&win Research
Bulletin,
Vol. 4, pp. 35-41. Printed
in the U.S.A.
A Note on the Distribution of Glial Cells in the Molecular Layer of the Dentate Gyrus’ K. KISHI,’
B. B. STANFIELD
AND W. M. COWAN
Department of Anatomy and Neurobiology Washington University School of Medicine, St. Louis,
(Received
6 September
MO
1978)
KISHI, K., B. B. STANFIELD AND W. M. COWAN. A note on the distribution of g/id cells in the moleculur lrryer of the dentote gww. BRAIN RES. BULL. 4(l) 35-41, 1979.-The distribution of the perikarya of astrocytes and other glial cells in the molecular layer of the dentate gyrus has been studied in gold chloride-sublimate preparations of rats and of normal and reeler mice, and in plastic embedded material from young adult rats. Contrary to previous reports (Rose et (11. [71),we have found no evidence for a distinct “line” or “band” of astrocytic cell bodies along the interface between the zones of termination of the entorhinal and hippocampal tierents to the dentate gyrus. Indeed, apart from a conspicuous accumulation of astrocytes immediately beneath the pial surface and hippocampal fissure, the distribution of glial cells across the extent of the molecular layer appears to be more or less random. In view of this it is difficult to ascribe a critical role to the astrocytes in either determining the normal distribution of afferent fibers to the dentate gyrus or in promoting their re-organization following its partial deafferentation. Glial cells
Astrocytes
Dentate gyrus
Molecular layer
WHILE studying the reaction of astrocytes in the molecular layer of the rat dentate gyrus to the selective removal of the afferents that arise within the entorhinal area Rose, Lynch and Cotman [7], noted that in many (but not all) sections from normal brains, the distribution of astrocytes stainable with the Cajal [4] gold chloride-sublimate method, was nonuniform across the thickness of the molecular layer. In sections prepared in this way, they claimed to observe a distinct “line” of astrocytic somata just beyond the proximal onefourth of the molecular layer (proximal, that is with respect to the bodies of the granule cells in the underlying stratum grurrulosum). Since this “line” appeared to coincide rather precisely with the boundary between the afferents from the medial entorhinal area (in the so-called entorhinul zone) and those from the hippocampus of the two sides (in the so-called hippoccrmpul zone that contains the ipsilateral associational and commissural afferents) they implied that the distribution of astrocytes at this interface was not fortuitous, but in some way related to the selective laminar arrangement of the afferents in the entorhinal and hippocampal zones of the molecular layer. Furthermore, they claimed that after removal of the entorhinal afferents there was a tendency for the astrocytes within the non-detierented hippocampal zones to migrate out into the denervated area, and for those along the “entorhinal-hippocampal” interface to move peripherally as (I unit. Since these events preceded the earliest indication of axonal sprouting of the afferents in the hippocampal zone they further suggested that the astrocytes may be critically implicated in the sprouting process and in the redistribution
Plasticity
of axon terminals that is seen in this region after the various types of deafferentation. In view of the important implications that these observations may have for the whole phenomenon of central neural plasticity, and for the localization of afferent fiber systems in cortical areas, we re-examined our collection of normal rat brains, and of brains from animals with various experimental manipulations of the hippocampal formation. Since this provided no clear confirmation of the pattern reported by Rose et a/. [7], we prepared a number of rat brains for staining with the Cajal gold chloride-sublimate method which they had used, and a number of 1 pm thick sections from plastic embedded blocks of the rat dentate gyrus. In addition we have examined the distribution of astrocytes in gold chloride-sublimate stained preparations of the dentate gyrus in normal and reeler mice. This last group of brains is of particular interest because as we have shown elsewhere, in the reeler mouse the afferents from the entorhinal cortex occupy the entire thickness of the molecular layer and the commissural and associational afferents are distributed, for the most part, among the granule cells and throughout the hilar region of the area dentata [8]. METHOD
From a number of rat brains stained for astrocytes by the gold chloride-sublimate method (following the procedure outlined by Rose et al. [7], as closely as possible) two brains of adult albino rats, were selected for special study; one had
‘This work was supported in part by grant NS10943 from the National Institutes of Health. ‘On leave of absence from Tokyo Medical and Dental University, Bunkyo-Ku, Tokyo, Japan.
Copyright 0 1979 ANKHO International
Inc.-0361-9230/79/010035-07$01.20/O
36
KISHI, STANFIELD
FIG. 1. A low power photomicrograph of a gold chloride-sublimate preparation of the dentate gyrus of a young adult rat. In fields such as this it appears that there may be a distinct line of astrocytic somata at the junction between the inner one-fourth and the outer three-fourths of the molecular layer (at the level marked by the arrows). Scale: 100 pm.
FIG. 2. A comparable low power photomicrograph to that shown in Fig. 1, but from an adjoining section. In this field there is no indication of a “line” of astrocytic somata near the junction between the “hippocampal” and “entorhinal” zones of the molecular layer of the dentate gyrus. Scale: 100 pm.
