NINDS Tuberous Sclerosis Complex Conference Summary

New Perspectives In Tuberous Sclerosis Complex

Conference Summary
September 19-22, 2002
Chantilly, Virginia

• I. Introduction
• II. Highlights Of Recent Research
• Epilepsy
• Cognition and Behavior
• Renal Development
• Pulmonary Function
• Dermatologic Involvement
• Animal Models of TSC
• Molecular Biology and Biochemical Pathways
• III. Recommendations for Future Research
• Tissue bank
• Molecular and cellular mechanisms
• Clinical trials/databases
• Attracting new investigators
• 
  

I. Introduction

Tuberous sclerosis complex (TSC) is a multi-system disease characterized by abnormal growths (hamartomas). These growths are distributed at multiple sites throughout the body. The severity and symptomatic expression of TSC is highly variable, making the disease a challenging one to diagnose. While severe manifestations may be seen in individuals diagnosed in childhood, mild forms of the disease may be observed in women and men diagnosed in adulthood. TSC is a genetic disorder but more than half the individuals with TSC have a spontaneous rather than an inherited mutation. The major organ systems involved include brain, kidneys, lungs, heart, eyes, and skin. Central nervous system involvement is the most common feature. The majority of patients first come to the clinic for treatment of seizures and TSC is one of the most common neurogenetic syndromes associated with epilepsy and autism.

Tuberous sclerosis shows an autosomal dominant pattern of inheritance. Two genes have been identified within the last 11 years (TSC1 in 1997 and TSC2 in 1992). The TSC1 gene is located on chromosome 9, and encodes a 130 kD protein called hamartin. The TSC2 gene, located on chromosome 16, is contiguous with the gene for polycystic kidney disease. The TSC-2 gene product, a protein called tuberin, is a GTPase activating leucine zipper protein. Hamartin and tuberin bind to each other, and are likely to function as tumor suppressors.

Research in tuberous sclerosis is as broad as the disease itself, addressing the minutiae of molecular genetics to the behavioral studies of cognition and from clinical observations in the operating room to drug discovery and development of new therapies. This meeting was unusual in that it bridged many disciplines and brought together a diverse group of experts. Critical issues in all areas of tuberous sclerosis were presented and discussed. Participants were asked to reflect on future directions and to develop a template for further research initiatives.

Top

II. Highlights Of Recent Research

Epilepsy

Epilepsy is the most common neurological symptom among individuals with TSC, and the management of epilepsy accounts for a large percentage of the care that they require. Approximately 90% of patients have seizures and an astounding 85% of seizures begin in the first year of life. Infantile spasms occur at a very high rate, and are associated with especially poor prognosis.

Hamartomas in the brain resemble tubers, hence the name tuberous sclerosis. Cortical lamination is disrupted within cortical tubers. Abnormal cells found in tubers include dysplastic neurons, large astrocytes, and giant cells displaying markers of neuronal, glial, and immature phenotypes. The number of tubers and their location varies, although there is a preferential accumulation within frontal and parietal cortex. The abnormal placement of neurons and the disorganized lamination that are typical of tubers suggest that the underlying developmental defect may be in neuronal migration. Neuronal maturation may also be incomplete, as about 70% of neurons in tubers have either fewer or thinner spines than normal.

Epilepsy in the majority of individuals with TSC is intractable. Currently available antiepileptic drugs are usually ineffective, particularly in the case of infantile spasms. Surgery thus becomes an important option. Michael Duchowny (Miami Children's Hospital) discussed the fact that not all tubers are the same, and some are more epileptogenic than others. Using depth electrodes, Duchowny and collaborators have shown that the core of a tuber can be epileptogenic. Propagation of electrical activity spreads to surrounding tissue and, due to aberrant connections, to other parts of the brain. If a particular tuber is shown to be an epileptic focus, surgical removal of that tuber can provide relief from seizures even while other tubers remain. The reasons underlying epileptogenesis in tubers in particular, and in epilepsy in general, are unclear. Intrinsic hyperexcitability of the neurons is likely to be a contributing factor. PET scans have shown that uptake of alphamethyltryptophan is higher in epileptogenic tubers than in nonepileptogenic tubers. This suggests that alterations in serotonin receptor activity may be involved. Seizure intractability may be partially explained by expression of multidrug resistance genes, which have been documented in TSC cells. When expressed in brain capillary endothelial cells, multidrug resistance gene products may lower the concentration of antiepileptic drugs so that they are below therapeutic levels in the brain even if systemic drug levels are high.

