|
Reprinted from: Journal of Comparative Neurology, 194, 413-426, 1980.
ABSTRACT
ABSTRACT
The effects of monocular lid-suture deprivation on development were evaluated by measurement of cell sizes in the lateral geniculate nuclei of six deprived and three normally reared galagos. In all animals autoradiographic demonstration of the retino-thalamic projections from one eye was used to define the lamination and distinguish the monocular from the binocular segment of the nucleus. Our results indicate that monocular deprivation significantly affects cell growth in both the binocular and monocular geniculate segments, with the greater change occurring in the binocular segment, suggesting that both visual experience and binocular competitive interactions influence geniculate cell growth in these primates. In animals forced to use their deprived eye for 2 months or more by reverse suturing, disparity of cell sizes is reduced in the monocular segment, while differences in binocular segment cell sizes are maintained. Our results also show that monocular deprivation with or without later reverse suture has an unequal influence on different geniculate layers, such that cells in laminae 4 and 5 are not as severely affected as the remaining layers. This differential influence could relate either to the unique pattern of projection of these layers to cortex or to functional differences between layers.
There is now ample evidence in a variety of species demonstrating that geniculate cell growth is affected by binocular visual interactions. Hence, when one eye is sutured the competitive balance is upset such that an asymmetry in cell size in the binocular segment of the nucleus develops (Wiesel and Hubel, 1963, 1965; Guillery and Stelzner, 1970; Guillery, 1972, 1973; Headon and Powell, 1973; Sherman et al). In the act and monkey these changes in geniculate cell size have been found to correlate with changes in the distribution of geniculocortical axon terminals in layer IV of the visual cortex. Thus, monocular deprivation produces a decrease in terminal arbors of deprived geniculate cells and a corresponding increase in terminal arbors of nondeprived cells (Wiesel and Hubel, 1974; Hubel et al).
In both cat and monkey the competitive interactions which produce these changes appear to involve neighbors in both thalamus and cortex; i.e., the geniculate layers receiving from the left and right eyes send input to adjacent cortical columns (Hubel et al., 1976; Shatz et al., 1977). At present it is unclear whether such competitive interactions depend upon competitor proximity in the geniculate nucleus, visual cortex, or both. For example, we have shown that monocular deprivations by lid suture can produce unequal effects on geniculate layers in a species (the tree shrew) in which cells in one geniculate layer (layer 3) are further from possible competitive interactions within the geniculate nucleus and have different projections to cortex (see Casagrande et al., 1978; Carey et al., 1979). Thus, in the tree shrew geniculate layer 3 is bounded on both sides by layers receiving input from the same eye: the remaining geniculate layers all lie adjacent to a layer or layers receiving input from the opposite eye. In turn cells in geniculate layer 3 appear to project to layers or sub-layers of striate cortex which do not receive input from the remaining geniculate layers. Therefore, one could argue that cells in layer 3 in this species are morphologically excluded from the competitive interactions affecting the other layers simply by their seduced position and differential projections.
In the present study we wished to examine further the importance of proximity between competing cell layers using the galago, a species in which geniculate layers 3 and 6 are bounded by laminae innervated by the same eye and thus may be excluded from competitive influences. In the galago, the three middle layers 2, 3, and 4 are innervated by the ipsilateral eye: the remaining layers receive a contralateral retinal input. If proximity in the geniculate nucleus is an important factor in binocular competition, one might predict that in this species cells in layer 3 and 6 might show less of a reaction to abnormal binocular competitive interactions, since they lie further from potential competition. On the other hand, if proximity between geniculocortical afferents are crucial in such competitive interactions, then one might not expect to see differences in cell reaction, since in galago geniculate, cells innervated by the two eyes project to matching regions of adjacent ocular dominance columns (Glendenning et al., 1976; Hubel et al., 1976; Casagrande et al., 1977).
The second and more general objective of this study was to correlate geniculate cell size changes in monocularly deprived galagos with the visual-motor and perimetry deficits discovered earlier in these same animals (Joseph and Casagrande, 1980). Preliminary reports of our findings have appeared earlier (Casagrande et al., 1977, 1978; Joseph and Casagrande, 1978).
METHODS AND MATERIALS
Subjects
Three normal and six monocularly deprived galagos served as subjects for these experiments. Seven of these animals belonged to the species Galago crassicaudatus, and two (76-104S; 76-106S) were Galago senegalensis.
