Brain Research, 595 (1992) 171-174 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00
171
BRES 25428
APV prevents the elimination of transient dendritic spines on a population of retinal ganglion cells K a m C h e u n g L a u , K w o k - F a i So a n d D a v i d Tay Department of Anatomy, Universityof Hong Kong, Pokfulam (HongKong) (Accepted 11 August 1992)
Key words: Retinal ganglion cell; Transient dendritic spine; Development; NMDA receptor; DL-2-amino-5-phosphonovalericacid
Blockade of the N-methyl-D-aspartate (NMDA) receptors on retinal ganglion cells (RGCs) during development prevents the elimination of the exuberant spine-like processes in a population of Type 1 RGCs in hamsters. During the development of RGCs, exuberant dendritic spines have been observed which disappear during maturation. Blocking the NMDA receptors on developing RGCs with the antagonist, DL-2-amino-5-phosphonovaleric acid (APV) and the subsequent retention of some of the normally transient dendritic spines suggest that the morphological development of post-synaptic neurons may be affected by this treatment. Our result further suggests that the elimination of exuberant spines during normal development requires interactions between receptors on the spines and neurotransmitters released b~ the pre-synaptic inputs.
N-methyl-D-aspartate (NMDA) receptors on neurons have been found to play an important role in neural plasticity and development. They are involved in the long-term potentiation of synaptic connections in the hippocampus 3'8'9'2° and the visual cortex 2'~. Studies on the visual system indicate that NMDA receptors on post-synaptic neurons are important for plasticity during development, and blocking the receptors by applying the antagonist, DL-2-amino-5-phosphonovaleric acid (APV) prevents the segregation of the eye-specific layers and matching of binocular maps in tectum in frogs5'2~, and the segregation of the on-off sublaminae in the dorsal lateral geniculate nucleus in ferrets 7. All these studies had concentrated on the effects of blocking NMDA receptors on the morphological development of the pre-synaptic axon arbors, but whether such treatment has any effect on the development of post-synaptic neurons has not yet been investigated. We report here that after blocking the NMDA receptor of CNS neurons during development, the morphology of post-synaptic neurons may be affected. Intracellular injection of Lucifer yellow (LY) was used to study the morphology of superior colliculus projecting Type I retinal ganglion cells (RGCs) from
two groups of young adult hamsters which had been injected intraocularly with APV (experimental) or buffer (control) during development. Intraocular injection of APV, or buffer, in neonatal hamsters was started at postnatal day 9 (P9). In the first group, two litters of time-mated hamsters were used. Intraocular injection of 1/zi of 0.1 or 1 mM APV in 0.1 M citrate buffer (pH 4.8) was performed bilaterally every 5 days. The results on the RGCs treated with the two concentrations of APV were the same, therefore, they were pooled as a group. In the second group, two litters of time-mated hamsters were injected with 1/zl of 0.1 M citrate buffer (pH 4.8) to both eyes of the animals every 5 days. The injection regime was continued until the animals were sacrificed at P30-P32. Ether was used as the anesthetic during intraocular injection. Two days before sacrifice, the animal was anesthetized with Sagatal (sodium pentobarbitai). The superior colliculus projecting RGCs were pre-labelled by applying Gel Foam (UpJohn) soaked with 3% Fast blue to the surface of both superior coUiculi, as described previously ~4. After 2 days survival, the animal was killed by an overdose of Sagatal. Both retinas were dissected and fixed in 2% paraformaldehyde in phosphate buffer (pH 7.4). Intracellular injection of LY and the recon-
Correspondence: K.C. Lau, Department of Anatomy, The University of Hong Kong, Li Shu Fan building, 5 Sassoon Road, Hong Kong.
