Long Noncoding RNA pancEts-1 Promotes Neuroblastoma Progression through hnRNPK-Mediated β-Catenin Stabilization Free

Neuroblastoma is a malignant tumor arising from primitive neural crest, and accounts for more than 7% of malignancies and around 15% of cancer-related mortality in childhood (1). The clinical features of neuroblastoma are heterogeneous, ranging from spontaneous regression to rapid progression or resistance to multimodal therapy (1). For high-risk neuroblastoma patients, tumor invasion and metastasis are the main causes of death (1). Although various genetic abnormalities, including 1p or 11q deletion, 17q gain, and MYCN amplification, have been proposed as indicators of poor prognosis (2), the mechanisms underlying the progression of neuroblastoma still remain to be determined to improve the therapeutic efficiency.

Recent evidence has shown the crucial roles of long noncoding RNAs (lncRNA) in the tumorigenesis and aggressiveness of neuroblastoma (3). For example, lncRNA cyclin-dependent kinase inhibitor 2A/ARF intron 2 (CAI2) contributes to the paradoxical overexpression of p16 in neuroblastoma, and serves as a useful biomarker of high-risk neuroblastoma (4). Linc00467, a novel lncRNA negatively regulated by MYCN, promotes the survival of neuroblastoma cells through reducing the activity of dickkopf-related protein 1 (5). In a model of human metastatic neuroblastoma, homeobox D (HOXD) antisense transcript 1 controls the retinoid acid-induced cell differentiation (6). High levels of noncoding RNA expressed in aggressive neuroblastoma (ncRAN) are associated with poor outcome of neuroblastoma patients, and knockdown of ncRAN inhibits the growth of neuroblastoma cells (7). Our previous studies show that MYCN opposite strand (MYCNOS) cooperates with CCCTC-binding factor to promote neuroblastoma progression through facilitating MYCN expression (8). Thus, it is currently necessary to further investigate the roles of lncRNAs in neuroblastoma progression.

In the current study, through mining of a public microarray dataset, we identified Ets-1 promoter–associated noncoding RNA (pancEts-1) as a novel 1395-nucleotides (nt) lncRNA associated with poor outcome of neuroblastoma. We demonstrate that pancEts-1 is upregulated in neuroblastoma tissues and cell lines. In addition, pancEts-1 binds to heterogeneous nuclear ribonucleoprotein K (hnRNPK) to facilitate its interaction with β-catenin, resulting in increased stability and transactivation of β-catenin, transcriptional alteration of downstream genes, and promotion of the growth, invasion, and metastasis of neuroblastoma cells in vitro and in vivo, indicating the crucial roles of the pancEts-1/hnRNPK/β-catenin axis in neuroblastoma progression.

Human MYCN-amplified [NB-1643, SK-N-BE(2), NB-1691, IMR32, and BE(2)-C] and non–MYCN-amplified (SK-N-AS, SH-SY5Y, and SK-N-SH) neuroblastoma cell lines were obtained from Type Culture Collection of Chinese Academy of Sciences (Shanghai, China) and ATCC. Cell lines were authenticated by short tandem repeat profiling, and used within 6 months after resuscitation of frozen aliquots. Mycoplasma contamination was regularly examined using the Lookout Mycoplasma PCR Detection Kit (Sigma). Cells were grown in RPMI-1640 medium containing 10% fetal bovine serum (Life Technologies, Inc.), incubated in serum-free RPMI-1640 medium for 4 hours, and treated with cycloheximide or MG132 (Sigma) as indicated.

The 138-bp probe was generated by PCR DIG Probe Synthesis Kit (Roche) with primer set indicated in Supplementary Table S1 and Supplementary Fig. S1A. Northern blot was performed as previously described (9).

Gene set enrichment analysis was performed as previously described (10). The published gene sets were used as indicated. Datasets were generated from published microarray (GSE16476) results.

To characterize the 5′ and 3′ ends of pancEts-1, total RNA extracted from BE(2)-C cells was used to generate Rapid amplification of cDNA ends (RACE)-ready cDNA, and PCR was performed using the SMARTer RACE cDNA Amplification Kit (Clontech). The cDNA ends were amplified with universal and gene-specific primers (Supplementary Table S1).

