Jump to
Abstract
The purpose of this study was to evaluate the DNA methylation levels of steroidogenic enzyme genes in aldosterone-producing adenoma (APA) and the effects of gene mutations in APA on the DNA methylation levels. DNA methylation array analysis was conducted using nonfunctioning adrenocortical adenoma (n=12) and APA (n=35) samples, including some with a KCNJ5 mutation (n=21), an ATP1A1 mutation (n=5), and without the known mutations (n=9). The quantitative polymerase chain reaction assay was performed for the detection of CYP11B2 and CYP11B1 expression levels in nonfunctioning adrenocortical adenoma and APA. We introduced the KCNJ5 T158A mutation using lentivirus delivery in the human adrenocortical 15 cell line, and analyzed the effects of the mutation on DNA methylation levels. We analyzed the 83 presumed DNA methylation sites of steroidogenic enzymes. In APA, we found 7 hypomethylated sites in CYP11B2 and 1 hypomethylated and 6 hypermethylated sites in CYP11B1. There were no differences in the steroidogenic enzymes gene DNA methylation of peripheral leukocytes between nonfunctioning adrenocortical adenoma and APA. No CYP11B2 methylation level was associated with CYP11B2 transcription levels in APA. All methylation sites, except for a CYP11B2 region, showed no difference among APAs with or without gene mutations. Human adrenocortical 15 cells with the KCNJ5 mutation showed no changes in CYP11B2 or CYP11B1 methylation levels compared with control cells. We demonstrated that CYP11B2 in APA was extensively hypomethylated, and CYP11B2 methylation in the region with hypomethylation was not induced by KCNJ5 or ATP1A1 mutations that cause aldosterone overproduction in APA and a KCNJ5 mutation human adrenocortical 15 cells.
Introduction
Primary aldosteronism (PA) is the most common cause of secondary hypertension and occurs by an excessive and autonomous production of aldosterone independent of the renin–angiotensin system.1 Patients with PA have an increased risk of cardiovascular and cerebrovascular complications compared with those with essential hypertension.2,3 Therefore, the elucidation of the molecular mechanisms of aldosterone production is crucially important for the development of therapeutic agents for PA.
The 2 most common forms of PA are aldosterone-producing adenoma (APA) and idiopathic hyperaldosteronism, also called as bilateral adrenal zona glomerulosa hyperplasia.1 Aldosterone is produced from the precursor cholesterol via a series of steroidogenic enzymes, and APA has higher levels of aldosterone synthase (CYP11B2) that catalyzes the final steps of aldosterone biosynthesis.4,5 Ca2+ signaling is upregulated in APA,6 and it plays a pivotal role in the activation of CYP11B2 transcription.7 Recently, exome sequencing analyses demonstrated that 50% to 80% of APA harbored somatic mutations in KCNJ5, ATP1A1, ATP2B3, CACNA1D, or CTNNB1.8–12 KCNJ5, ATP1A1, ATP2B3, and CACNA1D mutations lead to an increased intracellular Ca2+ concentration, an activation of Ca2+ signaling, and an increase in CYP11B2 transcription.8–11 High CYP11B2 expression levels were found in APA with somatic CTNNB1 mutations that activated Wnt signal.12,13
Epigenetic regulation, such as DNA methylation as well as intracellular signaling cascades, is important for tumor progression, cell survival, DNA damage repair, and hormone secretion via their influence of gene transcription.14,15 DNA methylation by DNA methyltransferase enzymes occurs at the cytosine of 5′-cytosine-guanine-3′ dinucleotides (CpG), where the methylation of the promoter region typically acts to repress the gene transcription.14 Previous reports demonstrated that methylation status might be regulated in APA,16–18 and a methylation site at the CYP11B2 promoter might be associated with CYP11B2 expression in adrenal tumors.17 However, there have been no reports on all presumed methylation sites of every adrenal steroidogenic enzymes in APA. The Infinium HumanMethylation450 BeadChip array enabled whole-genome analysis and detection with more accurate results,19 and analysis using the method with a large sample size in APA has not yet been reported.
