Methylation of IGF2 regulatory regions to diagnose adrenocortical carcinomas

in Endocrine-Related Cancer
View More View Less
  • 1 Department of Internal Medicine, Division of Endocrinology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
  • 2 Department of Pathology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
  • 3 Department of Histopathology, King’s College Hospital, Denmark Hill, London, UK
  • 4 Department of Surgery, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
  • 5 Department of Pathology, VU University Medical Center, Amsterdam, The Netherlands

Adrenocortical carcinoma (ACC) is a rare malignancy with a poor prognosis. Discrimination of ACCs from adrenocortical adenomas (ACAs) is challenging on both imaging and histopathological grounds. High IGF2 expression is associated with malignancy, but shows large variability. In this study, we investigate whether specific methylation patterns of IGF2 regulatory regions could serve as a valuable biomarker in distinguishing ACCs from ACAs. Pyrosequencing was used to analyse methylation percentages in DMR0, DMR2, imprinting control region (ICR) (consisting of CTCF3 and CTCF6) and the H19 promoter. Expression of IGF2 and H19 mRNA was assessed by real-time quantitative PCR. Analyses were performed in 24 ACCs, 14 ACAs and 11 normal adrenals. Using receiver operating characteristic (ROC) analysis, we evaluated which regions showed the best predictive value for diagnosis of ACC and determined the diagnostic accuracy of these regions. In ACCs, the DMR0, CTCF3, CTCF6 and the H19 promoter were positively correlated with IGF2 mRNA expression (P<0.05). Methylation in the most discriminating regions distinguished ACCs from ACAs with a sensitivity of 96%, specificity of 100% and an area under the curve (AUC) of 0.997±0.005. Our findings were validated in an independent cohort of 9 ACCs and 13 ACAs, resulting in a sensitivity of 89% and a specificity of 92%. Thus, methylation patterns of IGF2 regulatory regions can discriminate ACCs from ACAs with high diagnostic accuracy. This proposed test may become the first objective diagnostic tool to assess malignancy in adrenal tumours and facilitate the choice of therapeutic strategies in this group of patients.

Abstract

Adrenocortical carcinoma (ACC) is a rare malignancy with a poor prognosis. Discrimination of ACCs from adrenocortical adenomas (ACAs) is challenging on both imaging and histopathological grounds. High IGF2 expression is associated with malignancy, but shows large variability. In this study, we investigate whether specific methylation patterns of IGF2 regulatory regions could serve as a valuable biomarker in distinguishing ACCs from ACAs. Pyrosequencing was used to analyse methylation percentages in DMR0, DMR2, imprinting control region (ICR) (consisting of CTCF3 and CTCF6) and the H19 promoter. Expression of IGF2 and H19 mRNA was assessed by real-time quantitative PCR. Analyses were performed in 24 ACCs, 14 ACAs and 11 normal adrenals. Using receiver operating characteristic (ROC) analysis, we evaluated which regions showed the best predictive value for diagnosis of ACC and determined the diagnostic accuracy of these regions. In ACCs, the DMR0, CTCF3, CTCF6 and the H19 promoter were positively correlated with IGF2 mRNA expression (P<0.05). Methylation in the most discriminating regions distinguished ACCs from ACAs with a sensitivity of 96%, specificity of 100% and an area under the curve (AUC) of 0.997±0.005. Our findings were validated in an independent cohort of 9 ACCs and 13 ACAs, resulting in a sensitivity of 89% and a specificity of 92%. Thus, methylation patterns of IGF2 regulatory regions can discriminate ACCs from ACAs with high diagnostic accuracy. This proposed test may become the first objective diagnostic tool to assess malignancy in adrenal tumours and facilitate the choice of therapeutic strategies in this group of patients.

Introduction

Adrenocortical carcinoma (ACC) is a rare malignancy, with an estimated incidence of 0.7–2.0 cases per million population each year (Kebebew et al. 2006, Golden et al. 2009, Fassnacht et al. 2013, Kerkhofs et al. 2013). The prognosis is poor, with overall 5-year survival rates of 16–44% (Fassnacht et al. 2010). The likelihood of malignancy of an adrenal mass is based on the CT/MRI phenotype (e.g. size, Hounsfield units, contrast washout) and is decisive for surgery (Lacroix 2010). However, the discrimination of ACCs from adrenocortical adenomas (ACAs) still remains challenging on both imaging and postoperative histopathological grounds. The current pathological scoring system (Weiss score) is limited by a lack of reproducibility and lacks diagnostic accuracy in so-called ‘borderline malignant’ tumours with a Weiss score of 2 or 3 (Lau & Weiss 2009, de Krijger & Papathomas 2012). The assessment of malignant potential in adrenocortical tumour variants and/or paediatric adrenocortical neoplasms can be challenging and requires an experienced histopathologist; accordingly, applying the Weiss scoring system could tilt the diagnosis toward malignancy in oncocytic tumours and/or tumours in the paediatric setting (Dehner & Hill 2009, Papotti et al. 2014), whereas underdiagnosis could possibly be the case when confronted with myxoid variants (Lau & Weiss 2009, de Krijger & Papathomas 2012, Papotti et al. 2014). A biomarker that determines malignancy in adrenal tumours would be an important improvement in diagnostic certainty that can underpin therapeutic strategies such as adjuvant mitotane treatment and intensity and duration of follow-up (Young 2007).

The adrenal gland produces various growth factors, including the insulin-like growth factor 2 (IGF2). IGF2 is an imprinted gene primarily expressed from the paternal allele (Ohlsson et al. 1994) and is correlated with malignancy in adrenal tumours (Erickson et al. 2001, Giordano et al. 2003, de Fraipont et al. 2005, Almeida et al. 2008, Wang et al. 2014). However, there is a large variability in IGF2 expression and IGF2 expression levels are not fully discriminative for a distinction between ACC and ACA (Schmitt et al. 2006, Wang et al. 2014). By the concept of DNA methylation, several factors are supposed to regulate IGF2 expression (Fig. 1).

Figure 1
Figure 1

Overview of the imprinted region of the IGF2 gene and H19 gene on the maternal and paternal allele in the normal situation. Lines with dots represent known sites of DNA methylation. (arrow) expression of the gene. (stub-end arrow) repression of the gene. Methylation status of the DMR0 remains unclear. Figure is not at scale. DMR0, differentially methylated region 0; DMR2, differentially methylated region 2; ICR, imprinting control region, consisting of CTCF3 and CTCF6. Data from Tabano et al. (2010).

