Malformations of neocortical development are associated with cognitive dysfunction and increased susceptibility to epileptogenesis. Rodent models are widely used to study neocortical malformations and have revealed important genetic and environmental mechanisms that contribute to neocortical development. Interestingly, several inbred mice strains commonly used in behavioral, anatomical, and/or physiological studies display neocortical malformations. In the present report we examine the cytoarchitecture and myeloarchitecture of the neocortex of 11 inbred mouse strains and identified malformations of cortical development, including molecular layer heterotopia, in all but one strain. We used in silico methods to confirm our observations and determined the transcriptional profiles of cells found within heterotopia. These data indicate cellular and transcriptional diversity present in cells in malformations. Furthermore, the presence of dysplasia in nearly every inbred strain examined suggests that malformations of neocortical development are a common feature in the neocortex of inbred mice.
Neocortical lamination emerges from the precisely orchestrated migration of newly generated neurons away from proliferative regions of the embryonic forebrain (Angevine and Sidman 1961; Rakic 1974). During corticogenesis and early postnatal periods, migrating neurons are sensitive to a number of genetic (Gleeson and Walsh 2000; Bielas et al. 2004) as well as environmental (Schmidt and Lent 1987; Ang et al. 2006) perturbations, which alter their migratory routes and affect their final positioning. Malformations of neocortical development resulting from altered migration include phenotypes with fewer neuronal lamina, inverted lamina, band or periventricular heterotopia, or phenotypes lacking lamination (reviewed in Gupta et al. 2002; Olson and Walsh 2002).
Developmental disruption of neocortical lamination is often accompanied by seizure susceptibility and/or cognitive deficits (Schwartzkroin and Walsh 2000; Schwartzkroin et al. 2004). For example, a defining feature of brains of dyslexic patients is the presence of neocortical heterotopia consisting of clusters of misplaced neurons found in the molecular layer (layer I; Galaburda and Kemper 1979; Galaburda et al. 1985; Humphreys et al. 1990). Cognitive disruption and learning impairment, defining phenotypes of dyslexia, may therefore result from altered neuronal migration and cortical circuit reorganization (reviewed in Galaburda 2005; Galaburda et al. 2006). Whether molecular layer heterotopia is also accompanied by other anatomical deficits such as dendritic and spine growth or ion channel and neurotransmitter receptor expression remains unknown.
A number of inbred mouse strains with congenital autoimmunity, exhibit neocortical heterotopia with striking similarity to those found in the dyslexic brain (Sherman et al. 1985, 1987; Sherman, Morrison, et al. 1990). Supporting the link between deficits in neuronal migration with cognitive disruption, the presence of heterotopia in these mice is associated with impaired performance on spatial and nonspatial memory tasks (Denenberg et al. 1991; Boehm et al. 1996; Balogh et al. 1998) as well as deficits in sensory processing (Clark et al. 2000; Frenkel et al. 2000). Interestingly, mice with heterotopia were recently found to have lower thresholds for convulsant-induced seizures (Gabel and LoTurco 2002). Thus, cognitive deficits and increased seizure susceptibility in mice with heterotopia is suggestive of circuit reorganization in neocortex.
Given the widespread use of inbred mice in the investigation of the genetic basis of brain development (Seecharan et al. 2003; Wahlsten et al. 2006), cognitive function (Bolivar et al. 2001; Wahlsten et al. 2005), and sensation and perception (Zheng et al. 1999; Wong and Brown 2006), we sought to determine whether heterotopia or other malformations of neocortical development are widely present in the inbred neocortex. We examined brains of 11 different inbred and 2 outbred strains of mice including several priority strains defined by the Mouse Phenome Database (http://www.jax.org/phenome). Surprisingly, we observed dysplasia of the neocortex in 10 of 11 inbred strains of mice and in the present report, we describe the cytoarchitecture and areal distribution of these malformations. Malformations were found in early postnatal brains examined. No malformations were found in outbred mice. In order to describe the transcriptional profiles of cells within neocortical malformations in C57BL/6J mice, we used the Allan Brain Atlas, a database containing in situ hybridization data assayed for >20 000 genes (Lein et al. 2007). Our database search identified the expression of a number of cell-type specific genes, ion channel and neurotransmitter receptors genes, and cytoskeletal and transcription factor genes present in cells found in malformations. These data provide greater insight into the transcriptional diversity and cellular constituents of malformations of neocortical development. Moreover, the presence of neocortical dysplasia in many different inbred mice strains and in early postnatal brains, suggests that deficits in neocortical development are a characteristic feature of the inbred neocortex.
Neocortical Histology of Inbred/Outbred Mice Brains
Adult mice (outbred CD1 and CFW, Charles River, Wilmington, MA; inbred, Jackson Laboratories, Bar Harbor, ME; Table 1) were deeply anesthetized with pentobarbital and perfused through the heart with 0.9% saline followed by a phosphate-buffered (PB; 0.1 M) fixative containing 4% paraformaldehyde. Brains were removed from the skull and postfixed with 4% paraformaldehyde at 4 °C. Brains were sectioned (50–60 μm) in the coronal plane on a vibratome and free-floating sections were then collected in PB. Sections were mounted onto gelatin-coated slides and dried overnight. Following Nissl staining, slides were dehydrated and coverslipped with Permount. Myelin staining was performed on floating sections with gold-chloride according to Wahlsten et al. (2003).
|Strain||N = male/female||Total (N)||Males w/dysplasia||% Males||Females w/dysplasia||% Females||Combined w/dysplasia||Total (%)|
|Strain||N = male/female||Total (N)||Males w/dysplasia||% Males||Females w/dysplasia||% Females||Combined w/dysplasia||Total (%)|
Three breeding pairs of C57BL10/J mice were purchased from Jackson Laboratories and maintained in our colony. Eleven pups from 3 litters were euthanized and whole-heads were placed in 4% paraformaldehyde for 3 days. Heads were transferred to 30% sucrose for an additional 5 days and then cut on a cryostat at 40 μm. Every section was collected, mounted onto gelatin-coated slides and dried overnight. Following Nissl staining, slides were dehydrated and coverslipped with Permount.
