[1] 11w Developmental lineage of cell types in cortical dysplasia with balloon cells

[1] 300w Focal cortical dysplasia type IIB with Ballon cells is a developmental malformation of the cerebral cortex highly associated with epilepsy. As a strategy to define the embryonic origin and neurochemical phenotype of cells in this disease, we probed specimens (n = 10) resected during epilepsy surgery with a panel of 13 antibodies recognizing proteins associated with (i) specific progenitor cell types including brain lipid binding protein (BLBP), collapsin response mediator protein 4 (CRMP4), Dlx1, Dlx2, GFAPd, MASH1, Otx1, Pax6, vimentin and phosphorylated vimentin and (ii) excitatory or inhibitory neurochemical phenotypes such as the vesicular glutamate transporters-1 and 2 (VGLUT-1,VGLUT-2), or the vesicular GABA transporter (VGAT). Balloon cells and large dysplastic neurons in all specimens expressed Otx1, phospho-vimentin, Pax6 and BLBP, proteins normally expressed by cells in the embryonic ventricular zone. A subpopulation of balloon cells expressed MASH-1 also expressed in the ventricular zone. Most balloon cells and dysplastic neurons were VGLUT2 immunoreactive, whereas none expressed Dlx1 or Dlx2, markers for inhibitory cells derived from the medial ganglionic eminence and few expressed VGAT, found in GABAergic interneurons. Otx1 mRNA expression and Dlx1 mRNA absence was confirmed by single cell RT-PCR. A subpopulation of balloon cells was labelled with CRMP4 and GFAPd, markers specific for newly generated cells derived from the adult subventricular zone. Detection of Otx1, phospho-vimentin, Pax6 and BLBP expression but absence of Dlx1/Dlx2 expression suggests that balloon cells and dysplastic neurons derive from radial glial cells in the telencephalic ventricular zone and not the medial ganglionic eminence. VGLUT expression argues that dysplastic neurons may be glutamatergic. CRMP-4 and GFAPd expression suggests that new cells may arrive in focal cortical dysplasia, perhaps deriving in part from the subventricular zone. These findings provide a developmental lineage model in which balloon cells and dysplastic neurons are derived from radial glial progenitor cells.

[1] 335w Focal cortical dysplasia type IIB with balloon cells (FCDIIB) is a focal developmental malformation of the cerebral cortex that affects children and adults, and is highly associated with intractable epilepsy (Mischel et al., 1995;Palmini et al., 2004;Fauser et al., 2006). FCD is characterized histologically by disorganized or absent cortical lamination and the presence of cells exhibiting cytomegaly known as balloon cells (BCs). Cytomegalic dysmorphic neurons (DNs; Andre et al., 2004) within FCD are distinct from BCs in that they extend dendrites and axons and are morphologically similar to neurons. In contrast, BCs are larger than DNs (often exceeding 130 microns), exhibit a laterally displaced nucleus, and lack their neuronal morphology. Defining the cellular origin and molecular phenotype of BCs and DNs may provide critically important insights into how FCDs form during brain development. Indeed, a pivotal question regarding the developmental pathogenesis of FCD is whether BCs and DNs derive from cells within the telencephalic ventricular zone (VZ) or the medial ganglionic eminence (MGE) during cortical development and hence, whether BCs and DNs are derived from progenitor cells destined to be excitatory or inhibitory neurons. Most cortical excitatory neurons derive from progenitor cells in the VZ, whereas inhibitory interneurons largely derive from the MGE (McConnell, 1995;Anderson et al., 2001). In human brain, a small proportion of inhibitory cells derive from the VZ (Letinic et al., 2002). Cells in the VZ and MGE have completely divergent cell lineages, follow distinct migratory pathyways and lamination cues, and express different transcription factors, neurotransmitters, and neurotransmitter receptors when compared with progenitor cells in the MGE (Gotz and Sommer, 2005). Thus, cell autonomous events affecting the normal development or migration of progenitors in the MGE very likely would have completely distinct effects compared with similar events in a VZ progenitor. Furthermore, defining the lineage for cells in FCD i.e. VZ versus MGE, could be important in identifying causative gene mutations for FCD and perhaps identifying potential target molecules that are unique to cells in the VZ or MGE for future therapy.