A brief history of each subject is given in Table 1. Four (77-59; 77-60; 78-26; 78-27) of the nine galagos used also served as behavioral subjects (see Joseph and Casagrande, 1980); the remaining animals were used only for anatomical study. Rearing conditions for all animals and methods used for lid suture are described elsewhere (Joseph and Casagrande 1980).
Surgical procedures
In order to aid in laminar and monocular segment identification all animals were injected in one eye with H praline in saline (500 uCi in 25 ul in 78-26 and 78-27, and 2.5 to 3.0 mCi in 25-60 ul for the remaining animals). In four cases (760-124; 77-59; 78-26; 78-27) the deprived eye was injected and in two (76-104; 77-60) the open eye was injected. The animals receiving the smaller doses of tracer were then allowed to survive for 2 to 4 weeks prior to death. The longer survivals allowed for identification of transneuronally transported label to cortex. Preliminary results of these data have already been reported (Casagrande et al., 1977) and a longer report is in preparation.
Historical procedures
The brains were fixed by transcardiac perfusion of 0.9% saline followed by 10% formalin or 2% buffered paraformaldehyde and frozen sectioned sagittally at 25 um or 30um. Every section through the lateral geniculate nucleus was stained with cresyl violet. For the rest of the brain at least every tenth section was mounted and stained. Selected sections throughout the brain were also prepared for autoradiography (Cowan et al., 1972; Edwards, 1972).
Cell Measurements
The cross-sectional areas of 350 cells in both lateral geniculate nuclei of each normal animal and 350 or 700 cells in corresponding geniculate regions in each deprived galago were measured as previously described (see Guillery and Stelzner, 1970; Casagrande et al., 1978). Measurements from the binocular segment were taken from selected matching regions on each side near the middle of the nucleus defined as that region extending from the optic disc representation to approximately 300 um medial to this landmark. Either 25 or 50 cells were measured in each of the six layers. Measurements were taken in each layer at points equidistant from the optic disc representation –i.e., at points which should represent the same point in visual space (Kaas et al., 1972). Within the monocular geniculate segment, located at the rostrolateral margin of the nucleus, either 25 or 50 cells were measured within matching regions on each side of the brain. Within the monocular segment, approximately equal numbers of cells were selected from each layer where these layers were distinguishable.
Cross-sectional areas of individual perikarya were measured by tracing the outlines at 1000x with nucleoli in focus using a Zeiss camera lucida drawing tube. Final measures were made using an electronic planimeter.
RESULTS
The normal geniculate nucleus
The organization and input to the lateral geniculate nucleus in galagos has been previously described at length (Campos-Ortega and Glees, 1967; Ionescu and Hassler, 1968 et al.). In this species six layers can be identified in Nissl preparations. As in other primates. Layers 1 and 2 are magnocellular and form the ventral-most layers. On the basis of cytoarchitecture, the four remaining laminae can be divided into two sets: Layers 3 and 6 are broad layers containing darkly staining cells of medium size, and layers 4 and 5 are narrow layers containing small pale-staining cells. In agreement with previous investigators (Hassler, 1966; Kaas set al., 1978; Norden 1979) we suggest that layers 3 and 6 are likely homologous to the parvocellular layers of simian primates but that layers 4 and 5 are a peculiarity of galagos and other prosimians.
As mentioned earlier, the three central layers 2, 3, and 4 in galago receive afferent input from the ipsilateral eye while the remaining layers (1, 5, and 6) receive a contralateral retinal input (see Fig. 1A). Interruptions in layers 6, 5, and 1 (see Fig 1B) indicate the position of the optic disc representation within the nucleus. At the rostral border of the nucleus, all three of the contralaterally innervated layers extend beyond the remaining layers and join together to form the monocular segment of the nucleus (arrows in Figure 1A indicate the borders between the binocular and monocular segments). Within this segment layers 5 and 6 rapidly fuse but appear to remain segregated from layer 1 except at the rostral tip.
In galago, several investigators have indicated the presence of an extra or seventh layer referred to as layer “S” or “O” (Tigges and Tigges, 1970; Campos-Ortega and Hayhow, 1970; Kaas et al., 1978; NOrden, 1979). It has recently been suggested by Kaas et al. (1978) that on the basis of position, this layer is homologous to one of a set of cell groups in simian primates located between the optic tract and layer 1. Deprivation effects on this magnocellular layer are of interest as it represents an incomplete ipsilaterally innervated layer within the binocular segment of the nucleus with no contralaterally innervated counterpart.
