Light and electron microscopic analysis of two divisions of the suprachiasmatic nucleus in the young and aged rat.код для вставкиСкачать
THE ANATOMICAL RECORD 237:71-88 (1993) Light and Electron Microscopic Analysis of Two Divisions of the Suprachiasmatic Nucleus in the Young and Aged Rat WILLIAM H. WOODS, ERVIN W. POWELL, ANNETTE ANDREWS, AND CHARLES W. FORD, JR. Department of Biology, Philander Smith College and Department of Anatomy, University of Arkansas for Medical Sciences, Little Rock, Arkansas ABSTRACT The suprachiasmatic nucleus (SCN) is a principal controller of mammalian circadian rhythms. However, in spite of documented disturbance of biological rhythms in old animals, few significant age-related changes have been observed in this nucleus. This study examined agerelated differences in SCN volume, neuronal number, density, and ultrastructural features in the entire rat SCN and in its two divisions, the denser ventromedial (compacta)and less dense dorsolateral (dissipata). Light and electron microscopic morphometric techniques were utilized in weanlings (21-28 days), young adults (3-6 mo), and aged ( 3 M 6 mo) animals. The total SCN volume, as well as volumes of the compacta region, were significantly greater in young adult and aged rats than in weanlings. Thus, as the rat ages the SCN increases in total size. However, the dissipata region appears to decrease in volume while the compacta increases. Even though the total number of SCN neurons was quite constant in the three age groups, the number of neurons in the dissipata region was decreased significantly in the young adult and aged groups as compared to the weanling. Neurons in the compacta region were usually spindled-shaped with two dendritic processes, while oval to spheroidal cells with 3-4 processes predominated in the dissipata. Nuclei of SCN cells were often invaginated. In weanlings, more SCN neuronal nuclei had invaginated nuclei in the dissipata region (66%) compared to the compacta (37%). In the two older age groups of rats, a higher percentage of invaginated neuronal nuclei were found in both regions. However, more were still found in the dissipata (90%) compared to the compacta (72%),even though the number of these cells in the compacta doubled. Thus, there was a large increase in the number of invaginated nuclei, as well as the number of invaginations, in the young adult rats compared to the weanling group, and this increase persisted in aged rats. SCN neurons usually had nuclei surrounded by a thin perimeter of cytoplasm containing sparse mitochondria and granular endoplasmic reticulum, multiple Golgi regions, and a moderate number of free ribosomes. In weanlings, mitochondria contained dense cristae and the granular endoplasmic reticulum was relatively prominent. Degenerative ultrastructural changes which included mitochondria1 enlargementlvacuolation, Golgi vacuolation, lysosome, and lipofuscin development occurred in less than 10%of young adult SCN cells, and were more frequently found in the dissipata. In aged, rats 30%of the neurons showed degenerative changes in the dissipata compared with 18%in the compacta. Degenerative changes appeared highly correlated with the degree of membrane folding. Ultrastructural degenerative cell changes and light microscopic morphometric observations are discussed in relation to loss of circadian rhythms with advancing age. o 1993 WiIey-Liss, Inc. Key words: Rat, Aging, Light morphometric, Degeneration, Ultrastructural 0 1993 WILEY-LISS. INC Received April 20, 1992; accepted April 15, 1993. Address reprint requests to Dr. William H. Woods, Department of Biology, Philander Smith College, 812 W. 13th, Little Rock, AR 72202. 72 W.H. WOODS ET AL. Since the early reports by Moore and Eichler (1972) and Stephan and Zucker (1972) which showed a circadian rhythm control exerted by the SCN over adrenocorticosterone, drinking, and locomotion activity rhythms, many studies have followed which tend t o support this concept. Two subsequent reviews (Raisman and Brown-Grant, 1977; Rusak and Zucker, 1979) documented evidence for the SCN involvement in other major circadian rhythms including the estrous rhythm, pineal N-acetyltransferase activity, sleep, water and food intake, and others. More recent studies in the rat have confirmed and extended the concept of SCN influence on biological rhythms related to drinking (Drucker-Colin et al., 1984), insulin and glucose homeostasis (Yamamoto et al., 1984), sleep (Tobler et al., 1983; Eastman et al., 1984),temperature (Powell et al., 1979; Eastman et al., 19841, mitotic activity (Powell et al., 1980) and neuro-transmitter receptors (Kafia et al., 1981, 1985). Additional evidence in support of the central influence of the SCN on circadian periodicity has been provided by transplantation experiments in which fetal SCN tissue introduced into the third ventricle of SCN-lesioned, arrhythmic animals restored the lost drinking behavior in the rat (Drucker-Colin et al., 1984; Ralph et al., 1990) and activity rhythm in hamsters (Lehman et al., 1987; Martin et al., 1990). In view of the numerous studies which support the SCN as the principal controller of mammalian circadian rhythm and the fact that many major rhythmic activities are fragmented in advancing age in humans and laboratory animals (see Van Goo1 and Mirmiran, 1986; Miles and Dement, 1980; Ingram et al., 1982, for reviews), several studies have examined cell loss and/or degenerative changes in the SCN of aged animals. Peng et al. (1980) showed that running activity and food intake decreased with age in the rat, but were unable to show a significant loss in total SCN neurons. Although this has been confirmed for the total cell number, a decrease in a specific small population of cells, i.e., vasopressin (VP) cells, within the SCN along with an increase in nuclear volume and diameter and a decrease in cell density occurred in the rat with age (Roozendaalet al., 1987).A similar decrease in VP cells in the SCN of aged human subjects and a “marked” decrease in the volume of the nucleus has also been observed (Fliers and Swaab, 1986). Ogata (1986) has demonstrated a number of age-related changes in the dorsomedial part of the rat SCN which he correlated with age-related loss in circadian rhythmicity. These included a decrease in the number of neurons and mitochondria and an increase in the number of lysosomes, lipofuscin, and dark degenerating cells. The purpose of the present investigation was to determine if age related differences in nuclear volume, neuronal number, density, and ultrastructure