Tuberous sclerosis complex (TSC): An autosomal dominant disorder caused by a mutation in either the TSC1 or TSC2 genes and characterized by the development of hamartomatous growths in multiple organ systems.
Tuberous sclerosis complex (TSC) is an autosomal dominant disorder that occurs in up to 1 in 6,000 live births, without apparent ethnic clustering. TSC is characterized by the development of unusual tumor like growths, called hamartomas, in a variety of tissues and organs. Subependymal giant cell tumors in the brain occur in about one in ten individuals with TSC, with reported incidence ranging from 6.1% to 18.5%. On serial imaging, these tumors appear to arise from subependymal nodules which are believed to be present in 88–95% of individuals with TSC. Involvement of the brain is associated with some of the most problematic clinical manifestations of TSC, including intellectual handicap, epilepsy, and abnormal behavioral phenotypes, particularly autism and attention deficit disorder with hyperactivity. Other organs commonly and significantly involved in TSC include the skin, kidneys, and heart, where associated hamartomatous growths include facial angiofibromas, subungual fibromas, forehead plaques, shagreen patches, renal angiomyolipomas and cysts and cardiac rhabdomyomas (Fig. 1). Renal cell carcinoma (RCC) occurs at an earlier age and is more frequently multifocal in individuals with TSC than in the general population, although it is still uncommon (<5%).
Identification and characterization of the TSC genes
Linkage of TSC to chromosome 9q34 was reported in 1987 (locus termed TSC1); however, subsequent studies provided strong evidence for locus heterogeneity and led to the identification of a second locus at 16p13.3 (TSC2). Among families large enough to permit linkage analysis, approximately half show linkage to 9q34 and half to 16p13, and there is no evidence for a third locus. Initial definition of the 1.5 Mb TSC1 candidate region on chromosome 9q34 was achieved by identification of key meiotic recombination events in large TSC1 families. Two putative recombinants in unaffected individuals then narrowed this region to 900 kb. Large deletions and other rearrangements of the candidate locus were sought in patients with TSC, but no abnormalities were detected. Complete genomic sequencing of the region was initiated and GRAIL2 (Gene Recognition and Assembly Internet Link) and BLAST (Basic Local Alignment Search Tool) were employed to predict putative coding exons and genes. Systematic amplification and mutation screening of exons using TSC patient DNA samples revealed mobility shifts corresponding to small truncating mutations in the 62nd exon screened. This exon corresponded to a previously identified cDNA clone and various techniques were used to define the remainder of the open reading frame (ORF). Comparison of cDNA and genomic sequences revealed 23 exons, the first two of which were untranslated. The 8.6 kb full-length transcript was predicted to encode a novel 1,164 amino acid/130 kDa protein termed hamartin. During 1992–1993, linkage studies identified an ~1.5 Mb region of chromosome 16p as likely to contain the TSC2 gene. At the same time, a family with both tuberous sclerosis and autosomal dominant polycystic kidney disease was found to segregate a translocation between chromosomes 16p and 22q. The translocation breakpoint on chromosome 16 in this family was shown to disrupt the previously unidentified PKD1 gene and the TSC2 gene was predicted to lie telomeric to this breakpoint. The breakpoint was mapped to 150 kb telomeric to 16 AC2.5 (the most centromeric flanking marker then identified for TSC2). The telomeric limit of the candidate region was greatly reduced by the position of a second breakpoint in a previously reported patient who had a de novo truncation of 16p but no clinical or radiological evidence of TSC. The deletion in this patient effectively excluded ~1.1 Mb of the remaining 1.4 Mb TSC2 candidate region. A cosmid contig was constructed for the remaining 300 kb candidate region and probes generated from it were used to analyze a panel of TSC patients for rearrangements by pulsed field gel electrophoresis and Southern blotting. Five TSC patients were found to have genomic deletions of between 30 and 100 kb, which involved the same 120 kb interval. cDNA clones were isolated corresponding to four genes in the interval, and one was found to be disrupted by all five deletions, making it a strong candidate for TSC2. Four smaller intragenic deletions were then identified in TSC patients, including a de novo deletion that was associated with a truncated TSC2 transcript. These findings confirmed the identity of the TSC2 gene. The 5.5 kb TSC2 transcript was predicted to generate a novel 1,807 amino acid protein product of ~198 kDa termed tuberin.
