Abstract
Adrenocortical carcinoma (ACC) is an aggressive malignancy with high rates of recurrence following surgical resection. Long noncoding RNAs (lncRNAs) play an important role in cancer development. Pathogenesis of adrenal tumours have been characterised by mRNA, microRNA and methylation expression signatures, but it is unknown if this extends to lncRNAs. This study describes lncRNA expression signatures in ACC, adrenal cortical adenoma (ACA) and normal adrenal cortex (NAC) and presents lncRNAs associated with ACC recurrence to identify novel prognostic and therapeutic targets. RNA was extracted from freshly frozen tissue with confirmation of diagnosis by histopathology. Focused lncRNA and mRNA transcriptome analysis was performed using the ArrayStar Human LncRNA V3.0 microarray. Differentially expressed lncRNAs were validated using quantitative reverse transcriptase-PCR and correlated with clinical outcomes. Microarray of 21 samples (ten ACCs, five ACAs and six NACs) showed distinct patterns of lncRNA expression between each group. A total of 956 lncRNAs were differentially expressed between ACC and NAC, including known carcinogenesis-related lncRNAs such as H19, GAS5, MALAT1 and PRINS (P≤0.05); 85 lncRNAs were differentially expressed between ACC and ACA (P≤0.05). Hierarchical clustering and heat mapping showed ACC samples correctly grouped compared with NAC and ACA. Sixty-six differentially expressed lncRNAs were found to be associated with ACC recurrence (P≤0.05), one of which, PRINS, was validated in a group of 20 ACCs and also found to be associated with metastatic disease on presentation. The pathogenesis of adrenal tumours extends to lncRNA dysregulation and low expression of the lncRNA PRINS is associated with ACC recurrence.
Introduction
Adrenocortical carcinoma (ACC) is an aggressive malignancy with limited treatment options and an overall survival of <35% (Abiven et al. 2006). ACC commonly recurs following surgical resection and the use of adjuvant mitotane is currently under investigation in the hope of reducing recurrence rates (http://clinicaltrials.gov, NCT00777244: Efficacy of Adjuvant Mitotane Treatment (ADIUVO); available at http://clinicaltrials.gov/show/NCT00777244, accessed 9/8/2014). Unfortunately, mitotane is associated with significant side effects, limiting its use, and a better understanding of ACC pathogenesis has therefore been identified as a key component in improving outcomes (Ronchi et al. 2014). Understanding of ACC pathogenesis is being greatly advanced by the advent of high-throughput genomic medicine, which has shown that ACCs have distinct genome-wide expression, microRNA expression and methylation profiles compared with adrenal cortical adenomas (ACAs) and normal adrenal cortex (NAC) (Giordano et al. 2009, Soon et al. 2009, Assie et al. 2012, Rechache et al. 2012).
An additional genomic marker of ACC pathogenesis could be long noncoding RNAs (lncRNAs). The discovery of the existence of thousands of lncRNAs and their diverse functions is causing a shift in the knowledge of biology (Cech & Steitz 2014, Morris & Mattick 2014). The lncRNAs are defined as RNA transcripts longer than 200 nucleotides that do not encode protein and can localise to the nucleus or cytoplasm (Li & Chang 2014). The expression of lncRNA is more tissue specific than protein-coding genes, suggesting that they have distinct functions in specific cells (Derrien et al. 2012). The functions of lncRNA include epigenetic silencing, transcriptional regulation, splicing regulation and acting as decoys for proteins and microRNAs (Morris & Mattick 2014). Furthermore, lncRNA dysregulation is being recognised as an important part in the development of cancer (Prensner & Chinnaiyan 2011) and recent exome sequencing and SNP array analysis has supported that lncRNA dysregulation may play a role in the pathogenesis of ACC (Assie et al. 2014). However, with these studies, lncRNAs have not been the intended target of the original analysis and as lncRNAs generally have a lower transcription abundance than other RNAs, they can be more difficult to detect using next-generation sequencing technology (Ørom et al. 2010, Du et al. 2013).
Therefore, in this study, we performed focused lncRNA expression profiling using lncRNA microarrays to identify differentially expressed lncRNAs in ACCs compared with NAC and ACAs. We further investigated whether these lncRNAs expression levels are associated with ACC clinical outcomes such as recurrence. We have shown that multiple lncRNAs are dysregulated among ACC, NAC and ACA, and that individual lncRNAs expression levels can be associated with ACC outcome. This study identifies multiple potential diagnostic, prognostic and therapeutic targets for further functional study to better understand and improve outcomes in ACC.
Patients and methods
Patients and samples
Ethics approval was obtained from the Northern Sydney Area Health Service Human Research Ethics Committee and informed consent was obtained from all patients who participated in this study. Samples were obtained during surgery, snap frozen in liquid nitrogen and stored at −80 °C in the Neuroendocrine Tumour Bank of the Kolling Institute of Medical Research. The diagnosis of each sample was confirmed by an experienced endocrine pathologist (A J Gill) who was blinded to other data. Tumours with a Weiss score of <3 were classified as ACA and those with a Weiss score of ≥3 as ACC (Weiss 1984, Weiss et al. 1989). The ratio of tumour to non-neoplastic cells was calculated by histopathology and only samples with at least 80% tumour cells were included for analysis. NAC samples were obtained from non-paired adrenalectomy samples for non-functioning adrenal adenomas as described previously (Soon et al. 2009). The NAC was dissected by the operating surgeon who ensured that it was not adjacent to the adenoma and snap frozen. Histopathology was performed on the samples before analysis to ensure diagnosis of the samples and samples included only if they contained at least 80% NAC. Survival and recurrence outcomes were collated from the date of the surgery until a censor date of 1st May 2014. Recurrence was defined as a binary outcome (yes/no).
