Frequent amplification of a chr19q13.41 microRNA polycistron in aggressive primitive neuroectodermal brain tumors

Abstract

We discovered a high-level amplicon involving the chr19q13.41 microRNA (miRNA) cluster (C19MC) in 11/45 ( approximately 25%) primary CNS-PNET, which results in striking overexpression of miR-517c and 520g. Constitutive expression of miR-517c or 520g promotes in vitro and in vivo oncogenicity, modulates cell survival, and robustly enhances growth of untransformed human neural stem cells (hNSCs) in part by upregulating WNT pathway signaling and restricting differentiation of hNSCs. Remarkably, the C19MC amplicon, which is very rare in other brain tumors (1/263), identifies an aggressive subgroup of CNS-PNET with distinct gene-expression profiles, characteristic histology, and dismal survival. Our data implicate miR-517c and 520g as oncogenes and promising biological markers for CNS-PNET and provide important insights into oncogenic properties of the C19MC locus.

Figure 1. Frequent chr2 gains and chr19q13.41 amplication in CNS-PNET

Global DNA copy number profiles of 39 CNS-PNET were generated using the Affymetrix 500K SNP arrays (Fig S1). Fig A and B shows inferred DNA copy numbers as determined using dChip for SNPs on chromosomes 2 and 19 with DNA gains and losses indicated by shades of red. (A) Expanded copy number profiles of chr2 showing frequent whole chr2 gains in tumors; amplicons encompassing MYCN and candidate oncogenes, CREB1 and FZD5, in 2 individual tumors are shown on the right. (B) Expanded view illustrating overlapping boundaries of the chr19q13.41 amplicon which is flanked by the DPRX and LILRP2 loci in 8/39 tumors (red labels). Copy number plot of an individual tumor is shown on the right. (C) Inferred copy number plots generated using dChip pair-matched analysis reveals the chr19q13.41 amplicon in tumor but not matched normal lymphocyte (PBL) DNA of PNET54. (D) Fluorescence in situ hybridization confirms high level chr19q13.41 DNA copy number gains and amplification respectively in tumors PNET3 and PNET5, 6, as indicated by genotyping analysis.

Figure 2. MicroRNAs are the oncogenic target of the chr19q13.41 amplicon

(A) Genomic organization of the chr19q13.41 amplicon, which maps between the DPRX and LILRP2 loci, is schematized. Relative map positions of the chr19 microRNA cluster (C19MC) (red), the MIR371-373 cluster (blue) and select coding genes within the amplicon boundaries is shown. * mark candidate oncogenic MIR-520g and 517c; MIR-512-1 and MIR-516-1a mark the limits of the locus. BAC clone RP11-381E3, used for FISH analysis is indicated in green. All map positions are based on the hg18 build from the UCSC Genome Database (http://genome.ucsc.edu/). Lower panel shows a DNA copy number plot of SNPs indicating overlapping boundaries of the chr19q13.41 amplicon in 9 tumors. (B) Confirmation of chr19q13.41 amplicon boundaries in 9 tumors by qPCR copy number analysis of select loci. Result is representative of 2 independent experiments. (C) Heat map showing expression levels of 34 coding genes on chr19q13.41 as determined from BeadChiP expression analysis of 33 tumors; q-values indicate significance of gene expression changes relative to chr19q13.41 amplification, * indicates the CNOT3 locus which showed a modest 2-fold gene expression change in chr19q13.41 amplified tumors, and ‡ indicates select genes validated by qRT-PCR (Fig S2A). (D) TaqMan® MicroRNA array analysis of 7 tumors with chr19q13.41 amplification indicate high expression of several C19MC miRNAs (miR-512-3p, 517a, 520g, 517c,,519a) but not miR371-373 in tumors relative to fetal brain (Fig S2B-D). (E) Box and whisker summary plot showing expression levels of 10 select C19MC miRNA relative to chr19q13.41 amplification in tumors; qRT-PCR analysis of 8 amplified and 15 non-amplified tumors is shown. Paired boxes represent the 25-75^th centile expression values for each miRNA in tumors without (−) and with (+) the C19MC amplicon. Horizontal bar indicate median while vertical lines indicate maximum and minimum miRNA expression values for each tumor group; * denote significant correlation of miRNA expression levels with C19MC amplification ; error bars represent SEM n≥ 2 with 3 replicas. (F) Gene specific qRT-PCR confirms high expression of select C19MC miRNAs in chr19q13.41 amplified tumors relative to normal fetal brain (Fig S2); miR-520g and 517c had highest tumor-specific expression (red arrows) in seven tumors analysed, error bars represent SD of n≥ 2 experiments (3 replicas/experiment); miRNA levels shown in Figs D-F were determined relative to RNU6B control.

Figure 3. MiR-520g and 517c have oncogenic effects in vitro and in vivo

(A) Quantitative RT-PCR confirmation of stable miR-517c or 520g expression in PFSK, Daoy, and NIH3T3 cell lines; miRNA levels are relative to RNU6B. Results are shown as means ± SD.; n ≥2 3. Cell line pools were used for experiments shown in Figs B-D. (B) Growth curves of miR-520g or 517c expressing PFSK, Daoy, and NIH3T3 stable cell lines indicate significant growth advantage in miR-517c (upper panel) but not miR-520g expressing (lower panel) cells in 10% serum conditions. Data represent 3 independent experiments with 3 replicas/data point, error bars = SD. (C) MicroRNA 520g and 517c expression significantly enhances colony formation in Daoy but not PFSK or NIH3T3 cells; 2 independent soft agar growth assays with 3 replicas/data point are summarized, error bars denote SD. (D) Control and miR-520g or 517c expressing Daoy and NIH3T3 stable cell lines were injected into contralateral flanks of nu/nu mice as described in methods. Histogram represents tumor weights from 8 mice determined 6 weeks post-injections; error bars represent SEM. Right panel shows representative mice with Daoy or NIH3T3 miR-520g or 517c associated xenografts (red arrow); *indicates control injection sites.