[2] 344w Numerous animal studies have taken advantage of lineage marker protein expression to define the origin and phenotype of cortical neurons in both normal human cortex as well as human cortical malformations and animal models of malformations (see review, Hevner, 2007). Thus, as a strategy to study the origin and lineage of cells comprising human FCDIIB, tissue sections were probed with a panel of 13 antibodies that define select cell types in the developing cortex. For example, orthodenticle-1 (Otx1) is a transcription factor that is expressed by cells in layers V and VI of cortex that were generated in the VZ (Frantz et al., 1994;Acampora et al., 2001). Dlx1 and Dlx2 (distaless-and -2 homologs) are homeobox transcription factors expressed by inhibitory neurons that have their origin in the MGE (Anderson et al., 1997) or in the human VZ (Letinic et al., 2002). Mammalian achaete-scute homolog-1 (MASH1) is a helix-loop-helix transcription factor expressed by progenitors in the human VZ (Letinic et al., 2002;Yun et al., 2002;Ross et al., 2003). Pax6 is a paired box transcription factor that is expressed by radial glial cells in the developing VZ (Gotz et al., 1998). Brain lipid binding protein (BLBP) is linked to the Notch1 signalling pathway and is also expressed in radial glial cells in the VZ (Feng et al., 1994;Howard et al., 2006). Vimentin, and its phosphorylated isoform, is an intermediate filament protein identified in radial glial cells (Zecevic, 2004). The delta isoform of GFAP (GFAPd) is a distinct splice variant isoform of GFAP expressed in a specific astrocyte population in the adult human subventricular zone (SVZ; Roelofs et al., 2005) which contains neuronal stem cells (Sanai et al., 2004). Similarly, collapsin response mediator protein (CRMP4) is a marker for newly generated neurons derived from the SVZ (Seki, 2002). The vesicular glutamate transporters (VGLUT1, VGLUT2) are expressed by glutamatergic neurons and facilitate synaptic glutamate re-uptake into neurons (Minelli et al., 2003a). The vesicular GABA transporter (VGAT) mediates reuptake of GABA into pre-synaptic vesicles and serves as a marker for GABAergic neurons (McIntire et al., 1997;Minelli et al., 2003).

[1] 186w Cell counting identified a total of 72 750 BCs and 96 500 DNs across the 1 cm 2 ROIs measured in five immunolabelled sections from the 10 FCD cases (n = 50 ROI). The mean (AESEM) number of BCs per ROI was 291 AE 37 and DNs was 386 AE 42. The mean maximal cell soma diametre of the BCs was 127.2 microns (range 120-165 microns) and the mean calculated somatic area of BCs (estimated as roughly spheroid shapes) was 12661.26 microns 2 . The majority of BCs were measured within the 120-140 micron range. The mean maximal cell soma diametre of the DNs was 113.7 microns (range, 110-140 microns; excluding dendritic or axonal segments; mean area of DNs was not calculated because a clear distinction between the cell soma and proximal dendritic shaft was not easily discernable). Thus, a small population of very large DNs measured approximately the same size as smaller BCs. However, DNs extended definitive axonal or dendritic processes from the soma and could be morphologically distinguished from BCs. These morphometric parameters were used to generate cell counts for each individual protein marker.

[2] 140w The expression profile for each protein marker was heterogeneous across the 10 FCD cases. In some specimens, even when corrected for differing numbers of BCs or DNs across each case, there was robust expression of individual proteins while in others fewer numbers of cells were labelled (Table 3). The exceptions to this finding were Otx1 and vimentin expression, which was robust across all cases (see below). For some proteins detailed subsequently, there was enriched expression in BCs in deeper portions of the FCD i.e. within the subcortical white matter (Table 3, WM), whereas for others, protein expression was observed in BCs throughout the thickness of the lesion (Table 3, ALL). When protein expression was compared in BCs versus DNs, there was a statistically significant difference in the numbers of BCs expressing BLBP, GFAPd, MASH1, Pax6, phospho-vimentin and vimentin (Table 3).

