- Email: [email protected]
Neuron and glial cells in neocortex after methylazoxymethanol treatment in early development
Mechanisms of Ageing and Development 100 (1998) 299 – 311
Neuron and glial cells in neocortex after methylazoxymethanol treatment in early development Sandra Ciaroni a,*, Ornella Buffi a, Patrizia Ambrogini b, Tiziana Cecchini a, Paolo Del Grande a a
Istituto di Scienze Morfologiche, Campus Scientifico Localita` Crocicchia, Uni6ersita` di Urbino, I-61029 Urbino, Italy b Istituto di Anatomia e Fisiologia, Uni6ersita` di Urbino, I-61029 Urbino, Italy Received 9 July 1997; received in revised form 27 October 1997; accepted 29 October 1997
Abstract The quantitative changes were investigated in neuron and glia density in the different cortical layers of the frontal cortex of 3 and 12 month old mice, exposed to methylazoxymethanol on embryonic day 13 (MAM13). No loss of cortical neurons was found between young and adult animals. MAM exposure on the 13th day of development induced a neuron density decrease throughout on the entire cortical depth and did not produce changes in the density of glial cells with respect to the controls and to age. Consequently, at 3 months of age we observe a glia/neuron ratio greater than that of controls and at 12 months a similar value. In the neocortex of MAM-mice at this numerical uniformity of glial cell density, did not correspond to a similar proportional composition: the frequency of the astrocytes is lower, adapting to the decreased neuron density; the greater oligodendrocyte percentage may be related to disturbed layering and to the hyperinnervation of the hypoplastic cortex; the microglia shows a trend similar to that of the controls. These results, together with those of other studies, suggest that prenatal exposure to MAM causes a cortical compensatory response regulating glial cells proliferation. © 1998 Elsevier Science Ireland Ltd. All rights reserved.
* Corresponding author. Tel.: +39 722 304270; fax: + 39 722 304242. 0047-6374/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 0 4 7 - 6 3 7 4 ( 9 7 ) 0 0 1 4 3 - 7
300
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Keywords: Neocortex development; Neurons; Glia; Methylazoxymethanol
1. Introduction Methylazoxymethanol (MAM) is a potent alkylating agent that kills dividing cells (Matsumoto et al., 1972) and has been shown to have neurotoxic effects when administered during the period of active neurogenesis in different species (Haddad et al., 1972). Micrencephaly, induced by prenatal injection of Methylazoxymethanol (Spatz and Laquer, 1968), has been used as a useful experimental model for the study of behavioral impairments, neuroanatomical and neurochemical abnomalies, by several investigators (Kabat et al., 1985; Virgili et al., 1989; Rabe and Lee, 1990; Lee and Rabe, 1992; Collier and Ashwell, 1993). Previous studies have observed the effects of prenatal exposure to MAM on neocortical lamination, showing characteristic reductions in cortical layers (Haddad et al., 1972; Dambska et al., 1982) and neuroanatomical disorganization in selective brain structures for given gestational days of administration and range of dose (Rodier, 1977; Johnston et al., 1981; Gardette et al., 1982; Johnston and Coyle, 1982; Yurkewicz et al., 1984). In particular as recently reported by Cattaneo et al. (1995), this drug ablates the cells undergoing the last mitosis while cells capable of additional divisions are spared. In fact, proliferating astrocytes are not affected. Neuron-astroglia interaction during development was also studied in MAM micrencephalic rats by GFA immunohistochemistry suggesting an astrocytic readjustment in the changed neuronal development (Eriksdotter-Nilson et al., 1986). Therefore MAM producing a selective ablation of precursor cells destined for different brain regions, was employed to study prenatal neurogenesis and CNS development in general. In earlier studies using this alkylating agent, we obtained selective elimination of cells which depended on the date of MAM injection in relation to the schedule of neurogenesis (Ciaroni et al., 1989, 1995). Moreover, when MAM was injected on the 15th embryonal day we noted a neuronal loss in the upper cortical layers, a decrease in neuron nucleus size not coupled with differences in glial cell density. In fact, the glia showed a trend similar to that of the controls. After MAM administration later in development (17th embryonal day), the changes in neuron nucleus size were paralleled by an increase in glial density, in particular in microglia, in all cortical layers, and the glia did not change from 3 to 12 months of age (Ciaroni et al., 1996). In this investigation, we examine neurons, glial cells and neuron nucleus size in neocortex after MAM injection during early development (13th embryonal day). When MAM was injected on the 13th embryonal day caused a severe cortical atrophy and an adjustment of the synaptic number and of the extent of neuropil to the induced hypoplasia was found (Ciaroni et al., 1992). This study was performed to provide further information about the possible cell compensatory responses that may follow an alteration of normal development and the possible involvement of spared neurons in controlling glia proliferation.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
301
2. Materials and methods
2.1. Animals Methylazoxymethanol acetate (MAM, Aldrich Chemical Co. St. Louis, MO) was injected into the peritoneum of three pregnant CD1 mice on the 13th day of gestation at the rate of 6 mg/kg in physiological solution (NaCl 0.9%). At the same time, three pregnant females were given a single intraperitoneal injection of physiological solution in order to use their young as controls. Two groups of mice aged 3 months and 12 months (n= 3 for each treatment group for each age) were used in this study. The animals, chosen at random from each litter were deeply anesthetized with sodium pentothal and transcardially perfused with 0.9% NaCl containing heparin, followed by a solution of 1.25% glutaraldehyde and 1% paraformaldehyde in phosphate buffer (0.1 M pH 7.4) at 4°C. The brains were removed, and blocks of neocortex were cut in the coronal plane, postfixed for 4–6 h in the same fixative and processed both for light and for electron microscopy.
2.2. Light microscopy The region of neocortex situated between the anterior commissure and the first appearance of the hippocampus in coronal semithin sections, was considered. It was taken medially in a part of the neocortex that corresponds to the area 6, according to Caviness (1975). The blocks were orientated on the microtome so that the semithin sections included the entire depth of the cortex, from the pial surface through to the white matter, in a vertical plane parallel to the lengths of the apical dendrites of the layer V pyramidal cells. Since the MAM administration affects the cortical layering, it was necessary to group cortical layers as layer I, layers II–IV together, layer V and layer VI. Thus, the above-mentioned orientation of blocks allowed us to identify the precise location of the cortical layers and to obtain thin sections of tissue blocks including either layers II–IV, or layer V, or layer VI. On the semithin sections, the density of neuron and glial cells were estimated using a Leitz microscope and camera lucida drawing tube to count the cell number in a sampling field of 100× 70 mm under a oil-immersion 100× objective. Cells intersected by the right vertical field bar were included in the count; those intersected by the left vertical bar were not. The upper edge of the field was placed parallel to the pial surface and lowered successively throughout the depth of the layer used for cell-count determination. The procedure was repeated through all cortical laminae (layer I excluded) until all neuron and glial cells in a 100-mm wide sample probe, extending from pia to white matter, had been analyzed. Five vertical strips were counted in each layer of each section examined. To avoid incorrect results, seeing as the total volume of neocortex changes, all the numerical estimates were converted into number of cells per 0.05 mm2 of cortical layers examined (in this study layer I has not been considered). Although all of our observations were concerned with relative cell density rather than with absolute density or cell number, we determined the area of neurons and glial cells in the neocortex of 3 and
302
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
12 month old mice in order to avoid over- or undersampling of cells due to size differences. Having not found any statistical difference in the cell area between the treatment groups (glia p B0.1; neurons p B0.2), no stereological correction for sampling errors was applied (Abercrombie, 1946). The cells were identified as neuron and glia according to the criteria of Ling et al. (1973) (Fig. 1). Glial counts included astrocytes, oligodendrocytes and microglia; other non neuronal cells were excluded.
