We have previously demonstrated that the antiproliferative agent methylazoxymethanol acetate (MAM) is able to induce in rats cerebral heterotopia that share striking similarities with those observed in human periventricular nodular heterotopia (PNH), a cerebral dysgenesis frequently observed in human patients affected by drug-resistant focal epilepsy. In this study, we investigated the time-course of neurogenesis in the cerebral heterotopia of MAM-treated rats, with the idea of understanding why PNH develop in human patients. For these goals, we analyzed the cytoarchitectural features, the time of neurogenesis and the cellular phenotype of the heterotopia, by means of BrdU immunocytochemistry and confocal immunofluorescence experiments. Our data demonstrate that the different types of heterotopia in MAM-treated rats are formed through the same altered neurogenetic process, which follows quite organized neurogenetic gradients. The MAM-induced ablation of an early wave of cortical neurons is sufficient to alter per se the migration and differentiation of subsequently generated neurons, which in turn set the base for the formation of the different heterotopic structures. The neurogenesis of MAM-induced heterotopia may explain the origin and intrinsic epileptogenicity of periventricular nodular heterotopia in human patients.
Cerebral dysgeneses are developmental malformations of the brain structure determined by impairment of the spatially organized and time-regulated processes of neurogenesis, migration, neuron/glia interaction and cell differentiation, which take place in the developing brain. In recent years, it has been increasingly recognized that cerebral dysgeneses are a relevant cause of mental and neurological deficits in humans, including epilepsy (Galaburda et al., 1985; Kuzniecky et al., 1993; Barkovich et al., 1994). In particular, human patients with drug-resistant focal epilepsy are frequently (in ~20–40% of cases) affected by developmental malformations of the cortex and they frequently undergo epilepsy surgery for the relief of their intractable seizures (Hardiman et al., 1988; Meencke and Veith, 1992; Mischel et al., 1995).
The mechanisms leading to the genesis of the mature brain have been extensively investigated and the processes of neurogenesis, radial and tangential migration, gliogenesis and axonogenesis have been studied in detail in the normal developing brain (Rakic, 1972, 1995; Hatten, 1990; O’Rourke et al., 1992; Bayer and Altman, 1995; Zerlin et al., 1995; Tamamaki et al., 1997; Tan et al., 1998; Lavdas et al., 1999). In contrast to this, with a few exceptions of rare genetically proven cases determined by impairment of genes involved in the processes of neuronal migration (des Portes et al., 1998; Fox et al., 1998), the etiology and pathogenic mechanisms responsible for the development of most cerebral dysgeneses in humans are largely unknown.
The present study is part of a larger project investigating the mechanisms underlying the genesis and the hyperexcitability of neurons of cerebral heterotopia induced by the pre-natal administration of the antimitotic agent methylazoxymethanol acetate (MAM) (Colacitti et al., 1998, 1999; Sancini et al., 1998; Battaglia et al., 2002). MAM is an alkylating agent that has long been used to induce developmental brain dysfunction in rodents (Singh, 1977; Cattabeni and Di Luca, 1997). In vivo, MAM is rapidly converted to methyl-diazonium, which damages DNA by methylating the O6 or N7 positions of guanine nucleic acids (Matsumoto et al., 1972). Actively dividing neuroepithelial cells during the S-phase are affected, whereas post-mitotic neurons or neuroblasts in the G0 phase are spared (Johnston and Coyle, 1979). The narrow time-window of biological activity of MAM (2–24 h, with maximal activity at 12 h after administration) affects the proliferation of specific neuronal cell populations (Cattaneo et al., 1995). We have previously demonstrated that a double MAM administration in rats on embryonic day 15 is able to induce cerebral heterotopia made up by hyperexcitable neurons (Sancini et al., 1998; Colacitti et al., 1999) that share striking similarities with those observed in human peri-ventricular nodular heterotopia (PNH), a cerebral dysgenesis characterized by nodular masses of gray matter located in close apposition to the periventricular germinative neuroepithelium (Battaglia et al., 1996, 1997). In this study, we investigated neurogenesis in the cerebral heterotopia of MAM-treated rats, with the idea of understanding why periventricular heterotopia develop in human patients. For these goals, we have employed a combined approach, by analyzing: (i) the cytoarchitectural features of the heterotopia in the early post-natal period; (ii) the birthdating of cells within the heterotopia, by means of the incorporation of the thymidine analog bromodeoxyuridine (BrdU); and (iii) the cellular phenotypes of the heterotopia, by means of confocal immunofluorescence experiments. Our data demonstrate that all heterotopia in MAM-treated rats share a common neurogenesis and suggest that MAM-induced ablation of early generated waves of neurons is sufficient to deeply alter migration and differentiation of the subsequent waves of newly generated neurons, leading to the formation of the different types of heterotopia.
