Introduction

Primary aldosteronism (PA) is characterized by autonomous aldosterone production with increased aldosterone to renin ratio1. The increased level of aldosterone leads to abnormal retention of sodium and water and potassium excretion, causing hypervolemic hypertension and in more severe cases, hypokalemia2. Once thought to be a rare disease, PA is now being recognized as the most common cause of secondary hypertension with a prevalence of 5–10% in hypertensive individuals and up to 20% in therapy-resistant hypertension3,4,5,6. Most commonly, PA is either caused by unilateral aldosterone producing adenoma (APA) or bilateral adrenal hyperplasia (BAH)3,4. Unilateral APAs represent about 30% of PA cases3,4,5. Dominant unilateral aldosterone production caused by either APA or asymmetrical aldosterone release in the context of BAH can be cured by surgery, while patients with symmetrical BAH should be recommended mineralocorticoid receptor antagonists7. The importance of detecting PA patients among hypertensive individuals is not only due to the possibility for specific treatment and potential cure, but also because PA is an independent risk factor for increased cardiovascular morbidity, caused by excess aldosterone, independent of blood pressure8,9,10,11.

The WNT signaling pathway is important for normal development and maintenance of the adrenal cortex and more specifically, the zona glomerulosa (ZG) within the cortex12,13. Exon 3 of the CTNNB1 gene (encoding β-catenin), contains specific serine and threonine residues that, when phosphorylated, mark β-catenin for degradation14,15. Mutations affecting these amino acids, or deletions of exon 3, subsequently preclude phosphorylation of β-catenin, leading to aberrant activation of WNT signaling16,17,18,19,20.

Active WNT signaling has been observed in APAs21,22,23. The Secreted Frizzled-related protein II (SFRP2) is a negative regulator of the WNT pathway24 and its down-regulation has been proposed as one of the mechanisms for this increase in WNT signaling21. Mice with abnormal WNT signaling, due to alterations in exon 3 of the CTNNB1 gene, develop hyperaldosteronism and adrenocortical tumors23,25. Mutations in CTNNB1 also cause increased, abnormal WNT activation in human adrenocortical tumors26,27. Two independent exome sequencing experiments have demonstrated CTNNB1 mutations in two APAs28,29. The aim of this study was to determine the prevalence of CTNNB1 mutations in APAs.

Results

Mutation status of CTNNB1

The included APA samples exhibited mutations in KCNJ5 (n = 92, 46.5%), ATP1A1 (n = 6, 3.0%), ATP2B3 (n = 3, 1.5%), CACNA1D (n = 3, 1.5%), while 94 (47.5%) did not harbor mutations in known genes30,31. Somatic missense mutations in CTNNB1 were observed in 10 APAs (5.1% in the total cohort and in 10.6% of APAs without known mutations). No large deletions of exon 3 were observed in investigated non-mutated tumors (n = 7, Supplementary Fig. S1 online). The prevalence of CTNNB1 mutations was similar among different national cohorts (Supplementary Table S1 online). The detected mutations all occurred in conserved serine/threonine residues in exon 3 (Supplementary Fig. S2, Supplementary Table S2 online). The mutations were absent from constitutional DNA in all available specimens (n = 7), confirming their somatic nature. All mutations in CTNNB1 occurred mutually exclusive to KCNJ5, ATP1A1, ATP2B3 and CACNA1D (p = 1.47e-14, Fisher’s exact test).

Aldosterone production in CTNNB1 mutant tumors

We analyzed CYP11B2 mRNA expression by semi-quantitative RT-PCR in CTNNB1 mutated and KCNJ5 mutated tumors, as well as in NHPAs and CPAs. The investigated CTNNB1 mutants displayed higher CYP11B2 expression than NHPAs, CPAs and also significantly higher than tumors with KCNJ5 mutations (p = 0.011) (Fig. 1). CYP11B2 protein expression was analyzed by IHC using a mouse monoclonal antibody. Staining of normal adrenal tissue revealed heterogeneous expression selective for ZG (Fig. 2a). All investigated CTNNB1 mutated tumors (n = 8) exhibited expression of CYP11B2. Four had diffuse expression throughout the tumor (Fig. 2b), while four displayed limited heterogeneous staining (Fig. 2c). To further confirm the ability of CTNNB1 mutated tumors to produce aldosterone, aldosterone concentration were directly measured in tumor tissue lysates utilizing an automated immunoassay. The investigated CTNNB1 mutated APA lysates (n = 3) contained aldosterone at similar levels as KCNJ5 mutated tumor lysates (p = 0.48, Fig. 3).