2.3. Neuronal nucleus size measuring At least 150 neurons for each animal were used for nuclear-size measurements in the same semithin sections chosen for counting. These slices were chosen because the staining was optimal for visualizing the nuclear membrane. The outlines of neuronal nuclei from different cortical layers containing at least one nucleolus were drawn with a drawing tube at a final magnification of 985× . The microscopic stage was advanced laterally in each layer in steps until the final number of profiles was reached. From the drawing obtained, the mean nuclear area was calculated by means of OPTILAB software for image analysis.
2.4. Electron microscopy In order to count the three glial cell types separately, thin sections were used. Three blocks of cortex were selected for each of the three groups of layers from each animal (supragranular layers II and III, and granular layer IV, taken together;
Fig. 1. Semithin section from the frontal cortex of 3 month old MAM13 showing (N) neurons and (G) glia. Scale bar = 10 mm.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
303
layer V; layer VI). For each block, counts were made in a well-orientated ultrathin section revealing the entire layer depth. Thin sections were cut and carefully mounted on 150-mesh grids so that the depth of layer or layer groups was parallel to the grid bars. Each grid opening, approximately 112× 112 mm, served as the standard area for making counts of neuroglial cells. Since the areas of the openings often differed between different batches of grids, and in order to determine the area of the section that had been observed, after each grid had been counted the size of its openings was measured with a drawing tube. Counts were made on a row of grid openings starting in the upper openings. It was easy to orientate the layers and to distinguish the upper part of the layers II–IV because of the presence of the pial surface; the layer VI because of the underlying corpus callosum and the layer V because of the shape of the ultrathin section. In fact, all the blocks were cut so that the upper edge of the thin sections, corresponding to the upper part of layers, was shorter. Only astrocytes, oligodendrocytes and microglial cells containing nuclei were counted. Cells intersected by the right vertical and lower field bars were included in the count, whereas those intersected by the left vertical and upper bars were not. After a count had been made in the row containing the outermost part of layer, the row of openings immediately below was examined and the cells were counted until the entire depth of layer or group of layers had been covered. The three types of neuroglial cells were distinguished on the basis of their fine structure as frequently described in literature (Vaughan and Peters, 1974; Peters et al., 1976, 1991) (Fig. 2).
2.5. Statistics The results are expressed as mean9 S.E.M. The Chi-square test was applied in order to compare the proportions of the three types of glial cells and to compare the glia/neuron ratio. We used ANOVA test for comparison between the means after the assumptions of normality and homogeneity of variances were verified respectively by Kolmogorov-Smirnov and F-tests.
3. Results
3.1. Light microscopy 3.1.1. Neurons The counts of neurons and glial cells of cortical layers of controls and MAM injected mice are given in Table 1. As we have already observed in previous studies, MAM administration on the 13th embryonal day, when neocortical layering is not yet patterned, caused a significant decrease in neuronal density in all the cortical layers. No neuronal loss was found from 3 to 12 months of age both in controls and in treated mice; therefore, in MAM-injected mice the lower neuronal density persist with age. Two-way ANOVA showed no significant influence of age but a significant influence
33.3390.17 31.009 0.29 6.88 90.07 7.67 90.44
Neurons Control MAM13 Glia Control MAM13 8.55 9 0.28 8.18 9 0.13
34.67 9 0.67 31.77 9 0.23 8.00 9 0.50 8.78 9 0.64
28.00 90.29 22.83 91.17
3 months
3 months
12 months
Layer V
Layers II – IV
9.80 9 0.12 9.17 9 0.44
28.98 90.69 22.70 90.98
12 months
8.87 90.13 11.50 9 1.04
28.38 9 0.87 24.00 9 0.75
3 months
Layer VI
11.439 0.66 11.329 0.69
27.679 0.33 25.879 0.13
12 months
7.859 0.21 9.269 0.43
30.0190.18 25.969 0.55
3 months
9.82 9 0.31 9.56 9 0.08
30.649 0.38 26.71 9 0.37
12 months
Unit cortical column
Table 1 Values (mean 9S.E.M.) of the neuron and glial density in cortical layers and in a cortical column of 3 and 12 month old mice
304 S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
305
Fig. 2. Electron micrographs from the frontal cortex of 3 month old MAM13 showing the different types of glial cells. A: (O) oligodendrocyte lying next to a (N) neuron (10 000 × ). B: (M) microglia, (A) astrocyte (9600 ×).
of MAM treatment on the neuron density when comparison was made with controls (in all cortical layers and in the unit column pB 0.005).
3.1.2. Glial cells In all cortical layers there were no differences between control and treated mice of 3 months of age. Aging however showed a different effect on control and treated mice: while in control animals glial density increased significantly in all the cortical
306
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Table 2 Glia/neuron ratio within the cortical layers in treated and control mice of different ages Age (months) 3
12
Group
Control
MAM13
Control
MAM13
Layers II–IV Layer V Layer VI Unit cortical column
0.206 90.003 0.286 90.021 0.31090.01 0.256 90.002
0.243 9 0.014 0.383 9 0.008 0.473 9 0.029 0.356 9 0.008
0.246 9 0.013 0.340 9 0.011 0.410 90.01 0.323 9 0.01
0.260 9 0.005 0.406 9 0.02 0.436 9 0.026 0.356 9 0.001
layers between 3 and 12 months it increased slightly but not significantly in MAM13 with age. Two-way ANOVA revealed a significant influence of age on glial density in layers II – IV (p B 0.005), in layer V (pB 0.05) and in the total cortical depth (p B0.005); a significant interaction of age and MAM treatment in the cortical column was also showed (pB 0.025). As a consequence, in the controls, the glia/neuron ratio were different in the young with respect the adult (unit cortical column: chi-square = 8.120, d.f. = 1, pB0.005; layer VI: chi-square= 4.145, d.f.= 1, pB 0.05; layers II–IV: chisquare =3.603, d.f.= 1, layer V: chi-square= 3.721, d.f.= 1, 0.05 B pB 0.1), while in MAM13 no differences were observed between young and adult mice. Moreover, the glia/neuron ratio in the youngest MAM13 was significantly greater when compared to the age-matched controls (unit cortical column: chi-square= 9.506, d.f.=1, p B0.005; layer V: chi-square= 5.495, d.f.= 1, pB 0.025; layer VI: chisquare = 9.953, d.f.=1, p B 0.005), but very similar to that of the 12 month old controls (Table 2).