Materials and Methods
Animal Handling and Cerebral Tissue Processing
All experimental procedures were carried out with care to minimize discomfort and pain to treated rats, in accordance with the guidelines of the European Communities Council (Directive of 24 November 1986, 86/609/EEC). Pregnant Sprague–Dawley rats received two MAM acetate doses (15 mg/kg maternal body wt, i.p. in sterile saline) on E15, the first injection at 12.00 a.m. and the second at 12.00 p.m., as previously reported (Colacitti et al., 1999). On the same day, control pregnant rats were sham injected with the vehicle alone. A group of MAM-treated and sham-operated pregnant rats were then subjected to injections of 5-bromo,2′-deoxy-uridine (BrdU, 50 mg/kg maternal body wt, i.p. in sterile saline) at different embryonic ages (see below). The day after conception (as determined by vaginal smear) was designated embryonic day 1 (E1). Litters were born on day 22 or 23 of gestation and the day of birth was designated post-natal day 1 (P1). The pups were housed under standard conditions, as previously reported (Colacitti et al., 1999).
At various post-natal ages (1 day to 4 months of age), MAM/ BrdU-treated and sham-operated control rats were deeply anesthetized with ice (from P1 to P7) or chloral hydrate (from P15 to adulthood; 1 ml/100 g body wt of a 4% solution) and perfused with 1% paraformaldehyde followed by 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) at pH 7.2. Brains were removed from the skull, post-fixed overnight in 4% paraformaldehyde and cut with a vibratome into 40–50 mm thick coronal sections. Sections were collected in PBS and sodium azide (0.01%) in serial order. One out of three sections (for developing rats) or five sections (for adult rats) were reacted for immunocytochemistry (ICC) as outlined below. The sections adjacent to the immunoreacted ones were counterstained with 0.1% thionine.
To investigate the time of neurogenesis of the MAM-induced heterotopic neurons, seven different time points of BrdU injection were chosen, at 24, 36, 48, 60, 84, 108 and 132 h after the last MAM administration, referred to as E16/24, 17/12, 17/24, 18/12, 19/12, 20/12 and 21/12, respectively. Earlier BrdU administrations were not performed, given the already reported MAM-induced ablation of neuroblasts generated at the time of MAM administration or shortly thereafter (Gillies and Price, 1993; Noctor et al., 1999).
To optimize BrdU immunocytochemistry, we performed several pilot experiments according to protocols reported in the literature (Miller and Nowakowsky, 1988; Soriano and Del Rio, 1991; DeDiego et al., 1994). We obtained best results with the following protocol. Free-floating sections were initially pre-treated with 2 N HCl in PBS for 60 min to separate DNA strands and then with sodium borate 0.1 M for 10 min to neutralize the acid. Sections were then treated with 1% H2O2 in PBS for 20 min to neutralize the endogenous peroxidase activity, rinsed in PBS and incubated with 10% normal goat serum (NGS) and 0.2% Triton-X100 for 60 min, to mask non-specific adsorption sites and to increase the penetration of the reagents. Sections were then incubated overnight with anti-BrdU monoclonal antibodies (Becton-Dickinson, San José, CA; or Boehringer-Mannheim GmbH, Germany) diluted 1:75. After rinsing in PBS, the sections were incubated with biotinylated goat anti-mouse IgG (GAM, diluted 1:200; Jackson, PA), rinsed in PBS and then incubated with Extravidin (1:5000; Sigma-Aldrich, Milano, Italy). All immunoreagents were diluted in 1% NGS in PBS. Peroxidase staining was obtained by incubating the sections either in DAB (0.075%) and H2O2 (0.002%) or in DAB (1.25 mg/ml), NAS (0.04%), NH4Cl (0.004%), glucose (0.2%) and glucose oxidase (1.2 U/ml) in 0.05 M Tris–HCl at pH 7.6. The immuno-reacted sections were mounted onto gelatine-coated glass slides, air-dried, dehydrated and coverslipped with DPX. Slides were then analyzed and photographed with a Nikon Microphot FXA microscope.
