GLIA 5:lO-16 (1992)
Sexual Dimorphism in the Hamster Cerebellum Demonstrated by Glial Fibrillary Acidic Protein (GFAP) and Vimentin Immunoreactivity I. SUAREZ,~G. BODEGA,' M. RUBIO,~AND B. FERNANDEZ~ 'Departamento de Biologia Celular y Genetica, Universidad de Alcala, 28871 Madrid, and 'Departamento de Biologia Celular, Universidad Complutense, 28040 Madrid, Spain
KEY WORDS
Astroglia
ABSTRACT Male and female hamsters aged 1,4,and 10postnatal weeks were used to study the distribution of vimentin and glial fibrillary acidic protein (GFAP) in the cerebellum. Vimentin immunoreactivity exceeded that of GFAP during the first postnatal week, although GFAP was also observed in all cerebellar layers. Immunoperoxidase analysis revealed that by the fourth postnatal week vimentin was only detected in Bergmann fibers and the very scarce fibrous astrocytes located in the inner white matter. The Purkinje cell bodies were only coated with GFAP-immunopositive processes. At 10 weeks, vimentin immunoreactivity was reduced to thin Bergmann glial processes, whereas GFAP immunoreactivity had greatly increased in the whole cerebellum. The GFAP immunostaining was denser in males than in females; however, in females, the Bergmann fibers were heavily immunostained with anti-vimentin in contrast to the males. The results described in the present paper indicate a sex difference in vimentin and GFAP immunoreactivities in the cerebellar astrocytes at 4 weeks of age, which persisted in the oldest hamsters in this study. The existence of sexual dimorphism might suggest that the expression of both gliofilament proteins could be influenced by circulating sex steroids.
INTRODUCTION The cerebellum has a well-organized structure that facilitates the identification and localization of the various cell types. The rodent cerebellum has been well studied (Altman, 1972a-c; Herndon, 1964; Palay and Chan-Palay, 1974) and the morphology and the topographical distribution of cerebellar glial cells are also well known (Bovolenta et al., 1984; Herndon, 1964; Palay and Chan-Palay, 1974; Ramon y Cajal, 1911). The cerebellum, therefore, provides a useful model for analyzing the development and distribution of the intermediate filaments present in the different cerebellar astroglial cells. It has been established that vimentin and glial fibrillary acidic protein (GFAP) represent the principal constituents of the intermediate filaments found in astro@ 1992 Wiley-Liss, Inc
cytes (Dahl, 1981). In rats, vimentin is replaced by GFAP during the second and third postnatal weeks (Dahl, 1981; Pixley and De Vellis, 1984), although the two proteins can coexist in mammalian adult glia, such as Bergmann glia, Muller glia, and fibrous astrocytes from large myelinated tracts (Bovolenta et al., 1984; Schnitzer et al., 1981), as well as in cultured astroglial cells (Bignami and Dahl, 1989; Fedoroff et al., 1983). The postnatal development of GFAP-immunoreactive radial glial cells into GFAP-immunopositive astrocytes (Benjelloun-Touimi et al., 1985; Bignami and Dahl, 1973; Suarez et al., 1987), as well as the increase of the GFAP immunoreactivity with age (Bignami and Dahl, Received June 6,1990; accepted April 26,1991. Address reprint requests to Dr. I. Suarez, Departamento de Biologia Celular y Genetica, Universidad de Alcala, Alcala de Henares, 28871 Madrid, Spain.
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM
11
Fig. 1. Vimentin (A)and GFAP (B)immunoreactivities in the molecular (ml)and the granular (gl)layers a t 1 week of age. Vimentin-positive Bergmann processes (A) are thicker than GFAP-positive Bergmann fibers (B).Both types of immunopositive processes form the subpial glia limitans (arrows).Counterstained with toluidine blue. X305.
1973; Suarez et al., 1987; Weir et al., 1984), have been described in rodents. However, factors regulating GFAP levels during development or in the intact adult organism have not been demonstrated. During development, steroid hormones participate in sexual differentiation of the brain (Arnold and Gorski, 1984; Dohler et al., 1984). Studies on sex differences have mainly focused on neurons, but little attention has been paid to sex differences in glial cells. Recently, it has been shown that sex steroids affect the redistribution of GFAP immunoreactivity in both the intact rat brain (Garcia-Segura et al., 1988; Tranque et al., 1987) and in cultured astrocytes (Garcia-Segura et al., 1989). The purpose of this work was to study the distribution of vimentin and GFAP immunoreactivities in male and female hamster cerebella and compare the results between the two sexes, in order to know the possible influence of circulating sex steroids on the expression of both proteins.
MATERIALS AND METHODS Twelve male and twelve female hamsters aged 1, 4, and 10 weeks were used in our study. The animals were
housed under controlled lighting conditions (12 h light, 12 h dark), and given chow and water ad libitum. Hamsters were anesthetized with ether and decapitated, and the cerebella were quickly removed and placed in B5 fixative (Bullon et al., 19841,for 5 h at 4°C. The cerebella were dehydrated in graded concentrations of ethanol and embedded in paraffin; tissue sections were cut at 10 pm. Consecutive sections were immunostained with GFAP and vimentin antibodies. The PAP method using polyclonal GFAP antiserum (Dakopatts, Denmark), diluted 1500 in Tris-buffer, was performed as previously described (Suarez et al., 1987). For the vimentin immunohistochemical method, the sections were incubated for 30 min in normal serum at a 1:30 dilution. Without washing, sections were incubated overnight at 4°C with monoclonal vimentin antiserum (Dakopatts), diluted 1:400 in Tris-buffer, pH 7.6. Following three 10 min washes in Tris-buffer, the sections were incubated 1 h a t 30°C in peroxidaseconjugated rabbit anti-mouse immunoglobulin (Dakopatts) a t a 1 5 0 dilution and rinsed twice in Tris-buffer. Peroxidase activity was revealed with 0.025% 3,3’-diaminobenzidine in Tris-buffer with 0.005% H202for 10 min; the sections were then washed with distilled water, counterstained with toluidine blue, dehydrated in
12
S U a E Z ET AL.