AND COWAN
GLIAL CELLS IN THE DENTATE
CYRUS
been sectioned (at 40 pm) in the standard frontal plane and the other horizontally. Seven normal (or heterozygous +/?) and seven littermate reeler (rlirl) mouse brains were similarly prepared, the sections in these cases being cut at 30-40 pm. One micron thick plastic embedded sections were also cut from blocks of tissue taken from the hippocampus of a young adult rat which had been perfused with a mixture of 2% paraformaldehyde and 1.25% glutaraldehyde in a 0.1 M phosphate buffer. In each case sections were taken from near the middle of the septal, intermediate and temporal thirds of the hippocampal formation, and included the entire cross-sectional area of the dentate gyrus. The 1 pm sections were stained with toluidine blue and coverslipped in the usual way. For the illustrations the outlines of selected sections were traced with the aid of an Olympus microscope and drawing tube at a magni~cation of 400X, and the location of each impregnated astrocyte (in the case of the gold chloridesublimate material) and of each glial cell and superficially located neuron (in the case of the 1 pm plastic sections) carefully
marked
on the tracing. RESULTS
Figure 1 is a low-power photomi~rograph of a Cajal gold chloride-subliminate preparation comparable to those used by Rose et cd. [7]. In the field illustrated the density of astrocyte cell bodies in the molecular zone immediately adjacent to the layer of granule cells (the strcrtum grcmulosum) is relatively low, and appears to increase significantly at a distance of about 50 pm from the outer aspect of the .~tr~~tu~~
37 The abrupt increase in the number of astrocytic perikarya at this level clearly gives one the impression that it is marked by a distinct line of glial cells, and this impression is reinforced by the fact that most of the cells that are impregnated have ascending processes that extend superficially into the outer two-thirds or so of the molecular layer. Undoubtedly, it was fields like this, and the finding in some cases that the level at which the cell density increased corresponded more or less to the boundary between the zone in the molecular layer that receives afferents from the medial part of the entorhinal area supe~cially, and that receiving hippocampal afferents (from the same and opposite sides of the brain) internally, that led Rose and his colleagues to suggest that there is in fact a line of astrocytes at the interface between the two zones. However, as Fig. 2. shows, in other fields, one can see no such “tine” of astrocytic perikarya, and indeed if one examines the illustratjons in Rose ef crl.‘s paper [7], it is evident that in many of their preparations there was also no distinct border between the relatively astrocyte-poor inner part of the molecular layer and the astrocyte-rich outer two-thirds. It was this observation that first led us to question the conclusion that astrocytes are preferentially “lined-up” along the interface between the hippocampal and entorhinal zones of the molecular layer and to carry out the systematic quantitative analysis to be described below. But before presenting the quantitative data, it is worth pointing out that the distribution of the astrocytic perikarya, in the dentate gyrus, not only varies from section to section (as noted by Rose et rd. [7]), but even within one section one can find regions in which the density of glia is fairly uniform across the thick-
grunulostrm.
FIG. 3. Tracings to show the distribution of astrocytic somata over the entire cross sectional area of the dentate gyrus at each of two levels to the middle section of the central and caudal thirds of the gyrus) prepared from a well-impregnated gold chloride-sublimate
(corresponding
series of sections. Each filled-in circle marks the location of the perikaryon of an impregnated astrocyte.
38
KISHI,
ASTROCYTES (GOLD
SUBLIMATE)
4 2
20 PERCENTAGE
40
60
OF MOLECULAR
80
100
LAYER
FIG. 4. A histogram to show the relative distribution of the astrocytic somata in three representative sections from the rostral, middle and caudal thirds of the dentate gyrus. In constructing this histogram the position of each cell was expressed as a percentage of the perpendicular distance from the outer aspect of the .\t~t~/n ,~rc~~~~tb~s~~r~ to the pial surface or the hippocampal fissure, and the data from all three sections were pooled and grouped into 10 percentile “bins.”
ness of the molecular
layer immediately adjacent to others in which there is a much lower cell density in the inner part of the layer. Because of this one needs to examine the entire cross-sectional area of the gyrus at each of several levels to get a more accurate impression of the distribution of the glial cell somata. To do this we have systematically plotted the location of the somata of all astrocytes seen in three well-impregnated sections through the dentate gyrus of a young adult rat, taken from near the middle of the septal one-third of the hippocampus, near the center of the septo-temporal extent of the gyrus, and from the middle of its temporal third, respectively. Two of these three plots are shown schematically in Fig. 3: the third has been omitted for the sake of space. It is evident from these drawings that there are no marked heterogeneities in the distribution of the astrocytes across the thickness of the molecular layer although, as we shall point out later, in plastic embedded material it is evident that there is a somewhat higher density of cells immediately beneath the pia mater or the hippocampal fissure. As Rose P/ I//. [7l have emphasized that the highest proportion of as-
STANFIELD
AND
COWAN
trocytic somata occurred along a line 2755 of the distance from the outer edge of the .SVU~U/U ,~~~rrtir/osrl~l to the pial surface, we have measured the location of each cell body indicated in Fig. 3 (and in the third, more rostra1 section of the series) as a fraction of its distance along a vertical line perpendicular to the outer aspect of the strrrfrr/?r ,~rtr~~~losro~. When all these data were pooled, we could construct the histograms shown in Fig. 4. These confirm that in general the density of astrocytic perikarya is low in the inner one-third of the molecular layer, and that it rises appreciably towards the middle of this layer. In addition, they suggest that there is a somewhat higher proportion of astrocytes located between 30 and 4OZ of the way from the strvrtun? ~yw~~u/o.strm to the superficial limit of the molecular layer, although peaks like that shown in Fig. 4 are not always present. Furthermore, the “peaks” are not very prominent amounting to no more than a 3-5s increase in the number of astrocytic somata over that found in the deeper, or more superficial parts of the molecular layer. Since the gold chloride-sublimate method tends to be rather capricious, so that even different parts of the same section may show differing