To evaluate the cellular phenotype of BrdU-labeled cells, sequential double-labeling immunofluorescence experiments were performed with antibodies against BrdU and (i) the monoclonal antibody against glial fibrillary acidic protein (GFAP, diluted 1:1000; Swant, Bellinzona, Switzerland) as astroglial marker, or (ii) the monoclonal antibody against microtubule associated protein 2 (MAP2, diluted 1:1000; Sternberger Monoclonals Inc., Lutherville, MA), as neuronal marker. For BrdU/GFAP experiments, the tyramide signal amplification protocol (NEN Life Science, Boston, MA) was employed, following the manufacturer’s instructions and using biotinylated GAM (1:200), streptavidin-HRP (1:200) and Cy3-conjugated tyramide (1:100) to reveal BrdU-labeled nuclei and then FITC-conjugated GAM (1:200) to reveal the glial cellular phenotype. For BrdU/MAP2 experiments, the Alexa FluorR 546 GAM (diluted 1:2000; Molecular Probes, Eugene, OR) was used to reveal BrdU-labeled nuclei and the Alexa FluorR 488 GAM (diluted 1:2000; Molecular Probes) was subsequently used to demonstrate MAP2-positive cell bodies.
To evaluate the cellular phenotype of the white matter bands close to the heterotopia in early post-natal rats, double-labeling experiments were performed with antibodies against neuronal and glial cellular markers and the fluorescent Nissl stain NeuroTrace™ (Molecular Probes, Eugene, OR). Free-floating sections from P1–P7 MAM-treated rats were pre-treated with H2O2 and then with NGS/Triton before overnight incubation with the primary antibodies: (i) monoclonal anti-vimentin (Dako, Glostrup, Denmark), diluted 1:200, as an early glial marker (Zerlin et al., 1995); (ii) monoclonal anti-MAP2, diluted 1:1000 and anti-SMI 32, against non-phosphorylated neurofilaments, diluted 1:500 (Sternberger Monoclonals Inc., Lutherville, MA), as specific markers for early-generated cortical neurons (Del Rio et al., 2000). The sections were then incubated for 2 h with Cy2-conjugated GAM (1:200) and then in NeuroTrace™ 530/615 red fluorescent Nissl stain, according to the manufacturer’s instructions. After all experiments, sections were repeatedly rinsed, mounted on slides, coverslipped with Fluorsave (Calbiochem, Darmstadt, Germany) and examined on a Radiance 2100 confocal microscope (Bio-Rad). Images were saved in TIFF format and then elaborated by means of Adobe Photoshop software.
Quantification of BrdU Immunocytochemistry
To determine the percentage of BrdU labeled neurons after BrdU injections at different time points, double labeling BrdU/MAP2 confocal immunofluorescence experiments were used. BrdU-labeled, MAP2-labeled and BrdU/MAP2-double-labeled neurons were counted in three non-adjacent immunoreacted sections through different periventricular heterotopia at different rostro-caudal levels from five rats after E17/12, E18/12, E19/12, E20/12 and E21/12 BrdU injections. Labeled cells were counted without correction, since only estimates of ratios were sought — BrdU/MAP2-double-labeled neurons versus the total number of MAP2-labeled neurons (Saper, 1996). Only MAP2-positive neurons displaying the nucleus surrounded by immunoreactive cell body in the plane of the confocal sections were considered. Neurons with partially but clearly BrdU-labeled nuclei, as well as neurons with fully labeled nuclei were considered as BrdU-positive. The outer borders of the periventricular heterotopia were always easily recognizable from the surrounding white matter and lateral ventricle. Data from cell counts were obtained by using Microsoft Excel. Results were averaged and expressed as percentages of BrdU/MAP2-double-labeled neurons (±standard deviation) on the total number of MAP2-labeled neurons.
