Sporadic schwannomas and neurofibromas

Pathogenesis and pathophysiology
Article section 6 of 16.  Previous  Next

By Herbert B Newton MD

Schwannomas and neurofibromas are classified as peripheral nerve sheath tumors (Russell and Rubinstein 1989; Parisi and Mena 1993). Schwannomas are composed of a homogeneous mass of transformed Schwann cells in a collagenous background. Neurofibromas also contain numerous transformed Schwann cells, but also include an admixture of transformed perineurial cells and fibroblasts.

Schwannomas typically arise from cranial nerves, spinal nerves and nerve roots, autonomic nerves, and peripheral nerves. The tumors always arise from the nerve at the transition zone between the central glial nerve sheath and the peripheral Schwann cell nerve sheath. Rarely, schwannomas can occur within the substance of the brain or spinal cord (Aryanpur and Long 1988; Russell and Rubinstein 1989; Sharma et al 1993; 1998; Singh et al 1993). There is a predilection for schwannomas to affect sensory nerves, although motor nerves can be involved as well.

On gross pathologic inspection, schwannomas appear as discrete, rounded, firm, encapsulated masses of a semitranslucent or milky white color arising from a nerve fascicle. The tumors may have variable amounts of cyst formation, yellowish areas of xanthomatous changes, and hemorrhage. During the early "intraneural" phase of growth, the tumor is fusiform in shape, similar to neurofibromas (Parisi and Mena 1993). As the tumor enlarges, the adjacent nerve fascicles are compressed and displaced eccentrically. The nerve fascicles are usually not infiltrated by, or encased within the mass, although they may be incorporated superficially into the tumor capsule. Large tumors may distort and compress other surrounding neural structures (eg, brainstem, spinal cord) without infiltration or invasion. The gross cut surface of the tumor may demonstrate regions that have a whorling appearance and will usually not contain nerve fibers deep to the capsule.

Sporadic neurofibromas (ie, occurring in patients without neurofibromatosis type 1 or 2) typically arise from small cutaneous terminal nerves; less commonly, they develop in large peripheral nerves, spinal nerves, or spinal nerve roots (Russell and Rubinstein 1989; Parisi and Mena 1993). Intracranial neurofibromas of the cranial nerves are extremely rare in patients without neurofibromatosis type 1 or 2. On gross pathologic examination, neurofibromas are soft, well-circumscribed, pedunculated, and unencapsulated gelatinous masses of a whitish or opalescent color. Regions of cyst formation, xanthomatous changes, and hemorrhage are not seen as commonly as they are with schwannomas. During initial phases of growth, the tumor infiltrates the parent nerve, causing a localized, fusiform swelling. As the tumor enlarges, the parent nerve and those nerves around it may develop gross alterations of shape (eg, "bag of worms"), and can become encased within the mass. The gross cut surface of the tumor is devoid of the whorling texture noted in schwannomas, and contains more frequent nerve fibers deep to the capsule.

On microscopic histological examination, "classic" schwannomas are composed of a heterogeneous, biphasic architecture that contains 2 distinct regions: Antoni A and Antoni B (Russell and Rubinstein 1989; Parisi and Mena 1993; Macfarlane and King 1995; Strauss and Post 1995). In most tumors, the Antoni A regions predominate, and are organized into dense, compact rows or arrays of elongated, spindle-shaped cells that have hyperchromatic, rod-shaped nuclei and eosinophilic cytoplasm. The nuclei are often aligned into palisades that alternate with dense, anuclear zones of fibrillar eosinophilic material; these structures are called Verocay bodies. Verocay bodies are less common in vestibular schwannomas than spine or peripheral nerve schwannomas. In some tumors, the cell bundles form whorls of various sizes. The Antoni B regions are loosely organized and composed of large, vacuolated, pleomorphic stellate cells with pyknotic or irregular nuclei. Areas of microcystic change, hyalinization of blood vessels, hemorrhage with perivascular hemosiderin deposition, lipid accumulation, and nuclear degenerative atypia are common. Mitoses and nuclear pleomorphism can be seen on occasion, but do not imply malignant potential. Cell cultures derived from Antoni A and B areas produce distinctive types of Schwann cells, as confirmed by electron microscopic studies. More detailed electron microscopic analysis of Antoni A regions demonstrates a lamellar pattern of numerous thin, elongated, cytoplasmic processes that are coated by a dense basal lamina and separated from each other by intercellular basement membrane material (Russell and Rubinstein 1989). The tissue in Antoni B regions is characterized by large numbers of organelles (eg, mitochondria, lysosomes, osmiophilic bodies) and vacuoles. Immunohistochemical typing of schwannomas demonstrates strong reactivity to S-100 protein, as well as Leu-7 and myelin basic protein. The S-100 protein is cytoplasmic in origin, and is often used to identify nerve sheath tumors (Parisi and Mena 1993; Macfarlane and King 1995). In contrast, meningiomas stain weakly for S-100 protein. A small proportion of schwannomas also stain for glial fibrillary acidic protein (Parisi and Mena 1993).