Immunocytochemistry was used to reveal the presence of neurons in heterotopia as previously described (Ramos et al. 2006). Briefly, free-floating sections were collected into different wells and washed with phosphate-buffered saline (PBS; 3 times). Sections were permeablized and blocked in 5% normal goat serum (NGS) and 0.2% Triton X-100 for 1 h. Sections were incubated with antineuron-specific enolase (raised in rabbit; Chemicon, Temecula, CA) in 0.1% Triton X-100, 2.5% NGS, and PBS at 4 °C overnight. For use with light microscopy, sections were rinsed several times with 2.5% NGS in PBS and then incubated in biotinylated secondary antibodies (biotinylated goat anti-mouse, goat anti-rabbit; 1:200, Vector Labs, Burlingame, CA) for 2 h at room temperature. Sections were rinsed 3 times with PBS and then incubated for 1 h in an avidin–horseradish peroxidase mixture. Sections were rinsed in PBS 3 times and then reacted with 0.05% diaminobenzidine in the presence of 0.0015% H2O2. Sections were collected onto gelatin-coated slides, dried for several hours, and coverslipped with Permount.
Slides were examined and representative photomicrographs were taken on a Nikon Eclipse E500 or Olympus Bx51 equipped with digital cameras. Figures were prepared in Adobe Photoshop.
Allen Brain Atlas Database Search
The Allen Brain Atlas (ABA; www.brain-map.org) is a public database containing brains from adult, male, C57BL/6J mice processed for in situ hybridization according to methods described in Lein et al. (2007). Serial coronal and/or sagittal cryostat sections are found in the ABA (25-μm thickness) for each of over 20 000 genes (Lein et al. 2007). We used the ABA in order to examine the expression patterns of a number of genes involved in neocortical development and function. We first examined the sample data set provided by the ABA (www.brain-map.org), which included 98 genes. We next created a list of ∼1200 genes including (for example) those for ion channels (Na+, K+, Ca2+, Cl+, etc.), neurotransmitter receptors (glutamate, GABA, acetylcholine, etc.), cytoskeletal proteins (tubulin), apoptosis, and cell death associated proteins (annexin, caspase, etc.), and early immediate genes (Arc, Egr1, Fos, Jun), and cell-cycle proteins. Our list also included cell-type specific genes (neuron vs. glia), neocortical layer-specific genes (I–VI), and genes enriched in neocortex provided by Lein et al. (2007). Finally, we examined the list of genes with high expression levels provided on the ABA (Brain Structure Browser; 240 genes). We reviewed photomicrographs from brains found in the ABA, which were cut in the coronal plane. A smaller number of brains cut in the sagittal plane were also examined from this list. Supplementary Tables 1 and 2 lists all of the genes for which photomicrographs were examined.
According to methods described in the ABA homepage, approximately 8 series of sections were created per brain. However, at this time we do not know how many different genes were assayed per brain. In the present report, quantification of brains with/without malformations was performed assuming that no 2 hybridized cases were taken from the same brain. Thus, if multiple hybridized cases containing malformations come from the same brains, our quantification represents overestimates. However, if multiple hybridized cases without malformations come from the same brains, our quantification represents underestimates. As described below, percentages calculated from our histological material were similar to percentages found in the ABA.
The ABA provides colorimetric quantification of gene expression for every photomicrograph in the database. Eight different colors are used to code lowest-to-highest (blue-to-red) levels of expression. Based on this scale we could determine genes with high and low levels of expression in neocortical malformations.
Hemispheric bias in the occurrence of neocortical malformations was examined for those brains found in the ABA, which were cut along the coronal plane. According to methods posted on the ABA web site (http://community.brain-map.org/confluence/display/FORUM/Home; posted by Chinh Dang on 05/07/2007; ABA Community Site/User Forum), the left side of each photomicrograph displays the left side of each brain.
Photomicrographs containing the section with the largest extent of neocortical dysplasia were printed and archived. Representative photomicrographs taken from the ABA were imported into Photoshop for preparation of figures.
Malformations of Cortical Development in Inbred Mice
In order to determine the incidence of malformations in the inbred mouse neocortex, we examined brains from 11 inbred and 2 outbred mouse strains commonly used in behavioral and physiological strain surveys including priority strains defined by the Mouse Phenome Database (http://www.jax.org/phenome). Nissl and myelin-stained vibratome sections taken from these mice revealed malformations of cortical development, deficits in callosal development, and/or hydrocephaly in all inbred strains except for SWR/J mice. In contrast to our observations in inbred mice, we did not find any malformations in brains examined from CD-1 or CFW outbred mice.
The incidence of malformations among the inbred mice varied. These data are summarized in Table 1. Among the strains that did exhibit malformations, the C57BL/10J strain (9/11 brains; 81.82%) exhibited the highest total incidence, whereas the DBA/J strain exhibited the lowest incidence (7.69%). In addition to neocortical malformations, we observed varying incidences of callosal dysgenesis. The 129P3/J strain exhibited the highest incidence of callosal dysgenesis (6/13 brains) followed by Balb/cJ (1/10 brains) and DBA2/J (1/13 brains) strains. We observed 3/11 C57BL/10J mouse brains, which were hydrocephalic.