Taken together, we hypothesized that the steroidogenic enzymes in APA were regulated by not only intracellular Ca signaling but also DNA methylation and thus proposed to identify the methylation state of steroidogenic enzyme genes, including the steroidogenic acute regulatory protein (StAR) gene, in APA by DNA methylation array analysis. To examine the association between gene mutation and DNA methylation, we further aimed to investigate the effect of gene mutations known to produce hyperaldosteronism in APA on methylation levels in APA and the human adrenocortical carcinoma cell line (HAC15).
Methods
Patients and Tissue Collection
Forty-seven adrenal tumors including 12 nonfunctioning adrenocortical adenomas (NF) and 35 APA were obtained by surgery and stored at −80°C until used. The pathological diagnosis of APA was confirmed by detecting the expression of CYP11B2 by immunohistochemistry and quantitative polymerase chain reaction assays as previously reported.5,20 The clinical characteristics of the patients are shown in Table S1 in the online-only Data Supplement. The diagnosis and subtype diagnosis of PA were performed using the Japan Endocrine Society guideline,21 as previously described.20,22 NF was diagnosed by radiological findings that consisted of lipid tissue by computed tomography or magnetic resonance imaging and endocrinologic findings that did not show cortisol or aldosterone excess as previously reported.23 Our study was approved by the ethics committee of Hiroshima University, and written informed consent was obtained from all the patients.
DNA Extraction and Genotyping
Genomic DNA from adrenal tissues, peripheral leukocyte DNA, and HAC15 cells were isolated by the DNeasy blood and tissue kit (Qiagen, Hilden, Germany). Polymerase chain reaction–based direct sequencing for KCNJ5, ATP1A1, ATP2B3, and CACNA1D was performed as previously described.20
DNA Methylation Analysis
More than 500 ng of DNA, which underwent bisulfite conversion, was used for methylation analysis using the Infinium HumanMethylation450 BeadChip kit (Illumina, San Diego, CA) according to the manufacturer’s instruction. The BeadChip interrogated >485 000 methylation sites per sample at single-nucleotide resolution, and it covered 99% of the RefSeq genes and 96% of the CpG islands.19 Methylation levels for each CpG residue were presented as β values that were used to estimate the rate of the methylated signal intensity. The average β values were expressed as 0 to 1, representing completely nonmethylated to completely methylated values, respectively.
RNA Extraction and Quantitative Polymerase Chain Reaction Assays
Total RNA extraction and cDNA preparation were performed using the RNeasy Mini kit (Qiagen) and the Takara PrimeScript RT Master Mix (Takara Bio Inc, Shiga, Japan), respectively, as previously described.20 The mRNA expression levels of CYP11B2, CYP11B1, and GAPDH were determined using a Taqman Gene Expression Assay kit (Applied Biosystems, Waltham, MA). Gene expression levels were analyzed as arbitrary units normalized against GAPDH mRNA expression.
Cell Culture and Lentiviral Infection
The HAC15 cell line was provided by WE Rainey (University of Michigan, Ann Arbor, MI) and was cultured as previously reported.11 The plasmid pLX303 and pLX303-KCNJ5 T158A were prepared, and lentivirus production and infection in HAC15 cells were applied as previously reported.11
Statistical Analysis
Quantitative data are presented as the means and SD. Heat maps were depicted by the R software package (University of Auckland), and subsequent analyses were performed using SPSS for Windows (release 24.0; SPSS Inc, Chicago, IL). The comparison between 2 groups NF and APA was analyzed by Student t test or χ2 test. Simple regression analysis was conducted to evaluate the relationships of CYP11B2 or CYP11B1 mRNA expression levels with their methylation values. The relative expression of CYP11B2 and CYP11B1 was log transformed because they did not fit a normal distribution. The statistical significances among genotypes were determined by 1-way analysis of variance. In the case of statistically significant relations, the Bonferroni analysis was applied to assess the relationship between categories.
