Polymorphous Low-Grade Neuroepithelial Tumor of the Young (PLNTY): Molecular Profiling Confirms Frequent MAPK Pathway Activation

Abstract

Polymorphous low-grade neuroepithelial tumor of the young (PLNTY) is a recently described epileptogenic tumor characterized by oligodendroglioma-like components, aberrant CD34 expression, and frequent mitogen-activated protein kinase (MAPK) pathway activation. We molecularly profiled 13 cases with diagnostic histopathological features of PLNTY (10 female; median age, 16 years; range, 5–52). Patients frequently presented with seizures (9 of 12 with available history) and temporal lobe tumors (9 of 13). MAPK pathway activating alterations were identified in all 13 cases. Fusions were present in the 7 youngest patients: FGFR2-CTNNA3 (n = 2), FGFR2-KIAA1598 (FGFR2-SHTN1) (n = 1), FGFR2-INA (n = 1), FGFR2-MPRIP (n = 1), QKI-NTRK2 (n = 1), and KIAA1549-BRAF (n = 1). BRAF V600E mutation was present in 6 patients (17 years or older). Two fusion-positive cases additionally harbored TP53/RB1 abnormalities suggesting biallelic inactivation. Copy number changes predominantly involving whole chromosomes were observed in all 10 evaluated cases, with losses of chromosome 10q occurring with FGFR2-KIAA1598 (SHTN1)/CTNNA3 fusions. The KIAA1549-BRAF and QKI-NTRK2 fusions were associated respectively with a 7q34 deletion and 9q21 duplication. This study shows that despite its name, PLNTY also occurs in older adults, who frequently show BRAF V600E mutation. It also expands the spectrum of the MAPK pathway activating alterations associated with PLNTY and demonstrates recurrent chromosomal copy number changes consistent with chromosomal instability.

INTRODUCTION

Polymorphous low-grade neuroepithelial tumor of the young (PLNTY) is a newly recognized morphologically and molecularly distinct epileptogenic tumor type affecting primarily children and young adults (1, 2). Although predominantly occurring in the young, PLNTY has been recently reported in an older adult individual (3). Histologically, PLNTY is heterogeneous but characteristically shows an infiltrative growth pattern, an oligodendroglioma-like component and strong diffuse aberrant CD34 expression (1).

Molecular profiling of PLNTY has shown evidence of mitogen-activated protein kinase (MAPK) pathway activation through BRAF V600E mutation and fusions involving FGFR2 (FGFR2-KIAA1598 [FGFR2-SHTN1], FGFR2-CTNNA3 and FGFR2-INA) and FGFR3 (FGFR3-TACC3) in a mutually exclusive pattern (1, 3–7). By genome-wide DNA methylation profiling, PLNTY segregates from other low-grade neuroepithelial tumors, but its epigenetic signature is closely related to ganglioglioma and similar to pilocytic astrocytoma (1). Indeed, some of the MAPK pathway activating alterations observed in PLNTY have also been described in ganglioglioma and pilocytic astrocytoma (8–13), suggesting that although distinctive by morphology and DNA methylation profiling, these low-grade neuroepithelial tumors activate the MAPK pathway by similar genetic mechanisms.

We have recently described the radiological features of PLNTY in a series of 9 cases and have shown that PLNTY demonstrates characteristic radiological features such as inferior temporal lobe and cortical/subcortical location, dense calcification, cystic component, and infrequent contrast enhancement (14). Herein, we report the molecular profile of these 9 cases and 4 additional cases, totaling 13 histopathologically defined cases of PLNTY, using a neuro-oncology 219-gene next-generation sequencing (NGS) panel, RNA sequencing, chromosomal microarray, and/or immunohistochemistry.

MATERIALS AND METHODS

This study was conducted according to Institutional Review Board-approved protocols.

Case Selection

Since the description of PLNTY (1), 13 cases were prospectively identified during the routine diagnostic workup either via the consultative pathology practice (n = 11) or operated at the Mayo Clinic (n = 2) within a 3-year period (2017–2019). Of these 13 cases, 4 are being reported for the first time (cases 3, 5, 12, and 13) and include the 2 oldest patients in this series, a 44-year-old woman and a 52-year-old man. The diagnosis of PLNTY was established based on recently described characteristic histopathological and immunohistochemical features, including the presence of an infiltrative growth pattern associated with oligodendroglioma-like components, strong often diffuse CD34 immunostaining, and frequent intra-tumoral calcifications (in the absence of Rosenthal fibers, eosinophilic granular bodies, myxoid microcysts, ependymal rosettes, microvascular proliferation or necrosis) (1). Demographic, clinical presentation, and radiological findings from the 4 newly reported cases were obtained from electronic medical records (n = 1) or from referring pathologist (n = 3). Available follow-up data were extracted from electronic medical records (n = 3) or provided by referring clinician (n = 4).

