Multiplex Detection of Pediatric Low-Grade Glioma Signature Fusion Transcripts and Duplications Using the NanoString ncounter System

Transcription

1 J Neuropathol Exp Neurol Vol. 76, No. 7, July 2017, pp doi: /jnen/nlx042 ORIGINAL ARTICLE Multiplex Detection of Pediatric Low-Grade Glioma Signature Fusion Transcripts and Duplications Using the NanoString ncounter System Scott Ryall, BSc, Anthony Arnoldo, PhD, Rahul Krishnatry, MD, Matthew Mistry, MSc, Kangzi Khor, BSc, Javal Sheth, BSc, Cino Ling, BSc, Stephie Leung, MSc, Michal Zapotocky, MD, Ana Guerreiro Stucklin, MD, Alvaro Lassaletta, MD, Mary Shago, PhD, Uri Tabori, MD, and Cynthia E. Hawkins, MD, PhD Abstract Previous studies identified recurrent fusion and duplication events in pediatric low-grade glioma (plgg). In addition to their role in diagnosis, the presence of these events aid in dictating therapy and predicting patient survival. Clinically, BRAF alterations are most commonly identified using fluorescent in situ hybridization (FISH). However, this method is costly, labor-intensive and does not identify nonbraf events. Here, we evaluated the NanoString ncounter gene expression system for detecting 32 of the most commonly reported fusion/duplication events in plgg. The assay was validated on 90 plgg samples using FISH as the gold standard and showed sensitivity and specificity of 97% and 98%, respectively. We next profiled formalin-fixed paraffin-embedded preserved biopsy specimens from 429 plgg cases. 171 (40%) of the cases within our cohort tested positive for a fusion or duplication event contained within our panel. These events, in order of prevalence, were KIAA1549-BRAF 16;9 (89/171, 52.0%), KIAA1549-BRAF 15;9 (42/171, 24.6%), KIAA1549-BRAF 16;11 (14/171, 8.2%), FGFR1- TACC1 17;7 (13/171, 7.6%), MYBL1 duplication (5/171, 2.9%), KIAA1549-BRAF 18;10 (4/171, 2.3%), KIAA1549-BRAF 15;11 (2/171, 1.2%), FAM131B-BRAF 2;9 (1/171, 0.6%), and RNF130-BRAF 3;9 (1/ From the Department of Cancer and Stem Cell Biology, Hospital for Sick Children (SR, RK, MM, KK, CL, SL, AGS, UT, CEH), Toronto, Ontario, Canada; Department of Laboratory Medicine and Pathobiology, University of Toronto (SR, SL, MS, CEH), Toronto, Ontario, Canada; Divisions of Pathology and Genome Diagnostics, Department of Paediatric Laboratory Medicine (AA, JS, MS, CEH); and Division of Hematology and Oncology, The Division of Hematology and Oncology should be followed by Department of Pediatrics, Hospital for Sick Children, Toronto, Ontario, Canada. Hospital for Sick Children (RK, MZ, AGS, AL, UT), Toronto, Ontario, Canada; Department of Radiation Oncology, Tata Memorial Hospital (RK), Parel, Mumbai, India; and Institute of Medical Science, University of Toronto (UT), Toronto, Ontario, Canada. Send correspondence to: Cynthia Hawkins, MD, PhD, FRCPC, Department of Pediatric Laboratory Medicine Division of Pathology Hospital for Sick Children The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G1X8; cynthia.hawkins@sickkids.ca Financial Support: Canadian Institute of Health Research (CIHR) Graduate Scholarship (Scott Ryall) and Government of Canada through Genome Canada and the Ontario Genomics Institute (OGI-121). The authors have no duality or potential conflicts of interest to declare. Supplementary Data can be found at 171, 0.6%). This work introduces NanoString as a viable clinical replacement for the detection of fusion and duplication events in plgg. Key Words: Duplication, Fusion, Low-grade glioma, NanoString, Pediatrics. INTRODUCTION Pediatric low-grade gliomas (plgg) are the most common pediatric brain tumor and account for 35% of central nervous system tumors in children (1 4). plgg includes a range of histologic types with WHO grade I pilocytic astrocytoma being the most frequent (3). Clinically, plggs are generally slow growing lesions that can occur throughout the central nervous system and are surgically removed when possible.complete surgical resection is the most significant predictor of improved survival in plgg (5). However, the extent of resection is often dependent on tumor location and hence, only feasible for a subset of cases. For example, tumors arising in the posterior fossa can be completely resected and thereby boast an excellent progression free survival (6). However, for plggs located in the optic pathway, brainstem, or spinal cord, attempted resection can often lead to surgical-related morbidity (7 8). Historically, radiotherapy was used as the primary treatment for unresectable plgg. However, the long-term effect of radiotherapy on plgg patients is a major concern (9 11) and has recently been linked to delayed patient mortality (12). This has resulted in a more conservative low-dose chemotherapy and debulking approach as the primary strategy for unresectable plgg management, but this has shown variable levels of effectiveness (13 15). Clinically, this is often attributed to the high degree of heterogeneity within this group which results in clinical courses ranging from prolonged growth arrest to continuous progression and potential recurrence (16, 17). The high degree of heterogeneity and the resultant variability of clinical outcome generated an urgent need to identify biological and genetic markers capable of risk stratification. The involvement of RAS-MAPK pathway dysregulation in plgg was first suggested upon the identification of a novel duplication in chromosome 7q34 (18, 19). Later, it was shown 562 VC 2017 American Association of Neuropathologists, Inc. All rights reserved.

