MiR-483-5p and miR-139-5p promote aggressiveness by targeting N-Myc Downstream- Regulated Gene family members in adrenocortical cancer
Claire Agosta1,2,3,4, Jonathan Laugier2,3,4,#, Laurent Guyon2,3,4,#, Josiane Denis2,3,4, Jérôme Bertherat5,6,7, Rossella Libé5,6,7, Bruno Boisson8, Nathalie Sturm8, Jean-Jacques Feige2,3,4, Olivier Chabre 1,2,3,4 and Nadia Cherradi2,3,4 :2,3,4* ID
Centre Hospitalier Universitaire Grenoble Alpes, Service d’Endocrinologie, F-38000 Grenoble, France
2Institut National de la Santé et de la Recherche Médicale, Unité 1036, F-38000 Grenoble, France
3 Commissariat à l’Energie Atomique, Institut de Biosciences et Biotechnologies de Grenoble, Biologie du Cancer et de l’Infection, F-38000 Grenoble, France
4Université Grenoble Alpes, Unité Mixte de Recherche-S1036, 38000, Grenoble, France Université Paris Descartes, F-75006 Paris, France
Institut National de la Santé et de la Recherche Médicale, Unité 1016, Institut Cochin, Centre National de la Recherche Scientifique UMR 8104, F-75014 Paris, France
“Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Département d’Endocrinologie, Centre Expert Cancers Rares de la Surrénale, F-75014 Paris, France
Centre Hospitalier Universitaire Grenoble Alpes, Institut de Biologie et de Pathologie, F-38000 Grenoble, France
Running title: miR-483/NDRG2 and miR-139/NDRG4 in ACC aggressiveness
Keywords: adrenocortical cancer; miR-483-5p; miR-139-5p; N-Myc Downstream-Regulated Gene family; NDRG2; NDRG4; Invasion
*Corresponding author: Nadia Cherradi, Laboratoire Biologie du Cancer et de l’Infection Unité Mixte de Recherche INSERM-CEA-UGA UMR 1036 Institut de Biosciences et Biotechnologies de Grenoble, CEA 17, rue des Martyrs. 38054 Grenoble Cedex 09 France
Phone : 00 33 438 78 35 01 ; Fax : 00 33 438 78 50 58 ; E-mail : nadia.cherradi@cea.fr
# “Authors with equivalent contributions
Research Article
RE
Novelty and Impact
Little is known about the contribution of microRNA dysregulation to the pathogenesis of adrenocortical carcinoma (ACC), a highly aggressive malignancy. This work provides first evidence that two overexpressed microRNAs, miR-483-5p and miR-139-5p, respectively target two members of the N-Myc Downstream-Regulated Gene family NDRG2 and NDRG4 in adrenocortical cancer. The authors demonstrate that miR-483-5p and miR-139-5p promote adrenocortical cancer cell migration and invasion by suppressing expression of NDRG2 and NDRG4. Both target genes are strikingly under-expressed in patients with aggressive tumors. Thus, miR-483-5p/NDRG2 and miR-139-5p/NDRG4 axes may represent novel targets for therapeutic intervention in ACC with poor outcome.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an ‘Accepted Article’, doi: 10.1002/ijc.31363
Abbreviations:
ACC: Adrenocortical carcinoma
naACC: non-aggressive ACC
aACC: aggressive ACC
ACA: Adrenocortical adenoma
DMEM: Dulbecco’s Modified Eagle Medium
FBS: fetal bovine serum FFPE: Formalin-Fixed Paraffin-Embedded miRNA: microRNA
NC miR Inhib : Negative control miRNA inhibitor
NC miR: Negative control miRNA mimic
NDRG2: N-Myc Downstream-Regulated Gene member 2
NDRG4: N-Myc Downstream-Regulated Gene member 4
RT-qPCR: Reverse Transcription quantitative PCR
C
This manuscript contains 5255 words and 6 Figures. Additional Supporting Information includes 7 Supporting Information Tables, 10 Supporting Information Figures, and Supporting Material and Methods.
Accepted
Abstract
Adrenocortical carcinoma (ACC) is a tumor with poor prognosis in which overexpression of a panel of microRNAs has been associated with malignancy but a very limited number of investigations on their role in ACC pathogenesis have been conducted. We examined the involvement of miR-483-5p and miR-139-5p in adrenocortical cancer aggressiveness. Using bioinformatics predictions and mRNA/miRNA expression profiles, we performed an integrated analysis to identify inversely correlated miRNA-mRNA pairs in ACC. We identified N-Myc Downstream-Regulated Gene family members 2 and 4 (NDRG2 and NDRG4) as targets of miR-483-5p and miR-139-5p, respectively. NDRG2 and NDRG4 expressions were inversely correlated respectively with miR-483-5p and miR-139-5p levels in aggressive ACC samples from two independent cohorts of 20 and 44 ACC. Moreover, upregulation of miR-139-5p and downregulation of NDRG4 demonstrated a striking prognostic value. A direct interaction between miR-483-5p or miR-139-5p and their targets was demonstrated in reporter assays. Downregulation of miR-483-5p or miR-139-5p in the ACC cell lines NCI-H295R and SW13 increased NDRG2 or NDRG4 mRNA and protein expression, compromised adrenocortical cancer cell invasiveness and anchorage-independent growth. MiR-483-5p or miR-139-5p overexpression and NDRG2 or NDRG4 inhibition produce similar changes, which are rescued by NDRG2 or NDRG4 ectopic expression. We established that key factors mediating epithelial-to-mesenchymal transition are downstream effectors of miR-483-5p/NDRG2 and miR-139-5p/NDRG4 pathways. Collectively, our data show for the first time that miR-483- 5p/NDRG2 and miR-139-5p/NDRG4 axes promote ACC aggressiveness, with potential implications for prognosis and therapeutic interventions in adrenocortical malignancies.
Introduction
Adrenocortical carcinoma (ACC) is a rare cancer with an incidence of 0.7-2 cases per million people per year. Nevertheless, ACC is a highly aggressive tumor with a very poor prognosis which is in part due to the fact that many ACCs are not detected until they are at advanced stage. 36% to 47% of patients are stage IV at the time of diagnosis, as defined by the presence of distant metastases according to the European Network for the Study of Adrenal Tumors (ENSAT) classification of two independent cohorts from the German adrenocortical carcinoma registry and the U.S. National Cancer Institute SEER (Surveillance, Epidemiology and End Results) database, respectively 1,2. The 5-year disease-specific survival rate for stage IV patients in those cohorts was 13% and 6.9% respectively, and around 38.1% in the overall population (ACC stage I to IV). The only curative treatment for localized ACC (ENSAT stage I, II and III) is radical surgery, but postoperative recurrence is frequent. The treatment for aggressive forms of ACC is limited and still ineffective 3.
☒
☒
☒
☒
Over the last decade, pan-genomic analyses of genetic and epigenetic alterations and genome- wide mRNA expression profile studies allowed major advances in the understanding of the molecular genetics of adrenocortical carcinoma 46. Besides the well-known dysfunctional molecular pathways in adrenocortical tumors such as the IGF2, the Wnt/B-catenin and the TP53 or cell cycle control pathways, high-throughput technologies enabled a more comprehensive genomic characterization of adrenocortical cancer 7-9. Integration of mRNA expression profiling with exome sequencing, SNP array analysis, methylation and microRNA (miRNA) profiling led to the identification of two subgroups of malignant tumors with distinct molecular alterations and clinical outcomes 7. Both ACCs with poor outcome (aggressive ACCs) and ACCs with better prognosis (non-aggressive ACCs) displayed deregulations of miRNA expression 7. MiRNAs are small non-coding RNAs that regulate the expression of more than 50% of coding genes at the post-transcriptional level through their
☒
☒ binding to the 3’-untranslated region (UTR) of the target mRNA 10. MiRNA expression signatures have emerged as promising tools for the diagnosis and prognosis of adrenocortical tumors 11, 12. MiR-483-5p, which is located in the 11p15.5 locus within intron 2 of the IGF2,
☒ ☒
was consistently found to be overexpressed in ACC and to be associated with poorer disease- specific survival 7, 13-17. Its expression is correlated with the expression of its host gene IGF2 which is upregulated in almost 90% of ACCs 16. Importantly, circulating levels of miR-483-5p are increased in ACC patients 14, 18, 19 and predictive of recurrence risk 14. MiR-139-5p is located in the 11q13.4 locus within intron 2 of the phosphodiesterase 2A gene (PDE2A). Microarray and Reverse Transcription quantitative PCR as well as Next Generation Sequencing analyses revealed that miR-139-5p is overexpressed in ACCs with poor prognosis (recurring or metastatic tumors) but not in ACCs with good prognosis (non-recurring tumors) 7, 14.