[3] 71w We were unable to make any correlations between clinical phenotype (age, age at seizure onset, duration of seizures or surgical outcome) and protein marker expression. In addition, we did not detect any differences in protein expression in the two control groups across distinct brain regions nor between the epilepsy surgical and postmortem control groups to suggest an artifactual effect of anti-seizure medications, recurrent seizures, age or postmortem interval on protein expression.

[4] 11w Lineage Markers: Otx1, BLBP, Pax6, MASH1, vimentin, phospho-vimentin, Dlx1/2, CRMP-4, GFAPd

[5] 186w Otx1 expression was detected in numerous BCs (88%) across all FCD specimens (Fig. 2). Otx1 labelled BCs were observed throughout the full thickness of each dysplasia extending into the subcortical white matter. DNs (76%) and heterotopic neurons also expressed Otx1. In epilepsy and post-mortem control specimens, Otx1 labelling was observed within pyramidal cells in layer V (Fig. 2). Otx1 transcript was identified in single micro-dissected Otx1 immunolabelled BCs and DNs (Fig. 2). BLBP and Pax6 proteins were detected immunohistochemically in BCs and to a lesser extent DNs in all FCD specimens (Fig. 3). Cell counts revealed that 74% of BCs and 19% of DNs expressed BLBP and 58% of BCs and 12% of DNs expressed Pax6. There was minimal expression of BLBP and Pax6 in control cortex specimens (data not shown). MASH1 expression was detected in a small number of BCs and DNs (by quantitative count, 34% of BCs were MASH-1 labelled and 8% of DNs; Fig. 4). The distribution of MASH-1 expression was in deep and superficial portions of the lesion. Small multipolar MASH1 labeled cells were noted in the epilepsy and post-mortem control specimens.

[6] 62w In contrast, none of the BCs expressed Dlx1 or Dlx2 (Fig. 5). However, Dlx1 and Dlx2 labelled DNs were observed especially in the subcortical WM (heterotopic cells) and in small, multipolar cells in control cortices (Fig. 5). The absence of Dlx1 protein in BCs was supported by the inability to amplify Dlx1 mRNA from these cells by single cell RT-PCR (Fig. 5).

[7] 278w As previously reported, numerous vimentin labelled BCs (83%) were noted in all FCD specimens across all layers (Fig. 6). In addition, we found that the FCD specimens exhibited phospho-vimentin labelling in BCs (Fig. 6) and to a lesser extent DNs across all layers but enriched in the subcortical WM. Overall, 66% of morphologically defined A B C D Fig. 3 Expression of Pax6 and BLBP in FCD. (A) low and (B) higher magnification photos of BLBP expression in BCs (large arrows) and DNs (small arrows). (C) low and (D) higher magnification photos of Pax6 expression in FCD. In D, BCs (large arrow) and DNs (small arrow) were Pax6 labelled. Scale bar for B and D = 100 mm. A B C D E Fig. 5 Dlx1 and Dlx2 expression in FCD. (A) Dlx1 labelled small multipolar cells in post-mortem control cortex (arrows). Inset, Dlx1 labeled cells in post-mortem control cortex (arrowheads). (B) note absence of labelling in balloon cells (arrows in phase contrast image, B), and light micrograph,C (arrowhead demarcates pial surface). D, heterotopic cells exhibiting Dlx1 labelling (arrowheads) in WM (scale bar = 300 mm). E, Dlx2 labeled cells in WM. Inset, single cell RT-PCR depicting absence of Dlx 1 transcript in single microdissected BCs (lanes 2 and 3) but detection of Dlx1 transcript in Dlx1 immunolabelled cell (lane 4). A B Fig. 4 MASH1 labeled cells in FCD. (A) BCs (arrows) across all layers expressed MASH1. Inset, nearly absence of MASH1 labelling in post-mortem control cortex (arrowhead denotes pial surface). (B) Note labelling of BCs (arrows) including BC with dual nuclei (asterisk) and smaller cells which reflect dysmorphic astrocytes (short arrows). Scale bar = 150 mm.

