Finding targets and treatments for Neurofibromatosis Type II

Reto Baumann and Philip Zimmermann
© NEBION AG. April 11, 2013


Neurofibromatosis type II is a rare genetic disease resulting in benign tumors in the vicinity of the auditory-vestibular nerve. Common symptoms are auditive loss and desequilibrium, both of which strongly impact the quality of life of the patients. In this study, we used GENEVESTIGATOR to compare the expression regulation in this disease with other cancer types and to identify genes specifically regulated in the disease. Although at the time of analysis only very few transcriptomic studies on NF2 were published, we found that, on the transcriptional level, NF2 resembles most to breast cancer and to ERBB2+ cancers. We also identified several genes specifically and strongly up-regulated in NF2, opening new avenues for therapeutic intervention.


Neurofibromatosis type II (NF II), also known as MISME (Multiple Inherited Schwannomas, Meningiomas and Ependyomas) syndrome, is a rare hereditary, autosomal dominant disease that can be passed on from parents to their children or occur spontaneously due to de novo acquired autosomal genetic mutations (Evans, 2009). One hallmark of this rare disease is the development of benign (non-malignant) bilateral tumors known as vestibular Schwannomas in the region of the auditory-vestibular nerve (cranial nerve VIII, nervus vestibulocochlearis) which is required for transmitting sensory information from the inner ear to the brain (Evans et al., 2000). These tumors originally descend from Schwann cells, the myelinating cells of the peripheral nervous system. In the course of expansive tumor growth, symptoms such as progressive hearing loss and disequilibrium are very common. People affected by NF II additionally develop a wide range of other problems such as tumors in other cranial nerves and Meningiomas, eye lesions, spinal lesions (e.g. spinal Astrocytomas and Ependyomas), and skin alterations. NF II caused by various different mutations in the gene coding for a protein called Schwannomin or Merlin (an acronym for Moesin-Ezrin-Radixin-Like protein), whereby patients with frameshift and nonsense mutations generally suffer poorer prognosis than patients carrying missense mutations in the coding sequences. Additionally, this specific protein is usually inactivated in sporadic unilateral Schwannomas of patients not affected by NF II (Striedinger et al., 2008). Merlin is a protein of 70 kDa and belongs to the ERM (Ezrin-Radixin-Moesin) protein family. In humans, 10 different isoforms of Merlin exist, but the function seems to be the same for all isoforms. They act as molecular linkers connecting the cytoskeleton to the cell membrane or the membrane-embedded glycoproteins. In doing so, they affect cell morphology, polarity, migration and signal transduction (McClatchey and Giovannini, 2005). Supposedly, Merlin is involved in contact-mediated inhibition of cellular growth and proliferation. Human Merlin is mainly produced in tissues of the nervous system and localizes preferably to adherens junctions, structures which connect cells to each other (den Bakker et al., 1999). Merlin is a so-called tumor suppressor since loss of function leads to aberrant cell behavior and the subsequent formation of expansive neoplastic tissue. As mentioned previously, one defective copy of this gene is sufficient to cause disease. Obviously, defective Merlin is somehow able to sequester or disturb the healthy protein population that is still present in the cells. Due to the complexity of the underlying disease causing mechanism, no cure for NF II has yet been found. Current therapeutic approaches mainly focus on the microsurgical removal and radiological treatment of neoplastic tissue developing at different locations throughout the body (Regis et al.; Regis et al., 2007). Since growth of these benign but space-occupying masses causes severe problems, an efficient systemic medical treatment is urgently needed. Studies investigating the effects of the anti-cancer drug Bevacizumab (distributed under the brand name Avastin by F. Hoffmann-La Roche) on NF II patients delivered promising results (Mautner et al., 2010). Bevacizumab uniquely binds to and neutralizes human VEGF-A, thereby reducing angiogenesis. Tumor tissue deprived of oxygen and nutrients is more prone to be negatively affected than healthy tissues and shrinks during the course of the treatment. Sunitinib (marketed as Sutent by Pfizer), a small molecule multi-targeted receptor tyrosine kinase (RTK) inhibitor, is being studied for treatment of menangioma (, NCT00589784). Since this compound inhibits, among others, c-KIT (CD117), all receptors for platelet-derived growth factor (PDGFRs) and vascular endothelial growth factor receptors (VEGFRs), it causes reduced tumor vascularization and cancer cell death, resulting in tumor shrinkage. Lapatinib (marketed as Tykerb/Tyverb by GlaxoSmithKline), an inhibitor of EGFR/ErbB2 signaling is also under current investigation (, NCT00863122 and NCT00973739). By reducing mitogenic signaling, scientists hope to decelerate and restrict growth of the tumors. To treat this disease with maximum efficiency, one has to learn about the basic cellular mechanisms affected by loss of Merlin function. Our growing arsenal of highly specific anti-cancer drugs gives us today the possibility to target single proteins in complex signaling cascades, thereby partially circumventing the massive side effects of conventional chemotherapy. Since these drugs are usually very expensive and only work under special cellular contexts, it is important to know the pathways affected in NF II associated tumors such as Schwannomas. In the following analytical in silico approach, we will use different tools from GENEVESTIGATOR to compare Schwannomas with better characterized neoplasms at the level of gene expression and dissect intracellular signaling pathways, thereby focusing on the discovery of novel potential therapeutic targets.


