Meningiomas are found in half of NF2 patients and are frequently multiple

Neurofibromatosis type 2 (NF2, OMIM #101000) is a dominantly inherited tumor predisposition syndrome caused by inactivating mutations of the NF2 tumor suppressor gene on chromosome 22q12. NF2 is rare genetic disorder with a birth incidence of 1/33,000, and a prevalence of 1/56,161 [1]. NF2 patients are predisposed to develop multiple benign tumors including schwannomas, meningiomas and ependymomas. Half of cases harbor de novo mutation without any family history of NF2, and penetrance is almost 100% by 60 years of age [2, 3]. The cardinal feature of NF2 is the development of schwannomas on the vestibular branches of both eighth cranial nerves. Meningiomas are the second most frequent tumor type in NF2, affecting about half of NF2 patients. Spinal lesions are frequent in NF2, concerning 90% of patients [4], and include schwannomas, meningiomas, and ependymomas. Non-tumoral features of NF2 include polyneuropathy [5], skin and ophthalmic manifestations. About 70% of NF2 patients develop skin lesions: the most frequent are intracutaneous plaque-like lesions, which are slightly raised, pigmented and often hairy; more deep-seated nodular tumors with thickened nerve palpable on either side usually correspond to schwannomas [2]. Ophtalmic examination can show juvenile posterior subcapsular cataracts and epiretinal membranes, highly suggestive of NF2 [6, 7]. Manchester set of diagnostic criteria, widely used for the diagnosis of NF2 (Table 1), recognizes as NF2 a patient harboring multiple meningiomas when associated with schwannoma, glioma, neurofibroma, or cataract.

Table 1 Manchester clinical diagnostic criteria for Neurofibromatosis type 2
Full size table

The incidence of cranial meningiomas in published cohorts of NF2 patients is about 50%: from 28/74 (38%) in Japan [8], 54/120 (45%) in Manchester [2], 31/63 (49%) in NIH [3] to 28/48 (58%) in Hannover [4], the discrepancy in incidence depending in part on the screening protocols. Meningiomas in NF2 are frequently multiple and occur either in cranial (Fig. 1) or in spinal location (Fig. 2). The female predominance observed in sporadic meningioma cases is also observed in NF2 patients [9], although NF2 affects equally both sexes.

Fig. 1
figure 1

T1-weighted MRI with Gadolinium enhancement showing aspects of cranial meningiomas in NF2 patients. a right frontal convexity meningioma responsible for intra cranial hypertension (b) coronal view showing the association of a large left parasagittal meningioma and an intraventricular meningioma. Of note, a large left vestibular schwannoma is also visible (c) multiple medium sized meningiomas involving the convexity, the falx cerebri, and sagittal sinus

Full size image
Fig. 2
figure 2

MRI showing aspects of spinal meningiomas in NF2 patients. a axial T1-weighted sequence with Gadolinium enhancement showing a right anterior lateral thoracic meningioma compressing the spinal cord. b, c sagittal T2-weighted sequence (b) anterior C2 meningioma. Of note, an intra spinal cystic tumor corresponding to an associated spinal ependymoma (c) two thoracic meningiomas are shown in T5 and T9/10. Of note, multiple schwannomas are visible along the spinal roots

Full size image

In the pediatric age group, meningiomas are often the first sign of NF2 [10]. In a series of 21 NF2 patients and 18 pediatric cases, 83% of patients with early onset NF2 harbored meningiomas [11]. Furthermore, NF2 has to be ruled out in sporadic pediatric patients harboring meningioma, both multiple and single tumors: NF2 is diagnosed in 10–29% of patients in pediatric meningioma series [12–15].

Meningiomas in NF2 patients are associated with disease severity

Baser et al. [16] analyzed the factors associated with mortality in 368 patients from 261 families in the United Kingdom NF2 registry. Age at diagnosis, intracranial meningiomas, and specialization of treatment center were informative predictors of the risk of mortality. The relative risk of mortality was 2.51-fold greater in people with meningiomas compared with those without meningiomas. Thus, presence of meningiomas in NF2 patients is a marker of disease severity.