Table 3 Values (mean 9 S.E.M.) of the neuronal nucleus size in control and MAM mice Layers
II–IV
V
VI
Age (months)
3
12
3
12
3
12
Control
97.299 2.22
108.539 2.74
114.59 9 6.03
127.98 9 5.41
69.44 9 2.38
78.34 91.05
MAM13
98.389 2.23
109.90 9 8.23
120.26 9 3.02
137.26 9 9.91
70.06 9 1.75
79.31 9 4.77
Areas were obtained by averaging mean values of each animal.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
307
Table 4 Percentage of three types of neuroglial cells in control and treated mice Layers
II–IV
Age (months)
3
12
3
12
3
12
3
12
Astrocytes Control MAM13
80.4 67.9
81 70.5
74.7 72.2
72.2 62.4
73.6 71.3
69 66.2
75.7 70.6
74 65.9
Oligodendrocytes Control MAM13
9.8 22.3
7.7 16.8
18.6 18.2
14.3 20.8
19.8 19.3
21 20.5
16.9 19.8
14.5 19.7
9.8 9.8
11.3 12.6
6.7 9.6
13.5 16.8
6.5 9.4
10 13.3
7.4 9.6
11.5 14.4
Microglia Control MAM13
V
VI
Unit cortical column
3.1.3. Neuronal nucleus size measures Neuronal nucleus areas, expressed in mm2, are presented in Table 3. Statistical analysis revealed no significant differences between control and treated mice. However, nucleus size in old mice tended to be larger than that of young mice. Based on the mean areas shown in Table 3 neuron nucleus in older brains was 11– 13% larger than in younger brains. In fact, two-way ANOVA revealed significant differences between young and old animals (layers II–IV and V: pB0.05; layers VI: p B 0.025). 3.2. Electron microscopy The fine structure of the three types of neuroglial cells observed in the different treatment groups was similar. The analysis of proportion of each neuroglial cell type however showed some changes in treated mice with respect to the controls and to age (Table 4). When the different cortical layers were examined separately, the mean frequency of astrocytes was constantly lower in MAM13 than in the control mice both at 3 and 12 months of age in all cortical layers. The frequency of oligodendrocytes and microglia showed some variations among the treatment groups so that sometimes they have greater frequencies in treated mice, and sometimes similar to the controls. However, if we consider the entire depth of cortex we note a lower percentage of astrocytes and a greater percentage of oligodendrocytes and microglia in treated mice with respect to the controls both at 3 months and at 12 months. We applied log-linear showing an association between the different cellular types and treatment, and between the cellular types and age. On this basis, we used the chi-square test to examine the above-mentioned associations. The frequency of the three glial cell types was significantly different in the cortical layers considered together both at 3 months (chi-square= 6.416, d.f.= 2, p B 0.04) and at 12 months (chi-square = 16.444, d.f. = 2, pB 0.0002). In particular the change for astrocytes was highly significant at 3 months (chi-
308
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
square =6.189, d.f.=1, p B0.01) and at 12 months (chi-square= 14.855, d.f.= 1, pB 0.0001). The percentage of oligodendrocytes in treated mice at 3 months was slightly greater but this difference was not significant, whereas at 12 months it was (chi-square = 12.578, d.f. =1, p B 0.0003). Likewise the small increase observed in the percentage of microglia in MAM-mice was not significant. Moreover, the frequency of the three glia types changes significantly between 3 and 12 months both in controls (chi square = 7.212, d.f. = 2, pB 0.02) and in MAM-mice (chisquare = 14.991, d.f.=2, p B 0.0005). In the controls: the astrocytes and the oligodendrocytes remained unchanged, the microglia increased significantly during this time (chi-square =12.414, d.f.= 1, pB 0.0004). In treated mice, the astrocytes decreased (chi-square =4.445, d.f. = 1, pB 0.03); the oligodendrocytes remained similar and the increase of the microglia was highly significant (chi-square= 9.906, d.f.= 1, p B0.001).
4. Discussion In the present study, no neuronal loss was found from 3 to 12 months of age in control mice, in agreement with previous findings (Curcio and Coleman, 1982; Vincent et al., 1989; Tigges et al., 1990). The measurements of neuronal nucleus size indicates that nucleus size increase as the mice mature. This behavior may be caused by the need for supplying a more complex dendritic tree as Coleman and Flood (1987) and Flood and Coleman (1988) have already suggested. The glial cells increase during adult life in mice (Vaughan and Peters, 1974; Terry et al., 1987; Peters, 1991; Peinado et al., 1993). In fact, in agreement with data from the literature and our previous study, a greater number of glial cells, was found in older rather than in younger control mice. The ultrastructural analysis demonstrated that this increase in glial density with aging, is related to a significant increase in microglia frequency. On the 13 – 14th embryonal day the primordial plexiform zone undergoes a bilaminar subdivision with the emergence of the cortical plate therefore almost all precursors of neurons of cortical layers are affected by MAM injection on the 13th day of development. In fact the treatment induced neuron decrease throughout the entire cortex, which persists with age. Moreover, the glial cells are quantitatively similar to the controls in younger animals, and the increase observed in controls with age was not found in treated mice. Consequently, at 3 months of age, we observed a glia/neuron ratio greater than that of controls, and at 12 months treated mice and controls showed a similar glia/neuron ratio. It is interesting to note that the close resemblance of the glia/neuron ratio between 3 month old MAM13 and 12 month old controls was due to a decrease in the number of neurons in MAM-mice and to an increase in number of glial cells in older controls. The ultrastructural analysis showed that the frequency of astrocytes was lower with respect in the controls at 3 months of age and it further diminished at 12 months of age. The oligodendrocyte frequency resulted greater in some cortical layers and in the entire dept of the neocortex of the older MAM13. The microglia showed a trend similar to that observed in the controls.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
309
In a previous study, we showed a significantly greater numerical density of synapses and no significant differences in the average percentage of neuropil in MAM-mice injected on the 13th day of development (Ciaroni et al., 1992). Furthermore, according to the literature, these animals showed an altered mobility, and when they underwent learning tests such as the water-maze, they showed a learning deficit in the first test sessions, later they improved to the level of the controls (data not shown). Our present finding, coupled with previous studies which have reported a relative hyperinnervation of noradrenergic and cholinergic afferents to the cortex of MAM treated rats (Sanberg et al., 1987), are consistent with the hypothesis that the remaining neurons give an important compensatory repair response secondary to cell loss. This ability to remodel their dendrites arbor (McNeill and Koek, 1990) and to increase synaptic connections to compensate for occurring changes in the brain, plays an important role in maintaining a normal cortical synaptic neurochemistry. Measurements of level of neurotransmitters in the MAM-treated cortices demonstrated an increased concentration of norepinephrine, acetylcholine, serotonin and dopamin (Johnston and Coyle, 1979). A review by Lauder (1993) provides evidence that neurotransmitters such as the one mentioned above and others, regulate proliferation of non neuronal vertebrate cells and the findings on the excitatory neurotransmitter L-glutamate further strengthen the concept that neurotransmitters play roles in controlling glial cell number and development (Steinha¨user and Gallo, 1996). Most of glial cells are formed within the later proliferation phase and after birth and so their proliferation is not affected by MAM treatment. The greater glia/neuron ratio observed in MAM mice at 3 months of age can be interpreted as a compensatory response secondary to increased activity of neurons spared by MAM treatment. As regards the proportional composition of the glial population, the frontal cortex of MAM mice is somewhat different from that of the controls. In principle, the low frequency of astrocytes can be explained as an effect on radial glial cells caused by MAM exposure on the 13th embryonal day. In fact this cell class probably becomes a source of other cells of astroglial lineage. However, when the action of MAM was tested during embryonal development (Eriksdotter-Nilson et al., 1986) and in culture (Cattaneo et al., 1995), it was observed that astrocyte proliferation was not affected directly. Consequently, on the basis of our results, it might be speculated, that the development of astrocytes adjusts itself to the neuron number, that is their frequency is influenced by neuron reduction according to their function (Nedergaard, 1994). The higher frequency of oligodendrocytes in layers II– IV, can be believed to result from disturbed layering of treated cortices due to disruption of the radial glial scaffolding, which has been shown to provide a pathway for neuronal migration to the cortical plate (Gressens and Evrard, 1993). Moreover, this higher oligodendrocyte frequency is in agreement with the hyperinnervation quantified in the cerebral cortex of MAM animals in previous studies. The conclusion of the present study is that the neocortex appears as a plastic structure which responds to hypoplasia induced by MAM injection, producing a normal density of glial cells and a possible involvement of spared neurons in glial cell proliferation controlling may be suggested.
310
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Acknowledgement The authors would like to thank Dr Marco Oochi for statistics.