Figure 1
figure 1

CYP11B2 expression in APAs.

Relative mRNA expression of CYP11B2 in APAs with CTNNB1 mutation (n = 5) compared to APAs with KCNJ5 mutation (n = 16), NHPA (n = 4) and CPA (n = 4) using the 2(−ΔΔCT) method. GAPDH was used as the reference gene. Values indicate median with range. Y-axis shows logarithmic values. NHPA = Non-hormone producing adenoma. CPA = Cortisol producing adenoma. Statistical analysis using Mann -Whitney U-test.

Full size image
Figure 2
figure 2

CYP11B1/B2 protein expression.

Immunohistochemical analysis of CYP11B1/B2 expression. (a) CYP11B2 staining of normal adrenal gland. (b) CYP11B2 staining of a tumor with CTNNB1 mutation displaying diffuse expression. (c) CYP11B2 staining of a tumor with CTNNB1 mutation displaying weak heterogeneous expression. (d) CYP11B1 staining of normal adrenal gland. (e) CYP11B1 staining of a tumor with CTNNB1 mutation displaying weak heterogeneous expression. (f) CYP11B1 staining of a tumor with CTNNB1 mutation displaying diffuse expression.

Full size image
Figure 3
figure 3

Aldosterone concentration in tumor lysate (pmol/mg).

Aldosterone concentration in tumor lysates using a chemiluminescent immunoassay. Statistical analysis using Mann -Whitney U-test.

Full size image

Histologic appearance and CYP11B1/B2 protein expression in APAs with CTNNB1 and KCNJ5 mutation

Histologic examination of CTNNB1 mutated tumors did not reveal any specific cell morphology (Supplementary Table S3 online). CYP11B1 antibody specificity was confirmed in normal adrenal tissue (Fig. 2d). Staining of eight CTNNB1 mutated tumors revealed two distinct subgroups. In tumors with high CYP11B2 expression the CYP11B1 expression was also low (Fig. 2b,e). Similarly, in those with low CYP11B2 expression the CYP11B1 staining was strong (Fig. 2c,f). Comparing with KCNJ5 mutated APAs, no difference in CYP11B1 expression levels were observed, but a significantly higher CYP11B2 expression in CTNNB1 mutated tumors were observed (p = 0.02, Fig. 4a,b).

Figure 4
figure 4

Comparison of CYP11B1/B2 expression score in APA.

Immunohistochemical expression score of CYP11B2 and CYP11B1 in tumors with CTNNB1 and KCNJ5 mutation.Y-axis display the expression score of CYP11B2, calculated using a modified H-score. (a) CYP11B2. (b) CYP11B1. Median with range. Statistical analysis using Mann-Whitney U-test.

Full size image

β-catenin accumulation and activated WNT signaling in tumors with CTNNB1 mutation

All investigated APAs with CTNNB1 mutation displayed cytoplasmic and/or nuclear β-catenin staining (n = 8, Fig. 5) and displayed high levels of active β-catenin by western blot (n = 3, Fig. 6). No apparent nuclear staining was observed in tumors without CTNNB1 mutation, although 11/24 (45.8%) displayed cytoplasmic staining (H-score ≥50, Supplementary Table S4 and Supplementary Fig. S3 online). In the tumors without CTNNB1 mutation, β-catenin expression correlated with ZG-like cell content (r2 = 0.61, p < 0.0001, Supplementary Fig. S4 online). We did not observe any genotype-expression correlation as tumors with KCNJ5 mutation and ZG-like cell content displayed cytoplasmic expression of β-catenin (Supplementary Fig. S3 online). Western blot in one tumor with CTNNB1 mutation (p.Ser45Pro) and four without CTNNB1 mutation, validated antibody specificity for β-catenin (Supplementary Fig. S5 online). To analyze downstream targets of β-catenin, AXIN2 mRNA expression was measured. A significantly higher expression of AXIN2 in tumors with CTNNB1 mutation compared to those with KCNJ5 mutation was observed (p = 0.0029, Fig. 7). In five tumors without known mutations we observed high levels of AXIN2 suggesting active WNT signaling.