Top

Cognition and Behavior

Behavioral and cognitive dysfunction affects more than half of all individuals with TSC. Significant developmental delays, with or without autism spectrum disorders, are seen in about half of all children with TSC. In fact, autism was first diagnosed in a child with TSC.

Given the high prevalence of autism spectrum disorders in individuals with TSC (up to 60% in children with TSC compared to 0.15% in the general population worldwide), tuberous sclerosis may be a useful model system with which to investigate autistic spectrum disorders. One theory, proposed by Patrick Bolton (Cambridge UK) is that epileptic seizures in the temporal lobes, perhaps related to the presence of tubers, disrupt the development of key social representations during an early critical period. A simple, very early sign of these disruptions is the inability to tell where someone else's eyes are directed. These early changes are thought to lead to later impairments in the ability to infer the intent and emotions of other people.

There is significant risk for specific learning disabilities, even in the 50% of TSC patients whose development is within normal limits. Neuropsychological testing of patients 3 months to 45 years old at the TSC Comprehensive Clinic at the MGH (Boston MA) has shown that consistent deficits in attention, retrieval, and integration can be identified in individuals who otherwise tested in the overall normal cognitive range. This is consistent with the distribution of tubers in TSC, which are most commonly found in frontal, parietal, and subcortical regions of the brain. The development of children with TSC should be followed closely. Specifically, Dr. Prather used her recent findings to recommend that all children should receive a cognitive assessment at ages 5-6 (before entering school) and at age 9 (at the end of the primary elementary grades). As with all children with attentional difficulties, children with TSC are likely to benefit from more structure and repetition in the early elementary grades to reinforce basic skills, followed with help on organization and planning in the upper primary grades and accommodations and extra support through high school.

Top

Renal Development

TSC is also associated with lesions in the kidneys and lungs. Hamartomas in these organs consist of smooth muscle cells, blood vessels, and, in kidneys, fat tissue. Renal disease is a significant source of mortality in TSC, with the most life-threatening risk due to the rupture of vessels within the hamartomas.

Work in the Herzlinger laboratory (NYU, New York NY) has shown that renal angiolipomas are not the result of dedifferentiation of mature renal epithelia to an embryonic pluripotent cell. This is supported by the fact that hamartin and tuberin, the protein products of the TSC1 and TSC2 genes, are thought to be involved more in regulating cell growth rather than cell differentiation. Instead, it may be stromal cells that are the culprit cell type responsible for generating renal angiolipomas. Stromal cells differentiate into fibroblasts and vascular smooth muscle. They express TSC1 and are extremely sensitive to deregulated growth upon loss of TSC function. In addition, they secrete paracrine factors (notably FGF-7) that serve to pattern the size and number of renal tubules. Overgrowth of stromal cells when TSC1 is lost could lead directly to angiolipomas due to excess fibroblasts and smooth muscle cells, as well as oversecretion of paracrine factors that could result in the formation of renal cysts.

Top

Pulmonary Function

Pulmonary involvement in TSC consists primarily of cystic lesions and nodules in the lungs. These occur independently of each other; cysts are almost exclusively found in women but nodules occur in both men and women. Abnormal growths and cystic destruction of lung tissue is characteristic of lymphangioleiomyomatosis (LAM ), a rare lung disease that occurs as an isolated disorder (sporadic LAM) and in women with TSC.

In a large scale screen of women with TSC, Joel Moss (NIH, Bethesda MD) found that lung function was preserved even if cystic lesions were present, suggesting that a large population of individuals with TSC are asymptomatic or have very mild lung disease. The prevalence of LAM in women with TSC is estimated to be about 34%, although severe respiratory disease occurs in less than 1% of women with TSC. In screening women with LAM for TSC, brain scans revealed meningiomas in 10 women out of 250 who were screened; 3 of these women met the diagnostic criteria for TSC. Eventual respiratory failure due to destruction of lung tissue may require lung transplantation, which remains the most viable option for end-stage disease. As predictors of time to transplant or death, histological assessment of disease (LAM histology score), DLCO (in the case of mild disease only), the cardiopulmonary exercise test, and high resolution CT scans are all useful.