☒
☒
☒ To date, only a few miRNA targets have been identified or validated experimentally in adrenocortical cancer 15, 20-23. In this study, we performed in silico identification of inversely correlated miRNA-mRNA pairs in ACC. We provide evidence that the two members of the N- myc downstream-regulated gene family, NDRG2 and NDRG4 are relevant targets of miR-483- 5p and miR-139-5p, respectively. Importantly, depleting miR-483-5p or miR-139-5p levels in ACC cells potently suppress their invasive capacities by directly up-regulating NDRG2 or NDGR4 and reversing epithelial-mesenchymal transition, ultimately decreasing aggressiveness ☒
of ACC.
☒
Material and Methods ☒
Patient cohorts
Determinations of miRNA and NDRG2/NDRG4 mRNA expressions were performed in adrenocortical tumor samples from a cohort of 30 patients that has been described in details previously 14. Clinical parameters of the patients are summarized in Supporting Information
5
Table S1. Ten adrenocortical adenomas (ACA) and 20 ACC were collected from the national tumor bank of the Cortico-Medullo Tumeurs Endocrines (COMETE) network (Cochin Hospital, Paris, France). Informed signed consent for the analysis of the tumor tissue and access ☒ ☒ ☒ to the data collected was obtained from all the patients and the study was approved by the Institutional Ethics Committees (Comité de Protection des Personnes, CPP Ile de France). ACC samples were divided into two groups: a first group consisting of aggressive tumors (aACC: tumors recurring within 3 years after surgery with a median recurrence-free survival of 19 months, or tumors that were already metastatic at diagnosis, n=11), and a second group consisting of non-aggressive tumors (naACC: non-recurring tumors over a time period of 36 months, with a median follow-up of 6 years (range 3.5-12.6 years, n=9)14. Recurring carcinoma was defined as a histologically verified tumor at the same localization as the surgically removed first tumor. Three normal adrenal cortex tissues (NA) were obtained from nephrectomized patients.
Additional public data from Assié et al. 7 were used to compare with our own dataset. These data were retrieved from Gene Expression Omnibus, accession number GSE49279 for the microRNA sequencing data from 78 patients and GSE49278 for microarray data from 45 patients. The raw count data per microRNA were further log2 transformed. The collected mRNA data were already RMA (Robust Multi-array Average) normalized and log2 transformed. Clinical data were retrieved from supplementary material of Assié et al. 7 for the 45 ACC samples. The aggressive status was set similarly using clinical data, leading to 24 aACC and 20 naACC, 19 of which have been characterized for their mRNA content. One patient was discarded as its aggressive status was unknown because the follow-up was less than 36 months. Kaplan-Meier curves were generated using the survival library version 2.39.5 in R version 3.2 24
☒
☒
Immunohistochemical analyses of NDRG2 and NDRG4 were performed in FFPE samples from a third cohort of patients, which is described in Supporting Information Material and Methods.
Prediction of miRNA target genes
☒
The target genes of miR-483-5p and miR-139-5p were predicted using the previously published expression data 6 available at ArrayExpress (E-TABM-311) and miRwalk database which is based on comparative analysis of different algorithms 25. Several target prediction algorithms were combined for potential miRNA-mRNA target gene interactions: TargetScan and miRanda which are based on sequence complementarity 26, 27, RNA22 which finds putative microRNA binding sites in the sequence of interest then identifies the targeting microRNA 28 and RNAHybrid which is based on thermodynamic stability 29. We selected target genes which were provided by these prediction softwares and had an inversely correlated expression level with miR-483-5p and miR-139-5p levels in ACC (details provided in Figure 2). Statistically significant miRNA-mRNA relationships were extracted from the MiR Walk results using two criteria: a p value <0.05 and identification by at least three of the five selected target prediction algorithms.
Cell culture
☒ Cell lines were authenticated by Short Tandem Repeat profiling (ATCC). The human adrenocortical carcinoma NCI-H295R cell line was a gift from Dr William E. Rainey (University of Michigan, USA) 30. Cells were cultured as a monolayer on rat tail collagen I- coated plates (0.07 mg/ml, BD Biosciences), in DMEM:F-12 (Life Technologies) supplemented with 1% (v/v) Insulin, Transferrin, Selenium Premix (Corning Discovery Labware), 5% (v/v) Cosmic Calf Serum (HyClone) and antibiotics (100 U/mL penicillin, 100 µg/mL streptomycin
and 30 µg/mL gentamicin, Life Technologies). SW13 cells were purchased from European Collection of Authenticated Cell Cultures (ECACC 87031801) and cultured in Leibovitz’s medium containing 10% (v/v) Foetal Bovine Serum. HEK293T cells were purchased from ATCC and cultured in DMEM GlutaMAX (Invitrogen) containing 10% of FBS (GE Healthcare) and antibiotics. NCI H295R and HEK293T cells were cultured in a humidified incubator at 37°℃ in a 5% CO2-95% air atmosphere. SW13 cells were cultured in the absence of CO2.
Cell transfection
NCI H295R or SW13 cells were transiently transfected using mirVANA miRNA inhibitors or miRNA mimics (Thermo Fisher Scientific). MiRNA inhibitors of hsa-miR-483-5p (ID: MH11749) and hsa-miR-139-5p (ID: MH12629) or miRNA mimics of hsa-miR-483-5p (ID: MC12629) and hsa-miR-139-5p (ID: MC11749) were transfected at final concentration of 50 nM using 5 ul per well (6-well plate) or 2 ul per well (12-well plate) of Lipofectamine RNAiMAX (Invitrogen, Life Technologies) according to the manufacturer’s instructions. mirVANA miRNA Negative Control was transfected as a non-specific random sequence miRNA validated to not produce measurable effects on known miRNA function. Opti-MEM I Reduced Serum Medium (Invitrogen, Life Technologies) was used to dilute miRNAs and Lipofectamine. The efficiency of miRNA overexpression or miRNA knockdown was assessed by RT-qPCR 24 h, 48 h, and 72 h after transfection. Expression of NDRG2 or NDRG4 was inhibited by transfection of their specific siRNA (siRNAs s33036 and s35158 for NDRG2 and NDRG4, respectively; Negative control siRNA Cont #1, 4390843; Thermo Fisher Scientific) using the above-mentioned protocol.
☒
☒ For pCMV6-NDRG2 and pCMV6-NDRG4 plasmid (Origene) transfections, 2 x 106 NCI H295R cells were electroporated with the Cell Line Nucleofector® Kit, Program T-16
according to the manufacturer’s recommendations, in the presence of plasmids at different concentrations. Cells were seeded in duplicate into 6-well plates in DMEM/F12 medium containing 5% of Cosmic Calf Serum.
Luciferase plasmid construction and reporter gene assay
☒
☒
☒
The 3’-UTR of NDRG2 and NDRG4 mRNAs containing the predicted miR-483-5p and miR- 139-5p target sequences, respectively, were amplified from cDNA extract of healthy human trophoblasts with High Fidelity Advantage Genomic LA Polymerase (Sigma-Aldrich). The primers used for amplification were designed by adding a restriction site when necessary (Supporting Information Table S3). The PCR products (708 bp for NDRG2-3’UTR and 1158 bp for NDRG4-3’UTR) were cloned into the pMIR-Report luciferase vector (Promega) using the Multi-Cloning Site, downstream of the Firefly luciferase reporter gene and between Sac-I and Spe-I for NDRG2-3’-UTR or between Sac-I and Hind-III for NDRG2-3’-UTR. All the plasmids were fully sequenced (GATC Biotech). Nucleotide-substitution mutations were carried out using site-directed mutagenesis at the 3’-UTR regions of NDRG2 and NDRG4 (QuickChange XL site-directed mutagenesis kit, Agilent Technologies) with specific primers (Supporting Information Table S3) and the template corresponding to wild-type 3’-UTRs. Post- transcriptional inhibition of luciferase reporter gene by miR-483-5p and miR-139-5p was assayed in HEK293T cells. 1.5×105 cells were seeded in duplicate into 12-well plates one day before transfection and then co-transfected with either miR-483-5p mimics, miR-139-5p mimics or miRNA mimic negative control at a final concentration of 50 nM, together with 1ng of pMIR-Luc Report construct containing NDRG2-3’-UTR, NDRG4-3’-UTR or Mutant-3’-UTR, and with 50 ng of pRL-Tk-Renilla luciferase plasmid (Promega). Lipofectamine 2000 Transfection Reagent (Invitrogen, Life Technologies) was used according the manufacturer’s instructions. Activities of Firefly and Renilla luciferase were measured sequentially 48 h post-
☒
☒
☒
transfection using the Dual-Luciferase reporter assay system (Promega) on a LUMAT LB 9507 luminometer (EGG-Berthold). Results are presented as relative light units of Firefly luciferase activity over relative light units of Renilla luciferase activity to compensate for variations in transfection efficiency and were plotted as a percentage of the activity measured in the appropriate control.