Multiple meningiomas is a rare entity largely overlapping NF2

The proportion of patients harboring multiple meningiomas in the literature ranges from 1 to 12% in surgical series, and from 8 to 16% in radiological or autopsy series. Most of the studies refer to cases identified before NIH Consensus Statement on Neurofibromatosis, and NF2 patients are difficult to segregate from other multiple meningiomas patients. Antinheimo et al. [17] analyzed the frequency of multiple meningiomas in a well-defined population of 1,713,000 people in Finland. Seven out of 823 (1%) patients with meningioma had multiple meningiomas in association with NF2, and 29 of 823 (4%) had multiple meningiomas without obvious NF2.

Somatic mosaicism is a pivotal issue in multiple meningiomas patients. Somatic mosaicism is caused by postzygotic mutations in the early stage of embryo development. Only a subpopulation of the normal cells of a mosaic subject carries the constitutional mutation. Therefore, screening of a non-tumor tissue such as leucocytes may result in failure of detection when the mutation level of that tissue is low. Somatic mosaicism is of clinical interest because it may correlate with a mild form associated with a favorable course of the disease; it may confound linkage testing of offspring, and decrease genetic risk to the next generation. Somatic mosaicism is frequent in NF2, affecting up to 33% of patients [18, 19], and may explain a mild phenotype in some patients [18, 20]. Risk of transmission to offspring is low in a somatic NF2 patient with a mutation only detectable in tumor [19]. Evans et al. found that 8% of multiple meningiomas in adult patients may be caused by mosaic NF2 [21].

In contrast to mosaicism, data support the hypothesis that it is not rare that multiple meningiomas in the same patient share the same clonal origin. Larson et al. showed inactivation of the same X chromosome in every sample from the same patient and thus concluded that multiple meningiomas could arise from the uncontrolled spread of a single progenitor cell [22]. Studies addressing only NF2 somatic mutations cannot differentiate NF2 mosaicism from clonal spread of a meningioma across the meninges. Somatic mosaicism can be proven if the same NF2 mutation is identified in two different tumor types (e.g. meningioma and schwannoma) or when the mutation found in a meningioma specimen is detected at a low level in blood sample [18].

Apart from NF2, other genes are probably involved in meningioma development. Few familial cases of multiple meningiomas have been reported with an autosomal dominant transmission. Two families with multiple meningiomas have been reported and provide evidence for a locus outside NF2 involved in multiple meningiomas families. Pulst et al. [23] reported a family with multiple meningiomas and ependymomas in two generations. By linkage analysis with DNA markers known to flank the NF2 locus, they excluded NF2 as the candidate locus for the hereditary condition. Maxwell et al. [24] described a family with two members harboring meningiomas; NF2 protein immunoreactivity was present in meningioma samples, suggesting that alternate gene might be involved in tumorigenesis. Shen et al. analyzed by array comparative genomic hybridization the genomic profile of a large series of meningiomas, have recently confirmed this hypothesis. In contrast to sporadic meningiomas, they found that familial multiple meningiomas displayed no chromosomal imbalance events, supporting a distinct mechanism for the origin for these tumors [25]. Similarly to glioblastomas [26], whole genome sequencing of familial multiple meningiomas cases will help uncover new genes involved in meningioma genesis.

The most practical way to diagnose NF2 in multiple meningiomas patients, consist of a careful clinical examination with special emphasize on audiovestibular functions, skin examination, and slit lamp examination of the lens. The family history should be carefully investigated for previous surgery or symptoms suggestive of NF2. Cranial and spinal MRI is required and should include contrast medium injection and thin slices on cerebellopontine angles assessed by an experienced neuroradiologist [27]. Identifying the causative NF2 mutation by blood or tumor DNA sequencing can definitely be helpful to diagnose NF2 or NF2 mosaicism [28, 29]. The definitive diagnosis of NF2 mosaicism requires tumor tissue, underscoring the need to bank tumors for patients who might desire genetic diagnosis in the future.

Meningiomas are rare in Neurofibromatosis type 1 (NF1) and schwannomatosis

NF1 (MIM 162200) is a dominantly inherited disorder characterized by the presence of café au lait spots, peripheral neurofibromas, plexiform neurofibromas, Lisch nodules, axillary freckling, skeletal dysplasia, and optic gliomas. The incidence of meningiomas in NF1 patients seems to be that of general population (1/523 in [30], 2/158 in [31]).