LAM nodules in the lungs and AML (angiomyolipomas) in the kidney both contain abnormal smooth muscle cells that express smooth muscle and melanoma antigens. While similar to each other, cells in LAM and AML nodules are not like other smooth muscle or melanoma cells. Nodules have a characteristic composition, although quite a bit of variability is observed. The center of the LAM nodule contains proliferating cells, these are surrounded by epithelia-like cells and, these in turn may be surrounded by cells that are hyperplastic pneumocytes.

LAM nodules, like other growths in TSC, are not malignant.. However, they do possess some metastatic potential. Work from the Henske laboratory (Fox Chase Cancer Center, Philadelphia PA) demonstrates that recurrence of LAM in a healthy transplanted lung is due to metastasis of cells from the host (n=2).

Top

Dermatologic Involvement

Dermatologic hamartomas in individuals with TSC are harmless but nonetheless disfiguring, particularly if they are on the face. Skin tumors are frequently numerous and include facial angiofibromas, collagenomas (Shagreen patches), and periungal fibromas. They are extremely accessible to study, in contrast to the parallel internal tumors. In touch preparations of TSC skin lesions, Tom Darling (Uniformed Services University, Bethesda MD) reported that allelic deletion of the TSC2 gene was observed in nearly every sample. The tumors were a heterogenous mixture of cells with 18% to 50% showing allelic deletion of TSC2. These results are consistent with the two-hit theory, where individuals with a germ line defect in one allele lose the second allele in a tumor. Deletion of the TSC1 gene was not observed in any sample. Cells from TSC skin lesions can be cultured and have been propagated for up to 5 passages. These cells express higher levels of cytokines that may be one way that the neoplastic cells influence to growth of surrounding normal cells.

Most skin lesions on the face are angiofibromas that are either flat and red, or raised and unpigmented or red. The presence of facial angiofibromas can lead to bleeding, chronic irritation, and cosmetic disfigurement. Additional functional problems can arise including tooth maldevelopment, airway obstruction, or visual occlusion. Mark Mausner (Mausner Plastic Surgery Center, Rockville MD) described the results of treatment with a variety of specialty lasers that greatly improved appearance and alleviated functional problems. Many of the treated facial angiofibromas did not recur, and when they did, regrowth was extremely slow. Thus lesions should be treated early since it was noted in all individuals with TSC that facial angiofibromas increase rapidly in size and number during puberty.

Top

Animal Models of TSC

The Eker rat, first described in the 1960's by Eker and Mossige, spontaneously develops tumors in the kidney, uterus, and spleen. It is now known that these tumors are due to loss of function of the tsc2 gene. The Eker rat, and cell lines established from its tumors, are useful tools in understanding tuberin function and the renal manifestations of tuberous sclerosis. Interestingly, the Eker rat (bred on a Long-Evans background) rarely develops cortical tubers. To develop an Eker rat model for TSC seizures, various "second-hit" approaches have been tried. So far, work in Scott Baraban's laboratory (UC San Francisco) exposing Eker rats prenatally to hydroquinone has not resulted in tuber formation. A different "second-hit" was used by Philip Schwartzkroin (UC Davis) who gave young postnatal Eker pups a dose of radiation. The brains of these rats had a significantly lower seizure threshold and contained both dysmorphic neurons and large cells that were GFAP positive.

Several lines of transgenic mice have been developed. Jack Arbiser's laboratory (Emory University, Atlanta GA) generated a transgenic mouse line that expresses a dominant negative allele of tuberin behind a constitutive cytomegalovirus promoter. The mice are viable and fertile. They develop neoplastic nodules in the kidney as well as skin lesions that are reminiscent of collagenomas (Shagreen patch).

A tsc1 conditional knockout, under the control of the glial specific GFAP promoter, results in mice with abnormally high numbers of astrocytes (up to 6 times more than normal). David Gutmann (Washington University, St Louis MO) described how the astrocytes grow abnormally in vitro, having lost contact inhibition of growth. Interestingly, disorganized neuronal layers are evident in the hippocampus. Most interesting, the tsc1 conditional knockout mice develop seizures after about 1 month of age, and start to die after 3-4 months. The seizures occur with a frequency of 0.6/hour. Kevin Ess (Washington University, St Louis MO) described the seizures in more detail, noting that most seizures are generalized although focal onset in some seizures could be detected by hippocampal depth electrodes. Interrictal EEG and behavior were severely abnormal. Electrophysiology performed in the hippocampus of mutant mice showed that neither paired pulse facilitation or presynaptic facilitation were detectable, indicating that short term plasticity in these mice is impaired.

