Dendritic Atrophy in the Dentate Gyrus of the Senescent Rat YURI GEINISMAN, WILLIAM BONDAREFF AND JOHN T. DODGE Department of Anatomy, Northwestern University Medical School, 303 E. Chicago Auenue, Chicago, Illinois 6061 1
ABSTRACT Quantitative electron microscopic analysis of the supragranular zone of the dentate gyrus molecular layer has shown that the number, volume fraction and surface area of dendritic shaft profiles are significantly decreased in senescent rats, relative to young adults. These modifications of dendritic morphology, which are not associated with age-related changes in dimensions of the molecular layer or in numbers of granule cells, may result from a decrease in the number andlor length of dendrites. In either case, the decreases in the number, volume fraction and surface area of dendritic shaft profiles found in the dentate gyrus of senescent rats signify an age-related atrophy of dendrites. Comparison of changes in the number and volume fraction of dendritic shaft profiles has demonstrated that age-related dendritic atrophy involves predominantly smaller dendritic branches. Atrophy of dendrites has been shown to result from partial deafferentation of vertebrate neurons, brought about by various experimental procedures (for a comprehensive review of the literature see Globus, '75). Atrophy of dendrites might also result from partial deafferentation brought about by the process of aging. This possibility is suggested by the observation that partial deafferentation, indicated by a loss of axo-dendritic synapses, occurs in the senescent brain (Bondareff and Geinisman, '76; Geinisman et al., '77). Although dendritic atrophy in senescence has been inferred from investigations of Nissl and Golgi preparations (Mann and Yates, '74; Machado-Salas e t al., '75; Scheibel e t al., '75, '76; Vaughan, '771, the extent of this atrophy is not readily evaluated, because Nissl and Golgi staining methods used in earlier studies must be interpreted with caution. The Nissl stain is not specific for dendrites and it is not known whether the great selectivity of the Golgi method, on which its usefulness depends, varies with age. By means of electron microscopy, but not the Golgi technique, all dendrites are revealed unselectively. To our knowledge, there has been only one electron microscopic study showing an age-related atrophy of dendrites. Hinds and McNelly ('77) described a significant decrease in the total volume and length of mitral cell denAM. J. ANAT. (1978)152: 321-330
drites in the rat olfactory bulb with advancing age. In the study reported here an attempt has been made to determine, by means of a quantitative electron microscopic analysis, whether the partial deafferentation of dendrites in the senescent brain is accompanied by changes in the number, volume fraction and surface area of dendritic shaft profiles. We have used the Fischer-344 male rat, which has been shown (Coleman et al., '77) to be essentially free from cerebral pathology. Since the mean life-span of this strain of rats is 29 months (Coleman et al., '771, we compared young adult animals, three months of age, with senescent, 25-month-old animals, in which the probability of death is increased (cf. Birren and Renner, '77). In these rats the supragranular zone of the dendate gyrus molecular layer was chosen for electron microscopic analysis, because age-related partial deafferentation of dendrites, manifested by the loss of axo-dendritic synapses, had been demonstrated previously in this zone (Geinisman e t al., '77). In addition, the supragranular zone borders the layer of granule cells and contains segments of main dendritic shafts located relatively close to their perikaryal origin, which are readily identified in electron micrographs (cf. Peters et al., '76). Accepted November 16, '37
321
322
YURI GEINISMAN, WILLIAM BONDAREFF AND JOHN T. DODGE
nification of 10,000 x and enlarged photographically 2.4 x , These micrographs were Five male, young adult ( 3 months of age) taken successively in two scans (each no more and five male, senescent (25 months of age) than 12.5 wm in width) through the suprarats of the Fischer-344 strain were obtained granular zone of the molecular layer, just dorfrom the Charles River Breeding Laboratories sal and parallel to the granule layer. From (Wilmington, Massachusetts). The animals each scan, the first 10 electron micrographs were anesthetized with chloral hydrate (350- which contained no images of cell bodies or 500 mg/kg, intraperitoneally) and perfused blood vessels were selected. Profiles of denwith Karnovsky’s fixative. After perfusion, dritic shafts were identified according to well the brain