Transwell migration and invasion assay
In vitro cell migration assays were performed using translucent Transwell chambers (8 um-pore size, BD Biosciences). The lower face of the membrane was coated with rat tail collagen I (BD Biosciences). NCI H295R or SW13 cells were transfected in 12-well plates with 50 nM of miRNAs inhibitors or miRNA mimics. 24 h after transfection, the cells were trypsinized, resuspended in 0.5% serum containing medium and 15×104 cells (300 ul) were added to the upper chamber placed in a 24-well plate. Complete medium containing 15% of serum was added to the bottom chamber (750 ul). After 24 h of migration, cells that had not migrated to the lower surface of the membrane were removed using cotton swabs. Cells that had migrated were fixed for 20 min using 4% Paraformaldehyde solution and stained for 30 min using a Crystal Violet solution (0.01% in 10% methanol). Images of five random fields (×10) on each membrane were captured, and the number of migratory cells was counted; the mean values of three duplicate assays for each experimental condition were used for statistical analysis. Invasion assay was performed using inserts coated with 100 ul of Matrigel Basement Membrane Matrix (BD Biosciences) diluted at 0.4 mg/mL with DMEM/F-12. 3 × 104 cells were seeded in 300 ul of serum-free medium in the upper chambers whereas 15%-serum containing medium was added in the bottom chambers. Cells were fixed and stained 4-5 days later as described above.
Methods on RNA extraction, Reverse transcription and Taqman quantitative PCR, Western blot, cell proliferation, cell cycle and apoptosis analyses, anchorage- independent colony formation and immunohistochemistry are provided in Supporting Information Materials and Methods.
Statistical analyses
Statistical analyses were performed with GraphPad Prism software version 4.03 (San Diego, California, USA) using non-parametric Kruskall-Wallis or t tests where appropriate. Results are expressed as mean ± S.E.M. Statistical significance is indicated as * for p≤ 0.05; ** for p≤0.01; *** for p≤ 0.001.
Results
Upregulation of miR-483-5p and miR-139-5p in adrenocortical carcinoma is associated with advanced clinical stage
We have previously reported that a subset of miRNA was differentially expressed between aggressive (aACC) and non-aggressive ACCs (naACCs) 4. The aACC group was characterized by poor prognosis (defined as the C1A group in both studies by Reynies et al. and Assié et al. 6, 7), while the naACC group had a better outcome (defined as the C1B group in both studies by Reynies et al. and Assié et al. 6,7). We focused on two miRNAs, ☒
miR-483-5p and miR-139-5p, that we have found to be up regulated in ACCs as compared to ACAs or normal adrenal cortex and that appeared as potential biomarkers for diagnosis and prognosis 4. For consistency between miRNA and target gene expression measurements, we performed a new RT-qPCR re-analysis of tissue miR-483-5p and miR- 139-5p levels in this previously published adrenocortical tumor cohort 14. We confirmed that miR-483-5p expression was clearly increased in all ACCs as compared to ACA samples
☒
☒
☒
(p<0.01 and p<0.001 for naACC and aACC, respectively) and even more upregulated in a subset of aACCs corresponding to metastatic ACCs, as compared to naACC (Fig. 1A). Statistical analysis comparing miR-483-5p expression in aACC and naACC almost reached significance (p = 0.05, 95% Confidence Intervals of mean expression were [7.8-30.2] for naACC and [9.8-60.2] for aACC). MiR-139-5p was significantly upregulated in aACC tumors as compared to naACC (Fig. 1B). Kaplan-Meier analyses revealed that patients with high levels of miR-483-5p tended to have worse overall survival (hazard ratio (HR): 0.254; 95% CI: 0.066-1.108; log-rank test p=0.069) than those with low miR-483-5p levels (Supporting Information Fig. S1). Patients with high levels of miR-139-5p had a significant poorer survival (11.4 months, HR: 0; 95% CI: 0.011-0.230, Log-rank test p<0.0001) than those with low miR-139-5p levels (Supporting Information Fig. S1). To strengthen these findings, we compared our results with Assié et al data 7 as an ACC independent sample set (n=44). Again, miR-483-5p was significantly upregulated in both naACC (C1B group) and aACC (C1A group) as compared to ACA (with p=2.9 x 10-5 and p=2.2 x 10-9, respectively) (Supporting Information Fig. S2A). As observed with our cohort, miR-139-5p was found upregulated in aACC as compared to naACC (p=2.7 x 10-3) (Supporting Information Fig. S2B). Among ACCs, only miR-139-5p demonstrated a striking prognostic value (Supporting Information Fig. S3A and S3B), thus confirming our findings in this independent and larger cohort. We subsequently measured miR-483-5p and miR-139-5p expression in NCI H295R cells, a human cell line established from a large invasive ACC that was later reported to have metastasized to the lungs and liver 30. 90-fold and 18-fold higher levels of miR-483-5p and miR-139-5p were detected in NCI H295R cells, respectively, as compared to normal adrenal cortex (Fig. 1A and 1B), suggesting that NCI H295R is a relevant ACC cell line model to study miR-483-5p and miR-139-5p functions.
Identification of putative target genes
☒
☒
☒
To identify the biologically relevant target genes of miR-483-5p and miR-139-5p, we combined in silico prediction analysis using different bioinformatics algorithms and correlation analyses of the transcriptome and miRnome data of human adrenocortical tumors (Fig. 2). Inversely correlated miRNA-target mRNA expressions in ACC are shown in Supporting Information Tables S5 and S6. The N-myc Downstream Regulated Gene (NDRG) family member NDRG4 was the most downregulated potential target gene of miR- 139-5p. Interestingly, another member of the family, NDRG2, was a potential target of miR-483-5p. We therefore focused our attention on NDRG2 and NDRG4 whose function in the physiopathology of adrenocortical cancer was totally unknown (Supporting Information Tables S7).
NDRG2/NDRG4 and miR-483-5p/miR-139-5p expressions are negatively correlated in adrenocortical carcinoma clinical samples
☒
☒
☒
To validate our predictions, mRNA expression levels of NDRG2 and NDRG4 were determined by RT-qPCR in the cohort used for miRNA measurements. Both transcripts were down- regulated in aACCs as compared with naACC (Fig. 1C and 1D). Similar deregulations of NDRG2 and NDRG4 expression were found in ACC using the mRNA data collected from Assié et al 7 (Supporting Information Fig. S2C and S2D). Interestingly, a high level of NDRG4 but not of NDRG2 was significantly associated with better survival (Supporting Information Fig. S3C and S3D). Expression levels of NDRG2 and NDRG4 in NCI H295R cells were comparable to those measured in aACC tumors (Fig. 1C and 1D). We further studied the correlation between the expression of each miRNA and its putative target gene in each aACC or naACC sample. Inverse correlations between expression levels of NDRG2 and miR-483-5p, and between expression levels of NDRG4 and miR-139-5p were found in aACC (Fig. 1E-1H),
☒ suggesting that low expression of NDRG2 or NDRG4 is associated with ACC aggressiveness. Correlation analyses for miRNA and mRNA levels collected from Assié et al data 7 showed a negative correlation between NDRG4 and miR-139-5p expression in both naACC and aACC while there was a trend to a negative correlation between NDRG2 and miR-483-5p expression in aACC (Supporting Information Fig. S2E-S2H). Collectively, these results suggest that overexpression of miR-438-5p and miR-139-5p in ACC could lead to repression of NDRG2 and NDRG4 and to tumor aggressiveness.
NDRG2/NDRG4 protein expression in adrenocortical carcinomas
☒
☒
☒
☒
We next investigated whether NDRG2 and NDRG4 mRNA expressions were correlated to their respective protein abundance in an independent cohort of 12 aACC, 8 naACC and 10 ACA. Immunohistochemical analyses showed robust NDRG2 and NDRG4 expression in ACA and adjacent normal cortex (Nc) as represented by the brown staining (Fig. 3). An ACA specimen treated with nonspecific rabbit IgG was completely nonreactive (Supporting Information Fig. S4). In naACC, NDRG2 immunoreactivity was absent while heterogeneous expression pattern of NDRG4 was observed. By contrast, aACC featured complete loss of NDRG2 and NDRG4 expression compared with ACA and adjacent tissue. (Fig. 3). Statistical analyses of staining scores revealed that NDRG2 expression was significantly lower in ACC compared with ACA (p<0.001) but not significantly different between aACC and naACC (data not shown). NDRG4 displayed a significant lower expression in ACC compared with ACA (p=0.018) as well as in aACC compared with naACC (p=0.043). These findings demonstrate that NDRG2 and NDRG4 protein levels are downregulated in ACC in correlation with their respective transcripts and suggest that both proteins are potential immunohistochemical markers for ACC patients.