David Kwiatkowski's laboratory (Brigham & Women's Hospital, Boston MA) has generated 2 other conditional knockouts of the tsc1 gene controlled either by the synapsin promoter (expressed in all neurons) or by the the nestin promoter (expressed in neuroepithelial cells). The tsc1/synapsin mice die by 6 weeks of age. Their brains have mildly enlarged cells, no astrogliosis is present, and some disorganization in the hippocampus is observed. Seizures can be induced by physical manipulation (tail spinning). The tsc1/nestin mice develop spontaneous seizures and die by 4 months of age. Their brains show enlarged neuron-like cells and disorganized laminae in the cortex and hippocampus. Cells have also been cultured from these mouse embryos, and are proving to be useful in dissecting tsc1 and tsc2 intracellular signaling pathways.

Top

Molecular Biology and Biochemical Pathways

Tuberin (TSC2)is a GTPase activating protein, with homology to Rap1GAP. The kinase AKT, which is stimulated by growth factors, phosphorylates tuberin on 2 sites. Hamartin (TSC1)is not phsophorylated by AKT. Hamartin is thought to interact with the cytoskeletal proteins ezrin-radixin-moesin, binds to the light chain of neurofilament, and activates the small GTPase Rho. Tuberin and hamartin form a complex, and the presence of both proteins seems to be required in most of the pathways described below.

Hamartin and tuberin have been shown to be part of the growth factor (insulin) signaling pathway. This pathway is well conserved from flies to humans, and many of its molecular details are well understood. A pivotal component of the cascade, mTOR (mammalian Target Of Rapamycin), was identified because the immunosuppressant drug rapamycin inhibits its activity. The placement of tuberin and hamartin in this pathway upstream of mTOR is of intense interest due to the therapeutic possibilities. While many of the details concerning the role of hamartin and tuberin in this pathway have been elucidated, particularly for the Drosophila homologs of tsc1 and tsc2, many questions remain, particularly concerning exactly how directly hamartin and tuberin interact with mTOR.

Aside from the growth factor/mTOR pathway, hamartin and tuberin affect cell cycle control. Cyclin D levels increase when tuberin levels decrease. Cyclin D is regulated by Wnt/b-catenin, and Baldwin Mak (University of Washington, Seattle WA) presented evidence that hamartin/tuberin act at the level of the b-catenin degradation complex. A third pathway was described by Cheryl Walker (MD Anderson Cancer Center, Smithville TX), whose laboratory has demonstrated that tuberin leads to stabilization of the HIF2a subunit of HIF (hypoxia-induced factor). This results in elevated HIF activity, which increases the expression of VEGF (vascular endothelial growth factor), and ultimately leads to tumor angiogenesis.

Top

III. Recommendations for Future Research

Recommendations for future TSC research were made with the broad acknowledgment that collaborative efforts were going to be needed and that resources and reagents should be shared and made widely available to all interested investigators. Several common areas for future work were cited by investigators in each of the systems affected by TSC.

Tissue bank

Conference participants expressed overwhelming and unanimous support for the establishment of a tissue bank of TSC tissue from brain, lungs, skin, and kidneys. Both primary and immortalized cells would be useful. Skin samples in particular are readily obtainable but so far have not been widely available to researchers. Examples of the many ways that the tissue bank could be used are to characterize the different cell types in a tumor, to examine molecular pathways involving hamartin and tuberin, or to identify growth factors that stimulate angiogenesis and fibromagenesis. A collaborative group effort to perform electrophysiology on fresh tuber samples from human brains was also proposed.

Logistical considerations for the establishment and maintenance of a tissue bank were discussed. Coordinating surgery on patients with sending tissue samples to researchers seemed complicated, particularly if the patients and researchers are not at the same institution. There are also regulations about the sharing of tissue and data, especially across international borders. The tissue bank should probably be established with the idea that eventually the federal government will assume responsibility for running it. Finally, donors like to know what happened with the tissue that was donated, so a mechanism needs to be developed where the researcher can easily provide results.