TSC1 and TSC2 acting as tumor suppressor genes
Most individuals who have tuberous sclerosis carry a mutant tuberous sclerosis gene in each of their somatic cells. However, it is clear that a huge majority of these cells proliferate, differentiate, and function normally, while very occasionally, further localized events result in focal tumorigenesis. In 1971, Knudson proposed that inherited predisposition to tumors might reflect the germ-line mutation of “tumor suppressor genes” and that tumor development might be the result of somatic “second hit” mutations. Investigations of somatic mutations in a variety of tuberous sclerosis hamartomas supports classification of the TSC genes as tumor suppressor genes – several groups have reported evidence for large somatic deletions of the wild-type TSC1 or TSC2 allele, manifested as “loss of heterozygosity.” The observation of loss of heterozygosity implies clonality of hamartomas, and this has also been confirmed by demonstration of nonrandom X chromosome inactivation in hamartomas from female patients with tuberous sclerosis.
Lymphangioleiomyomatosis and pulmonary TSC
Lymphangioleiomyomatosis (LAM) is a disorder seen almost exclusively in females and is characterized by bronchiolar smooth muscle infiltration and cystic changes in the lung parenchyma. LAM patients often have angiomyolipoma of the kidneys and/or abdominal and hilar lymph nodes. Symptomatic LAM is estimated to occur in ~1 per million of the population without other evidence of TSC, but in up to 5% of females with TSC, implicating a role for the TSC genes in the etiology of LAM. It has recently been shown that somatic mutations of the TSC2 gene occur in the angiomyolipomas and pulmonary LAM cells of women with sporadic LAM, strongly supporting a direct role of TSC2 in the pathogenesis of this disease. Interestingly, a mutation in the TSC2 gene was identified in pulmonary and lymph node LAM cells from a patient with sporadic LAM prior to lung transplantation and the same mutation was subsequently found in the recurrent LAM. These data indicate that histologically benign LAM cells can migrate or metastasize in vivo to the transplanted lung. A model in which LAM cells migrate or metastasize to the lung challenges the boundary between benign and malignant diseases. There are other examples of histologically benign diseases in which cells appear to metastasize, including benign metastasizing leiomyoma, and disseminated peritoneal leiomyomatosis. If LAM cells containing TSC2 gene mutations have the potential to migrate in vivo, it may indicate that the TSC genes have key functional roles related to cellular metastasis.
Role of hamartin and tuberin in the mTOR pathway
Genetic screens for growth suppressors in Drosophila melanogaster showed that dTsc1 and dTsc2 play a pivotal role in the conserved insulin-signaling pathway to suppress cellular growth. Hamartin and tuberin function together to repress cell growth signaling through the mammalian target of rapamycin (mTOR). Tuberin possesses a guanine nucleotide triphosphate (GTP)-ase activating protein (GAP) domain which acts on the small G proteins Ras homologue enriched in brain (Rheb) and Rheb like 1 (RhebL1). Mutations in TSC2 that are associated with tuberous sclerosis often result in the loss of the GAP domain through the deletion of the C-terminus. Furthermore, a clustering of single amino acid mutations within the GAP domain of TSC2 has also been reported, implying that the GAP domain of tuberin is necessary for normal cell growth control. The normal function of hamartin and tuberin is to suppress the activity of Rheb and RhebL1: when in a complex with hamartin, tuberin reverts Rheb from an active GTP-bound state to an inactive guanine nucleotide diphosphate (GDP)- bound state through loss of the third phosphate on the guanine nucleotide. Active Rheb and RhebL1 have both been shown to associate with mTOR kinase and to specifically promote mTOR-directed phosphorylation of downstream signaling molecules involved in cell growth control that includes eukaryotic initiation factor 4E-Binding Protein 1 (4E-BP1) and ribosomal protein S6 kinase 1 (S6K1). mTOR also coordinates other critical cellular processes that if deregulated, can lead to cancer. These include cell cycle progression, autophagy (breakdown of cellular proteins), and nutrient uptake that feeds the growth of the cell. As a consequence, elevated mTOR signaling contributes to the pathology in numerous other human diseases associated with cancer.