Cells were seeded on coverslips and fixed with 4% paraformaldehyde. Biotin-labeled antisense and sense RNA probes for pancEts-1 were in vitro transcribed with the Biotin RNA Labeling Mix (Roche) and T7 RNA polymerase. Hybridization was undertaken in a humidified chamber at 37°C for 16 hours, with or without RNase A (20 μg) treatment. Cells were incubated with streptavidin-conjugated Cy3, and counterstained with 4′,6-diamidino-2-phenylindole (DAPI).

Human pancEts-1 cDNA (1383 bp), hnRNPK cDNA (1395 bp), and CTNNB1 cDNA (2346 bp) were amplified from neuroblastoma tissue (Supplementary Table S2) or kindly provided by Dr. Ralf Janknecht (11). Their truncations were amplified with primer sets (Supplementary Table S2, Supplementary Fig. S1A–S1C), and subcloned into pcDNA3.1 (Invitrogen), pCMV-3Tag-1A (Addgene), or pCMV-N-Myc (Beyotime Biotechnology), respectively. Oligonucleotides encoding short hairpin RNAs (shRNA) specific for pancEts-1 or hnRNPK (Supplementary Table S2) were subcloned into GV102 (Genechem Co., Ltd.). Stable cell lines were screened by administration of neomycin or puromycin (Invitrogen). Empty vector and scramble shRNA (sh-Scb) were applied as controls (Supplementary Table S2).

Total RNA of 1 × 10⁶ cells was isolated using TRizol reagent (Life Technologies, Inc.). Library preparation and transcriptome sequencing on an Illumina HiSeq X Ten platform were performed by Novogene Bioinformatics Technology Co., Ltd. to generate 100-bp paired-end reads. HTSeq v0.6.0 was used to count the read numbers mapped to each gene, and fragments per kilobase of transcript per million fragments mapped (FPKM) of each gene were calculated. Sequencing data have been deposited in Gene Expression Omnibus database (accession code GSE104950).

Nuclear and cytoplasmic RNA was isolated using RNA Subcellular Isolation Kit (Active Motif). Total RNA was isolated with RNeasy Mini Kit (Qiagen Inc.). Reverse transcription and real-time PCR were performed with Transcriptor First Strand cDNA Synthesis Kit (Roche), SYBR Green PCR Master Mix (Applied Biosystems) and primers indicated in Supplementary Table S1 and Supplementary Fig. S1A. The transcript levels were analyzed by 2^−ΔΔCt method.

Tissue or cellular protein was extracted with 1 × cell lysis buffer (Promega). Western blot analysis was performed as previously described (8, 12), with antibodies specific for hnRNPK (ab39975), β-catenin (ab32572), discoidin domain receptor tyrosine kinase 2 (DDR2, ab63337), discs large 1 (DLG1, ab60551), v-ets erythroblastosis virus E26 oncogene homolog 1 (Ets-1, ab26096), protocadherin 7 (PCDH7, ab139274), polyribonucleotide nucleotidyltransferase 1 (PNPT1, ab157109), FLAG (ab45766), Myc (ab9106, Abcam Inc.), glutathione S-transferase (GST, sc-33614), H3 histone (sc-10809), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH, sc-47724, Santa Cruz Biotechnology).

Coimmunoprecipitation (co-IP) was performed as previously described (13), with antibodies specific for hnRNPK (ab39975), β-catenin (ab32572), or FLAG (ab45766, Abcam Inc.). Bead-bound proteins were released and analyzed by Western blot analysis.

Tumor cells were plated on coverslips, and incubated at 4°C overnight with antibodies specific for hnRNPK (ab39975, Abcam Inc.; 1:100 dilution), or β-catenin (ab32572, Abcam Inc.; 1:100 dilution). Then, cells were incubated with Alexa Fluor 594 goat anti-rabbit immunoglobulin G (IgG, 1:1,000 dilution) and stained with DAPI (300 nmol/L; ref. 8).

The TOP-FLASH and FOP-FLASH reporter plasmids were obtained from Millipore. The luciferase reporters of E2F Transcription factor 1 (E2F1), nuclear factor kappa B (NF-κB), STAT3, and activator protein 1 (AP1) were obtained from Qiagen Inc. Dual-luciferase assay was performed as previously described (8, 13). Relative β-catenin activation was determined by the TOP-FLASH/FOP-FLASH ratio.

To restore the target gene expression induced by pancEts-1 overexpression, shRNA specific for hnRNPK (Supplementary Table S2) was transfected into tumor cells with Genesilencer Transfection Reagent (Genlantis). Empty vector and sh-Scb were applied as controls (Supplementary Table S2).

