performed under pentobarbital anaesthesia (35 mg/kg, i.p.). Injection

CL, central lateral nucleus; CM, centromedian nucleus; HRP, horseradish peroxidase; LGN. lateral geniculate nucleus; LP, lateral posterior nucleus; MIN. medial interlaminar nucleus; PO, posterior complex: WGA-HRP, wheat germ agglutinin conjugated with horseradish peroxidase.

Abhreck7tion.s:

and histology

Kittens were injected with wheat germ agglutinin conjugated with horseradish peroxidase (WGA-HRP) or HRP on the postnatal day shown in Table I. The kittens were anaesthetized with pentobarbital(35 mg/kg, i.p.) and placed on a stereotaxic holder. Respiration was always natural. For

the orthograde study, the bone and dura were removed over IO1

102

N. KATO

Table I. Postnatal

days on which the first and second suture. if any, and injection were carried out Animal code

Orthograde

WGA-HRP

Normal control

Binocular suture

Monocular suture

Reverse suture

study

NI N2 N3

First suture

-

-

Injection

30 43 60

were made by using either a camera lucida or a photographic enlarger. Nomenclature for the thalamic nuclei followed that of Berson and Graybiel.’ To delineate the cortical layers and areas, the criteria of Otsuka and Hassler for adult cats was used,” which can also be used for kittens. as shown in previous studies.5.‘j.Z0.Z?

BI B2 B3

7 7

7

39 39 44

Ml M2 M3 M4

7 7 4 4

-

43 43 50 50

Orthograde study with wheat germ agglutinin conjugated with horseradish peroxidase

RI R2 R3 R4 R5 R6

I IO 4 IO I I

43 34 41 41 38 38

50 41 48 48 45 46

-

60 45 45

44 44 38 38

51 51 45 45

binocular-sutured and monocular-sutured kittens, injected WGA-HRP was confined to the medial two-thirds of the LGN, covering both the A laminae (lamina A and lamina Al) and the C complex in both hemispheres, the terminal labelling in area 17 always resulted. There was no difference in the laminar distribution of geniculocortical afferents in area 17 among the three types of kittens. The labelled terminals were distributed densely in and around layer IV while they were sparse in layer I. In addition, terminals were probably labelled in layer VI; however, the presence of densely packed, retrogradely labelled neurons in this layer prevented clear observation. Retrogradely labelled corticogeniculate neurons were always found exclusively in layer VI, which confirms that the injected WGA-HRP was only taken up from the LGN, but not from adjacent parts of the thalamus around the LGN or the white matter, on the basis of the previous studies.“-” The injection site in a monocularly sutured kitten (M4 right) is shown in Fig. la. The injected enzyme densely covered both A laminae and C complex in the most medial part of the LGN. After this injection, orthogradely labelled terminals in area 17 were found distributed as shown in Fig. I b. The labelled terminals were dense in deep layers, but very sparse in layer I. In the other monocularly sutured kittens and all the binocularly sutured kittens, much the same laminar distribution was obtained. Thus, both binocular and monocular sutures left the normal development of the laminar distribution

Retrograde HRP study Normal control CR1

RSI RS2 RS3 RS4

-

-

CR2 CR3

Reverse suture

Second suture

frontal plane at 70pm. The sections were treated with benzidine dihydrochloride,’ and examined under a microscope without counter-staining. Later they were stained with Neutral Red and examined again. Drawings of the sections

5 5 IO IO

the middle suprasylvian gyrus. A micropipette filled with a 5% solution of WGA-HRP (Sigma) was inserted and its tip was placed in the lateral geniculate nucleus (LGN) with the aid of field potentials recorded from the LGN jn response to photic stimulation or optic nerve stimulation as described in the previous paper.2’ The enzyme was injected by passing a d.c. current through the microelectrode: 5 yA for 5-30 min in the monocularly sutured kittens, and 5 pA for S-IO min in the others. For the retrograde study, the bone and dura over the lateral gyrus were removed bilaterally. A micropipette filled with a 10% HRP (TOYOBO, G-l-C) solution in saline containing 0.5% poly+ornithine was inserted into the most medial part of the lateral gyrus, the presumed area 17, with the aid of a micro-driver (Narishige). At a depth of 50 pm, HRP was injected electrophoretically (0.5/A for OSmin, three injections for each hemisphere). The animals survived for another 24-48 h, they were then anaesthetized deeply with pentobarbital (60 mg/kg, i.p.). and perfused through the ascending aorta with 7% formalin in 0.1 M phosphate buffer and successively with a 10% sucrose solution for a few days, then frozen and cut in the