Results
DNA Methylation Profile of Steroidogenic Enzyme Genes
Eighty-three methylation sites (StAR: 9, CYP11A1: 22, HSD3B1: 3, HSD3B2: 4, CYP17A1: 7, CYP21A2: 18, CYP11B1: 12, and CYP11B2: 8) were analyzed using the HumanMethylation450 BeadChip. The heat map of methylation profiles of steroidogenic enzymes is depicted in Figure 1. The statistical significances between NF and APA were obtained after Bonferroni adjustment, and CYP11B2 and CYP11B1 had significant differences in the methylation rate between the 2 groups. Next, to confirm that the difference was not because of the germline, we compared the DNA methylation from peripheral leukocyte between NF (n=3) and APA (n=9). There were no significant differences of steroidogenic enzymes in peripheral leukocytes (Figure S1).
A heat map was generated from the HumanMethylation450 BeadChip array analysis of nonfunctioning adenoma (NF; n=12) and aldosterone-producing adenoma (APA; n=35). Eighty-three methylation sites (StAR: 9, CYP11A1: 22, HSD3B1: 3, HSD3B2: 4, CYP17A1: 7, CYP21A2: 18, CYP11B1: 12, CYP11B2: 8) were analyzed.
CYP11B2 Methylation Profile in APA
The conceivable methylation sites of CYP11B2 are shown in Figure 2A. Five sites are located in promoter region, and an exon, intron, and the 3′-untranslated region each has 1 methylation site. First, we compared the methylation rate of CYP11B2 between NF and APA, and all sites except for 1 promoter region were significantly hypomethylated in APA (Table 1). Second, the associations of CYP11B2 mRNA expression levels with the CYP11B2 methylation values were analyzed by simple regression analysis. APA had 478.1-fold higher CYP11B2 mRNA expression levels compared with NF (P<0.01; Figure S2A), however, there was no significant associations between CYP11B2 expressions and methylation in APA (Table 2).
DNA Methylation Value of CYP11B2 and CYP11B1 in NF and APA
The Relationship Between CYP11B2 or CYP11B1 Methylation Levels and Their mRNA Expression Levels in Aldosterone-Producing Adenoma
The presumed methylation sites of CYP11B2 (A) and CYP11B1 (B). The location from the transcription start site is indicated in parenthesis.
CYP11B1 Methylation Profile in APA
Figure 2B shows 12 presumed methylation regions in CYP11B1, including 3 in the promoter, 5 in exons, 3 in introns, and 1 in the 3′-untranslated region. The methylation levels of 6 sites in APA were higher than those in NF, whereas 1 site was hypomethylated in APA (Table 1). The CYP11B1 mRNA expression levels in APA had a 0.48-fold change compared with those in NF (P<0.01; Figure S2B). There were significant inverse correlations between methylation levels and CYP11B1 mRNA expression levels in 2 hypermethylated sites of CYP11B1 (Table 2). However, these were not located in the promoter region.
Bioinformatics Analysis of CYP11B2 Promoter Methylation Sites
We detected multiple potential transcription factors that bind to the CYP11B2 promoter methylation sites using The JASPAR database (http://jaspar.binf.ku.dk/; Figure S3). We analyzed the gene expression levels of transcription factors with our microarray data published recently.20 Because some genes were highly expressed in APA (Figure S3), their effects on CYP11B2 transcription might be regulated by methylation status.
The Relationship Between the Genotype of APA and DNA Methylation Levels
We compared DNA methylation values among APA with different gene mutations. A methylation site a of CYP11B2 in APA with the ATP1A1 mutation was significantly higher than those in APA with the KCNJ5 mutation (Table 3). However, the methylation site was different from the methylation sites that showed significant difference between NF and APA (Table 1) and did not have any correlation with CYP11B2 mRNA expression (Table 2). There were no significant differences in CYP11B1 methylation levels among APAs with different gene mutations (Table 3).