Silencing miR-483-5p and miR-139-5p restores endogenous NDRG2 and NDRG4 expression in adrenocortical carcinoma cells
In light of the preceding results, we hypothesized that NDRG2 and NDRG4 might be functionally relevant target genes of miR-483-5p and miR-139-5p, respectively. We first determined whether repression of miR-483-5p and miR-139-5p using miRNA inhibitors could restore endogenous expression of NDRG2 and NDRG4 mRNAs, respectively, in NCI H295R and SW13 cells. Intracellular levels of miR-483-5p and miR-139-5p in NCI H295R cells were efficiently decreased within 24 h, by 3.3 and 8.6-fold respectively and this inhibition was maintained for at least 72 h (Fig. 4A and 4B). Quantitative RT-PCR analysis revealed that miR- 483-5p and miR-139-5p blockade increased endogenous NDRG2 and NDRG4 mRNA levels (Fig. 4C and 4D). In accordance with these results, we observed an increase in NDGR2 and NDRG4 proteins in the presence of miR-483-5p and miR-139-5p inhibitors (Fig. 4E and 4F). Conversely, overexpression of miR-483-5p and miR-139-5p clearly decreased the expression levels of NDRG2 and NDRG4 proteins, respectively (Fig. 4E and 4F). Surprisingly, SW13 cells displayed extremely low levels of miR-483-5p and miR-139-5p in contrast to NCI H295R cells (Supporting Information Fig. S5A and S5B). Transfection of specific inhibitors decreased miR- 483-5p and miR-139-5p by 1.7- and 2.5-fold, respectively, and concomitantly increased NDRG2 and NDRG4 transcripts (by 1.7- and 1.9-fold, respectively, Supporting Information Fig. S5C and S5D). Conversely, NDRG2 and NDRG4 mRNA levels were reduced by 1.7- and 1.9-fold, respectively, in the presence of miR-483-5p and miR-139-5p mimics (Supporting Information Fig. S5E and S5F). These modulations of NDRG2 and NDRG4 mRNA expression by miR-483-5p and miR-139-5p inhibitors or mimics were correlated with NDRG2 or NDRG4 protein levels (Supporting Information Fig. S5G and S5H). Taken together, these results suggest that miR-483-5p and miR-139-5p negatively regulate NDRG2 and NDRG4 expression, respectively, in both NCI H295R and SW13 cell models.
MiR-483-5p and miR-139-5p directly target the 3’-untranslated region of NDRG2 and
NDRG4
☒ ☒ To investigate whether miR-483-5p and miR-139-5p directly target the 3’UTR of their putative mRNA targets, the 3’UTR of NDRG2 or NDRG4 were cloned downstream of the Firefly luciferase coding region in pMiR reporter vector (Fig. 5A and 5B). MiR-483-5p, miR-139-5p or negative control miRNA inhibitors were co-transfected with reporter plasmids into NCI H295R cells. Reporter assays revealed that miR-483-5p and miR-139-5p inhibition significantly increased NDRG2-3’UTR- and NDRG4-3’UTR-driven luciferase activity to 1.6- and 1.8-fold of the luciferase activity measured in control cells, respectively (Fig. 5C and 5D). Conversely, transfections of miR-483-5p, miR-139-5p or control miRNA mimics in the presence of reporter plasmids significantly decreased NDRG2-3’UTR- and NDRG4-3’UTR-driven luciferase activity to 0.6- and 0.7-fold of that measured in control cells, respectively (Fig 5E and 5F). To determine whether these effects were due to a direct interaction between NDRG2 3’UTR and miR-483-5p (at the 2081 bp position, NM_ 201535.1) and between NDRG4 3’UTR and miR- 139-5p (at the 1522 bp position, NM_020465.3), we mutated their putative binding sites (Fig. 5A and 5B). As shown in Fig. 5E and 5F, the ability of miR-483-5p or miR-139-5p to inhibit luciferase activity was completely abrogated. These results suggest that miR-483-5p and miR- 139-5p interact physically with their target gene 3’UTR and that their predicted binding sites contribute respectively to NDRG2 and NDRG4 repression.
MiR-483-5p and miR-139-5p do not affect cell cycle status, cell proliferation or cell apoptosis
☒
We next analyzed the role of miR-483-5p and miR-139-5p in shaping adrenocortical cancer cell phenotype. First, we evaluated the effect of miR-483-5p and miR-139-5p inhibition on cell cycle. NCI H295R cells were transfected with miR-483-5p and miR-139-5p inhibitors. No
☒
significant difference in the proportion of cells in G1, S, or G2-M phases was found between NCI H295R cells transfected with miR-483-5p or miR-139-5p inhibitors and cells transfected with negative control miRNA inhibitors or mock-transfected cells (Supporting Information Fig. S6A and S6B). In addition, miR-483-5p or miR-139-5p silencing had no effect on NCI H295R proliferation (Supporting Information Fig. S6C). We further explored the role of both miRNAs in NCI H295R cell apoptosis using flow cytometry. Annexin V-based cell apoptosis analyses showed no significant differences in the fraction of cells in early or late apoptosis between miR- 483-5p or miR-139-5p inhibitor-transfected cells and negative control miRNA-transfected cells (Supporting Information Fig. S6D). Collectively, these results indicate that miR-483-5p and miR-139-5p are not involved in the regulation of adrenocortical carcinoma cell proliferation or apoptosis.
☒
NDRG2 and NDRG4 downregulation mediates miR-483-5p and miR-139-5p-induced
migration and invasion of adrenocortical cancer cells
Because we observed that upregulation of miR-483-5p and miR-139-5p in adrenocortical cancer was more frequent in aggressive ACC (Supporting Information Fig. S2 and S7), we postulated that both miRNAs could be involved in ACC invasive capacities. NCI H295R cells were transfected with miR-483-5p or miR-139-5p inhibitors and analyzed for their migration and invasion using Transwell-based assay. We found that inhibition of miR-483-5p
☒
☒ or miR-139-5p significantly reduced the migration and invasion of NCI H295R cells through the insert membrane (Fig. 6 A-D). In SW13 cells, which express very low endogenous levels of miR-483-5p and miR-139-5p, transfection of each of these miRNAs promoted their migration and invasion (Supporting Information Fig. S8A-B). ☒
☒
It has been reported that NDRG2 is closely associated with tumor invasion and metastasis in human cancer 31 while the potential role of NDRG4 in such processes remains unknown. To
☒
determine whether the invasion-promoting effects of miR-483-5p or miR-139-5p depend on NDRG2 or NDRG4 suppression, we rescued the expression of NDRG2 or NDRG4 by using constructs that encode their entire coding sequences but lack the 3’UTR. The expression of both proteins was confirmed by western blotting (Supporting Information Fig. S9). Ectopic expression of NDRG2 or NDRG4 markedly decreased the invasive capacities of NCI H295R cells compared to cells transfected with empty vector (Fig. 6E-F). Conversely, siRNA- induced repression of NDRG2 or NDRG4 potentiated migration and invasion of NCI H295R and SW13 cells (Supporting Information Fig S10A-B and S8C-D), confirming that NDRG2 and NDRG4 downregulation is involved in the invasive phenotype of both cell lines.
☒
☒
☒ We hypothesized that miR-483-5p and miR-139-5p might promote invasion by regulating genes involved in epithelial-mesenchymal transition (EMT), a process in which epithelial cells lose their polarity and are converted into a mesenchymal phenotype 32. We therefore tested whether miR-483-5p and miR-139-5p were connected to EMT in adrenocortical cancer cells using NCI H295R cells transfected with miR-483-5p or miR-139-5p inhibitors. Assessment of the expression of EMT markers 48 h post-transfection revealed that downregulation of miR-483-5p or miR-139-5p leads to a substantial decrease in the levels of Snail, a major EMT-inducing transcription factor while Twist transcription factor was not affected (Fig. 6G). This was accompanied by a reduction in the expression of the mesenchymal proteins N-cadherin and vimentin (Fig. 6G). Surprisingly, no detectable changes in the cellular content of E-cadherin expression were observed (Fig. 6G), suggesting that 48 h of treatment with the miRNA inhibitors were not sufficient to achieve the expected increase in E-cadherin expression. In SW13 cells, both miR-483-5p and miR-139-5p mimics markedly increased vimentin expression while they modestly but consistently increased N- cadherin levels within 48 h (Supporting Information Fig. S8E). ☒
☒
☒
☒
To assess whether the loss of NDRG2 or NDRG4 could also modulate EMT marker expression, NDRG2 and NDRG4 were knocked down using increasing concentrations of siRNAs. NDRG2 and NDRG4 silencing in NCI H295R cells upregulated N-cadherin and vimentin expression with maximal effects at 50 nM and 10 nM of NDRG2 and NDRG4 siRNA, respectively (Supporting Information Fig. S10C). Collectively, these results indicate that upregulation of miR-483-5p and miR-139-5p promotes the aggressive phenotype of adrenocortical cancer cells through down-regulation of NDRG2 and NDRG4 and induction of epithelial-mesenchymal transition.