[8] 13w BCs expressed phospho-vimentin. Phospho-vimentin was not detected in epilepsy or post-mortem control cortex.

[9] 111w There was robust cytoplasmic CRMP-4 immunolabelling noted in each of the FCD specimens (Fig. 7). In seven specimens, CRMP-4 labelled BCs were noted throughout the thickness of the dysplasia extending into the subcortical white matter whereas in the remaining three specimens only a subpopulation of BCs confined to the subcortical WM expressed CRMP-4. Overall, 27% of morphologically defined BCs and 13% of morphologically defined DNs across all specimens exhibited CRMP-4 labeling. Of particular interest was the observation of 'halos' of CRMP-4 labelling around BCs, and around heterotopic neurons in the subcortical WM (Fig. 7). Rare CRMP4 labelling was noted in epilepsy and post-mortem controls in small cells scattered throughout all layers.

[10] 47w Many dysmorphic and enlarged cells exhibiting an astrocytic morphology were labelled with GFAPd (Fig. 8). To varying degrees in each specimen, scattered BCs were GFAPd immunoreactive while DNs were not labelled. Quantitative cell counts revealed that 42% of cells meeting morphological criteria for BCs were GFAPd labelled.

[11] 115w The expression of VGLUT1 and VGLUT2 was identified in all FCD specimens (Fig. 9). Specifically, a pattern of fine punctuate staining was detected throughout the tissue. The expression of VGLUT-1 was identified in axon terminals clustered around select BCs as well as heterotopic neurons in the subcortical WM (Fig. 9). Overall, there was a relative enrichment in VGLUT1 labelling in the superficial portions of the dysplasia compared with deeper regions. Many DNs and heterotopic neurons also expressed VGLUT1 but in view of the punctate staining pattern cell numbers were not quantified. Numerous BCs (by quantitative cell count, 71%) expressed VGLUT2 in their somatic cytoplasm. VGLUT2 expression was detected by pyramidal cells in control cortex specimens.

[12] 150w In contrast to VGLUT1 and 2, there were very few VGAT immunolabelled cells in the FCD specimens and only rarely (3%) did DNs expressed VGAT (Fig. 10). Small VGAT A B C D Fig. 9 VGLUT1 and two immunolabelling in FCD. (A) VGLUT2 expreession in pyramidal cells (arrows) in post-mortem control cortex. (B,C) perisomatic and dendritic punctate VGLUT1 labelling in DN (B) and heterotopic neuron (C) in subcortical white matter. Scale bar = 100 mm. (D) cytoplasmic VGLUT2 labeling in BCs. Scale bar = 200 mm. A B Fig. 6 Numerous DNs express (A) vimentin, (arrows; Scale bar = 200 mm). (B) BCs express phospho-vimentin, (large arrows; Scale bar = 70 mm). Astrocytes also express phospho-vimentin (small arrows). A B Fig. 7 CRMP-4 expression in BCs (A). Note 'halos' (arrows) of CRMP4 immunolabelling around BCs (B). Scale bar = 200 mm. Inset, near absence of CRMP-4 labelling in control cortex.

[13] 16w labelled multipolar cells were noted throughout each specimen and within the epilepsy and normal control specimens.

