**Abstract**

Atypical (WHO grade II) and malignant meningiomas (WHO Grade III) are a rare subset of primary intracranial tumors. Due to the high recurrence rate after surgical resection and radiotherapy, there has been a recent interest in exploring other systemic treatment options for these refractory tumors. Recent advances in molecular sequencing of tumors have elucidated new pathways and drug targets currently being studied. This article provides a thorough overview of novel investigational therapeutics, including targeted therapy, immunotherapy, and new technological modalities for atypical and malignant meningiomas. There is encouraging preclinical evidence regarding the efficacy of the emerging treatments discussed in this chapter. Several clinical trials are currently recruiting patients to translate targeted molecular therapy for recurrent and high-grade meningiomas.

**Keywords:** targeted therapy, molecular biology, progression free survival, overall survival, meningioma, genomics, angiogenesis, immunotherapy, outcomes

## **1. Introduction**

Meningiomas (MN) are a type of central nervous system (CNS) tumors that arise from the leptomeningeal arachnoid covering the encephalon and the spinal cord, more specifically, from the arachnoid cap cells [1]. In adults, MN accounts for approximately 37.6% of all primary brain tumors, and corresponds to the most common intracranial tumor in adults over 35 years [1, 2]. According to Ostrom et al., incidence of MN in the United States (US) is 8.83 per 100,000 per year [3]. Around 90% of all MN cases are diagnosed intracranially, with the rest arising from the spinal arachnoid [4]. The median age at diagnosis for MN is 65 years [4] with the majority of patients being in the range of 55–74 [4]. Cases in the pediatric population are extremely rare, corresponding only to 0.4–4.6% of all pediatric tumors [2]. There is a female predominance in case proportion, with a female:male ratio of 3:1 for all MN, and 9:1 for spinal cord MNs [2, 5]. MNs are characterized for being slow in growth and often not infiltrative, with an insidious development of symptoms. Clinical presentation of MN might vary from patient to patient, with tumor localization being the main determining factor of clinical features. Signs and symptoms might include headaches because of increased intracranial pressure, focal neurological deficits (mainly cranial nerve focalization), and seizures. In the case of MN developing in the frontal lobe, personality changes, altered mental status and mood disturbances might appear [6].

According to the World Health Organization (WHO), MN is classified in three subtypes: common type or WHO grade I, atypical/intermediate type or WHO grade II and the anaplastic/malignant type or WHO grade III. These high-grade tumors might develop *de novo* or as a transformation from a lower grade MN [7]. Approximately 70% of cases are WHO grade I, 28% are WHO grade II and only around 3% are classified as WHO grade III. According to a cohort of 992 patients with MN, the proportion of atypical and anaplastic MN was higher in males than females (*p* = 0.003) [4]. The more aggressive behavior in grade II and III MN is represented by a worse prognosis in terms of overall survival (OS) and recurrence risk after surgical resection (SR). In a cohort of 102 patients with grade II and III MN, 5-year OS (5-yOS) was 97.5% and 67.4% respectively, with a median OS (mOS) of 167 months and 72 months respectively [8]. These results showed a marked increase in survival over the last decades, arguably because of the introduction of better surgical techniques, radiation therapy and some forms of chemotherapy, as previous research showed a 5-yOS of 75% for grade II MN and 32% for grade III MN [9]. Tumor recurrence has been found to be considerably increased in high grade MN, with a 50% and 80% 5-year recurrence for grade II and grade III MN respectively, and only 5–10% for grade I MN [10, 11].

As high-grade MN continue to be a difficult to treat condition, with high recurrence and low response rates, molecular insights into precision medicine have been investigated in the last two decades. With a better understanding of the cellular and molecular pathways underlying MN pathophysiology, recurrence and malignancy, newer therapies have been considered as possible candidates for the treatment of these conditions. Some agents include newer systemic chemotherapeutic agents like trabectedin, inhibitors of the Epidermal Growth Factor Receptor (EGFR) like erlotinib and gefitinib, inhibitors of the Platelet-Derived Growth Factor Receptor (PDGFR), inhibitors of mTOR, especially from the complex 1 (mTORC1) as well as its upstream and downstream elements (AKT/PI3K and MEK). The biological process of angiogenesis is also under research, with ongoing trials with antiangiogenic agents from the Tyrosine Kinase Inhibitors (TKIs) targeting the Vascular Endothelial Growth Factor (VEGF) pathway, as well as antibody agents like bevacizumab. As it is expected, immunotherapy with checkpoint inhibitors is also under current investigation, with anti-PD1 and anti-PD-L1 monoclonal antibodies being tested in clinical trials. In this chapter we are going to cover the molecular biology of MNs, especially in the cases of grade II and grade III MN. We will also discuss the current knowledge in systemic treatments as well as therapies in clinical trials and possible candidates that are being tested *in vitro*.