MiR-483-5p and miR-139-5p favor anchorage-independent growth of NCI H295R adrenocortical carcinoma cells
Anchorage-independent growth is a hallmark of in vitro transformation 33. We therefore assessed the ability of NCI H295R cells transfected or not with miR-483-5p or miR-139-5p inhibitors to growth in an anchorage-independent manner (Fig. 6H). Negative control miRNA-treated cells formed numerous and large colonies in soft agar. In miR-483-5p inhibitor-transfected cells, the number of total colonies was significantly decreased while few large colonies were still detected. MiR-139-5p inhibitor-transfected cells generated only a few and very small colonies. Quantitative analysis of independent experiments revealed that inhibition of miR-483-5p and miR-139-5p decreased the colony number by 42 % and 75 %, respectively, when compared to negative control inhibitor (Fig 6I). These results indicate that antagonizing miR-483-5p and miR-139-5p overexpression impedes the transformed phenotype of NCI H295R adrenocortical carcinoma cells.
☒
☒
Discussion
ACC is rare and highly aggressive malignancy, associated with a pejorative prognosis due to
distant metastasis in most cases. Current treatments fail in many patients with advanced ACC 34. The underlying molecular mechanisms resulting in carcinogenesis and aggressiveness remain poorly defined and stand for the major limitation in finding novel molecular-targeted therapies. Significant progress has been made in the molecular characterization of ACC thanks to the European Network for the Study of Adrenal tumors (ENSAT) and The Cancer Genome Atlas (TCGA) international consortia. These studies improved the classification of adrenocortical tumors based on pan-genomic analysis of genetic mutations, chromosomal aberrations, DNA methylation, and mRNA and miRNA expression 7, 8. We and others have previously reported that a panel of miRNAs is altered in adrenocortical cancer 35. Among dysregulated miRNAs, miR-483-5p and miR-139-5p were found highly upregulated in ACC. Nevertheless, the regulatory mechanisms exerted by both miRNAs in ACC cells are unknown. In the present study, we report that miR-483-5p and miR-139-5p directly target two members of the N-Myc Downstream-Regulated Gene family, NDRG2 and NDRG4, which appear as newly identified tumor suppressors in adrenocortical cancer. New analyses of our previously published adrenocortical tumor cohort (10 ACA and 20 ACC) 14 as well as data mining of publicly available miRNA sequencing and microarray resources (45 ACC and 30 ACA) 7 revealed that miR-483-5p and miR-139-5p expressions were inversely correlated with NDRG2 and NDRG4 expressions respectively, in ACC. Moreover, inhibition of miR-483-5p and miR-139-5p or restoration of NDRG2 and NDRG4 effectively suppressed the invasive potential of adrenocortical carcinoma cells in vitro. Thus, miR-483-5p and miR-139-5p might act as oncomiRNAs in ACC pathogenesis.
☒
MiR-483-5p was first identified in human fetal liver 36 and subsequently found dysregulated and associated with poor disease-specific survival in certain cancers including ACC 16, 17, 37, 38 MiR- 483-5p resides within the second intron of IGF-2 gene. Its expression showed a significant positive correlation with IGF-2 transcript expression in ACC tumors 16. However, while IGF-2
☒
promote growth of ACC cells in vitro, no significant differences in the clinical, biological and molecular features were found between high- and low-IGF-2 expressing ACC 39. Our results showing that miR-483-5p promote invasiveness in the ACC cell models NCI H295R and SW13 suggest that miR-483-5p could target other cellular pathways than those regulated by it host gene IGF-2. Indeed, miR-483-5p has been shown to promote tumor invasion and metastasis in lung adenocarcinoma preclinical models 40. We show that silencing miR-483-5p in NCI-H295R cells did not affect cell proliferation. This is in contrast to a previous study showing that inhibition of miR-483-5p in NCI-H295R cells led to a modest reduction of cell proliferation 15. Although we do not have explanation for this discrepancy, one should mention that in Ozata’s study, the concentrations of miRNA inhibitors used are 20 times higher than those we have tested in the present work 15. MiR-139-5p, which resides within intron 2 of PDE2A gene, has been reported to be downregulated and to display anti-oncogenic and anti-metastatic activities in human cancers 41. We found that miR-139-5p is upregulated in ACC patient samples as well as in NCI H295R cells in which it acts as pro-invasive miRNA. Our result showing for the first time that miR-139-5p is a powerful prognostic factor for survival of ACC patients supports its role in ACC aggressiveness. Interestingly, amplification of the 11q13 chromosomal region which carries PDE2A gene occurs in several human cancers including ACCs 42. In addition, PDE2A has been shown to be overexpressed in ACC mutated for -catenin gene CTNNB1 43. Whether the expression of miR-139-5p is correlated with or independent from that of its host gene PDE2A in ACC remains to be determined.
☒
Because miRNAs regulate multiple genes, characterization of miRNA-mRNA interactions is a prerequisite to understand their role in tumor biology. Despite availability of numerous methods, attempts to connect mechanistically miRNAs with their mRNA targets and to study miRNA-mRNA integratively using their respective expressions in adrenocortical cancer are still sporadic. In ACC, the p53 upregulated modulator of apoptosis PUMA has been reported as a
☒ ☒ target of miR-483-3p, the complementary strand of miR-483-5p, while targets of miR-139-5p are still unknown. Integrating the transcriptome and the miRnome data of ACCs allowed us to identify NDRG2 and NDRG4 as target transcripts of miR-483-5p and miR-139-5p, respectively. To establish the functional link between miR-483-5p and NDRG2, and miR-139- 5p and NDRG4, we used the two available ACC cell lines NCI H295R and SW13. Importantly, we found that NCI H295R cells more closely resemble ACC patient tumor samples than SW13 in terms of miR-483-5p and miR-139-5p overexpression. Nevertheless, suppression of miR- 483-5p or miR-139-5p increased NDRG2 or NDRG4 mRNA and protein levels in both NCI- H295R and SW13 cells while transfection of exogenous miRNA mimics had opposite effects. Direct physical interactions between miR-483-5p and NDRG2, and between miR-139-5p and NDRG4, were demonstrated by reporter gene assays, indicating that each miRNA post- transcriptionally regulate its target transcript through its binding site in the 3’UTR.
☒
To clarify further the relationship between miR-483-5p and NDRG2 and between miR-139-5p and NDRG4, migration and invasion assays as well as EMT marker analyses were conducted in NCI-H295R and SW13 cells subjected to modulations of miRNA or NDRG2/NDRG4 levels. Silencing miR-483-5p and miR-139-5p in NCI H295R cells reduced their migration/invasion capacities and repressed EMT marker expression while increasing miR-483-5p and miR-139-5p levels in SW13 cells promoted their migration/invasion and induced EMT. Notably, downregulation of NDRG2 or NDRG4 using RNA interference enhanced migration and invasion of both cell lines, thus mimicking the effects of their regulatory miRNA overexpression. Therefore, miR-483-5p and miR-139-5p might be implicated in cell motility and dissemination from the primary tumor to metastatic sites through NDRG2 and NDRG4 repression. Our observations are in line with those reported by Song et al. on miR-483-5p-
☒ induced EMT in lung adenocarcinoma cell through targeting of RhoGDI and Alcam metastasis suppressors 40. Nonetheless, we could not observe an induction of the epithelial marker E-
cadherin 48 h post-transfection of miRNA inhibitors. Likewise, changes in E-cadherin cellular content were not detected during the early steps of TGFB-mediated EMT in mammary epithelial cells 44.
☒
NDRG2 and NDRG4 belong to the N-myc downstream-regulated gene family of proteins consisting of 4 members, NDRG1 to NDRG4, which share 53-65% amino acid identity. The NDRG genes are located on different chromosomes and display distinct subcellular (nuclear versus cytoplasm) and tissue distributions. Although they were found to play fundamental roles in physiological processes such as cell proliferation, differentiation, development, and stress responses, the molecular mechanisms of action of these proteins needs to be explored further 31. We found that NDRG2 mRNA levels are significantly downregulated in ACC as reported in other human malignancies 45. NDRG2 has been reported to interact with prenylated Rab acceptor-1 (PRA1), an ubiquitously expressed protein localized to Golgi and late endosomes, which regulates vesicular trafficking, lipid transport and cell migration through its binding to the prenylated Rab GTPase proteins 46. Interestingly, NDRG2 and PRA1 were shown to
☒
synergistically inhibit T-Cell Factor/B-catenin signalling by blocking GSK3B phosphorylation in colon carcinoma 46. Loss of NDRG2 expression in T-cell lymphoma and other cancers activates PI3K-AKT signalling through enhanced PTEN phosphorylation 47. As both Wnt/B- catenin and PI3K/AKT signalling pathways are activated in ACC, the relationships between NDRG2 and these pathways in ACC certainly deserves further investigations. Much less is known concerning NDRG4 expression and function in cancer, but it seems to have tumor- suppressive or oncogenic functions depending on the tissue type. Here, we show for the first time using immunohistochemistry that both NDRG2 and NDRG4 proteins are strikingly under- expressed in ACC samples. Moreover, re-expression of NDRG2 and NDRG4 was sufficient to impair the invasive capacities of NCI H295R cells. These results are in line with previous studies showing that downregulation of NDRG2 in malignant tumors is closely correlated with
☒
☒
☒ ☒
highly invasive phenotype, metastasis and poor overall survival 48, 49.