2 J Neuropathol Exp Neurol Volume 76, Number 7, July 2017 that this gain resulted in a gene fusion between BRAF and KIAA1549 (20). The breakpoints of the duplicated region fall within the BRAF and KIAA1549 genes, and result in a fusion between the 5 0 end of KIAA1549 and the 3 0 end of BRAF at the tandem duplication junction. This results in the N-terminal regulatory domain of BRAF being replaced, leading to constitutively enhanced kinase activity (21). Five separate KIAA1549-BRAF fusion configurations were originally described including between exons 16;9, 15;9, 16;11, 18;10, and 19;9 in order of prevalence (19). Fusions involving FAM131B and SRGAP3 with BRAF resulting in similar kinase activation have also been described, although at significantly lower frequencies (21, 22). Subsequent findings showed that RAS-MAPK activation through BRAF fusions are present in both grade I and grade II malignancies, although significantly less frequent in the latter (21, 23 27). The presence of BRAF fusion events in plgg has been shown to aid in tumor diagnosis (28), predict clinical outcome (29, 30), and dictate the most appropriate therapeutic approach (31 33), and, as such, represents a critical factor in the management of plgg. Besides BRAF, additional fusions such as those involving FGFR1, MYB/MYBL1, QKI, and NTRK2/3 have also been described in plgg (34 37). As is seen in the BRAF fusions, kinase activation is the resultant effect. Due to their size and location, plgg biopsies are generally processed as formalin-fixed paraffin-embedded (FFPE) tissue samples. As such, the use of these tissue samples for the purposes of BRAF fusion detection requires specially designed and validated assays that tolerate the degraded and fragmented material extracted from the FFPE specimens. Often fluorescent in situ hybridization (FISH) or real-time PCR based methodologies are used for clinical identification of BRAF fusion events in plgg. FISH is sensitive for the detection of BRAF duplication but may have problems with specificity, particularly in aneuploid samples, and is labor-intensive and thus expensive to implement. Further, it does not give information regarding the fusion partner and will not detect nonbraf fusions within the same test. Because of the number of possible partners, RT-PCR becomes impractical for identification of any but the most common fusion events. To develop and optimize a clinical grade test capable of robustly identifying a spectrum of plgg fusion and duplication events in FFPE-preserved tissue samples, we have taken advantage of the NanoString ncounter gene expression system (38 40) on which we developed a plgg-specific panel capable of detecting the majority of reported fusion and duplication transcripts described in plgg. The sensitivity and specificity of this method were compared against standard FISH analysis and shown to be comparable. Lastly, we utilized the Nano- String system to effectively characterize a cohort of clinical plgg to illustrate the effectiveness of the method and to interrogate the prevalence of certain fusion events. MATERIALS AND METHODS Patient Samples Upon institutional Research Ethics Board approval of the study, review of the pathology and oncology databases at Fusion/Duplication Detection in Pediatric Low-Grade Glioma the Toronto Hospital for Sick Children (SickKids) identified 634 patients diagnosed with plgg from 1985 to Fiftyeight lacked sufficient material quantity for RNA extraction and therefore 576 were used in this study. Cases were comprised of FFPE samples and consisted of biopsied or subtotally resected masses where available. Before analysis, a hematoxylin and eosin-stained section from each sample was centrally reviewed for pathological diagnosis and grading according to WHO criteria (41). RNA Isolation Total RNA was isolated from 5- to 10-mm scrolls of FFPE preserved tissue with the RNeasy FFPE extraction kit (QIAGEN, Valencia, CA) under the manufacturer s guidelines. RNA quantity and quality were assessed using the Nano- Drop 2000 (Thermo Scientific, Waltham, MA). Quality was based on 260/280 values and samples between 1.9 and 2.1 were considered to be sufficient. Fluorescent In Situ Hybridization The FISH test was designed to detect the tandem duplication using BAC clones within the duplicated region at 7q34 (RP11-248P7 and RP11-837G3) and 7p12.1 control probes (RP11-478M17 and RP11-876P22). Probes were obtained from The Centre for Applied Genomics (The Hospital for Sick Children, Toronto, Ontario, Canada). The 7q34 clones were directly labeled with Spectrum Green fluorochrome and the control clones were directly labeled with Spectrum Orange fluorochrome. Paraffin FISH analysis was performed on 4-mm tumor sections. Slides were baked overnight to fix the section to the slide and pretreated using a Paraffin Pretreatment kit (Abbott, Chicago, IL). Sections were dehydrated prior to slide/probe codenaturation on the Thermobrite (Intermedico, Markham, Ontario, Canada). Denaturation conditions used for paraffin embedded slides/probes were 85 C for 7 minutes, followed by overnight incubation at 37 C. Slides were washed in 0.4 SSC/0.3% NP40 at 73 C for 30 seconds, followed by 2 SSC/0.1% NP40 at room temperature for 30 seconds. Slides were counterstained with DAPI. Nuclei were analyzed using a Axioplan2 epifluorescence microscope (Zeiss, Jena, Germany). Images were captured by an Axiocam MRm Camera (Imaging Associates, Bicester, UK) and analyzed using an imaging system with MetaSystems Isis Software version (MetaSystems, Boston, MA). NanoString Fusion Detection Five hundred nanograms of total RNA input was used on the NanoString ncounter system (nanostring Technologies, Seattle, WA) utilizing our newly generated plgg panel (reagents are available from NanoString Technologies and codeset through the Hospital for Sick Children). Samples were processed on the ncounter Preparation Station and the cartridge scanned at 555 fields of view on the ncounter Digital Analyzer. Raw NanoString counts were background adjusted with a Poisson correction based on the negative control spikes included in each run. This was followed by a technical normalization using the 4 housekeeping transcripts included in each 563

