- Email: [email protected]
Synaptic density correlated with maze performance in young and aged rats. A preliminary study
Mechanisms of Ageing and Development, 21 (1983)245-255 Elsevier Scientific Publishers Ireland Ltd.
245
SYNAPTIC DENSITY CORRELATED WITH MAZE PERFORMANCE IN YOUNG AND AGED RATS. A PRELIMINARY STUDY
ALBERT WILLIAM KLEIN Department of Anatomy, Eastern Virginia Medical School, 700 Olney Road, Norfolk, Virginia 23.501 (U.S.A.) (Received August llth, 1980) (Revision receivedAugust 26th, 1982)
SUMMARY The aim of this study was to determine the effect of age both on maze performance and upon morphological changes in the brain. Young and senescent rats were differentiated into three groups by use of a complex 14-unit maze: young maze-bright rats, old maze-bright animals and old maze-dull rats. Layer III of the frontal and the occipital cortices of these animals was studied, employing coded electron micrographs, to determine the average number of synaptic profiles per area of thin section. It was observed that the young maze.bright rats had the highest synapse density, the old maze-bright rats had 85% of that amount while the old maze-dull animals had a density that was 54% that of the young animals.
Key words: Aged;Maze performance; Synaptic density
INTRODUCTION Several experimental studies have presented anatomical evidence for the decline of neuronal function with age. These studies have inferred that aging changes, such as loss of neurons, accumulation of lipofuscin pigment and loss of dendritic spines from cortical neurons, have functional correlations in aging brains [1-6]. The hypothesis is that these alterations hinder the total integrative potential of the aging brain and hence may contribute to clinical entities such as senile dementia. Apparently the number of synapses has great meaning for effective brain function. The careful work of Feldman and Dowd [4,5] has shown great variation as well as alterations m dendritic spine density that accompany aging. A specific alteration reported by Gemisman et al. [6] has been a decline in the number of axodenddtic synapses. Perhaps, because intercellular communication is so important in brain function, dendritic spines and synaptic frequencies play a major role in cerebral integration. The recent evidence of Buell and Coleman [7], who looked at dendritic arborization in man, supports the behavioral concepts and reports of wide differentials between cortical function in the aged human. 0047-6374/83/$03.00 Printed and Published in Ireland
© 1983 Elsevier ScientificPublishers Ireland Ltd.
246
The melding of behavioral and morphometric experimentation is an approach that has not been utilized to its fullest potential. The combination of these two disciplines is aimed at discovering those features of the aging cerebral cortex that are most likely to be involved with functional or behavioral brain deficits. Moreover, this method encourages speculation about possible, yet so far unproven, causal relationships I~etween neuronal morphology and senile dementia. This study was undertaken to examine the possible relationship between the effect of age on maze performance (learning ability), and on interneuronal communication, [e. synaptic frequency in aged rats. METHODS
Young and senescent female Wistar rats from the Gerontology Research Center in Baltimore, Maryland, were used in this investigation. Young animals were 8 months' old and senescent rats 25 months' old at the time of sacrifice. The mean life span of this strain of rats is 23 months. As described in preliminary work by Klein and Michel [8], performance on a 14-choice "T"-maze [9] behaviorally separated the rats into (1) a young maze-bright group, (2) an old maze-bright group, and (3) an old maze-dull group. These results agreed with the previous observations of Goodrick [ 10]. Ten young rats and 20 old rats began the training program. All 10 of the younger
MEDIAL
DORSAL
Fig. 1, Right telencephallc cortex of the rat.