The Difference of CYP11B2 and CYP11B1 DNA Methylation Levels Among Somatic Mutations in Aldosterone-Producing Adenoma
Effect of KCNJ5 Mutation on DNA Methylation in HAC15
We introduced the KCNJ5 mutation (T158A) using the lentiviral method into HAC15 cells. Transduction of HAC15 cells with KCNJ5 T158A lentivirus mutation increased aldosterone levels 5.4-fold in media compared with control HAC15 cells (P<0.01, data not shown). There were no significant differences in CYP11B2 and CYP11B1 DNA methylation between control and mutant KCNJ5 cells (Table 4).
The Difference of CYP11B2 and CYP11B1 DNA Methylation Levels Between Control and Mutant KCNJ5 Cells
Discussion
Detected in APA by DNA methylation array analysis, we demonstrated that 7 of 8 methylation sites of CYP11B2 were hypomethylated, and 1 and 6 sites of CYP11B1 were hypomethylated and hypermethylated, respectively. Four methylation sites were located in the promoter region of CYP11B2, although the methylation levels did not associate with the CYP11B2 transcription levels in APA. CYP11B1 had no regulated methylation region in its promoter, whereas some methylation rates in its exons or introns had an inverse relationship with CYP11B1 mRNA levels. All methylation sites, except for a CYP11B2 region, showed no difference among APAs with or without gene mutations, and an in vitro study also indicated that gene mutation did not influence the DNA methylation levels in CYP11B2 and CYP11B1.
We emphasize that most of the presumed CYP11B2 methylation sites were hypomethylated in APA compared with NF. Considering basic studies of DNA methylation,14 the hypomethylation in the promoter region could facilitate upregulation of CYP11B2 expression in APA. In addition, as our bioinformatics analysis showed, many potential transcription factors would be predicted to bind the sequence of the CYP11B2 promoter methylation sites. As we indicated that the methylation rate had no correlation with CYP11B2 expression levels in APA, it is suggested that total methylation status, not individual sites, might be important for CYP11B2 mRNA expression regulation.
We showed no relationship between CYP11B2 methylation levels and CYP11B2 expression in APA and compared the results with the previous studies. A previous study reported only one of the promoter methylation sites, which correspond to d in our study, and it had a significant relationship with CYP11B2 expression in all adrenal tissues combined with APA, NF, and normal adrenal gland.17 In applied same eligible tissues as the previous study, there was strong inverse correlation between the methylation rate and CYP11B2 expression (Figure S4). Therefore, our results that APAs are hypomethylated in d compared with NF are consistent with the previous report. Importantly, the relationship would be explained by the difference between APA and NF. Remarkably, in our study using a large sample of APAs, 7 methylation sites of CYP11B2 were hypomethylated, thus suggesting that full hypomethylation of CYP11B2 could influence the CYP11B2 transcription in APA.
CYP11B1 in APA showed a lower mRNA expression and higher DNA methylation rate compared with NF (Table 1; Figure S2), whereas there were no differences of serum cortisol levels by 1 mg dexamethasone suppression test between APA and NF (Table S1). Previous studies revealed that there were no differences in CYP11B1 mRNA expression between APA and normal adrenal gland,24,25 whereas in our study CYP11B1 expression seemed to be decreased in APA compared with NF (Figure S2). All of the CYP11B1 hypermethylation sites with hypermethylation existed outside the promoter region. It has been shown that exon or intron methylation can also regulate transcription by some possible mechanisms,26,27 and thus CYP11B1 transcription in APA might be regulated by nonpromoter DNA methylation. However, the CYP11B1 regulation machinery might not influence cortisol production in APA and NF, considering with the results of 1 mg dexamethasone suppression test.