Immunohistochemical Studies

CD34 (clone QBEnd/10; 1:50; Leica Novocastra, Buffalo Grove, IL), BRAF V600E (clone VE1; 1:50; Abcam, Cambridge, UK), IDH1-R132H (clone H09; 1:50; Dianova, Hamburg, Germany), ATRX (clone#D-5; 1:1000; Santa Cruz, Dallas, TX), GFAP (rabbit polyclonal; 1:8000; Dako, Santa Clara, CA), and synaptophysin (clone 27G12; 1:50; Leica Novocastra) immunostains were performed according to validated protocols. For CD34 (n = 13), BRAF V600E (n = 11), IDH1-R132H (n = 8), GFAP (n = 8), and synaptophysin (n = 1) immunostains, distinct cytoplasmic staining was considered positive. For the ATRX (n = 8) immunostain, the presence of nuclear staining was interpreted as normal retained protein expression; the lack of nuclear staining in tumor cells would be compatible with loss of protein expression only in the presence of internal control nuclear staining in non-neoplastic cells.

Molecular Testing Platforms

DNA and RNA were extracted from macrodissected formalin-fixed paraffin-embedded tumor tissue (15). The median visually estimated tumor percent within the areas delineated for macrodissection was 65 (range, 60–90).

Amplicon-based DNA and RNA NGS libraries were prepared for a custom neuro-oncology NGS panel (n = 10) to interrogate 219 unique central nervous system (CNS) tumor-associated genes, including the coding region and intron/exon boundaries of 150 genes in the mutation subpanel as well as 104 known gene fusions, 29 abnormal transcript variants and novel fusion transcripts containing targeted regions of at least one gene partner within 81 genes (FGFR2 gene not included) in the fusion subpanel as previously described (15). RNA sequencing libraries (n = 6) were prepared using the Agilent SureSelect Human All Exon V7 RNA Direct XT (Agilent, Santa Clara, CA). Paired-end 2 × 151 base-pair sequencing was performed on an Illumina HiSeq 2500 instrument (Illumina, Inc., San Diego, CA).

NGS data were processed through custom bioinformatics pipelines using GRCh37 (hg19) reference genome build to detect single nucleotide variants and small insertions/deletions (<50 base pairs) with at least 15% variant allelic frequency and 100× coverage and gene fusion events with at least 5 and 3 unique fusion reads for the neuro-oncology NGS panel and RNA sequencing, respectively. Identified alterations were visualized using Alamut Visual, version 2.7 rev. 2 (Interactive Biosoftware, Rouen, France) and/or integrative genomics viewer (16), and curated based on publicly available (17–19) and in-house genetic databases as well as literature search. Sequence alterations were classified as a pathogenic mutation, variant of unknown significance or benign polymorphism, and in-frame fusions were classified as oncogenic or of uncertain significance. Sequence variation nomenclature was in accordance with Human Genome Variation Society recommendations (20) using selected reference transcripts (Supplementary Table S1). Only sequence alterations and fusion events classified as pathogenic/likely pathogenic (i.e. mutation) and oncogenic/likely oncogenic, respectively, were considered clinically relevant and reported herein. Reported fusion events were validated by RT-PCR/targeted direct Sanger sequencing whenever possible as described elsewhere (21). Primers are listed in Supplementary Table S2.

RNA sequencing data were analyzed using the RNA-Seq bioinformatics pipeline MAP-RSeq (22), which uses STAR 2.6.1c for alignment (23) and featureCounts 1.5.1 (24) for gene expression quantification. Gene expression was obtained for all 6 PLNTY cases evaluated in addition to a comparison group consisting of 2 KIAA1549-BRAF fusion-positive pediatric pilocytic astrocytomas (KIAA1549 exon 16 fused to BRAF exons 10 and 11, respectively; one case with 3 replicates) and 16 adult IDH (IDH1 and IDH2)-wildtype glioblastomas. Normalized gene counts were also obtained (Fragments Per Kilobase pair per Million mapped reads) and used to perform unsupervised clustering by principal component analysis to evaluate how different the samples were from each other. Differentially expressed genes were defined as those that met a false discovery rate <5% and log2 fold change criteria of >1.5 or <−1.5 (25). Canonical pathway analysis was performed using the Ingenuity pathway analysis software IPA (Qiagen, Redwood City, CA). Biological functions and diseases information within the IPA software were used to investigate the canonical pathways of interest.