Figure 4. C19MC amplification and constitutive miR-517c and 520g expression is associated with a survival phenotype

(A) Expression profiles demonstrate significant enrichment of self renewal and survival genes in C19MC amplified tumors. Heat map shows up (red arrow) and down (green arrow) regulated genes identified using a moderated t-test statistic adjusted for multiple testing (FDR<0.05) and analysed for functional enrichment using GOSTAT (Table S5, S6). Significance and magnitude of gene expression changes between the 2 tumor groups are denoted respectively by p-value and fold change. (B-C) PFSK, Daoy and NIH3T3 cells with stable miR-520g or 517c expression were exposed to 0.1% serum and assayed for cell viability or apoptosis relative to similarly treated control cells. Fig B show a representative plot of relative cell viability at 72hrs determined by trypan-blue staining, error bars represent SD; n=3 with 3 replicas. Fig C shows results of TUNEL assays in miR-520g and 517c stable and control lines at 72hrs after serum starvation; 3 experiments (2 replicates/ data point) are summarized; error bars represent SD, % apoptosis equal ratio of TUNEL to DAPI positive cells.

Figure 5. MiR-520g and 517c expression drives growth and inhibits differentiation of human neural stem cells

(A) Quantitative RT-PCR confirmation of stable miR-520g or 517c expression in hNSCs. Results are shown as means ± SD.; n = 3; cell pools were used for all experiments. (B) hNSCs-miR-520g and hNSCs-517c cell lines were exposed to differentiation media (DMEM +10% FBS) and monitored for changes in confluence with a live cell imaging system. Photos on day 8 post-differentiation (bottom panel), shows no differences in cell mass, therefore cell confluence was used to estimate cell numbers. In contrast to control cells, hNSC with stable miR-520g, and 517c expression continued to grow in differentiation media. (C) FACS analysis of stable hNSC-miR-520g and 517c and control pcDNA cell lines stained for neuronal (Tubulin βIII) and astrocytic (GFAP) lineage markers after growth in proliferative and differentiation media. GFAP expression was not consistently altered in miR-520g and 517c expressing cells (data not shown); Tubulin βIII expression (red boxes and Fig D) was significantly diminished only in hNSCs-miR-520g cells grown in differentiation media. (D) and (E) Quantitative RT-PCR analysis of cell lineage and stem cell markers in stable hNSC-miR-520g, 517c and control hNSC-pcDNA cells grown under proliferative and differentiation conditions. Relative mRNA levels are normalized to 36B4 controls; * denotes p<0.05. Results shown are means ± SD.; n=2-3, with 2 replicas/experiment.

Figure 6. MiR-520g expression correlates with altered WNT signaling

(A) Summary of pathway enrichment analysis performed on gene expression profiles of hNSCs (left panel) and PFSK (right panel) with stable miR-520g expression. Expression profiles of stable cell lines generated using Illumina BeadChips were compared to identify genes most significantly correlated with miR-520g expression in hNSC or PFSK cells. The miR-520g associated gene sets were then analysed using the KEGG pathway database to determine significance of enrichment [−log10 (p-value)] for specific developmental signaling pathways (bars) (Table S7). Number of genes in each pathway that meet the p-value threshold [-log10 (p< 0.05), dashed line] is indicated. (B) and (C) Quantitative RT-PCR analysis confirms WNT pathway dysregulation in miR-520g expressing stable hNSC and PFSK cell lines. In hNSCs (Fig B) and PFSK (Fig C) cells, miR-520g expression correlated with upregulation of WNT ligand, WNT5A and WNT inhibitors - SFRP2/3 and DKK1. CCND1 and MYCC, expression was diminished in miR-520g expressing cells but did not correlate with canonical WNT reporter activity (Fig S5). Results shown are means ± SD with n=2, 2 replicas/experiment; mRNA levels are normalized to 36B4; * denotes p< 0.05. (D) Total lysates from control and stable PFSK-miR-520g cell lines were immunoblotted with phospho-specific antibodies for PKC (pPKC), CAMKII (pCAMKII), JNK1 & 2 (pJNK1, pJNK2) and antibodies for total JNK1&2; actin served as loading control. Arrows indicate higher levels of pJNK2, but not pJNK1, observed consistently in PFSK-miR-520g cells relative to controls without changes in total JNK1/2 levels; a representative of 4 independent experiments is shown.

Figure 7. Amplification of the C19MC locus identifies CNS-PNET with poor survival and distinct tumor pathology

(A) Kaplan-Meier survival estimates for patients stratified by C19MC amplification status in tumors; log-rank analysis correlate C19MC amplification with poorer overall survival. (B) Representative micrograph illustrating variant and classic CNS-PNET histology and corresponding frequency of C19MC amplication. Variant histology distinguished by ependymal or ependymoblastic differentiation and distinct rosette structures (arrows) correlate significantly with C19MC amplication. (C) Unsupervised hierarchical cluster analysis of expression profiles for 33 CNS-PNET is shown in relation to C19MC amplicon status and clinical characteristics of tumor (Table S1). Red and white boxes respectively denote presence or absence of clinical feature, grey boxes denote incomplete information. Cluster analysis show co-segregation (red cluster) and significant correlation of tumors with C19MC amplification or increased C19MC miRNA expression (Fig S7) with young age (<4yrs) and variant histology, but not with tumor metastasis. (D) Tabular summary indicating rarity of C19MC amplification in other pediatric brain tumors of different histologies, C19MC amplicon status was determined by FISH on tissue microarrays and/or SNP genotyping (CNS-PNET only).