numbers of impregnated glial cells, we have examined several representative 1 pm thick plastic sections stained with toluidine blue at high pH. These have two advantages: first, they display t//l the glial cells present in the tissue. so that one can be confident that the sample is not biased due to inadequate or variable staining; and second, the excellent resolution that such sections provide, enables one to clearly distinguish the various classes of glial cells. We have found that when using the morphological criteria suggested by Ling ct (I/. [6] all but a relatively small percentage of the glial cells in the molecular layer of the dentate gyrus can be assigned with confidence to one or another class, and different observers usually have little difficulty in assigning cells to the same categories. The results of a quantitative analysis of three such sections, taken again from the mid-region of the septal, middle and temporal thirds of the hippocampal formation, are shown in Fig. 5, and the actual distribution of the somata of the astrocytes seen in one of these sections is shown in Fig. 6. Merely by inspection it is difficult to perceive any consistent pattern in the distribution of the astrocytes in these preparations, except that they make it clear that a very high percentage of the astrocytes is aggregated immediately beneath the pial surface and the hippocampal fissure. It is also evident that when all the data are pooled in this way. there is a slight increase in the density of astrocytic perikarya in the region just beyond the normal zone of termination of the commissural and associational projections to the molecular layer of the dentate gyrua. between 30 and 4OV of the distance from the outer border of the .strutti/ll ~qrtr~~r~lo.\c~ur. From a comparison of the relative numbers of stained astroctyes in the thin plastic sections with the numbers of cells impregnated by the gold chloride-sublimate method, we have the distinct impression that the latter technique does not impregnate all the astrocytes in the tissue. We have already commented upon the variability of the impregnations from section to section, and even within different parts of the same section, and the failure of the gold chloride-sublimate method to impregnate most of the astrocytes located near the superficial limit of the molecular layer. But it is likely also that deep within the molecular layer only a proportion of the astrocytes is stained by the gold chloride-sublimate method. Since in the same preparations one can readily identify the other classes of glial cell\. we have also analyzed the
GLIAL CELLS IN THE DENTATE GYRUS
39
r
ASTROCYTES ( PLASTIC
SECTIONS
242220.
f
ALL
GLIAL
(PLASTIC
CELLS SECTIONS)
18. 16. 1412108-
r 1
6-
,..,..,.,
20
40
PERCENTAGE
60
20
86
OF
MOLECULAR
40
60
80
100
LAYER
FIG. 5. These two histograms show the relative distributions of the astrocytic somata and of the bodies of all the glial cells seen in a series of three representative 1 pm plastic embedded sections through the dentate gyrus. The histograms were prepared in the same way as that shown in Fig. 4, and are based on data derived from reconst~ctions comparable to that shown in Fig. 6. Note the promjnent peak in the 90-100% bins in both histograms. which reflect the high density of subpial astrocytes.
distribution of the three major types--oligodendrocytes (including both the light and dark varieties recognized by Ling CI trl. [6]), microglia and astrocytes. The dist~bution of these various glia in the molecular layer at one representative level is shown in Fig. 6. Again these other cell types seem not to be preferentially localized within the dentate gyrus and the only noteworthy point is that unlike the astrocytes, they show no specific tendency to be aggregated at the outer border of the molecular layer. It is perhaps also worth pointing out that there are quite a few neurons scattered throughout the molecular layer (their distribution is indicated by the symbol * in Fig. 6); these no doubt correspond to the various cell types recognized by Cajal [2,3] and others in Golgiimpregnated preparations. The absence of a distinct band or line of astrocytes at the junction between the “hippocampa1” and “entorhinal” zones of the molecular layer of the rat, is also evident in our gold chloride-sublimate preparations of normal and reeler mice as we have reported elsewhere [9]. Although we have not subjected these preparations to the same detailed quantitative analyses, it is evident from plots of the distribution of the somata of the impregnated astroctyes that although the cell density in the hippo~ampal zone (in normal mice) is low and one can occasionally see regions in which there is a fairly abrupt increase in the number of astrocytes close to the junctional region between the entorhinal and hippocampal afferents (which in the mouse occurs along a line about 20-2570 of the distance between the struturn grcltrul~sum and the pial surface 181)there are other regions in which there is
no such change in the glial cell density. In this respect our mouse material is comparable to that in the rat, and for this reason, the observations in the reeler mouse assume an added significance. As we have indicated above, in the reeler mouse the afferents from the medial and lateral parts of the entorhinal area to the dentate gyrus are laminated as in normal mice, but their combined zone of termination extends to the superficial border of the .~t~~~t~4tn ~r~i~~lil~sI~rn. The commissural and ipsilateral associational connections in the reeler are generally absent from the molecular layer, and are confined, for the most part, to the hilar region and the rather poorly-defined struturn gyunuhum [8,10]. One might expect, if there were a critical relationship between the distribution of astrocytes and the location of the interface between the entorhinal and hippocampal afferents, that in the reeler there should be an identifiable band of astrocytes along the outer aspect of the stratum grrmulosum. In fact no such aggregation of astrocytes is evident in our gold chloride-sublimate preparations of the reeler dentate gyrus, and within the somewhat thinned molecular layer, the only indication of an increased glial cell density is that found toward the upper limit of this layer. DISCUSSION
The major conclusion to be drawn from our observations is that while the density of astrocytes (and of glial cells in general) may be relatively low within the zone of termination of the commissural and associational connections to the den-
KISHI, STANFIELD
l
AND COWAN
ASTROCYTES
0 ~LI~O~ENDK~C
YT
A MICROGLIA * NEURONS ~NIOENTIFIED
FIG. 6. A plot to show the distribution of the astrocytic somata, and the location of the other &al cells and the occasional neurons seen in the molecular layer of the dentate gyrus, in a representative 1 pm plastic embedded section ihrough the rostra1 third of the gyrus.
