Papillary glioneuronal tumors: histological and molecular characteristics and diagnostic value of SLC44A1-PRKCA fusion

PGNT is a rare mixed glioneuronal tumor usually characterized by an indolent clinical behavior [21]. The rarity and the scarcity of tissue available for molecular and genetic analyses have hampered the identification of specific alterations. The diagnosis of PGNT is often difficult to establish both on MRI and histology. On MRI, PGNT is frequently described as a well-demarcated mass with a cystic component, contrast enhancement, isointensity on T1-weighted and a predominant location in the periventricular area [22]. However, there is no pathognomonic radiographic appearance and in the review by Carangelo and colleagues, around 80 % of PGNT showed a ventricular association [3]. All our cases show a ventricular association, one is clearly intraventricular, similar to the case reported by Matyja and colleagues [23] and the two others, the tumor have a direct contact with the temporal ventricle. The three cases present a nodular contrast enhancement. Cases 1 and 3 show a more typical appearance with a cystic component. According to the WHO 2007 classification, PGNT is histologically characterized by a distinct pseudopapillary architecture with layers of GFAP positive glial elements, lining hyalinized vessels and variable amount of interpseudopapillary spaces filling by synaptophysin-positive neurocyte-like, ganglioid, or ganglion cells [2]. However, microcystic oligodendroglial-like areas and astrocytic components could resemble to pilocytic astrocytoma or ANET. The diagnosis is thus based on a set of histological and radiological criteria. However, individual histological entities could classically overlap with other LGG entities (PGNT mimics). In particular, the distinction between PGNT and ganglioglioma is not yet completely clear as in WHO description, PGNT could contain a heterogeneously distributed ganglion cell tumoral population [2]. Bridge and colleagues identified a recurrent chromosomal translocation t(9;17)(q31;q24) in three PGNTs resulting in an in-frame fusion of SLC44A1 and PRKCA, with consequent generation of a constitutively expressed serine/threonine kinase fusion [12]. A recent study has reported the presence of the fusion in two additional PGNTs [13]. Consistent with this, our findings confirm the presence of SLC44A1-PRKCA fusion in PGNT. We took a particular attention in the group of PGNT mimics of two particular cases of ganglioglioma with pseudopapillary features (cases 5 and 6). Indeed, histological features of this entity are similar to those of PGNT. However, interestingly, the cases 5 and 6 did not show the SLC44A1-PRKCA fusion but were BRAF V600E mutation positive, as described in around 40 % of gangliogliomas [24]. Ganglioglioma is also composed of neoplastic mature ganglion cells in combination with neoplastic glial cells [25]. In our cases, a ganglion cell component was observed only in cases 5 and 6 associated with eosinophilic granular bodies. Conversely, positive cases for SLC44A1-PRKCA fusion consisted exclusively of small oligodendrocyte-like cells and neurocyte-like cells without gangliocytoid and ganglioid cell features. Furthermore, case 6 showed an extravascular expression of CD34, a marker consistently expressed in 70–80 % of gangliogliomas, especially those variants emerging from temporal lobe [26]. The neoplastic cells from case 5, located in chiasmatic region, did not express CD34. All these findings argue for definitively classifying these two tumors as papillary variant form of gangliogliomas, and to clearly separate them from PGNT. Notably, it could be interesting as ganglioglioma represent the single glio-neuronal tumoral entity susceptible to spontaneously present a malignant transformation [2]. Indeed, relatively little is known about malignant changes in PGNT [27]. Several cases with high proliferative activity have been reported but all showed malignant features at diagnosis and to our knowledge, no secondary malignant transformation has been described in PGNT [28, 29]. Besides, molecular data concerning malignant PGNT are very limited; BRAF V600E mutation was screened in two cases which were both negative [6]. Recurrent somatics alterations of FGFR1 are reported in pLGG [10] but no large pLGG cohort studying FGFR1 mutational status have included some PGNT. Contrary to the single case report describing a FGFR1 N546K mutation (located in the exon 14) [11], we did not find FGFR1 mutation in our cohort. No correlation seems to exist between the presence of a mutation and a positive immunostaining using phospho-FGFR1 antibody.

The identification of SLC44A1-PRKCA may also be useful to distinguish PGNT from other tumors with perivascular/angiocentric features.

In order to shed more light on the specificity of SLC44A1-PRKCA fusion and its potential value in the differential diagnosis of PGNT, we also investigated thirteen additional cases of PGNT mimics. Particularly, we studied five RGNTs cases that could be a differential diagnosis due to the periventricular location and the biphasic pattern. Two RGNTs cases have been previously tested for SLC44A1-PRKCA fusion and were negative [13]. In our cohort, none of ANETs, RGNTs, PRPs, PE, astroblastoma, neurocytomas studied showed the SLC44A1-PRKCA fusion, further supporting PGNT as a unique entity. PRKCA is the alpha isoform of protein kinase C, a family member of serine and threonine kinases that can be activated by calcium and the second messenger diacylglycerol. PRKCA can promote cell growth by phosphorylating and activating RAF1, which mediates the activation of the MAPK signaling pathway. Therefore, deregulation or high constitutive level of PRKCA expression secondary to fusion with SLC44A1 could directly result in the deregulation of the MAPK signaling pathway as observed in tumors showing BRAF V600E mutation [30], notably gangliogliomas, or KIAA1549-BRAF fusion as in pilocytic astrocytomas (PA) [8]. In our cohort, phospho-ERK analysis provides arguments for the activation of the MAPK signaling pathway.

MAPK pathway and BRAF status have been widely explored over the last few years showing that the histologic spectrum of low grade gliomas and glioneuronal tumors is vague but individualizing a group of low grade glioma with MAPK signaling pathway deregulation. PGNT is a rare entity which belongs to this group with MAPK signaling pathway deregulation. SLC44A1-PRKCA fusion alone could drive tumorigenesis, but additional studies are needed to confirm this hypothesis.