3 Ryall et al J Neuropathol Exp Neurol Volume 76, Number 7, July publications from 2014 with reported fusion/duplication events in plgg 13 publications involved analyzing previously described fusion/duplication events. In such cases, we cite the manuscripts originally identifying the aberration. 14 publications with unique fusions/duplications reported 4 publications did not report the exact sequence break-point of the fusion events run (ABCF1, ALAS1, CLTC, and HPRT1). Data is viewed using a box plot and the extreme statistical outlier method was used to detect the presence of an expressed fusion. A fusion or duplication transcript is designated as expressed if the raw count was above the average of the internal positive control raw counts plus 3xIQR as per manufacturer suggestion. RESULTS Identification of NanoString Fusion and Duplication Transcript Targets A comprehensive literature review of publications up until 2015 was completed to identify all fusion and duplication transcript events described in plgg (20 23, 34, 36, 42 45). Target inclusion was based on 2 requirements: the target must have been reported in a peer reviewed journal and the breakpoint sequence must be provided to allow for proper Nano- String capture probe design. A schematic depicting the literature review process and filtering can be seen in Figure 1.From this review, 1 duplication and 31 fusion events were selected for inclusion within the plgg panel (Table 1). These included all 9 previously reported KIAA1549-BRAF fusion events across all known breakpoints including 13;9, 15;10, 15;11, 15;9, 16;10, 16;11, 16;9, 18;10, and 19;9. Further, BRAF s other fusion partners were included: MACF1, CLCN6, FAM131B, FXR1, GNAI1, MKRN1, and RNF130. Additionally, nonbraf fusions were contained within the panel, including those involving FGFR1, MYB, MYBL1, RAF, SRGAP3, and QKI. Of note, the reported MYB and FGFR1 duplication events were not included in the final panel due to probe similarity between these and several other transcripts detected within the panel resulting in false positive detection (34). Protocol Overview and Result Interpretation A sample run on the NanoString plgg panel can be completed in 2 days. Sample RNA is mixed with the 10 publications with unique fusion/duplications with reported breakpoints 31 unique fusions and 1 unique duplication FIGURE 1. Flowchart depicting the literature breakdown followed to identify potential targets for the plgg NanoString Panel. 564 proprietary plgg panel CodeSet targeting the events described in Table 1 as well as both positive and negative technical control probes as previously described (40). CodeSet and sample are allowed to hybridize overnight for 20 hours. Excess CodeSet is then washed away and the hybridized CodeSet/ RNA complexes are next purified and immobilized onto the ncounter cartridge system to allow for data collection. The ncounter cartridge is then scanned to identify the unique fluorescent signatures (barcode) associated with each CodeSet probe. The barcodes are counted and a total of each target is tabulated. Raw NanoString counts are background adjusted and technically normalized based on the positive control spikes. Data is viewed using a box plot of the log values and the extreme statistical outlier method was used to detect the presence of an expressed fusion. A fusion or duplication transcript is designated as expressed if the raw count was above the average of the internal positive control raw counts plus 3xIQR. An example of the panel output is shown in Figure 2.In sample one, an instance of insufficient material quality is identified. This is a standard quality control step within the panel and is based on the detection of the 4 housekeeping transcripts included within the codeset. If the house-keeping control transcripts are not detected at sufficient levels (geometric mean for the 4 house-keeping genes >100 is the quality cut-off) and the issue is not deemed mechanical as per NanoString-specific internal quality control flags, the sample is concluded as having insufficient RNA quality. Our analysis identified that the age of the sample directly correlated with the likelihood of having insufficient RNA quality (Supplementary Fig. S1). All samples falling below the required material quality are unsuitable for this method of fusion detection and as such, we advise that samples no older than 13 years be used on this platform. Samples 2 6 tested positive for 5 different fusion transcripts contained within the plgg panel: KIAA1459-BRAF 16;11, FAM131B-BRAF 2;9, KIAA1459-BRAF 15;11, KIAA1459- BRAF 15;9, and KIAA1459-BRAF 16;9, respectively. As seen, each fusion call is easily made based on visualization of the results.