Figs. 2, 3.
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM
13
(Fig. 4), and the difference between male and female GFAP immunostaining in both gray and white matter was significant (Fig. 6). Vimentin-positive astrocytes could no longer be observed in the granule cell layer (Figs. 2A,3A94A).GFAP-positiveastrocytes in the white matter were very numerous (Fig. 4B) and they were densely immunostained; some of them surrounded blood vessels (Fig. 4B). No vimentin-positive astrocytes were observed in the white matter, with the exception of several perivascular astrocytes located in the inner white matter (Fig. 4A). At 10 weeks, the cerebellar pattern was similar to the one observed at 4 weeks, although the density of the GFAP-positive structures was higher than in previous ages (Figs. 5B,6). Vimentin-positive processes were reduced to the Bergmann fibers, and the few vimentinpositive perivascular processes observed at 4 weeks in the white matter could no longer be identified at 10 weeks postnatally (Fig. 5A). Astrocytes in the granule cell layer showed denser GFAP-positive processes at 10 weeks (Fig. 5B) than those at 4 weeks (Fig. 4B). GFAP-positive cells in the white matter showed intense immunoperoxidase reaction product in both their somata and processes RESULTS (Fig. 5B). After comparing the results of both vimentin and At 1 week, Bergmann cells and astrocytes showed GFAP immunoreactivities in the two sexes, we conboth vimentin and GFAP immunostaining. The molecu- cluded that the GFAP-containing cells were more lar and the granular layers displayed more vimentin densely distributed in the male than in the female than GFAP immunoreaction product (Fig. 11, but no cerebellum. On the contrary, vimentin-positive cell dissignificant difference between males and females could tribution was slightly denser in the female than in the be found, nor was a significant difference between male cerebellum at all the ages studied. When the vimentin and GFAP immunostaining observed in the surface densities of the two sexes were compared white matter astrocytes (Fig. 6). (Fig. 6), a sexual difference in the vimentin- and GFAPAt 4 weeks, vimentin immunoreactivity was still immunoreactive product reaction was not found in the higher than GFAP immunoreactivity in the Bergmann cerebellum at 1 week of age. However, the surface astroglial cells (Figs. 2,3). The immunostained Berg- densities of both vimentin and GFAP material in all mann processes were longer and thinner than at 1week. layers in the male hamster at 4 weeks and 10 weeks of Moreover, the immunoreactive processes reflected a age were significantly different from those of females at larger amount of vimentin in females than in males the corresponding age (P < 0.001). (Figs. 2A,3A), whereas the GFAP-positive Bergmann processes were denser in males than in females (Figs. DISCUSSION 2B,3B). The distribution of GFAP immunoreactivity was In the present study it has been shown that the markedly different from that of vimentin immunoreacdeveloping hamster cerebellum involves significant diftivity in both the granule cell layer and the white matter ferences in vimentin and GFAP immunoreactivities, which are dependent on age and sex. The Bergmann fibers stain with both vimentin and GFAP antibodies, but a decrease in vimentin immunostaining related to Fig. 2. Adjacent sections in the female molecular layer, at 4 weeks of age was significant. The Bergmann glia in the adult age, immunostained for vimentin (A) and GFAP (B). Numerous vimen- hamster showed more GFAP- than vimentin-immunotin-positive Bergmann cells send out radial processes towards the reactive material, whereas the adult rat Bergmann glia subpial surface, whereas they are weakly immunoreactive for GFAP (B). Note that some blood vessels are surrounded by vimentin- and stained equally with anti-vimentin and anti-GFAP GFAP-stained processes (arrows in A and B). Arrowheads, Purkinje (Dahl et al., 1981). The existence of two separate sysneurons; gl, granular layer. Counterstained with toluidine blue. x 600. tems of intermediate filaments in astroglia has been Fig. 3. Male molecular layer, at 4 weeks of age. A: Vimentin-positive suggested (Dahl et al., 1981) and the presence of both Bergmann processes are less numerous than in the female molecular vimentin and GFAP in adult Bergmann fibers could be layer (see Fig. 2A). B: GFAP-immunopositiveBergmann processes are denser than in the female cerebellum (see Fig. 2B). Blood vessels are explained by the permanent construction of filaments surrounded by vimentin- and GFAP-stained processes (arrows in A and comprised of both subunits (Bovolenta et al., 1984). The B). Arrowheads, Purkinje neurons; gl, granular layer. Counterstained with toluidine blue. X600. persistence of some vimentin-positive Bergmann fibers graded concentrations of ethanol, and mounted in DePeX. As with GFAP-immunostained sections, groups of vimentin-imunostained sections from all the different developmental stages studied were incubated in parallel on the same day. Some sections were incubated with preimmune serum at a 1:30 dilution as the primary antiserum; these control sections showed no immunoreactive product. Vimentin- and GFAP-immunostained sections were observed and photographed through a Zeiss microscope. Quantitative evaluation of immunoreactive material was performed utilizing a stereologic grid, according to the point-counting method of Weibel (19791, in which the ratio of the surface of imunoreactive profiles to the volume of a given structure [surface density (Sv)] is calculated by the following formula: Sv = 2I/L, where I is the number of points a t which the immunoreactive profiles cross the test grid lines and L is the test line length in the tissue (Weibel, 1979).The observer was not aware of the identity or sex of the cerebellum from which the sections were taken. Mean values were compared with the unpaired Student’s t-test.
14
SUAREZ ET AL.
Figs. 4,5.