occur during aging in the dense medial and the less dense lateral divisions of the rat SCN. imals were housed under the standard laboratory conditions of a 12/12 light-dark cycle with food and water ad libitum. Litters from breeding pairs were maintained under these conditions until they were weanlings (21-28 days) or young adult age (3-6 mo). Retired breeders were kept until old age (30-36 mo). Light Microscopic and Morphometric Procedures Seventeen animals were used for light microscopic procedures: the first group (n = 8) were 21-28-day-old weanlings which weighed 65 to 75 gm; a second group (n = 3) were young adults (3-6 mo of age) that weighed 250 to 350 g, and the third group (n=3) were aged animals (30-36 mo) weighing 300-400 g. The animals were sacrificed under deep nembutal anesthesia (50 mgkg of body weight) during the morning hours of 1O:OO to 12:OO and their brains were flushed by intracardial perfusion with 0.9% saline and fixed immediately by perfusion with 10% formalin in a normal saline solution. The brains were then removed and immersed in 10% formalin containing 20% sucrose for 3-7 days. The brains were frozen and 30 pm sections were cut through the SCN. The sections were serially mounted, stained with cresyl violet, and coverslipped for study. Drawings of the SCN were made from these slides. SCN neurons were counted utilizing a calibrated ocular grid a t a magnification of 250 x . The method used for determining the total number of neurons, nuclear volume, and cell density has been previously reported (Horikawa et al., 1988). Except for the fact that separate determinations were made for the ventromedial (dense or compacta) and the dorsolateral (less-dense or dissipata) parts of the SCN (Fig. l),the method was essentially identical. We prefer to call these two regions compacta and dissipata, respectively. These regional designations are similar to those used for other CNS areas, e.g., compacta and reticulata of the substantia nigra. Counts of neurons were made by taking 6 samples from each of the compacta and dissipata regions of the serially sectioned SCN; each sample was from a different region of a given section and had a volume of 0.000014 mm3. Counts from these six samples were averaged, and the average was divided by the volume of a sample to obtain an average density. Only neurons sectioned through a nucleolus were recorded to avoid counting a neuron more than once. The volume of each SCN (or its compacta and dissipata regions) were determined by first tracing the scaled (16 x ) image outline of the respective areas onto drawing paper as described by Konigsmark et al. (1969). All drawings were made by the same investigator who utilized the most prominent appearing cell free zone in establishing nuclear boundaries. Then, the area was determined by using a calibrated digitizing tablet (Sumagraphics) with a built-in microprocessor. The volume of a section of a nucleus was determined by multiplying the area by the thickness. The volumes of all the sections of the two MATERIALS AND METHODS regions were added to obtain their total respective volThis study utilized light microscopical and ultra- umes. The formula used to calculate the total number structural techniques, both of which involved quanti- of neurons, then, is tative procedures. A combined total of 27 SpragueN = Dc x Vc + Dd x Vd Dawley rats of both sexes, purchased from the Charles Rivers Laboratories as retired breeders (8-10 mo) and where N = total number of neurons; Dc = neuronal breeding pairs (3 mo), were used in this study. All an- density (average number of neurons per mm3 in the EM ANALYSIS OF RAT SCN 73 using the same procedure employed for nuclear area measurements. An average of 142 cells per animal were traced from the compacta and dissipata regions of the rostral, middle and caudal SCN. The Kruskal-Wallis non-parametric test was used to determine significance between the measured parameters in the three groups of rats. If significance existed, the groups were paired (weanlings vs. young adult; weanlings vs. old adult; and young adult vs. old adult) then, the Mann-Whitney U-test was used to determine which particular groups were significantly different from each other. Ultrastructural and Morphometric Procedures Fig. 1. Coronal, cresyl violet stained 30 pm sections through the SCN a t rostral (A), middle (B), and caudal (C) levels of the nucleus. The dense ventromedial compacta (C) and less dense dorsolateral dissipata (D) regions are evident at each level. V, third ventricle; OC, optic chiasm. x 250. compacta region); Dd = neuronal density in dissipata region; Vc = volume of compacta; and Vd = volume of dissipata. In addition, area determinations for 428 individual SCN neurons from 3 young adult rats were made by Electron microscopic (EM) procedures and analysis were carried out on 10 Sprague-Dawley rats which belonged to 3 age groups; 21-28 days (n=3), 3-6 mo (n = 3), and 30-36 mo (n = 4), and weighed 70-80,200300, 300-500 g, respectively. All animals were sacrificed as described above. After the saline flush, they were perfused with a mixture of 1.5% glutaraldehyde and 2% paraformaldehyde solution for 5 min followed by a 4% glutaraldehyde solution in 0.1 M phosphate buffer (pH 7.3) for 15 min. Brains of the animals were removed rapidly and immersed in 4% glutaraldehyde after which they were blocked into an area containing the hypothalamus. The latter was subdivided into medial and lateral hypothalamic nuclear groups and then into dorsal and ventral tiers consisting of strips 1-1.5 mm in depth, 1-1.5 mm in width, and 4 mm in length under a dissecting microscope with the use of an atlas of the rat brain (Konig and Klippel, 1963). Major structural land marks (anterior commissure, 3rd ventricle, optic chiasm, optic tracts, mammillary bodies, fornix, etc.) were used as guides during the dissecting procedures. Brain strips containing the SCN were rinsed in 0.2 M phosphate buffer, post-fixed in 1%osmium tetroxide in 0.1 M phosphate buffer, dehydrated, embedded in epon in flat molds and allowed to harden. The blocks were coronally sectioned in thick and thin sections on a Nova I1 ultratome. Thin sections (60 nm) were taken following each 1 p,m thick section through the rostrocaudal extent of the SCN. Thick sections were stained with toluidine blue and used for light microscope confirmation of location within the SCN. Thin sections were stained with uranyl acetate and lead citrate and studied with a Philips 300 electron microscope. More than a thousand (1,030) electron micrographs were