Our protocol of MAM administration consistently induced in the offspring of treated dams cortical hypoplasia with disruption of cortical layering and clusters of heterotopic neurons within the neocortex (cortical heterotopia), in the dorso-lateral part of the lateral ventricles (periventricular heterotopia) and within the hippocampal CA1 and CA2 regions [intra-hippocampal heterotopia; see also Colacitti et al. (Colacitti et al., 1999)]. The periventricular heterotopia were frequently in anatomical continuity with the lateral hippocampal border. In contrast, no cytoarchitectural abnormalities were ever found, either in sham-operated animals injected with saline or in rats treated with BrdU alone, given the low administered dose of BrdU (Miller and Nowakowski, 1988; Kolb et al., 1999). In the present study, we have specifically addressed the following issues: (i) the development of the heterotopia in the first two post-natal weeks; (ii) the neurogenesis of the heterotopia, by means of BrdU immunocytochemistry; and (iii) the cellular phenotypes within the heterotopia, by means of double-labeling immunofluorescence.
Early Post-natal Development of the MAM-induced Heterotopia
One of the main cytoarchitectural features of the MAM-treated rats during early post-natal development was the presence of elongated cellular bands of intensely thionine-stained cells within the subcortical white matter (Figs 1 and 2D, arrows). These bands were most probably clusters of young migrating neurons similar to those observed pre-natally in the intermediate and subventricular zones of the developing white matter (Bayer and Altman, 1995). In MAM-treated rats, however, numerous cellular bands were still present in the white matter of the early post-natal period, intermingled with many elongated cells with morphological features typical of young migrating neurons (Fig. 1B, arrowheads). More importantly, in MAM-treated rats the bands were located in close anatomical relationship to the heterotopia, i.e. in the white matter immediately beneath the cortical heterotopia (Fig. 1A–C), or overlying the periventricular and intra-hippocampal heterotopia (Figs 1D,E and 2D). They remained conspicuous during the first post-natal week, but progressively disappeared during the second post-natal week.
The cortical heterotopia overlying the white matter bands were already present at birth (P1), as round or elongated nodules of densely packed neurons of small size and ill-defined morphology (Fig. 1A). The neocortex around the heterotopia contained many small neurons of uniform morphology and the normal cortical layering was not recognizable. A clear reduction of cellular density was evident above the heterotopia (Fig. 1A, asterisk). During the first post-natal week the cortical heterotopia enlarged in size; the neurons at the borders of the heterotopia progressively displayed a less immature morphology (Fig. 1C, arrowheads), whereas the neurons within the heterotopia were still more densely packed than the surrounding neocortical neurons and intermingled with darkly stained small cells (Fig. 1C). At P15, the cortical heterotopia displayed the anatomical features already described for the adult rats (Colacitti et al., 1999).
The periventricular heterotopia were also already present at birth (Fig. 1D). They were located just dorsally to the germinative neuroepithelium and immediately beneath the bands of darkly stained cells (Fig. 1D,E). They progressively increased their size during the first post-natal week (cf. Fig. 1D and E). The core of the nodules was made up of densely packed round cells with ill-defined borders, whereas elongated neurons were marginally placed at the border of the nodules intermingled with more darkly stained small cells (Fig. 1F). At P5, a thin space, possibly tangentially running fibers, divided the border from the core of the nodules (Fig. 1F, asterisks). In contrast, the intra-hippocampal heterotopia were never present at birth (Fig. 2A,B). At P1, sectors of the germinative neuroepithelium dorsal to the developing hippocampus were particularly thick (Fig. 2A), with thin columns of dark migrating neurons extending to the already formed pyramidal cell layer of CA1 and CA2 (Fig. 2B, arrowheads). At P3, wedge-shaped masses of neurons filled the stratum oriens between the neuroepithelium and the pyramidal cell layer (Fig. 2C), whereas at P5 the intra-hippocampal heterotopia consistently disrupted the CA1 and CA2 pyramidal cell layer and invaded the stratum radiatum (Fig. 2D). Thick cellular bands were present in the subcortical white matter overlying the intra-hippocampal heterotopia (Fig. 2D, arrows). At P15, the subcortical as well as the intra-hippocampal nodules displayed the anatomical features of the adult rat (Colacitti et al., 1999).