15
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM No l e c u l a r Layer
W h i t e Maller
Granule Cell Layzr
1.5 1 .o
0.5 GF t iP
0.0-
~
lu
4u
1 ou
lu
1 ou
4u
WEEKS
lu
WEEKS
4u
Iou
WEEKS
A Grm,le
Molecular Layer
2.0
S"
1.5
-
Cei I Layer
While H a l I ? r
1.0
0.5 GF flf' mVirnent
0.0
lu
4u WEEKS
I ou
lu
4u WEEKS
1 ou
lu
4u
in
1 ou
WEEKS
B Fig. 6. Surface density (Sv) of vimentin- and glial fibrillary acidic protein (GFAP)-immunoreactive material in the male (A) and female (B) cerebellum, at 1 (lw), 4 (4w), and 10 (low) weeks of age. A significant difference in GFAF' immunoreactivity was found between 1 and 4 weeks ( P < 0.001) and between 4 and 10 weeks ( P < 0.001) in both male and female cerebellum.
and the emergence of GFAP-containing Bergmann glia of the mature brain remain to be explained. The decreased width and increased length of the Bergmann processes in hamsters during postnatal development are similar to the observations of Hanke and Reichenbach (1987) in rat cerebellum, concomitant to thickening of the molecular layer. The surface density of vimentin and GFAP is also dependent on sex, since the amount of vimentin is lower and GFAP is greater in males than in females.
Fig. 4. Vimentin (A) and GFAP (B)immunoreactivities in the male cerebellum at 4 weeks of age. Note the absence of staining for vimentin (A) in the granule cell layer (gl) and the white matter (wm), except in the inner perivascular astrocytes (arrow), while GFAP immunostaining (B)is present in both layers. Astrocytes in the white matter show an intense immunoreaction for GFAP. ml, molecular layer; V, IV ventricle. Counterstained with toluidine blue. ~ 2 7 7 . Fig. 5. Vimentin (A) and GFAP (B) immunoreactivities in the male cerebellum at 10 weeks of age. A Note the absence of vimentin-positive astrocytes in both the granular layer (gl)and the white matter (wm).B: GFAF' immunostaining is denser at 10 weeks when compared to 4 weeks (see Fig. 4B); blood vessels are surrounded by GFAP-positive astrocytic processes (arrow). ml, molecular layer. Counterstained with toluidine blue. x 317.
At 1week most of the cerebellar astroglial population contained both proteins, as has often been demonstrated in other astroglial cell types (Dahl, 1981; Yen and Fields, 1981).The astrocytes within the granule cell layer showed less GFAP than the rest of the cerebellum at this age. The GFAP content increased rapidly from the first postnatal week, with a corresponding decrease in vimentin. The vimentin-GFAP transition has been related to the period of rapid myelination (Dahl, 1981) and the increase in GFAP is thought to correspond with the astrocytic differentiation (Garcia-Segura et al., 1989). The functional significance of the differential expression of vimentin and GFAP in glial cells is unknown, but our results lead us t o suggest that the astrocytic maturation that takes place in the hamster astrocytes located within the granule cell layer, where complex synaptic connections are being established, is slower than in the axon-related astrocytes in the white matter. Consequently, the time of the vimentin-GFAP transition in astrocytes located in the granule cell layer is delayed when compared to fibrous astrocytes located in the white matter. From the fourth week on, the GFAP immunostaining
16
S U h E Z ET AL.
was higher in males than in females. Several considerations could account for the difference. It is known that astrocytes participate in synaptic reorganization by inserting or retracting their processes from between neuronal membranes (Meshul et al. 1987; Montagnese et al., 1988);furthermore, it has been shown that estradiol induces the redistribution of GFAP in the rat brain (Tranque et al., 1987) and can also modulate the morphology of cultured astrocytes (Garcia-Segura et al., 1989). If estrogen treatment has been reported to increase the number of synapses (Carrer and Aoki, 1982; Smith et al., 1989) and the number of synaptic contacts in females is higher than in males (Matsumoto and Arai, 19861, a decrease in GFAP immunoreactivity could be expected in the female hamster cerebellum, as has been observed in our study. It has been observed that postnatal treatment with an estrogen antagonist inhibited the differentiation of a sexually dimorphic nucleus in the rat (Dohler et al., 1984). If estrogens were one of the determining factors in sexual dimorphism, it would seem likely that the low GFAP-immunopositivesurface density observed in the female hamster cerebellum might depend on the exposure t o higher levels of estrogens than in males. Male and female differences in GFAP immunoreactivity have been observed in rats and it has been suggested that GFAP expression could be influenced by sex steroids (Garcia-Segura et al., 1988). It is clear that the external environment of the astrocytes varies from male to female in relation to the levels of circulating sex steroids, and this different environment might confer different properties on the astroglial cells. The nature of the signals that initiate and regulate the vimentin-GFAP transition is not known and further experiments are therefore needed to resolve this question.
ACKNOWLEDGMENTS We are grateful to C.F. Warren (ICE at U.A.H.)for her linguistic assistance. This study was partially supported by CAICYT grant PB86/0152.
