Introduction and Background

The identification of actionable mutations in cancer has heralded the development of highly effective targeted therapies that have made a dramatic impact on several cancers. One such target is BRAF, also known as v-raf murine sarcoma viral oncogene homolog B (B-raf), a proto-oncogene involved in the transduction of vital cellular signals via the mitogen-activated protein kinase (MAPK) pathway [1, 2]. BRAF aberrations cause abnormal cellular proliferation and survival, and are encountered in about 15% of all cancers [3]. These aberrations are relatively frequent in melanoma (66%), non-small-cell lung cancer (40%), colorectal cancer (12%), and hairy cell leukemia (100%) [1, 4]. Extensive genomic sequencing of brain tumors has revealed that although infrequent, BRAF aberrations do occur in these malignancies and provide an opportunity to target tumors in otherwise challenging locations that are not amenable to conventional treatments like surgery and radiation. In addition, recent insights into the mechanisms of resistance BRAF targeted treatments have allowed new approaches to be developed to enhance the activities of these agent with several clinical trials currently in progress evaluating such strategies in BRAF mutant tumors.

Structure and Biological Role of BRAF

The BRAF gene, located on chromosome 7q34 and consisting of 18 exons, encodes B-Raf, a 766 amino acid 94 kD cytoplasmic serine/threonine kinase—a member of the rapidly accelerated fibrosarcoma (Raf) kinase family. There are three isoforms in this family including B-Raf, C-Raf/Raf-1, and A-Raf which share common structural features, including three conserved regions (CR), CR1, CR2, and CR3, which are important for interaction of these proteins with their partners or encode their enzymatic function: CR1 is comprised a Ras-binding domain (RBD) and cysteine-rich domain (CRD) relevant to the Raf autoinhibition and has high homology across the isoforms, CR2 encompasses a conserved Akt phosphorylation site and is involved in negative regulation of Ras/raf activity, and CR3 consists of a catalytic domain which can be regulated by phosphorylation for kinase activation. Other than these three conserved regions, the Raf isoforms show little identity to each other and have different levels of ability to phosphorylate/activate MEK1/2 with the most active being B-Raf followed by C-Raf/Raf-1 and lastly A-Raf [5, 6]. The Raf kinases are components of the RAS/RAF/MEK/ERK/mitogen-activated protein kinase (MAPK) pathway [2], which is a critical signal transduction pathway that regulates several cellular functions such as proliferation, survival, differentiation, and senescence in response to activation of transmembrane tyrosine kinase (Fig. 1, left panel) [1, 2, 7, 8].

Fig. 1
figure 1

Reproduced from Zaman et al. Cancers 2019, 11(8), 1197, without modification under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/)

Illustration of MAPK pathway and mechanism of BRAF aberrations: in BRAF-driven cancers, mutant BRAF (BRAF*) can either act RAS independently as a monomer (class 1) and as a dimer (class 2) or act RAS dependently (class 3) to hyperactivate cellular growth.

Full size image

Categories of BRAF Aberrations

In human cancers, approximately 200 different BRAF mutant alleles have been identified. BRAF gene aberrations include activating mutations or duplications/fusions that lead to constitutive kinase activity causing the downstream cellular effects such as proliferation and tumor growth [9]. Based on the evidence from BRAF-mutated cancers, in particular melanoma, thyroid, and colon cancers, recent studies have proposed three classes of BRAF mutations [10•, 11]. These classes are based upon the ability of mutant BRAF to transduce signal as monomers or dimers and their dependency on RAS and kinase activity. Understanding the mechanism of these mutant classes has implications for predicting therapeutic response from targeted therapies.

Class I Mutations

These mutations are independent and do not require an upstream RAS activation or dimerization [11]. Point mutations leading to exchange of valine to a different amino acid comprise this class. Most commonly described mutation is c.1799T>A leading to substitution of valine by glutamic acid at position 600 (p.V600E) (Fig. 1, right panel). This leads to conformational change that permits BRAF monomers to an active configuration by releasing the auto inhibitory domain, and ultimately allowing activation of downstream effectors (MEK1/2) without the need for dimerization [11, 12].

Class II Mutations

This category comprises mutation of codons other than V600E with high or intermediate BRAF kinase activity undergoing constitutive, RAS-independent dimerization [10•]. BRAF fusion mutants, other point mutations, and deletions comprise this class (Fig. 1, right panel). A common fusion encountered is KIAA1549-BRAF. In this mutation, the N-terminal dimerization domain of KIAA fuses with C-terminal kinase domain of BRAF leading to loss of the regulatory domain of BRAF that promotes increase affinity for dimerization and independent BRAF kinase activity [13].