247 animals were alive at the conclusion of the maze training; however, due to the advanced age of the older group, only 11 were alive at the end of the maze training. This was expected based on the experiences of Goodrick [10]. Using a book of random numbers, animals were selected as samples for electron microscopy. Neocortical samples from the three groups of rats were prepared for both light- and electron-microscopic study first by perfusion with Karnovsky's fluid through the left ventricle as previously reported [8]. The external flesh was removed and the skulls were placed overnight in fixative. The following day the brains were removed and placed in fresh fixative. After a 4-tl period in the fixative, wedge-shaped samples of the frontal and occipital cortex were removed for study of the neocortex by light microscopy and transmission electron microscopy. The removal of the wedges was made consistent by the method previously reported by Klein and Michel [8]. Because the rat cortex is lissencephalic, the total length was marked off into eight segments anterior to posterior. The wedges removed were from segment number 2 (being frontal) and segment number 7 (occipital) and were the second millimeter lateral from the exact midline or sagittal surface (see Fig. I). The latter samples were post-fixed in osmium and processed for sectioning. Thick sections of 1 /an were cut in a plane perpendicular to the cortex and examined to determine the location of layer III. Layer III was further isolated by trimming. Thin sections were cut with a diamond knife and examined on a Philips 301 electron microscope. The brains of four animals from each group (young maze-bright, old maze-bright and old maze-dull) were processed. Two blocks of tissue were taken from the frontal cortex and two blocks of tissue were taken from the occipital cortex (hence four blocks per rat). The pyramidal cell layer (layer Ill) was isolated and sections from each block were mounted on grids (resulting in eight grids per animal). Four photographs were taken in a vertical line across the tissue surface of each grid (~ :ithout regard to tissue content). See the flow diagram (Scheme 1) for a summary of the samples. The synapse profiles were counted per unit area of section. Synapse counts were averaged per unit area, frontal and occipital regional counts were pooled for each of the three functionally differentiated groups. Because there were only three separate groups statistical comparisons were made using Student's t-test. RESULTS As previously reported by this laboratory [8] and Goodrick [10], the maze training effectively differentiated three groups of rats: (1) young animals, all of which learned
Each experimental animal
x 2 regions (frontal and occipital)/animal X 2 blocks/region X 2 grids/block x 4 photos/grid X 4 animals/group X 3 experimental groups Scheme 1. Flow diagram of samples.
Total samples
therefore therefore therefore
= = = = = =
2 regions/animal 4 blocks/animal 8 grids/animal 32 micrographs/animal 128 photomicrographs/group 384 photomicrographs used
248 the maze well (only one error on day 20, only one trial per day); (2) old animals which performed as well as the young rats (one error or less on day 20); and (3) old animals which committed approximately eight errors repetitively even on day 20 o f the trials. Of those old animals that did not survive the entire maze shaping and maze training paradigm some were from the bright group and some from the dull group.
Fig. 2. Layer III neuron from an 8-month-old rat specimen. Observe the presence of only slight amounts of lipofusein (Lf). Also seen are examples of smooth endoplasmic retieulum (SR), lysosomes (Ly), an axon (Ax) and a synapse (Sy). The reference bar is equivalent to 1 ~m (print magnification X8910).
249 All of the electron rnicrographs were examined while they were still coded so that examiners had no knowledge of the experimental groups. Upon decoding the data it was obvious that between the three groups there were many similarities and yet some striking differences. Lipofuscin pigment was variably present in all brain specimens. All neurons observed did not contain pigment, but all specimens, even the 8-monthold animals, did contain some lipofuscin pigment in some neurons (Fig. 2). Some neurons
Fig. 3. This view contains larger granules of the aging pigment lipofuscin (Lf) associated with lysosomal bodies (Ly). Note the membrane which seems to surromld the two substances. Ribosomes (Rb) are in abundance.
250 in sample sections of the older animals appeared to be free of pigment. Subjectively, it appeared that neurons of older animals contained more lipofuscin granules or that the granules were larger than in younger animals. Exact quantification of lipofuscin pigment was not attempted in this study. Qualitatively the deposits appeared to be the same. The lipofuscin was often adjacent to lysosomal bodies, and at higher magnifications it was actually seen to be contained within the lysosomal membranes (Fig. 3).
Fig. 4. Neuron and perineuronal glial cell degeneration was observed only in the older animals. Ineluded here are myelin whorls (MW) and granulovascnlar degeneration (Gv).