The regulation of NDRG family expression involves complex and various mechanisms including not only transcription repressors such as N-myc and C-myc proto-oncogenes, but also promoter hyper-methylation 50. Our data highlight additional mechanisms for the regulation of NDRG2 and NDRG4 expression in ACCs, through miR-483-5p- and miR-139-5p-mediated post-transcriptional control. As both miRNAs contribute to cell invasion processes, they represent potential therapeutic targets for these highly aggressive and therapy-refractory tumors.
Acknowledgments
This work was supported by the Institut National de la Santé et de la Recherche Médicale, the Ligue Départementale Contre le Cancer (Comité Loire), the Association Surrénales, the French Endocrine Society (SFE) and the Institut National du Cancer (INCA-DGOS-8663, Programme Hospitalier de Recherche Clinique COMETE-TACTIC). The authors thank Dr Aurélien de Reyniès (Ligue Nationale contre le Cancer) for his initial contribution in the integrated analysis of the miRnome and transcriptome of adrenocortical tumors, Veronique Collin-Faure for her technical assistance with the flow cytometry experiments, Fernande René-Corail and Karine Perlemoine for excellent technical assistance in clinical sample processing.
Conflict of Interest Statement:
The authors declare no conflict of interest
☒ References
1. Fassnacht M, Johanssen S, Quinkler M, Bucsky P, Willenberg HS, Beuschlein F, Terzolo M, Mueller HH, Hahner S, Allolio B, German Adrenocortical Carcinoma Registry G, European Network for the Study of Adrenal T. Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: proposal for a Revised TNM Classification. Cancer 2009;115: 243-50.
2. Lughezzani G, Sun M, Perrotte P, Jeldres C, Alasker A, Isbarn H, Budaus L, Shariat SF, Guazzoni G, Montorsi F, Karakiewicz PI. The European Network for the Study of Adrenal Tumors staging system is prognostically superior to the international union against cancer- staging system: a North American validation. Eur J Cancer 2010;46: 713-9.
3. Libe R. Adrenocortical carcinoma (ACC): diagnosis, prognosis, and treatment. Front Cell Dev Biol 2015;3: 45.
4. de Fraipont F, El Atifi M, Cherradi N, Le Moigne G, Defaye G, Houlgatte R, Bertherat J, Bertagna X, Plouin PF, Baudin E, Berger F, Gicquel C, et al. Gene expression profiling of human adrenocortical tumors using complementary deoxyribonucleic Acid microarrays identifies several candidate genes as markers of malignancy. J Clin Endocrinol Metab 2005;90: 1819-29.
5. Giordano TJ, Kuick R, Else T, Gauger PG, Vinco M, Bauersfeld J, Sanders D, Thomas DG, Doherty G, Hammer G. Molecular classification and prognostication of adrenocortical tumors by transcriptome profiling. Clin Cancer Res 2009;15: 668-76.
6. de Reynies A, Assie G, Rickman DS, Tissier F, Groussin L, Rene-Corail F, Dousset B, Bertagna X, Clauser E, Bertherat J. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol 2009;27: 1108-15.
7. Assie G, Letouze E, Fassnacht M, Jouinot A, Luscap W, Barreau O, Omeiri H, Rodriguez S, Perlemoine K, Rene-Corail F, Elarouci N, Sbiera S, et al. Integrated genomic characterization of adrenocortical carcinoma. Nat Genet 2014;46: 607-12.
8. Zheng S, Cherniack AD, Dewal N, Moffitt RA, Danilova L, Murray BA, Lerario AM, Else T, Knijnenburg TA, Ciriello G, Kim S, Assie G, et al. Comprehensive Pan-Genomic Characterization of Adrenocortical Carcinoma. Cancer Cell 2016;29: 723-36.
9. Gara SK, Wang Y, Patel D, Liu-Chittenden Y, Jain M, Boufraqech M, Zhang L, Meltzer PS, Kebebew E. Integrated genome-wide analysis of genomic changes and gene regulation in human adrenocortical tissue samples. Nucleic Acids Res 2015;43: 9327-39.
10. Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res 2009;19: 92-105.
11. Cherradi N. microRNAs as Potential Biomarkers in Adrenocortical Cancer: Progress and Challenges. Front Endocrinol (Lausanne) 2016;6: 195.
12. Igaz P, Igaz I, Nagy Z, Nyiro G, Szabo PM, Falus A, Patocs A, Racz K. MicroRNAs in adrenal tumors: relevance for pathogenesis, diagnosis, and therapy. Cell Mol Life Sci 2015;72: 417-28.
13. Duregon E, Rapa I, Votta A, Giorcelli J, Daffara F, Terzolo M, Scagliotti GV, Volante M, Papotti M. MicroRNA expression patterns in adrenocortical carcinoma variants and clinical pathologic correlations. Hum Pathol 2014;45: 1555-62. ☒
14. Chabre O, Libe R, Assie G, Barreau O, Bertherat J, Bertagna X, Feige JJ, Cherradi N. Serum miR-483-5p and miR-195 are predictive of recurrence risk in adrenocortical cancer patients. Endocr Relat Cancer 2013;20: 579-94.
15. Ozata DM, Caramuta S, Velazquez-Fernandez D, Akcakaya P, Xie H, Hoog A, Zedenius J, Backdahl M, Larsson C, Lui WO. The role of microRNA deregulation in the pathogenesis of adrenocortical carcinoma. Endocr Relat Cancer 2011;18: 643-55.
16. Patterson EE, Holloway AK, Weng J, Fojo T, Kebebew E. MicroRNA profiling of adrenocortical tumors reveals miR-483 as a marker of malignancy. Cancer 2011;117: 1630-9.
17. Soon PS, Tacon LJ, Gill AJ, Bambach CP, Sywak MS, Campbell PR, Yeh MW, Wong SG, Clifton-Bligh RJ, Robinson BG, Sidhu SB. miR-195 and miR-483-5p Identified as Predictors of Poor Prognosis in Adrenocortical Cancer. Clin Cancer Res 2009;15: 7684-92.
18. Patel D, Boufragech M, Jain M, Zhang L, He M, Gesuwan K, Gulati N, Nilubol N, Fojo T, Kebebew E. MiR-34a and miR-483-5p are candidate serum biomarkers for adrenocortical tumors. Surgery 2013;154: 1224-8; discussion 9.
19. Szabo DR, Luconi M, Szabo PM, Toth M, Szucs N, Horanyi J, Nagy Z, Mannelli M, Patocs A, Racz K, Igaz P. Analysis of circulating microRNAs in adrenocortical tumors. Lab Invest 2014;94: 331-9.
20. Caramuta S, Lee L, Ozata DM, Akcakaya P, Xie H, Hoog A, Zedenius J, Backdahl M, Larsson C, Lui WO. Clinical and functional impact of TARBP2 over-expression in adrenocortical carcinoma. Endocr Relat Cancer 2013;20: 551-64.
21. Doghman M, El Wakil A, Cardinaud B, Thomas E, Wang J, Zhao W, Peralta-Del Valle MH, Figueiredo BC, Zambetti GP, Lalli E. Regulation of insulin-like growth factor-mammalian target of rapamycin signaling by microRNA in childhood adrenocortical tumors. Cancer Res 2010;70: 4666-75.
22. Wu Y, Wang W, Hu W, Xu W, Xiao G, Nie Q, Ouyang K, Chen S. MicroRNA-205 suppresses the growth of adrenocortical carcinoma SW-13 cells via targeting Bcl-2. Oncol Rep 2015;34: 3104-10.
23. Glover AR, Zhao JT, Gill AJ, Weiss J, Mugridge N, Kim E, Feeney AL, Ip JC, Reid G, Clarke S, Soon PS, Robinson BG, et al. microRNA-7 as a tumor suppressor and novel therapeutic for adrenocortical carcinoma. Oncotarget 2015.
24. R Development Core Team. R: A language and environment for statistical computing Vienna, Austria: R Foundation for Statistical Computing., 2016.
25. Dweep H, Sticht C, Pandey P, Gretz N. miR Walk — database: prediction of possible miRNA binding sites by “walking” the genes of three genomes. J Biomed Inform 2011;44: 839-47.
26. Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 2005;120: 15-20.
27. John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS. Human MicroRNA targets. PLoS Biol 2004;2: e363.
28. Miranda KC, Huynh T, Tay Y, Ang YS, Tam WL, Thomson AM, Lim B, Rigoutsos I. A pattern-based method for the identification of MicroRNA binding sites and their corresponding heteroduplexes. Cell 2006;126: 1203-17.
