Tuberous Sclerosis 2

A number sign (#) is used with this entry because tuberous sclerosis-2 (TSC2) is caused by heterozygous mutation in the TSC2 gene (191092) on chromosome 16p13. The TSC2 gene product is known as 'tuberin.'

Tuberous sclerosis complex (TSC) is an autosomal dominant multisystem disorder characterized by hamartomas in multiple organ systems, including the brain, skin, heart, kidneys, and lung. These changes can result in epilepsy, learning difficulties, behavioral problems, and renal failure, among other complications (reviews by Crino et al., 2006 and Curatolo et al., 2008).

For a general phenotypic description and a discussion of genetic heterogeneity of tuberous sclerosis, see tuberous sclerosis-1 (191100), caused by mutation in the TSC1 gene (605284) on chromosome 9q34.

Approximately 10 to 30% of cases of tuberous sclerosis are due to mutations in the TSC1 gene: the frequency of cases due to mutations in the TSC2 gene is consistently higher. TSC2 mutations are associated with more severe disease (Crino et al., 2006) (see GENOTYPE/PHENOTYPE CORRELATIONS section).

Kumar et al. (1995) reported a 2-year-old Caucasian female with tuberous sclerosis-2. At birth, she had a hypopigmented patch on her left ankle and multiple hypopigmented patches on her back and trunk. She later developed facial plaques on her forehead but no facial angiofibromas or ungual fibromata. Onset of generalized seizures occurred at 7 months of age. CT scan of the brain demonstrated cerebral cortical tubers and subependymal nodules. Renal ultrasound showed multiple cysts in both kidneys. The parents were clinically normal.

Vrtel et al. (1996) reported a father and his son with tuberous sclerosis-2. The family was ascertained through a discovery of fetal bradycardia and arrhythmia in the proband at 20 weeks' gestation. At 24 weeks' gestation, an intracardiac mass suspected of being a rhabdomyosarcoma was detected by fetal ultrasound and the diagnosis of tuberous sclerosis was suggested. A boy, weighing 2,500 g, was delivered at 39 weeks. Postnatal ECG showed intermittent second and third degree atrioventricular block. Echography of the brain, liver, and kidneys showed no abnormalities, and the studies of the retina were also normal. At 3 months of age a hypomelanotic macule, 25 x 15 mm, was noted on the buttock using Woods light. The 30-year-old father showed no abnormalities on study of the brain, heart, skin, and retina, and the most questionable changes in the kidneys. All tooth surfaces showed pit-shaped enamel defects, corresponding to the dental pits described in patients with tuberous sclerosis. In addition, 2 gingival fibromas were found.

Khare et al. (2001) reported a 4-generation family with mild physical features of tuberous sclerosis-2, but in which there was significant clustering of neuropsychiatric disorders including mood disorder, anxiety disorder, and autism among affected individuals. A second family also had mild physical features of tuberous sclerosis, but no neuropsychiatric assessment had been performed.

Martin et al. (2003) reported a pair of twin boys discordant for tuberous sclerosis-2 in whom marker studies supported a probability of monozygosity greater than 99.9%. The twins had similar CNS features, as both were severely mentally retarded with motor delay. Obvious differences were seen in the skin, heart, and kidneys. Whereas twin T had a shagreen patch of the skin and a heart rhabdomyoma, twin M had none. Twin M was diagnosed early (at the age of 3 years) with renal lesions, namely, angiomyolipomas and cystic alterations. At 6 years of age, twin T also started to have the same types of renal lesions as twin M. Martin et al. (2003) suggested that the Knudson hypothesis (Knudson, 1971) explained the difference, assuming that many of the features such as the skin, cardiac, and renal alterations present a 2-hit phenomenon, the second hit depending on a random somatic event. Genetic analysis identified a germline mutation in the TSC2 gene (191092.0012) in both boys.