Chromosomal microarray analysis (n = 10) was performed using the OncoScan FFPE Assay Kit (Thermo Fisher Scientific, Waltham, MA) (26, 27), a highly multiplexed molecular inversion probe assay that targets unique single nucleotide polymorphic base pairs and interrogates copy number changes in a whole-genome array. Genome-wide functional resolution is approximately 100 kilobases for non-mosaic deletions and duplications. Copy number events considered clinically significant included deletions larger than 500 kilobases, duplications larger than 1 megabase (although in the context of some fusions smaller deletions and duplications were noted), and copy neutral loss of heterozygosity larger than 10 megabases.

RESULTS

Clinicopathological Findings

As summarized in Figure 1 and detailed in Supplementary Table S1, there were 10 female and 3 male patients with a median age at diagnosis of 16 years (range, 5–52). Most patients presented with seizures/epilepsy (9 of 12 with available history; 75%), and none had a known history of a hereditary cancer predisposition syndrome. All tumors were supratentorial, 9 (69%) involving the temporal lobe and 3 (23%) the parietal lobe. The single non-hemispheric tumor occurred in the third ventricle (Supplementary Fig. S1). Calcifications were frequent (10 of 11 cases with available imaging; 91%).

FIGURE 1.

Summary of molecular and clinical findings.

All cases had an infiltrative growth pattern associated with oligodendroglioma-like components and strong, often diffuse, CD34 immunostaining. Representative histopathological findings are shown for cases 1, 2, 6, and 7 (Fig. 2). An astrocytic/cytologically ambiguous component was frequently observed and occasionally associated with nuclear pleomorphism as described in the original series (1). Pleomorphic cells were present in 3 cases as a predominant (case 1; Fig. 2B), significant (case 7; Fig. 2N), or focal (case 2) component. The case with a predominant pleomorphic component (case 1) also had increased mitotic activity (up to 3–4 mitoses per 10 high-power fields), some mimicking “granular mitoses” as previously reported and illustrated (14). Increased mitotic activity (up to 2 mitoses per 10 high-power fields) was also observed in the case with a significant pleomorphic component (case 7). In the remaining cases, mitotic activity was rare (up to 1 mitotic figure per 10 high-power fields) or absent. Intra-tumoral calcifications were frequent (10 of 13, 77%). One case (case 2) also had focal gemistocytic and neuronal components. Immunohistochemically, the gemistocytic component was strongly positive for CD34 and variably positive for GFAP; the neuronal component consisted of haphazardly arranged dysmorphic ganglion cells without definite binucleation that were negative for CD34 and positive for synaptophysin (Fig. 2D–I). Although gemistocytic and neuronal components were not originally described in PLNTY, it was felt that this case was still in keeping with PLNTY as these components were only focally observed. Rosenthal fibers, eosinophilic granular bodies, microcysts, ependymal rosettes, microvascular proliferation, or necrosis were absent.

FIGURE 2.

(A–C) Case 1: PLNTY with FGFR2-KIAA1598 (FGFR2-SHTN1) fusion and TP53 and RB1 inactivation showed areas with oligodendroglioma-like morphology (A), areas with cellular pleomorphism (B) and strong diffuse CD34 immunostaining (C). (D–I) Case 2: PLNTY with FGFR2-MPRIP fusion had strong, diffuse CD34-immunoreactive oligodendroglioma-like (D, G) and focal gemistocytic (E, H) components as well as focal synaptophysin-positive neuronal component (F, I). (J–L) Case 6: PLNTY with QKI-NTRK2 fusion showing the characteristic oligodendroglioma-like component and calcifications (J), an astrocytic component (K), and strong diffuse CD34 immunostaining (L). (M–O) Case 7: PLNTY with KIAA1549-BRAF fusion showing the characteristic oligodendroglioma-like component with calcifications (M), areas with cellular pleomorphism (N), and strong diffuse CD34 immunostaining (O). A–O, 200×.