tate gyrus, there is no clear and consistent line or band of astrocytes along the junctional border between this zone and that in which the afferents from the medial part of the entorhinal area terminate. In some sections there is a slight increase in the density of astrocytic perikarya just beyond the junctional region-between 30 and 40% of the distance from the outer edge of the srlat~m gr~uwk~src~~~to the top of the molecular layer, but this is not constant, and by itself hardly impelling enough to support the hypothesis that “the astroglia represent some type of limiting barrier for the fibers which innervate the inner molecular layer and [that] the extent of their outward migration determines the extent of axon sprouting by these fibers” as Rose et trl. [7] have proposed. Moreover, in the reeler mouse, in which the entire thickness of the molecular layer is occupied by the entorhinal afferents, there is no evidence of a comparable line of astrocytic perikarya along the outer edge of the rather poorly-defined strtrtrrrn ~~44t744/4~~s14t~t1 although this marks the junction be-
tween the commissural and associational afferents and those from the medial entorhinal area [8]. In the brains of normal rats and mice impregnated by Cajal’s gold chloride-sublimate method, one can occasionally see short regions of the molecular layer in which the astrocytes appear to be lined up near the border between entorhinal and the hippocampal a&rents, and this appearance is accentuated by the fact that most of the impregnated astrocytic processes ascend superficially from the perikarya. But such areas are too variable and too infrequent to warrant the assumption that the arrangement of the astrocytes is in some way responsible for the sharp segregation between the two classes of afferents. This is not to say that the glial cells in general, and the astrocytes in particular, play no role in the development or maintenance of the molecular layer; clearly their presence in the layer is important, but it seems doubtful that they could play such a critical morphogenetic role as Rose et ctf. ]7] have implied. The fact that in normal
GLIAL CELLS IN THE DENTATE
GYRUS
41
animals there is no marked or constant accumulation of astrocytes in the molecular layer along the border between the entorhinal and hippocampal afferents, makes it difficult to assess the claim that after selective removal of the entorhinal afferents the astrocytes along or near this band advance outwards in a line, in anticipation of the sprouting of the hippocampal afferents and their spread into the denervated “entorhinal zone” [7]. We are inclined to think that the apparent prominence of astrocytes near the interface between the hippocampal and entorhinal afferents after lesions of the entorhinal area or perforant tract, is associated with the general glial response to the degeneration of the terminals of the entorhinal fibers. In support of this view is the observation that the glial response occurs within 48 hr after the placement of the lesion, which is some time before the earliest signs of collateral sprouting from the hippocampal zone, but coincides rather precisely with the onset of terminal degeneration in the entorhinal zone. Since it is well known that astrocytes undergo a vigorous and prompt reaction to axonal degeneration in this region [ 1,5], and that they participate actively in the engulfment and removal of the degenerating axon terminals, it is perhaps not altogether unexpected that there should be some increase in either the number or the prominence of these cells in the area of axonal degeneration. And
to the degree that they are responsible for the removal of the degenerating axon terminals and the freeing up of the synaptic sites for subsequent re-occupation by collateral sprouts from neighboring axons, they may well be thought of as participating in the axonal re-organization that occurs after lesions of the entorhinal area. Furthermore, if during development there is some degree of overlap between the entorhinal and hippocampal afferents which is subsequently eliminated (perhaps as the result of some competitive interaction between the two classes of afferents) it would not be surprising if this too were marked by the persistence, in some regions of the molecular layer, of a number of astrocytes. But this is entirely speculative, and it must be admitted that at present we know all too little about the factors which determine either the number or the distribution of any type of glial or supporting cell in the central or peripheral nervous system.
ACKNOWLEDGEMENTS
We should like to thank Mr. Mark Connelly for his assistance in the preparation of the plastic embedded material, Mr. Marc Davis for help with the photography and Ms. Sharon Musgrove for secretarial help.
REFERENCES 1. Alksne, J. F., T. W. Blackstad, F. Walbergand L. E. White, Jr. Electron microscopy of axon degeneration: a valuable tool in experimental neuroanatomy. Ergehn. Antrf. Gesch. 39: 1-31, 1%6. 2. Cajal, S. Ramon y. Estructura de1 asta de Ammon. Anal. Sot. Esp. Histol. Nut. Madrid 22: 53-114, 1893. 3. Cajal, S. Ramon y. Histologic du Sysrkme Nerveux de I’Homme et des VertPhrPs, T.2. Paris: A. Maloine, 1911, 993 pp. 4. Cajal, S. Ramon y. Contribution al conocimiento de la neuroglia de1 cerebra humano. Truh. Lab. Itnwsr. Sol. Univ. Madrid 11: 255-315, 1913. 5. Laatsch, R. H. and W. M. Cowan. Electron microscopic studies of the dentate gyrus of the rat. II. Degeneration of commissural afferents. J. camp. Neural. 130: 241-262, 1967. 6. Ling, E. A., J. A. Paterson, A. Privat, S. Mori and C. P. LeBlond. Investigation of glial cells in semithin sections. I. Identification of glial cells in the brain of young rats. J. camp. Neural. 149: 43-72. 1973.
7. Rose, G., G. Lynch and C. W. Cotman. Hypertrophy and redistribution of astrocytes in the deafferented dentate gyrus. Brain Res. Bull. 1: 87-92, 1976. 8. Stanfield, B. B., V. S. Caviness, Jr. and W. M. Cowan. The organization of certain afferents to the hippocampus and dentate gymrusin normal and reeler mice. J. co&. Neuiol. In press. 9. Stantield, B. B. and W. M. Cowan. The mornholoev of the hippocampus and dentate gyrus in normal and >eele;‘mice. J. camp. Neural. In press. 10. Stirling, R. V. and T. V. P. Bliss. Observations on the commissural projection to the dentate gyrus in the reeler mutant mouse. Brrrin Res.
150: 447-465, 1978.