[1] 284w We show that BCs and DNs in FCD express protein markers that highlight developmental lineage and neurochemical phenotype. Specifically, the expression of Otx1 and VGLUTs in BCs and DNs suggests that these cells derive from radial glial progenitor cells in the VZ. The persistent expression of phospho-vimentin, MASH1, Pax6 and BLBP in BCs and a subpopulation of DNs, supports the hypothesis that these cells derive from radial glial cells in the VZ and further suggests that these cells retain an embryonic phenotype. In contrast, the virtual absence of Dlx1, Dlx2 and VGAT in BCs supports the hypothesis that the majority of BCs and DNs do not derive from the MGE. Expression of CRMP-4 in BCs, a marker protein found in newly generated or immature cell types derived from the SVZ, supports the hypothesis that FCD may contain cells that have recently divided. In addition, expression of GFAPd labeling suggests that numerous dysmorphic astrocytes and a subpopulation of BCs in FCD may derive from the proliferative regions of the SVZ. The differential expression profile for most of the proteins assayed in these FCD specimens suggests that the cell constituents or temporal assembly of FCD is likely to be heterogeneous during brain development. These results further support results of previous investigators (Sisodiya et al., 2002;Englund et al., 2005;Thom et al., 2005;Yamanouchi et al., 2005;Ying et al., 2006) demonstrating that the expression of marker proteins within FCD can be variable and suggests that the phenotypic features of cells in FCD are not uniform. Furthermore, for several marker proteins e.g. BLBP, GFAPd, MASH1, Pax6, the immunolabeling profiles of BCs and DNs were distinct (Table 3) suggesting that these cells may derive from the VZ through distinct developmental programs.

[2] 201w We acknowledge several caveats to our results. First, the expression of cell selective markers in FCD may not demarcate specific cell lineages but instead may reflect aberrant gene transcription or protein translation due to the abnormal phenotype of BCs or up-or downstream changes in expression of other genes that are distinct from lineage identities. Alternatively, the profound abnormality of cortical lamination in FCD may cause altered expression of lineage or neurochemical markers that do not reflect actual lineage origin. While it is possible that differential expression of each protein could represent regional expression variability in the normal cortex, there is no published evidence nor evidence from our series that any of the markers analyzed exhibit differential regional expression in normal neocortex when comparing the frontal versus temporal versus parietal cortices. Other cells distinct from BCs or DNs were observed in the sections but these cells were not studied as the focus of the experiments was to define lineage and origin of BCs and DNs. We do not believe that the presence of seizures alone or anti-epileptic drug exposure alone altered protein expression since the expression of each marker protein was similar in our epilepsy control and the non-epilepsy control cortex groups.

[3] 278w The expression profiles of BLBP, MASH1, Otx1, Pax6, vimentin and phospho-vimentin in BCs in all 10 FCD cases provides strong support to the hypothesis that BCs are derived from radial glial cells in the telencephalic VZ (Cepeda et al., 2006). With the exception of Otx1, expression of these marker proteins is shut off in mature cortical neurons and thus the detection of Pax6, MASH1 or BLBP provides support to the previous hypothesis that BCs retain an embryonic phenotype (see below; Crino et al., 1997;Sisodiya et al., 2002;Mizuguchi et al., 2002;Englund et al., 2005;Thom et al., 2005;Yamanouchi, 2005;Ying et al., 2005). Radial glia are heterogeneous population of cells that give rise to neurons and astrocytes during development (Noctor et al., 2001) and that can be identified in human brain by several marker proteins including vimentin, phospho-vimentin, GFAP, the glial glutamate transporter (GLAST), Pax6 and BLBP (Wilkinson et al., 1990;Howard et al., 2006). The expression of these marker proteins changes with advancing gestational age. For example, phospho-vimentin and Pax6 (Terzic et al., 1999) are expressed in the telencephalon at 6 gestational weeks, whereas GLAST and BLBP appear by 19 gestational weeks (Howard et al., 2006). The detection of vimentin, phospho-vimentin, Pax6 and BLBP in BCs argues that these cells may derive from a cell progenitor that coincides with the onset of neurogenesis in the primordial cortex ($6-7 weeks gestation) but before 19 weeks gestation. The detection of Otx1 in DNs supports the notion that DNs derive from VZ radial glia. However, the relatively fewer number of DNs expressing Pax6, MASH1, BLBP or vimentin when compared with BCs, suggests that DNs may be more a more differentiated cell type than BCs.