172 struction of the LY-injected cells was performed as described before t4•e3. Effort was made to inject into those Fast blue labelled cells with large somata. After reconstruction, only those cells that were judged to be Type I RGCs 14a8 by their morphology were used in the present study. Previous study had shown that dendritic spine-like processes can already be observed on P4 and disappear by P30 in hamsters ~5, and the appearance of synapses in the inner plexiform layer in hamsters starts at P l l 6. It is believed that the exuberant spine-like processes on dendrites of developing RGCs are potential synaptic sites for making contact with inputs from bipolar or amacrine cells 24. Therefore, starting the intraocular injection of APV at P9 would block the NMDA receptors before the formation of synapses between developing RGCs and their afferent neurons. A total of 154 Type I RGCs from 10 experimental (n --- 78) and 10 control (n -- 76) animals were used in the present study• The morpholo~ of the 76 Type I RGCs which had been treated with citrate buffer during development was similar to that of normal Type I RGCs, as de.~o ibed previously t4. When dendritic spines were defined as processes equal to or shorter than 5 /~m, as they were defined in developing RGCs tS'tg'26, the mean nunrber of spines in the control group was 2.8 (+ 4.9. S.D.), which is similar to the results of the normal Type ! RGCs (2.3 + 3.2) described previously is. Fig. IA and B are histograms illustrating the distribution of the number of spines of the normal and control groups of Type I RGCs. The number of spines of the Type 1 RGCs in the normal and control groups ranged from 0 to 20 and 0 to 30, respectively. Results of the Mann-Whitney U test suggested that they were not different from each other (Z = -0.1296, P = 0.8969). The general morphology of Type I RGCs of the experimental (APV-treated) group was similar to that of the control group. Table 1 illustrates some general morphological parameters measured on these two groups of ceils. It can be seen that they were almost identical to each other. However, the mean number of spines of the experimental population of Type I RGCs was found to be 16.2 + 29.6, which was significantly higher (Mann-Whitney U test) than that of the normal (Z = -3.1693, P = 0.0015) or the control (Z = -3.893, P = 0.0001) groups. Fig. 1C is the histogram showing the distribution of the number of sr ~nes of the experimental group of cells. The number of spines observed in 82% of the cells in this group had a range similar to that of the normal or control groups. However, about 18% of the Type I cells had an extraordinarily large number of spines on their dendrites (Fig. 2). The range for the number of spines of these 18% of Type I cells was 40-140, which was comparable to those developing
A
IIONormal
lo°
L |
i~o
sl
"
-
t|o
Buffer
Z
I
51
Ill
tSl
$0
t00
il0
c" 1 ¢|
•
0.a
0
Number of
spines
Fig. I. Histograms showing the distribution of the number of spines of the Type I RGCs obtained from normal hamsters (A) and hamsters which had been treated with buffer (B) or APV (C) during development. Note that two different scales are used for the Y-axis.
TABLE I
Morphological data of the Type 1 retinal ganglion cells in hamsters following the treatment of intraocular injection of APV or citrate buffer during development Note: all the parameters measured, except the mean number of spines, of the two groups of cells are similar to each other a~d to the normal Type I cells 14 although both the buffer and the APV-treated cells are slightly smaller than the normal cells. The repeated eye injections during development might have damaged the eyeball and limited the growth of the retina, in fact, the retinas of both control and experimental animals were only 80% of the normal retinal size at the same age. Numbers in parentheses are +S.D. n, number of cells.
Retinal ganglion cells
Mean soma size (/zm 2)
Mean dendritic field (p.m)
Mean no. of primary dendrites
Mean dendritic spread
Mean no. of spines
APV treated (n-- 78) Buffer treated (n = 76)
352.1 (+48.7)
413.4 (+71.9)
4.1 (+1.2)
151.0 (+34.2)
16.2 (4.29.6)
357,4 ( 4- 52,8)
422.2 (4- 72.6)
4.0 ( 4-1.0)
152.3 (4- 33.5)
2.8 (4- 4.9)
173
A ) \
l
Fig. 2. A: reconstructed drawing of an APV-treated Type I RGC with numerous spine-like processes on dendrites. The axon is denoted by an asterisk. B: photograph showing spine-like i~rocesses (arrows) on dendrites of the same cell shown i~t A. Bars = 30 gm.