RESULTS

Normal, binocular-sutured kittens. In all the normal,

and monocular-sutured

Fig. I. (a) Bright-field photomicrograph of the frontal section containing the centre of the injection site in the LGN of the monocularly sutured kitten (M4 right). The injected WGA-HRP densely covered the medial part of both the A laminae and C complex of the LGN. (b) Dark-field photomicrograph of area I7 in the same kitten (M4 right). Dense orthograde terminal labelling is seen in and around layer IV, while labelling in layer 1 is very sparse (arrowheads). There are probably labelled terminals in layer VI, but a further observation was prevented because of many retrogradely labelled neurons in this layer. Those retrogradely labelled cells were found exclusively in layer VI, which confirms the restriction of the injected WGA-HRP into the LGN according to Kato et u/.20.*2(c) Dark-field photomicrograph taken of area I7 in the reverse-sutured kitten (RI right). Orthogradely labelled terminals are distributed densely in layer 1 (arrowheads) as well as in and around layer IV. The injection site is illustrated in Fig. 2 (RI right). The presence of the retrogradely labelled neurons exclusively in layer VI confirms that the injected WGA-HRP was confined in the LGN. This photomicrograph was taken from Kato.r6 (d) High-power, dark-field photomicrograph illustrating the detail of the photomicrograph shown in c. Scale bars = 200 pm for a and c, IOOpm for b and d.

Geniculocortical

projection in visually deprived kittens

Fig. 1

103

of penicuiocorticai afferents unchanged: the density in layer I. very high for the first three weeks, becomes sparse by one month of age, and meanwhile the terminals in deep layers, initially distributed widely and sparsely, are segregated in and around layer VI and markedly increase in density.‘O- x Reverse-sutured kittens. In contrast, reverse suture produced a different laminar distribution of labelling. In reverse-sutured kittens, whenever injected WGAHRP covered at least the medial or intermediate part of the C complex, the density of ortho~radeIy labelled terminals in layer I noticeably increased. The terminals in deeper layers, on the other hand, were as segregated as observed in normal kittens of similar ages. The labelled terminals in a reverse-sutured kitten (RI right) are shown in Fig. lc, d. In these photomicrographs taken of area I7 on the mesial surface of the right cerebrum, retrogradely labelled corticogeniculate neurons were located only in layer VI. which again shows that the injected WGA-HRP was taken up exclusively from the LGN but not from adjacent parts of the thalamus or the white matter.” ” Since monocular suture failed to cause such increase in the density in layer I. this finding means that the high density had not been continuously maintained since the earlier period, but was newly produced during the period between the second suture and the WGA-HRP injection. Even after reverse suture, no increase in the density in layer I could be observed in the hemispheres where injected WGA-HRP failed to cover the medial or intermediate part of the C complex. Successful LGN injections and cortical labelling were obtained from I I out of the total 12 hemispheres of the six reversesutured kittens (Table 1). In three out of these 1I cases where the injected WGA-HRP was confined to the medial (R3 left), intermediate (R6 left) or lateral (R6 right) part of the A laminae and in one case where only the lateral third of both the A laminae and C complex was injected (Fig. 2, R5 right), no increase in the density in layer 1 resulted (Fig. 3, R6 left). By contrast. the density increased (Fig. 3. R5 left) in six cases where the media1 {RI left, R3 right, R4 left) or intermediate third (RI right, R2 right, R5 left) of both the A laminae and C complex was covered (Fig. 2) and in one case (R4 right) where the intermediate third of the C complex but not the A laminae was involved. In some cases, labelling was not only found in area I7 but also in area 18 as seen in Fig. 3. but the labelling in area 18 is not described further in this paper. For comparison, extensive injections were produced by passing a d.c. current for a long time (30min) in two monocularly sutured kittens (M3 right. M4 left). In both kittens, the LGN was injected with an obviously far larger amount of WGA-HRP than in any reverse-sutured kittens, and the injected tracer densely covered the medial or intermediate part of both the A laminae and C complex (Figs 4, 5). However, no increase was observed in the density in

layer I, unlike in the reverse-sutured for M4 left). Retrogrude horsemdish pemridme