Independent In Silico Confirmation of Neocortical Malformations in the C57BL6/J Inbred Mouse Strain
We preformed a database search of the ABA in order to confirm our findings of neocortical malformations in inbred mice, with the prediction that by chance, a percentage of mice used in the preparation of the ABA (C57BL/6J) would also contain malformations (Sherman and Holmes 1999). The ability to view photomicrographs found in the ABA with extremely high resolution, allowed us to examine material found in the ABA at high magnification, even at the single-cell level. Our database search of the ABA confirmed our observations of malformations in C57BL/6J mice as well as similar dysplastic phenotypes found in other strains (described below). From among 1244 hybridized cases we examined (>5000 photomicrographs), 385 displayed some type of neocortical malformation (30.95%; Supplementary Tables 1 and 2), a percentage similar to that observed in this strain previously (Sherman et al. 1987; Sherman and Holmes 1999). In light of the fact that our tissue sections were prepared using a vibratome (50-μm sections) and material found in the ABA were prepared using a cryostat (25-μm sections), the observation of near-identical malformations from both sets of data argues strongly against sectioning/histological artifact as the cause of the observed malformations.
Cytoarchitecture of Neocortical Malformations in Inbred Mice
Two distinct types of malformations of neocortical development were present in inbred mice from our histological data set, which were also present in material examined from the ABA. The first malformation type was characterized by the presence of clusters of cells in layer I disrupting the normal border between the molecular layer and gray matter (molecular layer heterotopia). Besides heterotopia, we observed a diverse class of malformations characterized by a more complex disruption of white and gray matter and included tissue invaginations or subduction of adjacent cortical areas.
Molecular Layer Heterotopia in Dorsal Neocortex
Molecular layer heterotopia was found in frontal, motor, and somatosensory cortices and was characterized by radial clusters of neurons in layer I extending up to the pial surface. We observed molecular layer heterotopia in 129P3/J, A/J, AKR/J, C57BL/6J, C57BL/10J, CBA/J, DBA/2J, and SJL/J strains. Large heterotopia was visible without magnification as bumps on the surface of the cortex. Representative photomicrographs of brains from C57BL/10J mice containing 2 distinct heterotopia are found in Figure 1A,B (arrows and arrowhead). As can be seen, heterotopia can be found on either hemisphere, vary in size, and generally have a circular shape. Coronal serial sectioning of the brain shown in Figure 1B revealed 2 heterotopia found exactly where we saw the bumps on the cortical surface (Fig. 1C,D). Examination of individual Nissl-stained cells within heterotopia (Fig. 1E,F) suggested the presence of both neurons and glia according to criteria used for stereological analyses of Nissl-stained tissue including cell size and shape (Ling et al. 1973; Satorre et al. 1986; Williams and Rakic 1988a, 1988b; Seecharan et al. 2003; Schmitz and Hof 2005).
Not surprisingly, we observed many hybridized cases found in the ABA also containing molecular layer heterotopia. From among 385 cases in the ABA, which exhibited malformations, 324 of these cases displayed heterotopia (84.16%; Supplementary Table 1; 114 dorsal cortical heterotopia; 210 midline heterotopia, see below).
Heterotopia was found in over much of the rostro-caudal and medial–lateral axes of the neocortex. In order to demonstrate the medial–lateral range of observed heterotopia, we plotted the position of heterotopia on an illustration of a coronal section at the level of sensory–motor cortices. As shown in Figure 2, we observed heterotopia in all locations of dorso-lateral neocortex including midline cortical areas (cingulate cortex, discussed below). In contrast, we did not observe heterotopia in lateral cortical areas such as piriform, perirhinal, and entorhinal cortices. As shown in Figure 2, heterotopia had generally similar cytoarchitecture regardless of location, with the exception of heterotopia found in midline cortical areas (discussed below).
Although the vast majority of hybridized cases we examined from the ABA were restricted to those for which coronal sections were available, we did review a number of cases cut along the sagittal plane, which also contained heterotopia. In order to demonstrate the rostro-caudal range over which heterotopia was observed, we plotted the position of heterotopia on an illustration of a mid-sagittal section. As shown in Figure 3, we observed heterotopia in much of neocortex including frontal/prefrontal cortex extending to visual cortex.
Several interesting observations could be made by examining cases in both the coronal and sagittal planes using our own material and that found in the ABA. First, heterotopia often had an inverted triangular shape with neurons becoming increasingly more “fanned-out” as they approached the pial surface. In addition, sections adjacent to those containing heterotopia often displayed increases in the thickness of the molecular layer.
We measured the size of heterotopia found in our histological database. Additionally, we estimated the rostro-caudal extent of heterotopia (not including midline heterotopia, see below) found in the ABA based on the thickness of sections used in the preparation of the ABA (25 μm) as well as the section number of each photomicrograph. Heterotopia ranged in size from 50 to 800 μm (mean = 364 μm; standard error of the mean [SEM] = 62.80 μm; measurements from 20 cases).
In order to determine whether malformations are present in the early postnatal cortex, we prepared cryostat sections of P2 heads from the C57BL10/J strain in order to examine brains in situ. We observed molecular layer heterotopia (detailed below) in 4/11 brains (36.36%) with cytoarchitecture identical to that observed in adult mice (described below). A representative example of these data is shown in Figure 1 (G,H; see also Supplementary Fig. 1). These data suggest that malformations in inbred mice are the result of a defect in neocortical development. These data are consistent with heterotopia observed in perinatal brains of inbred New Zealand black (NZB) mice as previously reported by Sherman et al. (1992).
Midline Molecular Layer Heterotopia
In the mammalian brain, midline cortical areas are characterized by the apposition of the molecular layers of both hemispheres at the interhemispheric fissure (IHF). In contrast to this normal organization, we observed molecular layer heterotopia along the IHF (Fig. 2F). Unlike dorsal cortical heterotopia, which contained clusters of cells in the molecular layer, midline heterotopia generally had diverse cytoarchitecture. Specifically, midline heterotopia often contained cells in the molecular layer as well as cells in between the molecular layers of each respective hemisphere. Representative examples of such heterotopia are found in Figure 4A (arrow). Heterotopia was also observed, which was devoid of cells in the molecular layer but had numerous cells in between the molecular layers of each respective hemisphere forming “islands” of cells. Representative examples of such heterotopia are found in Figure 4B where white matter can be clearly seen surrounding heterotopic cells. We also observed midline heterotopia with additional white matter disruption. For example additional layers of white matter surrounding heterotopic cells could be observed indicative of tissue folding and invagination (arrows in Fig. 4C).