Citation: Endocrine-Related Cancer 23, 9; 10.1530/ERC-16-0266

H19, a gene whose transcript is not translated, is reciprocally imprinted with IGF2 (Rachmilewitz et al. 1992, Zhang & Tycko 1992). Low expression of the H19 gene may play a role in the development of ACCs in two different ways: H19 per se is supposed to be involved in tumour suppression and methylation of the H19 gene regulates mRNA expression of H19 and IGF2 (Hao et al. 1993). Previous research revealed a higher mean degree of methylation in the H19 promoter in ACCs compared with ACAs (Gao et al. 2002), but again with a significant overlap between both tumour entities.

Also differentially methylated regions (DMRs) play a role in the regulation of IGF2 expression. DMRs are regions in the genome with high concentration of CpGs controlling imprinting. In other types of cancer, it has been proven that hypomethylation of the DMR0 correlates with IGF2 loss of imprinting (LOI) (Baba et al. 2010). Previous research also suggests that hypomethylation of DMR2 is associated with ACCs and increased IGF2 expression (Barlaskar 2011).

The imprinting control region (ICR) is another element that contributes to regulation of IGF2. Methylation of this region determines whether IGF2 or H19 is expressed from the allele (Dejeux et al. 2009), as a result of binding of the methylation-specific CTCF protein to this region (Herold et al. 2012). Two of the key CTCF-binding sites are the third (CTCF3) and sixth (CTCF6) sites (Boissonnas et al. 2010).

In this study, DNA methylation is assessed by pyrosequencing, the most sensitive and accurate method to detect methylation at a single CpG (Quillien et al. 2012). To the best of our knowledge, we analyse for the first time methylation in a number of different IGF2 regulatory regions in benign and malignant adrenal tumours. We aimed to identify specific patterns correlating with malignancy in adrenal tumours. As such, our main research aim is to evaluate whether the methylation patterns of IGF2 regulatory regions, individually or combined, could serve as a valuable biomarker in distinguishing ACCs from ACAs. Finally, to demonstrate that methylation is involved in regulation of IGF2 expression in ACC, we examined the effect of demethylation in ACC cell lines.

Methods

Adrenocortical tissues

We obtained normal adrenals and adrenocortical tumours from patients during surgeries performed at the Department of Surgery, Rotterdam, Erasmus MC (EMC) between May 1995 and March 2014. We embedded the specimens in Tissue-Tek after resection and stored them at –80°C until analysis. All normal adrenals were collected during nephrectomy and confirmed by the pathologist as being normal. We randomly selected ACAs based on availability of tumour tissue in the past 7years. We excluded nonprimary ACCs (n=22), ACCs with insufficient DNA or RNA yield (n=7), or tumours defined as ‘borderline malignant’ by the pathologist (n=2; Supplementary Fig. 1, see section on supplementary data given at the end of this article). We collected medical history information and tumour characteristics from electronic patient records (Table 1). Follow-up data were updated from the Pathological Anatomical National Automated Archives (PALGA). An independent pathologist collected all data for pathological features, including the Weiss score, the currently used pathological scoring system for determining malignancy in adrenocortical tumours (Lau & Weiss 2009). A cutoff of 3 criteria or more present in the tumour was considered malignant, although only tumours with a clear pathological diagnosis were included in the analysis. We used the ENSAT classification to stage the tumours (Lughezzani et al. 2010). To validate our findings in an independent cohort, we used a series of frozen specimens from the VU University Medical Center (VUMC), Amsterdam, the Netherlands. DNA isolation and methylation analyses were performed at the EMC. The Weiss score was independently determined both at the VUMC and at the EMC as a central review. The study was conducted under the guidelines that had been approved by the Medical Ethics Committee of the Erasmus Medical Center.

Table 1

Characteristics of patients and adrenocortical tissues of the EMC and the VUMC cohorts.

EMCVUMC
ACCn=24ACAn=14ACCn=9ACAn=13
Mean age at diagnosis (years)50 (range 9–74)45 (range 26–61)50 (range 22–73)55 (range 17–79)
Mean follow-up (months)42 (range 1–187)37 (range 1–83)20 (range 0–68)30 (range 0–87)
Male (%)7 (29%)3 (21%)2 (22%)4 (31%)
Mean tumour size cm (s.d.)13.00 (s.d. 6.9)3.05 (s.d. 2.0)16.44 (s.d. 6.2)3.33 (s.d. 2.1) (n=12)
Secretion
 Androgens9 (38%)0 (0%)1 (13%)1 (9%)
 Glucocorticoids13 (54%)6 (43%)5 (63%)3 (27%)
 Mineralocorticoids0 (0%)5 (36%)0 (0%)1 (9%)
 Precursors3 (13%)0 (0%)0 (0%)0 (0%)
 Estradiol4 (17%)0 (0%)0 (0%)0 (0%)
 Nonsecreting8 (33%)3 (21%)3 (38%)6 (55%)
 Weiss score (mean)6.00 (range 3–8)0.14 (range 0–1)5.33 (range 3–8)0.31 (range 0–2)
ENSAT
 I0 (0%)12 (86%)0 (0%)13 (100%)
 II13 (54%)2 (14%)4 (44%)0 (0%)
 III3 (13%)0 (0%)4 (44%)0 (0%)
 IV8 (33%)0 (0%)1 (11%)0 (0%)
 Metastasis17 (71%)0 (0%)6 (67%)0 (0%)

The two patients with tumours with an unclear diagnosis from the EMC cohort are not included in the table. In the VUMC cohort, one patient with ACC and two patients with ACA did not receive a thorough hormonal workup, so data are missing. EMC, Erasmus University Medical Center; ENSAT, European Network for the Study of Adrenal Tumors; VUMC, VU University Medical Center.

Processing of adrenocortical tissues

We used several cryostat sections of 20µm to isolate RNA and DNA from Tissue-Tek embedded tissues and 5µm sequential cryostat sections for haematoxylin–eosin (H&E) staining in order to confirm that we used representative tissue specimens. We considered a tissue as representative when it contained at least 80% of tumour cells in case of ACA or ACC, or when a clear normal adrenal structure was visible at microscopic evaluation.