251 TABLE I AVERAGE NUMBER OF SYNAPSES PER SQUARE MILLIMETER OF EXAMINEDCORTICAL TISSUE
Mean number ± S.D. Percentage (%)
Young adult maz e-brigh t
Old maz e-brigh t
Old ma z e 4 u i l
395.7 -+31 100
338.8* ±58 85.6
215.6"** -*53 54.5
*Different from the bright group of old animals at the p < 0.01 level. **When pooled, the old groups together differed from the young animals at the p < 0.05 level. Neurons from young specimens consistently appeared as healthy, intact cells while those in older brain samples consistently exhibited neurons in various stages of disarray. These older, degenerating neurons contained what appeared to be fatty granules, signs of granulovaseular degeneration and myelin whorls (Fig. 4). This type of extensive cell change was observed only in the neurons of older rats. Synapses were counted from the systematically sampled micrographs and their numbers were expressed per square mm of thin section as in Table I. Synapses with discontinuous pre- and postsynaptic membrane thickenings (double membrane interrupted as in Fig. 5) were counted as one synapse. Figure 5 is included to illustrate the type of sample photograph and the magnification from which synapse counts were made. To show a micrograph from a young animal with many synapses and to show a micrograph from an old, dull animal would be a subjective presentation. All micrographs were used and therefore no subjective selecting of rnicrographs occurred. Synaptic configurations which exhibited vesicles as well as both presynaptie and postsynaptic membrane thickenings were the only structures counted. The average number of synapse prof'des seen in 1 square mm of thin section was 395.7 for the young adult, maze.bright group, 338.8 for the old maze-bright animals, and 215.6 for the old maze-dull group. When synapses for the young adults were considered as I00%, the old-brights and old-dulls were only 85% and 54%, respectively, of the young bright group. When pooled, the two older groups of animals were significantly different from the young animals at the p < 0.05 level. The old maze-dull group was different from the old maze.bright group at the p < 0.01 level. Because section thickness, synaptic dimensions and random synaptic orientation were not taken into account, the numbers in Table I were intended to reflect relative synaptic numbers, present in any given volume of layer III of neocortex. DISCUSSION Verzar-McDougall [11], Goodrick [10] and Klein and Michel [8] have demonstrated that a complex 14.unit maze is a reliable discriminator of learning performance in aged rats. These investigators have been able to differentiate old maze.bright rats from old maze-dull rats. T h e criteria for differentiation is that the old maze.bright animals per-
252
Fig. 5. This illustration depicts the type of photomicrograph and the magnification from which the synaptic counts were tallied. Several, but not all, synaptic sites have been marked (Sy). Synapses with discon~tuous double membranes, as in the insert, were counted as a single synapse. form as well as the young animals. Human studies which have ruled out m o t o r coordination as a prime factor and which have required a person to learn and trace a complex maze with a stylus have resulted in similar findings. Younger subjects performed with fewer errors than did the average of the elderly subjects. However, some elderly subjects
253 did perform as well as the best younger subjects, while certain of the elderly never mastered the maze [12,13]. These separate studies suggegt a possible common central nervous system mechanism for decline of performance with aging. Also, these studies support the inequity of performance decline with age; that is, not all individuals become infirm with advanced age. The study of structural, functional and biochemical changes with aging has elucidated a number of possible explanations for decline of cerebral performance with senescence. Classically, Brody [1] has documented cell loss with age in the cerebral cortex, and more recently in the locus coeruleus (brain stem) [2], a system with widespread projections. Wisnicwski and Terry [14] have shown an increase in neurofibrillary tangles with increasing age and senility. Feldman and Dowd [5] have thoroughly outlined the loss of dendritic spines with aging. The age-related loss of dendritic spines and synapses has also been studied by Bondareff and co-workers and has been shown to effect losses of axodendritic synapses sometimes more selectively than axosomatic synapses [15]. Synapse loss has also been reported in brain structures implicated in recent memory [6]. Scheibel even suggests that impaired neuronal connections may result in inappropriate adaptive responses [16]. Plasticity studies have pointed out functional deficits with advancing age [17], and that these deficits may be the result of plastogenic action of hormones [18]. Lipofuscin pigment accumulation within neurons may influence the movement of RNA and lead to poor intracellular economy and thereby contribute to a decline in neuronal function [3]. Hammer et al. [19] and others [20] implicate lipofuscin pigment as a disruptive presence in neurons, specifically the motor neurons in amyotrophic lateral sclerosis. Neuropharmacological changes with age have been increasingly documented. Primarily these changes were seen as specific decrements in neurotransmitters or receptors [21] with age. Certain transmitters have also been implicated in disease states common to be aged, such as presenile and senile dementia as well as Parkinsonism. From an alternative clinical approach, physostigmine and other manip ulations of the cholinergic system can affect recent memory in a positive manner [22]. This investigation is one of the few to correlate a function, such as maze performance or animal learning, to a structural entity like the synapse. However, it should not be taken as prima facie evidence that synaptic loss caused declines in performance. The two groups of old rats which did differentiate themselves, based on maze performanceold maze-bright and old maze-dull-did subsequently reveal that their synaptic densities were reduced from those of the young and from each other, both by age and by poor maze performance. However, correlation does not necessarily prove a cause-and-effect rehtionship. The synapses counted were from layer III of the frontal and occipital cortices. Both of these regions had previously failed to show changes in neuron numbers at the light.microscopic level [8,23 ]. The suggestion which needs verification by other laboratories is that old maze-dull animals have a significant decrement in intercellular communication sites and while a cause-and-effect relationship should not be implied, it is possible that synaptic alterations and plasticity play a major role in the determination of performance.