We measured the size of midline heterotopia we found in our histological database and also estimated the rostro-caudal extent of those found in the ABA as described above (see Fig. 4D). Midline heterotopia ranged in size from 200 μm to 4 mm (average = 1.77 mm rostro-caudal; SEM = 103.43 μm; measurements from 46 cases).
Other Malformations of Cortical Development
In addition to molecular layer heterotopia we observed other malformations of cortical development with diverse cytoarchitecture and location. The feature common among many brains containing this type of dysplasia was the dramatic misalignment of parenchyma and the molecular layer resulting in areas of white matter in continuous register to gray matter. In some cases cells were found outside the limits of the pial surface of the normotopic cortex (subarachnoid space) resulting in gray and white matter, which folded over/onto normotopic cortex. More rare cases exhibited areas with complex cytoarchitecture including multiple neuronal layers separated by extra layers of white matter, or subduction of adjacent cortical areas. Including our examination of data from the ABA, we observed these types of malformations in 129P3/J, A/J, AKR/J, BALB/cJ, C3H/HeJ, C57BL/6J, CBA/J, and SJL/J strains. From among 385 cases in the ABA, which exhibited malformations, 107 of these cases displayed this type of malformation (27.79%; Supplementary Table 1).
Figure 5 contains representative examples of brains containing these diverse malformations of cortical development. Examination of brains cut along sagittal and coronal planes serves as an indicator that these malformations occur at multiple orientations such as medial–lateral as well as rostral–caudal. Examination of adjacent sections (Fig. 5A–D, E–G, H,I) clearly demonstrates the complex cytoarchitecture of these malformations. Data shown in Supplementary Figure 2 demonstrate representative examples of brains where tissue folds are evident containing labeled cells. Though often torn either during the process of brain removal or tissue sectioning, a clear pattern/trajectory of cells leading outside the molecular layer limits (arrows) could be observed resulting in a fold of tissue on top of the normotopic cortex. Adjacent sections are provided for clarity and help to demonstrate the 3-dimensional features of this complex malformation. Additional examples of malformations found in the ABA including phenotypes containing cells surrounded by white matter are found in Supplementary Figure 3. Closer examination of these cases indicates both gray and white matter on top of normotopic cortex (arrowheads point to pial surface).
Multiple Neocortical Malformations in Individual Brains
We observed that only a single malformation was present in most brains in our data set. Similarly, the vast majority of brains with neocortical malformations found in the ABA exhibited only 1 discrete malformation. However, a small percentage did display multiple malformations and generally exhibited 2 or more molecular layer heterotopia. Cases exhibiting both dorsal and midline heterotopia were found in 21 of 385 identified cases (5.45%). Cases exhibiting both midline heterotopia and other malformations were found in 15 of 385 identified cases (3.9%). Cases exhibiting both dorsal heterotopia and other malformations were found in 7 of 385 identified cases (1.82%). We observed 7 cases containing 2, distinct dorsal heterotopia. We found only a single case containing all 3 malformations. Representative photomicrographs of brains containing multiple heterotopia (dorsal and midline) evident within individual sections are shown in Figure 6. These data indicate that multiple malformations can be found within individual brains.
Laterality of Malformations
In order to determine whether there exists a hemispheric asymmetry in the incidence of neocortical malformation, we recorded the hemispheric location of each of the observed types of dysplasia found in the ABA data set. These data are shown in Supplementary Figure 4 and revealed a right hemisphere bias (61.22%) for dorsal cortical heterotopia (n = 98 samples; χ2 = 4.94, P < 0.05). In contrast we observed that midline heterotopia occur almost exclusively from the left hemisphere (99.54%; n = 219 samples; χ2 = 216.01, P < 0.001). Malformations with more complex cytoarchitecture were found to have a right hemisphere bias (58.82%; n = 119 samples; χ2 = 62.84, P < 0.001).
Transcriptional Profiles of Cells in Heterotopia
The identification of cases in the ABA containing neocortical malformations provides a novel means of identifying the transcriptional profiles of cells within malformations. We examined brains hybridized for established genes, which are known to display cell-type specificity, in order to determine the general cellular phenotypes in malformations such as neurons versus glia. Furthermore, we examined brains hybridized for established genes with neuronal subtype-specific expression, in order to identify the neuronal phenotypes in malformations such as excitatory and inhibitory neurons as well as neurons found in individual layers. Finally, we examined brains hybridized for established genes for ion channels and neurotransmitter receptors, in order to provide clues as to the intrinsic and synaptic physiology of cells in malformations.
The large number of genes for which hybridized tissue contained malformations precludes the description of them all (Supplementary Table 1). Consequently, we describe data from those brains where medium to high levels of expression (colorimetric scale provided by the ABA) were observed.
Cell-Type Specific Genes Expression in Heterotopia
Heterotopia was found in a number of cases hybridized for expression of established genes known to be enriched or exclusively expressed in neurons (Lein et al. 2007) such as Chgb, Disp2, Gpr162, Reps2. Figure 7 illustrates the expression of the neuron-specific genes, Mtap1a, and the Tubulin β2a gene (Tubb2a). To further confirm the presence of neurons in heterotopia, we performed immunocytochemisty for Neuron-specific enolase (NSE) expression. As shown in Supplementary Figure 5, NSE-labeled neurons were found in heterotopia. Together with our observations of Nissl-stained material and data described below demonstrating expression of neuron-specific ion channels and neurotransmitter receptors genes, these data indicate that neurons are found in neocortical heterotopia.