In addition to the "classic" microscopic appearance of these tumors, several less common forms exist, including the cellular, ancient, plexiform, melanotic, and malignant schwannoma variants (Russell and Rubinstein 1989; Parisi and Mena 1993; Macfarlane and King 1995; Strauss and Post 1995). The cellular schwannoma variant is characterized by a predominantly compact Antoni A pattern without well-formed palisades or Verocay bodies. Anaplastic features such as mitotic activity, high cellularity, and nuclear pleomorphism may occur, but are not associated with malignant potential (White et al 1990; Deruaz et al 1993; Parisi and Mena 1993; Casadei et al 1995). The ancient schwannoma variant demonstrates relative hypocellularity and extensive degenerative changes that include cyst formation, marked vessel hyalinization, calcification, hemorrhage, and nuclear atypia (Parisi and Mena 1993). Mitoses are usually not present. The plexiform schwannoma variant has typical histology except for a relative lack of the Antoni B component. These tumors grow in a characteristic plexiform or multinodular pattern and represent approximately 4.3% of all schwannomas, often affecting the head and neck region (Berg et al 2008). Melanotic schwannomas have melanin pigment that is abundantly scattered throughout the tumor (Parisi and Mena 1993). They typically arise in autonomic nerves and behave in a benign fashion. The malignant schwannoma variant usually arises de novo, not as anaplastic degeneration in a previously benign schwannoma (Russell and Rubinstein 1989). The characteristic features are high cellularity, increased mitotic activity, and the arrangement of spindle-shaped tumor cells into a "herringbone" pattern. Cellular and nuclear pleomorphism is more widespread and areas of necrosis are prominent. Positive reactivity to S-100 is retained in most tumors, confirming their origin from Schwann cells. Although rare, malignant schwannomas can derive from typical low-grade tumors, develop a more accelerated growth rate, and become locally invasive (Hanabusa et al 2001).

On microscopic histological examination, the typical cutaneous or spinal neurofibroma has interlacing bundles of fusiform Schwann cells with wavy nuclei within a matrix of collagen-rich and mucopolysaccharide-rich material (Russell and Rubinstein 1989; Parisi and Mena 1993). Because of the infiltrative growth pattern of neurofibromas, axons of nerve fibers are easily demonstrated after silver impregnation of tumor tissue. Features common to schwannomas such as palisading, Verocay bodies, and whorling are noted infrequently. The architecture and cellular appearance can vary, producing several subtypes: storiform perineural fibroma, pacinian neurofibroma, epithelioid neurofibroma, and pigmented neurofibroma (Russell and Rubinstein 1989; Parisi and Mena 1993). Rarely, neurofibromas can contain foci of classic schwannoma (Feany et al 1998). Ultrastructural studies demonstrate the presence of Schwann cells, perineurial cells, fibroblasts, and collagen. Immunohistochemical typing of neurofibromas shows strong reactivity for vimentin and S-100 in most tumors, with less frequent reactivity for Leu-7 (Russell and Rubinstein 1989; Parisi and Mena 1993). Studies of the vasculature of sporadic and neurofibromatosis type 1-associated neurofibromas reveal a high degree of vascularization (Arbiser et al 1998). In addition, abundant perivascular staining for vascular endothelial growth factor is common. Anaplastic degeneration of a previously benign sporadic neurofibroma into a malignant neurofibroma is rare in patients without neurofibromatosis type 1. Malignant neurofibromas have increased cellularity, nuclear and cellular pleomorphism, and frequent mitotic activity. Foci of necrosis, as well as regions of metaplasia (eg, cartilage, osteoid), may be present.