#### **2. Molecular biology**

Advancements in understanding the pathophysiology and molecular biology of MNs are critical for improving risk evaluation and prognosis. Similarly, to design novel treatments aimed at blocking canonical pathways involved in carcinogenesis and disease evolution. As molecular analyzes of meningiomas continue to evolve, several cytogenetic, genomic, epigenetic, and expression alterations associated with tumor aggressiveness and proclivity for recurrence have been identified as potential biomarkers to enhance risk stratification [12]. Recently, several seminal studies evaluating the genomics of intracranial meningiomas have rapidly changed the understanding of the disease. The importance of NF2 (neurofibromin 2), TRAF7

*High Grade Meningiomas: Current Therapy Based on Tumor Biology DOI: http://dx.doi.org/10.5772/intechopen.100432*

#### **Figure 1.**

*Main cytogenetic and recurrent genetic alterations in recurrent and high-grade meningiomas according to the WHO classification and anatomical location.*

(tumor necrosis factor [TNF] receptor-associated factor 7), KLF4 (Kruppel-like factor-4), AKT1, SMO (smoothened), PIK3CA (phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha), and POLR2 (RNA polymerase II subunit A) demonstrates that there are at least six distinct mutational classes of meningiomas. In addition, six methylation classes of meningioma have been appreciated, enabling improved prognosis prediction compared with traditional WHO grades. Genomic studies have shed light on the nature of recurrent meningioma, distinct intracranial locations and mutational patterns, and a potential embryonic cancer stem cell-like origin [13–16] (**Figure 1**).

#### **2.1 Cytogenetics and genomics**

A large number of meningiomas possess a normal karyotype, with an overall low incidence of genomic alterations (including somatic copy number alterations— SCNA, rearrangements, and low mutational burden) [17–19]. However, these disruptions increase following tumor grade, the number of recurrences, and biological aggressiveness. More than half of all identified genomic alterations involve the NF2, which underlies inherited Neurofibromatosis syndrome. Indeed, the most significant SCNA in meningioma is chromosome 22 monosomy, which is present in ~56% of cases and leads to losing the genomic locus containing NF2 (22q12.2) [20, 21]. Among grade I meningiomas, those carrying NF2 alterations are more likely to progress than those with a normal karyotype. In addition, the frequency of NF2 aberrations increases with tumor grade.

Loss of heterozygosity on chromosome 1p is present in 16% of MNs [22]. Characterization of the smallest region of overlapping deletion on this chromosome spans ~3.7 megabases and identified 59 genes, 17 of which have putative tumorsuppressive functions based on gene ontology. The protein methyltransferase and tumor suppressor RIZ1, is located on chromosome 1p, and studies implicate its loss of expression in meningioma progression [23]. Loss of the CDKN2A/CDNK2B locus on chromosome 9q is common in grade II meningiomas that transition to anaplastic lesions [24]. Additionally, a study showed that the levels of p16 and p15, the proteins encoded by CDKN2A and CDKN2B, may hold prognostic significance and/ or represent a promising therapeutic target [25]. Recently, Nassiri et al. described

four consensus molecular groups of MN by combining DNA somatic copy-number aberrations, DNA somatic point mutations, DNA methylation, and messenger RNA abundance in a unified analysis [26]. These molecular groups predicted clinical outcomes compared with existing classification schemes. Each molecular group showed distinctive and prototypical biology (immunogenic, benign NF2 wild-type, hypermetabolic and proliferative) that informed therapeutic options. Proteogenomic characterization reinforced the robustness of defined molecular groups and uncovered highly abundant and group-specific protein targets [26].

### **2.2 NF2-related meningiomas**

Globally, meningiomas have a low mutation rate (~3.5 mutations per megabase) compared to other cancers [25]. Various efforts to genotype the disease using NGS have identified NF2 mutations as the predominant alteration in spontaneous and Neurofibromatosis syndrome-associated tumors [24], at a frequency of ~40% in low grade and nearly 80% in high-grade tumors [27]. MNs related to alterations in NF2 were more common in the cerebral convexities and posterior skull base than those found in other anatomic locations, and up to 13% were associated with other co-mutations, including single mutations in CREBBP, PIK3CA (R108H), PIK3R1, BRCA1, and SMARCB1 [27]. Unfortunately, within NF2 mutated meningiomas, none of these identified mutations can predict the chance of recurrence, which can vary widely.

TERT promoter mutations have recently been reported in ~6% of all MNs, with ~80% of these also harboring alterations (mutations or deletions) at the NF2 locus [28]. Similar to overall mutational burden, TERT mutations increase with tumor grade. In grade I MN, TERT C228T and C250T mutations are linked with transformation to higher grades [28], prompting many neuro-oncologists to consider standardized testing for TERT promoter mutations. Further studies demonstrate that the presence of C228T and C250T correlates with increased TERT mRNA and functional increases in telomerase activity [29]. In grade II or III tumors, univariate analysis revealed a significant association with decreased PFS (progression-free survival; median 12.5 vs. 26 months, *p* = 0.004) and OS (overall survival; mean 26 vs. 46 months, *p* = 0.009) [30]. In vitro studies demonstrated that TERT mutated meningioma cells show decreased TERT activity in response to YK-4-279, a small molecule inhibitor of ETS transcription factor, suggesting a novel potential strategy for targeting this subgroup of tumors. In addition to individual TERT promoter mutations, recent efforts using targeted sequencing approaches identified an additional TERT promoter in the known hotspot G124A, which like other TERT mutations, seems to correlate with poor prognosis [31].