kittens

(Fig. 5

study

In the seven kittens used (Table I), one hemisphere in each of the three normal kittens and six hemispheres in the four reverse-sutured kittens were successfully injected with HRP into layer I (Table 2). Injected HRP was completely restricted to layer I in one normal kitten and one hemisphere of a reversesutured kitten while there was HRP spread into the upper part of layer II in the other two normal kittens and the other five cases obtained from reversesutured kittens. In all the cases, neurons were retrogradely labelled in the thalamus. Four out of these nine cases were chosen and grouped into two pairs. Each pair consists of one normal kitten and one reverse-suture kitten. and the injection sites in both kittens were of a comparable size but the injection site in the normal kitten is always larger than in the reverse-sutured one. Within each pair, injection sites and labelled neurons were compared between the reverse-sutured kitten and the normal partner. In the first pair (CR2 and RSZR), the injection sites were completely confined to layer I. Illustrations of all the sections that contain the injected HRP. obtained from the normal kitten (CR2) are shown in Fig. 6a. The injected HRP was detected in five sections. The section showing the largest HRP staining (arrow in Fig. 6a), the presumed injection centre. is photomicrographed and schematically drawn on a larger scale (Fig. 6b, c). No HRP spread can be seen underneath layer I. In the reverse-sutured kitten (RSZR) of the first pair, the injected HRP was only found in four sections. The largest HRP staining around the injection centre (Fig. 7b, c) occupied a smaller area of layer I than did the largest deposit for the normal kitten of this pair (Fig. 6b, c). The size of the second. third and fourth largest HRP staining, respectively. was also compared between the two kittens, and was always larger in the normal kitten than in the reversesutured partner. Therefore, the total amount of the injected HRP is most likely to be smaller in the reverse-sutured kitten. After these injections, thalamic neurons were labelled mainly in the C complex of the LGN and the lateral posterior nucleus {LP) in both kittens (Figs 6d, 7d). However, the number of the thafamic neurons labelled in each of the two nuclei was larger in the reverse-sutured kitten than in the normal kitten. In other words, a larger number of thalamic neurons were labelled in the reverse-sutured kitten even though the injection site was smaller than in the normal partner. A few labefled neurons were found in the A iaminae of RSZR and in the central lateral nucleus (CL) of CR2 as well (Table 2). Similar results were obtained from the second pair (CR3 and RS4), in which injected HRP spread into

ieniculocortical projection in visually deprived kittens

Fig. 2. Diagrams illustrating the injection sites in four reverse-sutured kittens (Rl, R3, RS, R6). In each case, sections taken from the more rostra1 to caudal part of the thalamus are arranged from top to bottom. In Figs 2, 4, 5, filled and dotted areas indicate dense WGA-HRP precipitate and diffusion zone, respectively. The intermediate or medial part of the LGN (including both the A laminae and C complex) was densely injected in RI (left and right), R3 (right) and R5 (left). Only the A laminae were injected in R3 (left) and R6 (left and right). Only the lateral part of the LGN was covered in RS (right).

106

N. KATO

Fig. 3. Illustrations showing orthogradely labelled terminals and retrogradely labelled neurons in RS (left) in which both the A laminae and the C complex were densely injected, and in R6 (left) in which the C complex was spared dense WGA-HRP deposit. Fine and coarse dots indicate orthograde and retrograde labelling, respectively. The orthograde labelling in layer I was much higher in RS than in R6. while that in layer IV was very dense in both cases. Retrogradely labelled neurons are seen only in layer VI. Orthograde and retrograde labelling similar to that observed in area 17 is also found in area 18.

the upper part of layer II. The largest amount of HRP was injected into these kittens in the respective experimental groups. Figure 8 shows the largest HRP staining around the injection site in the normal kitten (Fig. 8a. CR3) and in the reverse-sutured kitten (Fig. 8b, RS4). The other sections containing the injected HRP were carefully compared section by section between the two kittens as in the first pair, and again it was shown that a larger amount of HRP was injected in the normal kitten. As in the first pair, a larger number of thalamic neurons were labelled in the reverse-sutured kitten than in the normal one, and the distribution of the labelled thalamic neurons was much the same between the reverse-sutured and normal kittens (Table 2). The intrathalamic distribution pattern of the labelled neurons in this pair was very similar to that in the first pair, but the absolute number of the labelled neurons in each thalamic

nucleus was larger in the second pair than in the first pair, in which a much smaller amount of HRP was injected than in the second pair. In the other five cases (CR I, RSI L, RSI R, RSZL, RS5), the injected HRP was nearly confined to layer I except for a minor spread into the upper part of layer II. The number of labelled neurons presumably varied according to the amount of the injected HRP, but the intrathalamic distribution of the labelled neurons was similar to that in the two pairs already described. The number of the retrogradely labelled neurons in all cases is shown in Table 2. The C complex of the LGN and the LP are the major origins of the thalamocortical projections onto layer I of area 17 in both the normal and reverse-sutured kittens. In the A laminae of the LGN, labelled neurons were found in all the cases obtained from the reverse-sutured kittens