Class III Mutations

These mutations require an upstream input and are dependent on RAS for their activity (Fig. 1, right panel). These either have impaired or at times absent kinase activity. The mutant protein causes increased activity by binding tightly to activate RAS compared to the wild-type BRAF and upon dimerization leads to increased activation of the wild-type binding partner [14, 15].

BRAF Aberrations in Brain Tumors

Analysis of data from a large cohort (n = 1320) of adult and pediatric brain tumor patients demonstrated that BRAF aberrations are seen, in order of frequency, in pleomorphic xanthoastrocytoma (PXA) (66%), PXA with anaplasia (65%), ganglioglioma (18%), and pilocytic astrocytoma (PA) (9%) and less frequently in other glial tumors (< 3%) [16]. Another report analyzing a cohort of 969 patients with various types of brain tumors identified 36 cases (4%) with immunohistochemically positive BRAF V600E mutation [17]. Approximately 10% of PA harbor a BRAF V600E mutation (class I) and over 60–70% have KIAA 1549-BRAF fusions (class II) [18,19,20]; mutations and fusion aberrations are noted to be mutually exclusive in PA. PAs with a BRAFV600E mutations typically arise in a non-cerebellar location with a recent systematic review reporting presence of BRAF mutations within exon 18 in 158 cases: the frequency of BRAF mutations in different types of tumors included PA (24/187; 13%), ganglioglioma (54/115; 47%), PXA (49/81; 60%), PXA with anaplasia (9/24; 37%), and epithelioid GBM (22/38; 58%). On the other hand, the KIAA 1549-BRAF fusion was reported in 139 cases, with the following frequency: in PA (120/191; 63%), pilomyxoid astrocytoma (10/13; 77%) and ganglioglioma (9/49; 18%) [21]. Brastianos et al. identified BRAFV600E mutation in 95% all the papillary craniopharyngiomas (CP) in a cohort of 39 patients [22••]. Another report of 73 CP patients identified BRAFV600E mutations in 24.6% and all being papillary CPs [23]. BRAF mutations have been reported in 8.13% of glioma patients, with 6.5% of patients having the canonical BRAFV600E mutation [16, 19•]. Of the 1579 patients queried from various datasets, a BRAF aberration was seen in 116 (7%) with a somatic mutation reported in 6% [24, 25] (data from lower grade glioma: TCGA, PanCancer Atlas; low-grade gliomas: UCSF, 2014; glioma: MSKCC, 2018; glioblastoma: TCGA, PanCancer Atlas; anaplastic oligodendroglioma and anaplastic oligoastrocytoma: MSKCC, 2017; medulloblastoma: Broad Institute, 2012; medulloblastoma: ICGC, 2012; medulloblastoma: PCGP, 2012; medulloblastoma: Sick Kids, 2016; pilocytic astrocytoma: ICGC, 2013) on cBioPortal (Fig. 2).

Fig. 2
figure 2

BRAF aberrations in human gliomas (cBioPortal)

Full size image

BRAF Targeted Therapies in the Brain

Targeted approaches include selective inhibition of the BRAFV600E mutant, which has become the standard treatment of BRAF mutant melanoma, NSCLC, Erdheim–Chester disease, among others [26,27,28]. There are three FDA-approved drugs in this class: vemurafenib, dabrafenib, and encorafenib. Since reactivation of ERK signaling is a common mechanism of resistance to RAF inhibitors, RAF inhibitors have been combined with MEK inhibitors in patients with BRAF V600E-mutated tumors. FDA-approved MEK inhibitors include cobimetinib, trametinib, and binimetinib. MEK inhibition alone, on the other hand, is used for tumors with BRAF duplication/fusion [29].