References Abercrombie, M., 1946. Estimation of nuclear populations from microtome sections. Anat. Rec. 94, 239–247. Cattaneo, E., Reinach, B., Caputi, A., Cattabeni, F., Di Luca, M., 1995. Selective in vitro blockade of neuroepithelial cells proliferation by methylazoxymethanol, a molecule capable of inducing long lasting functional impairments. J. Neurosci. Res. 41, 640 – 647. Caviness, V.S. Jr, 1975. Architectonic map of neocortex of the normal mouse. J. Comp. Neurol. 164, 247–264. Ciaroni, S., Cecchini, T., Gazzanelli, G., Del Grande, P., 1989. Methylazoxymethanol acetate (MAMac) effects on the ontogenesis of the mouse neocortex. J. Hirnforsch. 30, 699 – 705. Ciaroni, S., Cecchini, T., Ambrogini, P., Minelli, G., Del Grande, P., 1992. Nerve cell rearrangement in neocortical layers of MAM treated foetal mice. J. Hirnforsch. 33, 303 – 309. Ciaroni, S., Cecchini, T., Ambrogini, P., Buffi, O., Del Grande, P., 1995. Cell death and cell number in the developing cerebral cortex of MAM-treated mice. J. Brain Res. 2, 161 – 170. Ciaroni, S., Buffi, O., Ambrogini, P., Cecchini, T., Del Grande, P., 1996. Quantitative changes in neuron and glial cells of neocortex following prenatal exposure to methylazoxymethanol. J. Brain Res. 37, 537–546. Coleman, P.D., Flood, D.G., 1987. Neuron numbers and dendritic extent in normal aging and Alzheimer’s desease. Neurobiol. Aging 8, 521 – 545. Collier, P.A., Ashwell, K.W., 1993. Distribution of neuronal heterotopiae following prenatal exposure to methylazoxymethanol. Neurotoxicol. Teratol. 15, 439 – 444. Curcio, C.A., Coleman, P.D., 1982. Stability of neuron number in cortical barrels of aging mice. J. Comp. Neurol. 212, 158–172. Dambska, M., Haddad, R., Kozlowski, P.B., Lee, M.H., Shek, J., 1982. Telencephalic cytoarchitectonics in the brain of rats with graded degrees of micrencephaly. Acta Neuropathol. 58, 203 – 209. Eriksdotter-Nilson, M., Jonsson, G., Dahal, D., Bjorklund, H., 1986. Astroglial development in microencephalic rat brain after fetal methylazoxymethanol treatment. J. Dev. Neurosci. 4, 353 – 362. Flood, D.G., Coleman, P.D., 1988. Neuron numbers and sizes in aging brain: Comparisons of human, monkey, and rodent data. Neurobiol. Aging 9, 453 – 464. Gardette, R., Courtois, M., Bisconte, J., 1982. Prenatal development of mouse central nervous structures: Time of neuron origin and gradients of neuronal production. A autoradiographic study. J. Hirnforsch. 23, 415–443. Gressens, P., Evrard, P., 1993. The glial fascicle: an ontogenic and phylogenic unit guiding, supplying and distributing mammalian cortical neurons. Dev. Brain Res. 76, 272 – 277. Haddad, R.K., Rabe, A., Dumas, R., 1972. Comparison of effects of MAM-ac in brain development in different species. Fed. Proc. 31, 1520– 1523. Johnston, M.V., Coyle, J.T., 1979. Histological and neurochemical effects of fetal treatment with methylazoxymethanol on rat neocortex in adulthood. Brain Res. 170, 135 – 155. Johnston, M.V., Coyle, J.T., 1982. Cytotoxic lesions and the development of transmitters systems. Trends Neurosci. 5, 153–156. Johnston, M.V., Carman, A.B., Coyle, J.T., 1981. Effects of fetal treatment with methylazoxymethanol acetate at various gestational dates on the neurochemistry of the adult neocortex of the rat. J. Neurochem. 36, 124–128. Kabat, K., Buterbaugh, G.G., Eccles, C.U., 1985. Methylazoxymethanol as a developmental model of neurotoxicity. Neurobehav. Toxicol. Teratol. 7, 519 – 525.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
311
Lauder, J.M., 1993. Neurotransmitters as growth regulatory signals: role of receptors and second messengers. Trends Neurosci. 16, 233 – 240. Lee, M.H., Rabe, A., 1992. Premature decline in Morris water maze performance on aging micrencephalic rats. Neurotoxicol. Teratol. 14, 383 – 392. Ling, E.A., Paterson, J.A., Privat, A., Mori, S., Leblond, C.P., 1973. Investigation of glial cells in semithin sections. I. Identification of glial cells in the brain of young rats. J. Comp. Neurol. 149, 43–72. Matsumoto, H., Spatz, M., Laqueur, G.L., 1972. Quantitative changes with age in the DNA content of methylazoxymethanol-induced microencephalic rat brain. J. Neurochem. 19, 297 – 306. McNeill, H.T., Koek, L.L., 1990. Differential effects of advancing age on neurotransmitter cell loss in the substantia nigra and striatum of C57BL/6N mice. Brain Res. 521, 107 – 117. Nedergaard, M., 1994. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 263, 1768–1771. Peinado, M.A., Martinez, M., Pedrosa, J.A., Quesada, A., Peinado, J.M., 1993. Quantitative morphological changes in neurons and glia in the frontal lobe of the aging rat. Anat. Rec. 237, 104 – 108. Peters, A., Palay, S.L., Webster, H.D., 1976. The neuroglial cells. In: The Fine Structure of Nervous System, ch. VII. W.B. Saunders, Philadelphia, pp. 231 – 263. Peters, A., Josephson, K., Vincent, S.L., 1991. Effects of aging on the neuroglial cells and pericytes within area 17 of Rhesus monkey cerebral cortex. Anat Rec. 229, 384 – 398. Peters, A., 1991. Aging in Monkey cerebral cortex. In: Peters, A., Jones, E.G. (Eds.), Cerebral Cortex, vol. 9, ch. 11. Plenum Press New York, pp. 485 – 510. Rabe, A., Lee, M.H., 1990. Visual discrimination by rats with transplacentally induced micrencephaly. Neurotoxycol. Teratol. 12, 399–403. Rodier, P.M., 1977. Correlation between prenatally-induced alterations in CSN cell populations and post-natal function. Teratology 16, 235 – 246. Sanberg, P.R., Moran, T.H., Coyle, J.T., 1987. Microencephaly: cortical hypoplasia induced by methylazoxymethanol. In: Animal Models of Dementia. A.R. Liss, New York, pp. 253 – 278. Spatz, M., Laqueur, G.L., 1968. Transplacental chemical induction of microencephaly in two strains of rats. Proc. Soc. Exp. Biol. Med. 129, 705 – 710. Steinha¨user, C., Gallo, V., 1996. News on glutamate receptors in glial cells. Trends Neurosci. 19, 339–345. Terry, R.D., DeTeresa, R., Lawrence, A., Hansen, M.D., 1987. Neocortical cell counts in normal human adult aging. Ann. Neurol. 21, 530–539. Tigges, J., Herndon, J.G., Peters, A., 1990. Neuronal population of area 4 during the life span of the Rhesus Monkey. Neurobiol. Aging 11, 201 – 208. Vaughan, D.W., Peters, A., 1974. Neuroglial cells in the cerebral cortex of rats from young adulthood to old age: an electron microscope study. J. Neurocytol. 3, 405 – 429. Vincent, S.L., Peters, A., Tigges, J., 1989. Effects of aging on the neurons within area 17 of Rhesus monkey cerebral cortex. Anat. Rec. 223, 329 – 341. Virgili, M., Bernabei, O., Contestabile, A., 1989. Regional- and age-specific neurochemical alterations in rats rendered microencephalic by differentially timed gestational methylazoxymethanol treatment. Eur. J. Neurosci. 1, 647–653. Yurkewicz, L., Valentino, K.L., Floeter, M.K., Fleshman, J.W., Jones, E.G., 1984. Effects of cytotoxic deletions of somatic sensory cortex in fetal rats. Somatosens. Res. 1, 303 – 327.
.