Bulletin,
Vol. 4, pp. 35-41. Printed
in the U.S.A.
A Note on the Distribution of Glial Cells in the Molecular Layer of the Dentate Gyrus’ K. KISHI,’
B. B. STANFIELD
AND W. M. COWAN
Department of Anatomy and Neurobiology Washington University School of Medicine, St. Louis,
(Received
6 September
MO
1978)
KISHI, K., B. B. STANFIELD AND W. M. COWAN. A note on the distribution of g/id cells in the moleculur lrryer of the dentote gww. BRAIN RES. BULL. 4(l) 35-41, 1979.-The distribution of the perikarya of astrocytes and other glial cells in the molecular layer of the dentate gyrus has been studied in gold chloride-sublimate preparations of rats and of normal and reeler mice, and in plastic embedded material from young adult rats. Contrary to previous reports (Rose et (11. [71),we have found no evidence for a distinct “line” or “band” of astrocytic cell bodies along the interface between the zones of termination of the entorhinal and hippocampal tierents to the dentate gyrus. Indeed, apart from a conspicuous accumulation of astrocytes immediately beneath the pial surface and hippocampal fissure, the distribution of glial cells across the extent of the molecular layer appears to be more or less random. In view of this it is difficult to ascribe a critical role to the astrocytes in either determining the normal distribution of afferent fibers to the dentate gyrus or in promoting their re-organization following its partial deafferentation. Glial cells
Astrocytes
Dentate gyrus
Molecular layer
WHILE studying the reaction of astrocytes in the molecular layer of the rat dentate gyrus to the selective removal of the afferents that arise within the entorhinal area Rose, Lynch and Cotman [7], noted that in many (but not all) sections from normal brains, the distribution of astrocytes stainable with the Cajal [4] gold chloride-sublimate method, was nonuniform across the thickness of the molecular layer. In sections prepared in this way, they claimed to observe a distinct “line” of astrocytic somata just beyond the proximal onefourth of the molecular layer (proximal, that is with respect to the bodies of the granule cells in the underlying stratum grurrulosum). Since this “line” appeared to coincide rather precisely with the boundary between the afferents from the medial entorhinal area (in the so-called entorhinul zone) and those from the hippocampus of the two sides (in the so-called hippoccrmpul zone that contains the ipsilateral associational and commissural afferents) they implied that the distribution of astrocytes at this interface was not fortuitous, but in some way related to the selective laminar arrangement of the afferents in the entorhinal and hippocampal zones of the molecular layer. Furthermore, they claimed that after removal of the entorhinal afferents there was a tendency for the astrocytes within the non-detierented hippocampal zones to migrate out into the denervated area, and for those along the “entorhinal-hippocampal” interface to move peripherally as (I unit. Since these events preceded the earliest indication of axonal sprouting of the afferents in the hippocampal zone they further suggested that the astrocytes may be critically implicated in the sprouting process and in the redistribution
Plasticity
of axon terminals that is seen in this region after the various types of deafferentation. In view of the important implications that these observations may have for the whole phenomenon of central neural plasticity, and for the localization of afferent fiber systems in cortical areas, we re-examined our collection of normal rat brains, and of brains from animals with various experimental manipulations of the hippocampal formation. Since this provided no clear confirmation of the pattern reported by Rose et a/. [7], we prepared a number of rat brains for staining with the Cajal gold chloride-sublimate method which they had used, and a number of 1 pm thick sections from plastic embedded blocks of the rat dentate gyrus. In addition we have examined the distribution of astrocytes in gold chloride-sublimate stained preparations of the dentate gyrus in normal and reeler mice. This last group of brains is of particular interest because as we have shown elsewhere, in the reeler mouse the afferents from the entorhinal cortex occupy the entire thickness of the molecular layer and the commissural and associational afferents are distributed, for the most part, among the granule cells and throughout the hilar region of the area dentata [8]. METHOD
From a number of rat brains stained for astrocytes by the gold chloride-sublimate method (following the procedure outlined by Rose et al. [7], as closely as possible) two brains of adult albino rats, were selected for special study; one had
‘This work was supported in part by grant NS10943 from the National Institutes of Health. ‘On leave of absence from Tokyo Medical and Dental University, Bunkyo-Ku, Tokyo, Japan.
Copyright 0 1979 ANKHO International
Inc.-0361-9230/79/010035-07$01.20/O
36
KISHI, STANFIELD
FIG. 1. A low power photomicrograph of a gold chloride-sublimate preparation of the dentate gyrus of a young adult rat. In fields such as this it appears that there may be a distinct line of astrocytic somata at the junction between the inner one-fourth and the outer three-fourths of the molecular layer (at the level marked by the arrows). Scale: 100 pm.
FIG. 2. A comparable low power photomicrograph to that shown in Fig. 1, but from an adjoining section. In this field there is no indication of a “line” of astrocytic somata near the junction between the “hippocampal” and “entorhinal” zones of the molecular layer of the dentate gyrus. Scale: 100 pm.