REFERENCES Altman, J. (1972a) Postnatal development of the cerebellar cortex in the rat. I. The external germinal layer and the transitional molecular layer. J. Cornp.Neurol., 145:353-398. Altman, J . (1972b) Postnatal development of the cerebellar cortex in the rat. 11. Phases in the maturation of Purkinje cells and of the molecular layer. J . Cornp. Neurol., 145:399-464. Altman, J . ( 1 9 7 2 ~Postnatal ) development of the cerebellar cortex in the rat. 111. Maturation of the components of the granular layer. J . Cornp. Neurol., 145465-514. Arnold, A.P. and Gorski, R.A. (1984) Gonadal steroid induction of structural sex differences in the central nervous system. Annu. Reu. Neurosci., 7:413-442. Benjelloun-Touimi, S., Jaque, C.M., Derer, P., De Vitry, F., Maunoury, R., and Dupouey, P. (1985) Evidence that mouse astrocytes may be derived from the radial glia. An immunohistochemical study of the cerebellum in the normal and reeler mouse. J. Neuroimrnunol., 9:87-97. Bignami, A. and Dahl, D. (1973) Differentiation of astrocytes in the cerebellar cortex and the pyramidal tracts of the newborn rat. An
immunofluorescence study with antibodies to a protein specific to astrocytes. Brain Res., 49:393-402. Bignami, A. and Dahl, D. (1989)Vimentin-GFAP transition in primary dissociated cultures of rat embryo spinal cord. Int. J . Deu. Neurosci., 7:343-35 7. Bovolenta, P., Liem, R.K.H., and Mason, C.A. (1984) Development of cerebellar astroglia: Transitions in form and cytoskeletal content. Deu. Biol., 102:248-259. Bullon, M.M., Alvarez-Gago, T., Fernandez-Ruiz, B., and Aguirre, C. (1984) Glial fibrillary acidic (GFA) protein in rat spinal cord. An immunoperoxidase study in semithin sections. Brain Res., 309~79-83. Carrer, H.F. and Aoki, A. (1982) Ultrastructural changes in the hypothalamic ventromedial nucleus of ovariectomized rats after estrogen treatment. Brain Res., 240:221-233. Dahl, D. (1981) The vimentin-GFA protein transition in the rat neuroglia cytoskeleton occurs at the time of myelination. J . Neurosci. Res., 6:741-748. Dahl, D., Rueger, D.C., Bignami, A,, Weber K., and Osborn, M. (1981) Vimentin, the 57,000 dalton protein of fibroblast filaments, is the major cytoskeletal component in immature glia. Eur. J . Cell Biol., 24~191-196. Dohler, K.D., Hancke, J.L., Srivastava, S.S., Hofmann, C., Shryne, J.E., and Gorski, R.A. (1984) Participation of estrogens in female sexual differentiation of the brain; neuroanatomical, neuroendocrine and behavioral evidence. Prog. Brain Res., 61:99-117. Fedoroff, S., White, R., Neal, J., Subrahmanyan, L., and Kalnins, V.I. (1983)Astrocyte cell lineage. 11. Mouse fibrous astrocytes and reactive astrocytes in cultures have vimentin and GFP containing intermediate filaments. Dew. Brain Res., 7:303-315. Garcia-Segura, L.M., Suarez, I., Segovia,S.,Tranque, P.A., Cales, J.M., Aguilera, P., Olmos, G., and Guillamon, A. (1988) The distribution of glial fibrillary acidic protein in the adult rat brain is influenced by the neonatal levels of sex steroids. Brain Res., 456:357-363. Garcia-Segura, L.M., Torres-Aleman, I., and Naftolin, F. (1989) Astrocytic shape and glial fibrillary acidic protein immunoreactivity are modified by estradiol in primary rat hypothalamic cultures. Deu. Brain Res., 47:298-302. Hanke, S. and Reichenbach, A. (1987) Quantitative-morphometric aspects of Bergmann glial (Golgi epithelial) cell development in rats. A Golgi study. Anat. Ernbryol., 177:183-188. Herndon, R.M. (1964) The fine structure of rat cerebellum. 11. The stellate neurones, granule cells and glia. J . Cell Biol., 23:277-293. Matsumoto, A. and Arai, Y. (1986) Male-female difference in synaptic organization of the ventromedial nucleus of the hypothalamus in the rat. Neuroendocrinology, 42:232-236. Meshul, C.K., Seil, F.J., and Herndon, R.M. (1987) Astrocytes play a role in regulation of synaptic density. Brain Res., 402:139-145. Montagnese, C., Poulain, D.A., Vincent, J.D., and Theodosis, D.T. (1988) Synaptic and neuronal-glial plasticity in the adult oxytocinergic system in response to physiological stimuli. Brain Res. Bull., 20:681-692. Palay, S.L., and Chan-Palay, V. (1974) Cerebellar Cortex: Cytology and Organization. Springer-Verlag, Berlin, pp. 288-321. Pixley, S.R.K. and De Vellis, J. (1984) Transition between immature radial glia and mature astrocytes studied with monoclonal antibody to vimentin. Deu. Brain Res., 15:201-209. Ramon y Cajal, S. (1911)Histologie du Systerne Nerueux de 1'Hornrne et des Vertebres, Vol. 11. Maloine, Paris, pp. 68-71. Schnitzer, J., Franke, W.W., and Schachner, M. (1981) Immunocytochemical demonstration of vimentin in astrocytes and ependymal cells of developing and adult nervous system. J. Cell Biol., 90:435447. Smith, S.S., Woodward, D.J., and Chapin, J.K. (1989) Sex steroids modulate motor-correlated increases in cerebellar discharge. Brain Res., 476:307-316. Suarez. I.. Fernandez, B.. Bodega. G., Tranaue, P., Olmos. G.. and Garcia-Segura, L.M.'(1987)Postnatal development of glial fibrillary acidic protein immunoreactivity in the hamster arcuate nucleus. Deu. Brain Res., 37:89-95. Tranque, P.A., Suarez, I., Olmos, G., Fernandez, B., and GarciaSegura, L.M. (1987) Estradiol-induced redistribution of glial fibrillary acidic protein immunoreactivity in the rat brain. Brain Res., 406:348-351. Weibel, E.R. (1979) Stereological Methods, Vol. I in Practical Methods for Biological Morphornetry. Academic Press, London. Weir, M.D., Patel, A.J., Hunt, A,, and Thomas, D.G.T. (1984) Developmental changes in the amount of glial fibrillary acidic protein in three regions of the rat brain. Deu. Brain Res., 15:147-154. Yen, S.H. and Fields, K.L. (1981) Antibodies to neurofilament, glial filament and fibroblast intermediate filament proteins bind to different cell types ofthe nervous sytem. J . Cell Biol., 88:115-126.