Jansen et al. (2006) reported a large 5-generation French Canadian family with a mild form of tuberous sclerosis-2 due to a heterozygous mutation in the TSC2 gene (R905Q; 191092.0013). The proband was a 25-year-old man who started having seizures at age 8 years. Family studies showed that 25 individuals carried the mutation, but only 5 had definite TSC according to diagnostic criteria and 11 did not meet any diagnostic criteria. Examination of mutation carriers showed hypomelanotic macules in 92%, epilepsy in 60%, learning difficulties or mild cognitive impairment in 52%, and brain imaging abnormalities (white matter lesions, subependymal nodule, or subependymal giant cell astrocytoma) in 24%. None had cortical tubers. Renal lesions were found in 8% and retinal abnormalities in 4%. Five additional families with the same mutation were found, and all had a similarly mild phenotype, although cortical tubers were present in some.

In a retrospective review, McMaster et al. (2011) identified 10 cases of chordoma associated with tuberous sclerosis complex, although only 3 patients had documented mutations: 2 in the TSC1 gene and 1 in the TSC2 gene. The median age at onset in TSC-associated chordoma was 6.2 months (range 0 to 16 years), with only 1 patient diagnosed with chordoma after age 5. Chordomas were skull-based in 50% and sacral-based in 40%; the 16-year-old had a spinal-based tumor. The 5-year survival was 83%. Molecular and immunohistochemical studies of the chordomas from 2 patients with identified mutations in the TSC1 and TSC2 genes, respectively, demonstrated that 1 tumor had loss of heterozygosity (LOH) for the wildtype TSC1 allele, while the other tumor had LOH for the wildtype TSC2 allele, suggesting a pathogenetic role for the TSC1/TSC2 genes in these chordomas. Comparison with 65 cases of non-TSC-associated pediatric chordoma (215400) showed important clinical differences. The latter patients had onset between ages 0 and 18 years (median age at diagnosis was 12 years). Most (64.1%) were intracranial, 26.6% were spinal, and 9.4% were sacral. Chordomas were exclusively skull-based in the youngest age tertile, while sacral chordomas were confined to patients in the oldest tertile. Survival was poorer, at 68.2% at 5 years and 53.1% at 20 years. The findings suggested that TSC-associated chordoma has an unusually early onset and/or rapid growth, and that chordoma can be a rare pediatric manifestation of TSC.

The tuberous sclerosis complex consensus conference (Roach et al., 1998) proposed major and minor diagnostic criteria. Since no single feature is diagnostic, an evaluation that includes consideration of all clinical features is necessary to make a correct diagnosis. The clinical manifestations of TSC appear at distinct developmental points, which may further complicate the clinical diagnosis. Major criteria include retinal hamartomas, renal angiomyolipomas, facial angiofibromas, and cortical tubers, among other features. Minor criteria include dental pits, bone cysts, and cerebral white-matter radial migration lines, among other features.

Brackley et al. (1999) reported detection by ultrasound of cranial abnormalities at 14 weeks' gestation in a fetus subsequently confirmed as having tuberous sclerosis using DNA linkage analysis within the affected family. The presence of asymmetric ventricular enlargement persisted antenatally. MRI at 26 weeks indicated the possibility of poor gyral formation consistent with a neuronal migration disorder. Cardiac rhabdomyomata were not visualized on ultrasound scan until 30 weeks' gestation. Postnatal cranial ultrasound confirmed the significant neuropathology that was manifested by severe developmental delay and intractable fits in the child. The right fundus of the patient showed multiple peripheral pigmented areas and a persistent pupillary membrane also consistent with TSC.

Liang et al. (2014) generated a mosaic Tsc1-knockout mouse model in which mutant mice developed renal mesenchymal lesions that recapitulated perivascular epithelioid cell tumors (PEComas) found in patients with TSC. The authors found that YAP (YAP1; 606608) was upregulated by MTOR (601231) in mouse and human PEComas. Genetic or pharmacologic inhibition of Yap blunted abnormal proliferation and induced apoptosis of mouse Tsc1/Tsc2-deficient cells in culture and in mosaic Tsc1-knockout mice. Yap accumulated in cells lacking Tsc1/Tsc2 due to impaired degradation of Yap by autophagy in an Mtor-dependent manner. Liang et al. (2014) proposed that YAP is a potential therapeutic target for TSC and other disease with dysregulated MTOR activity.