In order to test whether specific neuronal subtypes are found in malformations, we examined cases found in the ABA hybridized for expression of genes known to be enriched or exclusively expressed in specific neuronal subtypes. For example, a number of calcium-binding proteins and neuropeptides are exclusively found in GABAergic neocortical neurons (reviewed in Wonders and Anderson 2006). With this in mind, we included these genes in our search of the ABA. Brains hybridized for Parvalbumin (Pvalb), Somatostatin (SST), and Prodynorphin (Pdyn) were found to exhibit malformations containing cells with expression of these transcripts (data not shown), indicating that GABAergic neurons are found in neocortical malformations.
Malformations were found in brains hybridized for expression of genes known to be enriched or exclusively expressed in glia cells (Lein et al. 2007). For example, we observed midline heterotopia in brains hybridized for the astrocyte-enriched genes Gldc, S100β, Sox8, Sparc, and Sparcl1. We also observed midline heterotopia in brains hybridized for oligodendrocyte-enriched genes such as Car2, Fa2h, Mcam, Olig2, and Plekhb1. Representative photomicrographs of astrocyte (Sparcl1) and oligodendrocyte (Olig2) expression in heterotopia are shown in Supplementary Figure 6.
A dorsal cortical heterotopion was observed in the brain hybridized for the oligodendrocyte-enriched gene, Myelin basic protein (Mbp). Interestingly, the expression profile of Mbp+ cells within the malformations revealed a pattern similar to our results of the pattern of myelin staining found in heterotopia following gold-chloride staining (Fig. 8).
These data indicate that the 2 major glia cell-types, astrocytes and oligodendrocytes are found in heterotopia. Moreover, these data indicate a change in the pattern of oligodendrocytes and myelinated axons and suggest reorganization of axons entering and/or exiting heterotopia.
Ion Channels and Neurotransmitter Receptor Gene Expression in Heterotopia
Our list of genes from which we examined the ABA contained a number of ion channels and neurotransmitter receptors genes known to play important roles in intrinsic electrophysiological phenotypes (e.g., voltage-gated K+ and Na+ channels), synaptic transmission (e.g., GABA, acetylcholine, serotonin), and synaptic plasticity (glutamatergic NMDA receptors). As with all of the expression data found in the ABA, only a percentage of brains hybridized for any given gene will contain a malformation. Nevertheless, we identified a number of cases hybridized for ion channels and neurotransmitter receptors, which contained malformations. For example, we identified expression of Ca2+, Cl−, K+, and Na+ channel genes in cells within heterotopia, all of which are known to participate in the electrophysiological function of neurons. Representative photomicrographs demonstrating ion channel gene expression in heterotopic cells (Ca2+, Cl−, K+) are shown in Supplementary Figure 7. We also identified neurotransmitter receptors for acetylcholine, corticotrophin releasing-hormone, dopamine, GABA, glutamate (Glu), histamine, and serotonin in cells within malformations. Representative photomicrographs demonstrating expression of GABA, Glu, and acetylcholine receptors are shown in Figure 9. These data point to ion channel as well as neurotransmitter receptor diversity present within neocortical heterotopia.
Layer-Specific Gene Expression in Heterotopia
In order to estimate the neuronal birth dates of cells within malformations, we looked for malformations in brains hybridized for genes expressed only in discrete neocortical layers. This approach relies on the finding that 1) individual neocortical lamina are made up of neurons with similar birth dates and that 2) early born neurons come to populate deep layers, whereas subsequently generated neurons populate progressively more superficial layers (Angevine and Sidman 1961; Caviness and Sidman 1973; Rakic 1974). For example, examination of tissue in the ABA used to probe the Wolfram syndrome 1 gene (Wfs1) revealed a midline heterotopion. Closer examination and colorimetric analyses of Wfs1 expression revealed cells with strong expression exclusively in layer II of normotopic neocortex as well as within the heterotopion (Fig. 10A). Identical results were observed for tissue hybridized for the Start-domain containing 8 gene (Stard8), Dendrin (Ddn), and the RAS protein specific guanine nucleotide releasing factor 2 gene (Rasgrf2), which are preferentially found in layer II cells (Fig. 10B). These data are shown in Figure 10 and suggest that cells originally destined for layer II (late-born neurons) are among the cells found in midline heterotopia.
In contrast to results described above, tissue used to assay expression of genes found in deeper lamina were not observed within midline heterotopia. For example, the case hybridized for the Cysteine rich transmembrane BMP regulator 1 gene (Crim1), which is found in layer III cells, contained a midline heterotopion which lacked cells with detectable levels of expression. Similar results were observed in tissue hybridized for FEZ family zinc-finger 2 gene (Fezf2) and the chloride-channel gene, Clcn2, which are densely expressed in layer III cells and only moderately expressed in layer II cells. We observed that only lightly labeled Fezf2 or Clcn2 cells were found in the heterotopia found in this tissue; heavily labeled cells were not found in the malformation. These data are illustrated in Supplementary Figure 8. Our results suggest that layer III neurons (and perhaps deeper layer neurons) are not found in midline heterotopia. Note that we did not find malformations in hybridized cases probed for other established deep-layer genes (e.g. Ctip1, Er81, Otx1, Tbr1; reviewed in Hevner et al. 2003, Molnar and Cheung 2006).
We performed a similar analysis in cases containing dorsal cortical heterotopia hybridized for genes enriched in particular neocortical layers, in order to determine the laminar origin of cells found in heterotopia (Fig. 11). Genes specific to upper and lower layer neurons were both found in dorsal cortical heterotopia indicating that both early- and late-born neurons are present. For example, we identified cells expressing S100β in a heterotopion suggesting the presence of layer IV cells within the malformation (Fig. 11C).