We also showed the effects of KCNJ5 and ATP1A1 mutations on DNA methylation of CYP11B2 and CYP11B1. CYP11B1 methylation was not observed to be influenced by the mutations in APA, whereas a methylation site in CYP11B2, indicated as a in our study, significantly differed among APAs with different mutations. The methylation site was not regulated between APA and NF, and there was no effect of a KCNJ5 mutation on the methylation status in vitro study. Therefore, it was thought that the site was not important for CYP11B2 transcription in APA. However, we need further ATP1A1 mutant APAs to detect reliable associations and basic study in HAC15 introducing an ATP1A1 mutation. More importantly, according to our in vitro study, CYP11B2 and CYP11B1 methylation or demethylation might not be regulated by a KCNJ5 mutation. It is suggested that the regulation machinery of CYP11B2 transcription and methylation could be different, and intracellular Ca2+ signaling stimulates CYP11B2 transcription but not CYP11B2 DNA methylation. Taken together with our results and previous studies,14,15 CYP11B2 mRNA expression in APA would be basically regulated by transcription factors, and hypomethylation of CYP11B2 might facilitate or potentiate CYP11B2 transcription induced by transcription factors.
Limitations arise from the cross-sectional study design, which do not allow the analysis of methylation and demethylation causality in APA and NF. A previous study applied in vitro performed methylation or demethylation analysis using a methyltransferase or demethyltransferase, respectively.17 However, these agents could regulate other mRNA expressions that regulate the CYP11B2 transcription by altering their methylation status, and thus it did not address the direct analysis between methylation and mRNA expression. Further experiments will be needed to regulate DNA methylation or demethylation at targeted sites. This study included a larger sample size than previous studies using the HumanMethylation450 BeadChip,17,18 hence the results here study for the methylation state of steroidogenic enzyme genes in APA including somatic mutations are valuable.
Perspectives
APA exhibited hypomethylated CYP11B2, and CYP11B2 methylation in the region with hypomethylation was not induced by KCNJ5 or ATP1A1 mutations that cause aldosterone overproduction in APA and a KCNJ5 mutation in HAC15 cells. CYP11B2 mRNA expression in APA would be basically regulated by transcription factors, and hypomethylation of CYP11B2 might facilitate or potentiate CYP11B2 transcription induced by transcription factors. Therefore, our results about the CYP11B2 DNA methylation level in APA may be important for basic research helping to further clarify intracellular mechanisms of aldosterone production or to create therapeutic targets for aldosterone production regulation.
Acknowledgments
This work was performed with the kind cooperation of the Analysis Center of Life Science, Hiroshima University.
Sources of Funding
This study was financially supported by JSPS KAKENHI (Grants-in-Aid for Scientific Research from Japan Society for the Promotion of Science) Grant Number JP30638995 (K. Oki), The Salt Science Research Foundation Grant Number 1424 (K. Oki), Takeda Science Foundation (K. Oki), and National Institutes of Health grant HL27255 (C.E. Gomez-Sanchez).
Disclosures
None.
Footnotes
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.116.08313/-/DC1.
- Received August 13, 2016.
- Revision received June 30, 2016.
- Accepted September 25, 2016.
- © 2016 American Heart Association, Inc.
References
Novelty and Significance
What Is New?
This study is the first study showing that aldosterone-producing adenoma (APA) with a large sample size exhibited extensively hypomethylated CYP11B2 using DNA methylation array analysis
CYP11B2 methylation level was not associated with CYP11B2 transcription levels in APA.
HAC15 cells with a KCNJ5 mutation showed no changes in CYP11B2 or CYP11B1 methylation levels compared with control cells.
What Is Relevant?
Hypomethylated CYP11B2 might facilitate or potentiate CYP11B2 transcription induced by transcription factors in APA.
Summary
APA exhibited hypomethylated CYP11B2, and DNA methylation of CYP11B2 was not induced by KCNJ5 or ATP1A1 mutations that cause aldosterone overproduction in APA and a KCNJ5 mutation in HAC15 cells.
This Issue
Jump to
Article Tools
Share this Article
- Hypomethylation of CYP11B2 in Aldosterone-Producing AdenomaNovelty and Significance