Most patients with tuberous sclerosis-2 have de novo heterozygous mutations in the TSC2 gene. Patients with tuberous sclerosis-2 generally have more severe disease than patients with tuberous sclerosis-1, thus reducing the chance of these patients having a family (Curatolo et al., 2008).

Mosaicism

Verhoef et al. (1995) reported somatic mosaicism in the father of a 2-year-old boy with tuberous sclerosis 2: the father had subclinical signs of TSC and an apparently low proportion of cells with the TSC2 mutation.

Verhoef et al. (1999) identified 6 families with mosaicism in a series of 62 unrelated families with a mutation in either the TSC1 or the TSC2 gene. In 5 families, somatic mosaicism was present in the mildly affected parent of an index patient. In 1 family with clinically unaffected parents, gonadal mosaicism was detected after tuberous sclerosis was found in 3 children. The detection of mosaicism has obvious consequences for genetic counseling. Clinical investigation of the parents of patients with seemingly sporadic mutations is essential to determine their residual chance of gonadal and/or somatic mosaicism, unless a mosaic pattern is detected in the index patient, proving a de novo event. In the data set of Verhoef et al. (1999), the exclusion of signs of tuberous sclerosis in the parents of a patient with tuberous sclerosis reduced the chance of one of the parents to be a mosaic mutation carrier from 10% to 2%. In the 5 families with somatic mosaicism, the parent was given the diagnosis after the diagnosis was made in the child.

During TSC2 mutational analysis, Roberts et al. (2002) identified 10 single-nucleotide polymorphisms (SNPs) that occur within or close to exon boundaries at minor allele frequencies greater than 5%. The authors determined the haplotypes for 6 of these SNPs and the microsatellite marker kg8 in the 3-prime region of TSC2 in a set of 40 parent-child trios. The most common haplotypes accounted for 53%, 11%, 6%, and 5% of chromosomes. Thirty-eight TSC2 mutation-bearing haplotypes had a similar distribution, indicating that there was no haplotype that predisposed to mutation in this region in TSC2. Family analysis was possible in 12 sporadic cases, and indicated that the mother was the parent of origin in 7 cases (3 point mutations, 2 small deletions, 2 large deletions), while the father was the parent of origin in 5 cases (2 point mutations, 3 small deletions). Roberts et al. (2002) concluded that TSC2 mutations occur at substantial frequency on both the maternally and paternally derived TSC alleles, in contrast to many other genetic diseases. There was no major age effect for mutations of paternal origin. The observation was considered significant to genetic counseling of parents of a sporadic TSC case; each parent should be considered the potential parent of origin, and some alternative reproductive choices such as use of a sperm donor only or egg donor only are not guaranteed to prevent recurrence.

Using tuberous sclerosis families in which linkage to the TSC1 locus on chromosome 9 had been excluded, Kandt et al. (1992) demonstrated linkage with D16S283, the closest marker on the proximal side of the locus for polycystic kidney disease type 1 (173900), on chromosome 16p13. A lod score of 9.50 at theta = 0.02 was observed; 1 family independently presented a lod score of 4.44 at theta = 0.05. Kandt et al. (1992) estimated that about 40% of tuberous sclerosis families are linked to chromosome 9, and most of the rest, the majority, to chromosome 16. No clear evidence for a locus on other chromosomes was found. Confirmation of a tuberous sclerosis locus on chromosome 16 (TSC2) was provided by Pericak-Vance et al. (1992), Short et al. (1992) and Smith et al. (1992).

Povey et al. (1994) did linkage studies in 32 families of tuberous sclerosis, using genetic markers on chromosomes 9, 11, 12, and 16. Approximately half the families appeared to be linked to TSC1 on chromosome 9 between ASS1 (603470) and D9S298 and half to TSC2 on chromosome 16 close to D16S291.