Changes in Gene Expression Density in Heterotopia
Colorimetric analyses allowed us to examine possible changes in transcriptional activity for those genes hybridized in brains containing malformations. Two possible changes in transcriptional activity were hypothesized. Thus, 1) genes with low expression levels may be upregulated in cells found in malformations or 2) genes with high expression levels may be downregulated in cells found in malformations. Surprisingly, we did not find any evidence of significant changes of expression density in cells found within malformations.
Within the ABA, we identified malformations in both the sagittal and coronal-sectioned tissue hybridized for the same gene (Supplementary Table 1). In addition, as described above, individual brains with multiple malformations were often observed. This allowed for a unique test of the reproducibility of our observations of a lack of change in the expression density of cells in malformations. For example, Figure 12 contains photomicrographs of a midline and dorsal heterotopion found in coronal and sagittal-cut tissue hybridized for the voltage-gated K+ channel gene, Kcnj4. As can be seen, cells in the midline (A, lower panel) or dorsal cortical heterotopion (B, lower panel) do not exhibit changes in expression density compared with normotopic cortex. Similar results were obtained for tissue hybridized for Cox4i1, Clcn2, Grm5, and Plagl1 (data not shown). These data indicate no change in expression density among the genes we identified present in cells found in malformations.
The results from the present report detail the cytoarchitecture of neocortical malformations found in a number of inbred mice commonly used in behavioral, anatomical, and physiological studies. Molecular layer heterotopia was the most commonly observed malformation and could be classified according to spatial/areal distribution. Specifically, molecular layer heterotopia in dorsal cortex had very stereotyped cytoarchitecture and are similar to that previously observed in autoimmune mice (Sherman et al. 1985, 1987; Sherman, Morrison, et al. 1990). These heterotopia where observed across a wide range of rostro-caudal and medio-lateral positions and were often observable without magnification as bumps on the surface of the intact brain. Interestingly, heterotopia was never found in lateral cortical areas (rhinal cortices) indicating that heterotopia formation may be area specific.
Molecular layer heterotopia was also found in cortical areas along the IHF (e.g., cingulate cortex). In contrast to those observed in dorsal cortex, midline heterotopia had diverse cytoarchitecture. Cells were observed in the white mater as well as in between the 2 hemispheres. Cells were often surrounded by white matter forming “islands.” We do not know of any previous reports of midline heterotopia in inbred mice although similar malformations have been observed in mutant and transgenic mice (discussed below).
Other malformations of cortical development were also observed in a number of inbred strains. These malformations displayed a diverse cytoarchitecture and could be found in more spatially diverse locations including rhinal cortices. These malformations included phenotypes with tissue folds, invaginations, and aberrant white matter patterning.
Development of Neocortical Malformations
Molecular layer heterotopia in dorsal cortex has been studied in detail especially in autoimmune mice, which display a high incidence for neocortical malformation (Sherman et al. 1987; Sherman, Morrison, et al. 1990). These studies have revealed several structural deficits in the developing neocortex of mice with heterotopia, which likely contribute to their formation. First, radial fibers from radial glial cells are reduced in number and abnormally organized in areas underlying heterotopia (Sherman et al. 1992). Thus, migrating neurons in areas lacking the structural (and molecular) cues provided by radial glia might migrate past their intended destination and into layer I. Disruption of radial fibers along developing midline cortical areas may therefore contribute to the formation of midline heterotopia. These data emphasize the important role that radial fibers play in neuronal migration and the development of neocortical lamina (reviewed in Nadarajah and Parnavelas 2002; Nadarajah 2003).
Migration deficits where neurons fail to migrate, such as periventricular nodular heterotopia or subcortical band heterotopia, are often associated with radial fiber disruption. Interestingly, molecular layer heterotopia is also observed after genetic or environmental perturbation of radial glia. Overmigration into the molecular layer may be due to focal changes in radial glia structure or molecular signaling which dramatically affects migrating neurons. For example, changes in the terminal tufts of radial glia (which ramify in the molecular layer) may no longer provide structural and molecular “detachment” cues for migrating neurons. Changes in detachment cues may therefore result in neurons in the molecular layer.
A number of transgenic mice exhibit overmigration phenotypes such as molecular layer heterotopia, including mice with deletion of extracellular matrix proteins, glycoproteins, and transcription factors. Among many knockout mice worth mentioning, the Intergrin linked kinase (Ilk) deletion mice show a dramatic overmigration phenotype including molecular layer heterotopia in both dorsal and midline cortices (Niewmierzycka et al. 2005). Interestingly, Ilk knockouts exhibit disorganization of radial glia fibers and displacement of Cajal-Retzuis cells (Niewmierzycka et al. 2005). It is possible that one or more of these anatomical changes are also present in the developing brain of inbred mice and underlie the formation of heterotopia. Supporting the role of radial glia in overmigration phenotypes, it was shown recently that molecular layer heterotopia are found in mouse knockouts that have radial glial fibers that fail to attach to the pial/glial limitans (Haubst et al. 2006). Such is the case in mice knockouts for Laminin γ1III4, α6 integrin, and perlecan. In these mutants, overmigration phenotypes were observed despite no changes in the proliferative properties of radial glia (Haubst et al. 2006).