was dissected and a coronal slice established criteria (Peters e t al., ’76) and outabout 1 mm thick, which included the rostral lined in electron micrographs. The number of portion of the dentate gyri, was cut. From this profiles of dendritic shafts was counted per slice, the right hippocampal formation was ex- square area (66 pm2) of an electron microcised, post-fixed in the same fixative for an graph. The volume fraction and surface area additional hour, washed in 0.2 M cocadylate of dendritic shafts per volume of neuropil were buffer for 20 minutes, transferred into a 1% determined by means of stereological pointOsO, solution for one hour, dehydrated in a counting and intersection-counting prograded series of ethanol and propylene oxide cedures (Weibel, ’69; Elias et al., ’71; Weibel and embedded in Araldite-502. and Bolender, ’73). The dendritic volume fracThe supra- and infra-pyramidal limbs of the tion was estimated by superimposing a transdentate gyrus (Avgevine, ‘65; Chronister and parent lattice of regularly spaced test points White, ’75) were identified in 1-pm-thickcor- on electron micrographs and counting the onal sections. The region of the suprapyrami- number of test points falling on profiles of dal limb which is opposite the free, lateral end dendritic shafts. The total volume fraction of of the infrapyramidal limb was sectioned cor- dendritic shafts per unit volume of neuropil onally so as to obtain sets of 1-pm-thick and was calculated by dividing this number by the 75 nm-thick sections a t intervals of 150 pm. total number of test points of the lattice. The Sets of 3-5 sections 1p m thick were mounted dendritic surface area was estimated by on glass slides, stained with methylene blue superimposing a transparent lattice of paraland examined with a Zeiss light microscope. lel, equally spaced test lines on electron microSets of six to nine ultrathin sections were graphs and counting the number of intersecpicked up on 300-mesh copper grids, stained tions of dendritic shaft membranes with test with uranyl acetate and lead citrate and ex- lines. The total surface area of dendritic amined in a Hitachi HU-12 electron micro- shafts per unit volume of neuropil was calcuscope. A central section from each set was an- lated by the formula: Sv = 2 IL,where “IL”is alyzed. the number of intersections divided by the In sections prepared for light microscopy, total length of test lines of the lattice. the width of the molecular layer and the numIn preliminary experiments using two rats ber of granule cells were estimated. The width of each age, it was determined that when a latof the molecular layer was measured by means tice was positioned a t any of three equal of an ocular micrometer a t a magnification of angles around an axis through the center of an 100 x . In each section, the distance from the electron micrograph, the same mean values dorsal border of the granule cell layer to the for volume-fraction and surface area were obhippocampal fissure was determined at three tained in each reading. This means that the different points, a t the most medial, central, anisotrophy (i.e., the preferred orientation) of and most lateral parts of the sections. All dendrites in ventro-dorsal and medio-lateral granule cells sectioned through the nucleolus directions has no effect on measurements of were counted a t a magnification of 500 X volume-fraction and surface area of these neualong three successive, 45.7-pm-long seg- ronal processes. When the parameters under ments of the granule cell layer. The results of study were estimated in two ultrathin secthese counts were expressed as number of neu- tions, one from each opposite (rostral and caurons per 100-pm-longsegment of the granule dal) faces of the tissue block, or in six ulcell layer. trathin sections prepared a t a n interval of 150 Ultrathin sections were used to obtain elec- p m from one another, the results were not aftron micrographs taken a t a n initial mag- fected. Therefore, only two sections obtained MATERIALS AND METHODS
323
DENDRITIC ATROPHY IN SENESCENCE TABLE 1 Number, volume fraction and surface area of dendritic shafts in 3-month-old and 25-month-old rats Number of profiles
Mean 2 S.E.M. per animal
Mean 2 S.E.M. per group % o f3-monthold group Significance level
Volume fraction
I
Surface area
3 months
25 months
3 months
25 months
3 months
26.8%1.0 26.22 1.0 25.92 0.9 22.92 0.6 21.120.9
19.42 0.6 19.220.7 18.62 0.7 18.420.6 17.450.4
0.4072 0.016 0.421k 0.011 0.4222 0.015 0.365k0.013 0.4512 0.015
0.3462 0.012 0.406k0.015 0.3952 0.012 0.329k0.011 0.333*0.015
2.1320.07 1.902 0.06 1.77%0.07 1.78’0.08 1,9120.07
1.3120.08 1.632 0.06 1.46%0.06 1.21%0.04 1.29%0.06
