Volume 8 [Spring 2001]
|Home | Contents | Subscriptions | Archives|
A Genetic Basis for Schizencephaly
Lhx2 may play a role in Schizencephaly, septo-optic dysplasia, and Joubert Syndrome
by Allen Ho
Lhx2 is a LIM homeodomain gene expressed by proliferating cells in the developing cerebral cortex. Based on its expression pattern and similarities to Lhx2-/- knockout mice, Lhx2 was investigated as a candidate gene for the human disorder schizencephaly, a rare cortical malformation characterized by a full-thickness cleft within the cerebral hemispheres. Septo-optic dysplasia, a disorder often associated with schizencephaly, and Joubert Syndrome, a cerebellar disorder, were also investigated. The human Lhx2 genomic sequence was defined and PCR protocols were developed to amplify DNA samples from both blood and paraffin-embedded brain tissue. The human Lhx2 gene was sequenced in nine patients with one or more of these malformations, with no definite mutations found in the known functional domains sequenced so far. A heterozygous sequence alteration is reported involving a hydrophobic to hydrophilic amino acid change in a hydrophobic polyalanine stretch with unknown functional significance. Also, two other reported heterozygous sequence alterations are identified involving base changes not predicted to result in a change to the protein coding sequence. Further investigation and a larger patient cohort are required to conclusively determine the role of Lhx2 in these human brain diseases.
The primate cerebral cortex represents an unsurpassed level of complexity and integration. During its development, over one billion neurons position, network, and differentiate themselves into proper functional areas, harmonizing the cortex into a distinctive gateway to human mental capacity and reasoning. Cortical development consists of a fascinating array of genetic programs that lead progenitor cells through phases of fate determination, proliferation, migration, and differentiation. Spatiotemporal cues, support scaffolds from glial cells or axons, and other mechanisms help to smoothly guide neurons through these stages. In turn, master control of these mechanisms centers around transcription factors, DNA-binding proteins that activate cascades of gene expression. Whether expressed in gradients or localized compartments (Donoghue and Rakic 1999), transcription factors establish and regulate normal development (Rubenstein and Puelles 1994). Many essential transcription factors are encoded by homeodomain genes, which are conserved across species and are known to supervise body patterning and cell specificity.
While the past decade has seen many advances in understanding cortex formation at a molecular level, much remains to be studied before a clear, unified model emerges. One approach to studying human cortical development is to consider brain malformations: insights into the genetic origins of brain disorders may not only shed light on these diseases but also result in a greater understanding of the development process. By identifying specific genes associated with a malformation, researchers can better deduce or confirm roles that the gene might have in normal development. This research investigates the homeodomain transcription factor Lhx2 as a candidate gene for schizencephaly, a rare human cortical malformation characterized by a hole or cleft from the outer cortex surface to the lateral ventricles (fluid-filled chambers at the brain’s center) (Figure 1). Lhx2, a LIM homeodomain transcription factor, is expressed early in the embryonic forebrain during neuronal proliferation. In addition to the primary study on schizencephaly, the human brain malformations septo-optic dysplasia (SOD) and Joubert Syndrome were examined for Lhx2 mutations. Septo-optic dysplasia often accompanies schizencephaly cases and is characterized by absence of the septum pellucidum and hypoplasia (underdevelopment) of the optic nerve (Barkovich 1989). Joubert Syndrome is typified by hypoplasia of the cerebellum (Sztriha et al. 1999).
|Figure 1. The four clinical subtypes of schizencephaly. Representative transverse brain MRI images from four different schizencephaly patients are shown. Close-lipped clefts are indicated by arrows. For comparison, note the brain morphology on the normal side of the unilateral cases (A and B). (A) Unilateral close-lipped schizencephaly. (B) Unilateral open-lipped schizencephaly. (C) Bilateral close-lipped schizencephaly. (D) Bilateral open-lipped schizencephaly. ( Modified from Packard et al. 1997).|
Cerebral Cortex Development
The cerebral cortex arises with the differentiation of the anterior neural tube, the origin of the entire central nervous system. Of the three primary vesicles (forebrain, midbrain, and hindbrain) that arise from the neural tube, the forebrain region ultimately gives rise to the cortex, dividing anteriorally into the telencephalon (cerebral hemispheres) and posterially into the diencephalon (thalamic, hypothalamic brain regions). These divisions are thought to occur through a set of transcription factors that are expressed in localized expression domains (Porteus et al. 1991). The Prosomeric model proposes that the embryonic forebrain is a neuromeric (segmented) structure subdivided into unique gene expression regions by longitudinal columns and transverse segments (Rubenstein and Puelles 1994). With well-defined boundaries that organize positioning cues, these regulatory genes generate an organized framework within which the cortex can develop.
Cell Fate Determination and Proliferation. As the cell population continues to rapidly divide, the ventricular zone (VZ, located at the surface of the future lateral ventricles) arises from the original neural tube layer, while another layer of cells overlying the VZ becomes the preplate (Figure 2). In a repeated process called interkinetic nuclear migration, dividing cells migrate a short distance away from the VZ, then migrate back to divide. In division, the plane of cleavage indicates the daughter fates: a vertical plane of cleavage leads to two proliferative daughter cells equivalent to their parent, while a horizontal plane of cleavage leads to one apical migratory daughter (that navigates toward the preplate and eventual cortex) and one basal proliferative daughter (that remains in the VZ) (Chenn and McConnell 1995). The genes Notch1 and Numb are asymmetrically expressed, respectively, in the apical and basal sides of dividing cells (Chenn and McConnell 1995, Zhong et al. 1996). Thus, those daughters formed from vertical cleavage planes obtain equal amounts of Notch 1 and Numb proteins; those daughters formed from horizontal cleavage planes receive differing amounts, a possible mechanism directing the apical daughter to migrate. Other levels of restricted gene expression, such as Jnk1, Jnk2, and the caspace cascade, regulate the critical task of programmed cell death (PCD) for these cells. Selective PCD reflects differential densities of later cell subpopulations, and failure of even several targeted founder cells to die can exponentially result in hundreds more progenitors and thousands more neurons (Rakic 1995) that induce cortical aberrations (Haydar et al. 1999).
Laminar (Layered) Formation – Cell Migration and Differentiation. As a wave of determined neurons exit the ventricular zone, they form the preplate region above the proliferative region (Figure 2A). This region accumulates a second wave of post-mitotic neurons at its center, until the preplate is finally split apically into the marginal zone (MZ, ultimately layer I) and basally into the subplate. The expanding area containing this second wave becomes the cortical plate (ultimately layers II-VI). In a sequence best described as “inside-out,” neurons destined to make up layer VI first emerge and migrate to their final destination, followed by layers V, IV, III, and II in succession (Rakic 1972. These four layers must thus migrate past layer VI as well as any layer of higher designation, building the cortex layer by layer into its recognizable shape. The maturing neurons accordingly modify their base layers such that each differs in functional properties, neuronal types, and input/output connections. Neurons also form networked connections between their respective layers, creating compartments of unique functional domains comprised of all 6 layers. The resulting vertical stacking of layers and horizontal alignment of functional domains represents a dual organization of complexity found only in mammals, where each neuron simultaneously responds to both its laminar and compartmental position.
The means by which neurons originate, navigate to their correct “address,” and specialize to that address have been the subject of intense study. Recent findings have demonstrated that the two main classes of neurons in the cortex, pyramidal and non-pyramidal cells, derive from different areas of proliferation (Anderson et al. 1999, Rubenstein and Rakic 1999, Ware et al. 1999, Parnavelas 2000) (see Figure 2B). Pyramidal cells, which contain excitatory glutamate and project outside their cortical region, originate at the cortical VZ and migrate radially through the use of radial glia, which serve as convenient guides by extending radial processes through the 6 layers (Walsh and Cepko 1988) (Figure 2). The MZ/Layer I, which through the inside-out laminar construction contacts each new layer, provides additional direction through secretion of positional cues such as reelin, which may direct cell positioning (Frotscher 1998), and derailin, which may assist in disengaging migrating neurons from the glial fiber (Anton et al. 1996). On the other hand, non-pyramidal cells, which contain inhibitory GABA and become interneurons, originate at the lateral and medial ganglionic eminences (LGE and MGE, both precursors to the basal ganglia) and migrate tangentially to the cortical layers, possibly along corticofugal axons that tangentially extend to the MZ (Anderson et al. 1999, Eagleson and Levitt 1999, Parnavelas 2000).