AND COWAN
GLIAL CELLS IN THE DENTATE
CYRUS
been sectioned (at 40 pm) in the standard frontal plane and the other horizontally. Seven normal (or heterozygous +/?) and seven littermate reeler (rlirl) mouse brains were similarly prepared, the sections in these cases being cut at 30-40 pm. One micron thick plastic embedded sections were also cut from blocks of tissue taken from the hippocampus of a young adult rat which had been perfused with a mixture of 2% paraformaldehyde and 1.25% glutaraldehyde in a 0.1 M phosphate buffer. In each case sections were taken from near the middle of the septal, intermediate and temporal thirds of the hippocampal formation, and included the entire cross-sectional area of the dentate gyrus. The 1 pm sections were stained with toluidine blue and coverslipped in the usual way. For the illustrations the outlines of selected sections were traced with the aid of an Olympus microscope and drawing tube at a magni~cation of 400X, and the location of each impregnated astrocyte (in the case of the gold chloridesublimate material) and of each glial cell and superficially located neuron (in the case of the 1 pm plastic sections) carefully
marked
on the tracing. RESULTS
Figure 1 is a low-power photomi~rograph of a Cajal gold chloride-subliminate preparation comparable to those used by Rose et cd. [7]. In the field illustrated the density of astrocyte cell bodies in the molecular zone immediately adjacent to the layer of granule cells (the strcrtum grcmulosum) is relatively low, and appears to increase significantly at a distance of about 50 pm from the outer aspect of the .~tr~~tu~~
37 The abrupt increase in the number of astrocytic perikarya at this level clearly gives one the impression that it is marked by a distinct line of glial cells, and this impression is reinforced by the fact that most of the cells that are impregnated have ascending processes that extend superficially into the outer two-thirds or so of the molecular layer. Undoubtedly, it was fields like this, and the finding in some cases that the level at which the cell density increased corresponded more or less to the boundary between the zone in the molecular layer that receives afferents from the medial part of the entorhinal area supe~cially, and that receiving hippocampal afferents (from the same and opposite sides of the brain) internally, that led Rose and his colleagues to suggest that there is in fact a line of astrocytes at the interface between the two zones. However, as Fig. 2. shows, in other fields, one can see no such “tine” of astrocytic perikarya, and indeed if one examines the illustratjons in Rose ef crl.‘s paper [7], it is evident that in many of their preparations there was also no distinct border between the relatively astrocyte-poor inner part of the molecular layer and the astrocyte-rich outer two-thirds. It was this observation that first led us to question the conclusion that astrocytes are preferentially “lined-up” along the interface between the hippocampal and entorhinal zones of the molecular layer and to carry out the systematic quantitative analysis to be described below. But before presenting the quantitative data, it is worth pointing out that the distribution of the astrocytic perikarya, in the dentate gyrus, not only varies from section to section (as noted by Rose et rd. [7]), but even within one section one can find regions in which the density of glia is fairly uniform across the thick-
grunulostrm.
FIG. 3. Tracings to show the distribution of astrocytic somata over the entire cross sectional area of the dentate gyrus at each of two levels to the middle section of the central and caudal thirds of the gyrus) prepared from a well-impregnated gold chloride-sublimate
(corresponding
series of sections. Each filled-in circle marks the location of the perikaryon of an impregnated astrocyte.
38
KISHI,
ASTROCYTES (GOLD
SUBLIMATE)
4 2
20 PERCENTAGE
40
60
OF MOLECULAR
80
100
LAYER
FIG. 4. A histogram to show the relative distribution of the astrocytic somata in three representative sections from the rostral, middle and caudal thirds of the dentate gyrus. In constructing this histogram the position of each cell was expressed as a percentage of the perpendicular distance from the outer aspect of the .\t~t~/n ,~rc~~~~tb~s~~r~ to the pial surface or the hippocampal fissure, and the data from all three sections were pooled and grouped into 10 percentile “bins.”
ness of the molecular
layer immediately adjacent to others in which there is a much lower cell density in the inner part of the layer. Because of this one needs to examine the entire cross-sectional area of the gyrus at each of several levels to get a more accurate impression of the distribution of the glial cell somata. To do this we have systematically plotted the location of the somata of all astrocytes seen in three well-impregnated sections through the dentate gyrus of a young adult rat, taken from near the middle of the septal one-third of the hippocampus, near the center of the septo-temporal extent of the gyrus, and from the middle of its temporal third, respectively. Two of these three plots are shown schematically in Fig. 3: the third has been omitted for the sake of space. It is evident from these drawings that there are no marked heterogeneities in the distribution of the astrocytes across the thickness of the molecular layer although, as we shall point out later, in plastic embedded material it is evident that there is a somewhat higher density of cells immediately beneath the pia mater or the hippocampal fissure. As Rose P/ I//. [7l have emphasized that the highest proportion of as-
STANFIELD
AND
COWAN
trocytic somata occurred along a line 2755 of the distance from the outer edge of the .SVU~U/U ,~~~rrtir/osrl~l to the pial surface, we have measured the location of each cell body indicated in Fig. 3 (and in the third, more rostra1 section of the series) as a fraction of its distance along a vertical line perpendicular to the outer aspect of the strrrfrr/?r ,~rtr~~~losro~. When all these data were pooled, we could construct the histograms shown in Fig. 4. These confirm that in general the density of astrocytic perikarya is low in the inner one-third of the molecular layer, and that it rises appreciably towards the middle of this layer. In addition, they suggest that there is a somewhat higher proportion of astrocytes located between 30 and 4OZ of the way from the strvrtun? ~yw~~u/o.strm to the superficial limit of the molecular layer, although peaks like that shown in Fig. 4 are not always present. Furthermore, the “peaks” are not very prominent amounting to no more than a 3-5s increase in the number of astrocytic somata over that found in the deeper, or more superficial parts of the molecular layer. Since the gold chloride-sublimate method tends to be rather capricious, so that even different parts of the same section may show differing