Techniques that can measure the biological potential of schwannomas and neurofibromas include Ki-67, bromodeoxyuridine, MIB1, and proliferating cell nuclear antigen labeling studies, as well as flow cytometric analysis (Cho et al 1988; Nishizaki et al 1988; Nishizaki et al 1989; Louis et al 1991; Deruaz et al 1993; Casadei et al 1995). In general, these studies are consistent with benign neoplasms that usually have small growth fractions and limited aggressive biological potential. Labeling studies using Ki-67 and bromodeoxyuridine demonstrate indices of 0.5% to 1.5% in classic schwannomas (Nishizaki et al 1988; 1989; Louis et al 1991). Casadei and colleagues determined labeling indices for proliferating cell nuclear antigen and MIB1 in a cohort of nonrecurrent and recurrent cellular schwannomas (Casadei et al 1995). The mean proliferating cell nuclear antigen labeling index was approximately 5.6% for both nonrecurrent and recurrent lesions, whereas for MIB1 the mean labeling index was 6% in nonrecurrent tumors, and 8% in recurrent lesions. In contrast, a mean proliferating cell nuclear antigen labeling index of 19.5% was noted for a group of classic schwannomas by Louis and colleagues (Louis et al 1991). Deruaz and colleagues compared proliferating cell nuclear antigen labeling between classical and cellular schwannomas (Deruaz et al 1993). For their small sample of cellular schwannomas, the proliferating cell nuclear antigen labeling percentage was higher than that measured for classical schwannomas (44.6% vs. 8.1%); however, this was not correlated with more aggressive clinical behavior. DNA flow cytometric studies of classic schwannomas have shown that despite variability in ploidy status (ie, tumors may be diploid or contain aneuploid populations), the S-phase fraction is generally low: 1.0% to 2.0%; this is similar to labeling studies with Ki-67 and bromodeoxyuridine (Cho et al 1988; Nishizaki et al 1988; 1989). Flow cytometric studies of cellular schwannomas showed similar ploidy data; 62% of tumors were diploid and 30% to 35% were tetraploid or aneuploid (Casadei et al 1995). However, the overall mean S-phase fraction of 6% was higher than that seen in classic schwannomas. In the subset of aneuploid tumors, the mean S-phase fraction (11.4%) was even larger.

Cytogenetic and chromosomal studies of sporadic schwannomas reveal several consistent findings (Rey et al 1987; Couturier et al 1990; Lodding et al 1990; Fontaine et al 1991; Webb and Griffin 1991; Lanser 1992). Many of the karyotype analyses contain normal, diploid stem cells. The most common abnormality, as determined by loss of heterozygosity studies, is monosomy of chromosome 22, which has been noted in vestibular and spinal tumors (Seizinger et al 1986; Rey et al 1987; Fontaine et al 1991). In addition, the long arm of chromosome 22 can manifest deletions, inversions, and translocations. Cytogenetic analysis of cellular schwannomas has shown similar abnormalities of chromosome 22 (Lodding et al 1990). In addition, most of the abnormal clones detected were hypodiploid, including monosomy 15, loss of the X chromosome, and loss of the long arm of chromosome 3.