BRAF V600E immunostain was positive in 6 (of 11; 54%) cases evaluated. All tested cases were negative for IDH1-R132H immunostain (n = 8) and showed retained ATRX immunohistochemical protein expression (n = 8).

Follow-Up

Follow-up data (median, 24 months; range, 3–36) were available in 7 (of 13) cases (Supplementary Table S1). Most patients (n = 6) were seizure-free without evidence of disease. The single patient who continued with seizures had a superimposed developmental disorder with severe developmental delay; she underwent gross total tumor resection at 5 years of age and had a non-specific 5 mm non-enhancing signal abnormality in the deep resection cavity at 36-month postoperative imaging studies.

Molecular Findings

Molecular findings are summarized in Figure 1 and detailed in Supplementary Table S1. A BRAF V600E mutation was identified in 6 (of 13; 46%) cases. The BRAF V600E identified by immunohistochemistry was confirmed by NGS in 2 cases (cases 11 and 12) in which the neuro-oncology mutation subpanel could also be performed. The observed immunohistochemical pattern and the variant allele frequency, relative to the estimated tumor percent at which the BRAF V600E mutation was identified, supported the impression that this mutation was a clonal event, as this mutation seemed to be present as a heterozygous event in all tumor cells (Supplementary Table S1). Neither the 2 above mentioned BRAF V600E mutant cases (cases 11 and 12) nor a third case (case 13), also BRAF V600E mutant, in which only the fusion subpanel could be obtained (Fig. 1) showed any additional sequence alterations/fusion events.

An in-frame gene fusion was observed in the remaining 7 (of 13; 54%) cases and involved FGFR2 in 5 cases, and NTRK2 and BRAF in one case each. All 5 FGFR2 fusion-positive cases contained exons 1–17, including its kinase domain (Supplementary Table S1). There was an FGFR2-KIAA1598 (FGFR2-SHTN1) (case 1), an FGFR2-MPRIP (case 2; Supplementary Fig. S2), an FGFR2-INA (case 3), and 2 FGFR2-CTNNA3 (cases 4 and 5) fusions. The FGFR2-KIAA1598 (FGFR2-SHTN1) and FGFR2-INA fusions were identical, and the FGFR2-CTNNA3 fusions were similar to FGFR2 fusion isoforms reported in PLNTY (1, 6). The FGFR2-KIAA1598 (FGFR2-SHTN1) and FGFR2-CTNNA3 fusions were also associated with variable-sized 10q deletion events disrupting FGFR2 (Supplementary Fig. S3). No copy number events involving either 10q (FGFR2 and INA) or 17p (MPRIP) were observed, suggesting that the FGFR2-MPRIP (Fig. 4) and FGFR2-INA fusion events may have formed through a balanced translocation (Supplementary Fig. S3). The remaining 2 fusions identified consisted of a QKI-NTRK2 (case 6; Fig. 3A) and a KIAA1549-BRAF (case 7; Fig. 3C) fusion, which were confirmed by RT-PCR followed by Sanger sequencing (not shown). The QKI-NTRK2 fusion was associated with a 9q21 duplication of approximately 191 kilobases disrupting NTRK2 in the context of multiple whole chromosome gains, including chromosome 9 (NTRK2), resulting in a focal ×4 copy number state at 9q21 (Figs. 3B, 4). The KIAA1549-BRAF fusion occurred in the context of a 7q34 heterozygous deletion of approximately 785 kilobases disrupting the 3′ portion of BRAF (Fig. 3D) and multiple other whole chromosome losses (Fig. 4).

FIGURE 4.

Summary of genome-wide copy number changes by chromosomal microarray.

FIGURE 3.

(A, B) Case 6: Sequencing reads spanning the QKI-NTRK2 fusion junction on forward and reverse directions (A) associated with NTRK2 disruption by a 9q21 duplication (B, blue bar: chromosomal region with copy number gain, gray curtain: probes mapping to NTRK2). (C, D) Case 7: Sequencing reads spanning the KIAA1549-BRAF fusion junction on forward and reverse directions (C) associated with a 3′ BRAF disruption by a 7q34 deletion (D, red bar: chromosomal region with copy number loss, grey curtains: probes mapping to KIAA1549 and BRAF). (A, C) Neuro-oncology targeted NGS fusion subpanel; (B, D) chromosomal microarray.