Type I RGCs possessing exuberant dendritic spines ~s. These results indicate that the blockade of the NMDA receptors during development had led to the retention, on about 18% of Type I RGCs, of a significant number of transient dendritic spines which would have been eliminated normally by P30 is. These 18% of cells were observed on the temporal half and the upper nasal retina, with the distance from the optic disc ranging
from 800 to 3100/.tm. No significant number of spines was observed on cells in the lower nasal retina, although about 25% of the injected cells were obtained from this area. This might be due to the fact that this part of the retina is relatively further away from the limbus of the upper temporal part of the eye where the intraocular injection of APV was performed. Although both the application of either tetrodotoxin (TTX) or APV during development was found to prevent the formation of mature axon arbors of RGCs in the visual center, such as the dorsal lateral geniculate n u c l e u s 7'22, it is known that TI'X cannot prevent the elimination of the transient dendritic spines on developing R G C s 13'25. Therefore, the present study showing retention of dendritic spines after APV treatment indicates that the action of APV on the developing RGCs is different from that of TTX. The bipolar and most of the amacrine cells, which provide the majority of the afferent inputs to RGCs, are neurons incapable of producing sodium channel-mediated action potentials ~, therefore, the TTX applied during development may not have any effects on the influences of these neurons to the RGCs, and hence, they are still able to form synapses with RGCs. The NMDA receptor is one o f the three main excitatory amino acid receptors observed on RGCs ~°'j6. Our results indicate that APV applied during development /orevents the disappearance of a proportion of file spines on about 18% of Type I RGCs. Thus, it is reasonable to believe that this population of cells may have more NMDA receptors on their dendrites than the other cells. It has been suggested that RGCs produce ~ransient dendritic spines ~0 seek their appropriate afferent inputs from amacrine and bipolar cells during development. Once they make connections with appropriate afferents, the spines wil~ retract 24. An initial increase, followed by a decrease in the number of spines, has been observed in neurons of the developing visual cortex 4, and the formation of exuberant spine-like processes followed by shortening or retraction of these spine-like processes has been described in neurons of the cerebellum during development ~2. These spines, or spine-like processes, observed on the cortical and cerebellar neurons are tl,c sites for forming synapses or transient synapses 4'12. "~'hus, the disappearance of spines in RGCs could also b~ interpreted as a morphological chan~e~ inclicati:~g that synapses have been formed betwee~a t~'~-• RGCs arid the interneurons. In addition, it ha,~ recentl: ~beam shown that the NMDA receptor antagonist b!ocks the ch~th~¢s o f C a 2+ in dendritic spines of hippocampal neurons which receive pre-synaptic stimulation ~7. Therefc.re, it is reasonable to speculate that transient dendritic spines on developing RGCs may bear receptors for neuro-
174 transmitters on their surfaces. Blockade of the N M D A receptors with APV has prevented them from interacting with the neurotransmitter (glutamate). Therefore synapses between some RGCs and interneurons cannot be formed or become mature and hence the spines persist. This result suggests that the elimination of spines during development may require interactions between the receptors on the spines with neurotransmitters released by pre-synaptic elements. The other 82% of Type I RGCs might mainly have receptors of the other types which have not been blocked by the application of APV and therefore could receive appropriate neurotransmitters from their afferents; hence most of their spines could become retracted. These cells might also have spines bearing NMDA receptors, but tl',e amount may be too little to be distingushed from the ba~e!ine value of the normal and control cells. Another possib'~!ity is that the processes leading to spine retractior, are initiated at different times for different neurons or retinal regions, such that the treatment was too late to affect those neurons which have started the spine retraction earlier. Further experiments on the blockade of all or most of the receptors on RGCs at earlier developmental times may be helpful to resolve this issue. This study was supported by research grants from The University of Hong Kong. We thank Drs. G. Campbell, L.S. Jen, S.B. Udin, and S.M.S. Wu for their helpful comments on the manuscript. I Archer, S.M., Dubin, M.W. and Strark, L.A., Abnormal development of kitten retinogeniculate connectivity in the absence of action potentials, Science, 217 (1982)743-745. 2 Artbla, A. and Singer, W., Long-term potentiation and NMDA receptors in rat visual cortex, Nature, 330 (1987) 649-652. 3 Bliss, T.V.P. and I..¢mo, T., Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path, J. Physiol., 232 (1973) 331-356. 4 Boothe, R.G., Greenough, W.T., Lund, J.S. and Wrege, K., A quantitative investigation of spine and dendrite development of neurons in visual cortex (area 17) of Macaca aemestrina monkeys, J. Comp. Neurol., 186 (1979)473-490. 5 Cline, H.T., Debski, E.A. and Constantine-Paton, M., N-mcthylD-aspartate receptor antagonist desegregates eye-specific stripes, Proc. Natl. Acad. Sci. USA, 84 (1987) 4342-4345. 6 Greiner, J.V. and Weidman, T.A., Development of the hamster retina: a morphologic study, Am. J. Vet. Res., 39 (1978) 665-670. 7 Hahm, J.-O., Langdon, R.B. and Sur, M., Disruption of retinogeniculate afferent segregation by antagonists to NMDA receptors, Nature, 35 ! (1991) 568-570.