~enicui~ortic~l

projection in visually deprived kittens

107

Fig. 4. Diagrams illustrating the injection sites in the mon~ularly sutured kitten (M3f. Thalamic sections are arranged in the rostrocaudal order from top to bottom on the left side and then on the right side. A much larger amount of WGA-HRP was injected into the right LGN of this kitten than in any reverse-sutured kitten. Both injection sites cover the A laminae and C complex.

but not in all the normal kittens, which could possibly mean that there is a very sparse projection from the A laminae to layer I of area 17 in normal kittens and the density of such a projection increases after reverse suture. Alternatively, the labelling in the A laminae may be due to the HRP spread into the upper part of layer II. In the medial interlaminar nucleus (MIN), CL, and centromedian nucleus (CM), Ia~Iling was not always seen. In the posterior complex (PO), only two neurons were labelled among all the animals used. DfSCUSSlON

The present experiments showed: (1) binocular suture and monocular suture leave the normal postnatal development of the laminar distribution of geniculocortical afferents in area 17 unchanged, but reverse suture does increase the density of the afferents to layer I without affecting the normal laminar pattern of ~nicul~orti~l terminals below layer II; (2) the increased afferents to layer I in reverse-sutured kittens originate largely from the medial two-thirds of the C complex; (3) reverse suture increases the number of neurons projecting from the LGN and LP to layer I of area 17.

Consideration of the methods Orthograde method. In normal adult cats or kittens aged more than one month, a restricted injection of orthogradely transported tracers into the A laminae does not elicit terminal labelling in layer I but such an injection into the C complex does.m3*~23~26 For this reason, the present findings on layer I projections might possibly merely reflect a difference in the position of the WGA-HRP injection site centre between the monocularly sutured and reverse-sutured kittens. However, this is unlikely. Firstly, as long as the medial two-thirds of the C complex was covered with injected WGA-HRP, the labelling pattern in the reverse-sutured kittens was the same irrespective of whether the injection centred in the C complex or in the A laminae. Secondly, a far larger amount of WGA-HRP was injected into the C complex as well as the A laminae in two monocularly sutured kittens (M3, M4). In both animals, although a far larger amount of WGA-HRP was injected into the C complex than in any reverse-sutured kittens, the density in layer I was far lower. In some kittens, the injected WGA-HRP covered a part of the optic radiation just dorsal to the A laminae of the LGN. Therefore injected WGA-HRP

108

N. KATO

Fig. 5. (a) Diagrams illustrating the injection sites in the monocularly sutured kitten (M4). Thalamic sections are arranged in the rostrocaudal order from top lo bottom on the left then again on the right. A much larger amount of WGA-HRP was injected into the left LGN of this kitten than in any reverse-sutured kitten. Both injection sites cover the A laminae and C complex densely. (b) Diagrams illustrating representative sections taken from the left cortex of the same kitten (M4). Coarse and fine dots schematically indicate retrogradely labelled neurons and orthogradely labelled terminals, respectively. Terminals labelled in area 17 were distributed largely in and around layer IV, and very sparsely in layer I. Also there appeared to be many labelled terminals in layer VI, but further observation was prevented by the presence of labelled neurons in this layer. Labelled terminals were also seen in area 18.

might be taken up in the optic radiation by axons en passage. However, the possibility of such uptake can be excluded. Firstly; in the present study, there was

no difference in the laminar pattern of the labelled geniculocortical afferents in area 17 between the monocularly sutured kittens in which WGA-HRP

109

Geniculocortical projection in visually deprived kittens Table 2. Number of neurons retrogradely labelled in each thalamic nucleus by layer I injections in normal and reverse-sutured kittens Animal code Normal CR]* CR2 CR3 Reverse-sutured RSlL*t RSlR*t RS2Lt RS2Rt RS3 RS4