29. Rehmsmeier M, Steffen P, Hochsmann M, Giegerich R. Fast and effective prediction of microRNA/target duplexes. RNA 2004;10: 1507-17.
30. Wang T, Rainey WE. Human adrenocortical carcinoma cell lines. Mol Cell Endocrinol 2012;351: 58-65.
31. Melotte V, Qu X, Ongenaert M, van Criekinge W, de Bruine AP, Baldwin HS, van Engeland M. The N-myc downstream regulated gene (NDRG) family: diverse functions, multiple applications. FASEB J 2010;24: 4153-66.
32. Marcucci F, Stassi G, De Maria R. Epithelial-mesenchymal transition: a new target in anticancer drug discovery. Nat Rev Drug Discov 2016;15: 311-25.
33. Cifone MA, Fidler IJ. Correlation of patterns of anchorage-independent growth with in vivo behavior of cells from a murine fibrosarcoma. Proc Natl Acad Sci U S A 1980;77: 1039-43.
34. Fassnacht M, Terzolo M, Allolio B, Baudin E, Haak H, Berruti A, Welin S, Schade- Brittinger C, Lacroix A, Jarzab B, Sorbye H, Torpy DJ, et al. Combination chemotherapy in advanced adrenocortical carcinoma. N Engl J Med 2012;366: 2189-97.
35. Cherradi N. microRNAs as Potential Biomarkers in Adrenocortical Cancer: Progress and Challenges. Front Endocrinol (Lausanne) 2015;6: 195.
36. Fu H, Tie Y, Xu C, Zhang Z, Zhu J, Shi Y, Jiang H, Sun Z, Zheng X. Identification of human fetal liver miRNAs by a novel method. FEBS Lett 2005;579: 3849-54.
37. Wang L, Shi M, Hou S, Ding B, Liu L, Ji X, Zhang J, Deng Y. MiR-483-5p suppresses the proliferation of glioma cells via directly targeting ERK1. FEBS Lett 2012;586: 1312-7.
38. Veronese A, Lupini L, Consiglio J, Visone R, Ferracin M, Fornari F, Zanesi N, Alder H, D’Elia G, Gramantieri L, Bolondi L, Lanza G, et al. Oncogenic role of miR-483-3p at the IGF2/483 locus. Cancer Res 2010;70: 3140-9.
39. Guillaud-Bataille M, Ragazzon B, de Reynies A, Chevalier C, Francillard I, Barreau O, Steunou V, Guillemot J, Tissier F, Rizk-Rabin M, Rene-Corail F, Al Ghuzlan A, et al. IGF2 promotes growth of adrenocortical carcinoma cells, but its overexpression does not modify phenotypic and molecular features of adrenocortical carcinoma. PLoS One 2014;9: e103744.
40. Song Q, Xu Y, Yang C, Chen Z, Jia C, Chen J, Zhang Y, Lai P, Fan X, Zhou X, Lin J, Li M, et al. miR-483-5p promotes invasion and metastasis of lung adenocarcinoma by targeting RhoGDI1 and ALCAM. Cancer Res 2014;74: 3031-42.
41. Miyoshi J, Toden S, Yoshida K, Toiyama Y, Alberts SR, Kusunoki M, Sinicrope FA, Goel A. MiR-139-5p as a novel serum biomarker for recurrence and metastasis in colorectal cancer. Sci Rep 2017;7: 43393.
42. Dohna M, Reincke M, Mincheva A, Allolio B, Solinas-Toldo S, Lichter P. Adrenocortical carcinoma is characterized by a high frequency of chromosomal gains and high-level amplifications. Genes Chromosomes Cancer 2000;28: 145-52.
43. Durand J, Lampron A, Mazzuco TL, Chapman A, Bourdeau I. Characterization of differential gene expression in adrenocortical tumors harboring beta-catenin (CTNNB1) mutations. J Clin Endocrinol Metab 2011;96: E1206-11.
☒
44. Bakin AV, Tomlinson AK, Bhowmick NA, Moses HL, Arteaga CL. Phosphatidylinositol 3- kinase function is required for transforming growth factor beta-mediated epithelial to mesenchymal transition and cell migration. J Biol Chem 2000;275: 36803-10.
45. Lorentzen A, Lewinsky RH, Bornholdt J, Vogel LK, Mitchelmore C. Expression profile of the N-myc Downstream Regulated Gene 2 (NDRG2) in human cancers with focus on breast cancer. BMC Cancer 2011;11: 14.
46. Kim JT, Kim JW, Kang YH, Kim KD, Lee SJ, Choi SC, Kim KS, Chae SK, Kim JW, Lim JS, Lee HG. NDRG2 and PRA1 interact and synergistically inhibit T-cell factor/beta-catenin signaling. FEBS Lett 2012;586: 3962-8.
47. Nakahata S, Ichikawa T, Maneesaay P, Saito Y, Nagai K, Tamura T, Manachai N, Yamakawa N, Hamasaki M, Kitabayashi I, Arai Y, Kanai Y, et al. Loss of NDRG2 expression activates PI3K-AKT signalling via PTEN phosphorylation in ATLL and other cancers. Nat Commun 2014;5: 3393.
48. Cao W, Yu G, Lu Q, Zhang J. Low expression of N-myc downstream-regulated gene 2 in oesophageal squamous cell carcinoma correlates with a poor prognosis. BMC Cancer 2013;13: 305.
49. Oh SS, Kim D, Kim DH, Chang HH, Sohn KC, Kim KH, Jung SH, Lee BK, Kim JH, Kim KD. NDRG2 correlated with favorable recurrence-free survival inhibits metastasis of mouse breast cancer cells via attenuation of active TGF-beta production. Carcinogenesis 2012;33: 1882-8.
50. Barreau O, Assie G, Wilmot-Roussel H, Ragazzon B, Baudry C, Perlemoine K, Rene-Corail F, Bertagna X, Dousset B, Hamzaoui N, Tissier F, de Reynies A, et al. Identification of a CpG island methylator phenotype in adrenocortical carcinomas. J Clin Endocrinol Metab 2013;98: E174-84.
Acc
Figure legends
Tale
☒
Figure 1. Inverse correlation between miR-483-5p or miR-139-5p and NDRG2 or NDRG4 expression in human adrenocortical carcinoma samples. A-D, Expression of miR-483-5p (A), miR-139-5p (B), NDRG2 (C) and NDRG4 (D) was assessed by RT-qPCR in adrenocortical adenomas (ACA), non-aggressive adrenocortical carcinomas (naACC), aggressive adrenocortical carcinomas (aACC) as well as in the NCI H295R cell line and normal adrenal cortex. Values represent miR-483-5p, miR-139-5p, NDRG2 and NDRG4 expression levels in tumor samples as a ratio to their expression levels in a pool of 3 normal adrenocortical tissue samples. White squares correspond to metastatic ACCs in the aACC group. The miRNA/RNU48 or NDRG/(HPRT and RPL13A) ratio in normal adrenal cortex was assigned as 1.0. * , p<0.05, ** , p<0.01, *** p<0.001. Comparison of miR-483-5p expression in naACC and aACC almost reached statistical significance (p=0.05). E-H, Pearson correlation of miR-483-5p and NDRG2 mRNA expressions or miR-139-5p and NDRG4 mRNA expressions in naACC and aACC samples. MiRNA/RNU48 or NDRG/(HPRT and RPL13A) normalized expressions were used.
☒
Figure 2. Integrative approach for miR-139-5p and miR-483-5p target identification. A, Diagram depicting the approach used to identify inversely correlated miRNA-mRNA pairs. Analyses were based on transcriptome and miRnome profiling from de Reyniès et al. 6 and Chabre et al. 14, respectively. B, Venn diagrams showing the number of predicted target genes whose expression was inversely correlated to that of miR-139-5p or miR-483-5p and which fulfill selection criteria: log2 fold-change (LFC) ≤-1 (p<0.05) and identification by at least 3 prediction algorithms.
☒
☒ ☒ ☒ Figure 3. Immunohistochemical analysis of NDRG2 and NDRG4 protein expression in FFPE adrenocortical tumor samples. Shown are representative images from a cohort of 12 aACC, 8 naACC and 10 ACA, which is described in Supporting Information Material and Methods. Immunohistochemistry was performed as described in details in Supporting Information Material and Methods. A strong staining was observed in both adjacent normal adrenal cortex (Nc) and adrenocortical adenoma (ACA) consistent with retained expression. Non-aggressive adrenocortical carcinoma (naACC) and aggressive adrenocortical carcinoma (aACC) scored complete loss of expression of NDRG2 based on the absence of staining compared with ACA. naACC retained partial expression of NDRG4 in some areas of the tumor while loss of NDRG4 expression was seen in aACC. HES: haematoxylin/eosin staining; Tm: tumor; n: tumor nodule. Scale Bar: 1 mm.