24.62 1.1
18.620.4
0.413k0.014
0.36220.016
1.90%0.06
1.3820.07
25 months
75.6
87.6
72.6
P< 0.001
P
ABSTRACT Quantitative electron microscopic analysis of the supragranular zone of the dentate gyrus molecular layer has shown that the number, volume fraction and surface area of dendritic shaft profiles are significantly decreased in senescent rats, relative to young adults. These modifications of dendritic morphology, which are not associated with age-related changes in dimensions of the molecular layer or in numbers of granule cells, may result from a decrease in the number andlor length of dendrites. In either case, the decreases in the number, volume fraction and surface area of dendritic shaft profiles found in the dentate gyrus of senescent rats signify an age-related atrophy of dendrites. Comparison of changes in the number and volume fraction of dendritic shaft profiles has demonstrated that age-related dendritic atrophy involves predominantly smaller dendritic branches. Atrophy of dendrites has been shown to result from partial deafferentation of vertebrate neurons, brought about by various experimental procedures (for a comprehensive review of the literature see Globus, '75). Atrophy of dendrites might also result from partial deafferentation brought about by the process of aging. This possibility is suggested by the observation that partial deafferentation, indicated by a loss of axo-dendritic synapses, occurs in the senescent brain (Bondareff and Geinisman, '76; Geinisman et al., '77). Although dendritic atrophy in senescence has been inferred from investigations of Nissl and Golgi preparations (Mann and Yates, '74; Machado-Salas e t al., '75; Scheibel e t al., '75, '76; Vaughan, '771, the extent of this atrophy is not readily evaluated, because Nissl and Golgi staining methods used in earlier studies must be interpreted with caution. The Nissl stain is not specific for dendrites and it is not known whether the great selectivity of the Golgi method, on which its usefulness depends, varies with age. By means of electron microscopy, but not the Golgi technique, all dendrites are revealed unselectively. To our knowledge, there has been only one electron microscopic study showing an age-related atrophy of dendrites. Hinds and McNelly ('77) described a significant decrease in the total volume and length of mitral cell denAM. J. ANAT. (1978)152: 321-330
drites in the rat olfactory bulb with advancing age. In the study reported here an attempt has been made to determine, by means of a quantitative electron microscopic analysis, whether the partial deafferentation of dendrites in the senescent brain is accompanied by changes in the number, volume fraction and surface area of dendritic shaft profiles. We have used the Fischer-344 male rat, which has been shown (Coleman et al., '77) to be essentially free from cerebral pathology. Since the mean life-span of this strain of rats is 29 months (Coleman et al., '771, we compared young adult animals, three months of age, with senescent, 25-month-old animals, in which the probability of death is increased (cf. Birren and Renner, '77). In these rats the supragranular zone of the dendate gyrus molecular layer was chosen for electron microscopic analysis, because age-related partial deafferentation of dendrites, manifested by the loss of axo-dendritic synapses, had been demonstrated previously in this zone (Geinisman e t al., '77). In addition, the supragranular zone borders the layer of granule cells and contains segments of main dendritic shafts located relatively close to their perikaryal origin, which are readily identified in electron micrographs (cf. Peters et al., '76). Accepted November 16, '37
321
322
YURI GEINISMAN, WILLIAM BONDAREFF AND JOHN T. DODGE
nification of 10,000 x and enlarged photographically 2.4 x , These micrographs were Five male, young adult ( 3 months of age) taken successively in two scans (each no more and five male, senescent (25 months of age) than 12.5 wm in width) through the suprarats of the Fischer-344 strain were obtained granular zone of the molecular layer, just dorfrom the Charles River Breeding Laboratories sal and parallel to the granule layer. From (Wilmington, Massachusetts). The animals each scan, the first 10 electron micrographs were anesthetized with chloral hydrate (350- which contained no images of cell bodies or 500 mg/kg, intraperitoneally) and perfused blood vessels were selected. Profiles of denwith Karnovsky’s fixative. After perfusion, dritic shafts were identified according to well the brain was dissected and a coronal slice established criteria (Peters e t al., ’76) and outabout 1 mm thick, which included