Based upon reconstructions of autoradiographs from both normal and deprived animals as well as cell measurements, it became apparent to us that the extra layer should not be considered a separate layer but that it instead represents several disrupted displaced segments of the overlying ipsilaterally innervated magnocellular layer (layer 2). This situation is perhaps comparable to what has been described for the human geniculate nucleus. In humans, Hickey and Guillery (1979) observed a number of cases in which the magnocellular layers were disrupted and patches of these layers appeared to break through and lie above the overlying parvocellular layers. A comparable situation is seen galago (see Figure 2). In Figure 2 three sets of roughly matched autoradiographs show the pattern of label following an eye injection in a normal galago. Figure 2A and 2B show the laminar organization within the center of the nucleus while the remaining sets show the laminar pattern as one progress medially within the nucleus. It can be seen in Figure 2C that as layer 2 begins to show disruptions, matching islands or strips (the “S” layer) of ipsilateral label appear where layer 1 would normally be located.
In some portions of the nucleus, one can rather convincingly follow corresponding displacements of the magnocellular layers and show continuity between “S” strip and layer 2 (see Fig. 2E, F). In other parts of the nucleus (Fig. 2C, D) this relationship is less clear. Curiously, inspection of a number of brains indicates that considerable individual variability exists in the extent and arrangement of these magnocellular islands. In spite of this variability, however, our analysis of label pattern and size of cells in each case is consistent with the idea that the “S” layer in galago represents a disrupted and displaced segment of layer 2 rather than in a separate layer. Thus, we have not included separate measurements for this cell group in the quantitative descriptions which follow.
Table 2 shows mean cell areas obtained from three normal adult galagos. These measurements confirm the description given earlier showing that the six laminae of the geniculate nucleus can be divided on the basis of perikaryal size into three matched sets of ipsilaterally and contralaterally innervated layers; where layers 1 and 2 have large cells 3 and 6 have cells of medium size, and 4 and 5 have small cells. The proportionally smaller cells in case 76-106S (a galago senegalensis) can be attributed to smaller brain and body size of this species as compared to the remaining two adult galago crassicaudatus.
Deprivation effects on the lateral geniculate nucleus
A marked difference in cell size and appearance of deprived and nondeprived geniculate laminae is apparent in all of the monocularly lid-sutured galagos. Figure 3 shows the appearance of the geniculate nucleus in deprived galago 76-124. As can be seen from the photograph, cells in the deprived layers in the binocular segments are shrunken in comparison with their nondeprived counterparts. This can be seen most clearly by comparing the four main geniculate layers (1, 2, 3, and 6) on the deprived side with those on the nondeprived side. Changes are much less apparent in the normally smaller cells of layers 4 and 5.
As can be seen from Table 3, significant differences (p<0.01) exist between the cell sizes in layers innervated by the deprived and experienced eyes. Overall, these differences are greatest in the main geniculate layers (1, 2, 3, and 6) within the binocular segment and least (although still significant in some cases) in the monocular segment. Moreover these differences are as great for an animal sutured at 38 days (78-27) as for animals sutured within first week after birth.
Two other noteworthy points concern the cell measurements. The first relates to the individual variability in deprivation effects on monocular segment cells. As indicated in Table 3, significant asymmetries in monocular segment perikaryal size are present only in the animals (76-104S; 76-124) which were not used for behavior. Assuming this result is not simply an artifact of normal variability and sample size, it suggests that late visual experience with the closed eye might reverse deprivation effects produced earlier. Experience per se, however, does not appear to change the overall effect of deprivation on cells in the binocular segment, since asymmetries exist in animals that were used for behavior as well as those that were not.
A second point concerns the differential effect of monocular deprivation on the individual geniculate laminae in the binocular segment. Inspection of perikaryal size shifts by layer. (Table 3) indicates that the small cell laminae 4 and 5 show less mean change than the others. This is particularly striking for lamina 5, which shows no significant asymmetry in mean cell area in four out of the six deprived galagos. Even when there is a significant mean change in cell size in layers 4 and 5 as in 76-104S, histograms (Fig. 4) showing overall shifts in the frequency distribution of cell sizes in these layers are indicative of a milder change. Apparently, as in tree shrew (Casagrande et al., 1978), the smallest cell layers are less affected by deprivation than the others. Similar comparisons between mean perikayal size in the remaining parvocellular layers (3 and 6) and the two magnocellular layers (1 and 2) suggest that cells in all four of these laminae are about equally affected by the deprivation.