Figure 5
figure 5

Immunohistochemical expression analysis of β-catenin.

(a) APA with CTNNB1 mutation displaying nuclear and cytoplasmic β-catenin expression. (b) APA with CTNNB1 mutation displaying heterogeneous cytoplasmic expression. (c) APA with CTNNB1 mutation with membranous and nuclear expression.

Full size image
Figure 6
figure 6

Western blot analysis using an anti-active β-catenin antibody.

Western blot analysis on three APAs with CTNNB1 mutation (1–3), one APA without known mutation (4) and two APAs with KCNJ5 mutation (5–6). Top row shows signal for active β-catenin (97 kD), lower row shows actin signal (43 kD).

Full size image
Figure 7
figure 7

AXIN2 mRNA expression.

Relative mRNA expression of AXIN2 in APAs with CTNNB1 mutation (n = 5) compared with APAs with KCNJ5 mutation (n = 16), APAs with ATP1A1 and ATP2B3 mutation (n = 6) and seven APAs without known mutation, using the 2(−ΔΔCT) method. GAPDH was used as the reference gene. Values indicate median with interquartile range. Statistical analysis using Mann -Whitney U-test.

Full size image

Clinical characteristics of patients with tumors harboring CTNNB1 mutation

The clinical data of the patients with tumors harboring CTNNB1 mutation are summarized in Supplementary Table S5 online. CTNNB1 mutated APAs were more often found in female patients (60%). There were no significant differences in preoperative aldosterone levels, tumor size, or age at surgery in patients with tumors harboring CTNNB1 mutation compared to those with tumors harboring KCNJ5 mutation (Supplementary Table S5 online). However a significant difference in size between the tumors with CTNNB1 mutation and tumors without known mutation was observed, median 24 mm (range 10–45) vs median 14 mm (range: 3–45, mean value 14.7 mm ± 7.9SD) (p = 0.01).

Discussion

WNT signaling is essential for normal development and maintenance of the adrenal cortex13. In the canonical WNT signaling pathway, absence of WNT ligands induces formation of a complex consisting of Dishevelled, AXIN, APC, CKI and GSK3β which binds and phosphorylates β-catenin, leading to its ubiquitination and subsequent degradation by proteasomes32. In contrast, WNT ligand binding to the frizzled and LRP co-receptors causes sequestering of the destruction complex at the cytoplasmic membrane, leaving it unable to phosphorylate. This leads to β-catenin accumulation, translocation to the nucleus and altered transcription of downstream targets, including: T-cell factor/lymphoid enhancer factor (TCF/LEF-1), CYP21, the Angiotensin I receptor and CYP11B221,33,34. β-catenin activation in the adrenal gland is normally restricted to cells of the ZG22. Stabilizing mutations in the CTNNB1 gene increase the activity of the finely tuned WNT signaling pathway, leading to tumor formation16. Previous reports have shown that mutations in β-catenin occur in a small proportion of APAs26,28,29,35. Yet contradictorily, two different studies sequencing APAs did not detect any CTNNB1 mutations21,22. Here we performed CTNNB1 sequencing in a large cohort of clinically and molecularly diagnosed APAs. We detected CTNNB1 mutations in 5.1% of cases, all affecting previously described mutation “hotspot” amino acids p.Thr41 and p.Ser45. All CTNNB1 mutations occurred mutually exclusive to KCNJ5, ATP1A1, ATP2B3 and CACNA1D mutated tumors, implying that aberrant WNT activation may be enough for APA formation as indicated in mouse models23. Tumors with CTNNB1 mutations were associated with relatively large adenomas and were found both in female and male patients, however with a small overrepresentation in females.