Tissue microarray
RNA extraction
Total RNA was extracted from ∼30 mg of fresh frozen tissue using a Qiazol protocol (RNeasy Mini Kit, Qiagen) according to the manufacturer's instructions. RNA quality was assessed using the Agilent 2100 Bioanalyser (Agilent Technologies, Englewood, CO, USA) with a minimum RNA integrity number of seven required.
Microarray profiling
The ArrayStar Human LncRNA/mRNA Expression Microarray Version 3.0 (ArrayStar, Inc., Rockville, MD, USA) was used, which includes ∼30 586 lncRNA probes and 26 109 mRNA transcripts and was constructed from six publicly available transcriptome databases and selected lncRNA publications (Arraystar Human LncRNA Array V3.0, available at www.arraystar.com/microarray/service_main.asp?id=198, accessed 12/7/2014). In the microarray profiling, the RNA labelling, microarray hybridisation, slide washing and scanning were performed based on the standard protocols of ArrayStar. Acquired array images were analysed using the Agilent Feature Extraction Software (version 11.0.1.1). Quantile normalisation and subsequent data processing were performed using the GeneSpring GX Version 12.0 Software (Agilent Technologies). After quantile normalisation of the raw data, lncRNAs/mRNAs present in at least 15 out of 21 samples were chosen for further data analyses. A selection criterion of statistical significance calculated as a corrected P value of <0.05 using the Benjamini Hochberg false discovery rate method along with a minimum fold change cut-off of 2.0 (up- or downregulated) was applied to identify differentially expressed lncRNAs/mRNAs.
Quantitative reverse transcriptase-PCR
The expression levels of lncRNA were also measured by quantitative reverse transcriptase-PCR (qRT-PCR) using TaqMan lncRNA assays (Applied Biosystems). One microgram of total RNA of each individual sample was reverse transcribed using the High capacity RNA-to-cDNA Kit (Applied Biosystems) according to the manufacturer's instructions. Normalisation was performed by the global normalisation method or by endogenous control using the mRNA TaqMan probes of ACTB and GAPDH. The Applied Biosystems 7900HT Real-Time PCR System (Applied Biosystems) was used with standard TaqMan cycling conditions. All samples were run in triplicate. The relative expression (RQ) was obtained using the ΔΔCt method and the differences between groups were assessed statistically using DataAssist Version 3.01 (Applied Biosystems).
Microarray data
The microarray data used in this study are available from the Gene Expression Omnibus, series entry GSE61359.
Statistical analysis
For sample size calculation, a difference in lncRNA expression between controls (NAC) and study cases (ACC and ACA) of 30% was estimated. Using statistical significance of <0.05, a minimum of six cases were required for >90% power and a minimum of five cases for >80% power. For the analysis of clinically associated lncRNAs and clinical data, SPSS Statistics, release 21 (IBM Corp., Armonk, NY, USA) was used with a significance of <0.05 considered as statistically significant. Continuous variables were compared using the t-test or Mann–Whitney U test depending on distribution, and proportions of categorical variables of the clinical groups were compared using the χ2 test. For analysis of lncRNAs associated with outcome, calculations were performed using the GraphPad Software (La Jolla, CA, USA). Receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC) were plotted using GraphPad. Survival estimates were calculated using the Kaplan–Meier estimate and compared using the log-rank test on two groups splitting by median expression into low- and high-expression groups.
Results
Patient samples and clinical characteristics
The lncRNA microarray included ten ACC (cases), five ACA (cases) and six NAC (controls) samples. Twenty ACC samples from the Kolling Institute Neuroendocrine Tumour Bank were used to validate the microarray findings and test lncRNAs associated with recurrence. The clinical characteristics and outcomes of the ACC patients included in this study are given in Table 1. There was no significant difference in ACC clinical characteristics between the microarray and validation group.
Clinical characteristics of ACC patients
| Microarray (n=10) | Validation (n=20) | P value | |
|---|---|---|---|
| Mean age (range) | 44.9 years (22.6–69.6) | 45.4 years (17.8–75.1) | 0.94 |
| Mean Weiss score (range) | 5.1 (3–9) | 5.0 (3–9) | 0.90 |
| Tumour size (range) | 8.8 cm (2.5–12.0) | 7.8 cm (2.5–11.0) | 0.44 |
| Stage 4 | 33% | 26% | 0.94 |
| Functional | 50% | 53% | 0.90 |
| Alive | 80% | 60% | 0.44 |
| Recurrence | 30% | 55% | 0.77 |
| Median follow-up (range) | 1.4 years (0.1–4.2) | 1.74 years (0.3–5.1) | 0.37 |
All P values >0.05, referring to non-significant difference in clinical characteristics between each group.
Differential expression of lncRNAs in ACC compared with NAC
Microarray transcriptome analyses revealed a large amount of variation in lncRNA expression between ACC and NAC samples (Fig. 1A). Unsupervised hierarchical and heat map clustering showed systematic variations in the ACC samples compared with NAC samples with no samples incorrectly grouped (Fig. 1B). Nine hundred and fifty-six lncRNAs were differentially expressed in ACC compared with NAC, of which 476 were up-regulated and 480 were down-regulated (P≤0.05). The 956 differentially expressed lncRNAs corresponded to 809 described lncRNA genes. The percentages of lncRNA genes corresponding to differentially expressed lncRNAs by chromosome are shown in Fig. 1C. The greatest difference in lncRNAs expression corresponded to chromosome 1 which amounted to 11.1% of lncRNAs differentially expressed between ACC and NAC.