A+Al 1

ccx

5

3 10 67

3 1 8 11 18 25

34 20 71 50 174 146

MIN

Pul

LP

CL

CM

2

6 16 26

1 2 10

1

3

3

2 1 1

3

13

11

1

18 36 84 48 44 91

4 6 29

PO

1 5 20

1

*In CRI, neurons were looked for in only some sections. In RSIL and RSIR, neurons were counted in every third section. In the other cases, they were counted in all the sections. tL and R stand for the left and right hemispheres, respectively. Note that the amount of injected HRP varied considerably among the cases. A, lamina A; Al, lamina Al; CCX, C complex; Pul, pulvinar.

spread into the white matter, and those in which the white matter was completely spared. Secondly, retrogradely labelled neurons were distributed in layer VI alone irrespective of whether WGA-HRP injected

into the LGN spread into the white matter or not. This finding means that WGA-HRP was not taken up from the optic radiation, because the optic radiation contains corticofugal axons to the superior colliculus and the extrageniculate visual thalamus which originate from layer V.‘8.‘9.“7 Finally, a possible contribution of transneuronal transport of WGA-HRP should be considered. There is a possibility that WGA-HRP transported orthogradely or retrogradely from the LGN to deep layers might be released and taken up by cortical neurons projecting to layer I. However, trans-synaptic transport in such neuron networks should be negligible, if at all, for the following three reasons. Firstly, in the present experiments, the distribution pattern of the labelled terminals is basically the same among the kittens whose survival time after the injection varies from 24 to 48 h. This is most likely to indicate that transneuronal transport, if any, contributes little to the labelling. If transneuronal transport, which occurs subsequently to simple ortho- or retrograde transport, contributed noticeably to the labelling, the distribution patterns would differ after survival periods of 24 and 48 h. Secondly, HRP, regarded as non-trans-synaptically transported tracer, was injected into the LGN to orthogradely trace geniculocortical terminals in normal kittens in the previous study,2n but the labelling pattern was the same as that obtained by using WGA-HRP. Thus, it is most unlikely that transneuronal transport might contribute to the labelling pattern obtained in the particular projections that were investigated in the present orthograde study. Retrograde method. The injected HRP generally tends to differ in quantity from animal to animal even

though the parameters of the injection, e.g. intensity of current for electrophoresis and its duration, were adjusted to be the same. To compare the injected amount of HRP, a most reliable way would therefore be to carefully observe the HRP deposit around the injection site, section by section, animal by animal. In fact, this method was adopted in the present study. The HRP deposit was completely confined to layer I in one pair, while it spread into the upper part of layer II in the other pair. It was judged that the data obtained from the latter pair was also reliable because, firstly, in kittens aged more than one month and adult cats, the upper part of layers II/III is spared any geniculocortical afferents20*26and, secondly, the findings obtained from the present two pairs were qualitatively much the same. Quantitatively, reflecting the difference in the amount of injected HRP, the number of the retrogradely labelled neurons was much larger in the second pair. Variability of retrograde labelling in the LGN is large as seen in Table 2, and this is presumably due to the difference in the amount of the injected HRP. If the dense-core of the injection site is assumed to be completely spherical, it can be estimated by comparison between the injection sites of CR2 (Fig. 6) and CR3 (Fig. 8a) that approximately eight times as much HRP was injected in CR3 as in CR2. On the other hand, variability of the labelling in extrageniculate thalamic nuclei may be attributed to individual difference, or more likely to the sensitivity of the HRP method, as well as to the difference in amount of the injected HRP. Neurons giving rise to a projection may well be labelled in some animals but not in others if the density of the projection is so low as to be just above threshold for detection by the retrograde HRP method. Tong et aLI reported a similar variability of labelling in some of the thalamic nuclei after injections into the lateral suprasylvian visual cortex, and attributed it mainly to sparseness of the projections.

