Experimental Brain Research
Exp. Brain Res. 33,353-369 (1978)
@ Springer-Verlag 1978
Synaptic Development in the Rabbit Superior Colliculus and Visual Cortex L.H. Mathers, Jr., K.L. Mercer, and P.E. Marshall Department of Structural Biology, Fairchild Research Center, School of Medicine, Stanford University Medical Center, Stanford, CA 94305, U. S. A.
Summary. The development of synapses in the visual cortex (VC) and superior colliculus (SC) of the rabbit has been examined with the electron microscope. In both areas, the number of synapses reaches adult levels by 2 0 - 2 5 days of postnatal age, but the development in the visual cortex is delayed in comparison to that in the superior colliculus. When S synapses (spheroidal vesicles, asymmetric thickening) are compared with F synapses (flattened vesicles, symmetric thickening), even greater differences are seen. In both the VC and SC, S synapses develop earlier than F synapses, though there is considerable overlap. Of interest is the fact that synapses in the visual cortex seem to overshoot their adult levels late in development, suggesting that an excess of synapses may be formed in this system. Multiple synapses, probably of retinal origin, increase in the first 3 weeks of synaptic development in the SC, but never are present in significant proportions in the VC. Synapse formation most often is characterized by formation of a junction and a postsynaptic thickening, followed by acquisition of synaptic vesicles. After 15 days, there is only a small number of such "non-vesicle synapses" in either the SC or VC. Key words: Synaptic development - Visual system - Visual cortex Superior colliculus - Electron microscopy. Previous physiologic (Spear et al., 1972; Grobstein et al., 1973; Mathers et al., 1974; Fox et al., 1976) and morphologic (Mathers, 1977a, b; Mathers, in prep.) studies of development in the rabbit visual system have provided an extensive description of postnatal receptive field and neuronal development. These studies showed that there is a considerable postnatal period of continued development in both the superior colliculus and visual cortex. Whereas the superior colliculus possesses all adult receptive field types by 9-10 days postnatal and is physiologically equal to the adult at 18-20 days, the visual 0 0 1 4 - 4 8 1 9 / 7 8 / 0 0 3 3 / 0 3 5 3 / $ 3.40
354
L.H. Mathers, Jr. et al.
c o r t e x d o e s n o t a t t a i n all a d u l t r e c e p t i v e t y p e s u n t i l d a y 18 a n d is n o t physiologically equal to the adult until 20-25 days. The morphologic development of neurons in these two areas nicely complements the physiologic sequence, in that the larger relay neurons (vertical cell in the superior colliculus; l a r g e p y r a m i d a l cell i n t h e v i s u a l c o r t e x ) m a t u r e e a r l i e r t h a n t h e s m a l l e r n e u r o n s ( s t e l l a t e a n d p i r i f o r m cells i n t h e s u p e r i o r c o l l i c u l u s ; s t e l l a t e a n d s o m e s m a l l p y r a m i d a l cells i n v i s u a l c o r t e x . I n e a c h o f t h e s e r e g i o n s , t h e r a t e o f d e n d r i t i c spine growth was found to be the clearest feature distinguishing the earlier developing neurons from the later (Mathers, 1977b; Mathers, in prep.). In the present study we have extended the principle of correlating morphology with physiology to the ultrastructural level, in an effort to d e t e r m i n e w h e t h e r s i g n i f i c a n t f e a t u r e s o f s y n a p t i c d e v e l o p m e n t m a y also h e l p explain the different rates of physiologic maturation in the superior colliculus and visual cortex. A portion of this work has already been reported (Mathers, 1977a).