Schwannomatosis (MIM 162091) is characterized by the development of multiple spinal, peripheral, and cranial nerve schwannomas in the absence of vestibular schwannomas (which would otherwise classify patients as NF2) [32]. Although the condition can be difficult to clearly differentiate from NF2 [33], especially in younger patients, where vestibular schwannomas can develop secondarily, often in the second decade of life, few cases of schwannomatosis with associated meningiomas have been reported [34, 35]. Mutation of the SMARCB1 gene on 22q seems to be responsible for at least one part of familial schwannomatosis [36]. Recently, Bacci et al. [37] reported a family segregating a germline mutation in SMARCB1 exon 1 in four affected members: one member had multiple schwannomas, one multiple meningiomas and two with both schwannomas and meningiomas. It is therefore possible that meningiomas are an inconstant feature of SMARCB1 mutation and could be part of the schwannomatosis tumor phenotype. To rule out the implication of SMARCB1 in multiple meningiomas patients, Hadfield et al. screened the SMARCB1 gene in a panel of 47 patients with multiple meningiomas unrelated to NF2. No SMARCB1 germline mutation was found and they concluded that while meningiomas may be associated with the schwannomatosis phenotype, SMARCB1 is not a major contributor to multiple meningiomas disease [34].

Meningioangiomatosis can occur in association with NF2

Meningioangiomatosis is a rare entity characterized by a plaque-like, cerebral hemispheric mass, most often involving the temporal and/or frontal lobes. Histologically, meningioangiomatosis consists of an intracortical and leptomeningeal collection of small blood vessels with perivascular spindled cells and variable degrees of cellularity, hyalinization, calcification, and even ossification. The intervening glioneuronal parenchyma appears mature but apparently disorganized, and reactive gliosis varies [38]. The majority (75–80% of cases) occurs sporadically and seizures are the most common symptoms. Twenty to 25% of meningioangiomatosis cases occur in association with NF2; they are often multifocal, associated or not with meningioma and are typically asymptomatic, with most cases identified as incidental autopsy finding. They should be considered in the differential diagnosis of cortical lesions in NF2 patients. It might be difficult to differentiate meningioangiomatosis associated with meningioma in NF2 patients from true brain invasion, which has a more severe prognosis in WHO 2007 classification.

Meningiomas can develop throughout the meninges in NF2, but may be less frequent at skull base

Perry et al. [11] reported a series of 53 meningiomas resected in 40 pediatric and/or NF2 patients (including 21 NF2 patients). The sites of origin of meningiomas were convexity and parasagittal in 49%, intraventricular in 13%, skull base in 6%, posterior fossa 9%, orbit/anterior pathways in 8%. Antinheimo et al. [9] reported a series of 39 meningiomas resected in 23 NF2 patients, both from Finland and Hannover. The site of origin of meningiomas was convexity and parasagittal in 38%, skull base in 25%, posterior fossa in 23%, and 2% optic nerve sheath. The higher rate of posterior fossa and skull base meningiomas in this latter series could be biased by specific center recruitment.

Specific locations that seem more frequent in NF2 patients than in sporadic cases include intraventricular meningiomas, usually in close contact with choroid plexus, often in the lateral ventricles, and optic nerve sheath meningiomas [7] that can affect both optic nerves [39].

Eight [11] to 10% [9] of resected meningiomas in NF2 patients are spinal. Spinal meningiomas can be difficult to radiologically distinguish from spinal schwannomas. In the series published by Mautner et al. [4], 89% of NF2 patients harbored spinal tumors on MRI, and 37% (7/19) of intradural extramedullary tumors were meningiomas on pathological examination. These data suggest that 1/3 NF2 patients may have spinal meningiomas.