Find neoplasms with a gene expression signature most similar to vestibular Schwannomas

The vestibular Schwannoma is a rare type of tumor and has been only scarcely investigated by functional genomics approaches. Until the end of 2012, only three studies reported global gene expression profiles of vestibular Schwannomas, and these studies were based on data obtained on different microarray platforms. Even worse, apart from specific coincidences, these studies showed no common trends. In a recent study carried out by Torres-Martin et. al. (Torres-Martin et al., 2012), more meaningful results and deeper insights into the gene expression signature of Schwannomas could be generated. The results published by this team directly provide fold-change information for the top upregulated and top downregulated genes, in addition to the complete and processed microarray data which is available online. We used this expression signature in GENEVESTIGATOR to identify other conditions in which the same sets of genes are up- or down-regulated, respectively. At the time of analysis, the GENEVESTIGATOR database contained almost 32,000 datasets from hundreds of different healthy and diseased tissue types. This list of top-regulated genes, together with their relative expression values, were therefore a valuable starting point to find neoplasms and other diseased human tissues exhibiting similar gene expression patterns. Our experimental in silico approach was straightforward, since we did not have to import large amounts of primary microarray data. This approach also works with RT-qPCR fold change values whenever available. We therefore entered this list of genes and their corresponding log-ratios of expression into the GENEVESTIGATOR Signature tool and filtered for those probes being well detected and having a consistent pattern of expression.

Figure 1. Search for experimental conditions causing a similar expression signature as the most differentially expressed genes from vestibular Schwannomas. Among thousands of conditions screened by the SIGNATURE tool, ERBB+ type breast cancers appeared as being most similar.

From more than 1,000 different experimental conditions tested, the SIGNATURE tool identified breast cancer as the condition being the most similar to Schwannoma. This is in agreement with previous studies investigating cell signaling in vestibular Schwannomas. Both types of neoplasms are known to suffer from deregulated growth factor signaling mainly concerning EGF (or NRG1, respectively in Schwannomas) via receptors of the Her/ErbB receptor tyrosine kinase (RTK) family (Curto et al., 2007; Lallemand et al., 2009) (Beguelin et al.; Hartman et al.). The fact that various highly specific drugs approved for the clinical treatment of breast cancer are under current clinical investigation for use in NF II (see introduction) supports our findings. Furthermore, advanced breast cancer exhibits a loss of Merlin via post-translational mechanisms (Morrow et al.) and melanoma, which also appears in our results, is also known to exhibit alterations in Merlin protein levels (Li et al.).

Identify genes specifically regulated in ERBB2+ breast cancer

Increased growth factor signaling via the HER2/ErbB2 co-receptor has been found to be present in aggressive Schwannoma (Stonecypher et al., 2006; Hansen et al., 2006) as well as in a large percentage of malign breast cancers. ErbB2 is also very highly expressed in the only Schwannoma sample currently available in the GENEVESTIGATOR database. The increase in mitogenic signaling is usually a result of gene duplication and/or overexpression of the ErbB2 gene. GENEVESTIGATOR allows identifying genes specifically regulated under certain pathological conditions and experimental perturbations (with minimal regulation in all other conditions). Since changes in gene expression between breast cancer and Schwannomas seem to be comparable, we used this elegant in silico approach to identify additional therapeutic targets valid for the treatment of both malignancies. Therefore, we used the PERTURBATIONS tool selecting the pathological conditions in which ErbB2 is clearly overexpressed (Her2+ experimental breast cancer samples). We obtained a list of 50 genes that clearly have the same expression profile as ERBB2, i.e. up-regulated in these cancer types but minimally regulated across more than 1,000 other conditions. To obtain a more comprehensible representation of our results, we exported the newly generated data to the HIERARCHICAL CLUSTERING tool. With this tool, we grouped genes and conditions in our experimental section according to their expression across these relevant conditions. Three genes appeared to be closely related to ERBB2: SBK1, NVL and probeset 242216_at. In a further step, we extended the search for genes having high similarity with ERBB2 across all samples in which ERBB2 is significantly regulated (and not only across the breast cancer ERBB2+ samples). Clustering the resulting list of 10 genes revealed a tight cluster of five genes including ERBB2, GRB7, PGAP3, CDK12, GSDMB. Interestingly, the GRB7 gene is located on the same chromosome just next to ERBB2 and in this analysis exhibited the most similar expression profile.