Cell culture

We used three available human ACC cell lines: H295R and SW13 were obtained from the American Type Culture Collection and from ECACC (Salisbury, Wiltshire, UK), respectively. HAC15 cells were a kind gift from Dr W Rainey (Department of Physiology, Medical College of Georgia, Augusta, GA, USA). Culture conditions were described in detail previously (van Koetsveld et al. 2006). We performed incubations with and without the demethylating drug 5′-Aza-2′-deoxycytidine (AZA), purchased from Sigma-Aldrich. For treatment, we added AZA to the medium for 72 and 168h. For 168h of treatment, we refreshed the medium after 72h and added AZA again. Based on the concentration of AZA that induced 50% cell growth inhibition (IC50), we chose three different concentrations of AZA (0.01, 0.05 and 1µM) for the experiments. At the end of the incubation period, we removed medium and collected cells for DNA or total RNA isolation as described below.

mRNA expression analysis

We isolated total RNA from ACC cells using the High Pure RNA Purification Kit and total RNA from tissues using the High Pure RNA Tissue Kit (both from Roche) according to manufacturer’s protocol. To synthesize cDNA, we added 500ng mRNA template to 40µL Super RT buffer (HT Biotechnology Ltd, Cambridge, UK) containing 40nmol dNTP, 20U RNA’sin, 15ng oligo-dT, 4U Super RT. After 1-h incubation at 40°C, we diluted the cDNA five times. We mixed 7.5µL TaqMan Universal PCR Master Mix (Applied Biosystems) with concentrations of primers and probes described in Supplementary Table 1, with 5µL cDNA template. Real-time quantitative PCR was performed by TaqMan Gold nuclease assay (Perkin-Elmer) and the ABI-PRISM-7900 Sequence Detection System (Perkin-Elmer), according to manufacturer’s protocol. PCR conditions were 95°C for 10min, followed by 45 cycles of 95°C for 15s and 60°C for 1min. Hypoxanthine-guanine phosphoribosyl transferase 1 (HPRT1) was used to normalize mRNA levels. We calculated PCR efficiencies (E) for the primer–probe combinations (Supplementary Table 1, Sigma-Aldrich) and calculated relative expression of genes using the comparative CT method 2−ΔΔCt.

Bisulphite conversion and pyrosequencing

We isolated DNA from ACC cells and adrenocortical tissues using the Wizard Genomic DNA Purification Kit (Promega). We treated genomic DNA (1ng to 1µg) with sodium bisulphite using the EpiTect Plus DNA Bisulfite Kit (Qiagen) following manufacturer’s protocol.

Each PCR mix of 50µL contained 0.2µM of forward and reverse primer (one of them being biotinylated), 0.25mM dNTPs, 2mM MgCl, 15µL 10× PCR buffer, 1.5U enzyme and 10ng bisulphite DNA, using the FastStart High Fidelity PCR System kit (Roche). PCR programmes were as follows: initial denaturation at 95°C for 10min, 45 cycles of 95°C for 30s, primer-specific hybrid temperature (Supplementary Table 2) for 30s, 72°C for 30s and final extension at 72°C for 10min. We confirmed amplification of all samples with agarose gel (1%) electrophoresis.

We designed a pyrosequencing assay of previously reported CpGs involved in expression of IGF2 using Pyromark Assay Design (Supplementary Table 2). After binding of the PCR product to streptavidin-coated Sepharose beads (GE Healthcare), we washed the template, made it single-stranded and neutralized the sample. When the negative control showed a signal during test phase of the primers, we developed new primers in the same region. We performed pyrosequencing using the PyroGold SQA reagent kit (Qiagen) according to manufacturer’s protocol and performed analyses on the Pyromark Q24 system. To test reproducibility, we treated five times the same sample of high and low methylated DNA (Epigendx, Hopkinton, MA, USA) with bisulphite and analysed the assays.

Statistical analysis

We performed statistical analyses by SPSS version 21 and GraphPad Prism 3.0. In all analyses, P<0.05 was considered significant. Values represent mean±s.d., unless otherwise specified. One-way ANOVA, followed by Tukey’s multiple comparison test, was used to test for significant differences in methylation and expression after treatment with AZA. To assess significant differences in methylation per region and IGF2 mRNA expression in ACCs, ACAs and normal adrenals, we used the nonparametric Mann–Whitney U test. For all correlation analyses within the carcinoma group, we used Spearman’s correlation.

We quantified the mean methylation in the different regions and IGF2 expression in the tumour by creating a score for all tumours, defined as the absolute standard deviation score (|SDS|) per region compared with methylation or expression in normal adrenals. We used receiver operating characteristic (ROC) analysis to determine regions with an individually significant predictive value for the diagnosis. We included the three most discriminating CpGs in these regions in final analysis. To assess the explained variation of IGF2 expression by these factors, we used multiple linear regression. We determined diagnostic accuracy of the |SDS| by performing ROC analysis with mean |SDS| per region for predicting the diagnosis by the pathologist. We performed the same analysis when discriminating only metastatic ACCs versus all other tumours. We determined the performance of the predictive test by carrying out 10-fold cross-validation. The data were randomly partitioned into 10 subsamples, where after we performed ROC analysis 10 times, in which each subsample was used once as the validation set. An independent cohort was used to validate test performance and the proposed cutoff values. Interobserver variability of the Weiss score was quantified by using an intraclass correlation, while the interrater variability of the binary parameter of pathological diagnosis was quantified by the kappa coefficient.

Results

Methylation status of the IGF2 regulatory regions in adrenocortical tissue

For analysis, we included 24 ACCs, 14 ACAs and 11 normal adrenals. Methylation patterns are shown in Supplementary Fig. 2. A median methylation of 81% was found in DMR0 in ACCs, 77% in ACAs and 82% in normal adrenals (Fig. 2A). In DMR2, a median methylation of 24% was found in ACCs, 33% in ACAs and 33% in normal adrenals (Fig. 2B). In both regions, there were no statistically significant differences in median methylation percentages between ACCs and ACAs.

CTCF3 showed a median methylation of 78% in ACCs, 48% in ACAs and 51% in normal adrenals (Fig. 2C). Methylation in ACCs was significantly higher compared with methylation in ACAs (P=0.018). CTCF6 demonstrated a median methylation of 57% in ACCs, 40% in ACAs and 47% in normal tissue (Fig. 2D), where methylation in ACCs was significantly higher compared with ACAs (P=0.034). As expected, the H19 promoter showed a higher methylation in ACCs as well, namely 69% compared with 45% in ACAs and 45% in normal adrenals (Fig. 2E). The methylation difference between ACCs and ACAs was also statistically significant (P<0.001). For none of the regions there were significant differences in methylation between ACAs and normal adrenals.

Figure 2
Figure 2

Methylation in the specific regions per group. Every dot represents the mean methylation for every region in a tumour of an individual patient. Lines represent medians. Mean methylations in (A) DMR0, (B) DMR2, (C) CTCF3, (D) CTCF6, and (E) H19 promoter. *P<0.05, ***P<0.001. ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; NAT, normal adrenal tissue.