Sexual Dimorphism in the Hamster Cerebellum Demonstrated by Glial Fibrillary Acidic Protein (GFAP) and Vimentin Immunoreactivity I. SUAREZ,~G. BODEGA,' M. RUBIO,~AND B. FERNANDEZ~ 'Departamento de Biologia Celular y Genetica, Universidad de Alcala, 28871 Madrid, and 'Departamento de Biologia Celular, Universidad Complutense, 28040 Madrid, Spain
KEY WORDS
Astroglia
ABSTRACT Male and female hamsters aged 1,4,and 10postnatal weeks were used to study the distribution of vimentin and glial fibrillary acidic protein (GFAP) in the cerebellum. Vimentin immunoreactivity exceeded that of GFAP during the first postnatal week, although GFAP was also observed in all cerebellar layers. Immunoperoxidase analysis revealed that by the fourth postnatal week vimentin was only detected in Bergmann fibers and the very scarce fibrous astrocytes located in the inner white matter. The Purkinje cell bodies were only coated with GFAP-immunopositive processes. At 10 weeks, vimentin immunoreactivity was reduced to thin Bergmann glial processes, whereas GFAP immunoreactivity had greatly increased in the whole cerebellum. The GFAP immunostaining was denser in males than in females; however, in females, the Bergmann fibers were heavily immunostained with anti-vimentin in contrast to the males. The results described in the present paper indicate a sex difference in vimentin and GFAP immunoreactivities in the cerebellar astrocytes at 4 weeks of age, which persisted in the oldest hamsters in this study. The existence of sexual dimorphism might suggest that the expression of both gliofilament proteins could be influenced by circulating sex steroids.
INTRODUCTION The cerebellum has a well-organized structure that facilitates the identification and localization of the various cell types. The rodent cerebellum has been well studied (Altman, 1972a-c; Herndon, 1964; Palay and Chan-Palay, 1974) and the morphology and the topographical distribution of cerebellar glial cells are also well known (Bovolenta et al., 1984; Herndon, 1964; Palay and Chan-Palay, 1974; Ramon y Cajal, 1911). The cerebellum, therefore, provides a useful model for analyzing the development and distribution of the intermediate filaments present in the different cerebellar astroglial cells. It has been established that vimentin and glial fibrillary acidic protein (GFAP) represent the principal constituents of the intermediate filaments found in astro@ 1992 Wiley-Liss, Inc
cytes (Dahl, 1981). In rats, vimentin is replaced by GFAP during the second and third postnatal weeks (Dahl, 1981; Pixley and De Vellis, 1984), although the two proteins can coexist in mammalian adult glia, such as Bergmann glia, Muller glia, and fibrous astrocytes from large myelinated tracts (Bovolenta et al., 1984; Schnitzer et al., 1981), as well as in cultured astroglial cells (Bignami and Dahl, 1989; Fedoroff et al., 1983). The postnatal development of GFAP-immunoreactive radial glial cells into GFAP-immunopositive astrocytes (Benjelloun-Touimi et al., 1985; Bignami and Dahl, 1973; Suarez et al., 1987), as well as the increase of the GFAP immunoreactivity with age (Bignami and Dahl, Received June 6,1990; accepted April 26,1991. Address reprint requests to Dr. I. Suarez, Departamento de Biologia Celular y Genetica, Universidad de Alcala, Alcala de Henares, 28871 Madrid, Spain.
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM
11
Fig. 1. Vimentin (A)and GFAP (B)immunoreactivities in the molecular (ml)and the granular (gl)layers a t 1 week of age. Vimentin-positive Bergmann processes (A) are thicker than GFAP-positive Bergmann fibers (B).Both types of immunopositive processes form the subpial glia limitans (arrows).Counterstained with toluidine blue. X305.
1973; Suarez et al., 1987; Weir et al., 1984), have been described in rodents. However, factors regulating GFAP levels during development or in the intact adult organism have not been demonstrated. During development, steroid hormones participate in sexual differentiation of the brain (Arnold and Gorski, 1984; Dohler et al., 1984). Studies on sex differences have mainly focused on neurons, but little attention has been paid to sex differences in glial cells. Recently, it has been shown that sex steroids affect the redistribution of GFAP immunoreactivity in both the intact rat brain (Garcia-Segura et al., 1988; Tranque et al., 1987) and in cultured astrocytes (Garcia-Segura et al., 1989). The purpose of this work was to study the distribution of vimentin and GFAP immunoreactivities in male and female hamster cerebella and compare the results between the two sexes, in order to know the possible influence of circulating sex steroids on the expression of both proteins.
MATERIALS AND METHODS Twelve male and twelve female hamsters aged 1, 4, and 10 weeks were used in our study. The animals were
housed under controlled lighting conditions (12 h light, 12 h dark), and given chow and water ad libitum. Hamsters were anesthetized with ether and decapitated, and the cerebella were quickly removed and placed in B5 fixative (Bullon et al., 19841,for 5 h at 4°C. The cerebella were dehydrated in graded concentrations of ethanol and embedded in paraffin; tissue sections were cut at 10 pm. Consecutive sections were immunostained with GFAP and vimentin antibodies. The PAP method using polyclonal GFAP antiserum (Dakopatts, Denmark), diluted 1500 in Tris-buffer, was performed as previously described (Suarez et al., 1987). For the vimentin immunohistochemical method, the sections were incubated for 30 min in normal serum at a 1:30 dilution. Without washing, sections were incubated overnight at 4°C with monoclonal vimentin antiserum (Dakopatts), diluted 1:400 in Tris-buffer, pH 7.6. Following three 10 min washes in Tris-buffer, the sections were incubated 1 h a t 30°C in peroxidaseconjugated rabbit anti-mouse immunoglobulin (Dakopatts) a t a 1 5 0 dilution and rinsed twice in Tris-buffer. Peroxidase activity was revealed with 0.025% 3,3’-diaminobenzidine in Tris-buffer with 0.005% H202for 10 min; the sections were then washed with distilled water, counterstained with toluidine blue, dehydrated in
12
S U a E Z ET AL.