Biotin-labeled RNA probes for pancEts-1 truncates were in vitro transcribed as described above. Biotin-labeled RNA pull-down was performed as previously described (8). The retrieved protein was detected by Western blot or mass spectrometry analyses at Wuhan Institute of Biotechnology.

Cells were ultraviolet light crosslinked at 254 nm (200 J/cm²) and collected by scraping (8). RNA immunoprecipitation (RIP) assay was performed using Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore), with two antibodies specific for hnRNPK (ab39975, sc-28380, Abcam Inc. and Santa Cruz Biotechnology). Coprecipitated RNAs were detected by RT-PCR with specific primers (Supplementary Table S1; Supplementary Fig. S1A). Total RNAs (input) and isotype antibody (IgG) were applied as controls.

A series of hnRNPK truncates were amplified from neuroblastoma tissues (Supplementary Table S2; Supplementary Fig. S1B), subcloned into pGEX-6P-1 (Addgene), and transformed into E. coli to produce GST-tagged truncated hnRNPK protein (8). The pancEts-1 cRNA was in vitro transcribed with TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific, Inc.). The hnRNPK–RNA complexes were pulled down using GST beads (Sigma). Protein was detected by SDS-PAGE and Western blot analysis, whereas RNA was measured by RT-PCR with specific primers (Supplementary Table S1; Supplementary Fig. S1A).

Biotin-labeled RNA probes for 691–828 bp of pancEts-1 truncates were prepared as described above. RNA electrophoretic mobility shift assay (EMSA) using nuclear extracts or recombinant hnRNPK protein was performed according to the instructions of LightShift Chemiluminescent RNA EMSA Kit (Thermo Fisher Scientific, Inc.).

Chromatin immunoprecipitation (ChIP) assay was performed according to the manufacturer's instructions of EZ-ChIP kit (Upstate Biotechnology; refs. 8, 13). Real-time quantitative PCR (qPCR) was performed with SYBR Green PCR Master Mix (Applied Biosystems) and primers targeting gene promoters (Supplementary Table S1).

The MTT (Sigma) colorimetric (13), soft agar (8, 12), and Matrigel invasion (8, 12, 14) assays for measuring the viability, growth, and invasion capability of tumor cells were conducted as previously described.

All animal experiments were carried out in accordance with NIH Guidelines for the Care and Use of Laboratory Animals, and approved by the Animal Care Committee of Tongji Medical College (approval number: Y20080290). The in vivo tumor growth studies (1 × 10⁶ tumor cells per mouse, n = 5 per group) and experimental metastasis (0.4 × 10⁶ tumor cells per mouse, n = 5 per group) studies were performed with blindly randomized four-week-old male BALB/c nude mice as previously described (8, 12, 14).

The Institutional Review Board of Tongji Medical College approved the human tissue study (approval number: 2011-S085). All procedures were carried out in accordance with the guidelines set forth by Declaration of Helsinki. Written informed consent was obtained from all legal guardians of the patients. All cases received no preoperative chemotherapy or other treatment. Human embryonic tissues or normal dorsal root ganglia were obtained from therapeutic abortions (at day 50 of gestation) and interrupted pregnancies, respectively. Fresh tumor specimens were collected at surgery and stored at −80°C until use. Total RNAs of normal human dorsal ganglia pooling from 21 male/female Caucasians were obtained from Clontech.

IHC staining was performed as previously described (8, 12, 14), with antibody specific for hnRNPK (ab39975, Abcam Inc.; 1:200 dilution). Ten different high-power fields (×400) for each specimen were blindly evaluated, and staining intensity and percentage of positive cells were scored on a scale from 0 to 3 (0, negative; 1, weakly positive; 2, moderately positive; 3, strongly positive) and from 0 to 4 (0, negative; 1, positive in 1%–25%; 2, positive in 26%–50%; 3, positive in 51%–75%; 4, positive in 76%–100%), respectively. On the basis of the products of staining intensity multiplied by percentage of positive cells, the results of IHC were classified into negative (−, 0–1), mild (+, 2–3), moderate (++, 4–8), and strongly positive (+++, 9–12).

All data were shown as mean ± SD. The cut-off values were determined by average gene-expression levels. The Student t test, analysis of variance, χ² analysis, and Fisher exact probability analysis were applied to compare the difference in tumor cells or tissues. The Fisher exact test was applied to analyze the statistical significance of overlap between two gene lists. Pearson correlation coefficient was applied for analyzing the relationship among gene expression. The log-rank test and Cox regression models were used to assess survival difference and HRs. All statistical tests were two-sided.