Figure 2. Clustering of the top 10 genes most correlated with ERBB2 across different ERBB2+ cancers. A tight cluster involving ERBB2, PGAP3, GRB7, CDK12 and GSDMB is highlighted in red.


In the study mentioned above, genes differentially expressed between NF2 and healthy controls were identified. In Genevestigator, we can go one step further to look for genes highly expressed in NF2 but weakly or non-expressed in all other types of cancer or tissues. Using the GENE SEARCH Neoplasms tool, we identified several genes having a high specificity score for NF2. The search was performed across 1012 categories of neoplasms and normal tissues, with the exclusion of neoplasm cell lines and cell cultures, and choosing "cellular Schwannoma" as target category. Since this category is represented by a single sample, the results are to be interpreted carefully, but nevertheless it's a good starting point to identify NF2 specific genes. The top ten probesets identified represent the genes GFRA3, SCN7A, CDH19, NRXN1, ALB, RELN and SORCS1, with GFRA3 being the top candidate. Apart from GFRA3 and SCN7A, all of these candidates showed significant levels of expression in central nervous system tissues, suggesting that our target category "cellular schwannoma" may have contained cells from adjacent tissues. We therefore discarded these candidates as being unspecific for NF2. While SCN7A was not expressed in CNS tissues but was expressed above background levels in more than 50 other cancer categories, GFRA3, in contrast, was highly specific for Schwannoma and a small group of other cancers. In fact, GFRA3 had expression above background levels in only 27 from the 1012 neoplasm and normal tissue categories tested. In NF2, the expression of GFRA3 was on average about 10 times higher than in these 27 categories, except in ganglioneuroblastoma, endometrioid adenocarcinoma of the ovary (represented by a single sample), and neuroblastoma (NOS). The protein encoded by the GFRA3 gene is a GPI-linked cell surface receptor and a member of the GDNF receptor family. This protein forms a signaling complex with the RET tyrosine kinase receptor. The function of GFRA3 as a mediator of an artemin-induced autophosphorylation of the RET receptor tyrosine kinase suggest that this kinase is over-activated in NF2 and could play a role in the progression of the disease. If the extremely high expression of GFRA3 is confirmed in other NF2 samples, then we may conclude that both GFRA3 as well as the RET receptor tyrosine kinase could be valuable drug targets for NF2. Moreover, if specific inhibitors already exist for these two proteins, they could be valuable candidates for repositioning them towards NF2 treatment.

Figure 3. Top 50 categories of neoplasms and normal tissues with highest expression of GFRA3. The level of expression in NF2 is indicated with an arrow. Since this result is from a single sample, it needs to be verified in other, independent NF2 experiments..


In this case study we used different tools of Genevestigator to compare NF2 with other tumor types. Starting with very little initial data, we quickly managed to find pathological conditions with similar gene expression signatures using the SIGNATURE tool, revealing a close relationship with breast cancer, in particular with ERBB2 positive cancers. This finding is corroborated by recent clinical trials repositioning breast cancer drugs towards NF2 treatment. In a further step, we used the GENE SEARCH Perturbations tool to identify genes that are specifically regulated in ERBB2 positive breast cancers and clustered them using the HIERARCHICAL CLUSTERING tool. Finally, we used the GENE SEARCH Neoplasms tool to find genes specifically expressed in NF2. A highly interesting cell surface protein encoding gene was identified, opening new leads for drug discovery and repositioning. Knowing about these specific molecular signatures facilitates the development of novel drugs or the transfer established treatment regimens to diseases such as NF type II which are currently only difficult to manage.