Neuron and glial cells in neocortex after methylazoxymethanol treatment in early development Sandra Ciaroni a,*, Ornella Buffi a, Patrizia Ambrogini b, Tiziana Cecchini a, Paolo Del Grande a a
Istituto di Scienze Morfologiche, Campus Scientifico Localita` Crocicchia, Uni6ersita` di Urbino, I-61029 Urbino, Italy b Istituto di Anatomia e Fisiologia, Uni6ersita` di Urbino, I-61029 Urbino, Italy Received 9 July 1997; received in revised form 27 October 1997; accepted 29 October 1997
Abstract The quantitative changes were investigated in neuron and glia density in the different cortical layers of the frontal cortex of 3 and 12 month old mice, exposed to methylazoxymethanol on embryonic day 13 (MAM13). No loss of cortical neurons was found between young and adult animals. MAM exposure on the 13th day of development induced a neuron density decrease throughout on the entire cortical depth and did not produce changes in the density of glial cells with respect to the controls and to age. Consequently, at 3 months of age we observe a glia/neuron ratio greater than that of controls and at 12 months a similar value. In the neocortex of MAM-mice at this numerical uniformity of glial cell density, did not correspond to a similar proportional composition: the frequency of the astrocytes is lower, adapting to the decreased neuron density; the greater oligodendrocyte percentage may be related to disturbed layering and to the hyperinnervation of the hypoplastic cortex; the microglia shows a trend similar to that of the controls. These results, together with those of other studies, suggest that prenatal exposure to MAM causes a cortical compensatory response regulating glial cells proliferation. © 1998 Elsevier Science Ireland Ltd. All rights reserved.
* Corresponding author. Tel.: +39 722 304270; fax: + 39 722 304242. 0047-6374/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 0 4 7 - 6 3 7 4 ( 9 7 ) 0 0 1 4 3 - 7
300
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Keywords: Neocortex development; Neurons; Glia; Methylazoxymethanol
1. Introduction Methylazoxymethanol (MAM) is a potent alkylating agent that kills dividing cells (Matsumoto et al., 1972) and has been shown to have neurotoxic effects when administered during the period of active neurogenesis in different species (Haddad et al., 1972). Micrencephaly, induced by prenatal injection of Methylazoxymethanol (Spatz and Laquer, 1968), has been used as a useful experimental model for the study of behavioral impairments, neuroanatomical and neurochemical abnomalies, by several investigators (Kabat et al., 1985; Virgili et al., 1989; Rabe and Lee, 1990; Lee and Rabe, 1992; Collier and Ashwell, 1993). Previous studies have observed the effects of prenatal exposure to MAM on neocortical lamination, showing characteristic reductions in cortical layers (Haddad et al., 1972; Dambska et al., 1982) and neuroanatomical disorganization in selective brain structures for given gestational days of administration and range of dose (Rodier, 1977; Johnston et al., 1981; Gardette et al., 1982; Johnston and Coyle, 1982; Yurkewicz et al., 1984). In particular as recently reported by Cattaneo et al. (1995), this drug ablates the cells undergoing the last mitosis while cells capable of additional divisions are spared. In fact, proliferating astrocytes are not affected. Neuron-astroglia interaction during development was also studied in MAM micrencephalic rats by GFA immunohistochemistry suggesting an astrocytic readjustment in the changed neuronal development (Eriksdotter-Nilson et al., 1986). Therefore MAM producing a selective ablation of precursor cells destined for different brain regions, was employed to study prenatal neurogenesis and CNS development in general. In earlier studies using this alkylating agent, we obtained selective elimination of cells which depended on the date of MAM injection in relation to the schedule of neurogenesis (Ciaroni et al., 1989, 1995). Moreover, when MAM was injected on the 15th embryonal day we noted a neuronal loss in the upper cortical layers, a decrease in neuron nucleus size not coupled with differences in glial cell density. In fact, the glia showed a trend similar to that of the controls. After MAM administration later in development (17th embryonal day), the changes in neuron nucleus size were paralleled by an increase in glial density, in particular in microglia, in all cortical layers, and the glia did not change from 3 to 12 months of age (Ciaroni et al., 1996). In this investigation, we examine neurons, glial cells and neuron nucleus size in neocortex after MAM injection during early development (13th embryonal day). When MAM was injected on the 13th embryonal day caused a severe cortical atrophy and an adjustment of the synaptic number and of the extent of neuropil to the induced hypoplasia was found (Ciaroni et al., 1992). This study was performed to provide further information about the possible cell compensatory responses that may follow an alteration of normal development and the possible involvement of spared neurons in controlling glia proliferation.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
301
2. Materials and methods
2.1. Animals Methylazoxymethanol acetate (MAM, Aldrich Chemical Co. St. Louis, MO) was injected into the peritoneum of three pregnant CD1 mice on the 13th day of gestation at the rate of 6 mg/kg in physiological solution (NaCl 0.9%). At the same time, three pregnant females were given a single intraperitoneal injection of physiological solution in order to use their young as controls. Two groups of mice aged 3 months and 12 months (n= 3 for each treatment group for each age) were used in this study. The animals, chosen at random from each litter were deeply anesthetized with sodium pentothal and transcardially perfused with 0.9% NaCl containing heparin, followed by a solution of 1.25% glutaraldehyde and 1% paraformaldehyde in phosphate buffer (0.1 M pH 7.4) at 4°C. The brains were removed, and blocks of neocortex were cut in the coronal plane, postfixed for 4–6 h in the same fixative and processed both for light and for electron microscopy.
2.2. Light microscopy The region of neocortex situated between the anterior commissure and the first appearance of the hippocampus in coronal semithin sections, was considered. It was taken medially in a part of the neocortex that corresponds to the area 6, according to Caviness (1975). The blocks were orientated on the microtome so that the semithin sections included the entire depth of the cortex, from the pial surface through to the white matter, in a vertical plane parallel to the lengths of the apical dendrites of the layer V pyramidal cells. Since the MAM administration affects the cortical layering, it was necessary to group cortical layers as layer I, layers II–IV together, layer V and layer VI. Thus, the above-mentioned orientation of blocks allowed us to identify the precise location of the cortical layers and to obtain thin sections of tissue blocks including either layers II–IV, or layer V, or layer VI. On the semithin sections, the density of neuron and glial cells were estimated using a Leitz microscope and camera lucida drawing tube to count the cell number in a sampling field of 100× 70 mm under a oil-immersion 100× objective. Cells intersected by the right vertical field bar were included in the count; those intersected by the left vertical bar were not. The upper edge of the field was placed parallel to the pial surface and lowered successively throughout the depth of the layer used for cell-count determination. The procedure was repeated through all cortical laminae (layer I excluded) until all neuron and glial cells in a 100-mm wide sample probe, extending from pia to white matter, had been analyzed. Five vertical strips were counted in each layer of each section examined. To avoid incorrect results, seeing as the total volume of neocortex changes, all the numerical estimates were converted into number of cells per 0.05 mm2 of cortical layers examined (in this study layer I has not been considered). Although all of our observations were concerned with relative cell density rather than with absolute density or cell number, we determined the area of neurons and glial cells in the neocortex of 3 and
302
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
12 month old mice in order to avoid over- or undersampling of cells due to size differences. Having not found any statistical difference in the cell area between the treatment groups (glia p B0.1; neurons p B0.2), no stereological correction for sampling errors was applied (Abercrombie, 1946). The cells were identified as neuron and glia according to the criteria of Ling et al. (1973) (Fig. 1). Glial counts included astrocytes, oligodendrocytes and microglia; other non neuronal cells were excluded.