In addition to reduced and disorganized radial fibers, mice with heterotopia also demonstrate disruption of the glial limiting membrane (GLM; Rosen et al. 1992). Serving as an important boundary between the brain and the subarachnoid space, the GLM as well as the basement membrane (BM) keep neurons from migrating out of the neocortex. Breaches in the GLM and BM might enable neurons to migrate past layer I. Thus, neuronal “islands” along the IHF are likely formed by disruption of the GLM/BM along the midline. Likewise, tissue folds and neurons found above the pial surface, as was observed in more complex malformations, might be formed by neurons that migrate past breaches in the GLM/BM. Consistent with an important role in neuronal migration played by the GLM/BM, disruption of the GLM/BM by surgical puncture at P1, results in the neocortical heterotopia in mice (Rosen et al. 1992). Moreover, neocortical heterotopia is prevalent in the neocortex of the dreher mutant mouse (Lmx1a−/−) which exhibits abnormalities in both the GLM and in blood vessel organization in layer I (Costa et al. 2001). Likewise, bumps on the surface of the cortex as we show in the intact brain, were recently described in Laminin a6 knockouts (Georges-Labouesse et al. 1998). Recently, a number of mouse knockouts with deletion of important BM proteins demonstrate a neocortical phenotype similar (although more severe) to the malformations we observed (Halfter et al. 2002; Michele et al. 2002; Moore et al. 2002; Beggs et al. 2003; Niewmierzycka et al. 2005; Hu et al. 2007). Interestingly, heterotopic “islands” of cells in between cerebellar folia have been observed in dystroglycan knockout mice (Li et al. forthcoming). The presence of this heterotopia phenotype at the IHF and at cerebellar folia may indicate a common occurrence of heterotopia at sites of cortical apposition.
A number of knockouts display complex malformations similar to those we observed. For example, large neocortical “sulci-like” folds/invaginations and/or aberrant patterning of layer I have been observed after deletion of Cpp32 (Caspase-3; Kuida et al. 1996), Abi2 (Grove et al. 2004), and CgkI (Demyanenko et al. 2005). How perturbation of these different molecules result in the malformations of cortical development is unknown.
Cellular Constituents of Neocortical Malformations
We used the ABA to identify a number of cell-types present in heterotopia according to cell-type specific gene expression. These data demonstrate that both neurons and glia are present in heterotopia. Glial subtypes (astrocytes and oligodendrocytes) were identified as well as 2 major neuronal subtypes found in the cortex (GABAergic and glutamatergic neurons). These data are consistent with previous data using electrophysiological techniques to identify putative glutamatergic neurons as well as immunohistochemical methods to detect GABAergic neurons in heterotopia (Sherman, Stone, Rosen, et al. 1990; Gabel and LoTurco 2001).
Our search for malformations in tissue assayed for layer-specific genes, revealed the origin of neurons found in midline heterotopia. These data indicate that layer II neurons are found in midline heterotopia in contrast to deep-layer neurons, which were not observed. Thus, the presence of upper layer neurons in heterotopia suggests that these malformations may result from deficits in migration of late-born neurons. In contrast, we found evidence that neurons from both upper as well deeper layers may contribute to the cellular constituents of dorsal heterotopia. These data are consistent with the presence of layer V–like burst firing neurons in heterotopia (NZB mice; Gabel and LoTurco 2001). Future, birth-dating experiments will help determine the precise birth dates of neurons in brains with heterotopia.
Transcriptional Diversity and Density in Neocortical Malformations
Quantification of gene expression density (colorimetric) of tissue found in the ABA allowed for determination of transcriptional up- or downregulation by cells found within heterotopia compared with normotopic areas. Interestingly, we found no evidence of significant changes in expression density. In light of the fact that our data included tissue assayed for ion channel and neurotransmitter receptor genes, these results suggest that little, if any, changes in the synaptic and electrophysiological features of cells found in malformations. An important caveat to this interpretation is that not all the tissue examined for important transmitter receptor and ion channel genes contained malformations. Moreover, messenger RNA and/or protein level changes may take place during fetal and/or early postnatal periods. Supporting a model of little change in ion channel genes, recent electrophysiological recordings did not find major differences in the intrinsic firing properties of neurons found in heterotopia (Gabel and LoTurco 2001).
It remains unclear to what extent altered synaptic connections (Jenner et al. 2002) contribute to the decreased seizure thresholds in mice with heterotopia (Gabel and LoTurco 2002). Thus, even in the absence of any changes in the expression of ion channels/transmitter receptors, brains with heterotopia may exhibit excitability changes due to alteration of the balance between excitatory and inhibitory synaptic connections. Our observations of changes in the pattern of myelinated fibers in dorsal cortical heterotopia are in agreement with axonal changes in heterotopia found after neurofilament immunocytochemistry (Sherman, Stone, Press, et al. 1990) and suggest dramatic circuit and synaptic reorganization. Future studies using anterograde or retrograde tracing techniques will help elucidate circuit-level changes in the neocortex of inbred mice with malformations of cortical development.
Neocortical Malformations in Inbred Mice Used in Strain Studies and the Production of Transgenic Mice
Inbred mouse strains are an important tool toward greater understanding of the genetic and neural organization of behavior. Strain surveys coupled with genetic analyses such as quantitative trait loci (QTL) analysis are critical in the identification of genes and molecules involved in numerous behaviors and phenotypes. Recently, inbred strains have been used to study the genetic components of a wide range of behaviors such as food intake (Bachmanov et al. 2002), learning (fear: Bolivar et al. 2001; motor: McFadyen et al. 2003; spatial: Wahlsten et al. 2005), visual acuity (Wong and Brown 2006), grooming (Kalueff and Tuohimaa 2005), sexual behaviors (Burns-Cusato et al. 2004), and aggressive behaviors (Mineur et al. 2003). How neocortical malformations contribute to intrastrain variability as well as interstrain differences in these studies is unknown. However, results from behavioral and physiological studies in mice with molecular layer heterotopia are clear, indicating changes in cognitive function (Hyde et al. 2000, 2001, 2002) and sensory processing (Frenkel et al. 2000; Peiffer et al. 2001). These results emphasize the importance of histological examination of neocortical tissue from inbred mice used in strain surveys.