8 Harris, E.W., Ganong, A.H. and Cotman, C.W., Long-term potentiation in the hippocampus involves activation of N-methyi-Daspartate receptors, Brain Res., 323 (1984) 132-137. 9 Herron, C.E., Lester, R.A.J., Coan, E.J. and Collingridge, G.L., Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic mechanism, Nature, 322 (1986) 265 -268. 10 Ikeda, H., Robbins, J. and Kay, C.D., Excitatory amino acid receptors on sustained retinal ganglion cells in the kitten during the critical period of development, Dev. Brain Res., 51 (1990) 85-91. 11 Kimura, F., Nishigori, A., Shirokawa, T. and Tsumoto, T., Longterm potentiation and N-methyI-D-aspartate receptors in the visual cortex of young rats, J. Physiol., 414 (1989) 125-144. 12 Larramendi, L.M.H., Analysts of synaptogenesis in the cerebellum of the mouse. In R. Lli~s (Ed.), Neurobiology of Cerebellar Evolution and Development, American Medical Association, Chicago, 1969, pp. 804-843. 13 Lau, K.C., The Morphological Plasticity of Retinal Ganglion Cells During Development and Regeneration: A Lucifer Yellow Injection Study, Ph.D. Dissertation, University of Hong Kong, 1991. 14 Lau, K.C., So, K.-F. and Tay, D., Effects of visual or light deprivation on the morphology, and the elimination of the transient features during development, of Type I retinal ganglion cells in hamsters, J. Comp. Neurol., 300 (1990) 583-592. 15 Lau, K.C., Sc~, K.-F. and Tay, D., Postnatal development of Type 1 retinal ganglion cells in hamsters: a Lucifer yellow study, J. Comp. Neurol., 315 (1992) 375-381. 16 Miller, R.F. and Slaughter, M.M., Excitatory amino acid receptors of the retina: diversity of subtypes and conductance mechanisms, Trends Neurosci., 9 (1986) 211-218. 17 Miiller, W. and Connor, J.A., Dendritic spines as individual neuronal compartment for synaptic Ca 2+ responses, Nature, 354 (1991) 73-76. 18 Perry, V.H., The ganglion cell layer of the retina of the rat: a Golgi study, Proc. R. Soc. Lond. (Biol), 204 (1979) 363-375. 19 Ramoa, A.S., Campbell, G. and Shatz, C.J., Dendritic Growth and remodeling of cat retinal ganglion cells during fetal and postnatal development, J. Neurosci., 8 (1988) 4239-4261. 20 Sastiv, B.R., Goh, J.W. and Auyeung, A., Associative induction of posttetanic and long-term potentiation in CA1 neurons of rat hippocampus, Science, 232 (1986) 988-990. 21 Scherer, W.J. and Udin, S.B., N-methyI-D-aspartate antagonists prevent interaction of binocular maps in Xenopus tectum, J. Neurosci., 9 (1989) 3837-3843. 22 Sretavan, D.W., Shatz, C.J. and Stryker, M.P., Modification of retinal ganglion cell axon morphology by prenatal infusion of tetrodotoxin, Nature, 336 (1988) 468-471. 23 Tauchi, M. and Masland, R.H., The shape and arrangement of the cholinergic neurons in the rabbit retina, Proc. R. Soc. Lond. (Biol), 223 (1984) 101-119. 24 Wiissle, H., Dendritic maturation of retinal ganglion cells, Trends Neurosci., 11 (1988) 87-89. 25 Wong, R.O.L., Herrmann, K. and Shatz, C.J., Remodeling of retinal ganglion cell dendrites in the absence of action potential activity, J. Neurobioi., 22 (1991) 685-697. 2~ Wong, R.O.L., Differential growth and remodelling of ganglion cell dendrites in the postnatal rabbit retina, J. Comp. Neurol., 294 (1990) 109-132.