With progenitor cells selectively dividing and younger neurons accurately migrating through an increasingly crowded field of positioned cells, it is clear that both intrinsic influences from the proliferative areas (a Protomap model) and extrinsic factors from outside the cortex (a Protocortex model) that control regionalization. The Protomap model likely involves patterning centers (tissues that produce morphogens) and transcription factors (Rubenstein et al. 1994), as evidenced by gene expression gradients and compartments that exist well before any outside influences penetrate the cortex (Donoghue and Rakic 1999). The Protocortex model likely plays a role in finalizing neuronal identity: axons from the thalamus, for example, arrive as late as Layer IV formation but nonetheless interactively synapse with maturing neurons (Mackarehtschian et al. 1999). It should be emphasized, however, that a host of transcription factors discretely expressed even earlier in neurogenesis plays a key though largely undefined role in early development. One of these transcription factors is Lhx2, a LIM-homeodomain gene expressed in progenitor cells and required for cortical neurogenesis.
Lhx2 as a Candidate Gene for Brain Malformations
|Figure 4. Representative MRI of Patient MR (parasaggital view) with unilateral open-lipped schizencephaly. The cleft (arrow) extends from the outside cortical surface to the lateral ventricles (center) and is filled with cerebrospinal fluid. The patient is an adult with relatively mild symptoms.|
There are no less than 25 homeodomain genes expressed in the developing forebrain (Rubenstein and Puelles 1994). The LIM homeodomain family has been implicated in cell fate determination, with members of the family expressed in motor neurons, the pituitary gland, and the developing neuraxis. The LIM homeodomain proteins contain the conserved homeodomain as well as two unique LIM domains: while the homeodomain encodes a DNA-binding domain, the LIM domains are double-zinc finger motifs thought to function as protein-protein interaction modules, mediating protein complex formation and modulating specific protein activity (Dawid et al. 1998). The LIM domain serves two divergent purposes in relation to the homeodomain. By default, the LIM domain binds to the homeodomain intramolecularly, thus preventing inappropriate DNA binding and activation of transcription. Yet the LIM domain releases the homeodomain in the presence of relevant proteins, forming a transcriptional complex that aids homeodomain binding and gene expression (Curtiss and Heilig 1998, Dawid et al. 1998). Thus, the LIM domain both inhibits and enhances homeodomain activity.
Lhx2 is one of several LIM homeodomain genes known to be expressed preferentially in the developing forebrain and dorsal midline (Xu et al. 1993, Porter et al. 1997). Like other regulatory transcription factors, Lhx2 is detected in discrete patterns: it is expressed spatially in the proliferation zones and temporally when progenitor cells commit to a specific neuronal fate. Moreover, expression continues in postmitotic neurons in layers II-VI of the cortex after birth (Xu 1993), though it is most concentrated in Layers II-III (Donoghue and Rakic 1999). Other studies indicate that overexpressed Lhx2 increases progenitor cell number, while underexpressed Lhx2 in turn results in a decrease in progenitors and inhibition of neuronal differentiation (Walsh et al., unpublished). Morphologically, Lhx2-/- knockout mice have missing hippocampi and hypoplasia of both the cerebral cortex and basal ganglia (Porter et al. 1997). Thus, Lhx2 plays a critical though not entirely defined role in cortical development. Specifically, Lhx2 almost certainly interacts with other proteins in complex modules, and thus has the potential to help form a “combinatorial code” that could determine cell fate. For example, Lhx2 might selectively form complexes with different proteins to activate different genes. The absence or malfunction of Lhx2 could drastically affect these processes, making it an appropriate candidate gene for human brain disorders involving defective neuronal proliferation or differentiation. The purpose of this study is to investigate its potential role in relevant human diseases.
Brain Malformations: Schizencephaly, Septo-optic Dysplasia, and Joubert Syndrome
The classification of human diseases involving cerebral cortical malformations has been limited in the past to morphologic observation of the brain post-mortem. Yet these disorders are generally considered to be genetically heterogeneous, where different genes can cause the same malformation. Conversely, one gene could be responsible for different malformations. However, with evolving technology and methods, progress is being made in understanding the causative roles of novel genes and correlating them more accurately with physical appearance. Schizencephaly, septo-optic dysplasia, and Joubert Syndrome are of interest because of their hypothesized link to Lhx2.
Schizencephaly. Schizencephaly is characterized by a cleft stretching from the pial surface to the lateral ventricles. There are two major forms of this rare cerebrum defect: closed-lip schizencephaly (Type I), with narrow clefts and lips fused in certain areas (pial-ependymal seam); and open-lip schizencephaly (Type II), with widely separated walls encompassing an excess cerebrospinal fluid space. Clefts may be unilateral or bilateral in both types, thus creating four distinct subtypes (Figure 1). The sides of the clefts are generally lined with heterotopic gray matter (an abnormal accumulation of neurons) (Granata et al. 1997). Associated human malformations that commonly accompany schizencephaly include hypoplasia of the corpus callosum and septum pellucidum, focal cortical dysplasia, coloboma of the retina, and hydrocephalus (Capra et al. 1996, Granata et al.1996). Clinically, most schizencephaly patients experience motor difficulties, mental deficits, and seizures that can range from mild to drastic. Though it should be noted that studies on schizencephaly are limited by relatively small sample sizes, comparison across studies indicates that the degree of motor and mental deficiencies is proportional to the severity of the malformation but not to the severity or presence of epileptic seizures (Granata et al.1996, Packard et. al. 1997, Capra et. al. 1996, Kuban et al. 1989). Thus, open-lipped, bilateral schizencephaly patients display the worst clinical symptoms.
Based on its 4 different phenotypes and associated malformations, schizencephaly was originally thought to have multiple nongenetic causes, including toxic, infectious, and metabolic (some likely in combination). One accepted hypothesis involves a developmental defect in the blood vessels supplying the cerebral cortex (Barkovich and Kjos 1992), resulting in tissue death and cleft formation due to lack of oxygen (in utero vascular insufficiency). More recently, however, schizencephaly has been proposed to have purely genetic origins resulting in abnormal proliferation of progenitor cells during cortical development. The detection of mutations in the homeodomain gene Emx2 of some schizencephaly patients (Brunelli et al. 1996, Brunelli et al. 1997) has confirmed that at least some schizencephaly cases result from germline mutations. Emx2 is expressed in restricted areas of the developing mammalian forebrain, including areas that develop into the cerebral cortex (Yoshida et al., 1997).
The discovery of Emx2 mutations in both sporadic and familial cases of schizencephaly (Brunelli et al. 1996, Brunelli et al. 1997, Granata et al. 1997) marked an important advance in establishing genetic causes for brain malformations. Accordingly, this finding has spurred interest in the potential roles of similar candidate genes that may cause specific cortical disorders, including schizencephaly cases that cannot be explained by Emx2 malfunction. The lack of Emx2 mutations in most schizencephaly patients has increased the likelihood of other genes being involved. Like Emx2-/- knockout mice, the Lhx2-/- knockout mice show evidence of reduced cerebral hemispheres, drawing comparisons with human schizencephaly.
Septo-Optic Dysplasia. Septo-optic dysplasia (SOD), often associated with schizencephaly, is generally characterized by an absence of the septum pellucidum, hypoplasia of the optic nerve, and hypothalamic-pituitary dysfunction leading to hormonal disturbances (Barkovich 1989). SOD has proved difficult to classify due to its wide variety of symptoms that include gonadotropin insufficiency and early onset of puberty. Like schizencephaly, some cases of SOD have been proposed to be a vascular condition (Lubinsky 1997). However, mice deficient in the Hesx1 gene have phenotypes analogous to human SOD, and promising Hesx1 mutations have been discovered in some SOD patients. Lhx2, like Hesx1, is expressed in prospective forebrain tissue, and so could account for a number of SOD cases (Dattani et al. 1998). More importantly, Lhx2 knockout mice possess underdeveloped optic nerves and missing septum pellucida, (Monuki, personal communication) the cardinal features of SOD. Depending on Lhx2’s relationship to schizencephaly, this likelihood is raised since nearly half of SOD patients also suffer from schizencephaly.