Because of the frequent loss of heterozygosity of chromosome 22 in sporadic and inherited schwannomas, and the results of linkage analysis studies that implicated the long arm of the chromosome, investigators began to search for a suspected tumor suppressor gene in this location that might be involved in Schwann cell growth control (Seizinger et al 1986; Rouleau et al 1987; 1993; Trofatter et al 1993). Further studies in patients with neurofibromatosis type 2 localized the candidate neurofibromatosis type 2 gene to a 6 Mb region of the q12 band of the long arm of chromosome 22 (Rouleau et al 1993; Trofatter et al 1993; Lutchman and Rouleau 1996; Gusella et al 1999; Xiao et al 2003; Welling et al 2007). The neurofibromatosis type 2 gene has 17 exons that encode for 595 amino acids and produce a protein (called either “merlin” or “schwannomin”) that has extensive sequence homology with membrane-interactive cytoskeletal proteins (eg, ezrin, moesin, radixin) (Rouleau et al 1993; Trofatter et al 1993; Gusella et al 1999; Xiao et al 2003; Beltrami et al 2013; Pecina-Slaus 2013). Some data demonstrate that the expression of ezrin, moesin, and radixin remains intact in schwannomas, despite the absence of merlin, suggesting that schwannoma tumorigenesis is not associated with loss of other ezrin, radixin, or moesin proteins (Stemmer-Rachamimov et al 1997). It is postulated that merlin plays a critical role in the formation of plasma membrane and cytoskeletal networks necessary to regulate cell adhesion and cellular proliferation (Huynh and Pulst 1996; Lutchman and Rouleau 1996; Xiao et al 2003; Welling et al 2007; Chang and Welling 2009). In support of this potential mechanism, schwannoma tumor cells have enhanced adhesion that depends on integrin chains alpha 6 beta 1 and alpha 6 beta 4 (Utermark et al 2003). Data suggest that the NF2 gene product, merlin, is under phosphorylative control and is active when serine 518 is hypophosphorylated (Surace et al 2004). Merlin loses its ability to inhibit cell growth and cell motility when S518 becomes hyperphosphorylated. Merlin appears to control the activity of PAK1, a Rac/CDC42-dependent serine/threonine kinase (Hirokawa el al 2004). Loss of functional merlin may activate PAK1, leading to transformation of NF2-deficient cells. Analysis of the neurofibromatosis type 2 gene often contains germline mutations in patients with neurofibromatosis type 2, and acquired, somatic mutations only within tumor tissue in patients with sporadic schwannomas. Numerous investigators have now revealed a wide spectrum of neurofibromatosis type 2 mutations in sporadic schwannomas (ie, 60% to 75% of analyzed tumors) that can cause abnormal expression of the merlin protein (Irving et al 1993; Bianchi et al 1994; Irving et al 1994; Jacoby et al 1994; Sainz et al 1994; Twist et al 1994; Jacoby et al 1996; Sainz et al 1996; Welling et al 1996; Welling 1998; Chang and Welling 2009). Alterations of the neurofibromatosis type 2 gene can include deletions (ranging from 1 bp to 79 bp), point mutations, splice site mutations, nonsense mutations, insertions, and missense mutations. The most common mutations are small deletions that result in frameshifts, leading to truncation of the C-terminal region of the protein product and abrogation of the ability to interact with cytoskeletal proteins (Bianchi et al 1994; Irving et al 1994; Jacoby et al 1994; Sainz et al 