Interaction between Sox-10 and the tumour suppressor Merlin in tumours of the nervous system.

Tumour suppressors are the brakes of the cell that prevent abnormal cell division in our bodies. Loss of the tumour suppressor Merlin leads to tumours in multiple cell types within our nervous systems. We have two copies of a tumour suppressor (one on each chromosome we inherit from our parents), and Merlin deficient tumours can commonly arise either by sporadic loss of both copies in a single cell, or in familial cases of cancer (eg. Neurofibromatosis type 2, NF2) by inheriting one abnormal copy and the second copy been lost in a cell during our lifetime. As would be predicted in patients with NF2, the onset of disease is much earlier and patients have multiple tumours. Tumours can arise from the Schwann cells that wrap and myelinate axons (schwannomas), meningeal cells that form the membrane around the brain and spinal cord (meningiomas), and ependymal cells that line the ventricles of the brain (ependymomas). These tumours grow slowly, are resistant to chemotherapy, and require multiple difficult surgical operations to remove the tumour tissue. In NF2 patients this means multiple tumours and surgeries, starting in their teenage years and continuing through their lifetime. Patients with NF2 typically present with hearing loss and balance problems due to schwannoma tumours on the vestibulocochlear nerve which carries hearing and balance information to the brain. We use primary human schwannoma cells from patients as an in vitro model of the disease, and study interactions between Merlin and other proteins, specifically transcription factors, which bind DNA and regulate other genes, control the normal stop of proliferation and promote differentiation of Schwann cells in our peripheral nervous system. Analysis of human schwannoma tumours, both in vitro and in vivo, shows that the level of a transcription factor called Sox-10 is reduced or absent in all schwannoma tumours analysed (n=14). Sox-10 is important not only in the control of Schwann cell-axon interaction, but also in stopping Schwann cells from proliferating and drives the induction and maintenance of these cells to a differentiated, quiescent state. Re-introduction of Sox-10 alone into Merlin null human schwannoma cells allows them to induce a differentiated quiescent state. In humans, Merlin and Sox-10 lie close together on chromosome 22 and, in at least 60% of tumours, the two genes are lost together in the second 'hit' that occurs in NF2. We believe that this is one mechanism by which tumours lose Sox-10 function, but there may be others involving the regulation of the Sox-10 gene itself. We plan to comprehensively study the role of Sox-10 in the pathogenesis of Merlin null tumours in both human primary schwannoma cells and a transgenic knockout approach in mice. For the work with primary human schwannoma cells, we will study both the regulation of Sox-10 and the effects of Sox-10 re-introduction into these cells. In the transgenic mice, these offer a unique way to study the interaction between Merlin and Sox-10 proteins. Previous attempts to model NF2 in mice have shown that loss of Merlin alone induces a very different spectrum of tumours compared to the human disease; we think that the untested role of Sox-10 in these tumours may lie behind these differences. Because Sox-10 and Merlin are on different mouse chromosomes, we can remove Merlin in Schwann cells on a Sox-10 wild-type, hemizygous (only 1 copy of the gene) or Sox-10 null background both in vitro and in vivo, and truly study the effects of Sox-10 function on the pathology of these tumours. We can then correlate these findings in the mouse with our findings in primary human schwannoma cells. The endpoint of these experiments will be twofold: Firstly, to understand how loss of Merlin and Sox-10 co-operate to cause schwannoma tumours. Secondly, and more importantly, to create a new more accurate model of disease for NF2 in mice to allow in vivo testing of new compounds for treatment of these tumours.

