Two predominant molecular subtypes of spinal meningioma: thoracic NF2-mutant tumors strongly associated with female sex, and cervical AKT1-mutant tumors originating ventral to the spinal cord

Spinal meningiomas (SM) comprise 5–10% of primary meningiomas and up to 30% of spinal intradural tumors. SMs are usually sporadic, but rarely, they can develop in association with genetic diseases like neurofibromatosis type 2 or schwannomatosis 4, 6]. While the mutational landscape of intracranial meningiomas has extensively 14], our understanding of the molecular profile of SM remains incomplete. To date, genomic studies in SMs have been underpowered to make significant conclusions about the correlations between main genomic driver alterations and clinical features of these tumors. Here, we sought to assess the mutational profile of WHO grade 1 SM and to investigate the clinical characteristics that correlate with the genomic status.

Spinal meningiomas (SM) comprise 5-10% of primary meningiomas and up to 30% of spinal intradural tumors. SMs are usually sporadic, but rarely, they can develop in association with genetic diseases like neurofibromatosis type 2 or schwannomatosis [2,4,6]. While the mutational landscape of intracranial meningiomas has been extensively studied [3,5,11,14], our understanding of the molecular profile of SM remains incomplete. To date, genomic studies in SMs have been underpowered to make significant conclusions about the correlations between main genomic driver alterations and clinical features of these tumors. Here, we sought to assess the mutational profile of WHO grade 1 SM and to investigate the clinical characteristics that correlate with the genomic status.
Targeted next-generation sequencing was performed using assays covering frequently mutated genes in meningiomas as previously described [8] and online resource. Moreover, we correlated clinical and imaging data with the molecular tumor status.
Our study cohort consisted of 50 patients with newly diagnosed SM WHO grade 1 (Table 1). Thirty-eight patients were female and 12 were male (female:male ratio of 3.2:1). The median age at diagnosis was 66 years (range 28-84 years). Cohort patients were included if they had suspected sporadic meningioma based on lack of family history of neurofibromatosis type 2, schwannomatosis, and/or lack of other meningiomas or CNS tumors. The mean follow-up time of our SM cohort was 60 months (range 6-288 months).
Our data demonstrate the primary role of NF2 and AKT1 alterations as driver events in WHO grade 1 SM, as 94% of the cases harbored mutations in these genes. AKT1 and NF2 mutations presented in a mutually exclusive pattern, identifying unique clinical cohorts. In concordance with our findings, Arslantas et al. investigated 16 SM samples and described NF2 mutations in 8 cases [1]. Remarkably, all NF2-mutant cases in the aforementioned study were located in the thoracic spine and six of them were female patients. Subsequently, Sahm et al. screened 1437 tumors for AKT1 mutations and found 65 mutant cases, including 6 AKT1-mutant cases in 57 SMs [10], and described a strong association between AKT1 E17K mutations and spinal tumor localization.
In contrast to AKT1-mutant intracranial meningiomas, where TRAF7 mutations are found to co-occur very frequently [5], we observed only a single AKT1-mutant SM case with a TRAF7 co-mutation. Combining our data with these prior reports, focusing on WHO grade 1 meningiomas only, Clark et al. found TRAF7 co-mutations in 50/68  . 1 Anatomic distribution of AKT1-and NF2-mutant meningiomas along the spinal column and in relation to the spinal cord AKT1-mutant intracranial meningiomas (15 were AKT1 "isolated", and 3 were co-mutant for AKT1/NF2), while we observed only 1 of 15 co-mutant tumors. Therefore, there appears to be a strong locational difference (p < 0.0001), suggesting that AKT1-mutant meningiomas arising from the ventral cervical spinal arachnoid are genetically distinct from their intracranial counterparts, based on the relative absence of TRAF7 co-mutation. We speculate that the absence of TRAF7 co-mutations in these tumors may potentially correlate with the more indolent clinical behavior that has been observed in AKT1-mutant SM compared to their intracranial counterparts. Furthermore, we found two SMARCB1 mutations that cooccurred in NF2-mutant meningiomas. SMARCB1 mutations have been associated with the development of SM [2,7]. Of note, our study included WHO grade 1 SMs exclusively and thus, did not include meningiomas with clear cell histology. This selection criterion may explain the absence of SMARCE1 mutations in our study, which are known driver events in clear cell meningiomas [13]. Nevertheless, the presence of cryptic inactivation of SMARCB1, NF2 or other genes cannot be definitely excluded in the three cases in our series that did not show known hot spot mutations in driver genes.
In summary, we have identified two predominant molecular subgroups in WHO grade 1 SM, characterized by AKT1 E17K and NF2 mutations. Both mutations are mutually exclusive and are associated with distinct patient characteristics and tumor features. AKT1-mutant meningiomas originate in the cervical spine ventrally to the spinal cord, are almost exclusively associated with meningothelial histology and exhibit no calcifications on imaging. In contrast, NF2-mutant meningiomas show strong female gender predominance, arise with a wider anatomic distribution, although most frequently in the thoracic spine dorsally to the spinal cord, and can be calcified while displaying variable histologic subtypes (Supplementary Figs. 2 and 3, online resource).