Principal component analysis of gene expression profiling using the 2,900 most differentially expressed genes (including 39 MAPK pathway-related genes; Supplementary Table S3) among PLNTY (n = 6), KIAA1549-BRAF fusion-positive pilocytic astrocytoma (n = 2), and IDH-wildtype adult glioblastoma (n = 16) showed that all 6 fusion-positive PLNTY cases with available RNA sequencing data (cases 1–5 and 7), including all cases with an FGFR2 fusion and the case with a KIAA1549-BRAF fusion, grouped in a cluster distinct from KIAA1549-BRAF fusion-positive pilocytic astrocytoma and adult IDH-wildtype glioblastoma (Fig. 5A). A heatmap of gene expression profiling using the 889 most differentially expressed genes (including 14 MAPK pathway-related genes; Supplementary Table S3) between PLNTY and pilocytic astrocytoma revealed relative transcriptional profile similarity among all fusion-positive PLNTY (Fig. 5B). The 2 cases with fusions not previously reported in PLNTY (FGFR2-MPRIP and KIAA1549-BRAF) showed greater similarity to the cases harboring the FGFR2-KIAA1598 (SHTN1)/CTNNA3 fusions previously reported in histomolecularly defined PLNTY than to KIAA1549-BRAF fusion-positive pilocytic astrocytomas (Fig. 5B).

FIGURE 5.

Unsupervised hierarchical clustering of RNA sequencing gene expression profiling data. (A) Principal component analysis (based on top 2,900 differentially expressed genes) showing that the 6 evaluated fusion-positive PLNTY formed clustered together and formed a group distinct from KIAA1549-BRAF fusion-positive pediatric pilocytic astrocytoma (PA) and adult IDH-wildtype glioblastoma (GBM). (B) Heatmap (based on top 889 differentially expressed genes) of fusion-positive PLNTY and pilocytic astrocytoma (PA) showing relative transcriptional profile similarity between the 6 evaluated fusion-positive PLNTY and that the FGFR2-MPRIP and KIAA1549-BRAF fusion-positive PLNTY had greater similarity to the PLNTY harboring the FGFR2-KIAA1598 (FGFR2-SHTN1) and FGFR2-CTNNA3 fusions respectively than to the KIAA1549-BRAF fusion-positive pilocytic astrocytoma.

Two fusion-positive cases had an apparently clonal mutation involving one or more tumor suppressor genes (Supplementary Table S1). A TP53 truncating mutation was identified in cases 1 (FGFR2-KIAA1598 [FGFR2-SHTN1] fusion-positive) and 7 (KIAA1549-BRAF fusion-positive). Both cases had evidence of TP53 biallelic tumor inactivation through whole chromosome 17 loss. Case 1 had an additional RB1 mutation associated with whole chromosome 13 loss also suggesting biallelic tumor inactivation of this tumor suppressor gene. Both patients are alive at 36 and 24 months after tumor resection respectively and do not show definitive evidence of recurrent/progressive disease (Supplementary Table S1).

It is noteworthy that the BRAF V600E mutation occurred predominantly in adult individuals (median age 30, range 17–52), whereas all fusions were observed in pediatric patients (median age 12, range 5–16). Additionally, individuals with BRAF V600E mutant tumors less frequently presented with seizures/epilepsy than patients with a fusion-positive tumor (3 of 6 vs 7 of 7, respectively). Other clinical and radiological parameters were similar between these 2 groups.

There was no evidence of an IDH mutation or ATRX inactivation in any of the 12 cases evaluated: 9 cases (2 BRAF V600E mutant and 7 fusion-positive) were IDH and ATRX wild-type by the neuro-oncology NGS mutation subpanel, and 3 cases (all BRAF V600E mutant) were negative for IDH1-R132H immunostain and showed retained ATRX immunohistochemical protein expression (Fig. 1; Supplementary Table S1).

Copy number changes were observed in all 10 tested cases (3 BRAF V600E mutant and 7 fusion-positive) and are shown in Figure 4. These changes predominantly involved whole chromosomes, and the most frequent (≥4 cases) recurrent events included loss of chromosomes 2, 13 and 22, and gain of chromosomes 5, 7, 8, 12, 16, 18, 19, 20, and X. No whole-arm 1p and 19q co-deletion was identified in any case. The copy number profile appeared generally similar between BRAF V600E mutant and fusion-positive cases, with the exception of case 7 (KIAA1549-BRAF fusion-positive), which only had chromosomal losses. The recurrent focal 10q losses in fusion-positive cases were associated with FGFR2 fusions as previously mentioned (Supplementary Fig. S3).