To investigate the lncRNA crucial for neuroblastoma progression, we first analyzed the microarray dataset (GSE16476; ref. 15) of 88 patients derived from Gene Expression Omnibus (GEO). We identified 93, 82, 271, and 167 lncRNAs that were differentially expressed (fold change >1.3, P < 0.05) in neuroblastoma specimens with varied status of death, clinical progression, and International Neuroblastoma Staging System (INSS) stages (stage 1 vs. 4 and stage 4S vs. 4), respectively (Fig. 1A). On the basis of overlapping analysis of these lncRNAs (P < 0.001), 9 lncRNAs were found to be consistently associated with death, clinical progression, and advanced INSS stages of neuroblastoma (Fig. 1A). Log-rank test analysis revealed top four lncRNA-predicting prognosis of neuroblastoma patients (Fig. 1A). Among them, the functions of three lncRNAs MYCNOS (8), DLX6 antisense RNA 1 (DLX6-AS1; ref. 16), and firre intergenic repeating RNA element (FIRRE; ref. 17) have been reported, whereas the roles of pancEts-1 (also known as uncharacterized LOC101929517) in the tumorigenesis and aggressiveness still remain elusive. Kaplan–Meier curves of these 88 well-defined neuroblastoma cases (GSE16476) showed highly significant difference in overall survival (P = 5.7 × 10⁻³) between the high and low pancEts-1 expression groups (Fig. 1B), which was in line with our findings in 42 neuroblastoma patients (Fig. 1B). Mining of a publicly available microarray dataset (GSE16476) revealed that pancEts-1 was highly expressed in neuroblastoma tissues with death (P = 0.0195), progression (P = 0.0414), or advanced INSS stages (P = 0.0208 and P = 0.0221, Fig. 1C). We validated higher pancEts-1 transcript levels in an independent cohort of 42 primary neuroblastoma tumors (Supplementary Table S3), than those in normal dorsal ganglia (P < 0.0001, Fig. 1D), especially in neuroblastoma cases with poor differentiation (P < 0.0001) or advanced INSS stages (P = 0.0024), but not with MYCN amplification (P = 0.3477, Fig. 1D). Real-time qRT-PCR indicated higher pancEts-1 expression in cultured neuroblastoma cell lines, especially in NB-1643, SK-N-BE(2), NB-1691, IMR32, and BE(2)-C, than that in normal dorsal ganglia (Fig. 1E). Gene set enrichment analysis (GSEA) on all genes correlated to pancEts-1 in 88 neuroblastoma specimens (GSE16476) yielded significant association with cancer metastasis gene signature [normalized enrichment score (NES) = 1.625, normalized P = 0.009; Fig. 1F]. These data indicated that pancEts-1, a novel lncRNA, was an independent prognostic marker for neuroblastoma progression.

The lncRNA pancEts-1, consisting of four exons, is located at chromosome 11q24.3. Northern blot analysis confirmed the existence of 1.3-kb pancEts-1 transcript in BE(2)-C cells (Fig. 2A). The 5′- and 3′-RACE analyses (Supplementary Fig. S2A) revealed that pancEts-1 was 1395-nt in length and polyadenylated (GenBank Accession No. KP742344). RNA-FISH and subcellular fractionation assays revealed nuclear and cytoplasmic localization and enrichment of pancEts-1 in BE(2)-C cells (Fig. 2B). High pancEts-1 levels were noted in the brain, lung, spleen, kidney, and muscle tissues of human embryos (at day 50 of gestation, Supplementary Fig. S2B). The Coding Potential Assessment Tool (CPAT; ref. 18) indicated a low value for pancEts-1 (coding probability = 0.0103). Ribosome profiling data (19) also indicated low protein-coding potential of pancEts-1 (Supplementary Fig. S2C). Mining of publicly available datasets (54 studies) derived from cBioPortal for Cancer Genomics (http://cbioportal.org) indicated low frequency of mutation, deletion, or amplification of pancEts-1 gene in most of human cancers (Supplementary Fig. S2D).