4 J Neuropathol Exp Neurol Volume 76, Number 7, July 2017 Fusion/Duplication Detection in Pediatric Low-Grade Glioma TABLE 1. Fusion Targets and Their Breakpoints Included on the NanoString plgg Panel Genes Exon 1 Exon 2 Event Gene ID KIAA1549-BRAF 13 9 Fusion ENST NM_ KIAA1549-BRAF Fusion ENST NM_ KIAA1549-BRAF Fusion ENST NM_ KIAA1549-BRAF 15 9 Fusion ENST NM_ KIAA1549-BRAF Fusion ENST NM_ KIAA1549-BRAF Fusion ENST NM_ KIAA1549-BRAF 16 9 Fusion ENST NM_ KIAA1549-BRAF Fusion ENST NM_ KIAA1549-BRAF 19 9 Fusion ENST NM_ BRAF-MACF Fusion NM_ NM_ CLCN6-BRAF 2 11 Fusion NM_ NM_ ETV6-NTRK Fusion NM_ NM_ FAM131B-BRAF 1 10 Fusion NM_ NM_ FAM131B-BRAF 2 9 Fusion NM_ NM_ FAM131B-BRAF 3 9 Fusion NM_ NM_ FGFR1-TACC Fusion NM_ NM_ FGFR3-TACC Fusion NM_ NM_ FXR1-BRAF Fusion NM_ NM_ GNAI1-BRAF 1 10 Fusion NM_ NM_ MKRN1-BRAF 4 11 Fusion NM_ NM_ MYB-MAML2 6 4 Fusion NM_ NM_ MYB-PCDHGA1 9 2 Fusion NM_ NM_ NACC2-NTRK Fusion ENST ENST NTRK3-ETV Fusion NM_ NM_ QKI-RAF Fusion NM_ ENST QKI-MYB 2 16 Fusion NM_ NM_ QKI-NTRK Fusion NM_ ENST RNF130-BRAF 3 9 Fusion NM_ NM_ SRGAP3-RAF Fusion ENST ENST SRGAP3-RAF Fusion ENST ENST ST6GAL1-WHSC1 2 4 Fusion NM_ NM_ MYBL1 / / Duplication / ABCF1 Housekeeping NM_ ALAS1 Housekeeping NM_ CLTC Housekeeping NM_ HPRT1 Housekeeping NM_ Comparison to Fluorescent In Situ Hybridization The current test used for the identification of BRAF 7q34 duplication events in most clinical labs is FISH. To compare the NanoString panel, we ran 90 fresh frozen or FFPEpreserved samples on the NanoString panel and compared the results to those observed via FISH (Supplementary Fig. S2). Results from the comparison can be seen in Table 2. Using FISH as the gold standard, we identified 2 discrepant cases using the NanoString panel, 1 false positive and 1 false negative. Both of these cases were further characterized using CytoScan HD (Affymetrix, Santa Clara, CA). In both cases, CytoScan confirmed the original FISH conclusion. It is important to further note that originally 95 samples were used for comparison, but 5 were excluded due to inconclusive FISH results. In these cases, NanoString was able to effectively identify a genotype. In addition to comparable results to FISH, the plgg panel provides a robust method of fusion and duplication detection in instances of inconclusive FISH results. Two examples of such are seen in Figure 3. InFigure 3A, a sample wild type for the KIAA1549-BRAF fusion yielded inconclusive FISH results, requiring detailed follow-up due to the presence of 3 copies of chromosome 7 within the sample (later identified via Q-PCR). Comparably, when run on the plgg panel, the sample was identified as wild type and required minimal analysis time for this conclusion. Likewise, in Figure 3B, the signal pattern identified via FISH suggested a tandem duplication of the region 7q34 of which 70% displayed a signal pattern suggestive of the presence of 3 chromosome copies, 2 with a duplicated 7q34 region, and 1 with a nonduplicated 7q34 region. A further 25% of nuclei appeared to have double chromosome 7 content as compared to the main clonal line with signals suggestive of 6 copies of chromosome 7, 3 of which harbored the fusion event. As seen, the plgg panel identified the KIAA1549-BRAF 16;9 fusion event. 565

5 Ryall et al J Neuropathol Exp Neurol Volume 76, Number 7, July 2017 FIGURE 2. Example of the NanoString output. Data is displayed as a boxplot with extreme outliers indicative of a positive fusion detection. Here lane 1 shows a degraded sample that failed to meet QC requirements, while lanes 2 10 show instances of fusion positive samples. TABLE 2. Sensitivity and Specificity of the NanoString plgg Panel as Compared to FISH as the Gold Standard Fluorescent In Situ Hybridization Sensitivity Specificity Test Positive Test Negative 97% 98% NanoString Test Positive True Positive 42 False Positive 1 Test Negative False Negative 1 True Negative 48 Prevalence of Fusion Events in plgg The NanoString plgg panel was used to identify the presence of fusion or duplication transcripts in plgg samples treated at the Hospital for Sick Children from 1985 to We identified 634 cases diagnosed as low-grade gliomas including pilocytic astrocytoma (PA), low-grade astrocytoma NOS (LGA, NOS), ganglioglioma (GG), dysembryoplastic neuroepithelial tumor (DNET), diffuse astrocytoma (DA), pilomyxoid astrocytoma (PMA), pleomorphic xanthoastrocytoma (PXA), oligodendroglioma, central neurocytoma, and subependymal giant cell astrocytoma (SEGA). Of the 634, 576 had sufficient tissue quantity for RNA extraction and were run on the plgg panel. 