N. KATO

I IO Interpretation

qf the result

The reverse-sutured kittens were allowed to survive for one week after the second suture until they were injected. What this survival period means should be discussed. Movshon and Blakemore and MovshonZR showed that one week after reverse suture at the age of four weeks, ocular dominance distribution in area 17 was almost completely changed in favour of the newly experienced eye, even though still slightly changing. Those authors also reported that when reverse suture was made at the age of five or six weeks, it took a longer time for the distribution to switch, although half the switching process was completed by the end of the first week and the process works most rapidly around one week after the second suture. Since the reverse-sutured kittens in the present experiments underwent the second suture at the age of five or six weeks, the survival time of one week means that the reversal process of ocular dominance columns was under way presumably at the highest rate when the kittens were injected. It would be interesting to examine whether or not the increased afferents in layer I are reduced in density several weeks after the second suture, when the switching process is already completed. There are several possible explanations for the neuronal changes responsible for the increase in thalamocortical afferents to layer I of area 17. Firstly, the population of neurons that normally project to layer 1 may increase in number after reverse suture. Secondly, some neurons which originally did not project to layer I but project to another layer or more may have come to send new axonal branches to layer I. Thirdly, sprouting of axonal branches may occur within layer I or below layer I. Then it would follow that axonal arbours originating from the same number of neurons become denser than before reverse suture and the terminal field of each axon becomes wider. Among these three possibilities, the first one is unlikely because cerebral cortical neurons are generally not supposed to undergo cell division after birth in cats. Possibilities of axonal sprouting seem more feasible. The present study showed that the increased geniculocortical afferents to layer I after reverse suture originate largely from the medial and intermediate parts of the C complex of the LGN. These parts of the LGN coincide with the binocular seg-

ment.‘.“’ This coincidence is in agreement with the view that increase of the afferents to layer I is related. directly or indirectly, to plastic changes caused by reverse suture, because it is in the binocular but not monocular segment of the LGN that morphological changes in the soma of geniculocortical neurons are known to take place after the eyelid suture.9.‘0 Since the W cells of the C complex, which give rise to layer I projections, are known to mature earlier than the X and Y cells,” the present result that layer 1 projections still have a great morphological malleability during the second postnatal month, may sound surprising. It is still difficult to settle the question as to why reverse suture but not simple monocular suture induced the increase in layer I projections. An attempt was made to address this question in the previous paper.” Briefly, the principal difference between monocular and reverse suture in anatomical terms is that while monocular suture has been shown to produce, primarily, shrinkage of terminal field of geniculocortical axons subserving the deprived eye,“,” reverse suture is considered to involve expansion of the terminal field of geniculocortical axonal arbours subserving the initially deprived, newly opened eye as we11.4,2” Such expansion of the ocular dominance columns, a proliferative rather than degenerative reorganization, may somehow be related to the increase in density of geniculocortical afferents to layer I. Two other lines of evidence suggest that some neuronal mechanisms underlying the changes after monocular suture and reverse suture differ. Ary et al.’ reported that noradrenaline, known to play a role in ocular dominance plasticity after monocular suture 2.‘4~15 appears to have little to do with the neuronal events after reverse suture. Murphy and Mitchel” showed that restoration of visual acuity from deprived conditions is much better after simple monocular suture than after reverse suture. Speculation

on functional

role

qf the

increased

layer

I

projection

It has been shown that some aspects of ocular dominance plasticity are of a Hebbian type.” In the Hebbian model,” the concurrent activation of the pre- and postsynaptic elements is the prerequisite for the occurrence of synaptic plasticity. Provided that the Hebbian model is applicable to ocular dominance

Fig. 6. Illustrations showing the injection site in area 17 and the lsbelled thalamic neurons in a normal kitten (CR2). (a) Illustrations showing the HRP injection site in area 17. For this figure and Fig. 78, all the sections containing the injected HRP are shown. The largest HRP deposit around the injection centre is indicated by the arrow. The injected HRP is confined completely to layer I. (b) Photomicrograph taken from the injection centre marked by the arrow in a. Scale bar = 200pm. (c) Diagram illustrating the injection centre shown in b. Dense HRP precipitate and HRP diffusion zone are indicated by filled and dotted areas, respectively. Roman numerals on the left indicate cortical layers. (d) Diagrams showing the distributions of thalamic neurons 18belled by the HRP injection illustrated in a, b, c. Three representative sections taken from the rostra1 (left), middle (centre) and caudal (right) parts of the thalamus illustrated. The labelled neurons are distributed mainly in the C complex of the LGN and the

are

LP.

Geniculocortical projection in visually deprived kittens

a

Fig. 6

a

d

Fig. 7. Illustrations showing the HRP injection site in area 17 and the labelled neurons in the thafamus in a reverse-sutured kitten (RSZR). (a-d) Conventions are the same as given in Fig. 6.