We further investigated the functional impact of pancEts-1 knockdown on neuroblastoma cells representing high expression levels. Depletion of pancEts-1 using two independent shRNAs, sh-pancEts-1 #1 and sh-pancEts-1 #2, in the BE(2)-C and IMR32 cell lines resulted in a dramatic decrease in the number of viable cells (Fig. 2C and D). In addition, stable knockdown of pancEts-1 led to decrease in invasiveness of BE(2)-C and IMR32 cells in vitro (Fig. 2E). Consistent with these findings, we observed a significant decrease in the growth and weight of xenografts formed by subcutaneous injection of neuroblastoma cells stably transfected with pancEts-1 specific shRNAs into nude mice (Fig. 2F). In experimental metastasis assay, athymic nude mice treated with tail vein injection of BE(2)-C cells stably transfected with sh-pancEts-1 #1 or sh-pancEts-1 #2 displayed less lung metastatic colonies and greater survival probability (Fig. 2G). Together, these findings demonstrated that knockdown of pancEts-1 suppressed the progression of neuroblastoma.

To identify the protein partner of pancEts-1, we performed the biotin-labeled RNA pull-down followed by a proteomic analysis of the RNA-associated protein complex in BE(2)-C cells. Mass spectrometry revealed that hnRNPK, a nucleocytoplasmic shuttling protein (20), was the protein with highest spectral counts (with 87 detected peptides) pulled down by biotin-labeled pancEts-1 (Fig. 3A; Supplementary Table S4). Western blot further validated that hnRNPK was readily detected in the pancEts-1 RNA pull-down complex, but not in the control samples pulled-down by pancEts-1 antisense RNA or beads only (Fig. 3A). RIP assay demonstrated an endogenous interaction between pancEts-1 and hnRNPK in BE(2)-C cells (Fig. 3B). In addition, deletion-mapping analyses indicated that exon 2 of pancEts-1, especially the 691–1060 nt region, was essential for its interaction with hnRNPK protein (Fig. 3B). In vitro binding assay indicated that K homology 2 (KH2, 106–210 amino acids), but not KH1 (1–105 amino acids), K interactive (KI, 211–329 amino acids), or KH3 (330–464 amino acids) domain, of GST-tagged hnRNPK protein was crucial for the interaction with pancEts-1 (Fig. 3C). Moreover, after incubation of cellular nuclear extracts with biotin-labeled pancEts-1 truncates, RNA pull-down assay indicated that exon 2 (especially the 691–1060 nt region) of pancEts-1 was able to interact with hnRNPK protein (Fig. 3D and E). Consistently, RNA EMSA using biotin-labeled probes (corresponding to exon 2) indicated the ability of endogenous or recombinant hnRNPK protein to interact with pancEts-1, and the formed complex was completely ablated upon competition with an excess of unlabeled homologous probe (Fig. 3F). These results confirmed the specific binding of pancEts-1 to hnRNPK protein in neuroblastoma cells.

To identify the putative targets of pancEts-1, we observed that pancEts-1 induced differentially expressed genes by RNA-seq analysis in SH-SY5Y cells. There were 1912 genes, including 1,035 upregulated and 877 downregulated ones, that showed differential expression (fold change > 2.0, P < 0.01, FDR < 0.05) upon pancEts-1 over-expression (Fig. 4A). Through overlapping (P < 0.001) with 657 positively and 514 negatively correlated genes (P < 0.05) in 88 neuroblastoma cases derived from a public microarray dataset (GSE16476), 255 genes were noted by consistent correlation tendency with pancEts-1 levels (Fig. 4A). Further analysis of these genes using ChIP-X database (21) revealed top 5 potential transcriptional regulators ranking by target gene numbers (Fig. 4A), including β-catenin, E2F1, NF-κB, STAT3, and AP1. Dual-luciferase reporter assay indicated that stable overexpression or knockdown of pancEts-1 altered the activity of β-catenin, but not of E2F1, NF-κB, STAT3, or AP1, in cultured neuroblastoma cells (Fig. 4B; Supplementary Fig. S3A and S3B). Notably, there were 48 target genes of β-catenin that were correlated with pancEts-1 levels (Fig. 4A). Among these target genes, the expression of DDR2, DLG1, Ets-1, PCDH7, and PNPT1 was significantly correlated with the pancEts-1 levels in neuroblastoma tissues (Supplementary Fig. S3C and S3D) and associated with the outcome of patients (Supplementary Fig. S3E). Importantly, stable overexpression or knockdown of pancEts-1 altered the transcript levels of these genes (Fig. 4C), and increased and decreased the nuclear translocation of β-catenin in neuroblastoma cells (Fig. 4D and E), respectively. Notably, ectopic expression or knockdown of pancEts-1 increased and decreased the protein levels, but not transcript levels, of β-catenin in neuroblastoma cells, respectively (Fig. 4F; Supplementary Fig. S4A). Pre-treatment of SH-SY5Y with cycloheximide, an established inhibitor of protein synthesis, prevented the increased β-catenin levels induced by ectopic expression of pancEts-1 (Fig. 4F). In addition, incubation of BE(2)-C cells with MG132, the proteasome inhibitor, prevented the degradation of β-catenin protein induced by knockdown of pancEts-1 (Fig. 4F). These results suggested that pancEts-1 regulated the β-catenin target gene expression by stabilizing β-catenin protein in neuroblastoma cells.