Joubert Syndrome. Joubert Syndrome is a rare autosomal recessive disorder characterized by a partial or complete agenesis of the midline region of the cerebellum (the cerebellar vermis). Consequently, abnormalities such as dysplastic axons connecting the cerebellum to the brainstem and cerebral cortex result in central nervous system, respiratory, renal and eye defects (Ni Scanaill et al. 1999). Associated symptoms sometimes include abnormal eye movements, developmental delay, and severe mental retardation (Saar et al. 1999). Like schizencephaly and SOD, Joubert Syndrome is likely to be genetically heterogeneous due to the fact that these associated symptoms range from nonexistent to acute.
Recently, one out of two Joubert Syndrome families studied showed linkage to the chromosome region 9q33-34.1 (Saar et al. 1999), the same expanse to which Lhx2 maps (Wu et al. 1996). Because Lhx2 knockout mice die before cerebellum development can be adequately observed, no absolute morphological evidence exists to elucidate Lhx2’s function in that area. Interestingly, however, Lhx2 expression domains also encompass the cerebellar dorsal midline region. Due to linkage analysis and its expression in the cerebellum, Lhx2 is investigated as a positional candidate gene for Joubert Syndrome.
Materials and Methods
Known human mRNA sequence was put through a BLAST search (http://www.ncbi.nlm.nih.gov) to find similar sequence information in the Genbank database. Relevant human genomic and mRNA submissions were identified and compiled to create a consensus sequence. This consensus was then used to identify and locate dispersed exon regions.
Primers were designed to produce products of no more than 250 base pairs. Exons were divided accordingly, with appropriate primers overlapping each other. Primers ideally were flanked 20 base pairs outside target regions that in sum encompassed the entire coding sequence, splice acceptor sites, and splice donor sites (common sites of mutation).
Primers were selected using Primer3, an online primer picking program located at the Whitehead Institute for Genomic Research website (http://www-genome.wi.mit.edu/cgi-bin/primer/primer3_www.cgi). Parameters were the following: oligo length between 17 and 22 base pairs, melting temperature between 530C and 630C, Max 3’ Stability of 5.0, Max Complementarity at 5.0, Max 3’ Complementarity at 3.0, GC% between 40% to 60%. Potential primers were screened against a human mispriming library for interspersed repeats and sequences to avoid. They also were reexamined for secondary structure, hairpin loops, and primer-dimer formation through Oligo, another primer picking program, and Sigma-Genosys/Fisher Scientific Ltd., the company from which the primers were ordered.
Because of the scarcity of schizencephaly patients, various approaches were used to find human patients. Prospective cases were reviewed courtesy of Dr. Edwin Monuki at the Boston Children’s Hospital Pathology Department, and relevant tissue embedded in paraffin blocks were pulled for analysis. A patient database maintained by Dr. Christopher A. Walsh was also referenced for patients with blood on file. Finally, a variety of doctors and researchers with schizencephaly patients were contacted and requests for blood samples were made. Patient or patient’s family consent was obtained. When possible, MRI scans or reports were obtained for review and confirmation of the schizencephaly diagnosis.
For blood and lymphocyte samples, DNA was isolated using the preparation and lysis protocols in the QIAGEN Genomic DNA Handbook (updated 8/99). DNA concentration was then measured by standard spectroscopy. For paraffin-embedded tissue, 2 different protocols were employed to extract DNA. Each had variable effectiveness depending on the sample.
Protoco1 1. For each sample, 10 sections of 10 um thickness were sliced from a microtome and placed in a 1.5 ml Eppendorf tube. 600 ul of xylene were added and the tube was mixed for 15 minutes to dissolve the paraffin. The mixture was centrifuged at 14,000 rpm for 5 minutes, and the supernatant discarded. This xylene step was repeated 3 more times.
To remove xylene, 600 ul of 99% ethanol was added to the pellet and mixed briefly. The tube was centrifuged at 14,000 rpm for 5 minutes and the supernatant discarded. This ethanol step was repeated 3 more times. The pellet was then dried in a speed vacuum.
A digestion mix was added to the pellet, consisting of 7.5 ul Tris 1M (pH 8.0), 15 ul EDTA (10 mM), 0.75 ul Tween 20, 3 ul Proteinase K (20 mg/ml), and 123.75 ul distilled H2O (dH2O). The tube was incubated at 370C for 24 hours and mixed intermittently to facilitate digestion.
The tube was then incubated at 950C for 10 minutes to inactivate Proteinase K. The sample was amplified via PCR in 1:1 and 1:4 dilutions to diminish the effect of inhibitors.
Protocol 2. For each sample, 10 sections of 10 um thickness were sliced from a microtome and placed in a 1.5 ml Eppendorf tube. 1 ml of xylene was added, and the tube was gently mixed for 5 minutes and then centrifuged at 14,000 rpm for 5 minutes. The supernatant was discarded. This xylene step was repeated 2 more times.
The tissue pellet was then resuspended in 100% ethanol and centrifuged at 14,000 rpm for 2 minutes. The supernatant was discarded and the pellet dried at 950C for 3 minutes to evaporate any remaining ethanol.
The pellet was then resuspended in 500 ul 10XTE, 15 ul Proteinase K (20 mg/ml), and 25 ul 10% SDS. The tube was placed in a 555C water bath overnight.
500 ul of phenol:chloroform:isoamyl alcohol (25:24:1) was added to each sample, mixed to emulsify the phases, and then spun at 14,000 rpm for 2 minutes. The supernatant was transferred to a new tube. This pheno:chloroform:isoamyl alcohol step was repeated.
Following this, an equal volume of chloroform:isoamyl alcohol (24:1) was added to each tube, mixed, and spun as before. The supernatant was again transferred to a new tube, and 100% ethanol equal to twice the supernatant volume was added. The tube was incubated at 700C for 30 minutes to deactivate the Proteinase K, and then centrifuged at 14,000 rpm for 10 minutes. The supernatant was discarded.
The pellet was washed in 70% ethanol and spun at 14,000 rpm for 10 minutes as before. The supernatant was discarded. The pellet was air dried, then resuspended in 30 – 50 ul 1 minute.
|Figure 5. Histologic section of Patient A82 (coronal view) with bilateral open-lipped schizencephaly. Both clefts (indicated in part by arrows) extend symmetrically from the cortical surface to the lateral ventricles (center). The patient was a premature new-born who died shortly after birth.|
Extracted genomic DNA was amplified for sequencing in 30 ul or 50 ul PCR reactions. Reaction mixtures were placed in PCR strip tubes without caps, and contained Qiagen 10X buffer (containing 15 mM MgCl2), dNTPs, Taq polymerase, distilled H2O, the primers (100 ng each) and DNA. Mineral oil was placed on top to prevent evaporation during cycles, and the heat lid of the thermal cycler was disabled. Denaturation was at 940C for 30 seconds, annealing was at 600C for 1 minute, and extension was at 720C for 1 minute. This cycle was repeated 34 more times, and followed by a final 10-minute extension period at 720C. Products were stored at 40C.
Relative quantities varied for each primer pair and for each DNA template, and in a number of cases the use of Qiagen Q solution and additional MgCl2 was employed. Also variable depending on primer pairs and DNA templates was the use of a “manual” versus “modified” hot starts, both of which theoretically reduce the formation of non-specific PCR products and primer-dimers. A manual hot start placed the reaction mixture minus the Taq and 10X buffer into the thermal cycler and heated it at 940C for 3 minutes before adding the 2 reagents. A modified hot start simply preheated an empty thermal cycler to 940C before adding the entire reaction mixture.
PCR products were run with 6X loading dye on 2% agarose gels stained with ethidium bromide. Products were visualized via a ultraviolet light source (Figure 6).
PCR products were run on a 2% agarose gel and band sizes were confirmed with a DNA ladder. Bands were cut out and the QIAquick Gel Extraction Kit (microcentrifuge protocol) was used to purify the DNA and prepare it for sequencing. Purified products were speed vacuumed to increase DNA concentrations.
Sequencing reactions were done in 10 ul volumes, containing 6 ul DNA template, 3 ul diluted Big Dye (containing an enzyme/dNTP/ddNTP mixture), and 1 ul of the appropriate primer (20 pmol). Denaturation was at 940C for 30 seconds, annealing was at 550C for 15 seconds, and extension was at 600C for 4 minutes. This cycle was repeated 29 more times, and the products were stored at 40C.