Geniculocortical projection in visually deprived kittens

Fig. 8. Photomicrographs showing the centre of the HRP injection sites in a normal (a, CR3) and a reverse-sutured (b, RS4) kitten. In both cases, the injected HRP spread into the upper part of layer II. Dense HRP precipitate shown in a is larger than in b. Scale bar = 2OOpm for a and b. after reverse suture, layer IV cells need to be activated, shortly after the second suture, in correlation with the input valley conveyed by the LGN afferents orginating from the formally deprived, newly opened eye. But, at that period, these afferents are presumed to suffer from competitive disadvantage.27-29 The increase of layer I projections may help to overcome such disadvantages as follows. Layer I projections are in a strategic position to regulate excitability of the whole area 17 mildly and uniformly because many neurons extend their axons to layer I and the excitation around the distal end of the dendrites causes a gradual, long-lasting, mild depolarization at the soma. 32Hence the increase of layer I projections may help to increase the excitability of the postsynaptic cell which the Hebbian model requires to be activated in time with the arrival of the input valley originating from the newly opened eye. plasticity

Thalamocortical projections onlo iayer I in normal kittens

Cunningham and LeVay6 showed that layer I of area 17 in cats receives thalamic afferents from the nucleus centralis including the CL, CM, and nucleus paracentralis without specifying which of these nuclei is responsible for the afferents to layer I. The present experiments confirmed their results and further showed that among the three nuclei at least the CL and CM project to layer I of area 17. Projections from the LP and the pulvinar onto layer 1 of area 17 were revealed in the present study. Acknou4edgemenrs-I express my gratitude to Prof. K. Sasaki for generous support, continuous encouragement, and helpful advice. The manuscript was prepared partly at the Max-Planck Institute for Brain Research (Frankfurt/ M), and I wish to thank Prof. W. Singer for use of laboratory facilities. Expert assistance by Frau G. Knott, Frau H. Reitz and Frau R. Ruhl is gratefully acknowledged.

REFERENCES

I. Ary M., Pettigrew J. D. and Kasamatsu T. (1979) Manipulation of cortical catecholamines fail to affect suppression

of deprived eye responses after reverse suture. Invest. Opthal. vis. Sci. 18, 136. 2. Bear M. F. and Singer W. (1985) Modulations of visual cortical plasticity by acetylcholine and noradrenaline. Nature 320, 172-176. 3. Eterson D. M. and Graybiel A. M. (1983) Organization of the striate-recipient zone of the cat’s lateralis posterior complex and its relations with the geniculostriate system. Neuroscience 8, 337-372.

114

N.

KATO

4. Blakemore C. and Van Sluyters R. C. (1974) Reversal of physiological effects of monocular deprivation in kittens: further evidence for a sensitive period. J. Physiol., Land. 237, 195-216. 5. Cragg B. G. (1975) The development of synapses in the visual system of the cat. J. contp. Neural. 160, 147-166. 6. Cunningham E. T. and LeVay S. (1986) Laminar and synaptic organization of the projection from the thalamic nucleus centrahs to primary visual cortex in the cat. J. eonzp. Neurol. 244, 65-77. 7 DeOlmos .J. and Heimer L. (1977) Mapping of collateral projections with the HRP method. Nptlrosci. &rr. 6, 107-l 14. 8 Friedtander M. J. (1982) Structure of physjologicaliy classified neurones in the kittens dorsal lateral geniculate nucleus. #uture 300, 180-183. 9. Guillery R. W. (1972) Binocular competition in the control of geniculate cell growth. f. eom,z. Nearo,? 144, I 17-f30. IO. Guillery R. W. and Stelzner D. J. (1970) The differential effects of unilateral lid closure upon the monocular and binocular segments of the dorsal lateral geniculate nucleus in the cat. J. camp. Neural. 139, 412-422. II. Hebb D. D. (1949) 7% Organizurion of Behaoior. Wiley, New York. 12. Hubel D. H. and Wiesel T. N. (1970) The period of susceptibility to the physiological effects of unilateral eye closure in kittens. f. P&_miol.,Lo&. 206, 419436. 13. Henderson Z. (1982) An anatomical investigation of projection from lateral genie&ate nucleus to visual cortical areas I7 and 18 in newborn kittens. &p/ Brain Res. 46, 177-185. 14. Kasamatsu T. and Pettigrew J. D. (1979) Preservation of binocularity after monocular deprivation in the striate cortex of kittens treated with 6-hydroxydopamine. J. contp. Neural. IX& 139-162. IS. Kasamatsu T., Pettigrew J. D. and Ary M. (1979) Restoration of visual cortical plasticity by local microperfusion of norepinephrine. J. conlp. Nemo/. 185, 163-182. 16. Kato N. ( 1986) Increase in the density of geniculocortical afferents in layer I after reverse suture in kittens, Brai0 Res. 386, 393-396. 17. Kato N. (1986) Dense geniculocortical afferents in layer I after reverse suture in kittens originate from the C complex. Neurosci. Lert., Suppl. 26, S56.