1994; Jacoby et al 1996; Sainz et al 1996; Welling et al 1996; Welling 1998; Welling et al 2007). Immunohistochemical and western blotting studies are consistent with the molecular data and reveal reduced or absent merlin expression in the majority of sporadic schwannomas (Gutmann et al 1997; Hitotsumatsu et al 1997; Stemmer-Rachamimov et al 1997; Harwalkar et al 1998). The absence of merlin expression appears to be universal in these tumors and is even noted in tumors that lack genetic evidence of complete neurofibromatosis type 2 gene inactivation (Stemmer-Rachamimov et al 1997). In addition, protein expression studies have not revealed the presence of truncated or abnormally sized merlin products, suggesting that mutant merlin proteins are unstable and may undergo rapid degradation (Harwalkar et al 1998). Epigenetic alterations of the NF2 gene may also play a role in schwannoma tumorigenesis (Gonzalez-Gomez et al 2003). In a series of 44 tumors, 18% had hypermethylation of the NF2 gene, resulting in silencing of gene expression. In addition, hypermethylation was noted in the retinoblastoma (Rb) and p16 genes in 15% of the tumors. Data by Kimura and colleagues implicate the calcium-dependent neutral cysteine protease calpain in the process of merlin proteolysis (Kimura et al 1998). In tissue cultures from explanted schwannomas, Schulze and colleagues induced viral transduction and stable expression of wild-type merlin (Schulze et al 2002). They noted reduced proliferation and G0/G1 arrest in affected cells. In addition, the rate of apoptosis was increased in transduced schwannoma cells. This is consistent with data from Utermark and colleagues, who noted an increased basal apoptotic rate in primary schwannoma cells in comparison to normal Schwann cells (Utermark et al 2005). Abnormalities of DNA repair mechanisms have also been implicated in the NF2 mutation process in sporadic schwannomas, with an increased ratio of somatic frameshift to nonsense mutations (Evans et al 2005). This effect appears to be age-related because it is more pronounced in older individuals. Analysis of epigenetic silencing of other genes has been attempted by Lassaletta and colleagues, including PTEN, Rb, MGMT, RASSF1A, von Hippel-Lindau, CASP8, and others (Lassaletta et al 2006). The results noted methylation of 12 of 16 genes, ranging from 9% to 27%, with a significant association between methylation and the CASP8 and RASSF1a genes. CASP8 was associated with patient age and tumor size, whereas RASSF1A was inversely correlated with the clinical growth index. More recent microarray studies by this same group have noted dysregulation of the MET pathway, as well as upregulation of the osteopontin gene (SPP1), which is involved in the degradation of the merlin protein (Torres-Martin et al 2013). A new study from Wong and colleagues has provided a link between NF-2 mutations and the angiogenic phenotype in schwannomas (Wong et al 2012). Schwannoma cells lacking merlin/NF-2 were noted to have significant downregulation of semaphorin 3F (SEMA3F). When SEMA3F is reintroduced into tumor cells, there is normalization of neoplastic vasculature, reduced tumor burden, and extended survival in nude mouse models. Merlin appears to regulate expression of SEMA3F through the Rho GTPase family member, Rac1. It was suggested that therapies designed to increase expression or activity of SEMA3F might be of benefit.

