Address correspondence to Peter B. Crino, MD, PhD, Department of Neurology, University of Pennsylvania, 3 West Gates Bldg., 3400 Spruce Street, Philadelphia, PA 19104, U.S.A. E-mail: email@example.com
Focal malformations of cortical development are highly associated with intractable epilepsy in children and adults. Most patients with focal cortical malformations and epilepsy will require epilepsy surgery. Recent studies have provided new insights into the developmental pathogenesis of cortical malformations specifically relating to alterations in cell signaling though the mammalian target of rapamycin (mTOR) pathway. Focal cortical dysplasias, hemimegalencephaly, and tubers in tuberous sclerosis complex all exhibit evidence for hyperactive mTOR signaling, suggesting that these disorders form a spectrum of malformations or “TORopathies” characterized by disorganized cortical lamination, cytomegaly, and intractable seizures. Alterations in mTOR activity in focal brain malformations provide a potential pathogenic pathway to investigate for gene mutations and to exploit for animal models. Most importantly, however, if select focal cortical malformations result from enhanced mTOR signaling, new therapeutic antiepileptic compounds, such as rapamycin, can be designed and tested that specifically target mTOR signaling.
Focal developmental malformations of the cerebral cortex or malformations of cortical development (MCD) are the most common cause of intractable epilepsy in children (Mischel et al., 1995; Krsek et al., 2008). In particular, focal cortical dysplasias (FCDs), tubers in the tuberous sclerosis complex (TSC), and hemimegalencephaly (HME) cause seizures that are resistant to medications and require epilepsy surgery. FCD has been classified histopathologically as either type I, characterized by subtle laminar disorganization of the cortex and enlarged neurons, or type II, characterized by gross disorganization or loss of cortical lamination and the presence of cytomegalic dysmorphic neurons (CDNs; Andre et al., 2004) and balloon cells (BCs; Palmini et al., 2004). Tubers in TSC represent an autosomal dominant form of type IIB dysplasia and are characterized by extensive laminar disruption, CDNs, and the presence of BCs or giant cells (GCs; Crino et al., 2006). HME often occurs in sporadic forms but can be seen in association with a variety of clinical syndromes such as hypomelanosis of Ito and linear sebaceous nevus syndromes (Flores-Sarnat et al., 2003). HME is characterized by unilateral hemispheric enlargement and severe cytoarchitectural abnormalities including laminar disorganization, BCs, CDNs, and astrocytosis (Yu et al., 2005). HME is associated with severe intractable neonatal seizures and infantile spasms and often requires urgent neurosurgical intervention.
Most focal MCD are detected by brain magnetic resonance imaging (MRI), although some can be subtle and radiographically inapparent (Lee et al., 2005). There is often, though not always, concordance between MRI findings and electroencephalography (EEG) localization of seizure onset. Many patients with focal MCD will require phase II presurgical evaluation with intracranial recording. Unfortunately, the surgical success rate in terms of seizure cure is lower than that of temporal lobectomy, and a proportion of patients will continue to have seizures despite surgery. A particularly compelling issue is that in some cases focal MCD may not be detected on preoperative MRI scanning, and thus a number of so-called nonlesional neocortical epilepsies actually reflect subtle focal MCD (Lee et al., 2005). Therefore, the role that focal MCD may play in intractable epilepsy may be even more extensive than currently believed.
Normal Cortical Development
It is important to view the molecular pathogenesis of focal MCD in the context of normal brain development, because this provides a mechanistic framework to understand how MCD might form. In particular, it is important to consider the cell types involved; the time course of development; the activity of particular cellular signaling cascades, for example, kinases and phosphatases; and the dynamics of cell proliferation (Hebert & Fishell, 2008). The human cerebral cortex forms approximately between weeks 7 and 20 of human gestation. During the early phases of corticogenesis, ongoing rounds of cellular mitosis in the proliferative telencephalic ventricular zone (VZ) leads to the formation of neurons that will migrate into the cortical plate (the nascent cerebral cortex) by an “inside-out” paradigm in which neurons destined for the deeper layers arrive first and those destined for more superficial layers arrive later. Cortical lamination proceeds in at least two predominant directions: radial and nonradial migration (Fishell & Hanashima, 2008). The majority of excitatory or projection neurons in the cortex follow radial glial cell pathways to the cortical plate from the VZ in a linear or radial course. In contrast, inhibitory interneurons follow tangential or nonradial glial-guided pathways from the ganglionic eminences (Ascoli et al., 2008). Once the cells have arrived to their appropriate laminar destination, they begin to extend dendrites and axons and make functional synaptic connections.