[4] 429w Otx1 encodes a transcription factor that is localized to the VZ (Frantz et al., 1994), in pyramidal cells in layers V and VI (Weimann et al., 1999), and specifically in layer V corticobulbar projection neurons (Hevner et al., 2003). Otx1 identifies a subset of cortical pyramidal neurons derived from the VZ that are among the first to enter the cerebral cortex. Thus, the expression of Otx1 in FCD identifies a putative set of progenitor cells that gives rise to BCs and DNs. VGLUT1 and VGLUT2 expression in BCs and DNs corroborates the profile of Otx1 expression since excitatory cells that express VGLUTs typically are derived from the VZ. The VGLUTs are responsible for presynaptic vesicular uptake of glutamate and are expressed almost exclusively in glutamatergic neurons (Fujiyama et al., 2001;Herzog et al., 2001;Boulland et al., 2004). Thus, as suggested previously (Crino et al., 2002), VGLUT1 and VGLUT2 expression in DNs may indicate lineage derivation from a glutamatergic cell phenotype. However, there is solid evidence that BCs lack excitatory synapses (Cepeda et al., 2003;Alonso-Nanclares et al., 2005) thus, VGLUT 1/2 expression in BCs may more accurately serve as a lineage marker rather than a marker for neurochemical phenotype. Dlx homeodomain transcription factors are essential during embryonic development for the production of forebrain GABAergic interneurons (Anderson et al., 1997;Pleasure et al., 2000). A recent study has demonstrated that >50% of cortical interneurons in the mature cortex express Dlx1 (Cobos et al., 2005) and Dlx1/2 serve as markers for GABAergic inhibitory interneurons in human brain (Letinic et al., 2002). We found no Dlx1/2 immunolabeled BCs or only rare Dlx1/2 labelled heterotopic DNs in FCD and these results are supported by low concomitant expression of VGAT, a marker for inhibitory neurons derived from the MGE (Minelli et al., 2003b), suggesting that BCs and DNs do not exhibit an inhibitory phenotype and do not originate in the MGE. Reduced GABA transporter expression and decreased IPSC frequency has been observed in FCD (Calcognotto et al., 2005) and previous studies have reported diminished numbers of calbindin or parvalbumin labeled cells in FCD (Garbelli et al., 1999;Alonso-Nonclares et al., 2005;Thom et al., 2003). The co-expression of Dlx1 and MASH1 has been reported in inhibitory interneurons derived from the human VZ (Letinic et al., 2002). However, the absence of Dlx1/Dlx2 labeling suggests that MASH1 expression may reflect a VZ lineage rather than an inhibitory neurochemical phenotype. Reduced numbers of inhibitory cells may reflect a failure of cell migration into FCD, diminished generation of inhibitory precursors, or perhaps selective death of inhibitory cells derived from the MGE or VZ.

[5] 143w GFAPd labelling provides a new view of dysmorphic astrocytes in FCD that likely derive from the cortical SVZ. Alternatively, GFAPd labelling may represent an influx of newly generated astrocytes into FCD from the SVZ. GFAPd is a GFAP protein isoform that is encoded by a C-terminal alternative splice variant of the GFAP gene (Roelofs et al., 2005). GFAPd is expressed specifically by a subpopulation of astrocytes located in the subpial zone of the cerebral cortex, but not within the cortical layers or subcortical white matter (Roelofs et al., 2005). In the SVZ, GFAPd specifically marks multipotent stem cells in the adult human brain. Interestingly, a recent study demonstrated expression of another stem cell marker, CD-133, in BCs (Ying et al., 2006). Thus, the expression of GFAPd in a subpopulation of BCs suggests that these cells may retain a progenitor cell phenotype (see below).