It is known that 1 function of merlin is to bind to PIKE-L, with subsequent binding and inhibition of the enzyme phosphoinositol 3-kinase (PI3K). Reduced levels of merlin should result in higher activity of PI3K and its downstream pathway components. Data from Jacob and colleagues have demonstrated that in schwannoma cells, the PI3K pathway is activated, with elevated activity of PI3K, total Akt, phospho-Akt, and m-TOR (Jacob et al 2008).

Involvement of other tumor suppressor genes has been unrevealing thus far. Irving and colleagues performed a loss of heterozygosity analysis on DNA from 41 sporadic vestibular schwannomas, looking for evidence of tumor suppressor genes on chromosomes 3p, 5q, 11p, 17p, 17q, and 22q (Irving et al 1993). No loss of heterozygosity was found, except for chromosome 22q (39% of cases). Bruder and colleagues studied 50 cases and reported a subset of tumors with deletions on chromosome 22q outside the neurofibromatosis type 2 locus that do not contain mutations within the neurofibromatosis type 2 gene (Bruder et al 1999). They conclude there may be heterogeneity in the oncogenesis of schwannomas and that additional genes on chromosome 22 may be important for tumor development. In a study of cellular schwannomas, Casadei and colleagues analyzed 51 tumors for the presence of elevated levels of p53 using immunohistochemical techniques (Casadei et al 1995). Twenty-six of 51 tumors (52%) stained for p53, although the percentage of positively staining cells was low in most of the tumors. Monoh and colleagues evaluated 21 tumors using restriction fragment length polymorphism analyses to evaluate for evidence that p53 is involved in schwannoma development (Monoh et al 1998). They were unable to detect loss of heterozygosity, deletions, or mutations, and concluded that p53 does not play a role in the oncogenesis of vestibular schwannomas. Cardillo and colleagues reported the expression of transforming growth factor-beta1 in a series of 31 vestibular schwannomas using immunohistochemical techniques (Cardillo et al 1999). Transforming growth factor-beta1 has been implicated in various processes, including extracellular matrix protein formation, angiogenesis, cell chemotaxis, and nervous system development. In 26 of 31 tumors (83.7%), transforming growth factor-beta1 was positively expressed within tumor cells with higher levels noted in Antoni A regions. The majority of tumors with positive expression were also found to stain strongly for transforming growth factor-beta1 within the vasculature. In a series of 30 sporadic vestibular schwannomas, Mawrin and colleagues evaluated the expression and presence of mutations in the PTEN gene, located on chromosome 10q23.3 (Mawrin et al 2002). PTEN expression was noted in 70% of the tumors and was more noticeable in the Antoni A regions. Mutations of PTEN were not detected using PCR and strand conformation polymorphism screening. Data from Thomas and colleagues revealed a 25% rate of loss of heterozygosity at the Rb gene locus, which can have a variable effect on Rb mRNA and protein levels (Thomas et al 2005). In this series of patients, Rb expression was usually increased 2- to 5-fold in comparison to controls. Levels of phospho-Rb were also frequently increased and may have an antiapoptotic function.

An analysis of the expression of the ErbB-1 (EGFR) and ErbB-2 receptor tyrosine kinases has been reported (Wickremesekera et al 2007). Using immunohistochemical and western blotting techniques, it was noted that ErbB-2 had increased expression in vestibular schwannomas. However, the expression was less than that documented in glioblastoma multiforme. Expression of ErbB-1 was minimal in these tumors. The lack of ErbB-1/EGFR expression has been corroborated by other investigators (Prayson et al 2007). These data are in contrast to data from other researchers who have documented upregulation of ErbB-1/EGFR in sporadic and NF2-related vestibular schwannomas (Doherty et al 2008). In their series, ErbB-1/EGFR was upregulated in 62% of sporadic tumors. Further work by Blair and colleagues demonstrated EGFR and bFGF expression in archived vestibular schwannoma tissues (Blair et al 2011). Using invasion assays, it was determined that more invasive behavior was related to activation of EGFR and bFGF as well as downstream factors such as phospho-Akt and phospho-Erk. Recently, the platelet-derived growth factor receptor-beta (PDGFR-beta) has been shown by Ammoun and colleagues to be highly overexpressed in schwannoma cells. In addition, there appears to be activation of the Raf/Erk/Akt pathway in these tumor cells. Inhibitors of these pathways were able to inhibit schwannoma cell proliferation (Ammoun et al 2008). Micro-RNAs are small noncoding RNA molecules that regulate gene expression through post-transcriptional control of mRNA concentration. It is now known that miR-21 can regulate the PI3-K/Akt signaling pathway. Cioffi and colleagues evaluated the role of miR-21 in primary human vestibular schwannoma cultures and noted consistent overexpression of miR-21 in comparison to normal vestibular nerve tissue (Cioffi et al 2010). Elevated levels of miR-21 correlated with reduced levels of PTEN. When an anti-miR-21 was transfected into the schwannoma cultures, cell proliferation was inhibited, and the frequency of apoptosis was increased. Using high-throughput miRNA expression profiling in a series of human vestibular schwannomas, Erkan and colleagues found evidence for involvement of miR-7 (Erkan et al 2011). In these tumors, miR-7 appeared to be functioning as a tumor suppressor gene, mainly affecting the activity of the EGFR, Pak1, and Ack1 oncogenic pathways. A recent report by Yi and colleagues evaluated the expression of stem cell genes (eg, Oct-4, Nanog) in vestibular schwannoma archived tissue, cell lines, and resection specimens, and how they respond to activation of PDGFR and EGFR (Yi et al 2012). Activation of both PDGFR and EGFR resulted in an increase in expression of the stem cell genes Oct-4 and Nanog, as well as an increase in tumorsphere-forming ability.