The normal patterning of the cerebral cortex provides numerous potential avenues to investigate what could account for the aberrant cortical architecture in focal MCD. For example, abnormalities in cell proliferation could lead to altered cell numbers destined for a particular region or cortical layer. Alternatively, defects in expression of guidance cues, trophic factors, or motility proteins could alter the way migrating cells move into the cortical plate or form layers. In either scenario, since the success of each cortical layer is dependent on that of the previous layer, it is clear that these changes could lead to disorganized or absent cortical lamination. Furthermore, because cell proliferation and cortical lamination are controlled in a temporal fashion, there may be critical developmental epochs in which focal MCD form. Changes in mechanisms that govern cell size or process (dendrite/axon) outgrowth could lead to changes in cell morphology seen in all focal MCD. A central question that remains is where in the embryonic brain focal MCD arise, for example, the VZ or ganglionic eminences. Abnormal cell types from either of these anatomic areas could lead to malformations with very distinct cellular properties. Lastly, alterations in pathways that control cell differentiation could lead to the clearly abnormal phenotype of cells in focal MCD. Each of these issues has direct relevance to understanding the potential molecular etiologies for each focal MCD subtype and clearly warrants future study.
Histopathologic Classification of Focal MCD
There are several unifying histopathologic characteristics of focal MCD, which suggest that these disorders may be linked on an etiologic or pathogenic continuum. For example, all affect, to a varying degree, a restricted cortical region and all exhibit disorganized cortical lamination. There is a variable degree of cellular dysmorphism in each MCD subtype, and in fact these morphologic differences can be used to distinguish certain subtypes. Changes in cell morphology include abnormally increased somatic size, altered dendritic arborizations, and loss of cell polarity. For example, dysmorphic neurons are characterized by abnormal orientation, altered/bifid apical dendritic arbors, tortuous axons, and the expression of the neurofilament protein SMI31 (Palmini et al., 2004). A particularly unique feature of some focal MCD is the presence of enlarged or “cytomegalic” cell types. These can be cells with a clear neuronal phenotype (cytomegalic neurons), or they may be cells that are referred to in FCD and HME as “balloon cells,” whereas in TSC they may also be referred to as “giant cells.” BCs are characterized by a thin membrane; pale, glassy, and eosinophilic cytoplasm; eccentric nucleus (or nuclei, as these cells may be multinucleated); and expression of the intermediate filament protein vimentin (see subsequent text).
Recent efforts have assembled a classification scheme for FCD into type I and type II subtypes based on morphologic differences (Palmini et al., 2004). Type I FCDs generally exhibit less disorganization of the cortical architecture and do not contain BCs. Enlarged (cytomegalic) neurons and subtle lamination defects are observed in type I FCD. In contrast, type II FCDs are characterized by more severe defects in lamination, for example, no laminar organization at all, and the presence of dysmorphic neurons and BCs. These pathologic classifications have clinical relevance, since differences in intraoperative approach and postoperative outcome have been shown to co-vary with type I or type II FCD. Widely accepted histopathologic classification schemes for tubers and HME have not been described. Of course, in view of the clear morphologic distinctions between type I and type II FCD, two pivotal questions arise: (1) are these malformations a spectrum of the same disorder; and (2) do they share molecular and developmental pathogenesis? Future studies will be necessary to answer these queries.
One interesting feature of several focal malformation subtypes is the cellular expression of protein markers, for example, vimentin, typically found in immature neurons or astrocytes, or even neuroglial progenitor cells. For example, nestin (Crino et al., 1996), MAP1B (Duong et al., 1994; Crino et al., 1997), vimentin (Lamparello et al., 2007), CD133 (Ying et al., 2005), doublecortin (Mizuguchi et al., 2002), collapsing response mediator protein (Lee et al., 2003), and Mcm2 (Thom et al., 2005) are all markers of neuroglial progenitor or stem cells and have all been reported in FCD, tubers, or HME. BCs express marker proteins found in early radial glial cells in the VZ of the developing cortex including phosphorylated vimentin, Pax6, and MASH1 (Lamparello et al., 2007). These results suggest that BCs may perhaps represent a unique immature or undifferentiated phenotype. This has potential relevance to mTOR pathway signaling, since expression of nestin, for example, is regulated via STAT3 and c-myc, by mTOR (see subsequent text).