Outcomes/Response Associated with the Targeted Therapies in Brain Tumors

Adult Glioma

VE-BASKET is a phase 2, non-randomized histology-agnostic study that included patients with BRAFV600 mutation-positive tumor types [30•]. The study included seven cohorts of patients with NSCLC, ovarian, colorectal, breast cancers, cholangiocarcinoma, multiple myeloma, and “other tumors.” Patients with brain tumors were required to have histologically confirmed glioma (any grade) and confirmation of BRAF V600E mutation. Patients received vemurafenib 960 mg twice per day continuously in 4-week cycles. Twenty-four patients with gliomas were enrolled: 11 with malignant diffuse glioma (6 with GBM and 5 with anaplastic astrocytoma), 7 with PXA, 2 with pilocytic astrocytoma, 3 with anaplastic ganglioglioma, and 1 with a high-grade glioma, not otherwise specified. One complete response (CR) was observed in a patient with PXA, and five patients achieved partial response (PR): two with PXA, one with anaplastic astrocytoma, one with anaplastic ganglioglioma, and one with pilocytic astrocytoma. The overall response rate in the group was 25% (95% CI 10% to 47%). CR lasted 25.9 months or more (censored at last assessment), and PRs lasted 13.1, 9.9, 7.5, 3.4, and 2.4 months. One patient achieved a PR in the diffuse malignant glioma subgroup and five patients had stable disease.

Wen et al. presented preliminary CNS results of a phase 2, open-label trial of dabrafenib and trametinib in patients with rare tumor types harboring the BRAF V600E mutation [31]. For the high-grade glioma cohort, eligible patients had histologically confirmed recurrent or progressive WHO grade 3 or 4 glioma (including PXA) and had prior treatment with radiotherapy and chemotherapy. The primary endpoint was investigator-assessed overall response rate (ORR) by RANO criteria. An abstract was presented with data on 31 of 37 patients. ORR was 26% (8/31; 95% CI 12–45%), including 1 complete response (CR). Five of 8 responding patients had a duration-of-response ≥ 12 months.

Gangliogliomas

In addition to the VE-BASKET trial summarized above that included three patients with anaplastic gangliogliomas, one of whom achieved a partial response, the evidence of activity of BRAF inhibitors in gangliogliomas is limited to case reports. One review of the literature estimated a complete response in 15% (3/20) and partial response in 50% (10/20) of reported pediatric and adult cases with ganglioglioma/anaplastic ganglioglioma at a median of 3.2 months after starting treatment and an estimated progression-free survival of 14 months. Some of those cases were treated with BRAF monotherapy, while others were treated with BRAF/MEK dual inhibition or a BRAF inhibitor and chemotherapy [32].

Pediatric Low-Grade Astrocytoma

In an abstract presented in 2016, Kieran et al. reported outcomes of treatment with dabrafenib in 32 pediatric patients with relapsed or refractory low-grade glioma [33]. Investigator confirmed overall response was 1 CR and 22 PRs (ORR 72% [95% CI 53–86%]). There were 13 patients with stable disease of ≥ 6 months. The Pediatric Brain Tumor Consortium then reported the results of a multi-center phase 2 study using the MEK1/2 inhibitor selumetinib in 25 patients with recurrent or progressive BRAF-aberrant or NF-1-associated low-grade glioma [34•]. Selumetinib was provided orally at 25 mg/m2 twice daily in 4-week cycles. Nine of 25 patients (36%, [95% CI 18–57]) with pilocytic astrocytomas with BRAF fusion or V600E mutation achieved a partial response with a median follow-up of 36.4 months. Ten of 25 patients (40%, 95% CI [21–61]) with NF-1-associated low-grade glioma achieved a partial response with a median follow-up of 48.6 months. Drobysheva et al. reported two cases of disseminated pilocytic astrocytoma with BRAF V600E and BRAF V600D mutations, treated with dabrafenib or dabrafenib and trametinib, respectively [35]. The former patient had resolution of her leptomeningeal disease after 3 months of therapy. The latter patient had stable disease after 11 months of the combination therapy.

Brain Metastases

The BREAK-MB phase 2 study first examined dabrafenib in 74 and 65 patients with BRAF-V600E mutant melanoma and brain metastases that were treatment naïve or previously treated, respectively. Twenty-nine (39%) and 20 (30%) achieved an overall intracranial response [36]. A pilot study then reported on 19 patients with previously treated symptomatic melanoma brain metastases treated with vemurafenib (960 mg twice daily): 7 patients (37%) achieved > 30% intracranial tumor regression. Three patients (16%) achieved intracranial PR and 13 achieved intracranial stable disease [37]. COMBI-MB is a phase 2 multicenter trial that looks at four cohorts of patients with melanoma and brain metastases [38••]. Preliminary results after a median follow-up of 8.5 months showed that 44 (58%) of 76 patients in cohort A [BRAFV600E-positive, asymptomatic melanoma brain metastases, with no previous local brain therapy, and an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1] achieved an intracranial response. Intracranial response by investigator assessment was also achieved in 9 (56%) of 16 patients in cohort B (BRAFV600E-positive, asymptomatic melanoma brain metastases, with previous local brain therapy, and an ECOG performance status of 0 or 1), 7 (44%) of 16 patients in cohort C (BRAFV600D/K/R-positive, asymptomatic melanoma brain metastases, with or without previous local brain therapy, and an ECOG performance status of 0 or 1), and 10 (59%; 33–82) of 17 patients in cohort D (BRAFV600D/E/K/R-positive, symptomatic melanoma brain metastases, with or without previous local brain therapy, and an ECOG performance status of 0, 1, or 2). There have also been reports of response to BRAF or BRAF/MEK inhibition in BRAF V600E mutant non-small-cell lung cancer metastatic to the brain (reported to occur in 1–4% of cases) [39, 40], as well as Erdheim–Chester Disease involving the brain [28].