DISCUSSION

In keeping with prior studies (1, 3–7), multiomics profiling of our series of PLNTY showed evidence of MAPK pathway activation. In fact, a MAPK pathway activating genetic alteration was identified in every case (13 of 13; 100%) of our series, lending further support to the notion that PLNTY is a MAPK pathway-driven tumor. In addition to previously reported MAPK activating alterations in PLNTY such as the BRAF V600E mutation and FGFR2-KIAA1598 (SHTN1)/INA/CTNNA3 fusions (1, 6), we observed 3 MAPK activating fusions not previously described with PLNTY, an FGFR2-MPRIP, a QKI-NTRK2, and a KIAA1549-BRAF fusion in one case each. Gene expression profiling of most of the fusion-positive cases, including cases harboring the FGFR2-MPRIP and KIAA1549-BRAF fusions, confirmed that such tumors share similar transcriptional profile, supporting the impression that despite the different fusion partners, they represent a closely related group of tumors.

The FGFR2-MPRIP fusion, to the best of our knowledge, has not been previously reported in CNS tumors or any other tumor type. In FGFR2 oncogenic fusions, activation of FGFR2 kinase and downstream signaling (including MAPK pathway) seems to be mediated by oligomerization induced by the gene partner dimerization domain (28, 29). MPRIP encodes an actin-binding protein with a C-terminal coiled-coil dimerization domain and has been previously observed as a gene partner in oncogenic fusions involving non-CNS tumors (30, 31). The FGFR2-MPRIP fusion includes the FGFR2 kinase domain and the coiled coil dimerization domain from MPRIP similar to FGFR2-KIAA1598(SHTN1)/CTNNA3/INA fusions (1, 6). Taken together, these findings suggest that the novel FGFR2-MPRIP fusion is likely oncogenic. Interestingly, this fusion was identified in the single case of our series that had minor gemistocytic and neuronal components. These components raised the possibility that this tumor could represent a ganglioglioma. FGFR2 fusions have been reported in a few ganglioglioma cases wherein the glial component was morphologically oligodendroglial (11, 12). However, the structural and gene expression profile similarity between the FGFR2-MPRIP fusion and the other FGFR2 fusions reported in PLNTY, in conjunction with the presence of otherwise characteristic histopathological features of PLNTY, such as the oligodendroglioma-like component and the strong and diffuse CD34 immunostaining pattern in both the oligodendroglial and gemistocytic components, were felt to support the diagnosis of PLNTY in our case.

The QKI-NTRK2 fusion was similar and the KIAA1549-BRAF fusion was identical to MAPK pathway activating oncogenic fusion isoforms previously described in pilocytic astrocytoma (8), underscoring the molecular similarity between PLNTY and pilocytic astrocytoma (1). The underlying mechanism leading to these overlapping fusion events in PLNTY, however, appears to be distinct from the mechanism observed in pilocytic astrocytoma. The QKI-NTRK2 fusion was associated with a focal duplication, whereas the pilocytic astrocytoma QKI-NTRK2 fusion occurred in the context of a balanced translocation (8). The KIAA1549-BRAF fusion was associated with a 7q34 deletion and multiple whole chromosome losses instead of the typical 7q34 tandem duplication that occurs in isolation or with additional whole chromosome gains in pilocytic astrocytoma (32–35). Moreover, this KIAA1549-BRAF fusion was associated with a transcriptional profile with closer similarity to the profile of fusion-positive cases harboring FGFR2 fusions previously observed in PLNTY rather than of KIAA1549-BRAF fusion-positive pilocytic astrocytoma. It is noteworthy that multiple 7q33q34 interstitial deletions resulting in the same KIAA1549-BRAF fusion isoform have been reported in a pediatric patient who presented with possible seizures and had a left temporal calcified tumor classified as “likely a pilocytic astrocytoma” (36). Also, identical KIAA1549-BRAF fusion isoform has been reported in 2 cases of pediatric oligodendroglioma presenting as calcified tumors and with seizures in one patient (37). These cases were described prior to the original series of PLNTY and were not evaluated for CD34 immunostain.

