MiR-486-3p was downregulated at microRNA profiling of adrenals of multiple endocrine neoplasia type 1 mice, and inhibited human adrenocortical carcinoma cell lines

Adrenocortical carcinoma is a rare aggressive disease commonly recurring regardless of radical surgery. Although data on genomic alterations in malignant tumors are accumulating, knowledge of molecular events of importance for initiation of adrenocortical transformation is scarce. In an attempt to recognize early molecular alterations, we used adrenals from young multiple endocrine neoplasia type 1 conventional knock-out mice (Men1+/−) closely mimicking the human MEN1 trait (i.e. transformation of pituitary, parathyroid, endocrine pancreatic, and adrenocortical cells). MicroRNA array and hierarchical clustering showed a distinct pattern. Twenty miRNAs were significantly upregulated and eleven were downregulated in Men1+/− compared to wild type littermates. The latter included the known suppressor miRNA miR-486-3p, which was chosen for transfection in human adrenocortical carcinoma cell lines H295R and SW13. Cell growth decreased in miR-486-3p overexpressing clones and levels of the predicted target gene fatty acid synthase (FASN) and its downstream product, palmitic acid, were lowered. In conclusion, heterozygous inactivation of Men1 in adrenals results in distinct miRNA profile regulating expression of genes with impact on tumorigenesis, e.g. transcription, nucleic acid and lipid metabolism. Low levels of miR-486-3p in the early stages of transformation may contribute to proliferation by increasing FASN and thus fatty acid production. FASN as a potentially druggable target for treatment of the devastating disease adrenocortical carcinoma warrants further studies.


Results
Differentially expressed miRNAs. The miRNA array profiling, performed on RNA prepared from the adrenal glands of ten Men1 +/− and ten wild type (Men +/+ ) mice revealed a distinct expression pattern. A hierarchical cluster analysis showed that nine out of ten Men1 +/− mice were in the same cluster and also nine out of ten Men1 +/+ mice were grouped together (Fig. 2). Eleven miRNAs were significantly downregulated and 20 were significantly upregulated in Men1 +/− compared to Men1 +/+ (P < 0.05) ( Table 1). The ratios of up-or down-regulation among significantly differentially expressed miRNAs ranged from 20 to 106%, mean ratio 42.3%. Seven of the differentially regulated miRNAs have earlier been reported as either suppressor or oncogene miRNAs. Among the eleven downregulated miRNAs in Men1 +/− three were known tumor suppressors (miR-486-3p, miR-330-5p and miR-214-5p) and these three were most highly ranked according to fold change (−1.33, −1.14 and −1.05, respectively). On the other hand, suppressor miR-497-5p and miR-195a-5p were significantly upregulated. Oncogenes miRNAs miR-494-3p and miR-132-3p were significantly enhanced in Men1 +/− . Selection of MiR-486-3p and six potential target genes. We selected the suppressor miR-486-3p for further investigation because it was top ranked among downregulated miRNAs in Men1 +/− , showing a foldchange of -1.33 (Table 1). In addition, among the multitude of predicted potential miR-486-3p target genes altogether six genes (ALDH2, FASN, GDI1, HINT1, KCND3 and MDGA1) ( Table 2) corresponded to results from proteomics profiling of adrenals of the same Men1 knock-out strain 24 . Increased expression of one or more of these proteins, as a result of decreased expression of miR-486-3p, may thus be of relevance for adrenocortical transformation. This hypothesis was further tested in expression studies in human adrenocortical cell lines (below).  Overexpression of miR-486-3p inhibited palmitic acid synthesis. To further understand how the increased levels of miR-486-3p followed by decreased expression of FASN may impact adrenocortical proliferation we set out to analyze the downstream product of FASN activity, palmitic acid. Palmitic acid synthesis was significantly reduced, both in H295R and SW13 cells expressing miR-486-3p compared to miR-NC cells (p < 0.05) (Fig. 8).