made from the SCN. Precise localization of the cells dorsally and ventrally in the two halves of the nucleus was confirmed from thick sections. Routinely, the grid section which included the greatest number of neurons under low magnification was photographed. Selected cell(s) from the group were then photographed at higher magnifications. Cells were analyzed for degenerative changes in the SCN, as defined under cytological age changes in the results section of this report. The ratio of degenerative cells to “normal” ones was compared between the compacta and dissipata regions of the SCN within a given age group as well as between the 3 different age groups. Based upon criteria detailed below (see Results) the invagination of the nuclear envelope is considered a degenerative change in SCN neurons. Therefore, the percentage of degenerative 74 W.H. WOODS ET AL. TABLE 1. Comparison of volume, cell number and density in weanling, young and aged rats* ~ Category Compacta volume ( ~ m ~ ) Dissipata volume (km3) Total SCN volume (pm3) Cell number for compacta region Cell number for dissipata region Total SCN cell number Compacta region density p~m-~) Dissipata region ~ m - ~ ) density Total SCN density (1OP urnp3) Weanling f S.E.M. n = 6 29,224,000 t 2,491,5781p2 8,365,000 ? 1,179,927182 37,542,332 t 2,083,3551.2 Young adult & S.E.M. n = 3 41,319,000 f 614,5001 4,230,667 2 335,909l 45,549,668 f 758,638l Aged adults 2 S.E.M. n = 3 43,683,000 218,011' 4,638,000 f 476,323' 48,321,000 279,086' * * 11,134 ? 860 11,834 2 987 11,373 f 3,013 1,219 & 285ls' 12,353 ? 960 403 f 15l 12,237 999 334 2 100' 11,706 3,112 * 386 ? 144 * 20 93 329 ? 15 269 14l,' 288 2 * 211 264 *4 2 18 2 71' 70 * 17 244 65 *Significant difference is indicated by and (P<0.025).The SCN compacta region volume (P<0.025) and total SCN volume (P<0.025)means of the young and aged animals were significantly greater than the means for weanlings. The dissipata region volume (P<0.025) and cell No. (P<0.025) and the compacta region density (FY0.025)for weanlings were significantly higher than those for the young and aged rats. cells with this characteristic was also determined in the young adult and aged rats. More than 1,400 cells (an average of more than 140 cells per animal) were used in this analysis. The Kruskal-Wallis non-parametric test was used to determine if the occurrence of nuclear membrane invaginations in SCN neurons was significantly different between the three groups of rats and between the compacts and dissipata regions of the SCN. The MannWhitney U-test was applied to see if significance was found when the groups were paired (weanling vs. young adult; weanling vs. old adult; young adult vs. old adult). RESULTS Light Microscopic Morphometrics The results of this study relating to total cell counts, nuclear volume, and neuronal density of the SCN are summarized in Table 1. Mean determinations for these parameters were essentially the same in the SCN on the right and left sides of the brain. Therefore, the figures in the table represent only those for the right SCN. The mean number of SCN compacta cells was 11,134, 11,834, and 11,373 in weanlings, young adults and old rats, respectively. On the other hand, the dissipata region contained 1,132, 403, and 334 cells in these respective groups. The total number of cells for the SCN (compacta and dissipata regions) was slightly less in the aged (11,706) as compared with weanlings and young adult rats (12,353 and 12,237, respectively). The difference in total cell number for the SCN between the three groups was not significant. However, the difference in the number of cells in the dissipata region of the nucleus was significant (P <0.025) in the weanling as compared with the young and old adult groups, but not between the young adult and old rats. Likewise, the mean density of cells in both parts of the SCN was reduced in young adult (288 cells x lop6 ~ m - and ~ ) old animals (264 cells x pmP3) as contrasted with the cell density of the weanlings (386 cells x ~ m - ~but ) , the reduction was significant only for the compacta region. The mean total volume for the combined compacta and dissipata parts of the SCN was calculated to be 0.038 mm3 in the weanlings as compared to 0.046 mm3 in young adults and 0.048 mm3 in the old rats. The mean compacta region volume (P <0.025) and total volume (P <0.05) of the young adult and old animals were significantly greater than the same means for the weanling group. However, the mean dissipata region volume (P ~ 0 . 0 2 5 of ) the weanling was significantly greater than those for the young adult and old animals. General morphology of SCN cells General morphological features of SCN cells observed from the compacta and dissipata regions of the nucleus (Fig. 1) as determined from light microscopic and low magnification of thin sections are briefly described here. The cells which dominated the compacta region of the SCN as determined from light and electron microscopy were small to moderate in size, averaging between 50 and 60 pm2 and had a mean average of 56.6 4.0 pm2. The large prominent nuclei of SCN neurons were usually surrounded by a small amount of cytoplasm, which was most abundant at cell poles or the site of origin of their processes. Characteristically, one primary process emanated from each pole of these spindle shaped cells (Fig. 2A) and less frequently, one process was observed. Although these cells were dispersed throughout the nucleus, they appeared more prevalent and concentrated in the ventral and medial compacta region of the nucleus. Larger oval to spheroid-shaped cells with up to 4 processes and oval to spheroid nuclei were more prevalent in the entire lateral division of the SCN but were especially obvious in the dissipata region of the nucleus (Fig. 2B). These were usually between 60 and 80 pm2 and had a mean average of 75.3 2.5 pm2. The cytoplasm and organelles were more abundant in these cells, thus accounting for their larger size. As a rule, the perikarya of medially located cells were closely packed and oriented in dorso-ventral rows, while those laterally had a dorsolateral orientation. It should be emphasized that in the extreme dorsomedial part of the * * EM ANALYSIS OF RAT SCN Fig. 2. A The cell in this electron micrograph was taken from the dissipata of a 5-mo-old rat. Note the spindle shape, the initial extension of two processes from the soma (arrows) and the more abundant cytoplasm at opposite poles of the cell. B A ventrolateral dissipata SCN cell of a 5-mo-old rat. The nucleus of this cell is surrounded by a 75 larger area of cytoplasm with more abundant organelles than the smaller medial cells. The site of origin of three processes is shown by arrowheads. A number of lysosomes (arrow) can be seen at this low magnification. 