[6] 154w Collapsin response mediator proteins e.g. CRMP-4, are expressed in the immature brain but are dramatically downregulated in the adult brain (Cnops et al., 2006), and have been identified as markers of recent cell proliferation (Nacher et al., 2002;Cnops et al., 2006) and in particular, for newly generated cells derived from the SVZ (Kee et al., 2001;Seki, 2002). The detection of CRMP4 labelled BCs supports previous studies demonstrating markers of cellular immaturity or proliferation such as MAP1B, cdk5, doublecortin, nestin, CD-133, Ki-67 and most recently Mcm2, in FCD (Crino et al., 1997;Sisodiya et al., 2002;Mizuguchi et al., 2002;Englund et al., 2005;Thom et al., 2005;Yamanouchi, 2005;Ying et al., 2006). These results suggest that either a subpopulation of CRMP4 labelled cells are actually newly generated as has been proposed for giant cells in tuberous sclerosis complex (Lee et al., 2003) or that there is a retention of an immature phenotype of BCs (De Rosa et al., 1992).

[7] 45w We show that there is a heterogeneous and differential profile of lineage and neurochemical marker proteins in human FCD that provide insights into the developmental and cellular phenotype of BCs and DNs. Further investigation into lineage marker expression may yield insights into epileptogenesis in FCD.

[1] 102w FCD specimens were obtained from 10 patients (average age 4.7 years; six males, four females; Table 1) undergoing surgery for the treatment of intractable epilepsy (Academic Medical Center, University of Amsterdam). Surgical localization of the resection site reflected the seizure focus as determined by scalp or intracranial EEG monitoring. Nine patients exhibited Engel class I or II outcome, one was deemed a class III outcome. Two neuropathologists reviewed all cases independently. Histological examination revealed abnormal cortical lamination, blurring of the grey-white junction, and the presence of DNs and BCs consistent with a pathological diagnosis of FCD, type IIB (Palmini et al., 2004).

[2] 227w Two control patient tissue groups were evaluated. First, neocortical specimens were obtained from focal cortical resection (mean age 5.2 years, including two temporal, two frontal and two parietal) for intractable complex partial seizures (Children's Hospital of Philadelphia, Academic Medical Center, University of Amsterdam). These samples did not differ in age of seizure onset, duration of seizures or numbers of antiepileptic drugs used prior to surgery from the FCD patients (Table 2). All patients exhibited Engel class I or II outcome. These patients had no radiographic evidence for FCD and the cortical cytoarchitecture was intact on histological examination of these specimens by a pathologist. These tissue samples were classified as epilepsy control specimens and were assessed as strategy to control for several variables including age, lobar location of resection, the presence of seizures, and the possible effects of anti-epileptic drugs. Second, neocortex was obtained from two temporal, two frontal and two parietal regions (Brain and Tissue Bank for Developmental Disorders, University of Maryland; Children's Hospital of Philadelphia) at necropsy from six patients who died of nonneurologic causes (mean age = 5.8 years; three males, three females; average interval to post-mortem exam <14 h). Seizures were not terminal events in these patients and none had a personal or family history of epilepsy. The cortical cytoarchitecture of these specimens was intact and these cases were classified as post-mortem control specimens.

[3] 30w All human tissue was obtained in accordance with protocols approved by Academic Medical Center, University of Amsterdam and the University of Pennsylvania Institutional Review Board and Committee on Human Research.

[4] 168w BCs and DNs were defined (Palmini et al., 2004) using morphometric parameters (maximal cell diameter, somatic area and process extension) based on cresyl violet and haematoxylin and eosin staining for quantitative cell counting analysis (Fig. 1; Baybis et al., 2004). Representative contiguous digital photos were obtained (Â20 magnification) from each tissue section using image acquisition and analysis software (Spot RT CCD camera, Diagnostic Instruments, Inc. and Phase 3 Imaging System integrated with Image Pro Plus; Media Cybernetics, Silver Spring MD). The contiguous images spanned a 1 cm 2 region of interest (ROI) and were generated under light microscopy and phase contrast optics to maximize identification of cell morphology. Prior to final assignment as BCs by the software, each ROI was visually inspected and any cellular elements erroneously included in the computerized analysis were deleted. The total numbers of morphologically identified BCs and DNs were determined in each ROI for each case and the mean (AESEM) number of BCs and DNs in ROIs were determined across All 10 cases.