Data from BrdU labeling in MAM-treated rats are illustrated in Figures 3–7 and summarized in the following, by grouping together E16/24–E17/12–E17/24 BrdU injections (i.e. from 24 to 48 h after the last MAM administration), E18/12–E19/12 BrdU injections (from 60 to 84 h after MAM) and E20/12–E21/12 BrdU injections (from 108 to 132 h after MAM). In general, the neocortical and hippocampal areas surrounding the heterotopia followed the previously reported time-course of cortical and hippocampal neurogenesis (Hicks and D’Amato, 1968; Bayer, 1980; Bayer and Altman, 1995). Indeed, inside-out, latero-medial and rostro-caudal gradients of neurogenesis were observed in the MAM neocortex after BrdU injections from E16 to E21. In the hippocampus, CA3 pyramidal neurons were generated before CA1 and CA2 neurons and granule cells in the dentate gyrus, and a superficial to deep (or external to internal) gradient of neurogenesis was clearly present (not shown). By contrast, the time of generation of neurons within the heterotopia and adjacent white matter bands was extended and it roughly overlapped that of superficial neocortical neurons. Indeed, as illustrated in Figure 3, after E18/19 BrdU injections labeled neurons were located in the more superficial cortical layers and in the cortical (Fig. 3A,B) and periventricular (Fig. 3B) heterotopia. In contrast, neurons in the deep cortical layers surrounding the heterotopia were mostly unlabeled (Fig. 3A,B).
After E16/24–E17/12–E17/24 BrdU injections, the cellular bands within the white matter were BrdU-negative, even if rare cells were BrdU-positive after E17/24 injections (not shown). In contrast, BrdU-positive neurons were found in the cortical and subcortical heterotopia (Figs 4–6) and their number increased progressively from E16/24 to E17/24 injections. They were mainly located at the periphery of cortical (Fig. 4A,B) and periventricular (Fig. 5A,C) heterotopia. The numbers of labeled cells and the intensity of labeling tended to decrease moving inside the core of the heterotopia. In the intra-hippocampal heterotopia, labeled cells were more evident dorsally and the numbers of labeled cells and intensity of labeling decreased moving ventrally (Fig. 6A,B).
After E18/12–E19/12 administration, BrdU-positive cells were numerous between P1 and P7 in the white matter bands close to the cortical and heterotopia (Fig. 7A) and overlying the subcortical heterotopia (Fig. 7B,D). They progressively disappeared after the first post-natal week, but some BrdU-positive neurons were still present in adulthood in the white matter close to the heterotopia (Fig. 7C). Within the heterotopia, many neurons were BrdU-positive during the post-natal development and in adult stages (Figs 4–6). Labeled heterotopic neurons were located throughout the extension of the heterotopia, in both the periphery and the core of the cortical (Fig. 4C,D) and periventricular (Fig. 5B,D) heterotopia and in both the dorsal and ventral aspects of the intra-hippocampal heterotopia (Fig. 6C,D).
After E20/12–E21/12 BrdU injections, the white matter bands close to the heterotopia were still BrdU-positive in P1–P7 MAM-treated rats, but less conspicuously if compared to those labeled after E18/19 injections. Some BrdU-positive neurons were still present in adult rats in the white matter close to the heterotopic nodules (not shown). In both developing and adult rats, labeled heterotopic neurons were confined inside the cortical (Fig. 4E,F) and periventricular (Fig. 5E,F) heterotopia and in the more ventral aspects of the intra-hippocampal heterotopia (Fig. 6E,F). Thus, taken together, these data indicate that the white matter bands begin to form after the heterotopia and that the heterotopia are formed through an outside-in (for cortical and periventricular heterotopia) and dorso-ventral (for intra-hippocampal heterotopia) neurogenetic gradient.
Most BrdU-labeled Heterotopic Cells are Neurons
To verify the cellular phenotype of the BrdU-labeled cells in the MAM-induced heterotopia, we performed double-labeling immunofluorescence experiments by combining BrdU staining with labeling for either GFAP, a well-known marker for astrocytes, or MAP2, a specific neuronal marker. At all considered time-points of BrdU administration, the vast majority of BrdU-positive cells within the heterotopia were characterized by large nuclei not surrounded by GFAP positive processes (Fig. 8A,B, arrowheads). In particular, double-labeled BrdU and GFAP-positive glial cells were very rare in the heterotopia after E17/12–E18/12 injections and only slightly more numerous after later BrdU injections (Fig. 8A). BrdU/GFAP-positive glial cells were present after E19/12 and later BrdU injections in the white matter adjacent to the heterotopic structures (Fig. 8B, arrows). By contrast, most BrdU-positive nuclei were surrounded by MAP2-positive cell bodies after BrdU injections at all considered time-points (Fig. 8C,D, arrows). After E19/12–E21/12 BrdU injections, some BrdU labeled nuclei not surrounded by MAP2 staining were present within the heterotopia and more conspicuously in the adjacent white matter (not shown). These findings clearly demonstrate that most BrdU-labeled cells within the heterotopia at the different considered time-points possess a neuronal phenotype and they also suggest that gliogenesis follows neurogenesis in the MAM-induced heterotopia.