2.3. Neuronal nucleus size measuring At least 150 neurons for each animal were used for nuclear-size measurements in the same semithin sections chosen for counting. These slices were chosen because the staining was optimal for visualizing the nuclear membrane. The outlines of neuronal nuclei from different cortical layers containing at least one nucleolus were drawn with a drawing tube at a final magnification of 985× . The microscopic stage was advanced laterally in each layer in steps until the final number of profiles was reached. From the drawing obtained, the mean nuclear area was calculated by means of OPTILAB software for image analysis.
2.4. Electron microscopy In order to count the three glial cell types separately, thin sections were used. Three blocks of cortex were selected for each of the three groups of layers from each animal (supragranular layers II and III, and granular layer IV, taken together;
Fig. 1. Semithin section from the frontal cortex of 3 month old MAM13 showing (N) neurons and (G) glia. Scale bar = 10 mm.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
303
layer V; layer VI). For each block, counts were made in a well-orientated ultrathin section revealing the entire layer depth. Thin sections were cut and carefully mounted on 150-mesh grids so that the depth of layer or layer groups was parallel to the grid bars. Each grid opening, approximately 112× 112 mm, served as the standard area for making counts of neuroglial cells. Since the areas of the openings often differed between different batches of grids, and in order to determine the area of the section that had been observed, after each grid had been counted the size of its openings was measured with a drawing tube. Counts were made on a row of grid openings starting in the upper openings. It was easy to orientate the layers and to distinguish the upper part of the layers II–IV because of the presence of the pial surface; the layer VI because of the underlying corpus callosum and the layer V because of the shape of the ultrathin section. In fact, all the blocks were cut so that the upper edge of the thin sections, corresponding to the upper part of layers, was shorter. Only astrocytes, oligodendrocytes and microglial cells containing nuclei were counted. Cells intersected by the right vertical and lower field bars were included in the count, whereas those intersected by the left vertical and upper bars were not. After a count had been made in the row containing the outermost part of layer, the row of openings immediately below was examined and the cells were counted until the entire depth of layer or group of layers had been covered. The three types of neuroglial cells were distinguished on the basis of their fine structure as frequently described in literature (Vaughan and Peters, 1974; Peters et al., 1976, 1991) (Fig. 2).
2.5. Statistics The results are expressed as mean9 S.E.M. The Chi-square test was applied in order to compare the proportions of the three types of glial cells and to compare the glia/neuron ratio. We used ANOVA test for comparison between the means after the assumptions of normality and homogeneity of variances were verified respectively by Kolmogorov-Smirnov and F-tests.
3. Results
3.1. Light microscopy 3.1.1. Neurons The counts of neurons and glial cells of cortical layers of controls and MAM injected mice are given in Table 1. As we have already observed in previous studies, MAM administration on the 13th embryonal day, when neocortical layering is not yet patterned, caused a significant decrease in neuronal density in all the cortical layers. No neuronal loss was found from 3 to 12 months of age both in controls and in treated mice; therefore, in MAM-injected mice the lower neuronal density persist with age. Two-way ANOVA showed no significant influence of age but a significant influence
33.3390.17 31.009 0.29 6.88 90.07 7.67 90.44
Neurons Control MAM13 Glia Control MAM13 8.55 9 0.28 8.18 9 0.13
34.67 9 0.67 31.77 9 0.23 8.00 9 0.50 8.78 9 0.64
28.00 90.29 22.83 91.17
3 months
3 months
12 months
Layer V
Layers II – IV
9.80 9 0.12 9.17 9 0.44
28.98 90.69 22.70 90.98
12 months
8.87 90.13 11.50 9 1.04
28.38 9 0.87 24.00 9 0.75
3 months
Layer VI
11.439 0.66 11.329 0.69
27.679 0.33 25.879 0.13
12 months
7.859 0.21 9.269 0.43
30.0190.18 25.969 0.55
3 months
9.82 9 0.31 9.56 9 0.08
30.649 0.38 26.71 9 0.37
12 months
Unit cortical column
Table 1 Values (mean 9S.E.M.) of the neuron and glial density in cortical layers and in a cortical column of 3 and 12 month old mice
304 S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
305
Fig. 2. Electron micrographs from the frontal cortex of 3 month old MAM13 showing the different types of glial cells. A: (O) oligodendrocyte lying next to a (N) neuron (10 000 × ). B: (M) microglia, (A) astrocyte (9600 ×).
of MAM treatment on the neuron density when comparison was made with controls (in all cortical layers and in the unit column pB 0.005).
3.1.2. Glial cells In all cortical layers there were no differences between control and treated mice of 3 months of age. Aging however showed a different effect on control and treated mice: while in control animals glial density increased significantly in all the cortical
306
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Table 2 Glia/neuron ratio within the cortical layers in treated and control mice of different ages Age (months) 3
12
Group
Control
MAM13
Control
MAM13
Layers II–IV Layer V Layer VI Unit cortical column
0.206 90.003 0.286 90.021 0.31090.01 0.256 90.002
0.243 9 0.014 0.383 9 0.008 0.473 9 0.029 0.356 9 0.008
0.246 9 0.013 0.340 9 0.011 0.410 90.01 0.323 9 0.01
0.260 9 0.005 0.406 9 0.02 0.436 9 0.026 0.356 9 0.001
layers between 3 and 12 months it increased slightly but not significantly in MAM13 with age. Two-way ANOVA revealed a significant influence of age on glial density in layers II – IV (p B 0.005), in layer V (pB 0.05) and in the total cortical depth (p B0.005); a significant interaction of age and MAM treatment in the cortical column was also showed (pB 0.025). As a consequence, in the controls, the glia/neuron ratio were different in the young with respect the adult (unit cortical column: chi-square = 8.120, d.f. = 1, pB0.005; layer VI: chi-square= 4.145, d.f.= 1, pB 0.05; layers II–IV: chisquare =3.603, d.f.= 1, layer V: chi-square= 3.721, d.f.= 1, 0.05 B pB 0.1), while in MAM13 no differences were observed between young and adult mice. Moreover, the glia/neuron ratio in the youngest MAM13 was significantly greater when compared to the age-matched controls (unit cortical column: chi-square= 9.506, d.f.=1, p B0.005; layer V: chi-square= 5.495, d.f.= 1, pB 0.025; layer VI: chisquare = 9.953, d.f.=1, p B 0.005), but very similar to that of the 12 month old controls (Table 2).