Top

Molecular and cellular mechanisms

Identification of the TSC1 and TSC2 genes has spurred great advances in understanding the disease mechanisms of tuberous sclerosis, particularly in the lung and kidney. More research is unquestionably needed on the multiple putative pathways that have been identified thus far. How these signaling pathways affect cell adhesion, migration, proliferation, growth and differentiation also need to be explored.

How hamartin and tuberin affect brain development, and whether the growth factor/mTOR pathway is involved in brain pathology, are questions that need to be answered. The origin of tubers needs to be understood. Understanding the electrophysiology of tubers is also a priority since this may lead to insights into epileptogenesis.

In addition to the basic science implications, understanding biological mechanisms is critical to developing new therapies. Drug discovery requires a molecular target, an enzyme or receptor whose function can be modified. The more clearly a disease process is understood at a molecular and cellular level, the more likely that key proteins are targeted for drug intervention.

Top

Clinical trials/databases

Several ideas for clinical trials and clinical databases were proposed.

Clinical Trial to Test Rapamycin in TSC Patients
This proposal was for a Phase 1/2 pilot study to investigate the feasibility and efficacy of treatment with rapamycin in individuals with TSC. Individuals with TSC who have renal angiomyolipomas renal would be enrolled, inclusion criteria would include angiomyolipomas (AML) 2-5 cm in size and good performance status. LAM patients with AML 2-5 cm in size would also be eligible. The trial would aim to start with about 20 individuals in the US and 20 people in the UK. The study participants would receive a clinical exam and MRI upon entry into the study. Rapamycin would be administered for 8 weeks, starting with 1 mg/day (renal transplant patients currently start with 2-5 mg/day). As the study progresses the dose could be increased if the safety profile is favorable. The primary endpoint would be AML size, and secondary endpoints would include the status of other organs that are involved, toxicity, and other biomarkers. A call was made to establish a working group to draw up a final protocol, secure funding, and develop a database in preparation for conducting the trial.

Longitudinal Study of Cognitive Outcome in Individuals with TSC
This proposal was for a multicenter, multidisciplinary, international, prospective, longitudinal study to identify the principal determinants of cognitive and behavioral outcome in TSC. Serial neurocognitive assessments would be made, along with neuroradiological and EEG assessments and behavioral assessments. Treatments and medications would be recorded. Centers in both the US and the UK would participate using a common core set of instruments but with different and complementary emphases on attention, autism spectrum disorder, and language. A call was also made to establish a working group to determine details of the study with respect to timing of visits, choice of assessment tools, and appropriate sources of funding.

Epilepsy Database
A multicenter effort is needed to collect data on epilepsy in individuals with TSC from as many clinicians as possible. Data on genotype, seizure subtypes, the use of antiepileptic drugs and their efficacy and side effects, tuber location as determined by MRI, SPECT, PET, surgical outcome and tissue pathology all need to be collected and systematized. Thus a picture of the natural history of the disease may emerge, along with treatments that are likely to be effective. A particular emphasis was placed on the need to focus on infantile spasms, as these individuals are the youngest and their seizures among the most intractable.

Clinical TSC Database
A database to collect information on multiple aspects of TSC was also proposed. The database would help to determine the natural variance in the disease, and to establish parameters for appropriate clinical endpoints that are easy to assay. The goal would be to enroll about 2000individuals with TSC, with a focus on adults with TSC. For proliferative and metastatic symptoms, understanding the progression and variance will help to quantify the expected placebo response. The visits would take place every 6 months to 1 year, depending on the individual's age (more frequent up to age 3). A web based electronic data capture system would be implemented. Questions that the database should address include the following: 1) How does tuber burden affect seizure refractoriness; 2) How does tuber burden correlate with cognitive abilities; 3) How dose seizure refractoriness correlate with cognition; 4) What is the rate of progression of AML lesions; 5) What is the rate of progression of LAM lesions after the first signs of subclinical evidence on CT; 6) Which serum based biomarkers correlate with disease status?

Top

Attracting new investigators

Research opportunities and training fellowships are needed to attract additional researchers and clinicians. TSC is a rare genetic disorder but should be positioned more prominently because of the possible insights that could be gained from understanding the disease mechanism(s). Informational sessions could be held to increase awareness of TSC, and workshops at professional meetings should be organized. A great deal of enthusiasm was also expressed for organizing a Gordon Conference devoted to tuberous sclerosis.