Cells within the White Matter Bands Possess a Neuronal Phenotype
To investigate the cellular phenotype of the white matter bands close to the heterotopia, we then performed double-labeling experiments with the fluorescent Nissl stain NeuroTrace™ and antibodies against neuronal and glial markers. In the first post-natal week, the white matter bands were made up of cells with a relatively large nucleus, intensely stained by NeuroTrace™ (Fig. 8H–L) and a thin rim of cytoplasm, immunoreactive for the anti-MAP2 (Fig. 8E) and anti-SMI 32 antibodies (Fig. 8F). In contrast, the bands were crossed by vimentin-positive fibers but they did not contain vimentin-immunoreactive cell bodies (Fig. 8G,L). Rare vimentin-positive cells were occasionally found at the borders of the white matter bands (Fig. 8G,L, arrow). Neurons within the cortical and periventricular heterotopia displayed a more evident MAP2 cytoplasmic staining if compared to neurons within the nearby white matter bands (Fig. 8E,H). Taken together, these double-labeling experiments demonstrate that the white matter bands close to the heterotopia are mostly made up by neurons.
Quantification of BrdU Immunocytochemistry
Finally, we used double labeling BrdU/MAP2 experiments to quantify the birthdating at different embryonic ages of neurons committed to form the heterotopia. We selected periventricular heterotopia for this analysis, since their anatomical borders were always easily distinguishable and BrdU/MAP2 immunofluorescence since single- and double-labeled neurons were recognizable with certainty when analyzed at the confocal microscope (Fig. 8C,D). Data are illustrated in Figure 9 as mean ratios (±standard deviations) of BrdU/MAP2 double-labeled neurons on the total number of MAP2-positive neurons within the heterotopia. Neurogenesis began in periventricular heterotopia soon after the last MAM administration (27.27% of double-labeled neurons after E17/12 BrdU injections) and it rapidly progressed to reach a neurogenetic peak around the 18th embryonic day for periventricular heterotopia (52.09% of double-labeled neurons). Interestingly, neurogenesis continued on the 19th and 20th embryonic days and still 10.55% of neurons forming periventricular heterotopia were generated at the 21st embryonic day (Fig. 8).
In the present study, by combining cytoarchitectural, BrdU and immunofluorescence analyses we have provided an assessment of the altered neurogenesis which underlies the formation of cerebral heterotopia in MAM-treated rats. We believe that the present data should provide insight not only into the process of cortical development, but also into the possible way by which periventricular nodular heterotopia develop in human patients.
Ontogenesis of Cerebral Heterotopia in MAM-treated Rats
Numerous cellular bands of migrating neurons are present in the white matter of MAM-treated rats during the first post-natal week, in close anatomical relationship with the newly forming heterotopia (Figs 1 and 2). During the process of migration in the normal rat brain, cells committed to the neocortex sojourn between E15 and E21 in cellular bands within the intermediate and subventricular zones before reaching their final cortical destination (Bayer and Altman, 1995). The white matter bands described here in the post-natal brain of MAM-treated rats likely correspond to the superior bands of the pre-natal brain of normal rats (Bayer and Altman, 1995), indicating that the MAM treatment determines a delayed post-natal maturation of the heterotopia. The bands are made up mostly of neurons, as demonstrated by their nuclear size (Robertson et al., 2000) and their perikaryal staining by anti-MAP2 and anti-SMI 32, but not by anti-vimentin antibodies (Fig. 8). MAP2-positive neurons were reported in the developing white matter of newborn MAM-treated ferrets (Noctor et al., 1999), but the neuronal bands were not described, possibly for the fact that the stage of cortical development in neonatal ferrets is more advanced in comparison to that in neonatal rats. Indeed, the MAP2-positive neurons within the white matter bands reported here are still-differentiating young neurons, since they display less intense MAP2 immunoreactivity than neurons already located within the heterotopia (Fig. 8). Their morphology and staining pattern, together with the close anatomical relationship with the different types of heterotopia, suggest that these bands function in the early post-natal period as a reservoir of young migrating neurons specifically committed to the newly forming heterotopia. They later disappear, possibly for the combination of migration into the heterotopia and apoptotic cell death (Del Rio et al., 2000), but quite a few neurons are still present within the white matter close to the heterotopia during adulthood (Fig. 7C), as described also for human PNH (Battaglia et al., 2002).