Differentially expressed lncRNAs between ACC and NAC. (A) Volcano plot of differentially expressed lncRNAs between ACC and NAC. Red plots represent RNAs with at least twofold change and corrected P value <0.05. (B) Unsupervised hierarchical clustering and heat map of lncRNA expression between ACC and NAC. Each column represents a sample and each row represents a gene. High relative expression is indicated in red and low relative expression in green. The bottom line represents the samples, with ACC represented in red (C) and NAC in blue (N). (C) Differentially expressed lncRNAs by chromosome for ACC and ACA compared with NAC. Percentage refers to proportion of lncRNAs differentially expressed for each chromosome by total amount of that group (ACC or ACA). Up-regulated lncRNAs are shown as a positive value and down-regulated lncRNAs as a negative value.
Citation: Endocrine-Related Cancer 22, 1; 10.1530/ERC-14-0457
Among the up-regulated lncRNAs, the maximum fold change was 30.7 for the lncRNA gene RP11-438N16.1. Among the down-regulated lncRNAs, the maximum fold change was 94.5 for the lncRNA gene RP11-64D22.2. A complete list of differentially expressed lncRNAs is attached separately as supplemental data (Supplementary Table 1a and b, see section on supplementary data given at the end of this article).
Notably, several lncRNAs, which have established functions in cancer development and cancer progression, such as growth specific arrest 5 (GAS5), a tumour-suppressive lncRNA known to be down-regulated in breast cancer (Cheetham et al. 2013), was also found to be down-regulated in ACC. Other known cancer-related lncRNAs with differential expression included H19, metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), psoriasis-associated RNA induced by stress (PRINS), guanine nucleotide-binding protein, alpha-stimulating-antisense 1 (GNAS-AS1) and maternally expressed 3 (MEG3) (Spizzo et al. 2012, Du et al. 2013). Array results of selected carcinogenesis-related lncRNAs are given in Table 2.
Selected carcinogenesis-related differentially expressed lncRNAs between ACC and NAC
| Sequence name | Gene symbol | Regulation | P value | Fold change | RNA length | Chromosome | Strand | Relationship |
|---|---|---|---|---|---|---|---|---|
| NR_002819 | MALAT1 (Lin et al. 2006) | Up | 0.0057 | 4.19 | 8708 | Chr11 | + | Intergenic |
| NR_002785 | GNAS-AS1 (Menigatti et al. 2013) | Up | 0.0473 | 2.29 | 1158 | Chr20 | − | Intronic antisense |
| ENST00000524035 | MEG3 (Zhou et al. 2012) | Up | 0.0032 | 2.63 | 2853 | Chr14 | + | Intergenic |
| ENST00000437331 | DLX6-AS1 (Wahlestedt 2013) | Up | 0.0242 | 2.40 | 580 | Chr7 | − | Intronic antisense |
| ENST00000439725 | H19 (Gao et al. 2002) | Down | 0.0178 | 25.31 | 1929 | Chr11 | − | Intergenic |
| ENST00000448718 | GAS5 (Pickard & Williams 2014) | Down | 0.0070 | 2.54 | 565 | Chr1 | − | Intergenic |
| ENST00000443921 | PRINS (Du et al. 2013) | Down | 0.0056 | 5.59 | 2195 | Chr10 | + | Intron sense-overlapping |
| ENST00000495032 | HOTAIRM1 (Zhang et al. 2009, 2014) | Down | 0.0036 | 4.66 | 163 | Chr7 | + | Natural antisense |
To test the validity of the microarray findings, we selected lncRNAs (CREM, PRINS, PHF14, MEG3, MALAT1, DLX6-AS1 and LOC10013120) among the differentially expressed lncRNAs and analysed their expression by qRT-PCR. The validation cohort included 20 ACC samples and six NAC samples from the Kolling Institute Neuroendocrine Tumour Bank.
MALAT1, DLX6-AS1 and LOC100131320 were found in the microarray to be overexpressed in ACC (P≤0.05). This was confirmed by qRT-PCR in the validation cohort, with these lncRNAs also overexpressed (P≤0.05; Fig. 2). CREM and PRINS expression was down-regulated in ACC by microarray (P≤0.05) and also by qRT-PCR (P≤0.05) (Fig. 2). The microarray results of PHF14 and MEG3 were not confirmed in the validation cohort. There was increased expression of these lncRNAs in the microarray (P≤0.05) while PHF14 and MEG3 were found to be non-significantly reduced by qRT-PCR (P=0.57 and P=0.92) in the validation cohort.
QRT-PCR validation of lncRNA microarray analysis. FC, fold change in comparison of ACC vs NAC, *P≤0.05 and **P≤0.01 in comparison of ACC vs NAC, NS indicates P value considered non-significant. Global normalisation method was used for analysis.