Relationship to Epileptogenesis
There have been numerous studies over the last decade attempting to define how laminar disorganization, abnormal cellular, or altered cell phenotype relates to intractable seizures that are so highly associated with focal MCD. One hypothesis is that the cytoarchitectural abnormality alone is responsible for seizure onset. Unfortunately, this does not account for severe seizures in the setting of mild (type I) FCD. More likely, it is a combination of laminar disarray, altered cell morphology, and abnormal synaptic connectivity resulting from abnormal cytoarchitecture, coupled with specific changes in ion channel or neurotransmitter receptor expression.
Whatever the underlying cause for focal MCD, there are clear neurochemical changes that occur as secondary or corollary effects in relation to the pathologic defect. For example, several studies have demonstrated alterations in glutamate [N-methyl-d-aspartate (NMDA), α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate (AMPA), and mGlu receptors] and/or γ-aminobutyric acid (GABA) receptor subunits in focal cortical dysplasia (Crino et al., 2001; Aronica et al., 2003; Andre et al., 2004). Translational interpretation of these results has proven to be complex. Glutamate and GABA-receptor subunit mRNA expression was defined in cortical tubers, first examining whole tuber homogenates and then single microdissected cell types (White et al., 2001). Interestingly, although altered expression of five distinct receptor subunits was observed in tuber homogenates, altered expression of 12 receptor subunits was identified when single giant cells or dysmorphic neurons were compared using microdissection techniques, suggesting that the profile of receptor subunit expression was cell autonomous and thus partially obscured by whole tissue analysis. In particular, an increase in expression of mRNA encoding the GluR4 and NMDA2B receptor subunit expression was identified. These findings were recently supported and extended by analysis of glutamate receptor subunit proteins by western blotting and immunohistochemical analysis in which highly cell specific changes in receptor subunit expression were identified, in particular the GluR4 and NMDA2B subunits (Talos et al., 2008). These investigators also reported changes in NMDA1 and NMDA3 subunit expression. Therefore, potential development of new drugs that target these receptors will need to account for cell-to-cell expression variability across lesion types.
Slice recordings in vitro from FCD taken directly from the operating room simultaneously foster and complicate our understanding of epilepsy in focal MCD (Mathern et al., 2000; Cepeda et al., 2003, 2005; Andre et al., 2004). Although BCs appear to be morphologically abnormal, they exhibit bland firing properties, for example, very high input resistance, a lack of definable voltage-gated Na+ and Ca2+currents, and electrical silence. Dysmorphic neurons show few features that are distinct from normal pyramidal cells. In contrast, it appears that cytomegalic neurons may be key culprits for seizure generation. These cells exhibit very high membrane capacitance, very low input resistance, and signs of hyperexcitability including repetitive, slowly inactivating Ca2+spikes when depolarized. These results highlight several important concepts. First, although changes in a broad range of receptors may be identified in focal malformations, a few, for example, GluR4 and NMDA2B may be particularly relevant. Second, our understanding of epileptogenesis in focal brain malformations must take into account the highly heterogeneous population of cell types present and how each makes differential contributions to the overall neurochemical and electrophysiologic profile of the lesion. Third, changes in neurotransmitter expression is but one aspect of the story, and other moieties such as ion channels, synthetic enzymes, growth factor effects, and cell signaling cascades must be considered. Finally, the unique neurochemical profiles that define focal MCD could in theory be exploited to design new lesion-specific or syndrome-specific therapies that surpass the existing compounds.
Genetic Links for Focal MCD
Defining the molecular pathogenesis of focal MCD remains a critical yet elusive goal. With the exception of TSC, an autosomal dominant disorder, most focal MCD occur sporadically with no known genetic etiology, association, or family pedigrees. Although there have been prior suggestions that focal MCD result from prenatal brain insults, for example, hypoxia or viral infection, the syndromic link of select focal MCD such as HME to certain genetic syndromes bolsters the hypothesis that these malformations ultimately have a molecular genetics etiology. For example, one study found increased frequency of TSC1 gene polymorphisms in FCDIIB (Becker et al., 2002). HME has been found in association with Proteus syndrome, which results from mutations in the PTEN gene, hypomelanosis of Ito, and linear sebaceous nevus syndrome (Tinkle et al., 2005). Rarely, HME has been seen in TSC. Previous reports have suggested the hypothesis that when these malformations occur sporadically, they result from somatic mutational events occurring during brain development.