Discussion
Over the past decade emerging evidence strongly indicate that miRNAs play important roles in virtually every biological process and many human diseases. Notably, all types of neoplasms analyzed so far, including benign tumors, reveal miRNA alterations. Dysregulation of miRNAs influences all stages of transformation, from tumor initiation to dissemination, by regulating not only protein translation but also transcription and post-transcription processes 25 , thus they impact both nuclear and cytoplasmatic activity. Many miRNAs have oncogenic effects and are overexpressed in cancer 26,27 , whereas others that show suppressor capabilities may be underexpressed or www.nature.com/scientificreports/ deleted, also in human adrenocortical carcinoma [28][29][30][31][32][33][34] . They can be released to the circulation, and since miRNAs are remarkably stable, also in blood, they have the potential to function as clinical biomarkers 35,36 .
In the present study, we aimed at identifying miRNAs that might be of importance for initiation of tumorigenesis in adrenocortical cells and thus chose to study adrenals from a Men1 knock-out mouse model showing few early signs of transformation. We chose to use a conventional knock-out mouse model, instead of a conditional, since the former mice have the same genotype as MEN1 patients, i.e. germ line heterozygous inactivation and preserved menin production from the retained wild type allele in the vast majority of adrenocortical lesions 2,37 . Since the goal was to recognize early molecular alterations more or less solely dependent on constitutionally low Men1 gene dosage (heterozygous mice), we chose to compare the miRNA array profile of the adrenals of 10 months old Men1 +/− mice to that of the adrenals of equally young Men1 +/+ animals. The Men1 +/− adrenals were not significantly larger than those of the wild type of the same strain, and they all expressed menin (Fig. 1).  We deliberately avoided to use older mice with long-standing more prominent adrenal enlargements that might have acquired a multitude of additional mutations not necessarily representing a direct result of Men1 being a haploinsufficient suppressor. Altogether 31 miRNAs were identified as significantly differentially expressed in Men1 +/− mice compared to Men1 +/+ . The ratios of individual miRNAs between the genotypes were not particularly high, mean 42.3%, (range 20-106%), but the sample size, ten of each genotype, was obviously enough to find a fair number of significant alterations in such early proliferations. The majority of these 31 miRNAs were not earlier known as to be involved in tumor development, and the actual biological relevance of these finding warrants further studies. However, seven of the dysregulated miRNAs in Men1 +/− adrenals were already suggested to play active roles in tumorigenesis of different cancer types. The oncogenic miRNAs, miR-132-3p and miR-494-3p were significantly enhanced in Men1 +/− adrenals. These two are known to participate in regulation of transcription and proliferation in pancreatic cancer and glioma 26,27 . A somewhat intriguing finding was the differential expression of the suppressor miR-15 family; miR-497-5p and miR-195a-5p were upregulated in proliferating Men1 +/− adrenals. These miRNAs are known to be downregulated in human adrenocortical carcinomas compared to benign adenomas 16 . On the other hand, in the same study the miRNA profile of benign adenomas was also compared to that of normal adrenal cortices, and the latter comparison showed results in concordance with ours; miR-497-5p and -195a-5p were upregulated in benign adenomas compared to normal adrenals 16 .
We also found that three earlier described suppressor miRNAs (miR-214-5p, miR-330-5p and miR-486-3p) were significantly downregulated in Men1 +/− adrenals. MiR-214-5p has been describes as a key hub in control of cancer networks and is frequently differentially regulated 38 , but it is also reported to be downregulated in childhood as well as adult adrenocortical carcinomas [29][30][31]34 . MiR-330-5p has revealed antiproliferative effects in colorectal cancer 32 and the tumor suppressor effect might be conducted via the mTOR/Akt-pathway 28 . Interestingly, the MEN1 protein itself has been suggested to be a negative regulator of AKT kinase activity 39 . The suppressor miR-486-3p has previously been studied in several tumor types; it regulated tumor progression in gastric cancer 40 and revealed anti-proliferative effects in breast cancer 33 . Moreover, miR-486-3p was found to be underexpressed    30 . In the present study, miR-486-3p held the highest rank by means of level of differential expression being downregulated by 60% compared to Men1 +/+ adrenals. One of many targets for miR-486-3p is the FASN gene. Accumulating evidence show that the enzyme FASN has huge impact on fatty acid synthesis pathways and links these processes to glucose metabolism [41][42][43] . The main function of FASN is to catalyze the synthesis of long chain saturated fatty acids, e.g. palmitate, from acetyl-CoA and malonyl-CoA 44 . Moreover, increased levels of FASN in cancer cells has been correlated to poor prognosis, and inhibition of FASN results in apoptosis of cancer cells [45][46][47] . We hypothesize that downregulation of miR-486-3p may be a key event in early adrenal tumorigenesis and a result of the MEN1 heterozygosity per se. Low levels of miR-486-3p might lead to increased levels of FASN and thus increased supply of lipids which in turn is a prerequisite for membrane biogenesis in proliferating cells. It would indeed have been interesting to test this    www.nature.com/scientificreports/ We conclude that heterozygous inactivation of Men1 in mouse adrenal tissue induces a specific miRNA profile, potentially regulating expression of target genes with effect on cellular processes in tumorigenesis, including transcription, nucleic acid metabolism, and lipid metabolism. We suggest that low levels of miR-486-3p might be an important early event in adrenocortical tumorigenesis and contributes to transformation by increasing the levels of FASN which in turn increases fatty acid production. Advanced adrenocortical carcinoma is a devastating disease with limited therapeutic options; FASN as a potential druggable target warrants further studies.