An analysis into the angiogenic mechanisms underlying schwannoma growth has focused on the expression of vascular endothelial growth factor and its receptor (VEGFR-1) (Cayé-Thomasen et al 2005). In a series of 27 patients, tumor tissue concentrations of VEGF and VEGFR-1 were analyzed using ELISA methods and correlated with clinical and neuroimaging parameters. The concentration of both vascular endothelial growth factor and VEGFR-1 correlated with tumor growth rate (p<0.001) but not with tumor size or symptom duration. Other authors have also noted high expression of VEGF and VEGFR-1 in schwannomas (Koutsimpelas et al 2007; Uesaka et al 2007). In both reports, high levels of VEGF expression correlated with higher tumor volumes. In addition, the report by Uesaka and colleagues noted that the expression of VEGF and VEGFR-1 were more prominent in recurrent tumors in comparison to primary tumors. Koutsimpelas and colleagues also analyzed the expression of basic-fibroblast growth factor (bFGF) and found it to be elevated in schwannomas and correlated with tumor volume and the microvessel density of tumors. Immunohistochemical staining for VEGF has also recently been shown to be relevant for malignant peripheral nerve sheath tumors. A higher degree of staining for VEGF was noted for these tumors in comparison to neurofibromas (p = 0.004) and schwannomas (p < 0.001). The high VEGF expression was positively correlated with poor survival in these patients (p = 0.015) (Wasa et al 2008). In a more recent analysis of the VEGF/VEGFR axis in a series of 182 sporadic vestibular schwannomas, Koutsimpelas and colleagues noted expression of the VEGF and its receptors in all tumor samples (Koutsimpelas et al 2012). There were significantly higher levels of expression of VEGF in the cohort of recurrent tumors (p = 0.038), as well as in the subgroup of preoperatively irradiated tumors (p = 0.02).

Analysis of cell cycle-related genes has also recently been investigated by several authors. Neff and colleagues evaluated the expression of cyclins D1 and D3 in a series of 15 sporadic vestibular schwannomas (Neff et al 2006). Cyclin D1 was not expressed in any of the specimens, whereas cyclin D3 was expressed in 7 of 15 of the tumors. A similar analysis of cell cycle and apoptosis gene expression was performed in aggressive and typical vestibular schwannomas (Seol et al 2005). The p27 expression was reduced in 67% of aggressive tumors and only 20% of typical cases. In an analysis of 21 vestibular schwannomas, Lassaletta and colleagues noted conflicting results, with the presence of cyclin D1 expression in 52% of cases (Lassaletta et al 2007). Cyclin D1 positive tumors was more likely in tumors with nuclear degenerative changes (p<0.0001). Other genes (such as p21, Bax, Bcl-2, and Fas) had similar expression between aggressive and typical schwannomas.