To analyze the effect on WNT signaling, we performed β-catenin IHC and quantitative PCR on AXIN2. We observed nuclear and/or cytoplasmic β-catenin staining in tumors with CTNNB1 mutation, indicating aberrant, increased protein accumulation. We also noted APAs without CTNNB1 mutation having cytoplasmic β-catenin expression and also those with relatively high mRNA expression of AXIN2. This suggests that WNT signaling is activated in a large subgroup of APAs and independent of genetic background22, raising the question if other genetic aberrations in the WNT pathway may be involved in development of APAs.

CYP11B2 and CYP11B1 are highly homologous enzymes expressed in ZF and ZG respectively. CYP11B1 converts 11-deoxycortisol to cortisol, whereas CYP11B2 converts deoxycorticosterone successively to corticosterone, 18-OH-corticosterone and finally aldosterone. CYP11B2 expression indicates aldosterone production and most APAs display expression, although in some cases heterogeneously36,37. Utilizing specific mRNA probes and primers for CYP11B238, we found expression of CYP11B2 mRNA in tumors with CTNNB1 mutations indicating ability to produce aldosterone. IHC utilizing an antibody against CYP11B2 showed high specificity for ZG. We verified expression of CYP11B2 in all CTNNB1 tumors. Four CTNNB1 mutated tumors demonstrated diffuse expression, while four tumors exhibited heterogeneous and relatively weak staining, similar to the pattern seen in many of the tumors with KCNJ5 mutation36,37. To further confirm aldosterone production, we measured aldosterone concentration in tumor lysates using a chemiluminescent immunoassay. The investigated mutants contained aldosterone and at similar levels to other APAs with KCNJ5 mutations indicating an active aldosterone production. Measuring aldosterone in tumor tissue may provide a diagnostic tool to discriminate functional aldosterone producing nodules from non-functional and with introduction of faster assays, could even be employed in the operating theatres.

APAs may contain a varying degree of ZG-like cells and ZF-like cells39. Some reports indicate that this phenotype depends on underlying genetic alteration29,37,40, whereas others do not41. Also, a difference in the protein expression pattern of CYP11B2/B1 has been observed between the different mutants36,37. By histologic examination, we observed a variable histologic appearance in CTNNB1 mutated tumors. Interestingly, we observed two subgroups of tumors; one with diffuse CYP11B2 expression with concomitantly low CYP11B1 expression and one with low CYP11B2 and high CYP11B1 expressions. Perhaps suggesting that these mutations mainly promote proliferation.

CTNNB1 mutations occur in cortisol producing adenomas (CPA). Recently, mutations in the catalytic subunit of Protein kinase A (PKA) were identified and shown to occur mutually exclusive to CTNNB1 mutations42,43,44,45. Activating mutations in PKA lead to a constitutively activated cAMP signaling, causing increased cortisol production and likely tumor formation. Expression analysis revealed increased expression of genes involved in biosynthesis and metabolism of steroids in tumors with PRKACA mutation42. This suggests two pathways for CPA formation: either alterations in cAMP signaling with increased steroidogenesis, or other mutations mainly stimulating proliferation. In APAs, most mutations observed affect the membrane potential and increase the likelihood of depolarization. These tumors harbor few mutations per megabase and seldom have CNV alterations28. In contrast, many APAs without known mutation have CNV alterations28, suggesting different pathways for APA development, similar to CPAs. We propose that CTNNB1 mutations belong to a group of alterations mainly affecting proliferation rate. The low prevalence of mutations in APAs compared to NHPAs or CPAs could reflect the low number of ZG cells expressing CYP11B2 in normal adrenal gland.