171
BRES 25428
APV prevents the elimination of transient dendritic spines on a population of retinal ganglion cells K a m C h e u n g L a u , K w o k - F a i So a n d D a v i d Tay Department of Anatomy, Universityof Hong Kong, Pokfulam (HongKong) (Accepted 11 August 1992)
Key words: Retinal ganglion cell; Transient dendritic spine; Development; NMDA receptor; DL-2-amino-5-phosphonovalericacid
Blockade of the N-methyl-D-aspartate (NMDA) receptors on retinal ganglion cells (RGCs) during development prevents the elimination of the exuberant spine-like processes in a population of Type 1 RGCs in hamsters. During the development of RGCs, exuberant dendritic spines have been observed which disappear during maturation. Blocking the NMDA receptors on developing RGCs with the antagonist, DL-2-amino-5-phosphonovaleric acid (APV) and the subsequent retention of some of the normally transient dendritic spines suggest that the morphological development of post-synaptic neurons may be affected by this treatment. Our result further suggests that the elimination of exuberant spines during normal development requires interactions between receptors on the spines and neurotransmitters released b~ the pre-synaptic inputs.
N-methyl-D-aspartate (NMDA) receptors on neurons have been found to play an important role in neural plasticity and development. They are involved in the long-term potentiation of synaptic connections in the hippocampus 3'8'9'2° and the visual cortex 2'~. Studies on the visual system indicate that NMDA receptors on post-synaptic neurons are important for plasticity during development, and blocking the receptors by applying the antagonist, DL-2-amino-5-phosphonovaleric acid (APV) prevents the segregation of the eye-specific layers and matching of binocular maps in tectum in frogs5'2~, and the segregation of the on-off sublaminae in the dorsal lateral geniculate nucleus in ferrets 7. All these studies had concentrated on the effects of blocking NMDA receptors on the morphological development of the pre-synaptic axon arbors, but whether such treatment has any effect on the development of post-synaptic neurons has not yet been investigated. We report here that after blocking the NMDA receptor of CNS neurons during development, the morphology of post-synaptic neurons may be affected. Intracellular injection of Lucifer yellow (LY) was used to study the morphology of superior colliculus projecting Type I retinal ganglion cells (RGCs) from
two groups of young adult hamsters which had been injected intraocularly with APV (experimental) or buffer (control) during development. Intraocular injection of APV, or buffer, in neonatal hamsters was started at postnatal day 9 (P9). In the first group, two litters of time-mated hamsters were used. Intraocular injection of 1/zi of 0.1 or 1 mM APV in 0.1 M citrate buffer (pH 4.8) was performed bilaterally every 5 days. The results on the RGCs treated with the two concentrations of APV were the same, therefore, they were pooled as a group. In the second group, two litters of time-mated hamsters were injected with 1/zl of 0.1 M citrate buffer (pH 4.8) to both eyes of the animals every 5 days. The injection regime was continued until the animals were sacrificed at P30-P32. Ether was used as the anesthetic during intraocular injection. Two days before sacrifice, the animal was anesthetized with Sagatal (sodium pentobarbitai). The superior colliculus projecting RGCs were pre-labelled by applying Gel Foam (UpJohn) soaked with 3% Fast blue to the surface of both superior coUiculi, as described previously ~4. After 2 days survival, the animal was killed by an overdose of Sagatal. Both retinas were dissected and fixed in 2% paraformaldehyde in phosphate buffer (pH 7.4). Intracellular injection of LY and the recon-
Correspondence: K.C. Lau, Department of Anatomy, The University of Hong Kong, Li Shu Fan building, 5 Sassoon Road, Hong Kong.