An important finding from the present report is the observation of neocortical malformations in nearly all inbred mouse strains examined, whereas outbred mice were normal. Although the incidences of malformations varied according to strain, these data point to malformations of cortical development as a characteristic feature of several inbred mice strains. How the diverse genotypes of the inbred mice examined result in similar malformation phenotypes is not known. Our data and that published previously by others points to a large list of inbred mice and mouse mutants with cortical malformations including many not found in our analysis such as NZB, BXSB, MRL, Snell dwarf mutant (dw/dw), and Dreher mutant (dr/dr; Sherman, Morrison, et al. 1990; Sherman, Stone, Press, et al. 1990). Interestingly, many different knockout (KO) mice have also been identified with neocortical heterotopia like that which we describe. These results point to potential overlapping/multigene regulation of neocortical malformation development. In the case of KO mice, is not clear why germline mutation or deletion of a given gene results in such focal disruption of neocortical lamination as is the case with molecular layer heterotopia. This might be the reason why, despite numerous descriptions of heterotopia in inbred/mutant/transgenic mice, there does not exist adequate models of the genetic origin of these neocortical malformations. Future studies, examining other inbred strains and employing crosses of mice with high and low incidence of heterotopia as well as QTL may provide clues to the genetic component of neocortical malformations.
Similar to data relating to neocortical malformations, it is well established that many different inbred mouse strains show callosal dysgenesis/agenesis (Wahlsten et al. 2003). Moreover, many different KO mice have been identified with callosal dysgenesis/agenesis including those with deletion of transcription factors, extracellular matrix molecules, growth factors, and intracellular signaling molecules (reviewed by Richards 2002). Thus, similar acallosal phenotypes in inbred and KO mice point to diverse genetic factors implicated in callosum development.
An interesting previous study is worth mentioning regarding the genetic origin of the neocortical malformations that we describe. For example, Sherman et al. (1992) reported that when inbred mouse embryos belonging to a strain with high incidence of heterotopia (NZB mice), are implanted into a pregnant female of a strain with 0% incidence (outbred mice), there is no change in the total incidence of malformations observed in the resulting litter. That is to say, the resulting pups of that litter (100% NZB genotype; brains examined as adults) have high incidence of heterotopia. Conversely, when embryos of an outbred strain are implanted into a pregnant NZB female, no animals in that resulting litter (100% outbred genotype) are observed to contain malformations (brains also examined as adults). Taken together, these data suggest that neither the maternal (in utero) nor early postnatal environment play a role in the etiology of malformations.
Mutant and transgenic mice have recently contributed a great deal to the identification of genes and molecules critical to brain development and plasticity. With the continued utilization of engineered mouse models it is becoming increasingly important to characterize the intrinsic anatomy, physiology, and behavior of the background strains from which these models are created (Bothe et al. 2004). Specifically, KO mice displaying phenotypes including neocortical heterotopia should be carefully compared against observed incidences of malformations in the background strain. Our results indicate that inbred mouse strains often have rather severe malformation of cortical development. Thus, histological examination can facilitate interpretation of results from studies with transgenic as well as inbred mice.
Malformation of Cortical Development in Humans
Malformations of cortical development in humans are often accompanied by developmental delay, cognitive deficits, and/or epilepsy (Schwartzkroin and Walsh 2000; Schwartzkroin et al. 2004). For example, post-mortem analysis of brains of dyslexic patients revealed the presence of numerous molecular layer heterotopia in nearly every brain examined which were identical to those heterotopia we observed in inbred mice (Galaburda and Kemper 1979; Galaburda et al. 1985; Humphreys et al. 1990). Molecular layer heterotopia has also been observed in post-mortem epileptic patients (Eriksson et al. 2001), implicating malformations of neocortical development with epilepsy. However, the role of heterotopia in epilepsy and cognitive impairment has remained controversial due to the presence of heterotopia in the general population (Meencke and Janz 1984; Lyon and Gastaut 1985; Kasper et al. 1999). For example, in addition to finding heterotopia in the brains of dyslexics, Galaburda and colleagues also found them in a sample of brains of clinically normal individuals (Kaufmann and Galaburda 1989). These authors identified heterotopia in approximately 30% of brains of clinically normal individuals which were identical in morphology and cytoarchitecture to those found in brains from dyslexics. However, the total number of heterotopia found in each affected brain (1–2 heterotopia) of clinically normal individuals was dramatically reduced compared with that found in brains of dyslexics (30–140 heterotopia; Kaufmann and Galaburda 1989).
It is not clear how to compare the incidences of malformations found in humans with those found in inbred mice or other animal models. The reported occurrence of malformations in inbred mice varied considerably according to strain and we did not find any malformations in brains from outbred mice or SWR/J inbred mice. Thus, based on our data and that reported from studies of brains of normal individuals, one would conclude that humans display greater incidence of heterotopia that that observed in outbred mice and SWR inbred mice. The incidences of malformations in A/J, C57BL/6J, C3H/HeJ, and CBA/J mice (Table 1) most closely match that observed in brains of clinically normal humans (e.g., Kaufmann and Galaburda 1989), whereas C57BL/10J mice had a very high incidence of malformation (∼80%) similar to that observed in dyslexic humans (Galaburda and Kemper 1979; Galaburda et al. 1985; Humphreys et al. 1990).
Whether our results in mice will serve as a good model for studies of malformations in humans with dyslexia and/or epilepsy will depend on further investigation of the relationship between heterotopia and these diseases. Nevertheless, the well-documented relationship between the presence of heterotopia in inbred mice and cognitive/sensory deficits (Frenkel et al. 2000; Peiffer et al. 2001) and excitability changes (Gabel and LoTurco 2002) serves as an important platform upon which to further investigate the effects of malformations of cortical development on cortical function.
Howard Hughes Medical Institute (Undergraduate Science Education Program) award to Queens College, CUNY (#52005118) supported R.L.R.; and (NS058758-01A1) to J.C.B.
We thank Sandra Giraldo and Jason Abramowitz for help with histology and Dr Ben Kest and Dr Richard Bodnar for the generous gift of inbred mice. We thank Gad Klein, Dr Lisa Gabel, Dr Matthew Sarkisian, and Dr Eric Richfield for helpful comments. Conflict of Interest: None declared.