Materials and methods
Animals. The conventional germline heterozygous Men1 mouse model was a kind gift by Professor Hayward of the Queensland Institute of Medical Research, Herston, Australia 5 . The phenotype of this strain mimics the human MEN 1 trait, i.e. development of macroscopically detectable classical Men1 target lesions in the young adult mouse, from about nine months of age, followed by onset of proliferation of the adrenal glands. Experiments and animal maintenance were approved by and performed according to the guidelines and regulations of the local committees for animal care at Uppsala University (Permit Number: C187/14).
Since we aimed at recognizing early effects of Men1 heterozygosity in the context of adrenal transformation, we chose ten months old animals to be compared to Men1 +/+ of the same age and strain. Ten of each genotype were analyzed; we used eight male and two female Men1 +/− , as well as six male and four female Men1 +/+ mice. Size of the dissected adrenal was assessed as the product of length of the adrenal body long-axis and maximum perpendicular width. As suspected, size of the adrenals varied between individuals and some of the Men +/− had slightly enlarged glands although the difference between genotypes was not significant; 8.6 ± 0.9 (SD) mm 2 in Men1 +/− compared to 6.9 ± 0.7 (SD) mm 2 in Men1 +/+ . Normal morphology and menin immunoreactivity (Fig. 1) were maintained in all specimens of both genotypes, also in the largest samples of Men1 +/− adrenal glands.
Immunohistochemistry. Sections of mouse adrenals were deparaffinized and rehydrated in alcohol and then heat-retrived in Tris-EDTA, pH 9.0. Endogenous peroxidase activity was blocked by incubating the slides with Peroxidazed (Histolab Products AB, Askim, Sweden). After using Background Sniper (Histolab) to block nonspecific staining, the primary antibody Rabbit anti-menin (Bethyl Laboratories, Texas USA) was applied in dilution 1:800. Incubation with secondary antibody Rabbit-on-Rodent HRP Polymer (Histolab) was followed by chromogen Betazoid DAB staining. The sections were counterstained with hematoxylin.
MiRNA preparation and miRNA array profiling. The dissected adrenals were immediately treated with RNAlater RNA Stabilization Reagent (QIAGEN, Hilden, Germany) in order to protect the RNA in samples. The miRNeasy Micro Kit (QIAGEN) was utilized for purification of total RNA, including miRNA from the samples. Five hundred nanogram of total RNA from each sample were used to prepare biotinylated RNA according to the FlashTag Biotin HSR RNA labeling kit (P/N 703095 Rev. 2). One hundred twenty microliters of each sample were loaded to the Affymetrix miRNA 4.1 Array Plates. Finally, the arrays were hybridized, washed, stained and scanned with the GeneTitan Multi-Channel Instrument, all according to the GeneTitan Instrument User Guide for Expression Arrays Plates (Affymetrix, Santa Clare, California, USA).
MiRNA array data analysis. The raw data was normalized in the free software Expression Console, provided by Affymetrix (http:// www. affym etrix. com), using the robust multi-array average method first suggested by Li and Wong in 2001 48 . Subsequent analysis of data set, containing only the mouse probes, was carried out in the freely available statistical computing language R (http:// www.r-proje ct. org). In order to search for differentially expressed genes between Men1 +/− and Men1 +/+ mice, an empirical Bayes moderated t-test was applied to employ the robust version of the lmFit function from the 'limma' package 49,50 . To address the problem with multiple testing, the p-values were adjusted using the method of Benjamini and Hochberg 51 .
Prediction of miR-486-3p target genes. Target genes of miR-486-3p, as the top-ranked differentially expressed miRNA in Men1 +/ -adrenal, was further investigated. The prediction was performed by using TargetScan (http:// www. targe tscan. org/) free online software program along with findings from proteomics profiling of adrenals of the same Men1 mouse model 24 . The biological functions of the potential miRNA targets were further investigated using Ensembl (http:// www. ensem bl. org/) gene ontology information.

