Neurosurg Focus 20 (1):E4, 2006 Tuberous sclerosis complex: molecular pathogenesis and animal models LEANDRO R. PIEDIMONTE, M.D., IAN K. WAILES, M.D., AND HOWARD L. WEINER, M.D. Division of Pediatric Neurosurgery, Department of Neurosurgery, New York University School of Medicine, New York, New York Mutations in one of two genes, TSC1 and TSC2, result in a similar disease phenotype by disrupting the normal inter- action of their protein products, hamartin and tuberin, which form a functional signaling complex. Disruption of these genes in the brain results in abnormal cellular differentiation, migration, and proliferation, giving rise to the charac- teristic brain lesions of tuberous sclerosis complex (TSC) called cortical tubers. The most devastating complications of TSC affect the central nervous system and include epilepsy, mental retardation, autism, and glial tumors. Relevant animal models, including conventional and conditional knockout mice, are valuable tools for studying the normal func- tions of tuberin and hamartin and the way in which disruption of their expression gives rise to the variety of clinical features that characterize TSC. In the future, these animals will be invaluable preclinical models for the development of highly specific and efficacious treatments for children affected with TSC. KEY WORDS • tuberous sclerosis • pathogenesis • animal model Tuberous sclerosis complex is an autosomal-dominant zures, mental retardation, and autism.1,51 The number and tumor predisposition syndrome that affects approximately location of cortical tubers and the patient’s age at seizure 1 in 7500 individuals worldwide.19 The TSC is character- onset are highly correlated with neurological outcomes.1,30 ized by benign hamartomatous growths in multiple organs, Epilepsy occurs in approximately 80% of affected indi- including the kidney, skin, retina, lung, and brain. Rarely, viduals. Infants affected with TSC commonly present with malignant tumors, including renal cell carcinoma, can partial motor seizures or infantile spasms within the 1st develop. Although it has been nearly a decade since link- year of life.9 An increase in seizure frequency and severity age analyses first revealed the two genetic loci associated is common during early childhood. With maturity, TSC- with TSC, the mechanism by which disruption of these associated infantile spasms frequently evolve into other seizure types, including partial motor, complex partial, and genes produces associated abnormalities remains poorly 8 understood. The current efforts aimed at developing pre- secondarily generalized ones. Epileptogenic foci identified by electroencephalography frequently correlate with corti- clinical animal models have provided important insights 9,20 into the pathogenesis of TSC and will continue to provide cal tubers visualized on magnetic resonance imaging. In useful tools for studying potential clinical therapies. addition, the progression from infantile spasms to other seizure types might reflect the formation of tubers at dif- ferent times during cortical development. Thus, tubers in CLINICAL ASPECTS areas of the brain that become functionally mature earlier Neurological Manifestations and Brain Lesions can become epileptogenic before tubers located in brain regions that mature more slowly. Finally, due to the prema- Clinically, cortical tubers (pathognomonic lesions) cause turity of the patient, some as young as 20 weeks of gesta- extremely disabling CNS complications, which character- tion, physicians cannot accurately visualize each of these ize the TSC. Tubers are detected as high-intensity signals lesions, thus making the progression of epilepsy in patients that are often located in the junctions between the gray and with TSC extremely difficult to predict. white matter on T2-weighted magnetic resonance images. Tuberous sclerosis complex–associated seizures are The size and location of cortical tubers have been suggest- often refractory to conventional pharmacological therapies.␥ ed to correlate with the major CNS manifestations: sei- Improved seizure control has been achieved using the - aminobutyric acid agonist vigabatrin in some patients with 23 Abbreviations used in this paper: CNS = central nervous system; TSC. Pathological evidence suggests that cortical tubers GDP = guanosine diphosphate; GTP = guanosine triphosphate; represent areas of abnormal differentiation and migration. GTPase = guanosine triphosphatase; mTOR = mammalian target of Nevertheless, because surrounding areas of cortex are not rapamycin; Rheb = Ras homolog enriched in brain; TSC = tuberous disrupted and because tubers often disrupt normal cortical sclerosis complex. lamination, it has been hypothesized that these lesions Neurosurg. Focus / Volume 20 / January, 2006 1 Unauthenticated | Downloaded 10/01/21 02:46 AM UTC L. R. Piedimonte, I. K. Wailes, and H. L. Weiner result from the defective differentiation, migration, and high frequency of TSC2 mutation might reflect a higher proliferation of a subpopulation of precursor cells (Fig. 1). intrinsic mutation rate for the TSC2 gene. Tubers contain dysmorphic neurons and an increased Individuals who are born with a mutated copy of the number of astrocytes; thus they are characterized as “giant TSC1 or TSC2 gene in all cells of their bodies are predis- cells.” The cellular origin of these giant cells is unknown; posed to the development of tumors in numerous organs; however, they may arise from a neuroglial precursor cell, this suggests that these genes act as tumor suppressors. Tu- because giant cells often express both mature neuronal and mors develop in these individuals when one or more somat- astrocytic proteins. Many giant cells also express proteins ic cells undergo a “second hit” and the remaining wild-type found in immature CNS cells, including nestin and TSC gene is inactivated. This loss of TSC1 or TSC2 expres- vimentin.6,35 Similarly, the morphological features of giant sion results in abnormal cell growth and proliferation. cells are variable.40 Recent evidence supports the proposal Indeed, hamartomas from patients with both familial and that cells of cortical tubers can undergo active proliferation sporadic forms of the disease have displayed a loss of the and that, despite their early formation, these lesions can be normal copy of TSC1 or TSC2, known as a loss of hetero- 35 zygosity, which indicates that these genes act as classic tu- more dynamic than previously assumed. 24,41,59 In addition to cortical tubers, low-grade astrocytic neo- mor suppressors. In vitro studies have also supported plasms develop in individuals affected with TSC. Subepen- the role of these genes as tumor suppressors. When either tuberin or hamartin is overexpressed in cultured cells, the dymal nodules are benign proliferative lesions that line the 2,37 surface of the lateral ventricles. Unlike the cortical tubers, cells undergo growth arrest. Similarly, downregulation of TSC2 expression in culture fibroblasts by antisense inhi- these lesions are often asymptomatic, but can develop into 54 subependymal giant cell astrocytomas. Although not frank- bition has been shown to increase cell proliferation. Tuberin and hamartin are highly expressed in the CNS of ly malignant, subependymal giant cell astrocytomas can 21,39 cause ventricular obstruction and hydrocephalus, which re- humans and mice in both neurons and glial cells. Both quire neurosurgical intervention.52 In addition to glial cells, of these proteins appear to be required for CNS develop- ment, and tuberin may be essential for neuronal differen- subependymal nodules and subependymal giant cell astro- 53 cytomas can also contain giant cells. These lesions, like cor- tiation. The effect these proteins have in the CNS is prob- tical tubers, are thought to develop early in life and have ably the same as they have where they act as tumor 38 suppressors. This is hypothesized because tuberin expres- been identified as early as 27 weeks of gestation. Like cor- 61 tical tubers, these lesions are thought to arise from abnormal sion is reduced or absent in 30% of sporadic astrocytomas. development of precursor cells during brain formation. There is conflicting evidence whether somatic mutations of the wild-type TSC1 or TSC2 allele in a precursor cell are required for cortical tuber formation in patients who have Genetics and Molecular Pathogenesis 24,41,49 only one normal copy of one of the two TSC genes. Linkage studies have identified two distinct loci that 48 Because these lesions are composed of many different cell undergo mutational inactivation in individuals with TSC. types, it can be difficult to identify the subset of cells that The phenotypic expression of TSC is demonstrated when has undergone a second mutational event. In support of the mutations occur in one of the two genes, TSC1 or TSC2. hypothesis that two genetic “hits” are required for TSC- The TSC1 gene is located on 9q43; this gene contains 23 59 associated tuber formation, recent evidence suggests that, exons and encodes the 130-kD hamartin protein. Hamar- in mice, neuroepithelial cell progenitors that lack TSC2 tin has little sequence homology to other known proteins. expression have many features of the giant cells found in The TSC2 gene on 16p13 encodes the 180- to 200-kD cortical tubers.42 tuberin protein. Tuberin contains a small region with The two abnormal cellular phenotypes in TSC involve sequence similarity to proteins with GTPase–activating an increase in both cellular proliferation and cell size. In-