Figs. 2, 3.
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM
13
(Fig. 4), and the difference between male and female GFAP immunostaining in both gray and white matter was significant (Fig. 6). Vimentin-positive astrocytes could no longer be observed in the granule cell layer (Figs. 2A,3A94A).GFAP-positiveastrocytes in the white matter were very numerous (Fig. 4B) and they were densely immunostained; some of them surrounded blood vessels (Fig. 4B). No vimentin-positive astrocytes were observed in the white matter, with the exception of several perivascular astrocytes located in the inner white matter (Fig. 4A). At 10 weeks, the cerebellar pattern was similar to the one observed at 4 weeks, although the density of the GFAP-positive structures was higher than in previous ages (Figs. 5B,6). Vimentin-positive processes were reduced to the Bergmann fibers, and the few vimentinpositive perivascular processes observed at 4 weeks in the white matter could no longer be identified at 10 weeks postnatally (Fig. 5A). Astrocytes in the granule cell layer showed denser GFAP-positive processes at 10 weeks (Fig. 5B) than those at 4 weeks (Fig. 4B). GFAP-positive cells in the white matter showed intense immunoperoxidase reaction product in both their somata and processes RESULTS (Fig. 5B). After comparing the results of both vimentin and At 1 week, Bergmann cells and astrocytes showed GFAP immunoreactivities in the two sexes, we conboth vimentin and GFAP immunostaining. The molecu- cluded that the GFAP-containing cells were more lar and the granular layers displayed more vimentin densely distributed in the male than in the female than GFAP immunoreaction product (Fig. 11, but no cerebellum. On the contrary, vimentin-positive cell dissignificant difference between males and females could tribution was slightly denser in the female than in the be found, nor was a significant difference between male cerebellum at all the ages studied. When the vimentin and GFAP immunostaining observed in the surface densities of the two sexes were compared white matter astrocytes (Fig. 6). (Fig. 6), a sexual difference in the vimentin- and GFAPAt 4 weeks, vimentin immunoreactivity was still immunoreactive product reaction was not found in the higher than GFAP immunoreactivity in the Bergmann cerebellum at 1 week of age. However, the surface astroglial cells (Figs. 2,3). The immunostained Berg- densities of both vimentin and GFAP material in all mann processes were longer and thinner than at 1week. layers in the male hamster at 4 weeks and 10 weeks of Moreover, the immunoreactive processes reflected a age were significantly different from those of females at larger amount of vimentin in females than in males the corresponding age (P < 0.001). (Figs. 2A,3A), whereas the GFAP-positive Bergmann processes were denser in males than in females (Figs. DISCUSSION 2B,3B). The distribution of GFAP immunoreactivity was In the present study it has been shown that the markedly different from that of vimentin immunoreacdeveloping hamster cerebellum involves significant diftivity in both the granule cell layer and the white matter ferences in vimentin and GFAP immunoreactivities, which are dependent on age and sex. The Bergmann fibers stain with both vimentin and GFAP antibodies, but a decrease in vimentin immunostaining related to Fig. 2. Adjacent sections in the female molecular layer, at 4 weeks of age was significant. The Bergmann glia in the adult age, immunostained for vimentin (A) and GFAP (B). Numerous vimen- hamster showed more GFAP- than vimentin-immunotin-positive Bergmann cells send out radial processes towards the reactive material, whereas the adult rat Bergmann glia subpial surface, whereas they are weakly immunoreactive for GFAP (B). Note that some blood vessels are surrounded by vimentin- and stained equally with anti-vimentin and anti-GFAP GFAP-stained processes (arrows in A and B). Arrowheads, Purkinje (Dahl et al., 1981). The existence of two separate sysneurons; gl, granular layer. Counterstained with toluidine blue. x 600. tems of intermediate filaments in astroglia has been Fig. 3. Male molecular layer, at 4 weeks of age. A: Vimentin-positive suggested (Dahl et al., 1981) and the presence of both Bergmann processes are less numerous than in the female molecular vimentin and GFAP in adult Bergmann fibers could be layer (see Fig. 2A). B: GFAP-immunopositiveBergmann processes are denser than in the female cerebellum (see Fig. 2B). Blood vessels are explained by the permanent construction of filaments surrounded by vimentin- and GFAP-stained processes (arrows in A and comprised of both subunits (Bovolenta et al., 1984). The B). Arrowheads, Purkinje neurons; gl, granular layer. Counterstained with toluidine blue. X600. persistence of some vimentin-positive Bergmann fibers graded concentrations of ethanol, and mounted in DePeX. As with GFAP-immunostained sections, groups of vimentin-imunostained sections from all the different developmental stages studied were incubated in parallel on the same day. Some sections were incubated with preimmune serum at a 1:30 dilution as the primary antiserum; these control sections showed no immunoreactive product. Vimentin- and GFAP-immunostained sections were observed and photographed through a Zeiss microscope. Quantitative evaluation of immunoreactive material was performed utilizing a stereologic grid, according to the point-counting method of Weibel (19791, in which the ratio of the surface of imunoreactive profiles to the volume of a given structure [surface density (Sv)] is calculated by the following formula: Sv = 2I/L, where I is the number of points a t which the immunoreactive profiles cross the test grid lines and L is the test line length in the tissue (Weibel, 1979).The observer was not aware of the identity or sex of the cerebellum from which the sections were taken. Mean values were compared with the unpaired Student’s t-test.
14
SUAREZ ET AL.
Figs. 4,5.
