A novel GIT2-BRAF fusion in pilocytic astrocytoma

PA is the most common glioma in the pediatric population [1] and it represents 5.4% of all gliomas in children and adults [2]. These tumors can arise anywhere within the neuraxis, however have a predilection for the posterior fossa, most classically the cerebellum [3]. On imaging, PA is usually well-circumscribed, cystic and solid, and demonstrates contrast enhancement [4, 5]. Histologically, PAs are composed of neoplastic astrocytes with mild-to-moderate atypia and bipolar piloid processes. Most PAs demonstrate a biphasic pattern of growth with areas of increased cellularity and a dense fibrillary background alternating with hypocellular areas with microcysts; interspersed Rosenthal fibers and/or eosinophilic granular bodies (EGBs) are seen in classic cases most typically in the cerebellum, but are not a specific diagnostic feature. Areas of infiltration, indistinguishable from those of a diffuse astrocytoma, may be present at the periphery of the tumor [5].

PAs generally have an excellent prognosis and are classified as WHO grade I neoplasms [5]. Surgical excision alone is frequently curative, with a survival rate of greater than 95% at 10-year follow-up [2, 6]. This benign clinical course underscores the importance of accurately distinguishing these tumors from other glial neoplasms. Occasionally, the histological diagnosis can be problematic because PAs can demonstrate mitoses, foci of increased proliferative labeling index by immunohistochemistry for MIB1, microvascular proliferation and sometimes geographic necrosis, all features that are also encountered in high-grade gliomas [5]. At the same time, midline glioma with H3 K27 M mutation, a WHO grade IV tumor with dismal prognosis, can present with low-grade histology, especially on biopsy specimens [7]. Most PAs have activation of mitogen activating protein kinase (MAPK) signaling through numerous described alterations, the most common being a tandem duplication at 7q34 that results in KIAA1459-BRAF fusion, therefore demonstration of a known alteration is helpful in supporting a diagnosis of PA [8‐10].

In this manuscript we describe a tumor with a novel fusion of BRAF with GIT2, hypothesized to result in activation of MAPK in a similar fashion as in PAs with the canonical KIAA1459-BRAF fusion.

H&E sections demonstrated a focally infiltrative tumor composed of mildly atypical cells (Fig. 2a) with prominent bipolar piloid processes (Fig. 2b). Areas of increased cellularity alternated with areas less cellular (Fig. 2a) and occasional multinucleated cells with eccentric nuclei (“pennies on a plate”) (Fig. 2c) were present. Some areas demonstrated oligodendroglioma-like morphology (Fig. 2d). Mitoses were difficult to find and there was no necrosis or microvascular proliferation. Unequivocal Rosenthal fibers or eosinophilic granular bodies were not identified. The tumor demonstrated extensive immunostaining with GFAP (Fig. 2e) and OLIG2 (not shown), and was negative for BRAF V600E and for IDH1 R132H mutant protein. A MIB-1 immunostain showed a proliferative labeling index of 1–2% (Fig. 2f). The tumor was provisionally signed out as low-grade astrocytoma with piloid features, W.H.O. grade I or II.

Molecular analysis using Oncopanel was performed, and the BreaKmer algorithm demonstrated a fusion involving intron 14 of GIT2 and intron 8 of BRAF, resulting in loss of BRAF exons 1–8 (Fig. 3a-c). Overall, 43 split reads were detected out of 259 total reads demonstrating a translocation at base 110,387,138 of chromosome 12 within intron 14 of GIT2, and 140,494,011 of chromosome 7 within intron 8 of BRAF. 29 discordant reads were also detected (Fig. 3d). There were no copy number changes besides low copy gain of SOX2 (3q26.33). There were no significant single nucleotide variants identified. Given the molecular test results, an integrative diagnosis of “low-grade astrocytoma with piloid feature most consistent with pilocytic astrocytoma, WHO grade I” was rendered.

PAs demonstrate MAPK signaling pathway activation through different genetic alterations. The first described and most common genetic event, KIAA1549-BRAF fusion, was described in 2008 by Bar et al. [8], who observed frequent gains at 7q34 and increased signaling in BRAF-MEK-ERK in PAs. More recent studies [9, 13, 14] describe numerous other genetic events that result in MAPK pathway activation: BRAF V600E mutations, BRAF intragenic deletions and insertions, NTRK2 fusions, FGFR1 and KRAS point mutations. Recently described BRAF fusion partners are FAM131B, RNF130, CLCN6, MKRN1, and GNAII. We report a novel inframe GIT2-BRAF fusion with breakpoints at intron 14 of GIT2 (band q24.11) and intron 8 of BRAF (band q34), with retention of exon 9 and higher in BRAF. There were no significant copy number changes observed, and no single copy gain at 7q34, therefore this fusion is the result of a deletion, and not tandem duplication; the same mechanism of MAPK activation has been described in pilocytic astrocytomas with FAM131B-BRAF fusions. Because the deleted region of BRAF is identical to the one affected in other described fusions, resulting in deletion of the inhibitory domain of the BRAF and constitutive activation, most likely this fusion activates the MAPK signaling pathway.

GIT2 encodes the ARF GTPase-activating protein GIT2, which plays a role in the regulation of cell-to-cell adhesion and cell motility. While overexpression of GIT2 has been described in the literature in breast and lung carcinomas [15, 16], this fusion with BRAF is novel in tumorigenesis and has not been described in any other type of tumor, hence it is difficult to predict the effect that the deletion will have on the GIT2 gene. Although the glioma described in this case report had focal infiltration, the histologic features were widely within the range seen in other pilocytic astrocytomas reviewed at our institution and reported in the literature; there were no specific histologic features that could be directly associated to BRAF-GIT2 fusion.

In the daily practice of neuropathology, it may be difficult to decide if the case presented is histologically WHO grade I or II due to its focally infiltrative nature. While areas of focal infiltration can be concerning for a histologic grade II neoplasm, many pediatric low-grade astrocytomas and glioneuronal tumors that are WHO grade I, including PA, have areas of infiltration. This is a feature seen in practice and mentioned in the literature, including in the revised WHO Classification of the Tumors of the Central Nervous System [5]. Such difficulty in precise histologic grading underscores the value of molecular findings in helping to predict clinical behavior. While there are infiltrative gliomas with MAPK pathway activation that may demonstrate progression over time, or have a first presentation with high-grade histology, these tumors usually have MAPK pathway activation through structural alterations in FGFR or NTRK family genes, or have pathogenic variants of NF1 or BRAF [17‐22]. Alternatively, structural rearrangements including fusions, duplications, and intragenic deletions involving the inhibitory domain of BRAF and thought to result in constitutive activation of the kinase domain, as seen in this case report, are very common in PA and possibly even specific to this neoplasm. They are associated with the increased survival expected in pilocytic astrocytomas and other WHO grade I pediatric gliomas, and therefore such a molecular finding may suggests a diagnosis of PA in the appropriate histologic context [23]. Although a body of literature suggests that a small percentage of diffuse astrocytomas, WHO grade II, have BRAF structural rearrangements, the progression to high-grade gliomas that is expected in diffuse astrocytomas is not expected in tumors with BRAF fusions [24].