Approximately 10 to 30% of tuberous sclerosis cases are due to TSC1 mutations, whereas the frequency of TSC2 mutations is consistently higher. TSC1 mutations account for 15 to 30% of familial cases and 10 to 15% of sporadic cases. The frequency of TSC2 mutations in sporadic cases ranges from 75 to 80%. About 15 to 20% of patients have no identifiable mutations, which may be due to mosaicism (Crino et al., 2006; Curatolo et al., 2008).

In a patient with tuberous sclerosis-2, Kumar et al. (1995) identified a de novo 1-bp deletion in the TSC2 gene (191092.0001).

In a father and son with tuberous sclerosis-2, Vrtel et al. (1996) identified a truncating mutation in the TSC2 gene (K12X; 191092.0003). The father had a milder phenotype, and was only diagnosed after his son was diagnosed. Flanking markers suggested that the mutated chromosome was of grandmaternal origin. The authors noted that it is possible that the mildly affected father was mosaic (although this was not detected), with the new mutation occurring by chance on the chromosome 16 he received from his mother. Vrtel et al. (1996) stated that the case illustrated the usefulness of mutation analysis in the diagnosis of families with an incomplete phenotype of tuberous sclerosis.

In 2 unrelated families with tuberous sclerosis-2, Khare et al. (2001) identified a missense mutation in the TSC2 gene (Q1503P; 191092.0011).

In 6 families with a mild form of tuberous sclerosis, Jansen et al. (2006) identified a heterozygous mutation in the TSC2 gene (R905Q; 191092.0013). The authors identified 2 additional mutations in the same codon, R905W (191092.0014) and R905G (191092.0015), in other families with a more severe phenotype, including cortical tubers, seizures, cognitive impairment, and severe skin lesions. Jansen et al. (2006) noted that the R905W and R905G substitutions resulted in the incorporation of nonpolar amino acids into the sequence, whereas the R905Q substitution introduced a polar amino acid with an amido functional group. Jansen et al. (2006) emphasized that patients with a mild form of tuberous sclerosis may not meet established diagnostic criteria, but can still have detrimental mutations in disease-associated genes.

Modifier of TSC2 Renal Angiomyolipomas

Because interferon-gamma (IFNG; 147570) is a useful mediator of tumor regression in animal models of kidney tumors and because there is a polymorphism within intron 1 of the IFNG gene for which 1 common allele (allele 2, with 12 CA repeats; 147570.0001) is associated with a higher expression of interferon-gamma in humans, Dabora et al. (2002) examined the relationship between the IFNG genotype and the severity of renal disease in patients with tuberous sclerosis who had TSC2 mutations. Patients were genotyped for the IFNG microsatellite polymorphism, allele 2, and its association with the development of kidney angiomyolipomas (which the authors called KAMLs) was examined. Both chi square analysis and the transmission/disequilibrium test (TDT) suggested an association between allele 2 and the absence of KAMLs in patients with known TSC2 mutations. Among the 127 patients who were more than 5 years old, KAMLs were present in 95 (75%) and absent in 32 (25%). In the group with KAMLs, the frequency of allele 2 was 56%; in the group without KAMLs, the frequency of allele 2 was significantly higher, at 78%. Family-based TDT analysis gave similar results. Subgroup analyses showed that both age and gender may influence the impact of this association. This study suggested that modifier genes play a role in the variable expression of tuberous sclerosis and also suggested a potential therapy for KAMLs in patients with tuberous sclerosis.

Lewis et al. (2004) used validated tools measuring intellectual function, depression, anxiety, and autistic and behavioral disorders to study the relationships between genotype, seizures, mental retardation, and behaviors in a cohort of 92 patients with mutations in the TSC1 or TSC2 genes. TSC2 but not TSC1 mutations were associated with autistic disorder (p = 0.001), infantile spasms (p = 0.001), and higher risk of low IQ (p = 0.0004) even after adjustment for a history of infantile spasms using logistical regression (OR, 3.50; 95% CI, 1.03-11.95). Previously unrecognized anxiety was frequently diagnosed in patients with mutations in either gene.

Au et al. (2007) performed mutation analyses on 325 individuals with definite tuberous sclerosis complex diagnostic status. The authors identified mutations in 72% (199 of 257) of de novo and 77% (53 of 68) of familial cases, with 17% of mutations in the TSC1 gene and 50% in the TSC2 gene. There were 4% unclassified variants and 29% with no mutation identified. Genotype/phenotype analyses of all observed tuberous sclerosis complex findings in probands were performed, including several clinical features not analyzed in 2 previous large studies. Au et al. (2007) showed that patients with TSC2 mutations have significantly more hypomelanotic macules and learning disability in contrast to those with TSC1 mutations, findings not noted in previous studies. The authors also observed results consistent with 2 similar studies suggesting that individuals with mutations in TSC2 have more severe symptoms. On performing metaanalyses of their data and the other 2 large studies in the literature (Dabora et al., 2001; Sancak et al., 2005), Au et al. (2007) found significant correlations for several features that individual studies did not have sufficient power to conclude. Male patients showed more frequent neurologic and eye symptoms, renal cysts, and ungual fibromas.

Jansen et al. (2008) compared the clinical features of 17 TS patients with mutations in the TSC1 gene and 31 patients with mutations in the TSC2 gene. Patients with TSC2 mutations tended to have an earlier onset of seizures, a higher incidence of infantile spasms, and lower cognition scores compared to those with TSC1 mutations. Patients with TSC2 mutations had more tubers and more tubers per brain proportion than those with TSC1 mutations, but the ranges overlapped. Patients with a mutation deleting or inactivating the GTPase-activating protein domain had more tubers than those with intact GTPase-activating domains. Despite some of these small differences, Jansen et al. (2008) concluded that there was considerable overlap between the groups and that prediction of the phenotype in patients with tuberous sclerosis should not be based on their particular TSC1 or TSC2 mutation.

Some patients with tuberous sclerosis develop pulmonary lymphangioleiomyomatosis (LAM; 606690), also known as pulmonary lymphangiomyomatosis, which has been reported in 34 to 39% of asymptomatic women and in some men with tuberous sclerosis. In a retrospective review of the chest CT scans of 45 female and 20 male patients with tuberous sclerosis, Muzykewicz et al. (2009) found cysts consistent with LAM in 22 (49%) women and 2 (10%) men. Among the women, changes consistent with LAM were observed in 6 (40%) of 15 with TSC1 mutations, 11 (48%) of 23 with TSC2 mutations, and 5 (71%) of 7 with no mutation identified. While the predominant size of cysts did not differ across these 3 groups, LAM women with TSC2 mutations had a significantly greater number of cysts than did women with TSC1 mutations. Some of the same mutations were identified in patients with LAM and in those without LAM. These findings suggested a higher rate of LAM in patients with TSC1 than previously recognized, as well as a fundamental difference in CT presentation between individuals with TSC1 and TSC2.

In a retrospective chart review of brain MRI scans of 173 patients with TSC, Chu-Shore et al. (2009) found that 46% of patients had at least 1 cyst-like cortical tuber. The tubers are called cyst-like because they presumably lack the inner endothelial lining seen in true cysts. Patients with TSC2 mutations were more likely to have a cyst-like tuber than patients with TSC1 mutation (p = 0.002) or patients with no mutation identified (p = 0.039). Patients with at least 1 cyst-like cortical tuber were more likely to have a history of infantile spasms (p = 0.00005; relative risk, 2.18), epilepsy (p = 0.0038; relative risk, 1.22), and refractory epilepsy (p = 0.0007; relative risk, 1.47) than patients without a cyst-like cortical tuber. Chu-Shore et al. (2009) concluded that cyst-like cortical tubers are strongly associated with TSC2 gene mutations and a more aggressive seizure phenotype in patients with tuberous sclerosis complex.