Automated sequencing was performed on an Applied Biosystems 377 Sequencer. Tracking lines were manually corrected to optimize the reading of lanes, and then submitted to the ABI Prism Sequencing Analysis program. The generated sequences were edited and used to form contigs on Sequencher 3.11 (Gene Codes Corporation). The sequences were aligned against the compiled consensus sequence to perform mutational analysis.
Clinical Findings of Patients
A total of seven schizencephaly patient DNA samples were obtained for sequence analysis, a sample size equivalent to similar studies such as those done on Emx2 (Brunelli et al. 1996). Of these seven, at least two are additionally inflicted with septo-optic dysplasia. Two Joubert Syndrome patient samples were also acquired. As this is an ongoing investigation in its beginning stages, the recruitment of additional patients for this study is not complete.
|Figure 6. Representative Image of PCR reaction products (Patient 231). See Materials and Methods section for more detail. For nearly all patients, all primers except for those flanking the Exon 3B region consistently produced the expected band size.|
Patient records were obtained where possible and reviewed for confirmation of the schizencephaly diagnosis, as well as for other associated malformations. Patients FM (courtesy of Dr. Michael Innes) and A82 (Dr. Edwin Monuki) suffered from open-lipped bilateral schizencephaly, while Patient MR (Dr. Christopher A. Walsh), S89 (Dr. Monuki), and S94 (Dr. Monuki) had close-lipped, unilateral schizencephaly. The reports of Patients JD (Dr. Andrew Blum) and AS (Dr. Michael Berg) were unavailable beyond gross diagnoses of schizencephaly. Available representative images are displayed for Patient MR (Figure 4) and Patient A82 (Figure 5). Additionally, Patients FM and A82 showed indications of septo-optic dysplasia. Patients varied in age and severity of symptoms, from newborns who died after premature birth to grown adults with relatively minor disabilities. All cases are apparently sporadic so an inheritance pattern could not be established, though Patient FM is known to have an unexamined maternal uncle inflicted with mental retardation and seizures. In addition, Patients 231 and 520 (both courtesy of Dr. Joseph Gleeson) represent cases of Joubert Syndrome. Though at least some instances of this disease are familial and perhaps inherited disorders, the parents of both patients studied here are normal.
DNA was extracted from blood in three patients (AS, MR, and JD) and paraffin-embedded tissue in three others (A82, S94, 989). DNA from Patients FM, 231, and 520 arrived already suspended in solution from collaborators.
Establishment of Human Lhx2 Genomic Sequence
Through searches on Genbank for human sequence submissions relevant to Lhx2, a sequence of the genomic Lhx2 region was compiled that included the transcription factor’s five exons and intervening introns. Although multiple mRNA sequences of lhx2 were registered (Genbank Accession Numbers U11701, NM_004789.1, AF124735.1, AF124734.1), the genomic sequence itself was unknown. Lhx2 maps to the human chromosome region 9q33-34.1 (Wu et al. 1996), and BLAST searches in that expanse revealed a BAC (Genbank Accession Number AC006450.13) in the process of being sequenced at the Sanger Center. The partially sequenced information contained the majority of the coding region and intervening introns. To eliminate sporadic sequence errors and discrepancies among the submissions, sequences were aligned and a consensus of the mRNA was generated. Moderate differences were found in the region of Exon 1, such that band size and sequence initially had to be estimated. However, subsequent sequence data of the nine patients confirmed the alignment. In addition, the genomic BAC clone ended within Exon 5. Coding sequence from the mRNAs U11701 and NM_004789 extended the consensus Exon 5 sequence further, but as they are incomplete entries part of the 3’ UTR (untranslated region) could not be realized and sequenced. However, it is generally believed that the 3’ UTRs of most mRNAs are poorly conserved and thus less likely to contain a relevant mutation.
Figure 7. Heterozygous sequence alterations in the Lhx2 gene from patient samples. Chromatogram data is generated from the ABI Prism Sequencing Analysis Program and aligned using Sequencher 3.11. Consistently spaced peaks with little background indicate strong signals. The nucleotide N here indicates two competing peaks, illustrating a heterozygous sequence alteration (highlighted N’s). Both DNA single strands were sequenced to confirm the base pair changes. In the case of Patient 989, the base change was in the region of Exon 4A and Exon 4B primer overlap, showing 3 independently sequenced chromatograms. (A) Patient AS, Exon 3. (B) Patient 989, Exon 4. (C) Patient 520, Exon 5. Colored peaks correspond to different bases: Green - A, Blue - C, Black - G, Red - T.
The identification and preliminary characterization of the human Lhx2 gene was critical for designing primers against intron regions not present in the mRNA submissions (Figure 3). Since sequencing often generates poor signals at the beginning and end of each PCR product, the ability to design primers 20 base pairs outside each exon ensured a higher quality signal in the important target regions. It also enabled study of splice acceptor and splice donor sites located at the intron/exon boundaries. As the regions of more and more patients were sequenced, the nucleotide information was used to create a more accurate consensus sequence ultimately comprised of more than ten different sources for most zones.
Figure 8. Comparison of Polyalanine String Across Species. The base change in Exon 3 of Patient AS causes a hydrophilic serine to be encoded in place of a hydrophobic alanine. The boxed areas indicate the amino acid and nucleotide of interest, which across these species are strictly conserved. However, the polyalanine tract in general is absent from homologous regions in the chick and Drosophila. Capitalized letters indicate amino acids, while lowercase letters indicate nucleotides.
Ten primer pairs designed to amplify regions of approximately 250 base pairs encompassed all five exons (Table I). With the exception of Exon 3B, all primers generally amplified products with expected band sizes (Figure 6). As this is an ongoing study, regions of some patients have not yet been sequenced adequately. Paraffin extracted DNA, for example, varies in quality depending on length of fixation prior to embedding because of the use of formalin, which cross-links proteins to DNA and thus obstructs Taq polymerase. In all cases, various inhibitors and sub-optimal conditions may also have complicated PCR using these primers. Exon regions amplified and sequenced thus far are on Table II.
Of the nine schizencephaly patients examined thus far, three single base pair changes in single alleles were discovered whose functional significance is not yet certain. No unambiguous mutations have yet been identified in the known functional domains of Lhx2, though important regions, including the Exon 3B expanse, remain to be sequenced.
A sequence alteration found in Exon 3 of Patient AS involves a conservative amino acid change from alanine (base sequence GCA) to serine (base sequence TCA) (Figure 7A). Alanine contains a hydrophobic aliphatic side chain, while serine is uncharged but polar with a hydrophilic hydroxyl side chain. Upon sequence comparison with other species that contain Lhx2 homologues (Figure 8), the mouse and rat were found to be conserved for this amino acid and nucleotide (GCA). The apparent rigidity of the base pair across species signifies the potential significance of alanine at this position, and points to the unusual nature of this heterozygous change. Since this change was located in the difficult-to-amplify Exon 3B region, it was not possible to adequately compare the alteration with other schizencephaly patients. The only other case with data in this other region, Patient MR, did not show this base change.
Two other heterozygous base changes were discovered, both located on wobble positions (the third position of the codon) and causing no amino acid change. One alteration was found on Exon 4 of Patient 989, altering the codon from CCC to CCG (Figure 7B). Both codons code for proline. Another change was found on Exon 5 of Patient 520 (a Joubert Syndrome case), changing the codon GAT to GAC (Figure 7C). Both codons code for asparatic acid. Because there is no evident change in the gene’s characteristic properties, the potential impact of both of these base changes is limited and likely to be neutral.
It should be mentioned that within the last weeks of experimentation, correctly sized bands appeared in negative PCR controls, suggesting DNA contamination. After running various controls, the most likely source seems to be distilled water used to suspend the primers in aliquots. Since negative PCR controls were not run on all ten primer pairs for each PCR reaction throughout the months of experimentation, it is uncertain as to when contamination occurred. The brightness of the contamination bands was approximated to be no more than 5% of the brightness shown by experimental bands.