To identify the protein partner of hnRNPK, antibody pull-down was performed with subsequent proteomic analysis of hnRNPK-associated protein complex in BE(2)-C cells stably transfected with sh-Scb or sh-pancEts-1 #2. Mass spectrometry revealed seven differential hnPRNK-interacting protein in BE(2)-C cells upon knockdown of pancEts-1, whereas β-catenin was the only transcriptional regulator (Fig. 5A; Supplementary Table S4). Co-IP and Western blot analysis indicated the endogenous physical interaction between hnRNPK and β-catenin in BE(2)-C and SH-SY5Y cells (Fig. 5A), and the KI (211–329 amino acids), but not KH1 (1–105 amino acids), KH2 (106–210 amino acids) or KH3 (330–464 amino acids) domain, of hnRNPK protein was essential for its interaction with β-catenin (Fig. 5A). Meanwhile, armadillo (ARM) domain of Myc-tagged β-catenin protein was crucial for the interaction with hnRNPK (Fig. 5A). Ectopic expression or knockdown of hnRNPK increased and decreased the β-catenin levels in SH-SY5Y and BE(2)-C cells, which was abolished by pretreatment with cycloheximide and MG132, respectively (Fig. 5B; Supplementary Fig. S4B). Then, we investigated the roles of pancEts-1 in hnRNPK-induced β-catenin transactivation, and found that ectopic expression or knockdown of pancEts-1 facilitated and attenuated the interaction and colocalization between hnRNPK and β-catenin in neuroblastoma cells, respectively (Fig. 5C and D). Notably, stable knockdown or ectopic expression of hnRNPK into SH-SY5Y and BE(2)-C cells resulted in decrease and increase of nuclear translocation and activity of β-catenin, which was abolished by over-expression or knockdown of pancEts-1, respectively (Fig. 5E and F). Collectively, these data demonstrated that pancEts-1 facilitated the hnRNPK-mediated stability and transactivation of β-catenin in neuroblastoma cells.

To further investigate the functional interplay of pancEts-1 and hnRNPK during the aggressiveness of neuroblastoma cells, we performed rescue studies in cultured SH-SY5Y and BE(2)-C cells with moderate pancEts-1 levels. As shown in Fig. 6A and Supplementary Fig. S5A, the enrichment of β-catenin and lymphoid enhancer binding factor 1 (LEF1) on the promoters of downstream genes DDR2, DLG1, Ets-1, PCDH7, and PNPT1 was significantly increased and decreased in neuroblastoma cells stably transfected with pancEts-1 or sh-pancEts-1 #2, which was rescued by knockdown and ectopic expression of hnRNPK, respectively. In addition, the expression of these β-catenin target genes was correspondingly altered in these neuroblastoma cells (Fig. 6B and C; Supplementary Fig. S5B and S5C). In soft agar and Matrigel invasion assays, stable over-expression of pancEts-1 facilitated the anchorage-independent growth and invasiveness of neuroblastoma cells (Fig. 6D and E; Supplementary Fig. S5D and S5E). Transfection of sh-hnRNPK #1 prevented the neuroblastoma cells from their changes in growth and invasion induced by stable transfection of pancEts-1 (Fig. 6D and E; Supplementary Fig. S5D and S5E). In mouse xenograft tumor assay, stable transfection of pancEts-1 into SH-SY5Y cells resulted in increased growth and tumor weight of subcutaneous xenograft tumors (Fig. 6F), and led to statistically more lung metastatic colonies and lower survival probability in athymic nude mice (Fig. 6F). Meanwhile, knockdown of hnRNPK attenuated the impacts of pancEts-1 on tumor growth and metastasis of neuroblastoma cells in vivo (Fig. 6F). These results suggested that pancEts-1 harbored oncogenic properties through its interplay with hnRNPK in vitro and in vivo.