MiRNA mimic transfection.
MiR-486-3p mimic and scrambled miRNA mimic (Thermo Fisher Scientific) were used to transfect H295R and SW13 cells. The mirVana mimic is double strand oligonucleotides mimicking mature miRNA and scrambled miRNA mimic is used as non-targeting negative control (miR-NC). The transfection experiment was performed by the reverse transfection procedure, according to the manufacturer's instructions, using Lipofectamine RNAiMAX (Thermo Fisher Scientific). A master mix including 25 pmol of miR-486-3p mimic or scrambled miRNA mimic (Thermo Fisher Scientific), 7.5 µL of Lipofectamine RNAiMAX www.nature.com/scientificreports/ and 250 µL Opti-MEM Medium (Thermo Fisher Scientific) per well was added to six-well plates, and then 6 × 10 5 H295R cells and 2 × 10 5 SW13 cells were seeded in each well. Cells were incubated in a 5% CO 2 -humidified atmosphere incubator at 37 °C for 48 h. The transfection experiments were performed at least three times and used for gene and protein expression analysis, cell proliferation and palmitic acid assay.

RNA extraction and quantitative real time PCR (QRT-PCR) analysis of miRNA and mRNA expression.
Total RNA from cells of the adrenocortical carcinoma cell lines H295R and SW13 were isolated by using the mirVana miRNA isolation kit (Thermo Fisher Scientific) according to the manufacturer's instructions. The purified RNA was eluted with nuclease-free water (Thermo Fisher Scientific) and was stored at -70ºC until further analysis. The RNA concentration was measured by using the NanoDrop 1000 (Thermo Fisher Scientific). One µg of total RNA from cell lines was reverse transcribed by the TaqMan MicroRNA Reverse Transcription Kit (Thermo Fisher Scientific) to detect miRNA expression. The primers used to analyze miRNA expression are described in the upper part of Supplementary Western blot analysis. Whole-cell protein lysates of H295R and SW13 cells were extracted by using radioimmunoprecipitation assay buffer (Sigma-Aldrich, St. Louis, Missouri, USA). Protein concentrations were determined using Coomassie Plus Better BradFord Assay (Thermo Fisher Scientific). Aliquots of 20 µg protein lysates were resolved by precast 4-20% Mini-PROTEAN TGX gels (Bio-Rad) and then transferred to 0.2-µm nitrocellulose membranes (Cell Signaling Technology, Danvers, Massachusetts, USA). PageRuler pre-stained protein ladder (Thermo Fisher Scientific) was used for the apparent size of proteins. The membranes were blocked with 5% milk in Tris-buffered saline solution containing 0.1% Tween-20 and then blotted with the primary antibody overnight at 4˚C. After washing, the membranes were incubated with horseradish peroxidase-conjugated antirabbit or anti-mouse IgG antibodies (GE Healthcare, Chicago, Illinois, USA), and proteins were visualized using ECL Plus Western Blotting Detection Systems (GE Healthcare) on a cooled charge-coupled device camera (Bio-Rad). Densitometrical analysis of the immunoblots was performed and quantified using the Imagelab software (Bio-Rad). Antibodies against Fatty acid synthase (#3180) and β-actin (#4970) were purchased from Cell Signaling Technology. Western blot analyses were performed three times.

Flow cytometry analysis. Click-IT Plus EdU Alexa Fluor 647 Flow Cytometry Assay Kit (Thermo Fisher
Scientific) was used to analysis cell proliferation in miRNA mimic transfected adrenocortical carcinoma cell lines H295R and SW13. In brief, the transfected cells were incubated with 10 µM EdU for two hours and the cells were then harvested for flow cytometry analysis. In addition, Click-IT palmitic acid azide (Thermo Fisher Scientific) was used to detect palmitic acid in miRNA mimic transfected H295R and SW13 cells. Briefly, transfected cells were incubated overnight in 25 mM palmitic acid azide and the cells then collected for flow cytometry analysis. The experiments were performed three times.