Additional morphological observations
As mentioned earlier (Methods and Materials) all of the galagos in the study had one eye injected with H praline and selected brain sections processed for autoradiography. In each animal we examined the distribution of retinal projections to all known central targets. Comparisons of projection patterns from normal, deprived, and nondeprived retinae reveal no obvious differences. Moreover there is no evidence of retinal projections to inappropriate geniculate layers.
DISCUSSION
General considerations concerning cellular changes in the lateral geniculate nucleus
The prevailing theory concerning the effects of monocular pattern deprivation on geniculate cell size is that abnormal binocular competitive interactions between the eye prevent normal cell growth within the binocular geniculate segment such that deprived binocular segment geniculate cells are more severely affected than their deprived counterparts in the monocular segment. The present results in deprived galagos provide further evidence in support of this view by showing major changes in mean geniculate cell areas in the binocular segment of the geniculate nucleus in contrast to minor size changes in the monocular segment of the nucleus
Several investigations have also shown that monocular eye closure produces pure deprivation effects in geniculate neurons in the form of changes in soma size n the monocular segment. These changes have been reported for cats (Hickey et al., 1977) and monkeys (von Noorden et al., 1976) but not tree shrews (Casagrande et al., 1978), dogs (Sherman and Wilson, 1975) or squirrels (Guillery and Kaas, 1974). In every case these pure deprivation effects are reported to be weaker than those produced by competitive interactions. Our results add to this evidence by showing a small but significant reduction in cell growth in the monocular geniculate segment in two out of six of our galagos. The puzzle concerning these results is that effects are not found in all of the subjects. One explanation for this variability could lie in a problem of sampling. In primates such as galago, it is critical to select exactly matched regions within the monocular segment, since magnocellular and parvocellular regions remain segregated within this segment except for the extreme rostral tip. To overcome loss of laminar borders we made a special effort to sample equal numbers of cells from regions which could be cytoarchitectonically defined as either parvocellular or magnocellular and to remain consistent in such selections between subjects.
Thus, it seems unlikely that the variability between our cases can be attributed to errors in sampling. A second possibility is that our samples were too small to detect significance in some subjects. In order to rule out this factor, we measured an additional 100 cells in each case, with no change in result. The most viable explanation for variability in monocular segment cell measurements, as mentioned previously, may lie in the experimental history of our subjects. Those animals in which a significant deprivation effect could be detected in the monocular segment had little or no visual experience using their deprived eye. In contrast, animals in which there was no significant effect were reverse-sutured at least 2 months prior to death. This additional visual experience may have modified or reversed original deprivation changes within the monocular segment. It is however noteworthy that the experienced animals were indistinguishable from the inexperienced in terms of deprivation effects on the binocular segment of the nucleus. Thus, assuming visual experience is a factor, it does not appear to reverse pure deprivation effects (if they exist) in areas where binocular competition still remains a factor. Finally, since the animals without visual experience show less variation about the mean, demonstration of a clear correlation with experience will obviously require additional cases for comparison.
Laminar differences in deprivation effect
The possibility that binocular competition disproportionately affects cell growth in different geniculate laminae is likely to be of some importance in understanding both basic mechanisms involved in binocular interactions and the origin of laminar differences in the geniculate nucleus itself. Our concern originally was based on the observation that in tree shrew perikaryal size in the geniculate layer with the smallest cells (layer 3) remains unchanged by monocular lid closure (Casagrande et al., 1978). Several alternatives exist as explanations for the results in tree shrew. For example, since cells in this layer and their connections in cortex are more distant from potential competing elements than are cells in other layers, it is possible that cells in layer 3 are secluded from competitive effects as a result of physical separation. Another possibility is that this layer is composed of a cell type not influenced by pattern deprivation (perhaps W-cells). The latter possibility is especially attractive because of our present results in galago and more recent results in cats (Hickey, 1980). Our reasoning is as follows: If proximity between competing elements in the geniculate nucleus were important in abnormal binocular interactions than one would expect layers 3 and 6 to be secluded from competitive interaction. This turns out not to be the case. If proximity between competing geniculocortical axons were crucial, one would expect cells in all geniculate layers in both galago and cat to be changed equally by competitive interaction. This turns out not to be the case either. In both galago and cat the smallest cell laminae (4 and 5 for galago and C1 and C2 for cat Hickey, 1980) are changed much less by deprivation than are the remaining layers.