[5] 173w For each immunohistochemical protein marker, determination of immunoreactivity was defined using a previously defined approach (Kyin et al., 2001). The relative optical density ratio (ODR) of labelled cells was calculated using Image Pro Plus software. The ODR serves as a labelling index for single cells versus the non-cellular background. The ODR is calculated by determining the level of pixel staining density in labeled cells versus the pixel density of the non-cellular background (the cells are digitally subtracted from the image). An ODR >3 was used as a threshold to define immunopositivity for a given antibody. The mean (AESEM) number of immunolabeled cells was expressed as a percentage of the total number of cell subtypes in all cases. Statistically significant differences in BCs expressing individual protein markers versus DNs were determined by Student's t-test (P < 0.05). Mean maximal diametre (cell diametre at its largest aspect) and somatic area (both AE SEM) were calculated using Image ProPlus software as expressed in pixel units, that were converted to microns by direct calibration with a micrometer.

[6] 204w Because Otx1 and Dlx1 are definitive lineage markers for the VZ versus MGE, we assayed the expression of each transcript in single immunolabeled microdissected cells. For successful single cell RT-PCR, cellular mRNA is first converted to cDNA beginning with in situ transcription directly on the tissue section. Following immunohistochemistry, tissue sections were treated with RNAse free Proteinase K (50 mg/ml) at 37 C for 30 min and then washed in diethylpyrocarbonate (DEPC)-treated water. To initiate in situ transcription, an oligo-dT primer coupled to a T7 RNA polymerase promoter was annealed to cellular poly (A) mRNA overnight at room temperature. cDNA synthesis was then performed with avian myeloblastosis reverse transcriptase (AMVRT; 0.5U/ml). Sections were washed in 0.5 Â SSC buffer and placed in RNAse-free water. Following in situ transcription, single Otx1 immunolabeled BCs (n = 10) were micro-dissected from FCD specimens under light microscopy using a stainless steel micro-scalpel and a joystick micromanipulator as described previously (Baybis et al., 2004). Single cells were aspirated into a pipette tip, transferred to a microfuge tube containing RT-PCR buffer, dNTPs, specific human Otx1 (NCBI accession number DQ896509) or Dlx1 (NCBI accession number BC036189) primers, and Taq polymerase. Following PCR amplification, amplicon size was analysed by agarose gel eletrophoresis.

[1] 214w Brain tissue samples were immersion fixed in 4% paraformaldehyde and embedded in paraffin. All fixed tissue sections were hydrated through graded ethanols, sectioned at 8 mm and mounted on poly-L-lysine coated coverslips. Five representative sections per case were probed with one of a panel of antibodies including BLBP (Chemicon, rabbit polyclonal, 1:500), CRMP-4, (courtesy S. Hockfield, Yale University; rabbit polyclonal, 1:100 dilution), Dlx1 or Dlx2 (courtesy D. Eisenstat; rabbit polyclonal, affinity purified, 1:100), GFAPd (rabbit polyclonal bleeding 100501, raised at the Netherlands Institute for Neuroscience; Roelofs et al., 2005, 1:250), MASH1 (Chemicon, rabbit polyclonal, 1:200), Otx1 (Santa Cruz, Burlingame, CA, USA; goat polyclonal, 1:100), Pax6 (Chemicon, mouse monoclonal; 1:250), VGLUT1 (Chemicon, guinea pig polyclonal; 1:1000) and VGLUT2 (Chemicon, Temecula, CA, USA; guinea pig polyclonal, 1:500), VGAT (Chemicon, guinea pig polyclonal; 1:100) and phospho-vimentin (MBL International, Woburn MA, USA; goat polyclonal; 1:500). All antibodies have been previously used to identify proteins in human brain tissue samples. Sections were probed with each antibody overnight at 4 C in TRIS buffer pH7.4/fetal bovine serum 2% solution followed by incubation with a biotinylated secondary antibody and visualized using avidin-biotin conjugation (Vectastain ABC Elite; Vector Labs, Burlingame, CA, USA) with 3,3 0 -diaminobenzidine. Sections used for histological analysis were dehydrated through graded ethanols and xylenes and coverslip mounted (Permount).