15
SEXUAL DIMORPHISM IN THE HAMSTER CEREBELLUM No l e c u l a r Layer
W h i t e Maller
Granule Cell Layzr
1.5 1 .o
0.5 GF t iP
0.0-
~
lu
4u
1 ou
lu
1 ou
4u
WEEKS
lu
WEEKS
4u
Iou
WEEKS
A Grm,le
Molecular Layer
2.0
S"
1.5
-
Cei I Layer
While H a l I ? r
1.0
0.5 GF flf' mVirnent
0.0
lu
4u WEEKS
I ou
lu
4u WEEKS
1 ou
lu
4u
in
1 ou
WEEKS
B Fig. 6. Surface density (Sv) of vimentin- and glial fibrillary acidic protein (GFAP)-immunoreactive material in the male (A) and female (B) cerebellum, at 1 (lw), 4 (4w), and 10 (low) weeks of age. A significant difference in GFAF' immunoreactivity was found between 1 and 4 weeks ( P < 0.001) and between 4 and 10 weeks ( P < 0.001) in both male and female cerebellum.
and the emergence of GFAP-containing Bergmann glia of the mature brain remain to be explained. The decreased width and increased length of the Bergmann processes in hamsters during postnatal development are similar to the observations of Hanke and Reichenbach (1987) in rat cerebellum, concomitant to thickening of the molecular layer. The surface density of vimentin and GFAP is also dependent on sex, since the amount of vimentin is lower and GFAP is greater in males than in females.
Fig. 4. Vimentin (A) and GFAP (B)immunoreactivities in the male cerebellum at 4 weeks of age. Note the absence of staining for vimentin (A) in the granule cell layer (gl) and the white matter (wm), except in the inner perivascular astrocytes (arrow), while GFAP immunostaining (B)is present in both layers. Astrocytes in the white matter show an intense immunoreaction for GFAP. ml, molecular layer; V, IV ventricle. Counterstained with toluidine blue. ~ 2 7 7 . Fig. 5. Vimentin (A) and GFAP (B) immunoreactivities in the male cerebellum at 10 weeks of age. A Note the absence of vimentin-positive astrocytes in both the granular layer (gl)and the white matter (wm).B: GFAF' immunostaining is denser at 10 weeks when compared to 4 weeks (see Fig. 4B); blood vessels are surrounded by GFAP-positive astrocytic processes (arrow). ml, molecular layer. Counterstained with toluidine blue. x 317.
At 1week most of the cerebellar astroglial population contained both proteins, as has often been demonstrated in other astroglial cell types (Dahl, 1981; Yen and Fields, 1981).The astrocytes within the granule cell layer showed less GFAP than the rest of the cerebellum at this age. The GFAP content increased rapidly from the first postnatal week, with a corresponding decrease in vimentin. The vimentin-GFAP transition has been related to the period of rapid myelination (Dahl, 1981) and the increase in GFAP is thought to correspond with the astrocytic differentiation (Garcia-Segura et al., 1989). The functional significance of the differential expression of vimentin and GFAP in glial cells is unknown, but our results lead us t o suggest that the astrocytic maturation that takes place in the hamster astrocytes located within the granule cell layer, where complex synaptic connections are being established, is slower than in the axon-related astrocytes in the white matter. Consequently, the time of the vimentin-GFAP transition in astrocytes located in the granule cell layer is delayed when compared to fibrous astrocytes located in the white matter. From the fourth week on, the GFAP immunostaining
16
S U h E Z ET AL.
was higher in males than in females. Several considerations could account for the difference. It is known that astrocytes participate in synaptic reorganization by inserting or retracting their processes from between neuronal membranes (Meshul et al. 1987; Montagnese et al., 1988);furthermore, it has been shown that estradiol induces the redistribution of GFAP in the rat brain (Tranque et al., 1987) and can also modulate the morphology of cultured astrocytes (Garcia-Segura et al., 1989). If estrogen treatment has been reported to increase the number of synapses (Carrer and Aoki, 1982; Smith et al., 1989) and the number of synaptic contacts in females is higher than in males (Matsumoto and Arai, 19861, a decrease in GFAP immunoreactivity could be expected in the female hamster cerebellum, as has been observed in our study. It has been observed that postnatal treatment with an estrogen antagonist inhibited the differentiation of a sexually dimorphic nucleus in the rat (Dohler et al., 1984). If estrogens were one of the determining factors in sexual dimorphism, it would seem likely that the low GFAP-immunopositivesurface density observed in the female hamster cerebellum might depend on the exposure t o higher levels of estrogens than in males. Male and female differences in GFAP immunoreactivity have been observed in rats and it has been suggested that GFAP expression could be influenced by sex steroids (Garcia-Segura et al., 1988). It is clear that the external environment of the astrocytes varies from male to female in relation to the levels of circulating sex steroids, and this different environment might confer different properties on the astroglial cells. The nature of the signals that initiate and regulate the vimentin-GFAP transition is not known and further experiments are therefore needed to resolve this question.
ACKNOWLEDGMENTS We are grateful to C.F. Warren (ICE at U.A.H.)for her linguistic assistance. This study was partially supported by CAICYT grant PB86/0152.