#### **2.3 Non-NF2 meningioma**

Non-NF2 mutated meningiomas, which generally have a benign behavior, are usually chromosomally stable, and often located in the anterior, medial, or skull base regions, possess a distinct mutational landscape [27]. Recent high throughput sequencing studies suggest an average of only 1.56 (SD ± 1.07) genomic alterations (GAs) per non-NF2 mutated tumor [31]. The pro-apoptotic E3 ubiquitin ligase, tumor necrosis factor receptor-associated factor 7 (TRAF7) is mutated ~25% of all meningiomas [31]. Such alterations occur in the C-terminal WD40 protein interaction domain, suggesting they may alter protein-protein interactions with MAPK and NF-kB family members [32]. While TRAF7 mutation is mutually exclusive with NF2 mutations, it is almost always correlated with PI3K and activating E17K mutation in AKT1, with the K409Q alteration of KLF4 [33].

AKT1, also referred to as protein kinase B, is a well-known oncogene. AKT activation relies on the PI3K pathway and is recognized as a critical node in the mTOR pathway. The E17 hotspot is the most characterized of AKT1 mutations and leads to constitutive activation of the protein. Mutations in AKT1 have also been shown to confer resistance to allosteric kinase inhibitors in vitro and are oncogenic in many solid tumors. Specifically, the E17K mutation is found in 7–12% of grade I meningiomas [34], is enriched in the meningothelial subtype [17], and is predictive of decreased PFS in olfactory groove tumors [35]. Altering the same signaling pathway PIK3CA mutations are also found in ~7% of non-NF2 tumors and are mutually exclusive with AKT1 mutation [36]. Targeted sequencing of this gene revealed novel non-synonymous mutations, A3140T and A3140G, which are reported as pathogenic, and C112T, which is also predicted to be pathogenic [31]. Indeed, increased PI3K signaling is related to aggressive behavior, especially within high-grade meningiomas [37], suggesting that therapeutics targeted toward this pathway may be a potential option.

Sequencing of 71 meningiomas genes recently identified two novel missense mutations in FGFR3, T932C, and G1376C, both of which were predicted to be pathogenic [31]. Identifying these mutations in patients with skull base low-grade tumors was associated with a good prognosis, given the absence of recurrence and the requirement of IMRT. KLF4 gene encodes a protein that belongs to the Kruppel family of transcription factors. The encoded zinc finger protein is required to control the G1-to-S transition of the cell cycle following DNA damage by mediating the tumor suppressor gene p53. In addition, KLF4 is involved in the differentiation of epithelial cells and may also function in skin, skeletal, and kidney development [38]. In meningiomas, KLF4 is thought to act as a tumor suppressor gene, expressed in low-grade tumors and downregulated in anaplastic tumors. At the genomic level, KLF4 is mutated in ~12% of grade I meningiomas, virtually all of which are of the secretory sub-type and harbor TRAF7 mutations [39]. All identified KLF4 mutations result in a K409Q substitution within the DNA binding domain, which likely alters several protein functions [40].

SMO (Smoothened, Frizzled Class Receptor) gene encoded a G protein-coupled receptor that interacts with the patched protein, a receptor for hedgehog proteins. Mutations in SMO, which result in L412F or W535L substitutions, lead to functional activation of Hedgehog signaling in meningioma [17, 41]. These mutations are present in ~5.5% of grade I meningiomas and are mutually exclusive with TRAF7, KLF4, and AKT1 mutations [27]. Meningiomas with the L412F mutation are more likely to recur (XX) and are enriched at the midline, perhaps due to the role that Hedgehog signaling plays in hemisphere separation during development [36]. Mutations in the Hedgehog family member SUFU are also found at low frequencies in sporadic meningiomas, and their germinal counterpart is also present in familial meningiomatosis [42]. Additional hedgehog family germline mutations occur in SMARCE1 and SMARCB1, though these carry less risk of recurrence than familial NF2 mutations [43, 44].

POLR2A (RNA Polymerase II Subunit A) catalyzes the transcription of DNA into RNA using the four ribonucleoside triphosphates as substrates. In addition, POLR2A is the largest and catalytic component of RNA polymerase II which synthesizes mRNA precursors and many functional non-coding RNAs. POLR2A encodes RPB1 (DNA-directed RNA polymerase II subunit), a gene found altered in about 6% of meningiomas [42]. From another perspective, inactivating somatic and germline mutations or gene deletions in the BAP1 tumor suppressor gene are explicitly found within high-grade rhabdoid meningioma [45]. Also, the loss of BAP1 is correlated with tumor aggressiveness and decreased time to progression. Alterations in the SWI/SNF pathway, specifically mutations in ARID1A, were

recently found in 12% of high-grade meningiomas. Other components of this canonical pathway, including SMARCB1, SMARCA4, and PBRM1, are altered in up to 15% of patients with non-NF2-dependent meningiomas [46].