Materials and Methods Twenty-four Dutch-belted rabbit pups were used in this study, ranging in age from 0 to 30 days. At least three animals from separate litters were studied at each age. Additional adult rabbits were also prepared. Rabbits were bred in the laboratory, and the day on which pups were first found (nests checked at 8-9 a.m.) was designated as day 0. Pups were removed from the nest and deeply anesthetized with sodium pentobarbital (this study was performed in accordance with the "Guiding Principles in the Care and Use of Animals" statement of the American Physiological Society). In many cases animals were artificially respired with room air during the initial phases of the perfusion procedure. The chest was then opened and 4% paraformaldehyde-l% glutaraldehyde in 0.tM Phosphate buffer was introduced through a 22 or 25 gauge needle into the left ventricle. No effort was made to monitor perfusion pressure, but the perfusion bottle was 70 cm above the animal, and care was taken to maintain low pressure so as to prevent hemorrhage from the nasal mucosa. The brain was quickly removed, and slices taken from both the visual cortex and the superior colliculns. Tissue in the visual cortex was removed from a point 1 mm lateral to the medial border of the area, and in the superior colliculus from the central mid-crown region. Tissues were then osmicated, block-stained in aqueous uranyl acetate, dehydrated, and embedded in Araldite 502. Sections were cut in the frontal and horizontal planes, and mounted on 150 mesh or 75 x 300 mesh grids. The grids were examined on a Siemens IA. Synapses were counted in linear or vertical sweeps across the tissue (Lurid and Lund, 1972a). A structure was counted a synapse only when a synaptic cleft and density were clearly present. Vesicles were usually but not always counted as necessary for the identification of a synapse, but at the earliest ages some structures were designated as non-vesicle synapses (NV) because their cleft and density morphologies were clearly those of a synapse and not of a desmosome or other adhesive junction. Each synapse or NV junction encountered in a sweep was recorded as to size, nature of pre- and postsynaptic structure, vesicle shape and number, and type of density. Each synapse was also charted for its position in the cortical or collicular tissue. For each animal studied, calculations of neuronal density per unit volume were calculated, using the methods of Abercrombie (1946) as further described by Cragg (1975). Synapses were counted according to their ratio to neurons (synapses/neuron). This method is more reliable than areal synaptic counts because this expression corrects for the expansion of tissue through development (Anker and Cragg, 1974). Synaptic density was calculated in this study without a correction for synapses of small size. Thus our results were probably low by 10-20% if Anker and Cragg's (1974) computations are correct. Since our purpose was principally to show change in synaptic ratios, this correction was not included, and in other respects the method was similar to that of Anker and Cragg (1974).
Synaptic Maturation in Visual System
355
The rationale behind making measurements of neuronal nuclei and synaptic thickenings is identical, and for a thorough discussion the reader is referred to Abercrombie (1946). We will present here a brief explanation. Neuronal nuclei are assumed to be near-spherical and randomly distributed in the tissue sampled. If our goal is to express nuclei/unit volume, then we cannot simply extrapolate from nuclear counts in the volume of tissue represented in our sample sections, because some nuclei will appear in more than one section, and counts will be inaccurate. Only as nuclear diameters approach zero do counts taken from one section become accurate. Therefore, we reason that the total number nuclei actually counted in a slab of tissue will include not only those nuclei wholly contained in the slab, but also any nucleus whose center was located less than a distance equal to the average radius of these nuclei above or below the section being studied. Therefore, we may imagine a theoretical slab of tissue whose volume is more accurately compared to the counts of nuclei. The thickness of this theoretical slab is the actual thickness of the measured slab plus one radius of the average nucleus above and below, or the thickness of the measured slab plus the diameter of the average nucleus. The remaining necessary step is to determine the average measured diameter of nuclei in a section and relate that to the actual average diameter of nuclei. The measured average diameter will of course be smaller than the actual, since grazing sections through points distant from the center will be averaged with measurements of nuclei at their centers. However, the average size of a set of random chords through a sphere is ~- • the actual diameter (Abercrombie, 1946), which represents only a 21% error. Abercrombie (1946) shows that since many nuclei are in fact wholly contained within the section, we are not in fact viewing a random series of chords, and the error is further reduced to about 10%. The degree of error decreases in proportion to the smallness of nuclear diameters with respect to the section thickness. So if actual nuclear diameters average 5-8 ~t, and they are counted in sections 10-15 ~t thick, the error is smaller than in 1 ~t plastic sections. In no case, however, does it exceed 21%. In our computations, we have assumed the full 21% error. This same reasoning can be applied to any structure within a tissue as long as it is assumed to be near spherical and randomly distributed. We have applied this analysis to both neurons and synapses, making the assumption that synaptic thickenings are discs transversely sectioned. For example, if a square of tissue 100 ~t on a side and 2 g thick contains 20 neuronal nuclei with an average diameter of 10 ~t, the Abercrombie (1946) calculation predicts that these nuclei are in reality 12.7 ~ average diameter ( 4 • 10)and are contained in a theoretical slab of tissue 100 p~on a side and 14.7 ~t thick. The volume of this theoretical slab is 147,000 ~t3. The neuronal density is then 20 neurons/147,000 ~a, or 1.36 x 10S/mm 3, or 1.36 • 1011/cm a. Synaptic measurements are made in similar fashion, first by measuring the average length of synaptic densities transversely sectioned, determining the theoretical slab of tissue in which such the observed number of synapses is contained, and arriving at a figure of density per unit volume. For example, if the average number of synaptic thickenings in a section 6 ~ x 6 ~t and 0.1 ~t thick is 5, and their average length is 0.31 ~x, the Abercrombie calculation tells us that the actual average length is 4/~ x 0.31 or 0.39 ~. The slab containing the centers of these synaptic discs is 0.1 ~t + 0.39 ~ or 0.49 ~t thick, and its volume is 0.49 x 6 ~t or 17.5 ~t3. The density of synapses is then 5/17.5 p? or 2.85 x 1014/cm 3. This number would then be expressed in relation to the number of neurons rather than to volume, because during development the tissue itself is expanding at a considerable rate. In Figures 3 and 4, examples of these calculations are shown.
Results 1. Synapses Synapses, for the purpose
of this report, were defined as membrane
appositions
between two neural structures where a recognizable density existed along or near the membrane apposition area, and where vesicles were accumulated near the presynaptic membrane. The membranous synaptic density was found on both the pre- and postsynaptic membrane. It was found adherent to the pre- and postsynaptic membrane equally (symmetric synapses), or the density was
22]
356
18 16128-
L.H. Mathers, Jr. et al.
~ NON-VESICLE SYNAPSES
\
Exp. Brain Res. 33,353-369 (1978)
@ Springer-Verlag 1978
Synaptic Development in the Rabbit Superior Colliculus and Visual Cortex L.H. Mathers, Jr., K.L. Mercer, and P.E. Marshall Department of Structural Biology, Fairchild Research Center, School of Medicine, Stanford University Medical Center, Stanford, CA 94305, U. S. A.
Summary. The development of synapses in the visual cortex (VC) and superior colliculus (SC) of the rabbit has been examined with the electron microscope. In both areas, the number of synapses reaches adult levels by 2 0 - 2 5 days of postnatal age, but the development in the visual cortex is delayed in comparison to that in the superior colliculus. When S synapses (spheroidal vesicles, asymmetric thickening) are compared with F synapses (flattened vesicles, symmetric thickening), even greater differences are seen. In both the VC and SC, S synapses develop earlier than F synapses, though there is considerable overlap. Of interest is the fact that synapses in the visual cortex seem to overshoot their adult levels late in development, suggesting that an excess of synapses may be formed in this system. Multiple synapses, probably of retinal origin, increase in the first 3 weeks of synaptic development in the SC, but never are present in significant proportions in the VC. Synapse formation most often is characterized by formation of a junction and a postsynaptic thickening, followed by acquisition of synaptic vesicles. After 15 days, there is only a small number of such "non-vesicle synapses" in either the SC or VC. Key words: Synaptic development - Visual system - Visual cortex Superior colliculus - Electron microscopy. Previous physiologic (Spear et al., 1972; Grobstein et al., 1973; Mathers et al., 1974; Fox et al., 1976) and morphologic (Mathers, 1977a, b; Mathers, in prep.) studies of development in the rabbit visual system have provided an extensive description of postnatal receptive field and neuronal development. These studies showed that there is a considerable postnatal period of continued development in both the superior colliculus and visual cortex. Whereas the superior colliculus possesses all adult receptive field types by 9-10 days postnatal and is physiologically equal to the adult at 18-20 days, the visual 0 0 1 4 - 4 8 1 9 / 7 8 / 0 0 3 3 / 0 3 5 3 / $ 3.40
354
L.H. Mathers, Jr. et al.
c o r t e x d o e s n o t a t t a i n all a d u l t r e c e p t i v e t y p e s u n t i l d a y 18 a n d is n o t physiologically equal to the adult until 20-25 days. The morphologic development of neurons in these two areas nicely complements the physiologic sequence, in that the larger relay neurons (vertical cell in the superior colliculus; l a r g e p y r a m i d a l cell i n t h e v i s u a l c o r t e x ) m a t u r e e a r l i e r t h a n t h e s m a l l e r n e u r o n s ( s t e l l a t e a n d p i r i f o r m cells i n t h e s u p e r i o r c o l l i c u l u s ; s t e l l a t e a n d s o m e s m a l l p y r a m i d a l cells i n v i s u a l c o r t e x . I n e a c h o f t h e s e r e g i o n s , t h e r a t e o f d e n d r i t i c spine growth was found to be the clearest feature distinguishing the earlier developing neurons from the later (Mathers, 1977b; Mathers, in prep.). In the present study we have extended the principle of correlating morphology with physiology to the ultrastructural level, in an effort to d e t e r m i n e w h e t h e r s i g n i f i c a n t f e a t u r e s o f s y n a p t i c d e v e l o p m e n t m a y also h e l p explain the different rates of physiologic maturation in the superior colliculus and visual cortex. A portion of this work has already been reported (Mathers, 1977a).
Materials and Methods Twenty-four Dutch-belted rabbit pups were used in this study, ranging in age from 0 to 30 days. At least three animals from separate litters were studied at each age. Additional adult rabbits were also prepared. Rabbits were bred in the laboratory, and the day on which pups were first found (nests checked at 8-9 a.m.) was designated as day 0. Pups were removed from the nest and deeply anesthetized with sodium pentobarbital (this study was performed in accordance with the "Guiding Principles in the Care and Use of Animals" statement of the American Physiological Society). In many cases animals were artificially respired with room air during the initial phases of the perfusion procedure. The chest was then opened and 4% paraformaldehyde-l% glutaraldehyde in 0.tM Phosphate buffer was introduced through a 22 or 25 gauge needle into the left ventricle. No effort was made to monitor perfusion pressure, but the perfusion bottle was 70 cm above the animal, and care was taken to maintain low pressure so as to prevent hemorrhage from the nasal mucosa. The brain was quickly removed, and slices taken from both the visual cortex and the superior colliculns. Tissue in the visual cortex was removed from a point 1 mm lateral to the medial border of the area, and in the superior colliculus from the central mid-crown region. Tissues were then osmicated, block-stained in aqueous uranyl acetate, dehydrated, and embedded in Araldite 502. Sections were cut in the frontal and horizontal planes, and mounted on 150 mesh or 75 x 300 mesh grids. The grids were examined on a Siemens IA. Synapses were counted in linear or vertical sweeps across the tissue (Lurid and Lund, 1972a). A structure was counted a synapse only when a synaptic cleft and density were clearly present. Vesicles were usually but not always counted as necessary for the identification of a synapse, but at the earliest ages some structures were designated as non-vesicle synapses (NV) because their cleft and density morphologies were clearly those of a synapse and not of a desmosome or other adhesive junction. Each synapse or NV junction encountered in a sweep was recorded as to size, nature of pre- and postsynaptic structure, vesicle shape and number, and type of density. Each synapse was also charted for its position in the cortical or collicular tissue. For each animal studied, calculations of neuronal density per unit volume were calculated, using the methods of Abercrombie (1946) as further described by Cragg (1975). Synapses were counted according to their ratio to neurons (synapses/neuron). This method is more reliable than areal synaptic counts because this expression corrects for the expansion of tissue through development (Anker and Cragg, 1974). Synaptic density was calculated in this study without a correction for synapses of small size. Thus our results were probably low by 10-20% if Anker and Cragg's (1974) computations are correct. Since our purpose was principally to show change in synaptic ratios, this correction was not included, and in other respects the method was similar to that of Anker and Cragg (1974).