Citation: Endocrine-Related Cancer 23, 9; 10.1530/ERC-16-0266

Correlations between DNA methylation in different IGF2 regulatory regions

The methylation status of CTCF3 and CTCF6, both located in the ICR, were strongly positively correlated in ACCs (ρ=0.614, P<0.001). Methylation in the H19 promoter was correlated with methylation in CTCF3, CTCF6 and the DMR0 (ρ=0.781, P<0.001; ρ=0.741, P<0.001; ρ=0.413, P=0.010, respectively). Methylation in DMR0 was also correlated with methylation in CTCF3 and CTCF6 (ρ=0.595, P<0.001; ρ=0.361, P=0.018, respectively), although strongly driven by two low methylated carcinomas. All correlation coefficients and P values are listed in Supplementary Table 3.

Expression of IGF2 mRNA in adrenocortical tissue and correlations with methylation status of different IGF2 regulatory regions

IGF2 expression was significantly higher in ACCs compared with ACAs (P=0.001, Fig. 3A). No statistically significant difference in IGF2 expression was found between ACAs and normal adrenals.

Figure 3
Figure 3

IGF2 mRNA expression in adrenal tissues and correlation of IGF2 mRNA expression in ACCs with methylation in IGF2 regulatory regions. Expression data are presented as scatter dot plot. Expressions are normalized to HPRT1.ρrepresents Spearman’s correlation coefficient. (A) Relative IGF2 mRNA expression in different groups on a log transformed scale. Lines represent medians. ***P<0.001. (B) Correlation of IGF2 mRNA expression in ACCs with DMR0 promoter methylation, (C) CTCF3 methylation, (D) CTCF6 methylation, and (E) H19 methylation. ACA, adrenocortical adenomas; ACC, adrenocortical carcinomas; NAT, normal adrenal tissue.

Citation: Endocrine-Related Cancer 23, 9; 10.1530/ERC-16-0266

In ACCs, the mean DMR0, CTCF3, CTCF6 and H19 promoter methylation were positively correlated with IGF2 mRNA expression (ρ=0.485, P=0.016, Fig. 3B; ρ=0.625, P=0.001, Fig. 3C; ρ=0.417, P=0.042, Fig. 3D; ρ=0.521, P=0.009, Fig. 3E, respectively).

Quantification of differences in methylation between carcinomas and adenomas

We calculated SDS per region for the different neoplasms (Supplementary Table 4). IGF2 expression, DMR2, CTCF3 and the H19 promoter individually showed a significant predictive value for the diagnosis of ACC (Fig. 4A). When using the mean |SDS| in the second until fourth CpG in DMR2, the fifth until seventh CpG in CTCF3, and the H19 promoter in a ROC curve for predicting the diagnosis an area under the curve (AUC) of 0.997 was found (Fig. 4B). Test characteristics included sensitivity of 96% and specificity of 100% for a cutoff of |SDS| 2.617, with positive predictive value of 100% and negative predictive value of 93%. These three factors together explained 45.1% (R2) of the variation in IGF2 expression in ACCs with df 3, F-statistic 5.470 and P=0.007.

Figure 4
Figure 4

Receiver operating characteristic (ROC) curve of mean absolute standard deviation score (|SDS|) compared with methylation in normal adrenals displayed as sensitivity (y-axis) and 1 – specificity (x-axis) for the diagnosis of adrenocortical carcinoma (ACC) or adrenocortical adenoma (ACA) according to the pathologist. The straight diagonal lines represent reference lines. (A) ROC curve of mean |SDS| compared with methylation in normal adrenals for methylation in all regions and IGF2 mRNA expression separately. DMR2 (P=0.001), CTCF3 (P<0.001), the H19 promoter (P<0.001), and IGF2 expression (P<0.001) showed individually significant predictive value for diagnosis of ACC or ACA. (B) ROC curve of the H19 promoter, second until fourth CpG in the DMR2 and fifth until seventh CpG in CTCF3 for the diagnosis of ACC or ACA according to the pathologist. The area under the curve (AUC) was 0.997±0.005, P<0.001. Cutoff value: mean |SDS| 2.617, sensitivity 96%, specificity 100%.

Citation: Endocrine-Related Cancer 23, 9; 10.1530/ERC-16-0266

Three tumours were difficult to classify based on (only) the Weiss score. Two tumours with a diameter of 9.5 and 6.5cm had both a Weiss score of 1. However, the pathologist considered the first tumour as borderline malignant, because it showed minor criteria in the Lin–Weiss–Bisceglia (LWB) system, a modified Weiss system for diagnosis of oncocytic neoplasms (de Krijger & Papathomas 2012). The second case was also difficult to classify, because some criteria, for example, capsular invasion and nuclear polymorphism, were only present in parts of the tumour. Necrosis and a high mitotic rate were not present. When these two tumours with uncertain pathological diagnosis, which were therefore not included in the ACA or ACC group, were tested with the methylation scorings system, they scored |SDS| of 4.29 and 4.20, respectively. This, in turn, would lead in our test to the classification of an ACC. During follow-up (27 and 8months, respectively), no recurrences or metastases were detected.

A third patient underwent surgery because of a cortisol producing adrenal tumour with the imaging phenotype of an adenoma. Pathological examination showed a Weiss score of 3 with a preferred diagnosis of ACC. However, not all negative prognostic factors (e.g. necrosis, capsular invasion, atypical mitosis) were present in this tumour. Although the diagnosis ACC was not certain, the patient was treated with mitotane as adjuvant therapy. After 20months, lymph node and subsequently bone metastases developed, confirming the diagnosis of ACC. The methylation score of this tumour showed a mean |SDS| of 4.68, which indicates malignancy based on our proposed test.

The mean |SDS| of the three regions did not significantly differ between ACCs with and without metastasis at the end of follow-up. When we segregated cases with adverse clinical outcome, that is, metastatic disease, from all other cases, an AUC of 0.900±0.049 (P<0.0001) was found.

A 10-fold cross-validation resulted in a good prediction of the diagnosis in 95% of the cases, with a mean cutoff of |SDS| 2.555±0.22. Mean sensitivity was 96±2.0%, specificity was 100±0.0% and AUC was 0.997±0.002. Mean coefficient of variation (CV) of regions included in final analysis in high methylated samples was 0.039 and in low methylated samples was 0.380. CVs for all regions are described in Supplementary Table 5.

Validation of the proposed diagnostic test in an independent cohort

After exclusion of four samples due to insufficient DNA yield, 9 ACCs and 13 ACAs were included in the validation cohort, based on the diagnosis determined by the pathologist at the VUMC. There was high agreement between the two pathologists (VUMC and EMC) regarding the Weiss score (r=0.914), with one discordant final pathological diagnosis (κ=0.904). The mean |SDS| was assessed in the three regions proposed on the basis of the EMC cohort (Fig. 5A and B). ROC analysis, using the diagnosis as determined by the pathologist from the VUMC, showed an AUC of 0.957±0.039 with a sensitivity of 89%, specificity of 92% and a cutoff value of |SDS| 2.331. When we applied the cutoff value obtained from the EMC cohort (2.617) to the validation cohort, a sensitivity of 75% and a specificity of 92% were found. This sensitivity increased to 88% when the pathological diagnosis obtained at the EMC was used. The mean |SDS| of the tumour with discordant pathological diagnosis by the two pathologists was 2.49, and follow-up was only 2months. Combination of the two cohorts (ACA n=27, ACC n=33) showed an AUC of 0.983±0.013 with a sensitivity of 94% and a specificity of 96% using a cutoff of mean |SDS| 2.442 (Fig. 5C and D).