Table 3 Values (mean 9 S.E.M.) of the neuronal nucleus size in control and MAM mice Layers
II–IV
V
VI
Age (months)
3
12
3
12
3
12
Control
97.299 2.22
108.539 2.74
114.59 9 6.03
127.98 9 5.41
69.44 9 2.38
78.34 91.05
MAM13
98.389 2.23
109.90 9 8.23
120.26 9 3.02
137.26 9 9.91
70.06 9 1.75
79.31 9 4.77
Areas were obtained by averaging mean values of each animal.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
307
Table 4 Percentage of three types of neuroglial cells in control and treated mice Layers
II–IV
Age (months)
3
12
3
12
3
12
3
12
Astrocytes Control MAM13
80.4 67.9
81 70.5
74.7 72.2
72.2 62.4
73.6 71.3
69 66.2
75.7 70.6
74 65.9
Oligodendrocytes Control MAM13
9.8 22.3
7.7 16.8
18.6 18.2
14.3 20.8
19.8 19.3
21 20.5
16.9 19.8
14.5 19.7
9.8 9.8
11.3 12.6
6.7 9.6
13.5 16.8
6.5 9.4
10 13.3
7.4 9.6
11.5 14.4
Microglia Control MAM13
V
VI
Unit cortical column
3.1.3. Neuronal nucleus size measures Neuronal nucleus areas, expressed in mm2, are presented in Table 3. Statistical analysis revealed no significant differences between control and treated mice. However, nucleus size in old mice tended to be larger than that of young mice. Based on the mean areas shown in Table 3 neuron nucleus in older brains was 11– 13% larger than in younger brains. In fact, two-way ANOVA revealed significant differences between young and old animals (layers II–IV and V: pB0.05; layers VI: p B 0.025). 3.2. Electron microscopy The fine structure of the three types of neuroglial cells observed in the different treatment groups was similar. The analysis of proportion of each neuroglial cell type however showed some changes in treated mice with respect to the controls and to age (Table 4). When the different cortical layers were examined separately, the mean frequency of astrocytes was constantly lower in MAM13 than in the control mice both at 3 and 12 months of age in all cortical layers. The frequency of oligodendrocytes and microglia showed some variations among the treatment groups so that sometimes they have greater frequencies in treated mice, and sometimes similar to the controls. However, if we consider the entire depth of cortex we note a lower percentage of astrocytes and a greater percentage of oligodendrocytes and microglia in treated mice with respect to the controls both at 3 months and at 12 months. We applied log-linear showing an association between the different cellular types and treatment, and between the cellular types and age. On this basis, we used the chi-square test to examine the above-mentioned associations. The frequency of the three glial cell types was significantly different in the cortical layers considered together both at 3 months (chi-square= 6.416, d.f.= 2, p B 0.04) and at 12 months (chi-square = 16.444, d.f. = 2, pB 0.0002). In particular the change for astrocytes was highly significant at 3 months (chi-
308
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
square =6.189, d.f.=1, p B0.01) and at 12 months (chi-square= 14.855, d.f.= 1, pB 0.0001). The percentage of oligodendrocytes in treated mice at 3 months was slightly greater but this difference was not significant, whereas at 12 months it was (chi-square = 12.578, d.f. =1, p B 0.0003). Likewise the small increase observed in the percentage of microglia in MAM-mice was not significant. Moreover, the frequency of the three glia types changes significantly between 3 and 12 months both in controls (chi square = 7.212, d.f. = 2, pB 0.02) and in MAM-mice (chisquare = 14.991, d.f.=2, p B 0.0005). In the controls: the astrocytes and the oligodendrocytes remained unchanged, the microglia increased significantly during this time (chi-square =12.414, d.f.= 1, pB 0.0004). In treated mice, the astrocytes decreased (chi-square =4.445, d.f. = 1, pB 0.03); the oligodendrocytes remained similar and the increase of the microglia was highly significant (chi-square= 9.906, d.f.= 1, p B0.001).
4. Discussion In the present study, no neuronal loss was found from 3 to 12 months of age in control mice, in agreement with previous findings (Curcio and Coleman, 1982; Vincent et al., 1989; Tigges et al., 1990). The measurements of neuronal nucleus size indicates that nucleus size increase as the mice mature. This behavior may be caused by the need for supplying a more complex dendritic tree as Coleman and Flood (1987) and Flood and Coleman (1988) have already suggested. The glial cells increase during adult life in mice (Vaughan and Peters, 1974; Terry et al., 1987; Peters, 1991; Peinado et al., 1993). In fact, in agreement with data from the literature and our previous study, a greater number of glial cells, was found in older rather than in younger control mice. The ultrastructural analysis demonstrated that this increase in glial density with aging, is related to a significant increase in microglia frequency. On the 13 – 14th embryonal day the primordial plexiform zone undergoes a bilaminar subdivision with the emergence of the cortical plate therefore almost all precursors of neurons of cortical layers are affected by MAM injection on the 13th day of development. In fact the treatment induced neuron decrease throughout the entire cortex, which persists with age. Moreover, the glial cells are quantitatively similar to the controls in younger animals, and the increase observed in controls with age was not found in treated mice. Consequently, at 3 months of age, we observed a glia/neuron ratio greater than that of controls, and at 12 months treated mice and controls showed a similar glia/neuron ratio. It is interesting to note that the close resemblance of the glia/neuron ratio between 3 month old MAM13 and 12 month old controls was due to a decrease in the number of neurons in MAM-mice and to an increase in number of glial cells in older controls. The ultrastructural analysis showed that the frequency of astrocytes was lower with respect in the controls at 3 months of age and it further diminished at 12 months of age. The oligodendrocyte frequency resulted greater in some cortical layers and in the entire dept of the neocortex of the older MAM13. The microglia showed a trend similar to that observed in the controls.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
309
In a previous study, we showed a significantly greater numerical density of synapses and no significant differences in the average percentage of neuropil in MAM-mice injected on the 13th day of development (Ciaroni et al., 1992). Furthermore, according to the literature, these animals showed an altered mobility, and when they underwent learning tests such as the water-maze, they showed a learning deficit in the first test sessions, later they improved to the level of the controls (data not shown). Our present finding, coupled with previous studies which have reported a relative hyperinnervation of noradrenergic and cholinergic afferents to the cortex of MAM treated rats (Sanberg et al., 1987), are consistent with the hypothesis that the remaining neurons give an important compensatory repair response secondary to cell loss. This ability to remodel their dendrites arbor (McNeill and Koek, 1990) and to increase synaptic connections to compensate for occurring changes in the brain, plays an important role in maintaining a normal cortical synaptic neurochemistry. Measurements of level of neurotransmitters in the MAM-treated cortices demonstrated an increased concentration of norepinephrine, acetylcholine, serotonin and dopamin (Johnston and Coyle, 1979). A review by Lauder (1993) provides evidence that neurotransmitters such as the one mentioned above and others, regulate proliferation of non neuronal vertebrate cells and the findings on the excitatory neurotransmitter L-glutamate further strengthen the concept that neurotransmitters play roles in controlling glial cell number and development (Steinha¨user and Gallo, 1996). Most of glial cells are formed within the later proliferation phase and after birth and so their proliferation is not affected by MAM treatment. The greater glia/neuron ratio observed in MAM mice at 3 months of age can be interpreted as a compensatory response secondary to increased activity of neurons spared by MAM treatment. As regards the proportional composition of the glial population, the frontal cortex of MAM mice is somewhat different from that of the controls. In principle, the low frequency of astrocytes can be explained as an effect on radial glial cells caused by MAM exposure on the 13th embryonal day. In fact this cell class probably becomes a source of other cells of astroglial lineage. However, when the action of MAM was tested during embryonal development (Eriksdotter-Nilson et al., 1986) and in culture (Cattaneo et al., 1995), it was observed that astrocyte proliferation was not affected directly. Consequently, on the basis of our results, it might be speculated, that the development of astrocytes adjusts itself to the neuron number, that is their frequency is influenced by neuron reduction according to their function (Nedergaard, 1994). The higher frequency of oligodendrocytes in layers II– IV, can be believed to result from disturbed layering of treated cortices due to disruption of the radial glial scaffolding, which has been shown to provide a pathway for neuronal migration to the cortical plate (Gressens and Evrard, 1993). Moreover, this higher oligodendrocyte frequency is in agreement with the hyperinnervation quantified in the cerebral cortex of MAM animals in previous studies. The conclusion of the present study is that the neocortex appears as a plastic structure which responds to hypoplasia induced by MAM injection, producing a normal density of glial cells and a possible involvement of spared neurons in glial cell proliferation controlling may be suggested.
310
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
Acknowledgement The authors would like to thank Dr Marco Oochi for statistics.