To investigate the expression of hnRNPK in neuroblastoma, fresh tissues and paraffin-embedded sections from 42 primary cases were collected. Immunohistochemical staining revealed that hnRNPK was expressed in the nuclei and cytoplasm of tumor cells (Fig. 7A). The hnRNPK expression was detected in 31/42 (73.8%) neuroblastoma cases, and higher in those with poor differentiation (P = 0.031), higher mitosis karyorrhexis index (MKI, P = 0.016), or advanced INSS stages (P = 0.006, Supplementary Table S5). Higher expression of hnRNPK was detected in neuroblastoma tissues and cell lines, than that in normal dorsal ganglia (Fig. 7B–D, and Supplementary Table S3). Higher hnRNPK transcript levels were observed in neuroblastoma cases with poor differentiation (P < 0.0001), advanced INSS stages (P = 0.001), but not with MYCN amplification (P = 0.0891, Fig. 7D). Kaplan–Meier survival curve of 42 (P < 1.0 × 10⁻⁴) and 88 well-defined neuroblastoma cases (GSE16476, P = 4.0 × 10⁻³) revealed that patients with high hnRNPK expression had lower survival probability than those with low expression (Fig. 7E). Multivariate Cox regression analysis of 88 neuroblastoma cases derived from GEO dataset (GSE16476) indicated that the patients' age (HR = 3.225, P < 0.001), MYCN amplification (HR = 2.215, P < 0.001), INSS stage (HR = 2.127, P < 0.001), pancEts-1 expression (HR = 2.066, P = 0.045), and hnRNPK expression (HR = 1.834, P = 0.027), but not gender (HR = 1.232, P = 0.567), were independent prognostic factors for unfavorable outcome of neuroblastoma patients (Supplementary Table S6). These results indicated that hnRNPK was highly expressed and predicted poor clinical outcome in neuroblastoma patients.

LncRNAs, a new class of noncoding RNAs with more than 200 nucleotides in length, play essential roles in regulating gene expression. For example, homeobox transcript antisense RNA (HOTAIR) silences transcription across 40 kb of the HOXD locus by inducing a repressive chromatin state in trans (22), whereas the X inactive specific transcript (XIST) is a key player in X-chromosome inactivation in cis (23). Recent evidence has shown that lncRNAs play both oncogenic and tumor-suppressive roles in the progression of cancers. However, the lncRNAs essential for the tumorigenesis and aggressiveness of neuroblastoma still remain largely unknown. In this study, we discover that the lncRNA pancEts-1 is an independent prognostic marker for progression and poor outcome of neuroblastoma. We demonstrate that pancEts-1 interacts with hnRNPK protein to increase the stability and nuclear translocation of β-catenin (Fig. 7F), which subsequently associates with LEF1 to regulate gene transcription (24). Our cell culture and mouse xenograft models clearly demonstrate that pancEts-1 possesses oncogenic properties to drive the progression of neuroblastoma. The discovery of such a lncRNA represents a promising step for the therapeutic intervention against neuroblastoma.

HnRNPK is one member of the heterogeneous nuclear ribonucleoprotein family protein. As an evolutionarily conserved nucleocytoplasmic shuttling protein, hnRNPK participates in the regulation of transcription, translation, mRNA splicing, mRNA stability, and chromatin remodeling (20). HnRNPK has been shown to regulate the expression of multiple genes, including c-Myc (25), c-Src (26), thymidine kinase 1 (27), androgen receptor (28), eukaryotic translation initiation factor 4E (29), and p53 target genes (30). In addition, hnRNPK enhances the translation of c-Myc mRNA (31), and plays a cytoplasm-based role in stabilizing gastrin (32). Emerging evidence shows that hnRNPK is highly expressed in many types of cancers, such as leukemia, esophagus cancer, lung cancer, nasopharynx cancer, and colorectal cancer (33), and is associated with the tumorigenesis, aggressive metastasis, poor prognosis, and adverse outcome of patients (33). Previous studies show that hnRNPK plays crucial roles in cellular proliferation, clonogenic activity, and migration (34), and knockdown of hnRNPK attenuates the angiogenic and migratory phenotypes of prostate cancer cells (35). In addition, hnRNPK exerts anti-apoptotic functions in cancer cells through decreasing the splice isoform levels of B-cell leukemia/lymphoma 2 like 1 (36). In the current study, we demonstrated that hnRNPK was associated with poor outcome of neuroblastoma patients. Gain- and loss-of-function studies indicated that hnRNPK promoted the growth and aggressiveness of neuroblastoma cells, suggesting the oncogenic roles of hnRNPK in the progression of neuroblastoma.