The discovery of Emx2 mutations in both sporadic and familial cases of schizencephaly (Brunelli et al. 1996, Granata et al. 1997) represented a significant breakthrough in establishing genetic causes for brain malformations, increasing the likelihood that other genes could cause similar defects. Like Emx2, Lhx2 is a homeodomain gene expressed in the developing forebrain and thought to be involved in controlling neuronal proliferation. Both Lhx2-/- and Emx2-/- knockout mice exhibited reduced cerebral hemispheres and hippocampi (Pellegrini et al. 1996, Porter et al. 1997). Furthermore, the optic nerve and septum pellucidum of Lhx2-/- mice were undersized or absent (Monuki, personal communication), cardinal features of septo-optic dysplasia (SOD) and common to many schizencephaly patients. The morphologic evidence through knockout mice, the confirmation of genetic causes for schizencephaly through Emx2 mutations, and the likelihood of schizencephaly as a genetically heterogeneous disease all establish Lhx2 as a strong candidate gene for this cortical disorder.
The approach and protocols established in this study serve several useful purposes for future Lhx2 studies, and moreover have potential applications for mutational analysis. As the first steps of a project that will involve multiple genes and human brain disorders, this study’s findings may also steer future related efforts in mutational analysis. The compilation of indiscriminate sequence fragment submissions, previously not attempted for this gene, marks the initial identification of the Lhx2 gene in the human genome. When compared against the compilation, the successful PCR design and sequencing data of nine independent human DNA patients allow the further development of a rigorous consensus sequence for Lhx2 specific to humans. For studies of Lhx2 in development, preliminary characterization of this gene based on this consensus confirms that the human Lhx2 gene is indeed evolutionarily conserved across species, with its characteristic functional domains similarly broken up among exons. For mutational studies, the improved base pair accuracy serves as an invaluable reference, as rigidly correct sequence is required for picking up differences as sensitive as single base changes. As the functional relevance of Lhx2 becomes more apparent and this transcription factor is implicated in more human brain malformations, the groundwork laid in this study will facilitate and perhaps confirm its characteristic role. In addition, the recruitment of nine patients with extremely rare brain disorders establishes a database of patients by which future candidate genes can be tested. The number, though small by statistical standards, is similar to that of other studies and relatively sizeable considering the rarity of the diseases. In the scenario that schizencephaly, SOD, and Joubert Syndrome are hypothesized to be the result of mutations in other genes that malfunction independent of or in conjunction with Lhx2, a source of testable DNA will be available. This assembly of samples included the successful extraction of DNA from paraffin-embedded tissue, a protocol not previously attempted in this laboratory that will open up a new source of patient DNA.
Mutational Analysis of Lhx2
Neither heterozygous nor homozygous mutations were found that affect the regions known to be most critical to Lhx2’s function as a transcription factor. These areas include the two LIM domains (Exons 2 and 3), which modulate DNA binding through protein-protein interactions; the putative nuclear translocation signal (Exon 3), which is required for Lhx2 protein entry into the nucleus; the homeodomain (Exons 4 and 5), which binds DNA; and the splice acceptor/donor sites (all intron/exon boundaries), which are required for accurate construction of processed mRNA. These areas require rigid sequence conservation for proper function. Moreover, a very high frequency of mutations found in other transcription factors linked to human diseases is found in equivalent areas (Semenza 1999), leading to well-defined consequences. For example, of the schizencephaly patients with confirmed Emx2 mutations, four had splice site mutations, while one had a frameshift mutation that altered the homeodomain. These changes lead to truncated or meaningless coding sequences.
Although no explicit mutations were found, three sequence alterations (one in a Joubert Syndrome patient) involving single alleles were discovered. These putative base pair changes are located in exon regions outside of known functional domains and as such, their consequences in causing malformations are not certain. One potential polymorphism changed the amino acid encoded (a replacement site alteration), while two others led to the same amino acid encoded (silent site alterations). The replacement site alteration (Patient AS) was found between the second LIM domain and the nuclear localization sequence, while the silent site alterations were immediately preceding (Patient 989) and following (Patient 520, a Joubert Syndrome patient) the homeodomain region (Figures 3, 7A-C).
The transversion in Patient AS is especially noteworthy for several reasons. First, the change at the codon’s first position from a purine to a pyramidine (a transversion) leads to a change from a hydrophobic alanine (GCA) to a hydrophilic serine (TCA) (Figure 7A). This fundamental difference in amino acid character could presumably affect the folding of the protein into its proper conformation. Significantly, the potential effect of this hydrophilic anomaly is enhanced because it occurs in an extremely hydrophobic region that includes a stretch of 10 consecutive alanines (discussed below). Second, serine is unique because of its uncharged but polar hydroxyl attachment, a functional group that is often the target of post-transcriptional modification, including phosphorylation and glycosylation. Phosphorylation by kinases not only adds mass to the protein but also introduces negative charges that produce electrostatic effects. Glycosylation, on the other hand, adds massive carbohydrate side chains to the protein and significantly affects protein folding during processing at the endoplasmic reticulum. The introduction of either group in that region could severely disrupt Lhx2 protein structure and function. Although phosphorylation and glycosylation of serine are also dependent upon the makeup of surrounding amino acids, making it unlikely that any randomly formed amino acid will be modified inappropriately, the possibility remains.
Lastly, Lhx2 in this region interestingly encodes a set of 10 consecutive alanines, of which the observed base change replaces the first (Figure 8). This atypical, hydrophobic region may be necessary for Lhx2 function. A number of transcription factors are known to contain similar polyalanine stretches outside their known binding domains, leading to much speculation about why they are conserved (Izpisua-Belmonte et al. 1990, Licht et al. 1990). Although their specific purpose is not well-defined, polyalanine domains have been shown to be required for several transcription repressors in Drosophila, including KRUPPEL (Licht et al. 1994). More importantly, mutations in a polyalanine stretch found in the HOXD13 transcription factor have been shown to cause synpolydactyly, a human autosomal disorder involving fusion of a supernumerary number of fingers or toes. While the general population has a conserved alanine stretch of 15 residues (Ala15) in HOXD13, pedigree analysis shows that affected individuals contained an extra seven to 14 residues (Ala22 to Ala29). These expansionary residues are stable over many generations, and the severity of disease was shown to be proportional to the number of extra alanines (Akarsu et al. 1996). The polyalanine stretches in both HOXD13 and Lhx2 are due to degenerate trinucleotide repeats (GCN, where N is any nucleotide), increasing the similarity between the two regions. Thus, a change in alanine number (a decrease in the case of this reported Lhx2 base change) might potentially lead to Lhx2 malfunction as well. Similarly expanded trinucleotide repeats involving polyglutamine repeats are linked to neurodegenerative diseases, including Huntington’s Disease.
The idea that polyalanine strings play important roles in transcription factors is not supported in all cases. Comparisons across families have shown that as a rule, many transcription factors actually contain polymeric strings of alanine (GCN), glycine (GGN), and proline (CCN) (Monuki et al. 1993) with no known function. These amino acids are all encoded by G+C rich codons, which are believed to be susceptible to expansion and contraction through slipped mispairing during DNA replication (Jeffreys et al. 1988). Additional evidence shows that while known functional domains such as the LIM domain in transcriptional factors are evolutionarily conserved across species, polyalanine stretches are not (Monuki et al. 1993). Accordingly, amino acid comparison of the human Ala10 string in Lhx2 indicates that while it is mostly conserved in the mouse (Ala8) and rat (Ala5) (Figure 8), it is mostly absent in the chick and undetected in Drosophila. Nucleotide comparison, however, indicates that the replaced nucleotide in question and its corresponding alanine are strictly conserved across the mouse, rat, and chick.
The paradoxical evidence from synpolydactyly and unstable DNA studies suggest that polyalanine stretches may have significant function only in specific cases. While the human Lhx2 polyalanine string is not as strictly conserved in other species as the LIM domains or homeodomain, it is possible that it may serve a function unique to those species in which it is present. A lack of detailed information on Lhx2 function unfortunately precludes any consequences of this alteration beyond speculation. The unusual nature of the alanine-rich region and the strict conservation of the position of this apparent polymorphism warrant further studies such as a series of functional assays testing DNA-binding, LIM domain interaction, and transactivation without the presence of the polyalanine stretch. Because the string is close to the nuclear localization signal, Lhx2 entry into the nucleus could be tested as well.