Figure 5
Figure 5

Mean absolute standard deviation scores (|SDS|) of the H19 promoter, second until fourth CpG in the DMR2, and fifth until seventh CpG in CTCF3 in the two different cohorts for adrenocortical adenomas (ACA) and carcinomas (ACC), as determined by the local pathologist on the basis of the Weiss score. Mean |SDS| in the EMC cohort (A), the VUMC cohort (B), and for the total cohort (C). Horizontal dotted lines represent the optimal cutoff value per cohort, as determined by receiver operating characteristic (ROC) curve. (D) ROC of the mean |SDS| for the diagnosis of ACC or ACA according to the pathologist in the total cohort, combining the specimens from the EMC and VUMC. The area under the curve (AUC) was 0.983±0.013 with a sensitivity of 94% and a specificity of 96% using a cutoff of mean |SDS| 2.442 for diagnosing malignancy. The straight diagonal line represents reference line. EMC, Erasmus University Medical Center; VUMC, VU University Medical Center.

Citation: Endocrine-Related Cancer 23, 9; 10.1530/ERC-16-0266

Effect of AZA treatment on DNA methylation and expression of IGF2 and H19 mRNA in human ACC cell lines

In order to determine the role of DNA methylation on IGF2 and H19 expression, we evaluated the effect of the demethylating drug AZA in three human ACC cell lines (H295R, HAC15 and SW13). After 7days of 1µM AZA treatment, the mean methylation in all regulatory regions in H295R decreased with 51.9±2.4%, P<0.0001 (mean±s.e.), in a both time-dependent and concentration-dependent manner (Supplementary Fig. 3A, B, C, D, E and H, I, J, K, L). Concomitant with the decrease in methylation, we observed a strong and statistically significant decrease in IGF2 mRNA expression after 7days of treatment with AZA in H295R (−79.4±3.8%, P<0.01, mean±s.e.m.), again in a both time-dependent and concentration-dependent manner (Supplementary Fig. 3F and M). The decrease in IGF2 mRNA expression was accompanied by strongly increased H19 mRNA expression (Supplementary Fig. G and N). We found similar results for methylation percentages and expression data in HAC15 cells (data not shown). SW13 methylation patterns were partly different compared with H295R and HAC15 and SW13 cells did not express H19 or IGF2 mRNA (data not shown).

Discussion

In the current study, we found that a combined DNA methylation score of three different IGF2 regulatory regions discriminates ACCs and ACAs with a sensitivity of 96% and specificity of 100%. For tumours of the adrenal cortex, the chance on malignant behaviour increases with tumour size. The differentiation between ACC and ACA on histopathological grounds can be challenging, by both a poor reproducibility and by tumours classified as being of uncertain malignant behaviour (Pohlink et al. 2004, Tan et al. 2005, Klibanski et al. 2006, Giordano 2010). Mitotane treatment is given to prevent disease recurrence in patients with malignant adrenal tumours but is associated with serious toxicity (Lacroix 2010, Fassnacht et al. 2013). A reliable biomarker that indicates malignant behaviour, particularly in ‘borderline malignant’ tumours, could be helpful for the decision on postoperative strategies, like the choice on adjuvant mitotane and the duration and intensity of follow-up.

If we would have used this test for the patient with a Weiss score of 3 who developed metastasis after 20months, the methylation score could have been supportive for the decision of adjuvant mitotane treatment. For the two cases originally classified as tumours with ‘uncertain behaviour’, but are identified as ACC according to our proposed test, close monitoring is currently performed. Further follow-up, which is now 27 and 8months from diagnosis, will have to elucidate whether these tumours will recur or metastasize and whether the methylation score indeed offers benefit compared with the Weiss score in these cases. At present, no recurrences or metastases developed in both patients.

In this study, we confirmed that IGF2 is an important marker for ACC, but as also previously described in a number of studies, does not fully discriminate ACCs from ACAs (Creemers et al. 2016). Mean sensitivity and specificity for diagnosing malignancy are both around 80%, with substantial variances between studies. H19 seems to be an important regulator, given the positive correlation of the H19 promoter methylation with IGF2 expression in our study, as well as in previous research (Gao et al. 2002). By pyrosequencing of different regions in a tumour of the same patient, we were able to define the most discriminating regions in this dataset and propose a diagnostic test with very high accuracy. Choosing a smaller subset of CpGs provided even better test characteristics.

We used three different adrenocortical cancer cell lines to demonstrate the mechanism of regulation of IGF2. By treating adrenocortical cancer cells with the demethylating agent AZA, we were able to strongly decrease the IGF2 mRNA expression and increase the H19 mRNA expression. Herewith, we confirmed DNA methylation as a strong regulatory element in IGF2 and H19 expression. Although it is still unclear whether IGF2 is mainly a contributor or an effect of malignancy (Drelon et al. 2012, Guillaud-Bataille et al. 2014), further research should be done in order to find new insights in therapeutic possibilities of demethylating drugs, considering the reversibility of IGF2 mRNA expression.

Pyrosequencing is a technology yielding accurate and highly reproducible knowledge about methylation percentages at a CpG (Tost & Gut 2007, Quillien et al. 2012). By analysing normal adrenals and neoplasms, we could quantify to which extent methylation in tumours deviated from normal. The methylation patterns in DMR2 and CTCF3 visually revealed three phenotypes in ACCs: hypomethylated, moderately methylated and hypermethylated (Fig. 2F and I, respectively). The concept of IGF2 regulation is based on instability of the IGF2/H19 locus. By transforming methylation in absolute SDS, we identified both hypomethylated and hypermethylated phenotypes. Additionally, methylation of DMRs can have various effects on gene expression, both silencing and activation (Moore et al. 1997). Methylation of CpGs in promoter regions generally cause repression of the gene (Bell & Beck 2010).

It is important to take into account heterogeneity of adrenocortical tumours, since this can provide different percentages of methylation in different parts of the tumour.