Meningioma neuropathology might be different in NF2 than in sporadic cases

Data about pathology of NF2 associated meningiomas are sparse, and no specific NF2 features have been identified so far. Perry et al. [11] reported pathological features of 30 NF2 associated meningiomas (15 pediatrics, 15 adults). Nineteen meningiomas (63%) were grade II or III, suggesting that meningiomas associated with NF2 are more aggressive than sporadic meningiomas. The main histological subtypes were transitional (60%), meningothelial (20%), fibroblastic, papillary and rhabdoid (7% each). For comparison, in sporadic meningiomas, the frequencies of histological subtypes have been described as transitional (28%), meningothelial (53%), fibroblastic (8%) [40]. Similarly, search for somatic NF2 mutations [41] or 22q loss [42] in sporadic meningiomas showed that meningothelial variants are often associated with unaltered NF2.

Antinheimo et al. [9] analyzed 35 NF2 patients’ meningiomas and compared them to 30 sporadic meningiomas, matched by age and gender. The most frequent histological subtypes in NF2 patients were transitional (53%), fibroblastic (30%), and meningothelial (17%). The high rate of fibroblastic meningiomas was also observed in the matched control group, suggesting that this difference in meningioma subtypes may reflect age-related differences rather than NF2 gene-related differences. They did not observe specific histological feature that would distinguish NF2 meningiomas from the sporadic control tumors. They observed a higher proliferation potential of the NF2 meningiomas, suggesting a more aggressive behavior of NF2 tumors. In summary, meningiomas resected in NF2 patients are more frequently of the fibroblastic subtype, and they often display aggressive features on pathological examination. However, these findings are based on the analysis of resected tumors. The majority of meningiomas in NF2 patients remain under surveillance, and these slow growing tumors are probably less histologically aggressive than tumors requiring surgery.

At the molecular level, NF2 associated meningiomas seem to share with sporadic meningiomas a similar spectrum and frequency of allelic deletions. Lamszus et al. studied 30 meningiomas from 22 NF2 patients by microsatellite analysis, and found LOH involving 22q12, 1p, 10q, 6q, 14q, 18q, and 9p in 100, 40, 27, 24, 24, 23, and 17% of the tumors respectively [43].

Treatment of meningiomas in NF2

Data about long-term natural history of meningiomas in NF2 are sparse in the literature. Surgery remains the core of the treatment of growing and/or symptomatic meningiomas in NF2. To our knowledge, no paper focused on meningioma surgery in NF2 in the literature nor described the surgical results in terms of morbidity/mortality and recurrence.

Stereotactic radiosurgery is a controversial treatment of vestibular schwannomas in NF2 [44], and is now being extensively debated. In contrast, stereotactic radiosurgery has rarely been reported for meningiomas in NF2. Kondziolka et al. reported 6 NF2 patients treated by Gamma knife for convexity meningiomas in a larger series of 125 patients. They have found that poorer overall survival was associated with NF2 [45]. Wentworth et al. published a highly controversial paper [46] reporting their experience with radiation therapy in the management of NF2-associated meningiomas. They treated 49 meningiomas in 12 NF2 patients, claming that the majority of their patients presented asymptomatic lesions that showed growth on serial imaging, located in eloquent locations in which further growth would lead to morbidity. A major concern for many physicians is that radiosurgery and radiotherapy may induce malignant transformation and adjacent tumor development in NF2 patients, particularly if used in the case of multiple asymptomatic meningiomas [reviewed in 47].

Finally, Plotkin et al. have reported recently the reduction in the volume of most growing vestibular schwannomas after Bevacizumab in a small series of 10 NF2 patients [48]. No results have been reported regarding the efficiency on meningiomas in this series. New promising drugs have been identified as candidate [49] in NF2 associated meningiomas, and have to be evaluated in clinical trials after validation in preclinical models [50, 51].

Conclusion

Meningiomas occur in about half NF2 patients, and are often multiple. NF2 and somatic mosaic NF2 should be ruled out in multiple meningiomas patients by detailed clinical and radiological examination; frozen samples should be preserved during tumor removal to search for somatic NF2 mutation. Meningiomas in NF2 patients are associated with a higher risk of mortality, and their treatment is challenging, but data about natural history of meningiomas in NF2 patients are sparse in the literature. Knowledge of tumor behavior is essential in slow growing tumors like meningiomas, to balance the risk of treatment against the natural history of the disease, and to evaluate the efficiency of alternative therapeutics (radiation therapy or new drugs). Further efforts are needed to delineate the best treatment modality and timing.