Recent evidence in TSC has allowed further consideration of focal MCD in the context of the mTOR pathway (Fig. 1). Studies beginning in Drosophila and culminating in human tissue analysis have demonstrated that TSC1 and TSC2 serve as critical negative regulators of the pivotal kinase mTOR. The mTOR pathway has been shown to govern cell growth, cell proliferation, protein translation, and autophagy in multiple organ systems [for review, see (Huang & Manning, 2008)]. mTOR is modulated via several upstream routes including via IGF-1, PI3Kinase, PDK1, and Akt, and via LKB1, STRAD, and AMPK. All of these signaling proteins converge on the TSC1–TSC2 protein complex and regulate mTOR signaling via differential phosphorylation of TSC1–TSC2. Loss of function mutations in either TSC1 or TSC2 lead to constitutive activation of mTOR and ongoing phosphorylation of several downstream proteins including p70S6 kinase, S6, and 4E-BP1. The net effect is to foster cell growth, cell size, and cell proliferation. These biochemical effects are supported by the pathologic phenotype of cells in tubers, that is, “giant” cell somata and enhanced numbers of astrocytes. In addition, the mTOR pathway is known to regulate expression of numerous proteins relevant to neural migration and dendrite outgrowth. Interestingly, targeted knockout of Tsc1 (Wong et al., 2003) or PTEN (Kwon et al., 2003), another negative regulator of mTOR, in the mouse, yields a phenotype demonstrating altered cortical lamination, enhanced cell size, increase astrocyte numbers, and seizures, suggesting that indeed the mTOR cascade and its regulation via TSC1–TSC2 is pivotal to understanding the pathogenesis of some focal MCD associated with epilepsy.
In view of the histopathologic similarities among tubers, FCD, and HME, several investigators have tested the hypothesis that the mTOR cascade is pathologically activated in these lesions as well. Using the aberrant phosphorylation of S6kinase, S6, and 4E-BP1 proteins as biomarkers, it became clear that FCD type II (Baybis et al., 2004; Miyata et al., 2004) and sporadic HME (Aronica et al., 2007) exhibited enhanced mTOR cascade signaling. Several immunohistochemical analyses revealed that the profile of phosphoprotein isoform labeling for S6kinase, S6, and 4E-BP1 was similar, although not identical, between tubers, FCD type II, and sporadic HME. These results suggest that tubers, FCD type II, and sporadic HME may fall along a continuum or spectrum with tubers in TSC as disorders of mTOR signaling, or so-called “TORopathies” exhibiting a triad of cortical dyslamination, cytomegaly, and seizures (Crino, 2007). Interestingly, we have recently found that activation of mTOR signaling distinguishes type II from type I FCD and further confirms the notion that aberrant mTOR signaling is central to the pathogenesis of sporadic brain malformations associated with cortical dyslamination, cytomegaly, and seizures (K. Orlova and P. Crino, unpublished observations).
Aberrant mTOR signaling could involve abnormalities in one of numerous proteins that modulate this pathway. Yet, FCD type II, and sporadic HME have no known family pedigrees and no other organ system involvement. Therefore, we have proposed that these focal MCD result from somatic mutations in genes that govern mTOR activation occurring within progenitor cells during brain development (Fig. 2). Cells that are progeny of the affected progenitor cell will contain the mutation and carry the effect forward into cortical development. Cells that are adjacent to the affected cells may experience secondary effects of the mutation, for example, altered expression of genes or proteins in response to the altered phenotype of the affected cell. This schema would account for the highly focal nature of focal MCD. Of course, the variable size of each MCD could depend on numerous factors such as how early in development the mutation occurred, the number of cells that carry forward the mutation, the differential migration patterns of the affected progenitor cell and its progeny, and the effects of the mutation on a cell’s ability to alter its environment. Mutations occurring earlier in cortical development or that are carried through multiple round of cell division would be expected to cause a more pervasive alteration in cortical cytoarchitecture.
Our understanding of the histopathology and pathogenesis of focal MCD associated with intractable epilepsy has advanced dramatically over the last decade. Clearly, it is now understood that the heterogeneous cell types found within focal MCD reflects disruption of the normal programs governing cortical development. Recent efforts to classify focal MCD have allowed progress in stratifying each MCD by morphologic and now cell signaling criteria. Most recently, the identification of aberrant mTOR signaling in several but not all subtypes of focal MCD suggests a pathogenic link between, for example, tubers in TSC, FCD type II, and sporadic HME. These findings are exciting because they provide new cellular pathways to study that will enhance our understanding of how focal MCD form and lead to epilepsy, but also new protein targets for therapeutic design.
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. This work was supported by R01NS045877 and Department of Defense CDMRP TSC Initiative. The author thanks L. Atkinson, M. Baybis, G. Heuer, K. Orlova, W. Parker, V. Tsai, and J. Yoon for their assistance.
Disclosure: The author has no conflict of interest to declare.