The close anatomical relationship between these bands and the newly forming heterotopia suggests that the heterotopia themselves are able to influence and direct the migration of later generated waves of neurons. The reduction of cellular density consistently observed just above the cortical heterotopia (Fig. 1A) also suggests that the absence of superficial neurons impairs the proper migration of later generated cortical neurons. Taken together, these data support the idea that early generated and properly migrated neurons are of key importance for the migration and differentiation of subsequent waves of neurons. A similar hypothesis has been recently put forward in a paper in which MAM was specifically administered in ferrets to affect the generation of subplate neurons (Noctor et al., 1999). In that paper, the MAM-induced dramatic alteration of cortical layering was taken to support the hypothesis that early generated cortical layers provided environmental factors necessary for the subsequent formation of cortical layers (Noctor et al., 1999). In keeping with this view, we hypothesized that the MAM-induced ablation of early generated cortical neurons determines the presence of heterotopic cortical neurons (Chevassus et al., 1998b; Colacitti et al., 1999) in the deep cortical layers and periventricular white matter and that, in turn, these heterotopic neurons influence the migration of later generated neurons to determine the genesis of the heterotopia.
Our birth-dating BrdU analysis has revealed a clear neurogenetic gradient in all MAM-induced heterotopia. Cortical and periventricular heterotopia are formed by progressive settling of neurons from the outside to the inside, whereas the intra-hippocampal heterotopia are progressively formed throughout a dorsal to ventral migration of neurons. A dorso-ventral neurogenetic gradient has been already reported in rats for the intra-hippocampal heterotopia induced by a single MAM dose (Chevassus et al., 1998b). Our data extend those findings in clearly demonstrating that in MAM-treated rats the neurogenesis is not chaotic nor rudimentary, as previously suggested (Chevassus et al., 1998b). Rather, it seems to follow the general ontogenetic rules of cortical layering formation. In MAM-treated rats, however, as a consequence of the MAM-induced ablation of early generated neuroblasts, heterotopically located neurons in the periphery of the cortical and periventricular heterotopia and in the dorsal part of the intra-hippocampal heterotopia set the base for the subsequent migration of neurons into the heterotopia.
Our morphologic analysis also demonstrate that intra-hippocampal heterotopia are not present at birth, but progressively formed in the early post-natal period, as already suggested (Singh, 1977; Chevassus et al., 1998b; Castro et al., 2001). In addition, our BrdU experiments demonstrate that neurons within the intra-hippocampal heterotopia display the same neurogenetic profile as neurons forming the cortical heterotopia, thus lending further support to the idea that these neurons are neurons committed to superficial cortical layers growing into the hippocampus (Chevassus et al., 1998b; Colacitti et al., 1999). Why do these heterotopia develop into the hippocampus and why do they develop post-natally? If already migrated neurons are able to direct subsequent migration of later generated neurons, we can hypothesize that hippocampal neurons may attract the young migrating neurons sojourning in the white matter bands toward the pyramidal cell layer. A common ontogenesis for all types of heterotopia is further supported by the fact that in rats deeply affected by MAM treatment it is not uncommon to find large cortical, periventricular and intra-hippocampal heterotopia fused together (data not shown). The post-natal development of intra-hippocampal heterotopia could be related to the pre-natal physical separation between the hippocampus and the subcortical white matter, which would make the migration into the hippocampus impossible (Chevassus et al., 1998b). Finally, these data together with the already reported anatomical and electrophysiological features of MAM-induced heterotopia (Chevassus et al., 1998a; Sancini et al., 1998; Colacitti et al., 1999) further demonstrate that heterotopic neurons maintain cortical features even if positioned outside the borders of the neocortex or within the hippocampus, thus supporting an early commitment of neocortical neurons (Tan et al., 1998).