18. Kato N. (1987) Postnatal development of the striate cortical projection onto the extrageniculate visual thalamus in the cat: an HRP study. Expl Brain Res. 67, 119-126. 19. Kato N. (1987) The postnatal development of thalamocortical projections from the pulvinar in the cat. Expl Brain Res. 68, 533-540.

20. Kato N., Kawaguchi S. and Miyata H. fl984) Genicul~orti~l projection to Iayer 1 of area I7 in kittens: orthograde and retrograde HRP studies. .I. romp. Neurol. 225, 441-447. 21. Kato N., Kawaguchi S. and Miyata H. (1986) Postnatal development of afferent projections onto the lateral suprasylvian visual area in the cat: an HRP study, J. camp. Neuroi. 228, 329-341. 22. Kato N., Kawaguchi S., Yamamoto T., Samejima A. and Miyata H. (1983) Postnatal development of the geniculocortical projection in the cat: electrophysiological and morphological studies. Expl Brain Res. 51, 65-72. 23. LeVay S. and Gilbert C. (1976) Laminar pattern of geniculo-cortical projection in the cat. Brain Res. 113, I-19. 24. LeVay S. (1984) The development of ocular dominance columns. In Neuronal Growth and Plusticity (ed. Kuno M.), pp. 271-291. Japan Scientific Societies Press, Tokyo. 25. LeVay S., Stryker M. P. and Shatz C. (3978) Ocular dominance columns and their development in layer IV of the cat‘s visual cortex: a quantitative study. J. camp. Nemo!. 179, 223-244. 26. Leventhal A. G. (1979) Evidence that the different classes of relay cells of the cat’s lateral geniculate nucleus terminate in different layers in the striate cortex. Espl Brain Res. 37, 349-372. 27. Mioch L. and Singer W. (1989) Chronic recordings from single sites of kitten striate cortex during experience-dependent modifications of receptive-field properties. J. Neurophysiol. 62, 185-197. 28. Movshon J. A. (1976) Reversal of the physiological effects of monocular deprivation in the kitten’s visual cortex. J. Physioi.. Lond. 261, 125-174. 29. Movshon J. A. and Blakemore C. (1974) Functional reinnervation in kitten visual cortex. Nurure 251, 504-505. 30. Murphy K. M. and Mitchel D. E. (1987) Reduced visual acuity in both eyes of monocularly deprived kittens following a short or long period of reverse suture. J. Neurosci. 7, 1526-1536. 31. Otsuka R. and Hassler R. (1962) Ober Aufbau und Ghederung der kortikalen Sehsphare bei der Katze. Arch. Psychiut. 2. ges. Neural. 203, 212-234.

32. Rall W. (1967) Distinguishing theoretical synaptic potentials computed for different soma-dendritic distributions of synaptic input. J. ,~eurap~~sjol~ 30, 1138-I 168. 33. Rauscheker J. and Singer W. (1981) The effect of early visual experience on the cat’s visual cortex and their possible explanation by Hebb synapses. J. Physiol., Land. 310, 215-239. 34. Shatz C.. Lindstrom S. and Wiesel T. N. (1977) The distribution of afferents representing the right and left eyes in the cat’s visual cortex. Brain Res. 131, 106-l 16. 35. Shatz C. and Stryker M. P. (1978) Ocular dominance in layer IV of the cat’s visual cortex and the effects of monocular deprivation. J. P/~ysioi., Lond. 281, 267-283. 36. Tong L., Kali] R. and Spear P. (1982) Thalamic projections to visual areas of the middle suprasylvian SU~.IS in the cat. J. eomp. Neuraf. 202, 103-117. 37. Tsumoto T., Suda K. and Sato H. (1983) Postnatal development of corticotectal neurones in the kitten striate cortex: a quantitative study with the horseradish peroxidase technique. J. camp. Nemo/. 21% 88-99. 38. Wjesel T. N. and Hubel D. H. (1963) Effects of visual deprivation on morphology and physiology ofcells in the cat’s lateral geniculate body. J. Neurophysioi. 26, 978-993. (Accepted 27 November 1989)