254 ACKNOWLEDGEMENTS The author is grateful for the suggestions of Drs. Donald Stein, R. Frederick Becker, Herbert Schapiro, Robert Brownson and Paul Coleman, and for the technical assistance of Ms. Sheila Rosenfield. Acknowledgement is made to the Gerontology Research Center, NIA and W. French for aged animals provided in the Guest Scientist Program. This work was funded in part by the Biomedical Research Support Grant RR09028 and in part by Eastern Virginia Medical School Institutional Research Support Funds. REFERENCES 1 H. Brody, Organization of the cerebral cortex. J. Comp. Neurol., 102 (1955) 511-556. 2 N. Vijayashbankar and H. Brody, Quantitative study of the pigmented neurons of the nuclear locus. Coeruleus and subcoeruleus in man as related to aging. Z Neuropathol. Exp. NeuroL, 38 (1979) 490-497. 3 D.M.A. Mann and P.O. Yates, Lipofuscin pigments their relationship to aging in the human nervous system: The lipofusein content in nerve cells. Brain, 97 (1974) 481-488. 4 M.L. Feldman and C. Dowd, Aging in rat visual cortex: light microscopic observations on layer V pyramidal apical dendrites.Anat. Rec., 178 (1974) 355. 5 M.L. Feldman and C. Dowd, Loss of dendrites spines in aging cerebral cortex. Anat. Erabryol., 148 (1975) 279-301. 6 Y. Geinisman, W. Bondareff and J. Dodge, Partial differentiation of neurons in the dentate gyrus of the senescent rat. Brain Res., 134 (1977) 541-545. 7 S.J. Buell and P.D. Coleman, Dendritic growth in the aged human brain and failure of growth in senile dementia. Science, 296 (1979) 854-856. 8 A.W. Klein and M.E. Michel, A morphological study of the neocortex of young adult and old maze-differentiated rats. Mech. Ageing Dev., 6 (1977) 44-452. 9 C. Stone, The age factor in animal learning: rats in the problem box and the maze. Genet. Psychol. Monogr., 5 (1929) 1-130. 10 C.L. Goodrick, Error goal-gradients of mature-young and aged rats during training in a 14-unit spatial maze. Psychol. Rep., 32 (1973) 359-362. 11 E. Verzer-McDougall, Studies in learning and memory in aging rats. Gerontologia, 1 (1957) 6585. 12 D. Arenberg, A longitudinal study of problem solving in adults. Gerontologia, 1 (1957) 65-85. 13 D. Friedman, Interrelations of two types of immediate memory in the aged. J. Psychol., 87 (1974) 177-183. 14 H. Wisnicwski and R.D. Terry, Neuropathology of the aging brain. In R. Terry and S.Gershon (Editors), Neurobiology o f Aging [Aging, Vol. 3], Raven Press, New York, 197"6, pp. 265-280. 15 R. Glick and W. Bondareff, Loss of synapses in the cerebellum of the senescent rat.,/. Gerontol., 34 (1979) 818-822. 16 J.P. Machado-Salas and A.B. Scheibel, Limbic system of the aged mouse. Exp. Neurol., 63 (1979) 347-355. 17 J.F. Marshall and N. Berries, Movement disorders of aged rats: reversal by dopamine receptor stimulation. Science, 206 (1979) 477-479. 18 J. Legrand, Morphogenetic actions of thyroid hormones. Trends NeuroscL, 9 (1979) 234-236. 19 R.P. Hammer, U. Tomlyasu and A.B. Scheibel, Degeneration of the human Betz cell due to amyotrophic lateral sclerosis. Exp. NeuroL, 63 (1979) 336-346. 20 H.E. Hirsch, The chemistry of motor neurons: Research strategies. In John M. Andrews (ed.), Amyotrophic Lateral Sclerosis, Recent Research Trends, Academic Press, New York, 1977, pp. 87-99. 21 H. M6hler and T. Okada, The benzodiazepine receptor in normal and pathological human brain. Br. J. Psychiatry, 133 (1978) 261-268.
255 22 R.T. Bartus, Physostigmine and recent memory: Effects in young and aged non-human primates. Science, 206 (1979) 1087-1089. 23 K.R. Brizzee, N. Sherwood and P.S. Timiras, Comparison of cell populations at various level of cerebral cortex of young-aduR and age Long-Evans rats. Y. Gerontol., 23 (1968) 289-297.