It is estimated that 1/300 individuals will develop a tumour due to Merlin loss in their lifetime. In individuals with neurofibromatosis type 2 or NF2 (incidence 1/35,000) inheritance of one mutant Merlin allele and somatic loss of the second leads to patients developing multiple tumours of the nervous system, typically presenting with bilateral vestibular schwannomas but also meningiomas, ependymomas and gliomas, which require repeated surgery for their removal during the lifetime of the patient. Typical life expectancy for a NF2 patient is less than 40 years. Attempts to model NF2 in mice, using conditional Merlin knockout mice, to develop therapies for the human disease have been difficult because the occurrence and tissue distribution of tumours following Merlin loss are different between mouse and man, and it is proposed there are additional genetic changes in human tumours. Many of the transcription factors that control Schwann cell proliferation and differentiation have now been characterised. These include the transcription factor Sox-10 which controls both axon-Schwann cell interaction and induction of Krox-20, the master regulator of differentiation, and together they control both proliferation and myelination of Schwann cells. Using primary human schwannoma cells we have discovered that Sox-10 is reduced or absent in these tumours in vitro and in vivo. We also find and that restoration of Sox-10 in Merlin null cells restores the ability of Merlin null cells to induce Krox-20 and differentiate. Merlin and Sox-10 are both located on the long arm of chromosome 22 and our analysis shows these two genes are lost together in at least 60% of the somatic mutations occurring in NF2. We propose that loss of Sox-10 function in the human Merlin null tumours is a novel and potential major modifier of Schwann cell phenotype in human tumours and plan to model this both in vitro in human schwannoma cells and in vivo using Merlin and Sox-10 conditional mouse strains.

The primary aim of this work is the greater understanding of the function of the Merlin tumour suppressor, and for a better understanding of tumours due to Merlin loss. Our preliminary findings, together with the experiments detailed in this proposal, is examining Merlin function in schwannoma cells in a novel way that has not been addressed previously. Although much is known about proteins that interact with Merlin, very little is known about what impact Merlin loss has upon the well characterised process of cell differentiation and control of proliferation in Schwann cells. The latter has been my area of research for the last 15 years and with access to both human tumour tissue and the appropriate conditional knockout mice, then we can begin to answer this question. The preliminary findings in this grant have already been presented at international conferences and have generated a lot of interest in the idea of connecting the fields of schwannoma tumour cell biology and Schwann cell development. This project would represent an excellent training and environment for a post-doctoral scientist, the opportunity to present their data both at national and international conferences, and the basis for a career in research in the area of Schwann cell or tumour biology.
Multiple types of tumours are caused by the loss of the Merlin tumour suppressor. All schwannomas, both sporadic and in cases of neurofibromatosis type2 (NF2) are Merlin null, and individuals with NF2 also develop meningiomas, retinal hamartomas and ependymomas. In the related disease schwannomatosis, whilst some cases are caused by mutations in another gene, INI1/SMARCB1, the majority of cases are also due to loss of Merlin function. In addition to this, it is thought that mutations in Merlin account for between 50-60% of sporadic meningiomas, which account for between 20-30% of all brain tumours, a small proportion of sporadic ependymomas as well as in a large proportion of mesotheliomas. As the majority of the tumours in the nervous system are benign, they do not respond well to chemotherapy and surgical removal is often the only available treatment. Particularly in patients with NF2, who may carry a large tumour load, this means multiple surgeries through their lifetime, starting in their teens, and meaning that the average life expectancy of a patient with NF2 is typically lower than 40 years. Clearly a greater understanding of the function of Merlin alone in cells together with how Merlin loss may co-operate with other proteins in the cell is vital to the design and eventual trial of new therapies for these tumours. A cytogenetic and microsatellite analysis has shown that the region encoding Sox10 is lost at the same time as the second Merlin allele in at least 60% of schwannoma tumours, but in cases where there is no apparent chromosomal loss, our work, if successful, may suggest that genetic screening of tumours for Sox10 mutations may be informative. Indeed, in collaboration with Dr Gareth Evans (Manchester, UK), we have begun to screen tumours with no apparent loss of Chr22 for Sox10 mutations. Using the NF2 and Sox10 conditional knockout mouse strains, we hope to be able to model more successfully the occurrence and location of tumours that are seen in patients with NF2. This, we hope will have several impacts. Firstly, we hope that this will generate a freely available and novel model of NF2 and schwannoma that may be used for the testing of new drugs and therapies. I am happy to talk to both colleagues and the public about how my work involves the use of animal model systems to understand human disease, and I hope that this project will lead to an excellent example of how basic research can lead to an improved understanding and modelling of an important human disease. Secondly, within the NF2 and Schwann cell research field, this will allow a greater understanding of how Merlin function impacts on both Schwann cell development and tumourigenesis.