76 W.H. WOODS ET AL. smooth nuclear membrane. Cells with smooth nuclear membranes were located throughout the SCN, but constituted the main type found in the compacta region (Fig. 5). As Table 2 shows, the nuclear membrane of neurons in the dissipata region of the SCN were usually mildly invaginated (58.7%) but often smooth (34%),while extensively invaginated ones were uncommon (7.3%). In the SCN compacta the percentage for neurons with mildly invaginated (35.6%) and smooth Ultrastructural features (63.2%)nuclear membranes was nearly the reverse of The most distinctive appearing organelles within the the dissipata. Only 1.2% of the cells in the compacta perikarya of SCN cells were mitochondria, the Golgi region showed an extensively invaginated nuclear apparatus, and the granular endoplasmic reticulum. membrane (Table 2). Except for a few minor differMitochondria were moderate in number and usually ences, the ultrastructural features of major organelles more heavily concentrated on one side of the cell. They in most SCN neurons of both the compacta and dissicontained dense transversely and sometimes horizon- pata regions (Figs. 5, 6) had essentially the same gentally oriented cristaes, and varied in shape from round eral morphology as described under general ultrastrucor oval to an elongated sausage form. The latter was tural features. The cristae of mitochondria were often constricted andlor bent in an arc (Figs. 3,4) such distinct and relatively dense but the organelles were that the two ends would occasionally meet and fuse rarely severely constricted, vacuolated or had unusual forming “dumbbell” circular or other configurations configurations. The Golgi apparatus tended toward the such as C or S shapes especially in adult and aged rats flattened or disk-shaped type. Granular endoplasmic (Fig. 3,4). Golgi bodies were multiple and usually well reticulum, although not abundant, appeared more developed. They were composed mainly of flattened prominently present and free ribosomes appeared to disk-shaped vesicles along with a small spherical type have a higher density than in older rats. Lysosomes of vesicle. In many instances the flattened vesicles and dense bodies were prominent in some cells but mawere in a circular or semicircular pattern and appeared ture lipofuscin granules were not observed. The cytoas a nearly concentrically laminated structure. The plasm was generally free of obvious vacuolations, exgranular endoplasmic reticulum was sparse to moder- cept for the flattened ones along the inner surface of ate in amount and located peripherally or at the poles the cell membrane. Nematosomes were occasionally of the cell from which primary dendritic processes observed as were nuclear vacuoles and inclusion bodarose (Fig. 3). Free ribosomes were rather evenly dis- ies. tributed in the perikaryon, but were denser in the narrow rim of cytoplasm of spindle-shaped cells than in the Young adult animals more abundant cytoplasm of the larger oval to spheroid Several ultrastructural differences were noted that or pear-shaped ones (Figs. 4, 11). This, in all probabil- distinguished many young adult rat SCN neurons from ity, accounts for dark and pale (light) cells reported by those in the SCN of weanlings. The most obvious others (Suburo and Pellegrino de Iraldi, 1969; Ogata, change was an increased incidence of nuclear mem1986). The cytoplasmic organelles and inclusions brane invagination (Table 2). There was a generalized which appeared frequently and showed a distinct age increase in the percentage of compacta cells having and regional correlation within the SCN were condens- invaginated nuclear membranes in the SCN, from ing vacuoles (vesicles), cytoplasmic (matrix) vacuoles, 36.8% in weanlings to 74.0% in young adults. In addilysosomes and lipofuscin granules. Cytoplasmic nema- tion, there was a greater percentage of extensively intosomes, nuclear vacuoles, and inclusions were seen vaginated cells (41.1%)in the dissipata region in cononly infrequently and hence showed no distinct age or trast to the compacta region (9.8%) (see Figs. 7, 8). In addition to nuclear membrane invagination, some disregional correlation within the nucleus. sipata SCN cells (<lo%)showed degenerative changes Age-related cytological changes in the SCN including a rather striking increase in vacuolation, lyThe most obvious and distinctive age-related cytolog- sosome and lipofuscin development (Fig. 8). Vacuolaical changes in SCN neurons were nuclear membrane tion occurred within the cytoplasmic matrix, in the invagination, organelle and cytoplasmic vacuolation, Golgi apparatus, and to a lesser extent in mitochonlysosomal aggregation and the development of mature dria. It was also rather obvious that free ribosomes lipofuscin granules or L3 pigment granules (Sekhon decreased in density and multivesicular bodies were and Maxwell, 1974). These features were used, collec- more frequent in the vacuolated cells of young adult tively, to define degenerative neurons. On the basis of SCN cells. nuclear membrane invagination, cells were categorized as smooth (no invaginations), mildly invaginated (1-2 Aged animals The most obvious change in 30-36-month-old rats invaginations), and extensively invaginated (3 or more invaginations) (Fig. 4). Age-related structural changes was a significant increase in the number of structurwere also observed in axons and dendrites, however, ally deteriorating cells. These characteristically contained 4 or more mature lipofuscin granules andlor agthese will not be addressed in the present study. gregates of lysosomes combined with cytoplasmic Weanlings vacuolation, mitochondria1 alterations (enlargement, The round to oval nucleus of the SCN neurons of vacuolation, and granulation), Golgi vacuolation and 21-28-day-old rats was surrounded by a predominantly occasionally dilated or vacuolated granular endoplas- SCN, and to a greater extent in its dissipata region, the cells were not as densely distributed as in the ventromedial part of the nucleus (Fig. 1). Another important morphological feature of SCN cells is the characteristic invagination of their nuclei. As will be explained below, this feature varied tremendously depending on the age of the animal and the region of the SCN in which the cell was found. EM ANALYSIS OF RAT SCN 77 Fig. 3. A low power electron micrograph