245
SYNAPTIC DENSITY CORRELATED WITH MAZE PERFORMANCE IN YOUNG AND AGED RATS. A PRELIMINARY STUDY
ALBERT WILLIAM KLEIN Department of Anatomy, Eastern Virginia Medical School, 700 Olney Road, Norfolk, Virginia 23.501 (U.S.A.) (Received August llth, 1980) (Revision receivedAugust 26th, 1982)
SUMMARY The aim of this study was to determine the effect of age both on maze performance and upon morphological changes in the brain. Young and senescent rats were differentiated into three groups by use of a complex 14-unit maze: young maze-bright rats, old maze-bright animals and old maze-dull rats. Layer III of the frontal and the occipital cortices of these animals was studied, employing coded electron micrographs, to determine the average number of synaptic profiles per area of thin section. It was observed that the young maze.bright rats had the highest synapse density, the old maze-bright rats had 85% of that amount while the old maze-dull animals had a density that was 54% that of the young animals.
Key words: Aged;Maze performance; Synaptic density
INTRODUCTION Several experimental studies have presented anatomical evidence for the decline of neuronal function with age. These studies have inferred that aging changes, such as loss of neurons, accumulation of lipofuscin pigment and loss of dendritic spines from cortical neurons, have functional correlations in aging brains [1-6]. The hypothesis is that these alterations hinder the total integrative potential of the aging brain and hence may contribute to clinical entities such as senile dementia. Apparently the number of synapses has great meaning for effective brain function. The careful work of Feldman and Dowd [4,5] has shown great variation as well as alterations m dendritic spine density that accompany aging. A specific alteration reported by Gemisman et al. [6] has been a decline in the number of axodenddtic synapses. Perhaps, because intercellular communication is so important in brain function, dendritic spines and synaptic frequencies play a major role in cerebral integration. The recent evidence of Buell and Coleman [7], who looked at dendritic arborization in man, supports the behavioral concepts and reports of wide differentials between cortical function in the aged human. 0047-6374/83/$03.00 Printed and Published in Ireland
© 1983 Elsevier ScientificPublishers Ireland Ltd.
246
The melding of behavioral and morphometric experimentation is an approach that has not been utilized to its fullest potential. The combination of these two disciplines is aimed at discovering those features of the aging cerebral cortex that are most likely to be involved with functional or behavioral brain deficits. Moreover, this method encourages speculation about possible, yet so far unproven, causal relationships I~etween neuronal morphology and senile dementia. This study was undertaken to examine the possible relationship between the effect of age on maze performance (learning ability), and on interneuronal communication, [e. synaptic frequency in aged rats. METHODS
Young and senescent female Wistar rats from the Gerontology Research Center in Baltimore, Maryland, were used in this investigation. Young animals were 8 months' old and senescent rats 25 months' old at the time of sacrifice. The mean life span of this strain of rats is 23 months. As described in preliminary work by Klein and Michel [8], performance on a 14-choice "T"-maze [9] behaviorally separated the rats into (1) a young maze-bright group, (2) an old maze-bright group, and (3) an old maze-dull group. These results agreed with the previous observations of Goodrick [ 10]. Ten young rats and 20 old rats began the training program. All 10 of the younger
MEDIAL
DORSAL
Fig. 1, Right telencephallc cortex of the rat.
247 animals were alive at the conclusion of the maze training; however, due to the advanced age of the older group, only 11 were alive at the end of the maze training. This was expected based on the experiences of Goodrick [10]. Using a book of random numbers, animals were selected as samples for electron microscopy. Neocortical samples from the three groups of rats were prepared for both light- and electron-microscopic study first by perfusion with Karnovsky's fluid through the left ventricle as previously reported [8]. The external flesh was removed and the skulls were placed overnight in fixative. The following day the brains were removed and placed in fresh fixative. After a 4-tl period in the fixative, wedge-shaped samples of the frontal and occipital cortex were removed for study of the neocortex by light microscopy and transmission electron microscopy. The removal of the wedges was made consistent by the method previously reported by Klein and Michel [8]. Because the rat cortex is lissencephalic, the total length was marked off into eight segments anterior to posterior. The wedges removed were from segment number 2 (being frontal) and segment number 7 (occipital) and were the second millimeter lateral from the exact midline or sagittal surface (see Fig. I). The latter samples were post-fixed in osmium and processed for sectioning. Thick sections of 1 /an were cut in a plane perpendicular to the cortex and examined to determine the location of layer III. Layer III was further isolated by trimming. Thin sections were cut with a diamond knife and examined on a Philips 301 electron microscope. The brains of four animals from each group (young maze-bright, old maze-bright and old maze-dull) were processed. Two blocks of tissue were taken from the frontal cortex and two blocks of tissue were taken from the occipital cortex (hence four blocks per rat). The pyramidal cell layer (layer Ill) was isolated and sections from each block were mounted on grids (resulting in eight grids per animal). Four photographs were taken in a vertical line across the tissue surface of each grid (~ :ithout regard to tissue content). See the flow diagram (Scheme 1) for a summary of the samples. The synapse profiles were counted per unit area of section. Synapse counts were averaged per unit area, frontal and occipital regional counts were pooled for each of the three functionally differentiated groups. Because there were only three separate groups statistical comparisons were made using Student's t-test. RESULTS As previously reported by this laboratory [8] and Goodrick [10], the maze training effectively differentiated three groups of rats: (1) young animals, all of which learned
Each experimental animal
x 2 regions (frontal and occipital)/animal X 2 blocks/region X 2 grids/block x 4 photos/grid X 4 animals/group X 3 experimental groups Scheme 1. Flow diagram of samples.
Total samples
therefore therefore therefore
= = = = = =
2 regions/animal 4 blocks/animal 8 grids/animal 32 micrographs/animal 128 photomicrographs/group 384 photomicrographs used
248 the maze well (only one error on day 20, only one trial per day); (2) old animals which performed as well as the young rats (one error or less on day 20); and (3) old animals which committed approximately eight errors repetitively even on day 20 o f the trials. Of those old animals that did not survive the entire maze shaping and maze training paradigm some were from the bright group and some from the dull group.
Fig. 2. Layer III neuron from an 8-month-old rat specimen. Observe the presence of only slight amounts of lipofusein (Lf). Also seen are examples of smooth endoplasmic retieulum (SR), lysosomes (Ly), an axon (Ax) and a synapse (Sy). The reference bar is equivalent to 1 ~m (print magnification X8910).
249 All of the electron rnicrographs were examined while they were still coded so that examiners had no knowledge of the experimental groups. Upon decoding the data it was obvious that between the three groups there were many similarities and yet some striking differences. Lipofuscin pigment was variably present in all brain specimens. All neurons observed did not contain pigment, but all specimens, even the 8-monthold animals, did contain some lipofuscin pigment in some neurons (Fig. 2). Some neurons
Fig. 3. This view contains larger granules of the aging pigment lipofuscin (Lf) associated with lysosomal bodies (Ly). Note the membrane which seems to surromld the two substances. Ribosomes (Rb) are in abundance.
250 in sample sections of the older animals appeared to be free of pigment. Subjectively, it appeared that neurons of older animals contained more lipofuscin granules or that the granules were larger than in younger animals. Exact quantification of lipofuscin pigment was not attempted in this study. Qualitatively the deposits appeared to be the same. The lipofuscin was often adjacent to lysosomal bodies, and at higher magnifications it was actually seen to be contained within the lysosomal membranes (Fig. 3).
Fig. 4. Neuron and perineuronal glial cell degeneration was observed only in the older animals. Ineluded here are myelin whorls (MW) and granulovascnlar degeneration (Gv).