Mechanisms of Resistance to BRAF Targeted Therapy

Mechanistically three different BRAF targeted therapy resistance patterns have been described in cancers harboring a BRAF aberration. These include intrinsic, adaptive, and acquired resistance mechanisms [41]. Presence of additional aberrations like copy number amplifications results in intrinsic resistance which results in lack of any response to initial treatment. Adaptive resistance or a partial response to the therapy can be seen due to de novo cellular epigenetic and transcription pathway alterations which may manifest as a partial and transient response with subsequent progression. Acquired resistance could result from selective pressures imposed by targeted therapy and resultant emergence of alternative clones within the tumor cell population with heterogeneous genetic alteration that can result in recurrence. It remains to be clarified whether all these resistance mechanisms are applicable to BRAF mutated brain tumor targeted treatments. In cases of PXA and PXA with anaplasia, resistance to targeted therapy with small molecule inhibitors like vemurafenib has been described. The typical BRAF mutation (class I) is often associated with sensitivity to targeted therapies; however, structural variations in the kinase domain such as β3-αC deletions and non-canonical BRAF mutations have been associated with resistant to such therapies [42, 43••]. For example, Wang et al. described that the emergence of secondary mutations in BRAF in addition to the canonical BRAFV600E mutation can confer resistance to dabrafenib in certain brain tumors (ganglioglioma) [44]. In this instance, the tumor showed initial response to dabrafenib but at progression, whole exome sequencing led to the identification of a new BRAF L514V mutation in addition to the initial V600E mutation that was not evident in the pretreatment tumor. The L514 residue is situated in the αC-β4-loop of the BRAF kinase domain. The BRAF kinase domain has a αC-β4-loop within the dimer interface that is crucial for the movements between α C-helix-in and α C-helix-out conformations [45]. This change represents an acquired resistance to targeted RAF inhibitor by inducing ERK signaling, promoting RAF dimer formation and thus promoting tumor growth. Resistance to RAF inhibitors has also been described in a cohort of PA cases due a novel recurrent BRAF insertion (p.V504_R506dup) [46].

Clinical Trials Currently Investigating BRAF Targeted Therapies for Brain Tumors

Presently, there are six clinical trials registered on ClinicalTrials.gov that are either active or in the recruiting stage evaluating BRAF targeted therapies in various primary and secondary brain tumors (ref: www.clinicaltrials.gov) summarized below (Table 1).

Table 1 Clinical trials against brain metastases active as of April 2021 in Clincaltrials.gov
Full size table

Summary and Future Directions

Certain subtypes of brain tumors, such as gangliogliomas and craniopharyngiomas, can be genetically simple and driven only by a BRAF V600E alteration. Higher grade gliomas, on the other hand, constitute a genetically complex disease with the hallmark of heterogeneity. Targeted therapy has had better success in simpler single-mutation-driven tumors, and more thoughtful data-driven combination approaches are needed for the more complex mutation-bearing tumors. Regardless, in light of the aggressiveness and lack of therapeutic options for high-grade gliomas, the presence of BRAF alterations does provide a unique potential for targeted therapy that has otherwise not been generally successful in gliomas. Due to the relative rarity of these events, multicenter collaborative studies will be essential in accruing sufficient numbers of patients and in assessing the potential for combination therapies in the setting of BRAF alterations in these tumors. Similarly, basket trials that are tumor agnostic have created an opportunity to study cancer patients with such rare mutations. Future basket trials and longitudinal natural history studies will be essential for better understanding of the prognosis and therapeutic implications of BRAF-altered gliomas in the pediatric and adult populations.