numbers of impregnated glial cells, we have examined several representative 1 pm thick plastic sections stained with toluidine blue at high pH. These have two advantages: first, they display t//l the glial cells present in the tissue. so that one can be confident that the sample is not biased due to inadequate or variable staining; and second, the excellent resolution that such sections provide, enables one to clearly distinguish the various classes of glial cells. We have found that when using the morphological criteria suggested by Ling ct (I/. [6] all but a relatively small percentage of the glial cells in the molecular layer of the dentate gyrus can be assigned with confidence to one or another class, and different observers usually have little difficulty in assigning cells to the same categories. The results of a quantitative analysis of three such sections, taken again from the mid-region of the septal, middle and temporal thirds of the hippocampal formation, are shown in Fig. 5, and the actual distribution of the somata of the astrocytes seen in one of these sections is shown in Fig. 6. Merely by inspection it is difficult to perceive any consistent pattern in the distribution of the astrocytes in these preparations, except that they make it clear that a very high percentage of the astrocytes is aggregated immediately beneath the pial surface and the hippocampal fissure. It is also evident that when all the data are pooled in this way. there is a slight increase in the density of astrocytic perikarya in the region just beyond the normal zone of termination of the commissural and associational projections to the molecular layer of the dentate gyrua. between 30 and 4OV of the distance from the outer border of the .strutti/ll ~qrtr~~r~lo.\c~ur. From a comparison of the relative numbers of stained astroctyes in the thin plastic sections with the numbers of cells impregnated by the gold chloride-sublimate method, we have the distinct impression that the latter technique does not impregnate all the astrocytes in the tissue. We have already commented upon the variability of the impregnations from section to section, and even within different parts of the same section, and the failure of the gold chloride-sublimate method to impregnate most of the astrocytes located near the superficial limit of the molecular layer. But it is likely also that deep within the molecular layer only a proportion of the astrocytes is stained by the gold chloride-sublimate method. Since in the same preparations one can readily identify the other classes of glial cell\. we have also analyzed the
GLIAL CELLS IN THE DENTATE GYRUS
39
r
ASTROCYTES ( PLASTIC
SECTIONS
242220.
f
ALL
GLIAL
(PLASTIC
CELLS SECTIONS)
18. 16. 1412108-
r 1
6-
,..,..,.,
20
40
PERCENTAGE
60
20
86
OF
MOLECULAR
40
60
80
100
LAYER
FIG. 5. These two histograms show the relative distributions of the astrocytic somata and of the bodies of all the glial cells seen in a series of three representative 1 pm plastic embedded sections through the dentate gyrus. The histograms were prepared in the same way as that shown in Fig. 4, and are based on data derived from reconst~ctions comparable to that shown in Fig. 6. Note the promjnent peak in the 90-100% bins in both histograms. which reflect the high density of subpial astrocytes.
distribution of the three major types--oligodendrocytes (including both the light and dark varieties recognized by Ling CI trl. [6]), microglia and astrocytes. The dist~bution of these various glia in the molecular layer at one representative level is shown in Fig. 6. Again these other cell types seem not to be preferentially localized within the dentate gyrus and the only noteworthy point is that unlike the astrocytes, they show no specific tendency to be aggregated at the outer border of the molecular layer. It is perhaps also worth pointing out that there are quite a few neurons scattered throughout the molecular layer (their distribution is indicated by the symbol * in Fig. 6); these no doubt correspond to the various cell types recognized by Cajal [2,3] and others in Golgiimpregnated preparations. The absence of a distinct band or line of astrocytes at the junction between the “hippocampa1” and “entorhinal” zones of the molecular layer of the rat, is also evident in our gold chloride-sublimate preparations of normal and reeler mice as we have reported elsewhere [9]. Although we have not subjected these preparations to the same detailed quantitative analyses, it is evident from plots of the distribution of the somata of the impregnated astroctyes that although the cell density in the hippo~ampal zone (in normal mice) is low and one can occasionally see regions in which there is a fairly abrupt increase in the number of astrocytes close to the junctional region between the entorhinal and hippocampal afferents (which in the mouse occurs along a line about 20-2570 of the distance between the struturn grcltrul~sum and the pial surface 181)there are other regions in which there is
no such change in the glial cell density. In this respect our mouse material is comparable to that in the rat, and for this reason, the observations in the reeler mouse assume an added significance. As we have indicated above, in the reeler mouse the afferents from the medial and lateral parts of the entorhinal area to the dentate gyrus are laminated as in normal mice, but their combined zone of termination extends to the superficial border of the .~t~~~t~4tn ~r~i~~lil~sI~rn. The commissural and ipsilateral associational connections in the reeler are generally absent from the molecular layer, and are confined, for the most part, to the hilar region and the rather poorly-defined struturn gyunuhum [8,10]. One might expect, if there were a critical relationship between the distribution of astrocytes and the location of the interface between the entorhinal and hippocampal afferents, that in the reeler there should be an identifiable band of astrocytes along the outer aspect of the stratum grrmulosum. In fact no such aggregation of astrocytes is evident in our gold chloride-sublimate preparations of the reeler dentate gyrus, and within the somewhat thinned molecular layer, the only indication of an increased glial cell density is that found toward the upper limit of this layer. DISCUSSION
The major conclusion to be drawn from our observations is that while the density of astrocytes (and of glial cells in general) may be relatively low within the zone of termination of the commissural and associational connections to the den-
KISHI, STANFIELD
l
AND COWAN
ASTROCYTES
0 ~LI~O~ENDK~C
YT
A MICROGLIA * NEURONS ~NIOENTIFIED
FIG. 6. A plot to show the distribution of the astrocytic somata, and the location of the other &al cells and the occasional neurons seen in the molecular layer of the dentate gyrus, in a representative 1 pm plastic embedded section ihrough the rostra1 third of the gyrus.