Human hnRNPK is a modular protein consisting of three conserved KH domains for RNA/DNA binding, and one KI region between KH2 and KH3 for protein interaction (37). Biologically, hnRNPK interacts with diverse proteins including transcriptional activators and repressors, and serves as a docking platform or scaffold that shuttles from cytoplasm to nucleus (33). Previous studies indicate that hnRNPK interacts with transcription factor SP1 to activate the c-src promoter (26). By serving as a cofactor for p53 during DNA damage, hnRNPK promotes the signaling of p53 downstream genes and induces cell-cycle arrest (30). In addition, hnRNPK physically interacts with other transcriptional regulators, such as Y-box binding protein 1 (38), zinc finger protein interacting with K protein 1 (39), CCAAT/enhancer binding protein beta (40), and purine rich element binding protein A (41). It has been implicated that several hnRNPs form complex with β-catenin to regulate the pre-mRNA splicing (42). However, the association of hnRNPK with β-catenin in human cancers still remains largely unknown. In this study, we identified an endogenous hnRNPK interaction with β-catenin in neuroblastoma. In addition, pancEts-1 was essential for the interaction between hnRNPK and β-catenin, which resulted in transcriptional regulation of β-catenin downstream target genes associated with the proliferation, invasion, and metastasis of tumor cells, such as DDR2 (43), DLG1 (44), Ets-1 (45), PCDH7 (46), and PNPT1 (47), shedding insights into novel mechanisms underlying the aggressiveness of neuroblastoma. We believe the different expression correlation values of these genes with pancEts-1 in our series and public microarray dataset may be due to varied detection methods or tumor heterogeneity, which warrants further investigation.

Recent evidence shows that hnRNPK interacts with Ewing sarcoma associated transcript 1 (EWSAT1), a EWS-FLI1–upregulated lncRNA, to affect the expression of genes involved in the pathogenesis of Ewing sarcoma (48). In colon cancer and prostate cancer, hnRNPK interacts with lncRNA MYCLo-2, and plays a crucial role in cancer transformation and tumorigenesis (49). In addition, hnRNPK facilitates the association of lincRNA-p21 with histone methyltransferase SET domain bifurcated 1 (SETDB1) and DNA methyltransferase 1 (50). In this study, we demonstrate that pancEts-1 binds to the KH2 domain of hnRNPK protein to increase its interaction with β-catenin, resulting in stabilization and transactivation of β-catenin. Because knockdown of hnRNPK abolishes the changes in biological features of neuroblastoma cells induced by pancEts-1, our findings indicate that the tumor-promoting functions of pancEts-1 are mediated, at least in part, through regulating the hnRNPK activity. Meanwhile, the other mechanisms contributing to the oncogenic roles of pancEts-1 in neuroblastoma progression warrant further investigation.

In summary, we have demonstrated that pancEts-1 is upregulated and serves as an independent prognostic factor for unfavorable outcome of neuroblastoma. LncRNA pancEts-1 directly interacts with hnRNPK to increase its interaction with β-catenin, resulting in stabilization and transactivation of β-catenin and promotion of the growth, invasion, and metastasis of neuroblastoma cells in vitro and in vivo. This study extends our knowledge about the crucial genes associated with neuroblastoma progression, and suggests that the pancEts-1/hnRNPK/β-catenin axis may be of potential value as a novel therapeutic target for neuroblastoma.

No potential conflicts of interest were disclosed.

Conception and design: L. Zheng, Q. Tong

Development of methodology: D. Li, X. Wang, H. Mei, E. Fang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D. Li, X. Wang, H. Mei, E. Fang, H. Song, F. Yang, H. Li

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D. Li, K. Huang, L. Zheng, Q. Tong

Writing, review, and/or revision of the manuscript: L. Zheng, Q. Tong

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Li, L. Ye, K. Huang

Study supervision: L. Zheng, Q. Tong

We appreciate Dr. Ralf Janknecht for providing vectors. This work was supported by National Natural Science Foundation of China [81272779, 81472363, 81402301, 81672500, 81772967 (to Q. Tong), 81372667, 81572423 (to L. Zheng), 81402408, 81773094 (to H. Mei)], Fundamental Research Funds for the Central Universities (2012QN224, 2013ZHYX003; to Q. Tong), and Natural Science Foundation of Hubei Province (2014CFA012; to Q. Tong).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.