172 struction of the LY-injected cells was performed as described before t4•e3. Effort was made to inject into those Fast blue labelled cells with large somata. After reconstruction, only those cells that were judged to be Type I RGCs 14a8 by their morphology were used in the present study. Previous study had shown that dendritic spine-like processes can already be observed on P4 and disappear by P30 in hamsters ~5, and the appearance of synapses in the inner plexiform layer in hamsters starts at P l l 6. It is believed that the exuberant spine-like processes on dendrites of developing RGCs are potential synaptic sites for making contact with inputs from bipolar or amacrine cells 24. Therefore, starting the intraocular injection of APV at P9 would block the NMDA receptors before the formation of synapses between developing RGCs and their afferent neurons. A total of 154 Type I RGCs from 10 experimental (n --- 78) and 10 control (n -- 76) animals were used in the present study• The morpholo~ of the 76 Type I RGCs which had been treated with citrate buffer during development was similar to that of normal Type I RGCs, as de.~o ibed previously t4. When dendritic spines were defined as processes equal to or shorter than 5 /~m, as they were defined in developing RGCs tS'tg'26, the mean nunrber of spines in the control group was 2.8 (+ 4.9. S.D.), which is similar to the results of the normal Type ! RGCs (2.3 + 3.2) described previously is. Fig. IA and B are histograms illustrating the distribution of the number of spines of the normal and control groups of Type I RGCs. The number of spines of the Type 1 RGCs in the normal and control groups ranged from 0 to 20 and 0 to 30, respectively. Results of the Mann-Whitney U test suggested that they were not different from each other (Z = -0.1296, P = 0.8969). The general morphology of Type I RGCs of the experimental (APV-treated) group was similar to that of the control group. Table 1 illustrates some general morphological parameters measured on these two groups of ceils. It can be seen that they were almost identical to each other. However, the mean number of spines of the experimental population of Type I RGCs was found to be 16.2 + 29.6, which was significantly higher (Mann-Whitney U test) than that of the normal (Z = -3.1693, P = 0.0015) or the control (Z = -3.893, P = 0.0001) groups. Fig. 1C is the histogram showing the distribution of the number of sr ~nes of the experimental group of cells. The number of spines observed in 82% of the cells in this group had a range similar to that of the normal or control groups. However, about 18% of the Type I cells had an extraordinarily large number of spines on their dendrites (Fig. 2). The range for the number of spines of these 18% of Type I cells was 40-140, which was comparable to those developing
A
IIONormal
lo°
L |
i~o
sl
"
-
t|o
Buffer
Z
I
51
Ill
tSl
$0
t00
il0
c" 1 ¢|
•
0.a
0
Number of
spines
Fig. I. Histograms showing the distribution of the number of spines of the Type I RGCs obtained from normal hamsters (A) and hamsters which had been treated with buffer (B) or APV (C) during development. Note that two different scales are used for the Y-axis.
TABLE I
Morphological data of the Type 1 retinal ganglion cells in hamsters following the treatment of intraocular injection of APV or citrate buffer during development Note: all the parameters measured, except the mean number of spines, of the two groups of cells are similar to each other a~d to the normal Type I cells 14 although both the buffer and the APV-treated cells are slightly smaller than the normal cells. The repeated eye injections during development might have damaged the eyeball and limited the growth of the retina, in fact, the retinas of both control and experimental animals were only 80% of the normal retinal size at the same age. Numbers in parentheses are +S.D. n, number of cells.
Retinal ganglion cells
Mean soma size (/zm 2)
Mean dendritic field (p.m)
Mean no. of primary dendrites
Mean dendritic spread
Mean no. of spines
APV treated (n-- 78) Buffer treated (n = 76)
352.1 (+48.7)
413.4 (+71.9)
4.1 (+1.2)
151.0 (+34.2)
16.2 (4.29.6)
357,4 ( 4- 52,8)
422.2 (4- 72.6)
4.0 ( 4-1.0)
152.3 (4- 33.5)
2.8 (4- 4.9)
173
A ) \
l
Fig. 2. A: reconstructed drawing of an APV-treated Type I RGC with numerous spine-like processes on dendrites. The axon is denoted by an asterisk. B: photograph showing spine-like i~rocesses (arrows) on dendrites of the same cell shown i~t A. Bars = 30 gm.