☒
Figure 4. miR-483-5p and miR-139-5p negatively regulate endogenous NDRG2 and NDRG4 expression, respectively, in adrenocortical carcinoma cells. A and B, NCI H295R cells were transiently transfected for 24 h, 48 h or 72 h with miR-483-5p or miR-139-5p inhibitors (Inhib), or Negative Control miR inhibitor (NC miR Inhib) then the efficiency of miRNA repression was checked by RT-qPCR. C and D, Endogenous NDRG2 and NDRG4 mRNA expressions were assessed by RT-qPCR 72 h post-transfection with miRNA inhibitors. Data are Means ± S.E.M from 3 independent experiments performed in duplicate .* , p<0.05, ** , p<0.01. E and F, Western-Blot analysis of NDRG2 and NDRG4 protein levels in NCI H295R transfected with negative control miRNA (NC miR), miR-483-5p mimics, miR-139-5p mimics, NC miR Inhib, miR-483-5p Inhib or miR-139-5p Inhib. a-Tubulin served as a loading control. Shown is a representative experiment of 3 independent ones.
☒
☒
Figure 5. MiR-483-5p and miR-139-5p directly target NDRG2 and NDRG4 3’UTR, respectively. A and B, Luciferase constructs and putative miR-483-5p and miR-139-5p binding sites in wild type (wt) NRDG2 3’UTR and NDRG4 3’UTR are shown. The 2 corresponding mutant binding sites (mt-3’UTR) are reported in bold. C and D, Luciferase activity in NCI H295R cells upon transfection of miR-483-5p and miR-139-5p inhibitors or NC miR Inhib and NDRG-3’UTR-luciferase constructs. E and F, Luciferase activity in HEK293T cells upon transfection of miR-483-5p and miR-139-5p mimics or NC miR and NDRG2 wt-3’UTR, NDRG4 wt-3’UTR, NDRG2 mt-3’UTR or NDRG4 mt-3’UTR. HEK293T cells were used for miRNA mimic transfections due to the already high levels of endogenous miR-483-5p and miR-139-5p in NCI H295R cells. Shown are the means ± S.E.M of 4-5 independent experiments performed in duplicate. ** , p<0.01; ns: non-significant.
☒
☒
Figure 6. Inhibition of miR-483-5p and miR-139-5p or ectopic expression of NDRG2 or NDRG4 prevent adrenocortical carcinoma cell migration and invasion. A and C, NCI H295R cells were transfected with NC miR Inhib, miR-483-5p or miR-139-5p inhibitors, then analyzed in Transwell migration (A) and invasion (C) assays. B and D, quantification of cell number in five random fields per well (mean ± S.E.M) in four independent experiments performed in duplicate (n=4). E, NCI H295R cell invasiveness is impeded by ectopic expression of NDRG2 or NDRG4. NCI H295R cells were transfected with empty pCMV6 plasmid or pCMV6 encoding for NDRG2 or NDRG4 coding sequences then analyzed in invasion assay. NDRG2 ☒ ☒ ☒
and NDRG4 expression was checked by western blot (Supporting Information Fig. S6). F, quantification of invaded cells in 3 independent experiments (n=3). * , p<0.05; ** , p<0.01; *** p<0.001; ns: non-significant. Scale bar, 20 um. G, Inhibition of miR-483-5p and miR-139-5p
☒ represses EMT markers in adrenocortical carcinoma cells. Western blotting of Snail, Twist, N- cadherin, E-cadherin and Vimentin in NCI H295R cells transfected with NC miR Inhib, miR-
483-5p or miR-139-5p inhibitors. ß-Actin was used as loading control. H, Inhibition of miR- 483-5p and miR-139-5p reduces anchorage-independent growth of adrenocortical carcinoma cells. NCI H295R cells were transfected as described above and analyzed in soft agar colony formation assays. Higher magnifications of the insets are shown in the lower panels. I, quantification of colony number in 4 independent experiments. * p<0.05, ** p<0.01, *** p<0.001.
Accepted Art
A
miR-483-5p
Normalized miRNA expression (tumor/normal adrenal ratio)
B
Normalized miRNA expression (tumor/normal adrenal ratio)
miR-139-5p
120
**
17.5
*
.
15.0
100
0
.
Q
80
12.5
10.0
O
60
7.5
40
5.0
20
2.5
0
4
ACA
naACC
aACC
NCI H295R
0.0
ACA
naACC
aACC
NCI H295R
C
Normalized mRNA expression (tumor/normal adrenal ratio)
NDRG2
D
2.5
Normalized mRNA expression (tumor/normal adrenal ratio)
NDRG4
10.0
.
2.0
1.0
.
1.5
.’
1.0
0.1
0.5
0000
O
0.0
ACA
naACC
aACC
NCI H295R
0.01
ACA
naACC
aACC
NCI H295R
E
F
5.0
3.5
NDRG2 expression
naACC
.
naACC
4.0
R =- 0.22
NDRG4 expression
3.0
R=0.03
p=0.57
2.5
p=0.93
3.0
2.0
.
2.0
1.5
.
1.0
-
1.0
0.5
0.0
0.0
0
1
2
3
4
5
0.00
0.05
0.10
0.15
0.20
0.25
0.30
miR-483-5p expression
miR-139-5p expression
G
H
1.75
0.20
NDRG2 expression
1.50
aACC
aACC
R =- 0.62
NDRG4 expression
.
1.25
p=0.04
0.15
.
R =- 0.67
p=0.02
1.00
0.75
0.10
0
0.50
0.25
0.05
O
0.00
O
1
2
3
4
5
6
7
0.00
00
0
1
2
3
4
5
miR-483-5p expression
miR-139-5p expression
A
Human adrenocortical tumor transcriptome analysis
Human adrenocortical tumor miRnome analysis
mRNAs downregulated in ACCs
miR-139-5p and miR-483-5p upregulated in ACCs
In silico predicted target mRNA of miR-139-5p or a miR-483-5p
mRNAs belonging to both groups
B
mRNAs downregulated in ACC LFC≤-1 and p<0.05
Predicted miR-139-5p target mRNAs ≥ 3 target prediction algorithms p<0.05
380
9
2605
mRNAs downregulated in ACC LFC≤-1 and p<0.05
Predicted miR-483-5p target mRNAs ≥ 3 target prediction algorithms p<0.05
359
30
1772
HES
NDRG2
NDRG4
tm
tm
tm
ACA
Nc
Nc
tm
tm
naACC
Nc
Nc
Nc
Nc
aACC
tm
n
tm
n
A
B
Normalized expression
1.50
miR-483-5p
Normalized expression
1.75
miR-139-5p
1.25
1.50
1.00
1.25
0.75
1.00
0.75
0.50
0.50
0.25
0.25
0.00
24 h
48 h
72 h
0.00
24 h
48 h
72 h
NC miR Inhib
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
C
NDRG2
D
NDRG4
*
2.5
p= 0.010
**
3.0
p=0.009
Normalized expression
Normalized expression
2.0
2.5
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0
NC miR Inhib
miR-483-5p Inhib
NC miR Inhib
miR-139-5p Inhib
E
F
NC miR Inhib
miR-483-5p Inhib
NC miR
miR-483-5p
NC miR Inhib
miR-139-5p Inhib
NC miR
miR-139-5p
40 kDa
NDRG2
45 kDa
NDRG4
55 kDa
a-tubulin
55 kDa
a-tubulin
NDRG2/tub ratio:
1.0
1.6
1.0
0.56
NDRG4/tub ratio:
1.0
1.5
1.0
0.41
CMV
Luciferase
NDRG2-3’UTR (nt 1432-2120)
3’UTR 5’ … UGUGGAGGGUGCCCUCCCGUCUC … 3’
C … 3’ 3’UTR 5’ … UGUGGAGGGUGCCCU
CMV
Luciferase
NDRG4-3’UTR (nt 1896-3043)
3’UTR 5’ … GAAUUCAUCCCUCUUAÇUGUAGA … 3’
5’
3’UTR 5’ … GAAUUCAUCCCUCUU .3’
11
5’
3’UTR
3’UTR
**
**
International Journal of Cancer
A
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
B
Migration
Migratory cells/field
70
**
60
50
40
*
30
20
10
0
NC miR Inhib
miR-483-5p
miR-139-5p Inhib
Inhib
C
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
D
Invasive cells/field
175
150
Invasion
125
100
75
ns
50
25
0
NC miR Inhib
miR-483-5p
miR-139-5p Inhib
E
Inhib
pCMV6
pCMV6-NDRG2
pCMV6-NDRG4
F
300
250
Invasion
Invasive cells/field
200
150
ns
100
50
0
pCMV6
pCMV6-
pCMV6- NDRG4p
NDRG2
G
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
Snail
N-cadherin
Vimentin
Twist
E-cadherin
ß-Actin
ß-Actin
ß-Actin
H
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib
I
250
**
Colony number
200
150
*
100
50
0
NC miR Inhib
miR-483-5p Inhib
miR-139-5p Inhib