The molecular basis underlying sporadic neurofibromas remains largely unexplored because virtually all tumors analyzed to date have been obtained from patients with neurofibromatosis type 1. In 1 report, a family was described with dysplastic nevus syndrome that had hereditary melanoma and neurofibroma tumors (Petronzelli et al 2001). Family members had germline splicing mutations associated with the CDKN2A locus, affecting both p16ink4 and p14arf mRNA processing. Loss of p16ink4 and p14arf function leads to increased activity of CDK4 and CDK6, more active phosphorylation of Rb, and loss of cell cycle control. A recent study by Koutsimpelas and colleagues reviewed the results of comparative genomic hybridization in a series of 20 cases of sporadic vestibular schwannoma (Koutsimpelas et al 2011). The most common loss was chromosome 22q, with additional losses of 9p. Genomic gains were noted on 17q, 19p, and 19q, as well as 16p and 16q. The authors suggested that these results were consistent with the role of other oncogenes and tumor suppressor genes in the genesis of vestibular schwannomas, in addition to NF2. Another study focusing on sporadic vestibular schwannomas evaluated the role of matrix metalloproteinase-2 (MMP-2), matrix metalloproteinase-9 (MMP-9), and tissue inhibitors of metalloproteinase-1 (TIMP-1) (Moller et al 2010). Using immunoenzymatic assays on resected samples from 12 patients, it was noted that MMP-9, MMP-2, and TIMP-1 were expressed in all tumors. The tumor concentration of MMP-9 correlated with absolute tumor growth rates, but not with any clinical parameters. Levels of MMP-2 and TIMP-1 did not correlate with any biological or clinical parameters. In a recent study specific to sporadic vestibular schwannomas, Lassaletta and colleagues evaluated a cohort of 51 patients and correlated the presence of NF-2 gene mutations with clinical parameters (Lassaletta et al 2013). An NF-2 mutation was noted in 49% of the cohort, whereas 22q chromosomal loss of heterozygosity was present in 57%, and amplification of the gene was noted in 13.7%. One mutational hit of NF-2 was present in 27% of tumors, whereas 2 mutational hits were noted in 45%. The presence of an NF-2 mutation was correlated with no complaint of hearing loss at the time of diagnosis (p = 0.023), as well as with the feeling of aural fullness (p = 0.029). In addition, the inactivation of the NF-2 gene was more frequent in smokers compared to those who had never smoked (p = 0.048).

Peripheral nerve sheath tumors have also been recently under investigation (Widemann 2009). Mawrin and colleagues evaluated the expression of somatostatin receptors (SST) in a series of schwannomas, neurofibromas, and malignant peripheral nerve sheath tumors (Mawrin et al 2005). The SST2A subtype was expressed often in schwannomas (89%), but not in neurofibromas (22%) or malignant nerve sheath tumors (15%). SST4 was almost exclusively expressed in malignant nerve sheath tumors (32%). Administration of an SST2A agonist was able to induce apoptosis in malignant nerve sheath tumor cells. The authors concluded that SST agonists might be useful for imaging or treatment of nerve sheath tumors. Microarray analysis of peripheral nerve sheath tumors was performed by Karube and colleagues (Karube et al 2006). Six genes were noted to be significantly upregulated in the malignant cases, including keratin 18, survivin, tenascin C, adenosine deaminase, collagen type VIa3,and collagen type VIIa1, whereas 1 gene (insulin-like growth factor binding protein 6) was downregulated. Protein immunochemistry verified the presence of increased amounts of survivin and tenascin C activity in malignant tumors.

The PI3K/Akt/mTOR signaling pathway was analyzed by Zou and colleagues in a series of 96 patients with malignant peripheral nerve sheath tumors (MPSNT) and compared to benign neurofibromas (Zou et al 2009b). The levels of p4EBP1, pS6Rp, and pAkt were all elevated in MPSNT in comparison to neurofibromas (p < 0.05), demonstrating an activation of these pathways. Although MPNST cells were sensitive to rapamycin (mTOR inhibitor), treatment resulted in enhanced pAkt and peIF4E expression. The use of PI-103 (dual PI3K/Akt/mTOR inhibitor) reduced cell growth and induced G1 cell cycle arrest, possibly through repression of cyclin D1. The same group evaluated 140 patients with MPSNT using tissue microarray and correlated these data with clinical parameters and survival (Zou et al 2009a). The survival rate at 10 years for patients with primary disease, recurrence, and systemic metastases was 31.6%, 25.9%, and 7.5%, respectively. Tumors greater than or equal to 10 cm at diagnosis, partial resection, and the presence of metastasis were all negative predictors of survival. Ki-67, vascular endothelial growth factor, p53, and pMEK were all overexpressed in MPNST in comparison to benign neurofibromas. On multivariate analysis, only tumor size and nuclear p53 expression were independent predictors of survival.

In This Article

Introduction
Historical note and nomenclature
Clinical manifestations
Clinical vignette
Etiology
Pathogenesis and pathophysiology
Epidemiology
Prevention
Differential diagnosis
Diagnostic workup
Prognosis and complications
Management
Pregnancy
Anesthesia
References cited
Contributors