147 failed to meet the RNA quality cut-off due to the age of the sample as discussed above. Of the 429 samples that yielded results from the plgg panel, 166 (39%) and 5(1%) samples tested positive for the presence of a fusion or duplication event, respectively (Fig. 4A). These events, in order of prevalence, were KIAA1549-BRAF ;9 (89/171, 52.0%), KIAA1549-BRAF 15;9 (42/171, 24.6%), KIAA1549-BRAF 16;11 (14/171, 8.2%), FGFR1-TACC1 17;7 (13/171, 7.6%), MYBL1 duplication (5/171, 2.9%), KIAA1549- BRAF 18;10 (4/171, 2.3%), KIAA1549-BRAF 15;11 (2/171, 1.2%), FAM131B-BRAF 2;9 (1/171, 0.6%), and RNF130-BRAF 3;9 (1/171, 0.6%)(Fig. 4B). Based on histology, pilocytic astrocytoma contained the most fusion events, while oligodendroglioma, PXA, and DNETs housed no fusions tested for on this panel. A full breakdown of the histological classifications of fusion/duplication positive samples is seen in Figure 4C. DISCUSSION Constitutive activation of the oncogenic RAS-MAPK pathway is involved in many cancers including both adult (46) and pediatric glioma (21, 23 27). Events leading to this activation are thought to be an early event in the development of these cancers, and are critical in aiding physicians in treating

6 J Neuropathol Exp Neurol Volume 76, Number 7, July 2017 Fusion/Duplication Detection in Pediatric Low-Grade Glioma A FISH Wild Type B FISH these tumors (18 21, 47). In this study, we developed a novel panel of fusion and duplication targets that are most commonly associated with plgg to be run on the NanoString ncounter system. This system provided high quality results on a range of samples, including those preserved in FFPE. When compared to FISH, the NanoString platform had comparable performance parameters while decreased processing and turnaround times and providing detailed results on multiple fusion targets and breakpoints within a single test. We conclude that the use of the NanoString plgg panel for fusion detection in plgg is poised to replace current FISH methodologies. Over the last decade, several seminal papers identifying multiple sources of RAS-MAPK pathway activation via fusion events in plgg have been identified (20 23, 34, 36, 42 45). In light of this, we aimed to generate a test capable of identifying all reported fusion and duplication targets (Fig. 1 and Table 1). NanoString, due to its proprietary reporter and capture probe designs is capable of up to 500 target detections per run (40), and therefore is able to detect multiple fusion/duplication events per run. Target inclusion required that the reported fusion be published and have a concisely defined breakpoint. In instances of homology, the fusion targets more commonly reported were included. The final plgg panel included 32 targets including 31 fusion and 1 duplication events. Importantly, the NanoString platform is amenable to expansion, meaning that newly identified fusions in plgg can be added to the panel, assuming that no cross reactivity with the current CodeSet is observed. An immediate benefit of the NanoString plgg panel over current methods is relieving the need of intense KIAA1549-BRAF 16:09 FIGURE 3. Examples of FISH analysis compared to NanoString. (A) Depicts a sample called inconclusive with FISH while NanoString concisely calls the samples WT. Likewise in (B), FISH concludes the sample to be inconclusive, while NanoString identifies the fusion present. interpretation of the results. As seen in Figure 2, results are displayed in a boxplot format, with an extreme outlier indicative of a fusion event. This method is preferred as it does not depend on a normal distribution of data, allows for a single sample comparative analysis, and is highly stringent in its call cut-offs. However, due to this tight stringency, this method may be prone to false negative results and must have the stringency adjusted accordingly, although this was not observed within our analysis. Comparatively, FISH is often determined based on experimenter confidence, which may lead to unintentional erroneous or inconclusive results in cases with complex signal patterns. Two such cases are highlighted in Figure 3, in which the NanoString panel robustly identified the aberration. These samples tested with FISH were difficult to interpret. In cases like these, secondary testing may be requested which is both costly and time-consuming. Our comparison of FISH to NanoString showed comparable results in terms of sensitivity and specificity (Table 2): 97% and 98% sensitivity and specificity were accomplished, respectively, using the NanoString panel, with 2 discrepant cases identified. With the false positive sample, NanoString identified a clear KIAA1549-BRAF 16;9 fusion, while FISH analysis appeared negative. Following up with CytoScan HD, a 658kb deletion between the KIAA1549 and BRAF genes was observed. This suggests that the fusion transcript is there, but generated by a rearrangement other than tandem duplication that was not detected using FISH. Although uncommon, a deletion-based mechanism of fusion formation has been previously reported (48). In the false negative, NanoString identified a clear normal result, while FISH was positive for the 567