251 TABLE I AVERAGE NUMBER OF SYNAPSES PER SQUARE MILLIMETER OF EXAMINEDCORTICAL TISSUE
Mean number ± S.D. Percentage (%)
Young adult maz e-brigh t
Old maz e-brigh t
Old ma z e 4 u i l
395.7 -+31 100
338.8* ±58 85.6
215.6"** -*53 54.5
*Different from the bright group of old animals at the p < 0.01 level. **When pooled, the old groups together differed from the young animals at the p < 0.05 level. Neurons from young specimens consistently appeared as healthy, intact cells while those in older brain samples consistently exhibited neurons in various stages of disarray. These older, degenerating neurons contained what appeared to be fatty granules, signs of granulovaseular degeneration and myelin whorls (Fig. 4). This type of extensive cell change was observed only in the neurons of older rats. Synapses were counted from the systematically sampled micrographs and their numbers were expressed per square mm of thin section as in Table I. Synapses with discontinuous pre- and postsynaptic membrane thickenings (double membrane interrupted as in Fig. 5) were counted as one synapse. Figure 5 is included to illustrate the type of sample photograph and the magnification from which synapse counts were made. To show a micrograph from a young animal with many synapses and to show a micrograph from an old, dull animal would be a subjective presentation. All micrographs were used and therefore no subjective selecting of rnicrographs occurred. Synaptic configurations which exhibited vesicles as well as both presynaptie and postsynaptic membrane thickenings were the only structures counted. The average number of synapse prof'des seen in 1 square mm of thin section was 395.7 for the young adult, maze.bright group, 338.8 for the old maze-bright animals, and 215.6 for the old maze-dull group. When synapses for the young adults were considered as I00%, the old-brights and old-dulls were only 85% and 54%, respectively, of the young bright group. When pooled, the two older groups of animals were significantly different from the young animals at the p < 0.05 level. The old maze-dull group was different from the old maze.bright group at the p < 0.01 level. Because section thickness, synaptic dimensions and random synaptic orientation were not taken into account, the numbers in Table I were intended to reflect relative synaptic numbers, present in any given volume of layer III of neocortex. DISCUSSION Verzar-McDougall [11], Goodrick [10] and Klein and Michel [8] have demonstrated that a complex 14.unit maze is a reliable discriminator of learning performance in aged rats. These investigators have been able to differentiate old maze.bright rats from old maze-dull rats. T h e criteria for differentiation is that the old maze.bright animals per-
252
Fig. 5. This illustration depicts the type of photomicrograph and the magnification from which the synaptic counts were tallied. Several, but not all, synaptic sites have been marked (Sy). Synapses with discon~tuous double membranes, as in the insert, were counted as a single synapse. form as well as the young animals. Human studies which have ruled out m o t o r coordination as a prime factor and which have required a person to learn and trace a complex maze with a stylus have resulted in similar findings. Younger subjects performed with fewer errors than did the average of the elderly subjects. However, some elderly subjects
253 did perform as well as the best younger subjects, while certain of the elderly never mastered the maze [12,13]. These separate studies suggegt a possible common central nervous system mechanism for decline of performance with aging. Also, these studies support the inequity of performance decline with age; that is, not all individuals become infirm with advanced age. The study of structural, functional and biochemical changes with aging has elucidated a number of possible explanations for decline of cerebral performance with senescence. Classically, Brody [1] has documented cell loss with age in the cerebral cortex, and more recently in the locus coeruleus (brain stem) [2], a system with widespread projections. Wisnicwski and Terry [14] have shown an increase in neurofibrillary tangles with increasing age and senility. Feldman and Dowd [5] have thoroughly outlined the loss of dendritic spines with aging. The age-related loss of dendritic spines and synapses has also been studied by Bondareff and co-workers and has been shown to effect losses of axodendritic synapses sometimes more selectively than axosomatic synapses [15]. Synapse loss has also been reported in brain structures implicated in recent memory [6]. Scheibel even suggests that impaired neuronal connections may result in inappropriate adaptive responses [16]. Plasticity studies have pointed out functional deficits with advancing age [17], and that these deficits may be the result of plastogenic action of hormones [18]. Lipofuscin pigment accumulation within neurons may influence the movement of RNA and lead to poor intracellular economy and thereby contribute to a decline in neuronal function [3]. Hammer et al. [19] and others [20] implicate lipofuscin pigment as a disruptive presence in neurons, specifically the motor neurons in amyotrophic lateral sclerosis. Neuropharmacological changes with age have been increasingly documented. Primarily these changes were seen as specific decrements in neurotransmitters or receptors [21] with age. Certain transmitters have also been implicated in disease states common to be aged, such as presenile and senile dementia as well as Parkinsonism. From an alternative clinical approach, physostigmine and other manip ulations of the cholinergic system can affect recent memory in a positive manner [22]. This investigation is one of the few to correlate a function, such as maze performance or animal learning, to a structural entity like the synapse. However, it should not be taken as prima facie evidence that synaptic loss caused declines in performance. The two groups of old rats which did differentiate themselves, based on maze performanceold maze-bright and old maze-dull-did subsequently reveal that their synaptic densities were reduced from those of the young and from each other, both by age and by poor maze performance. However, correlation does not necessarily prove a cause-and-effect rehtionship. The synapses counted were from layer III of the frontal and occipital cortices. Both of these regions had previously failed to show changes in neuron numbers at the light.microscopic level [8,23 ]. The suggestion which needs verification by other laboratories is that old maze-dull animals have a significant decrement in intercellular communication sites and while a cause-and-effect relationship should not be implied, it is possible that synaptic alterations and plasticity play a major role in the determination of performance.