References Abercrombie, M., 1946. Estimation of nuclear populations from microtome sections. Anat. Rec. 94, 239–247. Cattaneo, E., Reinach, B., Caputi, A., Cattabeni, F., Di Luca, M., 1995. Selective in vitro blockade of neuroepithelial cells proliferation by methylazoxymethanol, a molecule capable of inducing long lasting functional impairments. J. Neurosci. Res. 41, 640 – 647. Caviness, V.S. Jr, 1975. Architectonic map of neocortex of the normal mouse. J. Comp. Neurol. 164, 247–264. Ciaroni, S., Cecchini, T., Gazzanelli, G., Del Grande, P., 1989. Methylazoxymethanol acetate (MAMac) effects on the ontogenesis of the mouse neocortex. J. Hirnforsch. 30, 699 – 705. Ciaroni, S., Cecchini, T., Ambrogini, P., Minelli, G., Del Grande, P., 1992. Nerve cell rearrangement in neocortical layers of MAM treated foetal mice. J. Hirnforsch. 33, 303 – 309. Ciaroni, S., Cecchini, T., Ambrogini, P., Buffi, O., Del Grande, P., 1995. Cell death and cell number in the developing cerebral cortex of MAM-treated mice. J. Brain Res. 2, 161 – 170. Ciaroni, S., Buffi, O., Ambrogini, P., Cecchini, T., Del Grande, P., 1996. Quantitative changes in neuron and glial cells of neocortex following prenatal exposure to methylazoxymethanol. J. Brain Res. 37, 537–546. Coleman, P.D., Flood, D.G., 1987. Neuron numbers and dendritic extent in normal aging and Alzheimer’s desease. Neurobiol. Aging 8, 521 – 545. Collier, P.A., Ashwell, K.W., 1993. Distribution of neuronal heterotopiae following prenatal exposure to methylazoxymethanol. Neurotoxicol. Teratol. 15, 439 – 444. Curcio, C.A., Coleman, P.D., 1982. Stability of neuron number in cortical barrels of aging mice. J. Comp. Neurol. 212, 158–172. Dambska, M., Haddad, R., Kozlowski, P.B., Lee, M.H., Shek, J., 1982. Telencephalic cytoarchitectonics in the brain of rats with graded degrees of micrencephaly. Acta Neuropathol. 58, 203 – 209. Eriksdotter-Nilson, M., Jonsson, G., Dahal, D., Bjorklund, H., 1986. Astroglial development in microencephalic rat brain after fetal methylazoxymethanol treatment. J. Dev. Neurosci. 4, 353 – 362. Flood, D.G., Coleman, P.D., 1988. Neuron numbers and sizes in aging brain: Comparisons of human, monkey, and rodent data. Neurobiol. Aging 9, 453 – 464. Gardette, R., Courtois, M., Bisconte, J., 1982. Prenatal development of mouse central nervous structures: Time of neuron origin and gradients of neuronal production. A autoradiographic study. J. Hirnforsch. 23, 415–443. Gressens, P., Evrard, P., 1993. The glial fascicle: an ontogenic and phylogenic unit guiding, supplying and distributing mammalian cortical neurons. Dev. Brain Res. 76, 272 – 277. Haddad, R.K., Rabe, A., Dumas, R., 1972. Comparison of effects of MAM-ac in brain development in different species. Fed. Proc. 31, 1520– 1523. Johnston, M.V., Coyle, J.T., 1979. Histological and neurochemical effects of fetal treatment with methylazoxymethanol on rat neocortex in adulthood. Brain Res. 170, 135 – 155. Johnston, M.V., Coyle, J.T., 1982. Cytotoxic lesions and the development of transmitters systems. Trends Neurosci. 5, 153–156. Johnston, M.V., Carman, A.B., Coyle, J.T., 1981. Effects of fetal treatment with methylazoxymethanol acetate at various gestational dates on the neurochemistry of the adult neocortex of the rat. J. Neurochem. 36, 124–128. Kabat, K., Buterbaugh, G.G., Eccles, C.U., 1985. Methylazoxymethanol as a developmental model of neurotoxicity. Neurobehav. Toxicol. Teratol. 7, 519 – 525.
S. Ciaroni et al. / Mechanisms of Ageing and De6elopment 100 (1998) 299–311
311
Lauder, J.M., 1993. Neurotransmitters as growth regulatory signals: role of receptors and second messengers. Trends Neurosci. 16, 233 – 240. Lee, M.H., Rabe, A., 1992. Premature decline in Morris water maze performance on aging micrencephalic rats. Neurotoxicol. Teratol. 14, 383 – 392. Ling, E.A., Paterson, J.A., Privat, A., Mori, S., Leblond, C.P., 1973. Investigation of glial cells in semithin sections. I. Identification of glial cells in the brain of young rats. J. Comp. Neurol. 149, 43–72. Matsumoto, H., Spatz, M., Laqueur, G.L., 1972. Quantitative changes with age in the DNA content of methylazoxymethanol-induced microencephalic rat brain. J. Neurochem. 19, 297 – 306. McNeill, H.T., Koek, L.L., 1990. Differential effects of advancing age on neurotransmitter cell loss in the substantia nigra and striatum of C57BL/6N mice. Brain Res. 521, 107 – 117. Nedergaard, M., 1994. Direct signaling from astrocytes to neurons in cultures of mammalian brain cells. Science 263, 1768–1771. Peinado, M.A., Martinez, M., Pedrosa, J.A., Quesada, A., Peinado, J.M., 1993. Quantitative morphological changes in neurons and glia in the frontal lobe of the aging rat. Anat. Rec. 237, 104 – 108. Peters, A., Palay, S.L., Webster, H.D., 1976. The neuroglial cells. In: The Fine Structure of Nervous System, ch. VII. W.B. Saunders, Philadelphia, pp. 231 – 263. Peters, A., Josephson, K., Vincent, S.L., 1991. Effects of aging on the neuroglial cells and pericytes within area 17 of Rhesus monkey cerebral cortex. Anat Rec. 229, 384 – 398. Peters, A., 1991. Aging in Monkey cerebral cortex. In: Peters, A., Jones, E.G. (Eds.), Cerebral Cortex, vol. 9, ch. 11. Plenum Press New York, pp. 485 – 510. Rabe, A., Lee, M.H., 1990. Visual discrimination by rats with transplacentally induced micrencephaly. Neurotoxycol. Teratol. 12, 399–403. Rodier, P.M., 1977. Correlation between prenatally-induced alterations in CSN cell populations and post-natal function. Teratology 16, 235 – 246. Sanberg, P.R., Moran, T.H., Coyle, J.T., 1987. Microencephaly: cortical hypoplasia induced by methylazoxymethanol. In: Animal Models of Dementia. A.R. Liss, New York, pp. 253 – 278. Spatz, M., Laqueur, G.L., 1968. Transplacental chemical induction of microencephaly in two strains of rats. Proc. Soc. Exp. Biol. Med. 129, 705 – 710. Steinha¨user, C., Gallo, V., 1996. News on glutamate receptors in glial cells. Trends Neurosci. 19, 339–345. Terry, R.D., DeTeresa, R., Lawrence, A., Hansen, M.D., 1987. Neocortical cell counts in normal human adult aging. Ann. Neurol. 21, 530–539. Tigges, J., Herndon, J.G., Peters, A., 1990. Neuronal population of area 4 during the life span of the Rhesus Monkey. Neurobiol. Aging 11, 201 – 208. Vaughan, D.W., Peters, A., 1974. Neuroglial cells in the cerebral cortex of rats from young adulthood to old age: an electron microscope study. J. Neurocytol. 3, 405 – 429. Vincent, S.L., Peters, A., Tigges, J., 1989. Effects of aging on the neurons within area 17 of Rhesus monkey cerebral cortex. Anat. Rec. 223, 329 – 341. Virgili, M., Bernabei, O., Contestabile, A., 1989. Regional- and age-specific neurochemical alterations in rats rendered microencephalic by differentially timed gestational methylazoxymethanol treatment. Eur. J. Neurosci. 1, 647–653. Yurkewicz, L., Valentino, K.L., Floeter, M.K., Fleshman, J.W., Jones, E.G., 1984. Effects of cytotoxic deletions of somatic sensory cortex in fetal rats. Somatosens. Res. 1, 303 – 327.
.