In contrast to the replacement site alteration, the two silent site alterations both occur at the wobble position, causing no amino acid change. In Patient 989, the base change in Exon 4 (nine base pairs prior to the homeodomain) substitutes a pyramidine (CCG) in place of a purine (CCC) (Figure 7B), but both codons are synonymous and encode proline. In Patient 520 (a Joubert Syndrome case), the base change in Exon 5 (35 base pairs subsequent to the homeodomain) also introduces a synonymous codon (GAT to GAC), encoding aspartate (Figure 7C). The sequence change in theory could affect gene expression, with a change in secondary structure affecting transcription, processing, or translation. Additionally, a synonymous codon could recruit a different tRNA, affecting translation efficiency (Lewin 700). However, nearly all cases of third-position changes illustrate degeneracy, such that the position is irrelevant in determining protein structure or function. On amino acid comparison, only the Exon 4 amino acid is conserved across the mouse, rat, and chick, while the Exon 5 amino acid is conserved only in the mouse and rat. Nucleotide comparison across species, however, shows that in both cases only the mouse position is conserved. Due to the polymorphisms’ proximity outside Lhx2 hotspots and their relative lack of conservation across species, the potential impact of these alterations is probably neutral.
It is interesting to note that in Emx2 studies, similar sequence alterations were found in areas outside known functional domains (Brunelli et al. 1996). Four patients were found to have alterations in exon regions without known functional significance, three of which were in the same base position. The Lhx2 alteration noted here in Patient AS could not be observed in most of the other six schizencephaly patients sequenced because the region (Exon 3B) did not amplify consistently. This was due either to the primers or the region itself, which is G+C rich and therefore hard to amplify. The only other patient with sequence data in that expanse, Patient MR, did not show this base change. The interesting nature of this replacement site alteration merits further investigation, possibly through the use of different primers.
Possible Explanations and Expectations
The lack of explicit mutations discovered in the regions sequenced thus far provides important insights into the potential role of Lhx2 in schizencephaly and encourages several potential explanations based on the seven patients studied. One possibility is that if Lhx2 is involved in schizencephaly, the responsible mutation might lie in exon sequences not yet ascertained or intron sequences not amplified by this study’s primers. Mutations may also be present in the 5’ or 3’ UTR regions that logistically could not be covered in this study’s approach. Although these regions are not translated, the 5’ UTR contains enhancer and promoter regions necessary for transcription, while the 3’ UTR is involved in mRNA processing, as well as stability during translation. Neither region has been adequately characterized in Lhx2 as of this time, though poor conservation in these regions would make it difficult to prove bona fide mutations.
A more immediate consideration of this Lhx2 study is the genetically heterogeneous nature of schizencephaly. The malformation, similar to SOD and Joubert Syndrome, is defined morphologically and as such could have multiple genetic origins, of which Lhx2 represents only one. In conjunction with the unavoidably small sample size used for this thesis, it is possible that Lhx2 causes some cases of schizencephaly but not in these seven patients. Instead, the patients used for this study could have formed schizencephalies out of Emx2 mutations, alterations in other genes with roles in cortical proliferation (such as Otx1, Otx2, or Emx1), or even environmental defects developed in utero. Nongenetic published cases of schizencephaly include vascular obstructions in major cerebral arteries (Suchet 1994), leading to localized tissue death due to lack of oxygen (Landrieu and LaCroix 1994) or platelets (Kuijpers et al. 1994). Packard et al. speculate that schizencephaly is perhaps the end result of one or more genetic, toxic, vascular, infectious, or metabolic defects during critical stages of development (1997). The different possibilities would account for the wide spectrum of malformations frequently associated with schizencephaly.
A technical possibility that exists is the uncertain effect of PCR contamination discovered in the last weeks of experimentation. Negative water control lanes showed weak bands, sized correctly according to the primer pair used. Further investigation indicated that many of the different primers had been recently contaminated. Since it was impractical to run negative control lanes on all ten primer pairs for each PCR reaction throughout the period of experimentation, it is uncertain as to when each primer became contaminated and therefore what sequencing data might be affected. Contaminant DNA, presumably wild-type, could conceivably blemish sequence data by masking any real mutations. However, most weak signals of the contaminant negative controls are estimated to be no more than 5% of the brightness typically seen in successful PCR products. This indicates that any contribution to contaminant DNA when combined with targeted DNA template for amplification would be minimal. Additionally, nearly all chromatograms taken from sequence data indicate strong, even peaks with little distortion or background in relevant areas. These two factors suggest that any effect of contamination on mutation detection is likely to be negligible.
Because of the small sample size of the study, it is difficult to make definite conclusions about the lack of mutations. As more patients are recruited for this ongoing project, and assuming that other genes are responsible for schizencephaly patients not inflicted with Emx2 mutations, it would be interesting to speculate as to what sorts of Lhx2 mutations to expect if they indeed exist. One approach is to analyze Emx2, a gene with similar predicted functions and confirmed as a genetic cause for schizencephaly, as a precedent. As discussed earlier, both Lhx2-/- and Emx2-/- mouse knockout models are extremely comparable in their exhibition of reduced cerebral hemispheres. With regards to malformations often associated with schizencephaly, the Lhx2-/- mouse further resembles human patients because of its hypoplastic septum pellucidum and optic nerve accompanied by abnormal pituitary function (Monuki, personal communication). These additional malformations are also the definitive attributes of SOD.
However, although only heterozygous Emx2 mutations have been discovered in human schizencephaly patients, Emx2+/- (and Lhx2+/-) knockout mice actually display normal phenotypes. There are several likely reasons that explain the discrepancy between the heterozygous mouse (unaffected) and the heterozygous human (affected). One possibility suggests that the human Emx2 pathways or functionalities, though most likely similar to that of the mouse in CNS development, may be more sensitive to the loss of one allele. Thus, with only 50% of the Emx2 transcription factor protein functioning properly, schizencephaly patients with the heterozygous Emx2 mutation may lack sufficient Emx2 protein to handle its necessary roles in regulation (a haploinsufficiency or gene dosage effect). The mouse may in turn be less sensitive to this loss if it expresses another gene that plays a role redundant to that of Emx2. A second possibility suggests that instead of loss-of-function, the mutated protein may somehow interfere with the activity of the remaining wild-type protein (a dominant-negative or gain-of-function effect). This obstruction in fact often occurs with transcription factors that must dimerize with partner proteins in order to activate gene transcription (Semenza 1999). Thus, a mutated protein could conceivably have lost its DNA binding capability but still dimerize with an accessory protein, or vice versa. In either case, gene expression via the wild-type protein is slowed or impaired. A third possibility is that early in development during the proliferation phase, a second mutation or other event inactivates the wild-type allele in an individual neuron (a loss of heterozygosity effect). This effectively homozygous mutant proliferates exponentially as normal, creating a region of mutant cells. This mosaic possibility may explain why many unilateral cases of schizencephaly with genetic origin exist, where one hemisphere develops a cleft while the other grows normally. Loss of heterozygosity may also explain a well-documented case of familial schizencephaly (Granata et al. 1997), where two brothers with identical point mutations had different phenotypes.
Any of these explanations could account for the paradox between normal heterozygote Emx2 mice and inflicted heterozygote Emx2 humans. Because of the similarities Lhx2 has to Emx2 in terms of spatiotemporal expression and predicted involvement in regulation of proliferation, these three possibilities could also explain why any Lhx2 mutations in human schizencephaly patients might follow the example of Emx2 mutations. Thus, due to the similarities of the two knockout mice, and the fact that only heterozygous Emx2 mutations have been found in human schizencephaly patients (Brunelli et al. 1996), potential human Lhx2 mutations might be expected to be heterozygous as well.
In the case that an explicit mutation is found in the Lhx2 region, an array of different and independent approaches would need to be taken to first confirm and then characterize the change. Resequencing the region would first be necessary to rule out the possibilities of mutations introduced by Taq Polymerase, which has an error rate of 2 x 10-4 (Lewin 1997), or occasional artifacts introduced by PCR and automated sequencing. Comparing sequence data of the patient with that of the other cases might also render a pattern, especially in familial cases. Besides sequencing that patient, it would be important to sequence 100 or more known wild-type human samples to ensure that it is a nonconservative mutation rather than a neutral polymorphism. Assuming a heterozygous mutation, a third approach independent of sequencing would be to design a pair of primers, with one primer sequence theoretically able to anneal over the site of mutation. Individual alleles could be isolated and tested individually, such that only the mutant allele would produce a band. A final conditional approach requires that the putative mutation eliminates a known enzyme restriction site or creates a novel one. Assuming once again a heterozygous mutation, either change in the mutated region would generate a double band (one band from the normal allele, another from the mutant one).