tate gyrus, there is no clear and consistent line or band of astrocytes along the junctional border between this zone and that in which the afferents from the medial part of the entorhinal area terminate. In some sections there is a slight increase in the density of astrocytic perikarya just beyond the junctional region-between 30 and 40% of the distance from the outer edge of the srlat~m gr~uwk~src~~~to the top of the molecular layer, but this is not constant, and by itself hardly impelling enough to support the hypothesis that “the astroglia represent some type of limiting barrier for the fibers which innervate the inner molecular layer and [that] the extent of their outward migration determines the extent of axon sprouting by these fibers” as Rose et trl. [7] have proposed. Moreover, in the reeler mouse, in which the entire thickness of the molecular layer is occupied by the entorhinal afferents, there is no evidence of a comparable line of astrocytic perikarya along the outer edge of the rather poorly-defined strtrtrrrn ~~44t744/4~~s14t~t1 although this marks the junction be-
tween the commissural and associational afferents and those from the medial entorhinal area [8]. In the brains of normal rats and mice impregnated by Cajal’s gold chloride-sublimate method, one can occasionally see short regions of the molecular layer in which the astrocytes appear to be lined up near the border between entorhinal and the hippocampal a&rents, and this appearance is accentuated by the fact that most of the impregnated astrocytic processes ascend superficially from the perikarya. But such areas are too variable and too infrequent to warrant the assumption that the arrangement of the astrocytes is in some way responsible for the sharp segregation between the two classes of afferents. This is not to say that the glial cells in general, and the astrocytes in particular, play no role in the development or maintenance of the molecular layer; clearly their presence in the layer is important, but it seems doubtful that they could play such a critical morphogenetic role as Rose et ctf. ]7] have implied. The fact that in normal
GLIAL CELLS IN THE DENTATE
GYRUS
41
animals there is no marked or constant accumulation of astrocytes in the molecular layer along the border between the entorhinal and hippocampal afferents, makes it difficult to assess the claim that after selective removal of the entorhinal afferents the astrocytes along or near this band advance outwards in a line, in anticipation of the sprouting of the hippocampal afferents and their spread into the denervated “entorhinal zone” [7]. We are inclined to think that the apparent prominence of astrocytes near the interface between the hippocampal and entorhinal afferents after lesions of the entorhinal area or perforant tract, is associated with the general glial response to the degeneration of the terminals of the entorhinal fibers. In support of this view is the observation that the glial response occurs within 48 hr after the placement of the lesion, which is some time before the earliest signs of collateral sprouting from the hippocampal zone, but coincides rather precisely with the onset of terminal degeneration in the entorhinal zone. Since it is well known that astrocytes undergo a vigorous and prompt reaction to axonal degeneration in this region [ 1,5], and that they participate actively in the engulfment and removal of the degenerating axon terminals, it is perhaps not altogether unexpected that there should be some increase in either the number or the prominence of these cells in the area of axonal degeneration. And
to the degree that they are responsible for the removal of the degenerating axon terminals and the freeing up of the synaptic sites for subsequent re-occupation by collateral sprouts from neighboring axons, they may well be thought of as participating in the axonal re-organization that occurs after lesions of the entorhinal area. Furthermore, if during development there is some degree of overlap between the entorhinal and hippocampal afferents which is subsequently eliminated (perhaps as the result of some competitive interaction between the two classes of afferents) it would not be surprising if this too were marked by the persistence, in some regions of the molecular layer, of a number of astrocytes. But this is entirely speculative, and it must be admitted that at present we know all too little about the factors which determine either the number or the distribution of any type of glial or supporting cell in the central or peripheral nervous system.
ACKNOWLEDGEMENTS
We should like to thank Mr. Mark Connelly for his assistance in the preparation of the plastic embedded material, Mr. Marc Davis for help with the photography and Ms. Sharon Musgrove for secretarial help.
REFERENCES 1. Alksne, J. F., T. W. Blackstad, F. Walbergand L. E. White, Jr. Electron microscopy of axon degeneration: a valuable tool in experimental neuroanatomy. Ergehn. Antrf. Gesch. 39: 1-31, 1%6. 2. Cajal, S. Ramon y. Estructura de1 asta de Ammon. Anal. Sot. Esp. Histol. Nut. Madrid 22: 53-114, 1893. 3. Cajal, S. Ramon y. Histologic du Sysrkme Nerveux de I’Homme et des VertPhrPs, T.2. Paris: A. Maloine, 1911, 993 pp. 4. Cajal, S. Ramon y. Contribution al conocimiento de la neuroglia de1 cerebra humano. Truh. Lab. Itnwsr. Sol. Univ. Madrid 11: 255-315, 1913. 5. Laatsch, R. H. and W. M. Cowan. Electron microscopic studies of the dentate gyrus of the rat. II. Degeneration of commissural afferents. J. camp. Neural. 130: 241-262, 1967. 6. Ling, E. A., J. A. Paterson, A. Privat, S. Mori and C. P. LeBlond. Investigation of glial cells in semithin sections. I. Identification of glial cells in the brain of young rats. J. camp. Neural. 149: 43-72. 1973.
7. Rose, G., G. Lynch and C. W. Cotman. Hypertrophy and redistribution of astrocytes in the deafferented dentate gyrus. Brain Res. Bull. 1: 87-92, 1976. 8. Stanfield, B. B., V. S. Caviness, Jr. and W. M. Cowan. The organization of certain afferents to the hippocampus and dentate gymrusin normal and reeler mice. J. co&. Neuiol. In press. 9. Stantield, B. B. and W. M. Cowan. The mornholoev of the hippocampus and dentate gyrus in normal and >eele;‘mice. J. camp. Neural. In press. 10. Stirling, R. V. and T. V. P. Bliss. Observations on the commissural projection to the dentate gyrus in the reeler mutant mouse. Brrrin Res.
150: 447-465, 1978.