Our data and prior studies indicate that PLNTY is molecularly characterized by either a BRAF V600E mutation or a fusion event involving MAPK pathway-related genes in a mutually exclusive pattern (1, 6). Additionally, the median age of onset in our series was similar to the original series of PLNTY (16 vs 17.5 years) (1), and we identified tumors histopathologically compatible with PLNTY in older adult patients (>40 years) as recently reported (3). We observed that the BRAF V600E mutation predominantly occurred in adult individuals, whereas fusions were primarily observed in pediatric patients. However, our sample size is relatively small to support a conclusive association, and additional case series are needed to confirm our findings. From a diagnostic perspective, the expanded age spectrum in which PLNTY may be encountered supports the inclusion of PLNTY within the differential diagnoses of tumors with infiltrating growth pattern, oligodendroglial morphology and calcifications in an adult individual, especially when located in the temporal lobe. In such clinicopathological scenario, PLNTY would be a strong diagnostic consideration if an IDH mutation and whole arm 1p19q co-deletion are absent and a BRAF V600E mutation is present. From a therapeutic standpoint, even though PLNTY seems to be a low-grade indolent tumor (1, 6), targeted therapies could be considered for tumors that cannot be completely resected or that are located in surgically challenging locations (38, 39).

The clinical significance of biallelic inactivation of tumor suppressor genes (TP53 and RB1) in 2 of our fusion-positive cases (FGFR2-KIAA1598 [FGFR2-SHTN1] and KIAA1549-BRAF) is unclear but a recent report of an FGFR3-TACC3 fusion-positive and TP53 mutant PLNTY that underwent malignant transformation with acquisition of an RB1 mutation 17 months after gross total tumor resection suggests that TP53 and RB1 mutations may be markers associated with risk for malignant transformation (40). Tumor suppressor mutations, including TP53, have been reported in conjunction with KIAA1549-BRAF fusion in pilocytic astrocytoma but follow-up data were unavailable (41). Our 2 fusion-positive and tumor suppressor mutant tumors morphologically displayed a higher degree of cellular pleomorphism and increased mitotic activity when compared to cases without tumor suppressor mutations. Both patients are alive and without definitive evidence of recurrent/progressive disease at 24 and 36 months after tumor resection. Molecular profiling of additional cases with available clinical data and extended follow-up will help shed light into the significance of such findings.

In keeping with previously reported cases (5, 6), all tumors evaluated in our study had copy number changes predominantly involving whole chromosomes, suggesting that PLNTY is also characterized by chromosomal instability. The observed aneuploidy pattern is similar to the pattern described in other low-grade neuroepithelial tumors, such as ganglioglioma and dysembryoplastic neuroepithelial tumor, which primarily have whole chromosome gains, including chromosomes 5, 7, 8, 12, 19, 20, and X (9, 42–44). A similar pattern has also been observed in a subset of KIAA1549-BRAF fusion-positive pilocytic astrocytoma that had the characteristic 7q34 tandem duplication (34). It is noteworthy that recurrent 10q loss has been reported in ganglioglioma (42), that FGFR2 copy number alterations have been described in pediatric oligodendroglial tumors (10), and that chromosomal instability has been reported in a recent series of 3 BRAF V600E mutant pediatric tumors with oligodendroglial morphology, 2 of which had strong CD34 immunostaining and were described as possibly representing a subgroup of PLNTY (45). Although recurrent chromosomal copy number changes have not been clearly stated in the original series of PLNTY, these seemed present on available genome-wide copy number profiles derived from methylation array data (1). Leveraging the OncoScan microarray platform, which has high sensitivity and resolution for copy number changes, we were able to define the copy number profile and identify recurrent copy number patterns in our series of PLNTY.

Limitations of our study include the lack of long-term clinical follow-up and the small RNA sequencing data reference set. Despite these limitations, we believe that we were still able to provide additional insights and increase our understanding on this novel tumor type.

In conclusion, this study confirms PLNTY as a distinct low-grade neuroepithelial tumor with frequent MAPK pathway activation that also occurs in older adults and expands the spectrum of the underlying MAPK pathway activating alterations as well as demonstrates recurrent chromosomal copy number changes indicative of chromosomal instability.

 

Sources of support: Laboratory Genetics and Genomics Division, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota, USA.

Conflicts of interest: All authors have no conflicts of interests to disclose.

Supplementary Data can be found at academic.oup.com/jnen.