REFERENCES Altman, J. (1972a) Postnatal development of the cerebellar cortex in the rat. I. The external germinal layer and the transitional molecular layer. J. Cornp.Neurol., 145:353-398. Altman, J . (1972b) Postnatal development of the cerebellar cortex in the rat. 11. Phases in the maturation of Purkinje cells and of the molecular layer. J . Cornp. Neurol., 145:399-464. Altman, J . ( 1 9 7 2 ~Postnatal ) development of the cerebellar cortex in the rat. 111. Maturation of the components of the granular layer. J . Cornp. Neurol., 145465-514. Arnold, A.P. and Gorski, R.A. (1984) Gonadal steroid induction of structural sex differences in the central nervous system. Annu. Reu. Neurosci., 7:413-442. Benjelloun-Touimi, S., Jaque, C.M., Derer, P., De Vitry, F., Maunoury, R., and Dupouey, P. (1985) Evidence that mouse astrocytes may be derived from the radial glia. An immunohistochemical study of the cerebellum in the normal and reeler mouse. J. Neuroimrnunol., 9:87-97. Bignami, A. and Dahl, D. (1973) Differentiation of astrocytes in the cerebellar cortex and the pyramidal tracts of the newborn rat. An
immunofluorescence study with antibodies to a protein specific to astrocytes. Brain Res., 49:393-402. Bignami, A. and Dahl, D. (1989)Vimentin-GFAP transition in primary dissociated cultures of rat embryo spinal cord. Int. J . Deu. Neurosci., 7:343-35 7. Bovolenta, P., Liem, R.K.H., and Mason, C.A. (1984) Development of cerebellar astroglia: Transitions in form and cytoskeletal content. Deu. Biol., 102:248-259. Bullon, M.M., Alvarez-Gago, T., Fernandez-Ruiz, B., and Aguirre, C. (1984) Glial fibrillary acidic (GFA) protein in rat spinal cord. An immunoperoxidase study in semithin sections. Brain Res., 309~79-83. Carrer, H.F. and Aoki, A. (1982) Ultrastructural changes in the hypothalamic ventromedial nucleus of ovariectomized rats after estrogen treatment. Brain Res., 240:221-233. Dahl, D. (1981) The vimentin-GFA protein transition in the rat neuroglia cytoskeleton occurs at the time of myelination. J . Neurosci. Res., 6:741-748. Dahl, D., Rueger, D.C., Bignami, A,, Weber K., and Osborn, M. (1981) Vimentin, the 57,000 dalton protein of fibroblast filaments, is the major cytoskeletal component in immature glia. Eur. J . Cell Biol., 24~191-196. Dohler, K.D., Hancke, J.L., Srivastava, S.S., Hofmann, C., Shryne, J.E., and Gorski, R.A. (1984) Participation of estrogens in female sexual differentiation of the brain; neuroanatomical, neuroendocrine and behavioral evidence. Prog. Brain Res., 61:99-117. Fedoroff, S., White, R., Neal, J., Subrahmanyan, L., and Kalnins, V.I. (1983)Astrocyte cell lineage. 11. Mouse fibrous astrocytes and reactive astrocytes in cultures have vimentin and GFP containing intermediate filaments. Dew. Brain Res., 7:303-315. Garcia-Segura, L.M., Suarez, I., Segovia,S.,Tranque, P.A., Cales, J.M., Aguilera, P., Olmos, G., and Guillamon, A. (1988) The distribution of glial fibrillary acidic protein in the adult rat brain is influenced by the neonatal levels of sex steroids. Brain Res., 456:357-363. Garcia-Segura, L.M., Torres-Aleman, I., and Naftolin, F. (1989) Astrocytic shape and glial fibrillary acidic protein immunoreactivity are modified by estradiol in primary rat hypothalamic cultures. Deu. Brain Res., 47:298-302. Hanke, S. and Reichenbach, A. (1987) Quantitative-morphometric aspects of Bergmann glial (Golgi epithelial) cell development in rats. A Golgi study. Anat. Ernbryol., 177:183-188. Herndon, R.M. (1964) The fine structure of rat cerebellum. 11. The stellate neurones, granule cells and glia. J . Cell Biol., 23:277-293. Matsumoto, A. and Arai, Y. (1986) Male-female difference in synaptic organization of the ventromedial nucleus of the hypothalamus in the rat. Neuroendocrinology, 42:232-236. Meshul, C.K., Seil, F.J., and Herndon, R.M. (1987) Astrocytes play a role in regulation of synaptic density. Brain Res., 402:139-145. Montagnese, C., Poulain, D.A., Vincent, J.D., and Theodosis, D.T. (1988) Synaptic and neuronal-glial plasticity in the adult oxytocinergic system in response to physiological stimuli. Brain Res. Bull., 20:681-692. Palay, S.L., and Chan-Palay, V. (1974) Cerebellar Cortex: Cytology and Organization. Springer-Verlag, Berlin, pp. 288-321. Pixley, S.R.K. and De Vellis, J. (1984) Transition between immature radial glia and mature astrocytes studied with monoclonal antibody to vimentin. Deu. Brain Res., 15:201-209. Ramon y Cajal, S. (1911)Histologie du Systerne Nerueux de 1'Hornrne et des Vertebres, Vol. 11. Maloine, Paris, pp. 68-71. Schnitzer, J., Franke, W.W., and Schachner, M. (1981) Immunocytochemical demonstration of vimentin in astrocytes and ependymal cells of developing and adult nervous system. J. Cell Biol., 90:435447. Smith, S.S., Woodward, D.J., and Chapin, J.K. (1989) Sex steroids modulate motor-correlated increases in cerebellar discharge. Brain Res., 476:307-316. Suarez. I.. Fernandez, B.. Bodega. G., Tranaue, P., Olmos. G.. and Garcia-Segura, L.M.'(1987)Postnatal development of glial fibrillary acidic protein immunoreactivity in the hamster arcuate nucleus. Deu. Brain Res., 37:89-95. Tranque, P.A., Suarez, I., Olmos, G., Fernandez, B., and GarciaSegura, L.M. (1987) Estradiol-induced redistribution of glial fibrillary acidic protein immunoreactivity in the rat brain. Brain Res., 406:348-351. Weibel, E.R. (1979) Stereological Methods, Vol. I in Practical Methods for Biological Morphornetry. Academic Press, London. Weir, M.D., Patel, A.J., Hunt, A,, and Thomas, D.G.T. (1984) Developmental changes in the amount of glial fibrillary acidic protein in three regions of the rat brain. Deu. Brain Res., 15:147-154. Yen, S.H. and Fields, K.L. (1981) Antibodies to neurofilament, glial filament and fibroblast intermediate filament proteins bind to different cell types ofthe nervous sytem. J . Cell Biol., 88:115-126.