254 ACKNOWLEDGEMENTS The author is grateful for the suggestions of Drs. Donald Stein, R. Frederick Becker, Herbert Schapiro, Robert Brownson and Paul Coleman, and for the technical assistance of Ms. Sheila Rosenfield. Acknowledgement is made to the Gerontology Research Center, NIA and W. French for aged animals provided in the Guest Scientist Program. This work was funded in part by the Biomedical Research Support Grant RR09028 and in part by Eastern Virginia Medical School Institutional Research Support Funds. REFERENCES 1 H. Brody, Organization of the cerebral cortex. J. Comp. Neurol., 102 (1955) 511-556. 2 N. Vijayashbankar and H. Brody, Quantitative study of the pigmented neurons of the nuclear locus. Coeruleus and subcoeruleus in man as related to aging. Z Neuropathol. Exp. NeuroL, 38 (1979) 490-497. 3 D.M.A. Mann and P.O. Yates, Lipofuscin pigments their relationship to aging in the human nervous system: The lipofusein content in nerve cells. Brain, 97 (1974) 481-488. 4 M.L. Feldman and C. Dowd, Aging in rat visual cortex: light microscopic observations on layer V pyramidal apical dendrites.Anat. Rec., 178 (1974) 355. 5 M.L. Feldman and C. Dowd, Loss of dendrites spines in aging cerebral cortex. Anat. Erabryol., 148 (1975) 279-301. 6 Y. Geinisman, W. Bondareff and J. Dodge, Partial differentiation of neurons in the dentate gyrus of the senescent rat. Brain Res., 134 (1977) 541-545. 7 S.J. Buell and P.D. Coleman, Dendritic growth in the aged human brain and failure of growth in senile dementia. Science, 296 (1979) 854-856. 8 A.W. Klein and M.E. Michel, A morphological study of the neocortex of young adult and old maze-differentiated rats. Mech. Ageing Dev., 6 (1977) 44-452. 9 C. Stone, The age factor in animal learning: rats in the problem box and the maze. Genet. Psychol. Monogr., 5 (1929) 1-130. 10 C.L. Goodrick, Error goal-gradients of mature-young and aged rats during training in a 14-unit spatial maze. Psychol. Rep., 32 (1973) 359-362. 11 E. Verzer-McDougall, Studies in learning and memory in aging rats. Gerontologia, 1 (1957) 6585. 12 D. Arenberg, A longitudinal study of problem solving in adults. Gerontologia, 1 (1957) 65-85. 13 D. Friedman, Interrelations of two types of immediate memory in the aged. J. Psychol., 87 (1974) 177-183. 14 H. Wisnicwski and R.D. Terry, Neuropathology of the aging brain. In R. Terry and S.Gershon (Editors), Neurobiology o f Aging [Aging, Vol. 3], Raven Press, New York, 197"6, pp. 265-280. 15 R. Glick and W. Bondareff, Loss of synapses in the cerebellum of the senescent rat.,/. Gerontol., 34 (1979) 818-822. 16 J.P. Machado-Salas and A.B. Scheibel, Limbic system of the aged mouse. Exp. Neurol., 63 (1979) 347-355. 17 J.F. Marshall and N. Berries, Movement disorders of aged rats: reversal by dopamine receptor stimulation. Science, 206 (1979) 477-479. 18 J. Legrand, Morphogenetic actions of thyroid hormones. Trends NeuroscL, 9 (1979) 234-236. 19 R.P. Hammer, U. Tomlyasu and A.B. Scheibel, Degeneration of the human Betz cell due to amyotrophic lateral sclerosis. Exp. NeuroL, 63 (1979) 336-346. 20 H.E. Hirsch, The chemistry of motor neurons: Research strategies. In John M. Andrews (ed.), Amyotrophic Lateral Sclerosis, Recent Research Trends, Academic Press, New York, 1977, pp. 87-99. 21 H. M6hler and T. Okada, The benzodiazepine receptor in normal and pathological human brain. Br. J. Psychiatry, 133 (1978) 261-268.
255 22 R.T. Bartus, Physostigmine and recent memory: Effects in young and aged non-human primates. Science, 206 (1979) 1087-1089. 23 K.R. Brizzee, N. Sherwood and P.S. Timiras, Comparison of cell populations at various level of cerebral cortex of young-aduR and age Long-Evans rats. Y. Gerontol., 23 (1968) 289-297.