It is unknown if carcinomas of the adrenal cortex evolve in an adenoma-to-carcinoma sequence46. Evidence in favor of this hypothesis comes from experiments using β-catenin mutated mice that develop benign tumors that can progress to malignancy, suggesting requirement of additional genetic alterations23,25. This is in line with the multistep progression model seen in patients with familial adenomatous polyposis47. Moreover, signs of carcinomas inside benign adrenal masses have been described48,49,50. However, most APAs rarely increase in size and aldosterone producing carcinomas are exceedingly rare7. If adenoma to carcinoma sequences does occur, one could speculate that patients with tumors harboring CTNNB1 mutation have a small but increased risk of malignant transformation. Another interesting observation, coming from the β-catenin stabilized mouse model is that up to 10 months of age the mice produce aldosterone and show no signs of corticosterone production. Despite this, 17-month-old mice exhibit a trend towards lower aldosterone synthesis and increased corticosterone production, suggesting that the level of WNT activity determine the functional characteristics of the lesion23. This could also explain why mutations in CTNNB1 are less frequent in APAs.

In conclusion, we observed CTNNB1 mutations in a subset of APAs with aberrant β-catenin accumulation. Tumors harboring these mutations have a variable histological and CYP11B2/B1 expression pattern, produce aldosterone and show similar clinical characteristics as patients with tumors harboring KCNJ5 mutation. We speculate that CTNNB1 mutations lead to a proliferative advantage in the adrenal cell it occurs in. Because of the higher frequency of WNT activation in APAs than CTNNB1 mutations, future investigations of alterations in other members of the WNT pathway are warranted.

Materials and Methods

Tissues

Adrenocortical tumor tissues were collected from six different centers. According to routine protocols at the centers, a diagnosis of APA was established by these criteria: an elevated aldosterone/renin ratio together with positive confirmatory tests and lateralization studies by CT, MRI and/or adrenal vein sampling and a postoperative cure or considerable improvement. Patients with cortisol producing adenomas (CPA) had elevated serum cortisol, failure to suppress cortisol with dexamethasone, normal aldosterone levels and all were postoperatively cured. Non-hormone producing adenomas (NHPA) were detected en passant by abdominal CT scan in biochemically normal and asymptomatic patients and were removed due to large size or malignant features on CT.

DNA and RNA extraction

DNA and RNA from frozen tissues were extracted using Allprep DNA/RNA kit, (Qiagen, Hilden, Germany) or FFPE tissue using AllPrep DNA/RNA FFPE Kit (Qiagen, Hilden, Germany) and from blood using DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). Synthesis of cDNA was performed using the First-Strand cDNA Synthesis kit according to the manufacturer’s instructions (Fermentas, Thermo Fisher Scientific, Waltham, USA).

PCR amplification and DNA sequencing

Exon 3 of the CTNNB1 gene was amplified by PCR amplification, using primers: FW: 5′-TGTCTTTCAGATTTGACTTTATTT, RW: 5′-TCAAAACTGCATTCTGACTTTCA followed by a nested reaction with FW2: 5′-CAATGGGTCATATCACAGATTCTT, then Sanger sequenced. Traces were analyzed and compared to the CTNNB1 reference sequence NP_001895.1. Observed mutations were validated in an independent PCR reaction. cDNA was sequenced using mRNA specific primers, cDNA FW: 5′-GCTACTCAAGCTGATTTGATG and RW: 5′-CAGGACTTGGGAGGTATCC. Large deletions affecting exon 3 of the CTNNB1 gene were investigated by RT-PCR amplification using mRNA specific primers, FW primer: 5′-GAAGGTCTGAGGAGCAGCTTC and RW: 5′-GACATTAGATGAGGGCATGC. The obtained amplicons were run on a 2% agarose gel to visualize size of the products. All tumors had previously been sequenced for KCNJ5, ATP1A1, ATP2B3 and CACNA1D mutations31.