7 Ryall et al J Neuropathol Exp Neurol Volume 76, Number 7, July 2017 A Duplication 1.3% (6) C FGFR1-TACC1 6.4% (12) MYBL1 Duplication 3.2% (6) RNF130-BRAF 0.5% (1) Fusion 39.0% (183) Wild Type 59.7% (280) KIAA1549-BRAF 89.4% (169) SEGA 0.6% (1) Unknown 1.2% (2) PA 69.0% (118) duplication. RNA-SEQ follow-up identified a novel fusion event not previously described and hence, not included in our original panel and therefore not detected. Importantly, other clinically approved methods such as RT-PCR as described by Tian et al. (42), reliably report the presence of BRAF-KIAA1549 fusions in FFPE patient samples. However, as compared to NanoString, it is incapable of detecting multiple targets within 1 analysis. Therefore, the widespread breakpoints and fusion/duplication partners makes it a very time-consuming process. Alternatively, several laboratories utilize single-nucleotide polymorphism (SNP)-based assays to identify chromosomal abnormalities (49, 50). This method has been shown to robustly identify fusion events in a wide array of plgg samples with a straightforward analysis. However, SNP arrays are expensive, sometimes incompatible with FFPE preserved tissue, do not detect balanced rearrangements, and are often insensitive to single copy duplication events in samples frequently contaminated with normal tissue. The use of massively parallel sequencing based targeted panels such as the OncoPanel and OncoCopy have also been suggested for clinical use (51, 52). These platforms are ideal due to their ability to identify polymorphisms, indels, small copy number changes and fusion events. Despite the appeal of these B DA 4.7% (8) GG 4.7% (8) FAM131B-BRAF 0.5% (1) Glioneuronal 0.6% (1) LGA, NOS 14.0% (24) Neurocytoma 1.2% (2) PMA 4.1% (7) PXA 0% Oligodendroglioma 0% DNET 0% FIGURE 4. Incidence of fusions within cohort of plgg patients treated at the Hospital for Sick Children since (A) Depicts the incidence of fusion versus WT samples (B) Shows the histologies harboring the fusion/duplication events. (C) Shows the fusion breakdown of these tested positive on the plgg panel. 568 types of methods, it is important to consider the cost and time required, 2 factors critical to clinical use. The NanoString equipment is an FDA approved clinical grade machine on which this laboratory developed panel was developed. It takes 48 hours to complete and costs significantly less than other methods of fusion detection. Importantly, no library preparation or complex downstream analysis is required. Alternately, OncoPanel requires a library-preparation step, yielding turnaround times of 2 weeks (51). Further, the analysis of these platforms is extensive, furthering the required time for a conclusive mutation call (51). Lastly, the cost of OncoPanel and OncoCopy are in excess of the $100/sample pricing (reagent costs) associated with the NanoString panel, although exact costs vary based on the test provider. To show the utility of the platform, we profiled patient samples from the extensive cohort of plggs treated at the Hospital for Sick Children. This allowed us to test the prevalence of fusion events across a broad spectrum of low-grade gliomas. We identified a fusion or duplication event occurring in 40% of the 429 samples successfully tested (Fig. 4A). This percentage is lower than previous reports, although most published work focuses on specific subtypes of plgg, which yield higher frequencies of fusion events (21 24, 27, 29,

8 J Neuropathol Exp Neurol Volume 76, Number 7, July , 52). The lower percentage reported here is potentially due to the inclusion of all histologic low-grade tumors within our cohort including those known to infrequently harbor fusion/duplication events (Fig. 4B). In Figure 4C, we identify the specific genetic aberrations present in plgg with respect to the fusion partners and their breakpoints. This revealed a high level of variability of the specific genetic targets matching previous reports, further substantiating the value of multiplex capabilities (19, 21, 22). The identification of driving fusion events in plgg is beneficial both diagnostically and therapeutically. It has previously been reported that fusion/duplication events are almost exclusively seen in low-grade tumors, aiding in diagnosing small specimens and highly heterogeneous samples (21 24, 27, 29, 34 36). Further, certain fusion events and breakpoints have been shown to be enriched in particular histologic subtypes, especially in cases of rarer neuroepithelial tumors, and hence may aid in defining a tumor subclass (8). Reports have also shown a location saturation of specific fusion events, and as such may aid in concisely defining a midline versus hemispheric lesion (28, 35). While surgical resection remains the mainstay of treatment for patients with plgg, in cases where complete resection is not possible, fusion status aids in providing the basis for targeted therapies. MEK inhibitors have been shown to block oncogenesis in mouse models harboring BRAF or RAF fusions (50). Further, it has been shown that BRAF fusion positive patients are resistant to standard inhibitors of BRAFV600E (53). Currently, several clinical trials are taking place to investigate the potential exploitation of the fusion