of a cell from the ventrolateral SCN of a 5-mo-old rat which contained a single large dendrite (arrowhead). Note that the area of abundant cytoplasm which contains numerous variously shaped mitochondria (M)and other or- ganelles is more abundant at the site of origin of the process than other nuclear perimeters. The nuclear membrane is extensively invaginated with three relatively deep folds (arrows). mic reticulum. In extreme cases, lipofuscin granules crowded the cytoplasm, and the mitochondria appeared swollen and ruptured. The assumption of varied morphological configurations by mitochondria appeared to be accentuated in aged animals. Both light and dark cells displayed such features and were dispersed throughout the SCN but were more highly concen- trated in the dissipata region (30%)of the SCN than in its compacta region (18%),as is apparent in low power micrographs of the 2 regions (Figs. 9 , l O ) . However, the percentage of nuclear membrane invaginations in the dissipata SCN neurons of old rats were essentially the same as in young adults (Table 2). Also, the number of invaginations in the nucleus was highly corre- 78 W.H. WOODS ET AL. Fig. 4. A higher magnification of the cell in Figure 3 which shows several forms of mitochondria1 constriction and elongation. An “S,” “dumbbell,” and a “C-shaped form are shown at the lower left corner and middle of the cell, respectively (arrows). Free ribosomes (R) are rather evenly distributed in the cell cytoplasm. 79 EM ANALYSIS O F RAT SCN TABLE 2. Comparison of nuclear invaginations in the compacta and dissipata parts of the SCN in weanling, young adult and old rats* Compacta E S M Dissipata E S 22.8 8.6 25.5 t 3.4' * 48.0 t 12.8 58.4 t 3.0 12.8 2.8 16.1 t 1.0' * 7.0 2 2.3 11.2 t 3.1' 32.0 % 5.7 51.0 t 1.5 24.0 t 5.6 37.8 t 3.4' 20.3 t 0.9' 25.9 2.6l * 50.3 t 6.0 64.2 t 1.8 7.7 t 0.9' 9.8* 0.3' 5.0 2.6l 7.5 t 3.6l 33.7 2.9 51.4 & 4.3 27.0 t 5.1' 41.1 t 7.1' 46.3 t 6.9l 63.2 t 1.5' 25.7 t 2.0 35.6 & 1.8' 1.0 t 0.6 1.2 t 0.8 22.3 t 2.4: 34.0 t 2.1 39.7 7.3 58.7 2 5.0' 4.3 t 1.3 7.3 t 3.3 M Aged (n = 4) Mean number of cells t S.E.M. Mean percentages t S.E.M. Young adult (n = 3) Mean number of cells Mean percentages t S.E.M. Weanling(n = 3) Mean number of cells t S.E.M. Mean Dercentanes t S.E.M. * * *S = smooth (no invagination); M = mild (1-2 invaginations); and E = extensive (3 or more invaginations). Significant differences were found for compacta versus dissipata, e.g., weanling compacta smooth mean is compared to the weanling dissipata smooth mean. and indicatePs0.025. lated with cytological degeneration, whether it was found in the dissipata or the compacta SCN. On the average, 56% of the cells in the dissipata with extensively invaginated nuclear membranes exhibited obvious degenerative changes (organelle vacuolation, lysosomes, and lipofuscin granule formation) while only 17% with mildly invaginated nuclear membranes showed such changes in the aged rats. Less than 5%of the SCN neurons with smooth nuclear membranes showed characteristic deterioration regardless of location within the nucleus. We estimate that about twice as many cells with degenerative changes occurred in the middle and caudal SCN than in its rostra1 part. It may be noted that not only did the number of invaginations increase with age but the depth of the invaginations also appeared to increase, often spanning the entire depth of the nucleus (Fig. 11). DISCUSSION The results of light microscopic morphometric analysis showed essentially no difference in total SCN cell number between weanling and young adult animals (12,353 vs. 12,904). There was a slight decrease between these two groups and aged animals (11,7061,but the decrease was not significant. The relative consistency of SCN cell numbers in all ages of rats is in agreement with findings of Peng et al. (1980) and Roozendaal et al. (1987). The total SCN cell count approximates the mean cell counts of 10,823 found by Van den Pol (1980) and 11,650 (males and females combined) found by Guldner (1983). The total SCN volume of 0.041 mm3 in young animals and 0.046 mm3 in aged animals is less than the 0.050 mm3 reported by Guldner (1976), the 0.068 mm3 observed by Van den Pol (1980) and the 0.064 mm3 observed by Guldner (19831, but essentially the same as that found b? Roozendaal et al. (1987) of 0.039 mm3 and 0.045mm in young (7-8 mo) and old (32-33 mo) rats. The discrepancy in these figures could be explained in part by the difficulty one experiences in determining the dorsolateral boundary of the SCN between the compacta and dissipata regions (Van den Pol, 1980). In view of the sparseness of cells in this region, a larger area would not significantly change the total SCN cell number. Two important findings in the present study were a significant decrease in cell number in the dissipata re- gion (P <0.025) and a decrease in density in the compacts region (P <0.025) in young adults and old animals as compared to weanlings (Table 2). These findings will be discussed below in light of our ultrastructural observations. Cells of the SCN have spindle and oval shaped perikarya, from which 1 to 4 processes originate, as determined from thin and Nissl stained sections. The cells contain multiple Golgi bodies, a small amount of mitochondria and granular endoplasmic reticulum and a large spheroidal or oval nucleus. The nucleus contained multiple nucleoli and a varying number and depth of nuclear membrane invaginations. The most frequently observed neurons were small to moderate averaging 57 pm2 in size, spindle-shaped and possessed one to two primary processes. Their nuclei were spindle to ovoid-shaped and had none to several membrane invaginations. These seemed to be most highly concentrated in the ventral and medial parts of the compacta region of SCN. Larger oval-shaped (averaging 75 bm2) cells with more abundant cytoplasm, spheroidal nuclei and 2-4 processes were observed in the dissipata and, to a lesser extent, in the compacta region of the nucleus. In general, similar cellular features were described in young rats by Suburo and Pellegrino de Iraldi (1969)and briefly outlined by Rusak and Zucker (1979). In a detailed study of the SCN utilizing several light microscopic techniques (including Golgi impregnation) and electron microscopy, Van den Pol (1980) extended the list to five cell types. In addition to the characteristic perikarya, morphological features and processes of SCN cells described in the present study, he also described spiny and "curly" bipolar neurons (Van den Pol, 1980) based on dendritic trees of Golgi impregnated materials. In the absence of Golgi stained material, a complete correlation of cell types on this basis cannot be made. However, the three remaining cell types described were monopolar, simple bipolar and multipolar types. In all probability, these cell types observed correspond to those with 1, 2, and 3-4 processes seen in thick and thin sections in this study. The presence of four or more mature lipofuscin granules together with organelle and cytoplasmic vacuolation along with increased nuclear membrane invaginations were major criteria used to define degenerating SCN neurons. Except for nuclear membrane invagina- 80 W.H. WOODS ET AL. Fig. 5. An electron micrograph of neurons in the compacta region of the middle rostrocaudal extent of the SCN of a 28-day-old rat. The neurons are generally spheroidal to oval in shape and have granular endoplasmic reticulum (ger) with a characteristic polar or peripheral distribution. Mitochondria (M) are typically oval, short cylinders or round with no vacuoles and only occasional constrictions. The small Golgi (GI region contains spherical and flattened vesicles often taking a circular form. The nuclear membrane was almost always smooth within this region of weanling rats. tion, these criteria have been traditionally used to describe aged or degenerating features within neuronal somas including those of the hypothalamus (Hasan et al., 1974; Glees et al., 1975; Ogata, 1986).The degree of nuclear membrane invagination (number and depth) of SCN neurons appeared to show a high correlation with the other morphological correlations of aging in these cells. Therefore, we utilized the presence of an extensively invaginated nuclear membrane within SCN cell somas as an additional indicator of cell deterioration. EM ANALYSIS O F RAT SCN 81 Fig. 6. An electron micrograph of neurons taken from the same animal and rostrocaudal level as in Figure 5, but these neurons were taken from the dissipata region of the SCN. Note the three slight invaginations (arrows) in the cell with a complete nucleus and the single fold in the nucleus at bottom right of the picture. The morphology of mitochondria (M), Golgi (G),granular endoplasmic reticulum (per) of these cells are similar to that described for the compacta cells in Figure 5 except that they are more numerous. We have observed this phenomenon in other hypothalamic neurons but not to the same degree (unpublished observations). Matsumoto et al. (1982) also noted this phenomenon in the arcuate nucleus of aged rats, and Glees et al. (1975) noted i t in the hypothalamus of aged monkeys. An increased frequency of invaginated nuclei in the dissipata region of the SCN was observed by Van den Pol (1980). In general, these authors explained increased invagination of hypothalamic neuronal nuclei as a response to a n increased cell volume and also to a decrease in the distance between nuclear membrane and nucleolus. Since many of the nuclear membrane invaginations in this report did not appear to be asso- ciated with nucleoli, this morphological change is probably a response to some other change in the environment of the aging SCN. Although no significant total SCN cell loss was found in this study, cells showing structural features of degeneration occurred throughout the SCN, especially in the lateral part of the nucleus including its ventral half. Whether or not the development of these altered morphological features within SCN neurons is SUEcient to indicate severe physiological deficits within them is debatable. However, changes in the regular circadian pattern of glucose utilization have been shown to occur in middle aged ovariectomized (18-21 82 W.H. WOODS ET AL. Fig. 7.An electron micrograph of neurons in the compacta region with their characteristic spindle-shape and close soma1 apposition are shown from middle SCN of a 8-mo-old rat. The cytoplasm is abundant at one end of the spindle to oval-shaped nucleus. Mitochondria (MI, some of which are swollen and vacuolated, are oval to short cylinders with mainly transverse cristae. The Golgi (G) body primarily consists of flattened and spherical vesicles. Granular endoplasmic reticulum (per) is sparse and peripheral or polar in distribution. The nuclear membrane is generally smooth or mildly invaginated (1or 2 invaginations). E M ANALYSIS OF RAT SCN Fig. 8. An electron micrograph of neurons in the dissipata region of the SCN from the same animal as in Figure 7. Neurons of this region are usually oval to spheroid and have a larger volume and lighter density of cytoplasm than those of the denser compacta part of the nucleus. Although only about one-half of the nuclei of two of these 83 three larger cells is shown, each membrane was extensively invaginated (3-4 invaginations). Note the presence of vacuoles (V),several lysosomes (Ly), and isolated strands of endoplasmic reticulum (ger) and small Golgi bodies ( G )in these cells. 84 W.H. WOODS ET AL. Fig. 9. The four neurons shown in this electron micrograph were taken from the compacta region of the SCN of a 36-mo-old animal. The cells retain their usual spindle shape and are surrounded by a thin rim of dense cytoplasm. No severe signs of degeneration except occasional lipofuscin (Lf) granules or Lysosomes (Ly) are evident in these medial cells. No more than two invaginations (mildly invaginated condition) occur in either of the cells, the Golgi ( G ) contain normal, flattened and small spherical vesicles and mitochondria (M) are the regular spherical or cylindrical (tubular) forms. mo) as compared to young (3-4 mo) rats by Wise et al. (1987,1988). These authors associate their results with a decline in oxidative metabolism which in turn contributes to physiological deficits. The increased prevalence of altered SCN neuronal morphology, especially vacuolated, swollen and ruptured mitochondria, tends to support their concept. Another point that merits consideration is whether the loss of a small population of dissipata cells with a decline in functional efficiency of many of the remaining cells in the lateral division of the SCN could be associated with documented losses of circadian rhythms in laboratory animals (Miles and Dement, 1980; Ingram et al., 1982) and humans (Van Goo1 and Mirmiran, 1986). The SCN is complex in its organization and contains a significant number of immunocy- E M ANALYSIS OF RAT SCN Fig. 10. This electron micrograph was taken from the dissipata region of the SCN of the same animal in Figure 9. Cells at the top and right of the figure show degenerative changes. Note small Golgi bodies (G),vacuolated mitochondria (M), cytoplasmic vacuolation (V) and lipofuscin (Lf) granules. Although the complete nuclei of three cells are not shown, all were either mildly or extensively invaginated. 85 86 W.H. WOODS ET AL. Fig. 1 1 . A high power electron micrograph of an extensively invaginated nuclear membrane within a SCN neuron of the dissipata of a 36-mo-old rat. The membrane has six folds (arrows), one of which is relatively deep. Obvious lipofuscin granules (Lf), free ribosome (R),and a distinct Golgi apparatus (G) are present in the cytoplasm. tochemically identified cells which are involved in a n unusual degree of local circuitry (Van den Pol, 1980; Moore, 1983; Moore and Card, 1985; Van den Pol and Gorcs, 1986; Guy et al., 1987). Due to this extremely varied cell population, it is impossible to state with certainty, the immunocytochemical identity of the SCN neurons which we have shown to be more sensitive to the aging process. However, ultrastructural analysis of the present study revealed that in the aged rat pronounced degenerative changes occurred not only 87 EM ANALYSIS OF RAT SCN in the dissipata, but also in many cells of the ventrolateral compacta division of the SCN. In view of the fact that EM tissue samples from different age groups were frequently processed simultaneously and degenerative changes were prevalent only in aged animals, it is reasonable to conclude that such changes are age related. As a rule, the cells which displayed degenerative changes were larger than those of the ventromedial part of the nucleus, spheroidal in shape and contained large extensively invaginated nuclei. These general morphological features are essentially identical to those described for vasoactive intestinal peptide (VIP) containing neurons which have been shown to be highly concentrated in the ventrolateral SCN (Card et al., 1981; Maegawa, 1987; Chee et al., 1988). Vasoactive intestinal peptide neurons along with immunoreactive vasopressin, somatostatin and glutamic acid decarboxylase make up 50% or more of SCN neurons (Card and Moore, 1984) but only the VIP neurons predominate in the ventrolateral region of the nucleus (Card and Moore, 1982; Maegawa, 1987). The heavy concentration of VIP cells laterally in the SCN has lead to their being designated as the most likely positioned neurons to receive and integrate visual afferent projections (Card and Moore, 1982). Further, VIP cells of the SCN have been shown by Maegawa (1987)to be among the first to appear, developmentally, and along with vasopressin cells show a day-night rhythm of appearance. Hence, he indicated that VIP cells may function as a spontaneous oscillator displaying a light-dark rhythm. Finally, a few comments should be made as noted earlier in this discussion relative to the high degree of degenerative ultrastructural alterations and cell loss observed in the dissipata region of the aging rat SCN. The significant decrease in cell number in this region may be indicated by the predominance of a diffusely dispersed single neuronal cell type localized in this area following loss of the more susceptible cell type. As pointed out above, morphological features displayed by the majority of cells in this region are identical to those described for VIP cells by others. Specific cell-type loss has been reported for vasopressin (Roozendaal et al., 1987) and VIP cells (Chee et al., 1988). The latter study suggested “selective cell death” to explain their findings of a decreased vasopressin and VIP cell loss in the SCN of aged rats while the total cell number remained stable. Because the SCN dissipata is populated by fewer cell types than the compacta, a single cell type, e.g., VIP cells, that undergoes early degenerative changes would result in a significantly decreased cell number in this region and would be more readily detected than a decrease for the total SCN. Additional immunocytochemical studies a t the ultrastructural level and, possibly, electrophysiological studies of specific regional cell types in aged animals must be undertaken in order to obtain a definitive correlation between cell loss in the aging SCN and the loss or alteration of normal circadian rhythms. A move in this direction is indicated by the recent report of Miller and Fuller (1992), who identified a subpopulation of isoperiodic firing of rat SCN neurons in vivo. These authors also suggested that most, if not all, of these neurons were VIP cells. ACKNOWLEDGMENTS The authors would like to express their extreme gratitude to Mrs. Patricia Marks for her help with statistical data, Ms. Janice Gray for her clerical help, and Mrs. Evelyn Payne-Allen for her technical assistance. We are equally grateful for the care and time devoted to reading the manuscript by Dr. Robert Skinner, Dept. of Anatomy, University of Arkansas for Medical Sciences. This research was supported by National Institute of Health grant S14GM02716. LITERATURE CITED Card, J.P., and R.Y. Moore 1982 Ventral lateral geniculate nucleus efferents to the rat suprachiasmatic nucleus exhibit avian pancreatic polypeptide-like immunoreactivity. J. Comp. Neurol., 206t390-396. Card, J.P., and R.Y. Moore 1984 The suprachiasmatic nucleus of the golden hamster: immunohistochemical analysis of cell and fiber distribution. Neuroscience, 13t415-431. Card, J.P., N. Brecha, H.J. Karten, and R.Y. Moore 1981 Immunocytochemical localization of vasoactive intestinal polypeptide-containing cells and processes in the suprachiasmatic nucleus of the r a t light and electron microscopic analysis. J . Neurosci., 1~12891303. Chee, C.A., B. Roozendaal, D.F. Swaab, E. Goudsmit, and M. Mirmiran 1988 Vasoactive intestinal poly-peptide neuron changes in the senile rat suprachiasmatic nucleus. Neurobiol. Aging, 9t307312. Drucker-Colin, R., R. Aguilar-Roblero, F. Garcia-Hernandez, F. Fernandez-Cancino, and F.B. Rattoni 1984 Fetal suprachiasmatic nucleus transplants: diuranl rhythm recovery of lesion rats. Brain Res., 311t353-357. Eastman, C.I., R.E. Mistlberger, and A. Rechtschaffen 1984 Suprachiasmatic nuclei lesions eliminate circadian temperature and sleep rhythms in the rat. Physiol. Behav., 32t357-368. Fliers, E., and D.F. Swaab 1986 Neuropeptide changes in aging and Alzheimer’s diseases. In: Aging of the Brain and Alzheimer’s Disease, Progress in Brain Research, Vol. 70. D.F. Swaab, E. Fliers, M. Mirmiran, W. A. Van Gool, and F. Van Haaren eds. Elsevier, Amsterdam, pp. 141-152. Glees, P., P.E. Spoerri, and E. El-Ghazzawi 1975 An ultrastructural study of hypothalamic neurons in monkeys of different ages with special reference to age related lipofuscin. J . Hirnforsch., 16t379394. Guldner, F.H. 1976 Synaptology of the rat suprachiasmatic nucleus. Cell Tissue Res., 165t509-544. Guldner, F.H. 1983 Number of neurons and astroglial cells in the suprachiasmatic nucleus of male and female rats. Exp. Brain Res., 50t373-376. Guy, J., 0. Bosler, G. Dusticier, G. Pelletier, and A. Calas 1987 Morphological correlates of serotonin-neuropeptide Y interactions in the rat suprachiasmatic nucleus: combined radioautographic and immunocytochemical data. Cell Tissue Res., 25Ot657-662. Hasan, M., P. Glees, and E. El-Ghazzawi 1974 Age associated changes in the hypothalamus of the guinea pig: effect of dimethylaminoethyl p-chlorophenoxyacetate. An electron microscopic and histochemical study. Exp. Gerontol., 9.153-159. Horikawa, K., N. Kinjo, L.C. Stanley, and E.W. Powell 1988 Topographic organization and collateralization of the projections of the anterior and laterodorsal thalamic nuclei to cingulate areas 24 and 29 in the rat. Neurosci. Res., 6.31-44. Ingram, D.K., E.D. London, and M.A. Reynolds 1982 Circadian rhythmicity and sleep: effects of aging in laboratory animals. Neurobiol. Aging, 3t287-297. K a a a , M.S., P.J. Marangos, and R.Y. Moore 1985 Suprachiasmatic nucleus ablation abolishes circadian rhythms in rat brain neurotransmitter receptors. Brain Res., 327r344-347. Kafka, M.S., A. WirzJustice, and D. Naber 1981 Circadian and seasonal rhythms in alpha and beta adrenergic receptors in the rat brain. Brain Res., 207.409-419. Konig, J.F.R., and R. Klippell963 The Rat Brain: A Stereotaxic Atlas of the Forebrain and Lower Parts of the Brain Stem. Williams and Wilkins Co., Baltimore, MD. Kongismark, B.W., U.P. Kalyanaraman, P. Corey, and E.A. Murphy 1969 An evaluation of techniques in neuronal population estimates: the sixth nerve nucleus. John Hopkins Med. J., 125t146158. 88 W.H. WOODS ET AL. Lehman, M.N., R. Silver, W.R. Gladstone, R.M. Kahn, M. Gibson, and E. Bittman 1987 Circadian rhythmicity restored by neural transplant. immunocytochemical characterization of the graf€ and its integration with the host brain. J. Neurosci., 7t1626-1638. Maegawa, M. 1987 Development of architechtonics of the suprachiasmatic nucleus of rats. Immunohistochemical study. Tokushima J. Exp. Med., 34t29-43. Martin, R.R.. G.F. Russell. F.C. Davis. and M. Menaker 1990 Transplanted suprachiasmatic nucleus' determines circadian period. Science, 24 7r975-978. Matsumoto, A,, R. Okada, and Y. Arai 1982 Synaptic changes in the hypothalamic arcuate nucleus of old male rats. Exp. Neurol., 78: 583-590. Miles, L.E., and W.C. Dement 1980 Sleep and aging. Sleep, 3r119220. Miller, J.D., and C.A. Fuller 1992 Isoperiodic neuronal activity in suprachiasmatic nucleus of the rat. Am. J. Physiol., 263tR51R58. Moore, R.J., and V.B. Eichler 1972 Loss of a circadian adrenal corticosterone rhythm following suprachiasmatic lesions in the rat. Brain Res., 42t201-206. Moore, R.Y. 1983 Organization and function of a central nervous system oscillator: the suprachiasmatic hypothalamic nucleus. Fed. Proc., 42r2783-2789. Moore, R.Y., and J.P. Card 1985 Visual pathways and the entrainment of circadian rhythms. Ann. N.Y. Acad. Sci., 453t123-133. Ogata, R. 1986 Age-related morphological changes in the rat suprachiasmatic nucleus correlated with age-related changes in circadian rhythmicity of gross activity. Fukuoka Med. J., 77r437-457. Peng, M.T., M.J. Jiang, and H.K. Hsu 1980 Changes in running-wheel activity, eating and drinking and their daylnight distributions throughout the life span of the rat. J . Gerontol., 35r339-347. Powell, E.W., F. Halberg, J.N. Pasley, W. Lubanovic, B. Rockway, and L.E. Scheving 1979 Suprachiasmatic modification of circadian group rhythm of intraperitoneal temperature in inbred fisher rats. In: The Pineal Gland. R.J. Reiter and R.J. Wurtman, eds. Springer-Verlag, Wein, J. Neurotrans. Sup., 113. Powell, E.W., J.N. Pasley, L.E. Scheving, and F. Halberg 1980 Amplitude-reduction and acrophase-advance of circadian mitotic rhythm in corneal epithelium of mice with bilaterally lesioned nuclei. Anat. Rec., 197r227-281. Raisman, and K. Brown-Grant 1977 The suprachiasmatic syndrome: endocrine and behavioral abnormalities following lesions of the suprachiasmatic nuclei in the female rat. Proc. R. SOC.Lond. [Biol.], 198r297-314. Ralph, R., R.G. Foster, F.C. Davis, and M. Menaker 1990 Transplanted suprachiasmatic nucleus determines circadian period. Science, 247:975-978. Roozendaal, B., W.A. Van Gool, D.F. Swaab, J.E. Hoogendijk, and M. Mirmiran 1987 Changes in vasopressin cells of the rat suprachiasmatic nucleus with aging. Brain Res., 409r259-264. Rusak, B., and I. Zucker 1979 Neural regulation of circadian rhythms. Physiol. Rev., 59t449-526. Sekhon, S.S., and D.S. Maxwell 1974 Ultrastructural changes in neurons of the spinal anterior horn of ageing mice with particular reference to the accumulation of lipofuscin pigment. J. Neurocytol., 3r59-72. Stephan, F.K., and I. Zucker 1972 Circadian rhythms in drinking behavior and locomotion activity are eliminated by hypothalamic lesions. Proc. Natl. Acad. Sci. U.S.A., 54r1521-1527. Suburo, A.M., and A. Pellegrino de Iraldi 1969 An ultrastructural study of the rats suprachiasmatic nucleus. J. Anat., 105t439446. Tobler, I., A.A. Borbely, and G. Groos 1983 The effect of sleep deprivation on sleep in rats with suprachiasmatic lesions. Neurosci. Lett., 41r109-113. Van den Pol, A.N., 1980 The hypothalamic suprachiasmatic nucleus of the rat: intrinsic anatomy. J. Comp. Neurol., 191t661-702. Van den Pol, A.N., and T. Gorcs 1986 Synaptic relationships between neurons containing vasopressin, gastrin-releasing peptide, vasoactive intestinal polypeptide and glutamate decarboxylase immunoreactivity in the suprachiasmatic nucleus: dual ultrastructural immunocytochemistry with gold-substituted silver peroxidase. J. Comp. Neurol., 252t507-521. Van Gool, W.A., and M. Mirmiran 1986 Aging and circadian rhythms. Prog. Brain Res., 70t255-277. Wise, P.M., I.R. Cohen, N.G. Weiland, and E.D. London 1988 Aging alters the circadian rhythm of glucose utilization in the suprachiasmatic nucleus. Proc. Natl. Acad. Sci. U.S.A., 85t5305-5309. Wise, P.M., R.C. Walovitch, I.R. Cohen, N.G. Weiland, and E.D. London 1987 Diurnal rhythmicity and hypothalamic deficits in glucose utilization in aged ovariectomized rats. J . Neurosci., 7t34693473. Yamamoto, H., K. Nagai, and H. Nakagawa 1984 Additional evidence that the suprachiasmatic nucleus is the center for regulation of insulin secretion and glucose homeostasis. Brain Res., 304t237241.