Once the mutation is confirmed, characterization of the mutation to establish the mechanism by which the cleft might form could begin. To gain a general sense of the mutated region’s significance, the changed base(s) could be compared across species, with the knowledge that a mutation in a rigidly conserved region would likely be more significant than an alteration in a poorly conserved one. A series of assays could then be used to discover any changes in transcription activity or function. With this knowledge, targeted disruption of the mouse Lhx2 gene emulating the human mutation could be done to confirm the mutation in vivo. Clinical and morphologic features of the mouse could also be observed and compared with the original Lhx2 knockout mice (Porter et al. 1997).
It is important to keep in mind that the characterization of a gene mutation may not answer all the questions concerning a disease, especially one as complex as schizencephaly. Yet the approach established in this thesis can be considered a first attempt or approximation at studying the relationship between the human Lhx2 gene and relevant cortical malformations. The insights obtained through this investigation could influence issues as specific as genetic counseling (Robinson 1991, Hilburger et al. 1993) or as general as an improved understanding of cortex formation. For now, the genetically heterogeneous nature of schizencephaly and the small sample size of patients so far preclude any conclusive role or lack thereof of Lhx2 in schizencephaly, SOD, and Joubert Syndrome. It is reasonable to believe that future, more extensive patient screenings will contribute to describing Lhx2’s elusive role in human cortical development and malformation.
Akarsu, A.N., Stoilov, I., Yilmaz, E., Sayli, B.S., & Sarfarazi, M. (1994). Genomic structure of HOXD13 gene: a nine polyalanine duplication causes synpolydactyly in two unrelated families. Hum. Mol. Genet. 5, 945-952.
Anderson, S., Mione, M., Yun, K., & Rubenstein, J.L.R. (1999). Differential origins of neocortical projection and local circuit neurons: role of Dlx genes in neocortical interneuronogenesis. Cerebral Cortex. 9, 646-654.
Anton, E.S. et al. (1996). A regionallydistributed radial glial antigen: a candidate for signaling an end to neuronal migration. Soc. Neurosci. Abstr. 22, 1206.
Austin, C.P., Feldman, D.E., Ida, J.A., Jr., & Cepko, C.L. (1995). Vertebrate retinal ganglion cells are selected from competent progenitors by the action of Notch. Development. 121, 3637-3650.
Barkovich, A.J., & Kjos, B.O. (1992). Schizencephaly: correlation of clinical findings with MR characteristics. Ajnr American Journal of Neuroradiology. 13, 85-94.
Barkovich, A.J., Fram, E.K., & Norman, D. (1989). Septo-optic dysplasia: MR imaging. Radiology. 171(1), 189-192.
Boncinelli, E. et al. (1993). Emx and Otx homeobox genes in the developing mouse brain. Journal of Neurobiology. 24(10), 1356-1366.
Brodsky, M.C., Conte, F.A., Taylor, D., Hoyt, C.S., & Mrak, R.E. (1997). Sudden death in septo-optic dysplasia. Arch Ophthalmology. 115, 66-70.
Brunelli, S., Faiella, A., Capra, V., Nigro, V., Simeone, A., Cama, A., & Boncinelli, E. (1996). Germline mutations in the homeobox gene EMX2 in patients with severe schizencephaly. Nature Genetics. 12, 94-96.
Cameron, F.J., Khadilkar, V.V., & Stanhope, R. (1999). Pituitary dysfunction, morbidity and mortality with congenital midline malformation of the cerebrum.. Eur J Pediatr. 158, 97-102
Capra, V., De Marco, P., Moroni, A., Faiella, A., Brunelli, S., Tortori-Donati, P., Andreussi, I., Boncinelli, E., & Cama, A. (1996). Schizencephaly: Surgical features and new molecular genetic results. European Journal of Pediatric Surgery. 6 (Supplement 1), 27-29.
Chenn, A., & McConnell, S.K. (1995). Cleavage orientation and the asymmetric inheritance of Notch 1 immunoreactivity in mammalian neurogenesis. Cell. 82, 631-641.
Curtiss, J., & Heilig, J. DeLIMiting development. (1998). BioEssays. 20, 58-69.
Dattani, M.T., Martinez-Barbera, J.P., Thomas, P.Q., Brickman, J.M., Gupta, R., & Martensson, I., Toresson, H., Fox, M., Wales, J.K.H., Hindmarsh, P.C., Krauss, S., Beddington, R.S.P., & Robinson, I.C. (1998). Mutations in the homeobox gene HESX1/Hesx1 associated with septo-optic dysplasia in human and mouse. Nature Genetics. 19, 125-133.
Dattani, M.T., Martinez-Barbera, J.P., Thomas, P.Q., Brickman, J.M., Gupta, R., Wales, J.K., Hindmarsh, P.C., Beddington, R.S., & Robinson, I.C. (1999). HESX1: a novel gene implicated in a familial form of septo-optic dysplasia. Acta Paediatr Suppl. 88(433), 49-54.
Dawid, I.B., Breen, J.J., Toyama, R. (1998). LIM domains: multiple roles as adapters and functional modifiers in protein interactions. Trends Genet. 14(4), 156-62.
Donoghue, M.J., & Rakic, P. (1999). Molecular gradients and compartments in the embryonic primate cerebral cortex. Cerebral Cortex. 9, 586-600.
Eagleson, K.L, & Levitt, P. (1999). Complex signaling responsible for molecular regionalization of the cerebral cortex. Cerebral Cortex. 9, 562-568.
Eksioglu, Y.Z., Scheffer, I.E., Cardenas, P., Knoll, J., DiMario, F., Ramsby, G., Berg, M., Kamuro, K., Berkovic, S.F., Duyk, G.M., Parisi, J., Huttenlocher, P.R., & Walsh, C.A. (1996). Periventricular heterotopia: an X-linked dominant epilepsy locus causing aberrant cerebral cortical development. Neuron. 16, 77-87.
Faiella, A., Brunelli, S., Granata, T., D’Incerti, L., Cardini, R., Lenti, C., Battaglia, G., & Boncinelli, E. (1997). A number of schizencephaly patients including 2 brothers are heterozygous for germline mutations in the homoeobx gene EMX2. European Journal of Human Genetics. 5(4), 186-190.
Frotscher, M. (1997). Dual role of Cajal-Retzius cells and reelin in cortical development. Cell Tissue Res. 290, 315-322.
Gilbert, S. (1994). Developmental Biology, Fourth Edition. Sunderland, Massachusetts: Sinauer Associates, Inc.
Gleeson, J.G., & Walsh, C.A. (1997). New genetic insights into cerebral cortical development. In Normal and Abnormal Development of Cortex. Ed. Galaburda, A., & Christen, Y. New York: Springer-Verlag.
Granata, T., Farina, L., Faiella, A., Cardini, R., D’Incerti, L., Boncinelli, E., & Battaglia, G. 1997. Familial schizencephaly associated with EMX2 mutation. Neurology. 48(5), 1403-6.
Granata, T., Battaglia, G., D’Incerti, L., Franceshetti, S., Spreafico, R., Battino, D., Savoiardo, M., & Avanzini, G. (1996). Schizencephaly: neuroradiologic and epileptologic findings. Epilepsia. 37(12), 1185-1193.
Haydar, T.F., Kuan, C., Flavell, R.A., & Rakic, P. (1999). The roll of cell death in regulating the size and shape of the mammalian forebrain. Cerebral Cortex. 9, 621-626.
Hilburger, A.C., Willis, J.K., Bouldin, E., & Henderson-Tilton, A. 1993. Familial schizencephaly. Brain Development. 15(3), 234-236.
Ishibashi, M., Moriyoshi, K., Sasai, Y., Shiota, K., Nakanishi, S., & Kageyama, R. (1994). Persistent expression of helix-loop-helix factor HES-1 prevents mammalian neural differentiation in the central nervous system. EMBO. 13, 1799-1805.
Izpisua-Belmonte, J.C., Dolle, P., Renucci, A., Zappavigna, V., Falkenstein, H., & Duboule, D. (1990). Primary structure and embryonic expression pattern of the mouse Hox-4.3 homeobox gene. Development (Cambridge, UK). 110, 733-745.
Jeffreys, A.J., Royle, N.J., Wilson, V., & Wong, Z. (1988). Spontaneous mutation rates to new length alleles at tandem-repetitive hypervariable loci in human DNA. Nature (London). 332, 278-281.