A notable consideration in this study is the reference diagnosis, considering the only definite diagnosis of ACC after metastasis. We have set the diagnosis of the tumours as conclusive as possible, taking into account histopathological features and presence of metastasis (71% of all ACCs). The AUC of 0.900±0.049 for discriminating metastatic ACCs vs all other tumours requires the consideration that there are ACCs reported with no adverse outcome and long survival (>10years), in literature as well as in our cohort.

The aberrant methylation patterns in ACCs can be explained either by loss of heterozygosity or by de novo (de)methylation. Probably in most ACCs, aberrant methylation patterns are caused by deletion of the maternal allele, a structural abnormality mostly seen in malignant adrenal tumours (Gicquel et al. 1997). However, this cannot be the case in all ACCs based on the methylation patterns. It is important to notice that the underlying cause of these methylation patterns will not influence the test characteristics.

This diagnostic test, based on analysis in only DNA, has an advantage for both reproducibility and time consumption. Besides, it makes the test more practical and useful, since RNA is more instable and fragile. We do need to take into account that our analyses are performed on frozen specimens, which is at the moment not readily available in all centres. Future research could focus on the possibility of assessing methylation percentages on DNA obtained from formalin-fixed paraffin-embedded specimens, improving the diagnostic applicability of our proposed test.

Our findings were validated in an independent cohort, which largely increases the generalizability of the proposed diagnostic test. However, larger studies, requiring multicentre collaborations, due to the rare incidence of ACCs, will be necessary to fully validate the predictive value of these specific epigenetic changes and to also validate the required cutoff values for malignancy. These results should be compared with urine steroid metabolomics, a recently introduced new potential preoperative marker for malignancy in adrenal tumours (Arlt et al. 2011).

In conclusion, we show that specific methylation patterns of IGF2 regulatory regions are a promising tool to assess malignancy in adrenal tumours and could lead to the first objective diagnostic tool for identifying ACCs.

Supplementary data

This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-16-0266.

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

This research was financially supported by Erasmus MC Vriendenfonds and by a donation from Mr and Mrs Westra.

Acknowledgements

The authors thank Dr V L Wester for his advice regarding data analysis.

References

  • Almeida MQ, Fragoso MC, Lotfi CF, Santos MG, Nishi MY, Costa MH, Lerario AM, Maciel CC, Mattos GE & Jorge AA 2008 Expression of insulin-like growth factor-II and its receptor in pediatric and adult adrenocortical tumors. Journal of Clinical Endocrinology and Metabolism 93 35243531. (doi:10.1210/jc.2008-0065)

    • Search Google Scholar
    • Export Citation
  • Arlt W, Biehl M, Taylor AE, Hahner S, Libe R, Hughes BA, Schneider P, Smith DJ, Stiekema H & Krone N 2011 Urine steroid metabolomics as a biomarker tool for detecting malignancy in adrenal tumors. Journal of Clinical Endocrinology and Metabolism 96 37753784. (doi:10.1210/jc.2011-1565)

    • Search Google Scholar
    • Export Citation
  • Baba Y, Nosho K, Shima K, Huttenhower C, Tanaka N, Hazra A, Giovannucci EL, Fuchs CS & Ogino S 2010 Hypomethylation of the IGF2 DMR in colorectal tumors, detected by bisulfite pyrosequencing, is associated with poor prognosis. Gastroenterology 139 18551864. (doi:10.1053/j.gastro.2010.07.050)

    • Search Google Scholar
    • Export Citation
  • Barlaskar FM 2011 Identifying, Modeling and Targeting IGF-II Overexpression in Adrenocortical Carcinoma, pp 3650. Ann Arbor, MI, USA: The University of Michigan.

    • Search Google Scholar
    • Export Citation
  • Bell CG & Beck S 2010 The epigenomic interface between genome and environment in common complex diseases. Briefings in Functional Genomics 9 477485. (doi:10.1093/bfgp/elq026)

    • Search Google Scholar
    • Export Citation
  • Boissonnas CC, Abdalaoui HE, Haelewyn V, Fauque P, Dupont JM, Gut I, Vaiman D, Jouannet P, Tost J & Jammes H 2010 Specific epigenetic alterations of IGF2-H19 locus in spermatozoa from infertile men. European Journal of Human Genetics 18 7380. (doi:10.1038/ejhg.2009.117)

    • Search Google Scholar
    • Export Citation
  • Creemers SG, Hofland LJ, Korpershoek E, Franssen GJ, van Kemenade FJ, de Herder WW & Feelders RA 2016 Future directions in the diagnosis and medical treatment of adrenocortical carcinoma. Endocrine-Related Cancer 23 R43R69. (doi:10.1530/erc-15-0452)

    • Search Google Scholar
    • Export Citation
  • de Fraipont F, El Atifi M, Cherradi N, Le Moigne G, Defaye G, Houlgatte R, Bertherat J, Bertagna X, Plouin PF & Baudin E 2005 Gene expression profiling of human adrenocortical tumors using complementary deoxyribonucleic acid microarrays identifies several candidate genes as markers of malignancy. Journal of Clinical Endocrinology and Metabolism 90 18191829. (doi:10.1210/jc.2004-1075)

    • Search Google Scholar
    • Export Citation
  • de Krijger RR & Papathomas TG 2012 Adrenocortical neoplasia: evolving concepts in tumorigenesis with an emphasis on adrenal cortical carcinoma variants. Virchows Archiv 460 918. (doi:10.1007/s00428-011-1166-y)

    • Search Google Scholar
    • Export Citation
  • Dehner LP & Hill DA 2009 Adrenal cortical neoplasms in children: why so many carcinomas and yet so many survivors? Pediatric and Developmental Pathology 12 284291. (doi:10.2350/08-06-0489.1)

    • Search Google Scholar
    • Export Citation
  • Dejeux E, Olaso R, Dousset B, Audebourg A, Gut IG, Terris B & Tost J 2009 Hypermethylation of the IGF2 differentially methylated region 2 is a specific event in insulinomas leading to loss-of-imprinting and overexpression. Endocrine-Related Cancer 16 939952. (doi:10.1677/ERC-08-0331)

    • Search Google Scholar
    • Export Citation
  • Drelon C, Berthon A, Ragazzon B, Tissier F, Bandiera R, Sahut-Barnola I, de Joussineau C, Batisse-Lignier M, Lefrancois-Martinez AM & Bertherat J 2012 Analysis of the role of Igf2 in adrenal tumour development in transgenic mouse models. PLoS ONE 7 e44171. (doi:10.1371/journal.pone.0044171)

    • Search Google Scholar
    • Export Citation
  • Erickson LA, Jin L, Sebo TJ, Lohse C, Pankratz VS, Kendrick ML, van Heerden JA, Thompson GB, Grant CS & Lloyd RV 2001 Pathologic features and expression of insulin-like growth factor-2 in adrenocortical neoplasms. Endocrine Pathology 12 429435. (doi:10.1385/ep:12:4:429)