Synaptic Maturation in Visual System
355
The rationale behind making measurements of neuronal nuclei and synaptic thickenings is identical, and for a thorough discussion the reader is referred to Abercrombie (1946). We will present here a brief explanation. Neuronal nuclei are assumed to be near-spherical and randomly distributed in the tissue sampled. If our goal is to express nuclei/unit volume, then we cannot simply extrapolate from nuclear counts in the volume of tissue represented in our sample sections, because some nuclei will appear in more than one section, and counts will be inaccurate. Only as nuclear diameters approach zero do counts taken from one section become accurate. Therefore, we reason that the total number nuclei actually counted in a slab of tissue will include not only those nuclei wholly contained in the slab, but also any nucleus whose center was located less than a distance equal to the average radius of these nuclei above or below the section being studied. Therefore, we may imagine a theoretical slab of tissue whose volume is more accurately compared to the counts of nuclei. The thickness of this theoretical slab is the actual thickness of the measured slab plus one radius of the average nucleus above and below, or the thickness of the measured slab plus the diameter of the average nucleus. The remaining necessary step is to determine the average measured diameter of nuclei in a section and relate that to the actual average diameter of nuclei. The measured average diameter will of course be smaller than the actual, since grazing sections through points distant from the center will be averaged with measurements of nuclei at their centers. However, the average size of a set of random chords through a sphere is ~- • the actual diameter (Abercrombie, 1946), which represents only a 21% error. Abercrombie (1946) shows that since many nuclei are in fact wholly contained within the section, we are not in fact viewing a random series of chords, and the error is further reduced to about 10%. The degree of error decreases in proportion to the smallness of nuclear diameters with respect to the section thickness. So if actual nuclear diameters average 5-8 ~t, and they are counted in sections 10-15 ~t thick, the error is smaller than in 1 ~t plastic sections. In no case, however, does it exceed 21%. In our computations, we have assumed the full 21% error. This same reasoning can be applied to any structure within a tissue as long as it is assumed to be near spherical and randomly distributed. We have applied this analysis to both neurons and synapses, making the assumption that synaptic thickenings are discs transversely sectioned. For example, if a square of tissue 100 ~t on a side and 2 g thick contains 20 neuronal nuclei with an average diameter of 10 ~t, the Abercrombie (1946) calculation predicts that these nuclei are in reality 12.7 ~ average diameter ( 4 • 10)and are contained in a theoretical slab of tissue 100 p~on a side and 14.7 ~t thick. The volume of this theoretical slab is 147,000 ~t3. The neuronal density is then 20 neurons/147,000 ~a, or 1.36 x 10S/mm 3, or 1.36 • 1011/cm a. Synaptic measurements are made in similar fashion, first by measuring the average length of synaptic densities transversely sectioned, determining the theoretical slab of tissue in which such the observed number of synapses is contained, and arriving at a figure of density per unit volume. For example, if the average number of synaptic thickenings in a section 6 ~ x 6 ~t and 0.1 ~t thick is 5, and their average length is 0.31 ~x, the Abercrombie calculation tells us that the actual average length is 4/~ x 0.31 or 0.39 ~. The slab containing the centers of these synaptic discs is 0.1 ~t + 0.39 ~ or 0.49 ~t thick, and its volume is 0.49 x 6 ~t or 17.5 ~t3. The density of synapses is then 5/17.5 p? or 2.85 x 1014/cm 3. This number would then be expressed in relation to the number of neurons rather than to volume, because during development the tissue itself is expanding at a considerable rate. In Figures 3 and 4, examples of these calculations are shown.
Results 1. Synapses Synapses, for the purpose
of this report, were defined as membrane
appositions
between two neural structures where a recognizable density existed along or near the membrane apposition area, and where vesicles were accumulated near the presynaptic membrane. The membranous synaptic density was found on both the pre- and postsynaptic membrane. It was found adherent to the pre- and postsynaptic membrane equally (symmetric synapses), or the density was
22]
356
18 16128-
L.H. Mathers, Jr. et al.
~ NON-VESICLE SYNAPSES
\