Type I RGCs possessing exuberant dendritic spines ~s. These results indicate that the blockade of the NMDA receptors during development had led to the retention, on about 18% of Type I RGCs, of a significant number of transient dendritic spines which would have been eliminated normally by P30 is. These 18% of cells were observed on the temporal half and the upper nasal retina, with the distance from the optic disc ranging
from 800 to 3100/.tm. No significant number of spines was observed on cells in the lower nasal retina, although about 25% of the injected cells were obtained from this area. This might be due to the fact that this part of the retina is relatively further away from the limbus of the upper temporal part of the eye where the intraocular injection of APV was performed. Although both the application of either tetrodotoxin (TTX) or APV during development was found to prevent the formation of mature axon arbors of RGCs in the visual center, such as the dorsal lateral geniculate n u c l e u s 7'22, it is known that TI'X cannot prevent the elimination of the transient dendritic spines on developing R G C s 13'25. Therefore, the present study showing retention of dendritic spines after APV treatment indicates that the action of APV on the developing RGCs is different from that of TTX. The bipolar and most of the amacrine cells, which provide the majority of the afferent inputs to RGCs, are neurons incapable of producing sodium channel-mediated action potentials ~, therefore, the TTX applied during development may not have any effects on the influences of these neurons to the RGCs, and hence, they are still able to form synapses with RGCs. The NMDA receptor is one o f the three main excitatory amino acid receptors observed on RGCs ~°'j6. Our results indicate that APV applied during development /orevents the disappearance of a proportion of file spines on about 18% of Type I RGCs. Thus, it is reasonable to believe that this population of cells may have more NMDA receptors on their dendrites than the other cells. It has been suggested that RGCs produce ~ransient dendritic spines ~0 seek their appropriate afferent inputs from amacrine and bipolar cells during development. Once they make connections with appropriate afferents, the spines wil~ retract 24. An initial increase, followed by a decrease in the number of spines, has been observed in neurons of the developing visual cortex 4, and the formation of exuberant spine-like processes followed by shortening or retraction of these spine-like processes has been described in neurons of the cerebellum during development ~2. These spines, or spine-like processes, observed on the cortical and cerebellar neurons are tl,c sites for forming synapses or transient synapses 4'12. "~'hus, the disappearance of spines in RGCs could also b~ interpreted as a morphological chan~e~ inclicati:~g that synapses have been formed betwee~a t~'~-• RGCs arid the interneurons. In addition, it ha,~ recentl: ~beam shown that the NMDA receptor antagonist b!ocks the ch~th~¢s o f C a 2+ in dendritic spines of hippocampal neurons which receive pre-synaptic stimulation ~7. Therefc.re, it is reasonable to speculate that transient dendritic spines on developing RGCs may bear receptors for neuro-
174 transmitters on their surfaces. Blockade of the N M D A receptors with APV has prevented them from interacting with the neurotransmitter (glutamate). Therefore synapses between some RGCs and interneurons cannot be formed or become mature and hence the spines persist. This result suggests that the elimination of spines during development may require interactions between the receptors on the spines with neurotransmitters released by pre-synaptic elements. The other 82% of Type I RGCs might mainly have receptors of the other types which have not been blocked by the application of APV and therefore could receive appropriate neurotransmitters from their afferents; hence most of their spines could become retracted. These cells might also have spines bearing NMDA receptors, but tl',e amount may be too little to be distingushed from the ba~e!ine value of the normal and control cells. Another possib'~!ity is that the processes leading to spine retractior, are initiated at different times for different neurons or retinal regions, such that the treatment was too late to affect those neurons which have started the spine retraction earlier. Further experiments on the blockade of all or most of the receptors on RGCs at earlier developmental times may be helpful to resolve this issue. This study was supported by research grants from The University of Hong Kong. We thank Drs. G. Campbell, L.S. Jen, S.B. Udin, and S.M.S. Wu for their helpful comments on the manuscript. I Archer, S.M., Dubin, M.W. and Strark, L.A., Abnormal development of kitten retinogeniculate connectivity in the absence of action potentials, Science, 217 (1982)743-745. 2 Artbla, A. and Singer, W., Long-term potentiation and NMDA receptors in rat visual cortex, Nature, 330 (1987) 649-652. 3 Bliss, T.V.P. and I..¢mo, T., Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path, J. Physiol., 232 (1973) 331-356. 4 Boothe, R.G., Greenough, W.T., Lund, J.S. and Wrege, K., A quantitative investigation of spine and dendrite development of neurons in visual cortex (area 17) of Macaca aemestrina monkeys, J. Comp. Neurol., 186 (1979)473-490. 5 Cline, H.T., Debski, E.A. and Constantine-Paton, M., N-mcthylD-aspartate receptor antagonist desegregates eye-specific stripes, Proc. Natl. Acad. Sci. USA, 84 (1987) 4342-4345. 6 Greiner, J.V. and Weidman, T.A., Development of the hamster retina: a morphologic study, Am. J. Vet. Res., 39 (1978) 665-670. 7 Hahm, J.-O., Langdon, R.B. and Sur, M., Disruption of retinogeniculate afferent segregation by antagonists to NMDA receptors, Nature, 35 ! (1991) 568-570.