For the sake of our argument, it is now important to recall that in tree shrew the smallest cell layer is also affected by binocular competition (Casagrande et al., 1978). This then raises the possibility that the small cell layers in all three species share in common features which make them less susceptible to abnormal binocular competitive interactions. This point is further reinforced by several other morphological similarities in the small cell laminae which contrast them with the remaining geniculate layers in all three species. These similarities include a sparser projection from the retina and a heavier input from the superior colliculus and striate cortex (Niimi et al., 1970; Graybiel, 1972; Casagrande, 1974; Graham, 1977 et al). In light of these similarities it would be especially interesting to determine if the smallest geniculate cell layers in tree shrews and galagos receive W-cell input, as has been described for cat (Wilson and Stone, 1975; Cleland et al., 1976; Wilson et al., 1976).
It is noteworthy that our results provide no evidence of a differential deprivation effect on the remaining geniculate layers. Thus, cells in the main parvocellular (3 and 6) and magnocellular (1 and 2) layers how approximately the same mean shifts in size with deprivation. In monkey reports also indicate that cell growth in parvocellular and magnocellular layers is about equally affected by monocular lid closure (Headon and Powell, 1973; von Noorden, 1975; von Noorden and Middleditch, 1975; von Noorden et al., 1976 et al). Both of these findings are surprising in light of physiological and morphological data in cat which suggest that the largest geniculate cells are more vulnerable to pattern deprivation (Sherman et al., 1972; Hikcey et al., 1977, LeVay and Ferster, 1977 et al). For example, in cat Sherman et al. (1972) showed that one functional subclass of cells, namely Y-cells, were more affected by monocular suture than X-cells. Later several lines of evidence suggested that Y-cells might correspond to the largest geniculate cells in cat and furthermore that growth of the largest cells was in turn more affected by deprivation (Hickey et al., 1977; LeVay and Ferster, 1977; Lin and Sherman, 1978).
Based upon the above data and assuming comparable deprivation effects across species, one would predict that primates which (unlike cats) normally show segregation of X-cells and Y-cells to the parvocellular and magnocellular layers, respectively (Sherman et al., 1976; Dreher et al., 1976; Malpeli and Schiller, 1977) would show a correspondingly greater net change in cell growth in the magnocellular geniculate laminae following deprivation. Reasonable species comparisons, however, will require both physiological study of deprived Y-type geniculate cells in primates and definitive morphological identification of deprived and nondeprived X-cells and Y-cells in cats.
Behavioral considerations
In our previous paper we were able to demonstrate that monocularly deprived galagos develop visual deficits that can best be explained by some form of binocular competition (Joseph and Casagrande, 1980). Thus, tests of visual orienting ability show a normal visual field for the nondeprived eye and vision limited to the temporal periphery (monocular segment) for the deprived eye. In this paper we have presented measurements of geniculate cell sizes in four of these behavioral animals, all of which showed significantly smaller cells in deprived geniculate segments that did not function in visual orientation, i.e., the binocular segments. No significant cell size changes were revealed for those geniculate segments (the monocular segments) which did function in visual orientation. A similar positive correlation between geniculate cell changes and visual behavior has also been reported for cats (Sherman, 1973; Sherman et al., 1974) and dogs (Sherman and Wilson, 1975).
Two other features concerning the relationship between our anatomical and behavioral results merit discussion. First, the lack of significant cell size changes in the monocular segment of some of our behavioral cases contrasted with our findings for visually naïve animals. As discussed earlier, one explanation for the difference is that visual experience may reverse deprivation effects in the monocular segments under conditions of forced use. A second point concerns the critical period for producing morphological and behavioral deficits. In our behavioral animals a permanent loss of visual orienting ability within the binocular segment of the field was evident both n animals deprived starting within the first postnatal week or within the sixth postnatal week. In either case did reverse suture and forced use of the deprived eye by reverse suture cause obvious visual field recovery of an obvious reversal of morphological changes in mean cell size in the binocular segment of the geniculate nucleus. This suggest that as in cat (Wiesel and Hubel, 1965; Dews and Wiesel, 1970; Hubel and Wiesel, 1970; Movshon, 1976; Wan and Cragg, 1976; Blasdell and Pettigrew, 1978) 6 weeks postnatal in these primates is still within the critical period for producing deprivation-related changes and that many behavioral and morphological deficits produced by monocular lid closure at 6 weeks or earlier cannot be reversed by adult experience.