To try out GENEVESTIGATOR or to replicate/extend this analysis, please go here.


Beguelin, W., Diaz Flaque, M.C., Proietti, C.J., Cayrol, F., Rivas, M.A., Tkach, M., Rosemblit, C., Tocci, J.M., Charreau, E.H., Schillaci, R., et al. (2010) Progesterone receptor induces ErbB-2 nuclear translocation to promote breast cancer growth via a novel transcriptional effect: ErbB-2 function as a coactivator of Stat3 Molecular and cellular biology 30, 5456-5472.  [Abstract]

Curto, M., Cole, B.K., Lallemand, D., Liu, C.H., and McClatchey, A.I. (2007) Contact-dependent inhibition of EGFR signaling by Nf2/Merlin. The Journal of cell biology 177, 893-903.  [Abstract]

den Bakker, M.A., Vissers, K.J., Molijn, A.C., Kros, J.M., Zwarthoff, E.C., and van der Kwast, T.H. (1999) Expression of the neurofibromatosis type 2 gene in human tissues. J Histochem Cytochem 47, 1471-1480.  [Abstract]

Evans, D.G. (2009) Neurofibromatosis type 2 (NF2): a clinical and molecular review. Orphanet journal of rare diseases 4, 16.  [Abstract]

Evans, D.G., Sainio, M., and Baser, M.E. (2000) Neurofibromatosis type 2. Journal of medical genetics 37, 897-904.  [Abstract]

Hansen, M.R., Roehm, P.C., Chatterjee, P., and Green, S.H. (2006) Constitutive neuregulin-1/ErbB signaling contributes to human vestibular schwannoma proliferation. Glia 53, 593-600.  [Abstract]

Hartman, Z.C., Yang, X.Y., Glass, O., Lei, G., Osada, T., Dave, S.S., Morse, M.A., Clay, T.M., and Lyerly, H.K. (2011) Title HER2 overexpression elicits a proinflammatory IL-6 autocrine signaling loop that is critical for tumorigenesis.  [Abstract]

Lallemand, D., Manent, J., Couvelard, A., Watilliaux, A., Siena, M., Chareyre, F., Lampin, A., Niwa-Kawakita, M., Kalamarides, M., and Giovannini, M. (2009) Merlin regulates transmembrane receptor accumulation and signaling at the plasma membrane in primary mouse Schwann cells and in human schwannomas. Oncogene 28, 854-865.  [Abstract]

Li, W., Cooper, J., Karajannis, M.A., and Giancotti, F.G. (2012) Merlin: a tumour suppressor with functions at the cell cortex and in the nucleus. EMBO reports 13, 204-215.  [Abstract]

Mautner, V.F., Nguyen, R., Kutta, H., Fuensterer, C., Bokemeyer, C., Hagel, C., Friedrich, R.E., and Panse, J. (2010) Bevacizumab induces regression of vestibular schwannomas in patients with neurofibromatosis type 2. Neuro-oncology 12, 14-18.  [Abstract]

McClatchey, A.I., and Giovannini, M. (2005) Membrane organization and tumorigenesis--the NF2 tumor suppressor, Merlin. Genes & development 19, 2265-2277.  [Abstract]

Morrow, K.A., Das, S., Metge, B.J., Ye, K., Mulekar, M.S., Tucker, J.A., Samant, R.S., and Shevde, L.A. (2011) Loss of tumor suppressor Merlin in advanced breast cancer is due to post-translational regulation. The Journal of biological chemistry 286, 40376-40385.  [Abstract]

Regis, J., Carron, R., Moucharrafien, S., Delsantin, C., Porcheron, D., Thomassin, J.M., Murracciole, X., and Roche, P.H. (2012) Role of radiosurgery and stereotactic radiotherapy in the management of vestibular schwannomas Cancer Radiother 16 Suppl, S70-78.  [Abstract]

Stonecypher, M.S., Chaudhury, A.R., Byer, S.J., and Carroll, S.L. (2006) Neuregulin growth factors and their ErbB receptors form a potential signaling network for schwannoma tumorigenesis. Journal of neuropathology and experimental neurology 65, 162-175.  [Abstract]

Striedinger, K., VandenBerg, S.R., Baia, G.S., McDermott, M.W., Gutmann, D.H., and Lal, A. (2008) The neurofibromatosis 2 tumor suppressor gene product, merlin, regulates human meningioma cell growth by signaling through YAP. Neoplasia 10, 1204-1212.  [Abstract]