The neuronal phenotype of most cells in both heterotopia and white matter bands and the double-labeling GFAP/BrdU experiments (Fig. 8) clearly demonstrate that in MAM-induced heterotopia, neurogenesis precedes gliogenesis. These findings reflect the general rule in CNS development that astrocytes are mostly generated after neurons. In addition, they suggest that the abnormalities of glial cells described in other papers dealing with the prenatal effect of MAM (Collier and Ashwell, 1993; Zhang et al., 1995; Noctor et al., 1999) are most probably not the cause of the genesis of heterotopia, but more simply a consequence of the altered neuronal neurogenesis reported here.
Relevance of the MAM Model for the Ontogenesis of Heterotopia in Humans
We have previously demonstrated that MAM-induced heterotopia are characterized by anatomical features similar to those observed in human PNH (Colacitti et al., 1999). Similarities between MAM induced heterotopia and human PNH are also supported by the fact that heterotopic neurons in MAM-treated rats are hyperexcitable (Baraban and Schwartzkroin, 1995; Sancini et al., 1998; Baraban et al., 2000; Castro et al., 2001) and heterotopic nodules in human PNH can give rise to epileptic discharges (Dubeau et al., 1995; Kothare et al., 1998). In addition, recent data from our group suggest that heterotopia in both MAM-treated rats and human PNH are characterized by reduced expression and altered function of NMDA receptor complex and the alpha subunit of Ca2+/calmodulin-dependent protein kinase II (αCaMKII) (Battaglia et al., 2002; Gardoni et al., 2003). Therefore, anatomical, electrophysiological and molecular analyses indicate that heterotopia in MAM-treated rats and heterotopia in human PNH share common features. The discussion of the molecular and electrophysiologic similarities between MAM-induced heterotopia and PNH in humans are obviously beyond the scope of the present paper. However, these similarities can be taken to further validate MAM-treated rats as an experimental model for human PNH, thus allowing the use of the developmental profile of heterotopia in MAM-treated rats as clues to inform speculation about the origin and intrinsic epileptogenicity of human PNH.
If it is true, as suggested by the evidence provided in this paper, that ablation of early waves of neuroblasts is sufficient to alter the structure of selected sectors of the brain, it can be speculated that in humans noxious events around the 7th–8th week of gestation (Bayer et al., 1993), such as ischemic insults in limited parts of the germinative neuroepithelium, are sufficient to determine the genesis of periventricular nodules typical of PNH. The existence of these nodules may, in turn, determine altered axonogenesis, since, during their development, the heterotopic neurons send and receive axonal projections to and from the cortex. The ‘misplaced’ axonogenesis sets the base for the establishment of altered connections between heterotopia and neocortical and archicortical areas (Chevassus et al., 1998b; Colacitti et al., 1999; Hannan et al., 1999). In addition, even if the heterotopic neurons display normal expression of many neurochemical markers (Chevassus et al., 1998b; Colacitti et al., 1999), the process of neuronal differentiation is probably altered. Indeed, neurons in MAM-induced intra-hippocampal heterotopia are characterized by reduced expression of Kv4.2 A-type potassium channel subunits (Castro et al., 2001) and by altered GABA-mediated synaptic inputs (Calcagnotto et al., 2002) and neurons in both MAM-induced heterotopia and human PNH are characterized by reduced expression of the NMDA receptor complex and αCaMKII (Gardoni et al., submitted for publication). Even if the molecular mechanisms underlying the increased excitability are still unexploited, human PNH are likely characterized by hyperexcitable neurons inserted in a redundant neocortical/archicortical network that may facilitate the diffusion of single interictal discharges and determine the genesis of sustained epileptiform discharges.
This study was partially supported by grants ICS 030.3/RF98.36 and RF162/02 from the Italian Ministry of Health.
Address correspondence to Giorgio Battaglia, Molecular Neuroanatomy Laboratory, Department of Experimental Neurophysiology, Istituto Neurologico ‘C. Besta’, Via Celoria 11, 20133 Milano, Italy. Email: firstname.lastname@example.org.