    • Search Google Scholar
    • Export Citation
  • Fassnacht M, Johanssen S, Fenske W, Weismann D, Agha A, Beuschlein F, Fuhrer D, Jurowich C, Quinkler M & Petersenn S 2010 Improved survival in patients with stage II adrenocortical carcinoma followed up prospectively by specialized centers. Journal of Clinical Endocrinology and Metabolism 95 49254932. (doi:10.1210/jc.2010-0803)

    • Search Google Scholar
    • Export Citation
  • Fassnacht M, Kroiss M & Allolio B 2013 Update in adrenocortical carcinoma. Journal of Clinical Endocrinology and Metabolism 98 45514564. (doi:10.1210/jc.2013-3020)

    • Search Google Scholar
    • Export Citation
  • Gao ZH, 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 11701176. (doi:10.1210/jcem.87.3.8331)

    • Search Google Scholar
    • Export Citation
  • 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 25592565. (doi:10.1210/jc.82.8.2559)

    • Search Google Scholar
    • Export Citation
  • Giordano TJ 2010 Adrenocortical tumors: an integrated clinical, pathologic, and molecular approach at the University of Michigan. Archives of Pathology & Laboratory Medicine 134 14401443. (doi:10.1043/2010-0320-RA.1)

    • Search Google Scholar
    • Export Citation
  • Giordano TJ, Thomas DG, Kuick R, Lizyness M, Misek DE, Smith AL, Sanders D, Aljundi RT, Gauger PG & Thompson NW 2003 Distinct transcriptional profiles of adrenocortical tumors uncovered by DNA microarray analysis. American Journal of Pathology 162 521531. (doi:10.1016/S0002-9440(10)63846-1)

    • Search Google Scholar
    • Export Citation
  • Golden SH, Robinson KA, Saldanha I, Anton B & Ladenson PW 2009 Clinical review: prevalence and incidence of endocrine and metabolic disorders in the United States: a comprehensive review. Journal of Clinical Endocrinology and Metabolism 94 18531878. (doi:10.1210/jc.2008-2291)

    • Search Google Scholar
    • Export Citation
  • Guillaud-Bataille M, Ragazzon B, de Reynies A, Chevalier C, Francillard I, Barreau O, Steunou V, Guillemot J, Tissier F & Rizk-Rabin M 2014 IGF2 promotes growth of adrenocortical carcinoma cells, but its overexpression does not modify phenotypic and molecular features of adrenocortical carcinoma. PLoS ONE 9 e103744. (doi:10.1371/journal.pone.0103744)

    • Search Google Scholar
    • Export Citation
  • Hao Y, Crenshaw T, Moulton T, Newcomb E & Tycko B 1993 Tumour-suppressor activity of H19 RNA. Nature 365 764767. (doi:10.1038/365764a0)

  • Herold M, Bartkuhn M & Renkawitz R 2012 CTCF: insights into insulator function during development. Development 139 10451057. (doi:10.1242/dev.065268)

    • Search Google Scholar
    • Export Citation
  • Kebebew E, Reiff E, Duh QY, Clark OH & McMillan A 2006 Extent of disease at presentation and outcome for adrenocortical carcinoma: have we made progress? World Journal of Surgery 30 872878. (doi:10.1007/s00268-005-0329-x)

    • Search Google Scholar
    • Export Citation
  • Kerkhofs TM, Verhoeven RH, Van der Zwan JM, Dieleman J, Kerstens MN, Links TP, Van de Poll-Franse LV & Haak HR 2013 Adrenocortical carcinoma: a population-based study on incidence and survival in the Netherlands since 1993. European Journal of Cancer 49 25792586. (doi:10.1016/j.ejca.2013.02.034)

    • Search Google Scholar
    • Export Citation
  • Klibanski A, Stephen AE, Greene MF, Blake MA & Wu CL 2006 Case records of the Massachusetts General Hospital. Case 36-2006. A 35-year-old pregnant woman with new hypertension. New England Journal of Medicine 355 22372245. (doi:10.1056/NEJMcpbib69027)

    • Search Google Scholar
    • Export Citation
  • Lacroix A 2010 Approach to the patient with adrenocortical carcinoma. Journal of Clinical Endocrinology and Metabolism 95 48124822. (doi:10.1210/jc.2010-0990)

    • Search Google Scholar
    • Export Citation
  • Lau SK & Weiss LM 2009 The Weiss system for evaluating adrenocortical neoplasms: 25 years later. Human Pathology 40 757768. (doi:10.1016/j.humpath.2009.03.010)

    • Search Google Scholar
    • Export Citation
  • Lughezzani G, Sun M, Perrotte P, Jeldres C, Alasker A, Isbarn H, Budaus L, Shariat SF, Guazzoni G & Montorsi F 2010 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. European Journal of Cancer 46 713719. (doi:10.1016/j.ejca.2009.12.007)

    • Search Google Scholar
    • Export Citation
  • Moore T, Constancia M, Zubair M, Bailleul B, Feil R, Sasaki H & Reik W 1997 Multiple imprinted sense and antisense transcripts, differential methylation and tandem repeats in a putative imprinting control region upstream of mouse Igf2. PNAS 94 1250912514. (doi:10.1073/pnas.94.23.12509)

    • Search Google Scholar
    • Export Citation
  • Ohlsson R, Hedborg F, Holmgren L, Walsh C & Ekstrom TJ 1994 Overlapping patterns of IGF2 and H19 expression during human development: biallelic IGF2 expression correlates with a lack of H19 expression. Development 120 361368.

    • Search Google Scholar
    • Export Citation
  • Papotti M, Duregon E, Volante M & McNicol AM 2014 Pathology of the adrenal cortex: a reappraisal of the past 25 years focusing on adrenal cortical tumors. Endocrine Pathology 25 3548. (doi:10.1007/s12022-013-9291-6)

    • Search Google Scholar
    • Export Citation
  • Pohlink C, Tannapfel A, Eichfeld U, Schmidt F, Fuhrer D, Paschke R & Koch CA 2004 Does tumor heterogeneity limit the use of the Weiss criteria in the evaluation of adrenocortical tumors? Journal of Endocrinological Investigation 27 565569. (doi:10.1007/BF03347480)

    • Search Google Scholar
    • Export Citation
  • Quillien V, Lavenu A, Karayan-Tapon L, Carpentier C, Labussiere M, Lesimple T, Chinot O, Wager M, Honnorat J & Saikali S 2012 Comparative assessment of 5 methods (methylation-specific polymerase chain reaction, methylight, pyrosequencing, methylation-sensitive high-resolution melting, and immunohistochemistry) to analyze O6-methylguanine-DNA-methyltranferase in a series of 100 glioblastoma patients. Cancer 118 42014211. (doi:10.1002/cncr.27392)