mTOR and other hamartoma syndromes
Germline mutations that impair the normal tumor suppressor function of PTEN (phosphatase and tensin homolog), NF1 (neurofibromin 1), and LKB1 (also known as STK11 (serine/threonine kinase 11)) can also lead to inherited hamartoma syndromes. Loss of PTEN function causes Cowden’s disease, Lhermitte- Duclos disease and Bannayan-Zonana syndrome, while functional loss of either NF1 or LKB1 causes Neurofibromatosis “-type1” Peutz–Jeghers syndrome, respectively. Tuberin’s ability to function as a RhebGAP is regulated by multiple phosphorylation events that involve cell signaling pathways that are regulated by PTEN, LKB1, and NF1 (Fig. 2); therefore lesions associated with each of these hamartomas syndromes show activation of mTOR and enhanced cell growth signals. PKB and RSK kinase, which lie within the PI3K/PTEN and Raf/NF1 signaling pathways, phosphorylate tuberin on overlapping Serine and Threonine sites. PKB inhibits tuberin through the direct phosphorylation of Ser939 and Thr1462 that leads to elevated mTOR activity. Conversely, the LKB1 regulated energy-sensing pathway inhibits cell growth through AMPK-mediated phosphorylation of tuberin, which activates tuberin and represses mTOR. Interestingly, cancer cells deficient in PTEN have heightened sensitivity to the anti-proliferative and anticancer properties of rapamycin, which potently and specifically inhibits mTOR.
Rodent models of tuberous sclerosis
Genetically engineered mice that carry a single mutant Tsc1 allele (Tsc1+/- mice) develop renal cysts which develop into cystadenomas and RCCs. In one such model, the RCCs had a sarcomatoid morphology consisting of spindle cells with nuclear anaplasia arranged in whorled patterns and occasionally metastasized to the lungs (Fig. 3). Although the occurrence of RCC in humans with TSC is unusual, an association is recognized. The carcinomas are typically discovered at a young age and are thought to evolve from the lining of hyperplastic cysts. The proportion of sarcomatoid features in human TSC-associated renal carcinomas is far greater than in sporadic RCC. Tsc2+/- mice develop renal lesions by 6 months, which appear to grow progressively throughout the life of the mouse. Histologically, the lesions resemble cystadenomas consisting of a spectrum of lesions including pure cysts, cysts with papillary projections, and solid adenomas. Renal carcinoma, characterized by nuclear atypia, massive growth and metastatic disease developed in 5–10% of mice by 18 months, suggesting a very low rate of malignant progression for the cystadenomas (~1 in 1,000), indicating that additional genetic or epigenetic events are required for transformation. Liver hemangiomas, characterized by endothelial and smooth muscle proliferation with large vascular spaces, are seen in about half of Tsc2+/- mice by 18 months. Hemangiosarcomas also develop on the tail, paws, or mouth region in about 7% of Tsc2+/- mice by 12 months; these lesions do not metastasize but are malignant by cytologic criteria and exhibit bone invasion. The Eker rat harbors a naturally occurring mutation (insertion of a 6.3 kb intracisternal A particle) in one allele of Tsc2 and was first described as an autosomal dominant, hereditary model of predisposition to renal adenoma and carcinoma in 1954. Kidney lesions vary in morphology and include pure cysts, cysts with papillary projections and solid adenomas which can be seen as early as 4 months. A small minority of these tumors become malignant, with nuclear atypia, and expand to include the entire kidney and metastasize to the lungs, pancreas, and liver. Eker rats also develop pituitary adenomas, uterine leiomyomas and leiomyosarcomas, splenic hemangiomas, and, at a low frequency, brain hamartomas resembling human TSC subependymal nodules.
Fig. 1. Tuberous sclerosis complex is characterized by the development of hamartomas in a variety of tissues and organs. (a) A facial angiofibroma, (b) an abdominal MRI scan showing a renal angiomyolipoma (arrowed), (c) a cranial MRI scan showing a partly intraventricular subependymal giant cell astrocytoma (arrowed), and (d) a renal ultrasound scan showing a renal cell carcinoma (within a cyst) (arrowed)
Fig. 2. Converging signaling pathways that regulate hamartin and tuberin. The NF1, LKB1, and PTEN tumor suppressors negatively regulate mTOR via the PI3K/PKB, AMPK, and Ras/Erk/RSK signaling pathways. The hamartin/tuberin heterodimer act as a RhebGAP that converts the active Rheb˙GTP it to an inactive GDP-bound state. Rheb˙GTP activates mTOR, which promotes cellular growth
Fig. 3. Tsc1+/- mice described by Wilson et al. (2005) are predisposed to renal cysts that develop into RCCs that occasionally metastasize to the lungs. (a) Paired kidneys from a Tsc1+/- mouse with RCC ≥5 mm; (b) microscopic view of a sarcomatoid RCC with elongated sheets of spindle cells; (c) macroscopic and (d) microscopic RCC metastases in the lungs. Macroscopic bars are 1 cm and microscopic bars are 200 mm