Histology examination and immunohistochemistry

Tumors were evaluated to determine their zona glomerulosa (ZG) and zona fasciculata (ZF) cell content. Cells were classified as ZG-like if they had a high nuclear to cytoplasmic ratio and formed circular cell clusters. Cells were classified as ZF-like if they were lipid-rich with a high cytoplasm to nuclear ratio and a more uniform sponge-like phenotype39. IHC for β-catenin was conducted on eight tumors with CTNNB1 mutation and 24 tumors without CTNNB1 mutation, using an anti-β-catenin goat polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA; catalog no. sc-1496, dilution 1:5) as previously described18. Staining of CYP11B1 and CYP11B2 were conducted using rat and mouse monoclonal antibodies respectively51. Briefly, 5 μm sections were deparaffinized in xylene and rehydrated through graded ethanol. Endogenous peroxidases were inhibited by incubation in hydrogen peroxide solution. Antigen retrieval using 10 mM sodium citrate (pH6.0) was used. Sections were then incubated with 10% normal goat serum/horse serum after which the primary antibody was added (CYP11B1, 1:100 and CYP11B2, 1:250) overnight. For detection, biotinylated HRP conjugated goat anti-rat or horse anti-mouse secondary antibodies were used, 1:200 (Vector Laboratories, Burlingame, USA). The sections were developed with DAB (Vector Laboratories, Burlingame, USA) and counterstained with hematoxilin. Two researchers, unaware of the mutational status, conducted comparison of expression. A scoring system using a modified H-score was applied52. Five 20× fields and the entire sample were scored on intensity i (0 = negative, 1 = weak but detectable, 2 = distinct, 3 = strong) and percentage of stained cells Pi. A mean value of these six measurements was used in the formula ∑ = Pi(i + 1). Images were acquired using a Zeiss Observed Z1 microscope (Zeiss, Oberkochen, Germany) and imported using the AxioVision SE64 4.9.1 software (Zeiss, Oberkochen, Germany).

Western blot

To verify specificity of the β-catenin antibody used in the IHC experiment, western blot analysis was performed on one CTNNB1 mutated APA (p.Ser45Pro) and four APAs without CTNNB1 mutation (Santa Cruz-1496; dilution 1:1000), with anti-actin (Santa Cruz-1616) as loading control. Western blot was also performed on three tumors with CTNNB1 mutation, two with KCNJ5 mutation and one APA without known mutation, using an anti-active β-catenin antibody (Millipore 05–655; dilution 1:1000) and anti-actin (Santa Cruz-1616) as loading control.

Semiquantitative RT-PCR

The expression of CYP11B2 was measured using an mRNA specific custom assay (ThermoFisher Scientific, Waltham, USA)38. AXIN2 mRNA expression was quantified using SYBR green with FW primer: 5′-GGTCCACGGAAACTGTTGACA and RW: 5′-GGCTGGTGCAAAGACATAGCC. An mRNA specific GAPDH assay was used as internal control (4352934, ThermoFisher Scientific, Waltham, USA). The expression was measured in duplicates and a relative difference was obtained using the 2(−ΔΔCT) method53.

Chemiluminescent immunoassay

Sections (20 μm) of frozen tumor tissue from three tumors with CTNNB1 mutations, six with KCNJ5 mutation and one NHPA, were weighed with a highly sensitive scale (±0.01 mg) and homogenized in filtered water. The lysate was again weighed and a concentration of mg of tumor tissue per mg of water was calculated. The samples were then centrifuged for 20 minutes at 20000 RPM. The obtained supernatant was diluted 1:10 and subjected to an aldosterone measurement using a chemiluminescent immunoassay (DiaSorin, Italy) used in routine clinical practice at the Department of Clinical Chemistry at Uppsala University hospital. A concentration in pmol/L was obtained and adjusted to the concentration. This gave a concentration of aldosterone/mg tissue (pmol/mg).

Statistics

The overall group effect was analyzed using a Kruskal-Wallis test. If significant group effects were present the overall analysis was followed by post-hoc comparisons between groups using a Mann-Whitney U-test with Bonferroni correction. Categorical data were analyzed by Chi-squared tests. Data are expressed as either mean ± SD or median and range for non-parametric data. For statistical analysis SPSS 22 (IBM, NY, USA) was used.

Additional Information

How to cite this article: Åkerström, T. et al. Activating mutations in CTNNB1 in aldosterone producing adenomas. Sci. Rep. 6, 19546; doi: 10.1038/srep19546 (2016).