8 Harris, E.W., Ganong, A.H. and Cotman, C.W., Long-term potentiation in the hippocampus involves activation of N-methyi-Daspartate receptors, Brain Res., 323 (1984) 132-137. 9 Herron, C.E., Lester, R.A.J., Coan, E.J. and Collingridge, G.L., Frequency-dependent involvement of NMDA receptors in the hippocampus: a novel synaptic mechanism, Nature, 322 (1986) 265 -268. 10 Ikeda, H., Robbins, J. and Kay, C.D., Excitatory amino acid receptors on sustained retinal ganglion cells in the kitten during the critical period of development, Dev. Brain Res., 51 (1990) 85-91. 11 Kimura, F., Nishigori, A., Shirokawa, T. and Tsumoto, T., Longterm potentiation and N-methyI-D-aspartate receptors in the visual cortex of young rats, J. Physiol., 414 (1989) 125-144. 12 Larramendi, L.M.H., Analysts of synaptogenesis in the cerebellum of the mouse. In R. Lli~s (Ed.), Neurobiology of Cerebellar Evolution and Development, American Medical Association, Chicago, 1969, pp. 804-843. 13 Lau, K.C., The Morphological Plasticity of Retinal Ganglion Cells During Development and Regeneration: A Lucifer Yellow Injection Study, Ph.D. Dissertation, University of Hong Kong, 1991. 14 Lau, K.C., So, K.-F. and Tay, D., Effects of visual or light deprivation on the morphology, and the elimination of the transient features during development, of Type I retinal ganglion cells in hamsters, J. Comp. Neurol., 300 (1990) 583-592. 15 Lau, K.C., Sc~, K.-F. and Tay, D., Postnatal development of Type 1 retinal ganglion cells in hamsters: a Lucifer yellow study, J. Comp. Neurol., 315 (1992) 375-381. 16 Miller, R.F. and Slaughter, M.M., Excitatory amino acid receptors of the retina: diversity of subtypes and conductance mechanisms, Trends Neurosci., 9 (1986) 211-218. 17 Miiller, W. and Connor, J.A., Dendritic spines as individual neuronal compartment for synaptic Ca 2+ responses, Nature, 354 (1991) 73-76. 18 Perry, V.H., The ganglion cell layer of the retina of the rat: a Golgi study, Proc. R. Soc. Lond. (Biol), 204 (1979) 363-375. 19 Ramoa, A.S., Campbell, G. and Shatz, C.J., Dendritic Growth and remodeling of cat retinal ganglion cells during fetal and postnatal development, J. Neurosci., 8 (1988) 4239-4261. 20 Sastiv, B.R., Goh, J.W. and Auyeung, A., Associative induction of posttetanic and long-term potentiation in CA1 neurons of rat hippocampus, Science, 232 (1986) 988-990. 21 Scherer, W.J. and Udin, S.B., N-methyI-D-aspartate antagonists prevent interaction of binocular maps in Xenopus tectum, J. Neurosci., 9 (1989) 3837-3843. 22 Sretavan, D.W., Shatz, C.J. and Stryker, M.P., Modification of retinal ganglion cell axon morphology by prenatal infusion of tetrodotoxin, Nature, 336 (1988) 468-471. 23 Tauchi, M. and Masland, R.H., The shape and arrangement of the cholinergic neurons in the rabbit retina, Proc. R. Soc. Lond. (Biol), 223 (1984) 101-119. 24 Wiissle, H., Dendritic maturation of retinal ganglion cells, Trends Neurosci., 11 (1988) 87-89. 25 Wong, R.O.L., Herrmann, K. and Shatz, C.J., Remodeling of retinal ganglion cell dendrites in the absence of action potential activity, J. Neurobioi., 22 (1991) 685-697. 2~ Wong, R.O.L., Differential growth and remodelling of ganglion cell dendrites in the postnatal rabbit retina, J. Comp. Neurol., 294 (1990) 109-132.