Citation: Endocrine-Related Cancer 22, 1; 10.1530/ERC-14-0457
Differentially expressed lncRNA correlate with differentially expressed mRNAs
The array included probes for mRNAs as well as lncRNAs, which showed that 1523 mRNAs corresponding to 1341 protein-coding genes were significantly differentially expressed in ACC compared with NAC (Supplementary Table 2, see section on supplementary data given at the end of this article). To study correlation between differentially expressed lncRNAs and mRNAs, the Pearson correlation coefficients (PCC) were analysed for differentially expressed lncRNAs and mRNAs. Of the total differentially expressed 956 lncRNAs and 1523 mRNAs, 6396 matched lncRNA–mRNA pairs were identified with a PCC of >0.9. These correlated pairs consisted of 393 individual lncRNAs and 453 individual mRNAs (Supplementary Table 3). Among correlated genes, 96.4% showed a positive correlation between mRNA and lncRNA expression. The lncRNA–mRNA co-expression network generated by Cytoscape (Smoot et al. 2011) is shown in Supplementary Fig. 1. The lncRNA–mRNA co-expression networks can serve as a prediction platform to investigate the potential function of the lncRNAs, particularly in this study, to identify the differentially expressed lncRNA involved in carcinogenesis or cell development. Examples of correlated lncRNA–mRNAs with possible functional roles in ACC tumourigenesis are given in Table 3.
Examples of correlated lncRNA–mRNA with a potential functional role in ACC
| lncRNA gene | lncRNA sequence name | lncRNA array FC | mRNA | mRNA array FC | Correlation | Protein function |
|---|---|---|---|---|---|---|
| IKBKB | NR_033818 | 4.0 Up | IGFL2 (Emtage et al. 2006) | 3.8 Up | 0.9604 | Cell growth |
| PRINS (Du et al. 2013) | uc009xkj.1 | 5.3 Down | WFDC1 (Madar et al. 2008) | 15.9 Down | 0.9325 | Growth inhibitor |
| EMX2OS (Spigoni et al. 2010) | uc001ldf.3 | 3.2 Up | DAXX (Assie et al. 2014) | 3.1 Up | 0.9411 | Apoptosis regulation |
| RP11-181C3.2 | ENST00000546421 | 2.5 Up | BUB1 (de Reynies et al. 2009) | 7.2 Up | 0.9115 | Cell cycle regulation |
| H19 (Gao et al. 2002) | ENST00000439725 | 25.3 Down | SPON2 (Qian et al. 2012) | 3.1 Down | 0.9074 | Neuron development |
| H19 (Gao et al. 2002) | ENST00000446406 | 15.3 Down | AADAC (Probst et al. 1994) | 131.0 Down | 0.9176 | Carcinogen metabolism |
| ZEB2 | NR_033258 | 4.2 Down | IGFBP5 (Rho et al. 2008) | 5.8 Down | 0.9041 | Cell growth |
| EMX2OS (Spigoni et al. 2010) | uc001ldf.3 | 3.2 Up | VEGFB (Lautenschlaeger et al. 2013) | 2.7 Up | 0.9691 | Cell signalling |
FC, fold change of difference in array expression between ACC and NAC with corrected P≤0.05.
Differential expression of lncRNAs in ACC compared with ACAs
Eighty-five lncRNAs were differentially expressed (P≤0.05) in ACC compared with ACA (Fig. 3A), of which 37 were up-regulated and 48 were down-regulated (Supplementary Table 4, see section on supplementary data given at the end of this article). The 85 differentially expressed lncRNAs corresponded to 80 described lncRNA genes. Among the up-regulated lncRNAs, the maximum fold change was 47.2 for the lncRNA gene RP11-65L3.1. Among the down-regulated lncRNAs, the maximum fold change was 65.0 for the lncRNA gene RP11-64D22.2.
Differentially expressed lncRNAs between ACC and ACA. (A) Volcano plot of differentially expressed lncRNAs between ACCs and ACAs. Red plots represent RNAs with at least twofold change and corrected P value ≤0.05. (B) Unsupervised hierarchical clustering and heat map of lncRNA expression between ACC and ACA. Each column represents a sample and each row represents a gene. High relative expression is indicated in red and low relative expression in green. The bottom line represents the samples with ACC represented in blue (C) and ACA in red (A).
Citation: Endocrine-Related Cancer 22, 1; 10.1530/ERC-14-0457
Despite a lesser difference in the total number of differentially expressed lncRNAs between ACC and ACA compared with ACC and NAC, unsupervised hierarchical and heat map clustering also showed systematic variations in the ACC samples compared with ACA samples with no samples incorrectly grouped (Fig. 3B).
Differential expression of lncRNAs in ACAs compared with NAC
Between ACAs and NACs, 2655 lncRNAs were differentially expressed with 1999 lncRNAs up-regulated and 656 lncRNAs down-regulated (Supplementary Table 5, see section on supplementary data given at the end of this article). Among the up-regulated lncRNAs, the maximum fold change was 7.6 for the lncRNA gene RP11-65L3.1. The carcinogenesis-implicated lncRNA – MALAT1 (Lin et al. 2006) – was the second most up-regulated, with a fold change of 7.5. Among the down-regulated lncRNAs, the maximum fold change was 15.3 for the lncRNA gene RMST. The percentages of lncRNA genes corresponding to differentially expressed lncRNAs by chromosome are shown in Fig. 1C. The greatest difference in lncRNAs expression corresponded to chromosome 1, which amounted to 9.2% of lncRNAs differentially expressed between ACA and NAC.
Association of lncRNAs with ACC recurrence
To identify lncRNAs associated with ACC recurrence, the lncRNA expression levels were compared between samples associated with and without recurrence. For this analysis, outcome data were available for nine of the ten patient ACC samples included in the microarray. This analysis identified 66 lncRNAs that were associated with recurrence (Supplementary Table 6, see section on supplementary data given at the end of this article).
Of these lncRNAs, the carcinogenesis-implicated lncRNAs GAS5, FEZF1-AS1, RPL23AP82, GNAS-AS1, MALAT1 and PRINS were selected to test this association with the 20 ACC samples of the validation cohort (Table 1). In this analysis, qRT-PCR confirmed PRINS to have a higher expression for non-recurrent ACC samples compared with recurrent ACC samples (FC=2.69, P=0.008); however, the other five lncRNAs showed no significant difference between ACC with a clinical history of recurrence and that with non-recurrence (data not shown).
PRINS was found to have a decreasing level of expression among NACs, ACAs, ACC without recurrence and ACC with recurrence (Fig. 4), suggesting a possible role as a tumour suppressor. A low expression level of PRINS was also found to be associated with metastatic disease (P=0.030; Fig. 5A). To determine possible diagnostic accuracy of using PRINS as a marker of recurrence and/or metastatic disease, the AUC was determined (0.889, P=0.0044 for recurrence and 0.843, P=0.026 for metastatic disease; Fig. 5B). Using a sample splitting method, no difference was found in overall survival between high and low PRINS expression using the Kaplan–Meier method (P=0.17).
PRINS expression levels by adrenal group. Normalised expression microarray levels of the lncRNA PRINS by clinical group. Expression levels shown as median with error bars representing interquartile range, *P≤0.05 and **P≤0.01; rec, recurrence.
Citation: Endocrine-Related Cancer 22, 1; 10.1530/ERC-14-0457
PRINS expression levels in a validation cohort. (A) A comparison of the qRT-PCR lncRNA expression of PRINS in a validation cohort. Expression levels shown as median with error bars representing interquartile range, *P≤0.05 and **P≤0.01. The normalisation method used was endogenous control by ACTB and GAPDH. n=19 for metastatic group due to stage at presentation not being available for one patient. (B) Receiver operating characteristic (ROC) curve for PRINS. Area under the ROC curve (AUC)=0.889 for recurrence and AUC=0.843 for metastatic disease.
Citation: Endocrine-Related Cancer 22, 1; 10.1530/ERC-14-0457
Discussion
This study is the first to demonstrate a distinct lncRNA expression profile distinguishing among ACC, ACAs and NAC. Unsupervised hierarchical and heat map clustering shows that the ACC can be classified by lncRNA expression profile alone compared with both NACs and ACAs. These results indicate that lncRNA differential expression profiling could be used for diagnostic and therapeutic advantage in the management of adrenal tumours. This possibility is emphasised by the discovery that low expression of the lncRNA PRINS is associated with recurrence in ACC.
PRINS was described in 2005 when it was found to be overexpressed in epidermal cells of psoriatic patients (Sonkoly et al. 2005). PRINS is located on chromosome 10, is over 2 kb bases long and consists of two exons. PRINS is transcribed by RNA polymerase II and is spliced, polyadenylated and contains a high density of stop codons and lacks an extensive open reading frame (Sonkoly et al. 2005). In epidermal cells, PRINS has been shown to regulate G1P3 (IFI6), an anti-apoptotic gene (Szegedi et al. 2010). Silencing of PRINS in HeLa cells has been shown to lead to altered cell morphology and up-regulation of multiple protein-coding genes, including Homo sapiens B-cell/lymphoma 9 (BCL9), which forms part of the Wnt/β-catenin signalling pathway (Mani et al. 2009, Szegedi et al. 2010). Abnormal activation of the Wnt pathway has been associated with poor outcome in ACC and further research is needed to clarify whether PRINS contributes to this activation in a tumour suppressor manner (Ragazzon et al. 2010).
New techniques to predict recurrence are essential for the management of ACC, due to the aggressiveness of this disease and propensity to recur. In current clinical practice, following primary tumour resection, recurrence rates can be predicted by use of the Ki-67 index (>10%) and other factors (Glover et al. 2013). PRINS expression levels could be included in this algorithm, allowing high-recurrence-risk patients to be recommended for adjuvant radiotherapy and mitotane and allowing low-risk patients to avoid these potential morbid treatments. The association of PRINS with distant metastatic disease is also of great clinical interest. As with many malignancies, the natural history of ACC is varied, with some patients having long disease-free intervals and many having short disease-free intervals of <6–12 months. These differences are probably due to the underlying tumour biology. The ability for clinicians to predict patients with metastatic disease allows for patients to be identified who may benefit from more intensive monitoring, avoid potential morbid local therapies and be offered clinical trials of novel systemic treatments.
Larger cohort studies examining the use of PRINS as a biomarker are required to assess this potential clinical application. Despite the advances in knowledge of the underlying molecular pathology of ACC and ACAs, the carcinogenesis pathway of ACC remains controversial compared with many other malignancies (Lerario et al. 2014). Molecular medicine supports that both carcinomas and adenomas develop in a similar way, with activation of abnormal pathways such as the Wnt pathway observed as an early change in development of both adenomas and carcinoma (Heaton et al. 2012). It is with interest that this study has demonstrated that the highest number of differentially expressed lncRNAs were between ACAs and NACs (2655 lncRNAs), while 956 lncRNAs were dysregulated between ACC and NAC, compared with 85 lncRNAs dysregulated between ACC and ACAs. These findings suggest that changes in lncRNA expression may be an early part in the pathogenesis of both ACC and ACAs.
This study has demonstrated that there are over 1000 lncRNAs differentially expressed in adrenocortical tumours, providing a base for functional research to elicit if these changes are due to driver mutations with underlying RNA regulatory function or due to a consequence of the neoplastic process (Morris & Mattick 2014).
The lncRNAs can be both tumour suppressive and oncogenic depending on the cell type as shown by H19 (Prensner & Chinnaiyan 2011). In ACC, H19 is thought to act as a tumour suppressor and H19 down-regulation is observed in ∼90% of sporadic ACCs (Gicquel et al. 1997). In this study, H19 was reduced by a fold change of 25.3 in ACC compared with NAC.
The lncRNAs can act as a decoy for proteins. GAS5 binds directly to the DNA-binding domain of the glucocorticoid receptor, which allows it to act as a decoy, inhibiting glucocorticoid-regulated transcription in growth-arrested cells (Kino et al. 2010). As the glucocorticoid receptor is overexpressed in ACC (Tacon et al. 2009), down-regulation of GAS5 may play a role in this overexpression.
The lncRNAs can also act as translational regulators of mRNA (Morris & Mattick 2014). Our lncRNA–mRNA co-expression network analysis correlated a number of ACC-related genes that are of interest, including DAXX which has been recently discovered to be differentially expressed in a large series of ACCs analysed by exome sequencing and SNP analysis (Assie et al. 2012). DAXX was correlated with 41 different lncRNAs in our network, one of which, EMX2OS, has been shown to contribute to developmental regulation of transcription factors (Spigoni et al. 2010). Our co-expression network also identified other mRNAs of interest in ACC that were correlated with lncRNA expression including BUB1 (de Reynies et al. 2009), IGFL2 (Emtage et al. 2006) and IGFBP5 (Patel et al. 2014), which offers opportunity for further study to assess any functional relationship of this correlation.
Across all fields of molecular biology, due to the large number of lncRNAs identified, it is estimated that it will take a decade of analysis for specific lncRNA functions to be elicited (Cech & Steitz 2014).
The expression profiles of lncRNAs can distinguish among NAC, ACA and ACC. The lncRNA PRINS is a possible tumour suppressor in ACC. Lower expression levels of PRINS expression are associated with ACC recurrence and the presence of metastatic disease at presentation, identifying PRINS as a novel prognostic and therapeutic target for further investigation.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-14-0457.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
A R Glover is the holder of the 2014 Sir Roy McCaughey Surgical Research Fellowship (Royal Australasian College of Surgeons). S B Sidhu is a Sydney Medical School Foundation Fellow (University of Sydney).
Author contribution statement
A R Glover, J T Zhao, B G Robinson, P S H Soon and S B Sidhu designed the study. A R Glover, J T Zhao and S B Sidhu contributed to manuscript composition. A R Glover, J T Zhao, A J Gill and P S H Soon performed experimental techniques. A R Glover, J C Ip and J C Lee performed statistical analysis. A R Glover, J T Zhao, J C Ip, J C Lee, B G Robinson, P S H Soon and S B Sidhu revised the manuscript.
References
Abiven G, Coste J, Groussin L, Anract P, Tissier F, Legmann P, Dousset B, Bertagna X & Bertherat J 2006 Clinical and biological features in the prognosis of adrenocortical cancer: poor outcome of cortisol-secreting tumors in a series of 202 consecutive patients. Journal of Clinical Endocrinology and Metabolism 91 2650–2655. (doi:10.1210/jc.2005-2730).
Assie G, Giordano TJ & Bertherat JJ 2012 Gene expression profiling in adrenocortical neoplasia. Molecular and Cellular Endocrinology 351 111–117. (doi:10.1016/j.mce.2011.09.044).
Assie G, Letouzé E, Fassnacht M & Jouinot A 2014 Integrated genomic characterization of adrenocortical carcinoma. Nature Genetics 46 607–612. (doi:10.1038/ng.2953).
Cech TR & Steitz JA 2014 The noncoding RNA revolution – trashing old rules to forge new ones. Cell 157 77–94. (doi:10.1016/j.cell.2014.03.008).
Cheetham SW, Gruhl F & Mattick JS 2013 Long noncoding RNAs and the genetics of cancer. British Journal of Cancer 108 2419–2425. (doi:10.1038/bjc.2013.233).
Derrien T, Johnson R, Bussotti G, Tanzer A, Djebali S, Tilgner H, Guernec G, Martin D, Merkel A & Knowles DG et al. 2012 The GENCODE v7 catalog of human long noncoding RNAs: analysis of their gene structure, evolution, and expression. Genome Research 22 1775–1789. (doi:10.1101/gr.132159.111).
Du Z, Fei T, Verhaak RGW, Su Z, Zhang Y, Brown M, Chen Y & Liu XS 2013 Integrative genomic analyses reveal clinically relevant long noncoding RNAs in human cancer. Nature Structural & Molecular Biology 20 908–913. (doi:10.1038/nsmb.2591).
Emtage P, Vatta P, Arterburn M, Muller MW, Park E, Boyle B, Hazell S, Polizotto R, Funk WD & Tang YT 2006 IGFL: a secreted family with conserved cysteine residues and similarities to the IGF superfamily. Genomics 88 513–520. (doi:10.1016/j.ygeno.2006.05.012).
Gao Z-H, Suppola S, Liu J, Heikkila P, Janne J & Voutilainen R 2002 Association of H19 promoter methylation with the expression of H19 and IGF-II genes in adrenocortical tumors. Journal of Clinical Endocrinology and Metabolism 87 1170–1176. (doi:10.1210/jcem.87.3.8331).
Gicquel C, Raffin-Sanson ML, Gaston V, Bertagna X, Plouin PF, Schlumberger M, Louvel A, Luton JP & Le Bouc Y 1997 Structural and functional abnormalities at 11p15 are associated with the malignant phenotype in sporadic adrenocortical tumors: study on a series of 82 tumors. Journal of Clinical Endocrinology and Metabolism 82 2559–2565. (doi:10.1210/jcem.82.8.4170).
Giordano TJ, Kuick R, Else T, Gauger PG, Vinco M, Bauersfeld J, Sanders D, Thomas DG, Doherty G & Hammer GD 2009 Molecular classification and prognostication of adrenocortical tumors by transcriptome profiling. Clinical Cancer Research 15 668–676. (doi:10.1158/1078-0432.CCR-08-1067).
Glover AR, Ip JCY, Zhao JT, Soon PSH, Robinson BG & Sidhu SB 2013 Current management options for recurrent adrenocortical carcinoma. OncoTargets and Therapy 6 635–643. (doi:10.2147/OTT.S34956).
Heaton JH, Wood MA, Kim AC & Lima LO 2012 Progression to adrenocortical tumorigenesis in mice and humans through insulin-like growth factor 2 and β-catenin. American Journal of Pathology 181 1017–1033. (doi:10.1016/j.ajpath.2012.05.026).
Kino T, Hurt DE, Ichijo T, Nader N & Chrousos GP 2010 Noncoding RNA Gas5 is a growth arrest- and starvation-associated repressor of the glucocorticoid receptor. Science Signaling 3 ra8.
Lautenschlaeger T, George A, Klimowicz AC, Efstathiou JA, Wu CL, Sandler H, Shipley WU, Tester WJ, Hagan MP & Magliocco AM et al. 2013 Bladder preservation therapy for muscle-invading bladder cancers on Radiation Therapy Oncology Group Trials 8802, 8903, 9506, and 9706: vascular endothelial growth factor B overexpression predicts for increased distant metastasis and shorter survival. Oncologist 18 685–686. (doi:10.1634/theoncologist.2012-0461).
Lerario AM, Moraitis A & Hammer GD 2014 Genetics and epigenetics of adrenocortical tumors. Molecular and Cellular Endocrinology 386 67–84. (doi:10.1016/j.mce.2013.10.028).
Li L & Chang HY 2014 Physiological roles of long noncoding RNAs: insight from knockout mice. Trends in Cell Biology 24 p594–p602. (doi:10.1016/j.tcb.2014.06.003).
Lin R, Maeda S, Liu C, Karin M & Edgington TS 2006 A large noncoding RNA is a marker for murine hepatocellular carcinomas and a spectrum of human carcinomas. Oncogene 26 851–858. (doi:10.1038/sj.onc.1209846).
Madar S, Brosh R, Buganim Y, Ezra O, Goldstein I, Solomon H, Kogan I, Goldfinger N, Klocker H & Rotter V 2008 Modulated expression of WFDC1 during carcinogenesis and cellular senescence. Carcinogenesis 30 20–27. (doi:10.1093/carcin/bgn232).
Mani M, Carrasco DE, Zhang Y, Takada K, Gatt ME, Dutta-Simmons J, Ikeda H, Diaz-Griffero F, Pena-Cruz V & Bertagnolli M et al. 2009 BCL9 promotes tumor progression by conferring enhanced proliferative, metastatic, and angiogenic properties to cancer cells. Cancer Research 69 7577–7586. (doi:10.1158/0008-5472.CAN-09-0773).
Menigatti M, Staiano T, Manser CN, Komljenovic A, Jiricny J, Buffoli F & Marra G 2013 Epigenetic silencing of monoallelically methylated miRNA loci in precancerous colorectal lesions. Oncogenesis 2 e56. (doi:10.1038/oncsis.2013.21).
Morris KV & Mattick JS 2014 The rise of regulatory RNA. Nature Reviews. Cancer 15 423–437. (doi:10.1038/nrg3722).
Ørom UA, Derrien T, Beringer M, Gumireddy K, Gardini A, Bussotti G, Lai F, Zytnicki M, Notredame C & Huang Q et al. 2010 Long noncoding RNAs with enhancer-like function in human cells. Cell 143 46–58. (doi:10.1016/j.cell.2010.09.001).
Patel D, Ellis R, Howard B, Boufraqech M, Gara SK, Zhang L, Quezado MM, Nilubol N & Kebebew E 2014 Analysis of IGF and IGFBP as prognostic serum biomarkers for adrenocortical carcinoma. Annals of Surgical Oncology 21 3541–3547. (doi:10.1245/s10434-014-3768-5).
Pickard MR & Williams GT 2014 Regulation of apoptosis by long non-coding RNA GAS5 in breast cancer cells: implications for chemotherapy. Breast Cancer Research and Treatment 145 359–370. (doi:10.1007/s10549-014-2974-y).
Prensner JR & Chinnaiyan AM 2011 The emergence of lncRNAs in cancer biology. Cancer Discovery 1 391–407. (doi:10.1158/2159-8290.CD-11-0209).
Probst MR, Beer M, Beer D, Jenö P, Meyer UA & Gasser R 1994 Human liver arylacetamide deacetylase. Molecular cloning of a novel esterase involved in the metabolic activation of arylamine carcinogens with high sequence similarity to hormone-sensitive lipase. Journal of Biological Chemistry 269 21650–21656.
Qian X, Li C, Pang B, Xue M, Wang J & Zhou J 2012 Spondin-2 (SPON2), a more prostate-cancer-specific diagnostic biomarker. PLoS ONE 7 e37225. (doi:10.1371/journal.pone.0037225).
Ragazzon B, Libe R, Gaujoux S, Assie G, Fratticci A, Launay P, Clauser E, Bertagna X, Tissier F & de Reynies A et al. 2010 Transcriptome analysis reveals that p53 and β-catenin alterations occur in a group of aggressive adrenocortical cancers. Cancer Research 70 8276–8281. (doi:10.1158/0008-5472.CAN-10-2014).
Rechache NS, Wang Y, Stevenson HS, Killian JK, Edelman DC, Merino M, Zhang L, Nilubol N, Stratakis CA & Meltzer PS et al. 2012 DNA methylation profiling identifies global methylation differences and markers of adrenocortical tumors. Journal of Clinical Endocrinology and Metabolism 97 E1004–E1013. (doi:10.1210/jc.2011-3298).
de Reynies A, Assie G, Rickman DS, Tissier F, Groussin L, Rene-Corail F, Dousset B, Bertagna X, Clauser E & Bertherat JJ 2009 Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. Journal of Clinical Oncology 27 1108–1115. (doi:10.1200/JCO.2008.18.5678).
Rho SB, Dong SM, Kang S, Seo SS, Yoo CW, Lee DO, Woo JS & Park SY 2008 Insulin-like growth factor-binding protein-5 (IGFBP-5) acts as a tumor suppressor by inhibiting angiogenesis. Carcinogenesis 29 2106–2111. (doi:10.1093/carcin/bgn206).
Ronchi CL, Kroiss M, Sbiera S, Deutschbein T & Fassnacht M 2014 European Journal of Endocrinology Prize Lecture 2014: Current and evolving treatment options in adrenocortical carcinoma – where do we stand and where do we want to go? European Journal of Endocrinology 171 R1–R11. (doi:10.1530/EJE-14-0273).
Smoot ME, Ono K, Ruscheinski J, Wang P-L & Ideker T 2011 Cytoscape 2.8: new features for data integration and network visualization. Journal of Gerontology 27 431–432. (doi:10.1093/bioinformatics/btq675).
Sonkoly E, Bata-Csörgo ZZ, Pivarcsi A, Polyanka H, Kenderessy-Szabo A, Molnar G, Szentpali K, Bar L, Megyeri K & Mandi Y et al. 2005 Identification and characterization of a novel, psoriasis susceptibility-related noncoding RNA gene, PRINS. Journal of Biological Chemistry 280 24159–24167. (doi:10.1074/jbc.M501704200).
Soon PSH, Tacon LJ, Gill AJ, Bambach CP, Sywak MS, Campbell PR, Yeh MW, Wong SG, Clifton-Bligh RJ & Robinson BG et al. 2009 miR-195 and miR-483-5p identified as predictors of poor prognosis in adrenocortical cancer. Clinical Cancer Research 15 7684–7692. (doi:10.1158/1078-0432.CCR-09-1587).
Spigoni G, Gedressi C & Mallamaci A 2010 Regulation of Emx2 expression by antisense transcripts in murine cortico-cerebral precursors. PLoS ONE 5 e8658. (doi:10.1371/journal.pone.0008658).
Spizzo R, Almeida MI, Colombatti A & Calin GA 2012 Long non-coding RNAs and cancer: a new frontier of translational research? Oncogene 31 4577–4587. (doi:10.1038/onc.2011.621).
Szegedi K, Sonkoly E, Nagy N, Németh IB, Bata-Csörgo ZZ, Kemény L, Dobozy AA & Széll M 2010 The anti-apoptotic protein G1P3 is overexpressed in psoriasis and regulated by the non-coding RNA, PRINS. Experimental Dermatology 19 269–278. (doi:10.1111/j.1600-0625.2010.01066.x).
Tacon LJ, Soon PSH, Gill AJ, Chou AS, Clarkson A, Botling J, Stalberg PLH, Skogseid BM, Robinson BG & Sidhu S et al. 2009 The glucocorticoid receptor is overexpressed in malignant adrenocortical tumors. Journal of Clinical Endocrinology and Metabolism 94 4591–4599. (doi:10.1210/jc.2009-0546).
Wahlestedt C 2013 Targeting long non-coding RNA to therapeutically upregulate gene expression. Nature Reviews. Drug Discovery 12 433–446. (doi:10.1038/nrd4018).
Weiss LM 1984 Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. American Journal of Surgical Pathology 8 163–169. (doi:10.1097/00000478-198403000-00001).
Weiss LM, Medeiros LJ & Vickery AL 1989 Pathologic features of prognostic significance in adrenocortical carcinoma. American Journal of Surgical Pathology 13 202–206. (doi:10.1097/00000478-198903000-00004).
Zhang X, Lian Z, Padden C, Gerstein MB, Rozowsky J, Snyder M, Gingeras TR, Kapranov P, Weissman SM & Newburger PE 2009 A myelopoiesis-associated regulatory intergenic noncoding RNA transcript within the human HOXA cluster. Blood 113 2526–2534. (doi:10.1182/blood-2008-06-162164).
Zhang X, Weissman SM & Newburger PE 2014 Long intergenic non-coding RNA HOTAIRM1 regulates cell cycle progression during myeloid maturation in NB4 human promyelocytic leukemia cells. RNA Biology 11 777–787. (doi:10.4161/rna.28828).
Zhou Y, Zhang X & Klibanski A 2012 MEG3 noncoding RNA: a tumor suppressor. Journal of Molecular Endocrinology 48 R45–R53. (doi:10.1530/JME-12-0008).
(A R Glover and J T Zhao contributed equally to this work)