Kuban, K.C., Teele, R.L., & Wallman, J. (1989). Septo-optic-dysplasia-schizencephaly. Radiologic and clinical features. Pediatric Radiology. 19(3), 145-150.
Kuijpers, R.W., van den Anker, J.N., Baerts, W., & von dem Borne, A.E. (1994). A case of severe neonatal thrombocytopenia with schizencephaly associated with anti-HPA-1b and anti-HPA-2a. Br J Haematol. 87(3), 576-579.
Landrieu, P., & LaCroix, C. (1994). Schizencephaly, consequence of a developmental vasculopathy? A clinicopathological report. Clinical Neuropathology. 13(4), 192-6.
Lewin, B. (1997). Genes VI. New York: Oxford University Press, 1997. In Molecular and Cellular Approaches to Neural Development. Ed. W. Cowan et al. New York: Oxford University Press.
Licht, J.D., Grossel, M.J., Figge, J., & Hansen, U.M. (1990). Drosophila Kruppel protein is a transcriptional repressor. Nature (London). 346, 76-79.
Licht, J.D., Hanna-Rose, W., Reddy, J.C., English, M.A., Ro, M., Grossel, M., Shaknovich, R., & Hansen, U. (1994). Mapping and mutagenesis of the amino-terminal transcription repression domain of the Drosophila Kruppel protein. Mol. Cell. Biol. 14, 4057-4066.
Lubinsky, M.S. (1997). Hypothesis: septo-optic dysplasia is a vascular disruption sequence. American Journal of Medical Genetics. 69(3), 235-236.
Mackarehtschian, K., Lau, C.K., Caras, I., & McConnell, S.K. (1999). Regional differences in the developing cerebral cortex revealed by Ephrin-A5 expression. Cerebral Cortex. 9, 601-610.
Mark, M., Rijli, F.M., & Chambon, P. Homeobox genes in embryogenesis and pathogenesis.
Matsumoto, K., Tanaka, T., Furuyama, T., Kashihara, Y., Ishii, N., Tohyama, M., Kitanaka, J., Takemura, M., Mori. T., & Wanaka, A. (1996). Differential expression of LIM-homeodomain genes in the embryonic murine brain. Neuroscience Letters. 211, 147-150.
Monuki, E., Kuhn, R., & Lemke, G. (1993). Cell-specific action and mutable structure of a transcription factor effector domain. Proceedings of the National Academy of Sciences USA. 90, 9978-9982.
Ni Scanaill, S., Crowley, P., Hogan, M., & Stuart B. (1999). Abnormal prenatal sonographic findings in the posterior cranial fossa: a case of Joubert’s syndrome. Ultrasound Obstet Gynecol. 13(1), 71-74.
Packard, A.M., Miller, V.S., & Delgado, M.R. (1997). Schizencephaly: Correlations of clinical and radiologic features. Neurology. 48, 1427-1434.
Parnavelas, J.G. (2000). The origin and migration of cortical neurones: new vistas. Trends Neurosci. 23, 126-131.
Porter, F.D., Drago, J., Xu, Y., Cheema, S.S., Wassif, C., Huang, S.P., Lee, E., Grinberg, A., Massalas, J.S., Bodine, D., Alt, F., & Westphal, H. (1997). Lhx2, a LIM homeobox gene, is required for eye, forebrain, and definitive erythrocyte development. Development. 124(15), 2935-44.
Porteus, M.H., Bulfone, A., Ciaranello, R.D., & Rubenstein, J.L.R. (DATE?). Isolation and characterization of a novel cDNA clone encoding a homeodomain that is developmentally regulated in the ventral forebrain. Neuron 7, 221-229.
Rakic, P. (1995). A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution. Trends in Neuroscience. 18, 383-388.
Rakic, P. (1972). Mode of cell migration to the superficial layers of fetal monkey neocortex. Journal of Comparative Neurology. 145, 61-84.
Reid, C.B., Liang, I., & Walsh, C.A. (1995). Systematic widespread clonal organization in cerebral cortex. Neuron. 15, 299-310.
Reid, C.B., & Walsh, C.A. (1996). Early development of the cerebral cortex. In Progress in Brain Research. Vol. 108. Ed. R.R. Mize and R.S. Erzurumlu. New York: Elsevier.
Robinson, R.O. (1991). Familial schizencephaly. Dev Med Child Neurol. 33(11), 1010-1012.
Rodriguez-Esteban, C., Schwabe, J.W.R., De La Pena, J., Rincon-Limas, D.E., Magallon, J., Botas, J., & Belmonte, J.C.I. (1998). Lhx2, a vertebrate homologue of apterous, regulates vertebrate limb outgrowth. Development. 125, 3925-3934.
Rubenstein, J.L.R., Anderson, S., Shi, L., Miyashita-Lin, E., Bulfone, A., & Hevner, R. (1999). Genetic control of cortical regionalization and connectivity. Cerebral Cortex. 9, 524-532.
Rubenstein, J.L.R., & Pulles, L. (1994). Pattern formation in the vertebrate neural plate. Current Topics in Developmental Biology. 29, 1-63.
Rubenstein, J.L.R., & Rakic, P. (1999). Genetic control of cortical development. Cerebral Cortex. 9, 521-523.
Rubenstein, J.L.R., & Shimamura, K. Regulation of patterning and differentiation in the embryonic vertebrate forebrain. ??!???!?
Saar, K., Al-Gazali, L., Sztriha, L., Rueschendorf, F., Nur-E-Kamal, M., Reis, A., & Bayoumi, R. (1999). Homozygosity mapping in families with Joubert Syndrome identifies a locus on chromosome 9q34.3 and evidence of genetic heterogeneity. American Journal of Human Genetics. 65, 1666-1671.
Semenza, G.L. (1999). Transcription Factors and Human Disease. New York: Oxford University Press.
Suchet, I.B. (1994). Schizencephaly: antenatal and postnatal assessment with colour-flow Doppler imaging. Can Assoc Radiol J. 45(3), 193-200.
Walsh, C.A. (1999). Genetic Malformations of the Human Cerebral Cortex. Neuron. 23: 19-29.
Walsh, C.A. (1995). Neuronal identity, neuronal migration, and epileptic disorders of the cerebral cortex. In Brain Development and Epilepsy. Ed. Scwartzkroin. New York: Oxford University Press.
Walsh, C.A., & Cepko, C.L. (1988). Clonally related neurons show several patterns of migration in cerebral cortex. Science. 255: 1342-1345.
Wanaka, A., Matsumoto, K., Kashihara, Y., Furuyama, T., Tanaka, T., Mori, T., Tanno, Y., Yokoya, S., Kitanaka, J., Takemura, M., & Tohyama, M. (1997). LIM-Homeodomain Gene Family in Neural Development. Developmental Neuroscience. 19, 97-100.
Ware, M.L., Tavazoie, S.F., Reid, C.B., & Walsh C.A. (1999). Coexistence of widespread clones and large radial clones in early embryonic ferret cortex. Cerebral Cortex. 9, 636-645.
Willnow, S., Kiess, W., Butenandt, O., Dorr, H.G., Enders, A., Strasser-Vogel, B., Egger, J., & Schwarz, H.P. (1996). Endocrine disorders in septo-optic dysplasia (De Morsier syndrome) – evaluation and follow up of 18 patients. European J Pediatr. 155, 179-184.
Wu, H., Heng, H.H.Q., Siderovski, D.P., Dong, W., Okuno, Y., Shi, S., Tsui, L., & Minden, M.D. (1996). Identification of a human LIM-Hox gene, hLH-2, aberrantly expressed in chronic myelogenous leukaemia and located on 9q33-34.1. Oncogene. 12, 1205-1212.
Xu, Y., Baldassare, M., Fisher, P., Rathbun, G., Oltz, E., Yancopoulos, G., Jessell, T., & Alt, F.W. (1993). LH-2: a LIM/homeodomain gene expressed in developing lymphocytes and neural cells. Proceedings of the National Academy of Sciences USA. 90, 227-231.
Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N., Kuratani, S., & Aizawa, S. (1997). Emx1 and Emx2 functions in development of dorsal telencephalon. Development. 124(1), 101-11.
Zhong, W., Feder, J., Jiang, M., Jan, L., & Jan, Y. (1996). Asymmetric localization of a mammalian Numb homolog during mouse cortical neurogenesis. Neuron. 17, 43-53.