    • Search Google Scholar
    • Export Citation
  • Rachmilewitz J, Goshen R, Ariel I, Schneider T, de Groot N & Hochberg A 1992 Parental imprinting of the human H19 gene. FEBS Letters 309 2528. (doi:10.1016/0014-5793(92)80731-U)

    • Search Google Scholar
    • Export Citation
  • Schmitt A, Saremaslani P, Schmid S, Rousson V, Montani M, Schmid DM, Heitz PU, Komminoth P & Perren A 2006 IGFII and MIB1 immunohistochemistry is helpful for the differentiation of benign from malignant adrenocortical tumours. Histopathology 49 298307. (doi:10.1111/j.1365-2559.2006.02505.x)

    • Search Google Scholar
    • Export Citation
  • Tabano S, Colapietro P, Cetin I, Grati FR, Zanutto S, Mando C, Antonazzo P, Pileri P, Rossella F & Larizza L 2010 Epigenetic modulation of the IGF2/H19 imprinted domain in human embryonic and extra-embryonic compartments and its possible role in fetal growth restriction. Epigenetics 5 313324. (doi:10.4161/epi.5.4.11637)

    • Search Google Scholar
    • Export Citation
  • Tan HS, Thai AC, Nga ME & Mukherjee JJ 2005 Development of ipsilateral adrenocortical carcinoma sixteen years after resection of an adrenal tumour causing Cushing’s syndrome. Annals of the Academy of Medicine 34 271274.

    • Search Google Scholar
    • Export Citation
  • Tost J & Gut IG 2007 DNA methylation analysis by pyrosequencing. Nature Protocols 2 22652275. (doi:10.1038/nprot.2007.314)

  • van Koetsveld PM, Vitale G, de Herder WW, Feelders RA, van der Wansem K, Waaijers M, van Eijck CHJ, Speel EJM, Croze E & van der Lely AJ 2006 Potent inhibitory effects of type I interferons on human adrenocortical carcinoma cell growth. Journal of Clinical Endocrinology and Metabolism 91 45374543. (doi:10.1210/jc.2006-0620)

    • Search Google Scholar
    • Export Citation
  • Wang C, Sun Y, Wu H, Zhao D & Chen J 2014 Distinguishing adrenal cortical carcinomas and adenomas: a study of clinicopathological features and biomarkers. Histopathology 64 567576. (doi:10.1111/his.12283)

    • Search Google Scholar
    • Export Citation
  • Young WF Jr 2007 Clinical practice. The incidentally discovered adrenal mass. New England Journal of Medicine 356 601610. (doi:10.1056/NEJMcp065470)

    • Search Google Scholar
    • Export Citation
  • Zhang Y & Tycko B 1992 Monoallelic expression of the human H19 gene. Nature Genetics 1 4044. (doi:10.1038/ng0492-40)

If the inline PDF is not rendering correctly, you can download the PDF file here.

 

      Society for Endocrinology

Sept 2018 onwards Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 748 102 5
PDF Downloads 279 80 4
  • View in gallery

    Overview of the imprinted region of the IGF2 gene and H19 gene on the maternal and paternal allele in the normal situation. Lines with dots represent known sites of DNA methylation. (arrow) expression of the gene. (stub-end arrow) repression of the gene. Methylation status of the DMR0 remains unclear. Figure is not at scale. DMR0, differentially methylated region 0; DMR2, differentially methylated region 2; ICR, imprinting control region, consisting of CTCF3 and CTCF6. Data from Tabano et al. (2010).

  • View in gallery

    Methylation in the specific regions per group. Every dot represents the mean methylation for every region in a tumour of an individual patient. Lines represent medians. Mean methylations in (A) DMR0, (B) DMR2, (C) CTCF3, (D) CTCF6, and (E) H19 promoter. *P<0.05, ***P<0.001. ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; NAT, normal adrenal tissue.

  • View in gallery

    IGF2 mRNA expression in adrenal tissues and correlation of IGF2 mRNA expression in ACCs with methylation in IGF2 regulatory regions. Expression data are presented as scatter dot plot. Expressions are normalized to HPRT1.ρrepresents Spearman’s correlation coefficient. (A) Relative IGF2 mRNA expression in different groups on a log transformed scale. Lines represent medians. ***P<0.001. (B) Correlation of IGF2 mRNA expression in ACCs with DMR0 promoter methylation, (C) CTCF3 methylation, (D) CTCF6 methylation, and (E) H19 methylation. ACA, adrenocortical adenomas; ACC, adrenocortical carcinomas; NAT, normal adrenal tissue.

  • View in gallery

    Receiver operating characteristic (ROC) curve of mean absolute standard deviation score (|SDS|) compared with methylation in normal adrenals displayed as sensitivity (y-axis) and 1 – specificity (x-axis) for the diagnosis of adrenocortical carcinoma (ACC) or adrenocortical adenoma (ACA) according to the pathologist. The straight diagonal lines represent reference lines. (A) ROC curve of mean |SDS| compared with methylation in normal adrenals for methylation in all regions and IGF2 mRNA expression separately. DMR2 (P=0.001), CTCF3 (P<0.001), the H19 promoter (P<0.001), and IGF2 expression (P<0.001) showed individually significant predictive value for diagnosis of ACC or ACA. (B) ROC curve of the H19 promoter, second until fourth CpG in the DMR2 and fifth until seventh CpG in CTCF3 for the diagnosis of ACC or ACA according to the pathologist. The area under the curve (AUC) was 0.997±0.005, P<0.001. Cutoff value: mean |SDS| 2.617, sensitivity 96%, specificity 100%.

  • View in gallery

    Mean absolute standard deviation scores (|SDS|) of the H19 promoter, second until fourth CpG in the DMR2, and fifth until seventh CpG in CTCF3 in the two different cohorts for adrenocortical adenomas (ACA) and carcinomas (ACC), as determined by the local pathologist on the basis of the Weiss score. Mean |SDS| in the EMC cohort (A), the VUMC cohort (B), and for the total cohort (C). Horizontal dotted lines represent the optimal cutoff value per cohort, as determined by receiver operating characteristic (ROC) curve. (D) ROC of the mean |SDS| for the diagnosis of ACC or ACA according to the pathologist in the total cohort, combining the specimens from the EMC and VUMC. The area under the curve (AUC) was 0.983±0.013 with a sensitivity of 94% and a specificity of 96% using a cutoff of mean |SDS| 2.442 for diagnosing malignancy. The straight diagonal line represents reference line. EMC, Erasmus University Medical Center; VUMC, VU University Medical Center.