events such as the SIOP-E LOGGIC phase III trial. The findings from this study support that the NanoString system is a robust platform for the identification of fusion and duplication events in plgg. NanoString offers the ability to multiplex large numbers of genes, has reasonable input material requirements, is capable of complex gene analysis, and is functional with both frozen and FFPE preserved samples. We have successfully applied the NanoString panel to a large set of plggs treated at the Hospital for Sick Children. The resultant data identified that a large portion of these samples harbored fusion or duplication events involving a variety of gene partners and fusion breakpoints. To conclude, we believe that the NanoString plgg panel is a logical alternative for clinical characterization of fusion and duplication events in plgg. The observed prevalence and variety of fusion events in plgg further support the need for modernized methods of detection, and NanoString should be highly considered for clinical implementation in the future. REFERENCES 1. Ostrom QT, Gittleman H, Farah P, et al. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in Neuro Oncol 2014;16: Armstrong GT, Conklin HM, Huang S, et al. Survival and long-term health and cognitive outcomes after low-grade glioma. Neuro-Oncology 2011;13: Rickert CH, Paulus W. Epidemiology of central nervous system tumors in childhood and adolescence based on new WHO classification. Child s Nerv Syst 2001;17: Qaddoumi I, Sultan I, Gajjar A. Outcome and prognostic features in pediatric gliomas: a review of 6212 cases from the surveillance, epidemiology, and end results database. Cancer 2009;115: Fusion/Duplication Detection in Pediatric Low-Grade Glioma 5. Shaw EG, Wisoff JH. Prospective clinical trials of intracranial low-grade glioma in adults and children. Neuro Oncol 2003;5: Gajjar A, Sandford RA, Heiderman R, et al. Low-grade astrocytoma: a decade of experience at St. Jude Children s Research Hospital. J Clin Oncol 1997;15: Jallo GI, Danish S, Velasquez L, et al. Intramedullary low-grade astrocytoma: long-term outcome following radical surgery. J Neurooncol 2001; 53: McGirt MJ, Chaichana KL, Atiba A, et al. Neurological outcome after resection of intramedullary spinal cord tumors in children. Childs Nerv Syst 2008; 24: Erkal HS, Serin M, Cakmak A. Management of optic pathway and chiasmatic-hypothalamic gliomas in children with radiation therapy. Radiother Oncol 1997;45: Grabenbauer GG, Schuchardt U, Buchfelder M, et al. Radiation therapy of optico-hypothalamic gliomas (OHG): radiographic response, vision and late toxicity. Radiother Oncol 2000;54: Merchant TE, Conklin HM, Wu S, et al. Late effects of conformal radiation therapy for pediatric patients with low-grade glioma: prospective evaluation of cognitive, endocrine, and hearing deficits. J Clin Oncol 2009;27: Krishnatry R, Zhukova N, Guerreiro Strucklin A, et al. Clinical and treatment factor determining long-term outcomes for adult survivors of childhood low-grade glioma: a population-based study. Cancer 2016;122: Packer RJ, Ater J, Allen J, et al. Carboplatin and vincristine chemotherapy for children with newly diagnosed progressive low-grade gliomas. J Neurosurg 1997;86: Reddy AT, Packer RJ. Chemotherapy for low-grade gliomas. Childs Nerv Syst 1999;15: Lassaletta A, Scheinemann K, Zelcher S, et al. Phase II weekly vinblastine for chemotherapy naïve children with progressive low-grade glioma: a Canadian Pediatric Brain tumor Consortium Study. J Clin Oncol 2016; pii: JCO Nicolin G, Parkin P, Mabbott D, et al. Natural history and outcome of optic pathway gliomas in children. Pediatr Blood Cancer 2009; 53: Opocher E, Kremer LC, Da Dalt L, et al. Prognostic factors for progression of childhood optic pathway glioma: a systematic review. Eur J Cancer 2006; 42: Pfister S, Janzarik WG, Remke M, et al. BRAF gene duplication constitutes a mechanism of MAPK pathway activation in low-grade astrocytomas. J Clin Invest 2008;118: Tatevossian RG, Lawson AR, Forshew T, et al. MAPK pathway activation and the origins of pediatric low-grade astrocytomas. J Cell Physiol 2010; 222: Jones DT, Kocialkowski S, Liu L, et al. Tandem duplication producing a novel oncogenic BRAF fusion gene defines the majority of pilocytic astrocytomas. Cancer Res 2008;68: Forshew T, Tatevossian RG, Lawson AR, et al. Activation of the ERK/ MAPK pathway: a signature genetic defect in posterior fossa pilocytic astrocytomas. J Pathol 2009;218: Cin H, Meyer C, Herr R, et al. Oncogenic FAM131B-BRAF fusion resulting from 7q34deletion comprises an alternative mechanism of MAPK pathway activation in pilocytic astrocytoma. Acta Neuopathol 2011;121: Jones DT, Kocialkowski S, Liu L, et al. Oncogenic RAF1 rearrangement and a novel BRAF mutation as alternatives to KIAA1549:BRAF fusion in activating the MAPK pathway in pilocytic astrocytoma. Oncogene 2009;28: Jacob K, Albrecht S, Sollier C, et al. Duplication of 7q34 is specific to juvenile pilocytic astrocytomas and a hallmark of cerebellar and optic pathway tumours. Br J Cancer 2009;101: Dougherty MJ, Santi M, Brose MS, et al. Activating mutations in BRAF characterize a spectrum of pediatric low-grade gliomas. Neuro Oncol 2010;12: Sievert AJ, Jackson EM, Gai X, et al. Duplication of 7q34 in pediatric low-grade astrocytomas detected by high-density single-nucleotide polymorphism-based genotype arrays results in a novel BRAF fusion gene. Brain Pathol 2009;19: Korshunov A, Meyer J, Capper D, et al. Combined molecular analysis of BRAF and IDH1 distinguishes pilocytic astrocytoma from diffuse astrocytoma. Acta Neuropathol 2009;118:

