Article Text

Download PDFPDF

Recent advances
Advances in the understanding of tuberous sclerosis
  1. Finbar J O'Callaghan,
  2. John P Osborne
  1. Royal United Hospital, Combe Park, Bath BA1 3NG, UK
  1. Prof. Osborne email: mpsjpo{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Medical textbooks still define tuberous sclerosis (TSC) as a triad of clinical features, mental retardation, epilepsy and “adenoma sebaceum”, first described by Vogt in 1908.1 ,2 Not surprisingly therefore many clinicians still do not realise that half of TSC sufferers will have normal intellect, a quarter will not have epilepsy, and almost any organ in the body can be affected. Recent advances in molecular genetics and imaging have begun to unravel the complexity of the disease, explaining the mechanisms behind the clinical features and providing insights into how these patients should be managed.

Genetic heterogeneity of TSC

Gunther and Penrose described the autosomal dominant pattern of inheritance of TSC in 1935.3 We now know that there are two genes (and probably only two) causing the TSC clinical phenotype. TSC1 is found on the long arm of chromosome 9, and its protein product is called hamartin.4 The TSC2 gene's protein product is named tuberin and it is situated on the short arm of chromosome 16, only 48 base pairs of DNA from the gene for adult onset polycystic kidney disease (PKD1).5 This genetic heterogeneity in TSC raises the question of whether the clinical syndrome produced by the two different genes is the same. In some conditions, neurofibromatosis for example, genetic heterogeneity helps explain clinically distinct forms of the disease. However, all the complications of TSC have been seen in both TSC1 and TSC2 except for the contiguous gene deletion syndrome (see below). There may, however, be subtle differences in the phenotype produced by the two genes or by specific mutations. There is already some evidence from case series that mutations in TSC2 tend to produce more severe disease than TSC, but this needs to be confirmed in larger less biased studies.6 We may then see no significant differences in the phenotypes because hamartin and tuberin are thought to function as part of, or at adjacent steps within, the same intracellular pathway.7

Function of the genes

We know that both tuberous sclerosis genes are tumour suppressor genes. Their function is to help regulate cell growth and differentiation. When altered, by mutation, control of cell growth is disturbed and tumours form throughout the body. The strongest evidence for the tumour suppressor hypothesis was provided in 1994 when it was reported that some of the hamartomas of tuberous sclerosis patients showed loss of heterozygosity, either in the chromosomal region 9q34 or in 16p13.8 ,9 Markers on 9q34 or 16p13 that were heterozygous in the patient were homozygous in the tumour. The loss of heterozygosity implies that an individual with tuberous sclerosis inherits or acquires through an early embryonic mutation, a deletion in one copy of the gene but only develops lesions when there is a somatic mutation in the other previously normal copy. This two hit mechanism was first proposed by Knudson to explain the pathogenesis of retinoblastoma.10

The tumour suppressor hypothesis is further strengthened both by recent research in the Eker rat (an animal model of tuberous sclerosis), and by studies of the possible intracellular function of tuberin. The Eker rat, which has a mutation in the rat homologue of the TSC2 gene, suffers from dominantly inherited renal cell carcinoma and subependymal and subcortical hamartomata. Reintroduction of a wild type TSC2 gene suppresses the development of renal tumours in this model.11 A search for sequence homologies at the protein level has revealed a region of similarity between tuberin and the GTPase activating protein GAP3.12 The GTPases are known to be involved in the regulation of cell proliferation and differentiation and it is thought that tuberin may have a role in mediating this activity.

Contiguous gene deletions in TSC

Renal cysts are common in TSC. They occur in a substantial minority of patients and have been associated with both TSC1 and TSC2 disease. A particularly severe phenotype has been observed in some patients (very early onset polycystic kidney disease). This is caused by a contiguous deletion that affects both TSC2 and PKD1 genes.13 It is sensible to perform renal ultrasonography on all newly diagnosed children with TSC to identify this rare problem early. An alternative is to monitor blood pressure and serum creatinine.

Mosaicism in TSC

Some of the patients with contiguous deletions of the TSC2 and PKD1 genes have a milder clinical course than others and this is, in some cases, caused by the fact that these patients are somatic mosaics for the contiguous gene deletion. Somatic mosaics have two or more cell lines, and only one cell line possesses the disease causing mutation. Somatic mosaicism has been described in both TSC1 and TSC2 patients with and without the contiguous TSC2-PKD1 syndrome.14 ,15The severity of disease in mosaics is variable but can be severe.

If the abnormal cell line is confined to the gonad (gonodal mosaicism) then a phenotypically normal parent has a high risk of producing affected children: this situation does occur in tuberous sclerosis and we quote a risk of 2–3% for two apparently unaffected parents (after full investigation) having further affected offspring after the birth of one child with TSC.16

Cranial imaging and TSC

The advent of magnetic resonance imaging (MRI) has allowed clinicians to image many of the lesions of TSC with greater clarity. MRI also has the advantage of not exposing the individual tox rays. Theoretically radiation could cause a “second hit” in the second copy of the gene. Cerebral tubers are shown on T2 weighted images but may be more easily visualised on a fluid attenuated inversion recovery sequence (FLAIR).17The number of tubers is related to degree of learning difficulty but variation in the relationship does not allow prediction for the individual.18 ,19 Lesions in the temporal lobe have been shown to increase the risk of autism in TSC.20 However, MRI does not show the subependymal lesions that are pathognomonic of the condition as clearly as computerised tomography (CT). Therefore it is important that cranial CT is used when an individual is being investigated for a possible diagnosis of TSC.

Screening in TSC

There are common hamartomas that can develop potentially life threatening consequences. In the kidney, renal angiomyolipomas are common in affected individuals but infrequently cause serious problems (haemorrhage into renal substance, collecting system, or retroperitoneally). They are usually easily seen as hyperechoic lesions on renal ultrasound, but CT is required to confirm the fat content of any hypoechoic lesions which might otherwise be a renal cell carcinoma (a very rare but serious complication of TSC: such a lesion must be shown to be HMB 45 negative on biopsy before nephrectomy is performed or an angiomyolipoma may inadvertently be removed).21 In the brain, the giant cell astrocytomas occur in approximately 5% of patients. They arise in the lateral ventricles close to the Foramen of Monroe (or rarely in the fourth ventricle), giving rise to signs and symptoms of raised intracranial pressure. Untreated this will cause hydrocephalus, blindness, and ultimately death. No feature of a giant cell astrocytoma on imaging has yet been shown to differentiate it reliably from a subependymal nodule—it is obstruction of CSF flow which causes symptoms and which must be treated.

Given the potential seriousness of these complications it is understandable that clinicians wish to identify and treat at a presymptomatic stage those lesions that will later cause serious problems. However, we do not yet know the natural history of these lesions and therefore one of the most important prerequisites for a screening programme is not in place. Some longitudinal studies on small series of selected patients have been performed and claim to have identified characteristics that distinguish those renal and cerebral lesions that will cause future problems.22 ,23 The most serious problem that is common to all these studies is that they have been conducted on small selected groups of patients over short time periods, often retrospectively or with interventions. We would currently advise against presymptomatic treatment of these lesions. However, we would strongly urge vigilance in investigating and treating symptoms. For the kidney, this means looking out for any evidence of frank haematuria, anaemia, or loin pain, and being aware that any sudden deterioration in clinical condition in a TSC patient could be caused by renal bleeding. Problematic giant cell astrocytomas declare themselves because they give rise to raised intracranial pressure. In a patient with normal intellect this is easily picked up as they usually present with classical signs and symptoms. Diagnosis in the patient with learning difficulty requires more care as they may not be able to communicate their symptoms to the carer or clinician. In such patients, unexplained deteriorations in behaviour or gait in addition to symptoms of raised intracranial pressure should be taken seriously and the need for cranial scanning considered.

Diagnosis of TSC

Mutational analysis is expensive and many mutations cannot yet be detected as it is usual for mutations to be unique to each family or individual. For prenatal diagnosis of offspring of affected individuals, it is now possible to request this. But for diagnosis, clinical skill is still often required. The old method of minor and major criteria is more logically replaced by examining and investigating for hamartomas. These are individually rare, so the finding of two or more “independent” hamartomas (table 1) strongly suggests the diagnosis. Individuals with a single hamartoma might have developed this through the random occurrence of two somatic mutations in a cell in that organ—but that is unlikely to occur twice in different organs in the same individual without an inherited or early embryonic mutation. In practice, it is usually easy to find two hamartomas in affected individuals. For this purpose, angiomyolipomas and pulmonary lymphangiomyomatosis have to be considered dependent hamartomas as these two lesions often occur together in the absence of other proof of TSC (see table 1). To exclude the disease in an individual, a full clinical examination including fundoscopy and cranial imaging must be undertaken: renal imaging probably confuses as often as it helps and is not recommended for this purpose. Echocardiography is only helpful in young children. Other features, such as hypomelanic patches, should be taken as an indication for further investigation and not as a reliable sign of TSC. Non-penetrance is no longer thought to occur.24

Table 1

Lesions which are individually rare


Finbar O'Callaghan is supported by a Wellcome Trust Training Fellowship in Clinical Epidemiology. We would also like to acknowledge the financial support of the Bath Unit for Research in Paediatrics, and the Tuberous Sclerosis Association of Great Britain.