the rostral lined in electron micrographs. The number of portion of the dentate gyri, was cut. From this profiles of dendritic shafts was counted per slice, the right hippocampal formation was ex- square area (66 pm2) of an electron microcised, post-fixed in the same fixative for an graph. The volume fraction and surface area additional hour, washed in 0.2 M cocadylate of dendritic shafts per volume of neuropil were buffer for 20 minutes, transferred into a 1% determined by means of stereological pointOsO, solution for one hour, dehydrated in a counting and intersection-counting prograded series of ethanol and propylene oxide cedures (Weibel, ’69; Elias et al., ’71; Weibel and embedded in Araldite-502. and Bolender, ’73). The dendritic volume fracThe supra- and infra-pyramidal limbs of the tion was estimated by superimposing a transdentate gyrus (Avgevine, ‘65; Chronister and parent lattice of regularly spaced test points White, ’75) were identified in 1-pm-thickcor- on electron micrographs and counting the onal sections. The region of the suprapyrami- number of test points falling on profiles of dal limb which is opposite the free, lateral end dendritic shafts. The total volume fraction of of the infrapyramidal limb was sectioned cor- dendritic shafts per unit volume of neuropil onally so as to obtain sets of 1-pm-thick and was calculated by dividing this number by the 75 nm-thick sections a t intervals of 150 pm. total number of test points of the lattice. The Sets of 3-5 sections 1p m thick were mounted dendritic surface area was estimated by on glass slides, stained with methylene blue superimposing a transparent lattice of paraland examined with a Zeiss light microscope. lel, equally spaced test lines on electron microSets of six to nine ultrathin sections were graphs and counting the number of intersecpicked up on 300-mesh copper grids, stained tions of dendritic shaft membranes with test with uranyl acetate and lead citrate and ex- lines. The total surface area of dendritic amined in a Hitachi HU-12 electron micro- shafts per unit volume of neuropil was calcuscope. A central section from each set was an- lated by the formula: Sv = 2 IL,where “IL”is alyzed. the number of intersections divided by the In sections prepared for light microscopy, total length of test lines of the lattice. the width of the molecular layer and the numIn preliminary experiments using two rats ber of granule cells were estimated. The width of each age, it was determined that when a latof the molecular layer was measured by means tice was positioned a t any of three equal of an ocular micrometer a t a magnification of angles around an axis through the center of an 100 x . In each section, the distance from the electron micrograph, the same mean values dorsal border of the granule cell layer to the for volume-fraction and surface area were obhippocampal fissure was determined at three tained in each reading. This means that the different points, a t the most medial, central, anisotrophy (i.e., the preferred orientation) of and most lateral parts of the sections. All dendrites in ventro-dorsal and medio-lateral granule cells sectioned through the nucleolus directions has no effect on measurements of were counted a t a magnification of 500 X volume-fraction and surface area of these neualong three successive, 45.7-pm-long seg- ronal processes. When the parameters under ments of the granule cell layer. The results of study were estimated in two ultrathin secthese counts were expressed as number of neu- tions, one from each opposite (rostral and caurons per 100-pm-longsegment of the granule dal) faces of the tissue block, or in six ulcell layer. trathin sections prepared a t a n interval of 150 Ultrathin sections were used to obtain elec- p m from one another, the results were not aftron micrographs taken a t a n initial mag- fected. Therefore, only two sections obtained MATERIALS AND METHODS
323
DENDRITIC ATROPHY IN SENESCENCE TABLE 1 Number, volume fraction and surface area of dendritic shafts in 3-month-old and 25-month-old rats Number of profiles
Mean 2 S.E.M. per animal
Mean 2 S.E.M. per group % o f3-monthold group Significance level
Volume fraction
I
Surface area
3 months
25 months
3 months
25 months
3 months
26.8%1.0 26.22 1.0 25.92 0.9 22.92 0.6 21.120.9
19.42 0.6 19.220.7 18.62 0.7 18.420.6 17.450.4
0.4072 0.016 0.421k 0.011 0.4222 0.015 0.365k0.013 0.4512 0.015
0.3462 0.012 0.406k0.015 0.3952 0.012 0.329k0.011 0.333*0.015
2.1320.07 1.902 0.06 1.77%0.07 1.78’0.08 1,9120.07
1.3120.08 1.632 0.06 1.46%0.06 1.21%0.04 1.29%0.06
24.62 1.1
18.620.4
0.413k0.014
0.36220.016
1.90%0.06
1.3820.07
25 months
75.6
87.6
72.6
P< 0.001
P