9 Ryall et al J Neuropathol Exp Neurol Volume 76, Number 7, July Faulkner C, Ellis HP, Shaw A, et al. BRAF fusion analysis is pilocytic astrocytomas: KIAA1549-BRAF 15-9 fusions are more frequent in the midline than within the cerebellum. J Neuropathol Exp Neurol 2015;74: Becker A, Scapulatempo-Neto C, Carloni A, et al. KIAA1549:BRAF gene fusion and FGFR1 hotspot mutations are prognostic factors in pilocytic astrocytomas. J Neuropathol Exp Neurol 2015;74: Hawkins C, Walker E, Mohamed N, et al. BRAF-KIAA1549 fusion predicts better clinical outcome in pediatric low-grade astrocytoma. Clin Cancer Res17: Jeuken JWM, Wesseling P. MAPK pathway activation through BRAF gene fusion in pilocytic astrocytomas: a novel oncogenic fusion gene with diagnostic, prognostic, and therapeutic potential. J Pathol 2010; 222: Whittaker S, Menard D, Kirk R, et al. A novel, selective, and efficacious nanomolar pyridopyrazinone inhibitor of V600EBRAF. Cancer Res 2010;70: Bollag G, Hirth P, Tsai J, et al. Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature 2010;467: Zhang J, Wu G, Miller C, et al. Whole-genome sequencing identifies genetic alterations in pediatric low-grade gliomas. Nat Genet 2013;45: Qaddoumi I, Orisme W, Wen J, et al. Genetic alterations in uncommon low-grade neuroepithelial tumors: BRAF, FGFR1, and MYB mutations occur at high frequency and align with morphology. Acta Neuropathol 2016;131: Jones D, Hutter B, Jager N, et al. Recurrent somatic alterations of FGFR1 and NTRK2 in pilocytic astrocytoma. Nat Genet 2013;45: Singh D, Chan J, Zoppoli P, et al. Transforming fusions of FGFR and TACC genes in human glioblastoma. Science 2012;337: Reis PP, Waldron L, Goswami RS, et al. mrna transcript quantification in archival samples using multiplexed, color-coded probes. BMC Biotechnol 2011; 11: Northcott PA, Shih DJ, Remke M, et al. Rapid, reliable, and reproducible molecular sub-grouping of clinical medulloblastoma samples. Acta Neuropathol 2012;123: Geiss G, Bumgarner R, Birditt B, et al. Direct multiplexed measurement of gene expression with color-coded probes. Nat Biotechnol 2008;26: Louis D, Perry A, Reifenberger G, et al. The 2016 World Health Organization classification of tumors of the central nervous system. Acta Neuropathol 2016;13: Tian Y, Rich B, Vena N, et al. Detection of the KIAA1549-BRAF fusion transcripts in formalin-fixed paraffin-embedded pediatric low-grade glioma. J Mol Diag 2011;13: Lin A, Rodriguez F, Karajannis M, et al. BRAF alterations in primary glial and glioneuronal neoplasms of the central nervous system with identification of 2 novel KIAA1549: BRAF fusion variants. J Neuropathol Exp Neurol 2012; 71: Dahiya S, Yu J, Kaul A, et al. Novel BRAF Alteration in a sporadic pilocytic astrocytoma. Case Resp Med 2012;2012: Ramkissoon L, Horowitz P, Craig J, et al. Genomic analysis of diffuse pediatric low-grade gliomas identifies recurrent oncogenic truncating rearrangements in the transcription factor MYBL1. Proc Natl Acad Sci USA 2013;110: Furnari F, Fenton T, Bachoo R, et al. Malignant astrocytic glioma: genetics, biology, and paths to treatment. Genes Dev 2007;21: The Cancer Genome Atlas Research Network. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008;455: Roth JJ, Santi M, Pollock AN, et al. Chromosome band 7q34 deletions resulting in KIAA1549-BRAF and FAM131B-BRAF fusions in pediatric low-grade gliomas. Brain Pathol 2015;25: Dougherty MJ, Tooke LS, Sullivan LM, et al. Clinical utilization of high-resolution single nucleotide polymorphism based oligonucleotide arrays in diagnostic studies of pediatric patients with solid tumors. Cancer Genet 2012; 205: Palanisamy N, Ateeq B, Kalyana-Sundaram S, et al. Rearrangements of the RAF kinase pathway in prostate cancer, gastric cancer and melanoma. Nat Med 2010;16: Wagle N, Berger M, Davis M, et al. High-throughput detection of actionable genomic alterations in clinical tumour samples by targeted, massively parallel sequencing. Cancer Discov 2012;2: Ramkissoon S, Bandopadhayay P, Hwang J, et al. Clinical targeted exome-based sequencing in combination with genome-wide copy number profiling: precision medicine analysis of 203 pediatric brain tumors. Neuro Oncol 2017; pii: now Sievert AJ, Lang SS, Boucher KL, et al. Paradoxical activation and RAF inhibitor resistance of BRAF protein kinase fusions characterizing pediatric astrocytomas. Proc Natl Acad Sci USA 2013; 110: