Primary aldosteronism (PA) is a group of disorders caused by an autonomous overproduction of aldosterone from the adrenal glands, independent of the renin-angiotensin system (Stowasser and Gordon, 2016). PA is the most common form of secondary hypertension, with an estimated prevalence reaching more than 10% in the hypertensive population (Rossi et al., 2006) and even approximately 20% in resistant hypertension patients (Calhoun et al., 2002).

Conventional hematoxylin-eosin (HE) staining performed in most pathological laboratories provides only morphological information; however, it is unsuitable for functional histopathological analyses (Nakamura et al., 2014). For instance, the differential distinction between an adenoma and a nodular hyperplasia is unclear on HE staining because many aldosterone-producing adenomas (APAs) also exhibit peritumoral zona glomerulosa (ZG) hyperplasia (Ganguly, 1998; Enberg et al., 2004; Boulkroun et al., 2010; Yamazaki et al., 2017). CYP11B2 and CYP11B1, which catalyze the final biosynthesis steps of aldosterone and cortisol, respectively, share a 93% homogeneity at the amino acid sequence level (Gomez-Sanchez et al., 2014). With the successful development of specific antibodies against CYP11B2 and CYP11B1 (Gomez-Sanchez et al., 2014), immunohistochemistry is now a promising technique for precisely locating the culprit responsible for aldosterone excess and for broadening our knowledge regarding the spectrum of PA. Nishimoto et al. (Nishimoto et al., 2010) have revealed an entity of subcapsular CYP11B2-expressing cells, termed aldosterone-producing cell cluster (APCC), which are 200–1,300 μm wide and 100–500 μm deep and can sustain aldosterone secretion in spite of a suppressed renin-angiotensin-aldosterone system. Later, Omata et al. (Omata et al., 2018) demonstrated that bilateral idiopathic hyperaldosteronism may result from the accumulation and enlargement of APCCs which harbor somatic aldosterone-driver gene mutations rather than diffuse aldosterone-producing ZG hyperplasia. Hence, CYP11B2 immunostaining can markedly improve the accuracy of PA histopathological diagnosis.

The Wnt/β-catenin signaling pathway is critical for regulating the proliferation, differentiation, and tumorigenesis of the adrenal cortex (El Wakil and Lalli, 2011; Clevers and Nusse, 2012). And several studies have reported mutaion burden of CTNNB1 gene which encodes β-catenin in APA (Scholl et al., 2015; Akerstrom et al., 2016). The frequency of somatic CTNNB1 mutation in APA ranges from 2.1 to 5.1% (Scholl et al., 2015; Akerstrom et al., 2016). Most CTNNB1 mutations affect the serine/threonine residues in exon 3, which thereby hamper the proteasomal degradation of β-catenin under resting conditions, leading to abnormal intracellular accumulation of β-catenin. Actually, constitutive activation of Wnt/β-catenin signaling is associated with aldosterone overproduction and treatment of H295R cell line, which already has an activating mutation of CTNNB1 endogenously, with siRNA against β-catenin or β-catenin inhibitor can significantly decrease aldosterone secretion (Berthon et al., 2014). However, the mechanism underlying how β-catenin provokes mineralocorticoid overproduction is undetermined; possible explanations include the indirect activation of CYP11B2 promoters (Berthon et al., 2014) or increased expression of LHCGR and GNRHR by β-catenin in pregnant and postmenopausal patients (Teo et al., 2015).

Laparoscopic adrenalectomy is the first-line choice for treating unilateral PA and excess aldosterone secretion can be resolved in most cases (Stowasser and Gordon, 2016). But not all PA patients can achieve hypertension cure or remission after the surgery, the factors that may affect clinical outcome include age, gender, family history of essential hypertension, hypertension duration, BMI, antihypertensive drug dosage and so on (Steichen et al., 2012; Williams et al., 2017). Nevertheless, individual prediction of the clinical outcome based on the above factors is not completely accurate, and more predictors based on new emerging technology need to be identified (Stowasser and Gordon, 2016). To date, the association between immunostaining features and prognosis of unilateral PA after surgery still remains an uncharted area and warrants further investigation.

Therefore, the aim of our current study is to utilize CYP11B2 immunostaining to subtype clinically diagnosed unilateral PA and explore the functional and prognostic significance of CYP11B2, CYP11B1 and β-catenin immunostaining.

Herein, our study firstly evaluated the association between steroidogenic enzyme immunoreactivity, β-catenin immunostaining status, clinical parameters and outcome in patients with unilateral PA. Differential diagnosis between unilateral APA and UAH can sometimes be puzzling and difficult when using the traditional HE staining technique, because the ZG of the peritumoral tissue adjacent to APA can sometimes display different levels of hyperplasia, and HE staining cannot identify the precise lesion that is responsible for aldosterone overproduction (Ganguly, 1998; Enberg et al., 2004; Boulkroun et al., 2010; Nakamura et al., 2014; Yamazaki et al., 2017). In situ hybridization of CYP11B2 has been used to identify aldosterone-producing cells; however, this technique cannot be performed at a large-scale as it is time-consuming and difficult to implement. Using a specific monoclonal antibody for CYP11B2, immunohistochemistry staining of steroidogenic enzyme can be an advantageous tool to precisely localize the cells producing aldosterone and markedly improve the diagnostic accuracy as it is faster, more convenient and less expensive than in situ hybridization (Nanba et al., 2013; Gomez-Sanchez et al., 2014). Here, 30 patients were initially diagnosed with UAH based on HE staining, and then 27 of them were re-diagnosed with APA after CYP11B2 immunostaining, as positive nodular staining was observed. Moreover, we found more than one APAs on the tissue sections of 11 patients via CYP11B2 immunostaining, many of which were undetectable under CT examination. Yamazaki et al. (Yamazaki et al., 2017) classified CT-negative PA into two subtypes based on CYP11B2 immunostaining: multiple adrenocortical micronodules and diffuse hyperplasia. They also reported that the ZG of adjacent adrenals near micronodules often demonstrated various degrees of paradoxical hyperplasia, which is histopathologically considered hyperplasia but is negative for CYP11B2 staining. Although all patients in our study showed positive CT findings, our result is consistent with that of Yamazaki et al. Paradoxical hyperplasia of ZG cells was also observed in our study. Interestingly, neither diffuse CYP11B2-positive hyperplasia of ZG cells nor positive nodular staining was observed in the remaining three UAH patients; but several APCCs within the adrenal cortex were observed on their sections. Nishimoto et al. (Nishimoto et al., 2010) firstly identified a novel cell cluster positive for CYP11B2 immunostaining, which they termed as APCC. Several studies (Nishimoto et al., 2010; Nanba et al., 2017; Omata et al., 2017; Omata et al., 2018) have demonstrated that APCCs frequently harbor aldosterone-driver somatic mutations mostly in the CACNA1D gene, and can autonomously secrete mineralocorticoid independent of the renin-angiotensin system under normal and pathological conditions. Omata et al. (Omata et al., 2018) elucidated that idiopathic hyperaldosteronism (IHA) may result from the accumulation and enlargement of APCCs instead of hyperplasia of ZG cells. Moreover, Nanba et al. (Nanba et al., 2013) observed APCCs in unilateral PA patients with ipsilateral adrenocortical tumor negative for CYP11B2 staining. According to these findings, we hypothesized that APCCs may be the source of aldosterone excess in certain patients with unilateral PA rather than diffuse hyperplasia of aldosterone-producing ZG cells or adenomas. Additionally, one patient in our study had a CYP11B2-negative adrenocortical tumor and one APCC on his section, indicating that APCC rather than the adrenocortical tumor found by CT may be responsible for the aldosterone overproduction. Thus, our findings, in accordance with other studies, indicated that CYP11B2 staining can more precisely pinpoint the culprit responsible for aldosterone overproduction compared to traditional HE staining and effectively differentiate between different subtypes of unilateral PA. Further, APCC may play a crucial role not only in normal blood pressure maintenance (Nanba et al., 2017; Omata et al., 2017) but also in both unilateral and bilateral, CT-negative and CT-positive hyperaldosteronism (Nanba et al., 2013; Yamazaki et al., 2017; Omata et al., 2018).

Immunoreactivity was assessed using the McCarty H-score, which is now considered the best available approach for semi-quantitatively evaluating the immunoreactivity of steroidogenic enzymes in an objective manner in a specific tumor sample (McCarty et al., 1985). In our study, the CYP11B2 H-score correlated with serum aldosterone and serum potassium, but not with PRA, ARR, or serum cortisol. Meanwhile, the CYP11B1 H-score correlated with serum cortisol, but not with serum aldosterone, PRA, or ARR (data not shown). Several studies (Nanba et al., 2013; Nakamura et al., 2014; Ono et al., 2014) have shown that steroidogenic enzyme H-score level alone does not necessarily represent the overall steroid hormone production capacity of an adrenocortical tumor because it only represents the hormone production ability of the tumor per unit area and per cell, and there may be several other factors such as tumor area and upstream steroidogenic enzymes of CYP11B2/1 that may also play a role in hormone production. Therefore, the functional significance of CYP11B2/1 immunostaining was investigated by correlating the H-score adjusted for tumor area with the endocrine data. Through this adjustion, the adjusted CYP11B2 score was correlated with serum aldosterone, ARR, and serum potassium, and the adjusted CYP11B1 score was correlated with serum 8:00 am cortisol, serum aldosterone, ARR, and PRA. These correlationships provide further evidence that adjusted H-score can be used as a new marker to reflect clinical manifestations and severity of unilateral PA.

APA is a heterogenous lesion composed of cells originating from different adrenocortical zones. Both ZF-like cells (cells with large, vacuolated, lipid-laden cytoplasm and central, round nuclei similar to that of ZF) and ZG-like cells (cells with relatively small, lipid-poor cytoplasm and a higher nuclear to cytoplasmic ratio similar to that of ZG) are observed in the majority of APAs (Enberg et al., 2004; Nakamura et al., 2014; Gioco et al., 2015; Fallo et al., 2017). Our results were consistent with the histopathological findings and provide further immunohistochemical evidence on this perspective. Functionally, CYP11B2 is restricted in ZG, catalysing the final steps of mineralocortical biosynthesis, and CYP11B1 is restricted in ZF and zona reticularis (ZR), catalysing the final step of glucocorticoid formation (Stowasser and Gordon, 2016). Our results suggest that APA can simultaneously produce cortisol in patients with PA, which may lead to clinical or subclinical manifestation of Cushing syndrome (Goupil et al., 2015; Stowasser and Gordon, 2016; Fallo et al., 2017). A positive correlationship between the adjusted CYP11B1 score and serum cortisol was revealed in our study, which emphasized the importance of careful monitoring of the hypothalamic-pituitary-adrenal axis before and after surgery in patients with unilateral PA.

Interestingly, the adjusted CYP11B1 score not only correlated with serum cortisol, but also with serum aldosterone, ARR, and PRA. This correlation raised a question on the role of the CYP11B1 enzyme in the biosynthesis of aldosterone. We suggest that both CYP11B1 and CYP11B2 can convert deoxycorticosterone to corticosterone via 11β-hydroxylase, and corticosterone is the precursor of aldosterone. Although CYP11B2 is the only enzyme that can convert corticosterone to aldosterone, CYP11B1 may still contribute to aldosterone overproduction via synthesizing enough precursor, i.e. corticosterone, as a substrate for the biological activity of CYP11B2 (Nakamura et al., 2014; Ono et al., 2014).

The Wnt/β-catenin signaling pathway plays a pivotal role in the regulation of proliferation, differentiation, and function of the normal adrenal cortex and adrenocortical tumorigenesis (El Wakil and Lalli, 2011; Clevers and Nusse, 2012). Aberrant cytoplasmic and nuclear accumulation of β-catenin reflects the constitutive activation of this pathway (Tissier et al., 2005). The prevalence of abnormal β-catenin staining in our study was 35.6%, similar to that reported by Tissier et al. (Tissier et al., 2005). Further, we observed that the abnormal staining group secreted more aldosterone and cortisol than the wild-type staining group. Our results showed that the tumor area and CYP11B2 H-score were significantly higher in the abnormal staining group than those in the wild-type group. Several factors such as McCarty H-score and tumor area are involved in the overall steroid hormone generation of adrenocortical adenoma as described above. General steroid hormone production of APA simultaneously increases as tumor area becomes larger. Besides, higher CYP11B2 H-score implies that constitutive β-catenin activation increases the aldosterone synthesis capacity of APA per unit area and per cell. Studies have revealed that β-catenin can modulate the membrane potential and intracellular ion homoeostasis in different cell types through interacting with ion channels and transporters (Lesage et al., 2004; Wilmes et al., 2012; Zhao et al., 2014). Hence, we hypothesized that an aberrant cytoplasmic accumulation of β-catenin may depolarize the membrane potential and increase the cytosolic Ca²⁺ concentration of APA cells, thereby enhancing the steroidogenic ability. However, this hypothesis needs deeper exploration. Therefore, both the higher CYP11B2 H-score and tumor area effect synergistically give rise to higher serum aldosterone concentration in the abnormal staining group, whereas the higher serum cortisol in the abnormal staining group could mainly result from the tumor area effect rather than the similar CYP11B1 H-score compared with the wild-type group. Nevertheless, the underlying mechanism of why aberrant activation of the Wnt/β-catenin signaling pathway selectively enhances CYP11B2 expression rather than CYP11B1 in patients with APA warrants further investigation.

Laparoscopic adrenalectomy is now considered the standard approach in the treatment of unilateral PA as it is minimally invasive and can normalize aldosterone secretion, decrease blood pressure, and reduce antihypertensive medication dosage (Funder et al., 2008; Stowasser and Gordon, 2016). However, normotension without antihypertensive drug usage cannot always be achieved after adrenalectomy for unilateral PA. Complete clinical success was achieved in 37% of patients in the PASO study (ranging from 17 to 62% between different centers) (Williams et al., 2017). In the current study, complete clinical success was attained in 53.1% patients, similar with the PASO study. Laparoscopic adrenalectomy can confer a decrease in SBP about 25–40 mmHg and a reduction in the number of antihypertensive medications prescribed about 1–2 drug classes averagely (Steichen et al., 2012; Williams et al., 2017). Clinical characteristics predicting hypertension cure include sex, age, family history of hypertension, hypertension duration, BMI, estimated glomerular filtration rate, evidence of arteriolosclerosis and so on (Steichen et al., 2012; Stowasser and Gordon, 2016; Williams et al., 2017). Using the multivariate logistic regression model, our results proposed that sex, age and family history of hypertension were independently associated with clinical outcome. Patients who were younger, female and without a family history of hypertension had a higher likelihood of being cured of hypertension after adrenalectomy. Although the adjusted CYP11B2 and CYP11B1 H-scores were not independent predictors associated with clinical success based on multivariate analysis in our study, our results demonstrated that the adjusted H-scores in the complete success group were significantly lower than those in the incomplete success group, indicating that patients with lower immunostaining scores might benefit more from surgery for unilateral PA. It is reported that hyperaldosteronism patients have an increased rate of cardiovascular events than matched essential hypertension patients, showing that aldosterone per se can cause target organ damage independent from its hypertensive effect (Mulatero et al., 2013). The pathogenetic role of aldosterone excess include chronic inflammation, fibrosis, endothelial dysfunction, glucose and lipid metabolism disorder (Kozakova et al., 2003; Fallo et al., 2007; Staermose et al., 2009; Tsuchiya et al., 2009). Through these machanisms, aldosterone oversecretion can result in irreversible vascular and renal damage. Several studies have demonstrated media-to-lumen ratio of small arteries and estimated glomerular filtration rate are independently associated with the outcome following adrenalectomy for unilateral PA (Rossi et al., 2008; Group et al., 2009). As mentioned above, the adjusted H-scores are positively correlated with serum aldosterone. We postulate that patients with higher scores may develop more deleterious vascular remodeling and more impaired renal function due to the persistent secondary hypertension and thereby decrease the likelihood of complete clinical success (Rossi et al., 2008; Group et al., 2009). The main limitaion of our study is that the sample size is not large enough, which maybe undermines the statistical validity of multivariate analysis. Larger studies are needed to identify the association between immunohistochemical features and clinical outcome after adrenalectomy for unilateral PA.

In conclusion, CYP11B2 immunohistochemical staining is a useful histopathological tool for precisely locating the cells responsible for aldosterone production, and assists in the differential diagnosis and subtyping of unilateral PA. In a small portion of patients with unilateral PA, increased autonomous aldosterone production is not caused by the CT-detectable adenomas but might be driven by subcapsular APCCs. Adjusted CYP11B2 and CYP11B1 H-scores are correlated with serum aldosterone and other endocrine parameters in patients with APA. The aberrant constitutive activation of β-catenin can provoke the proliferation activity and enhance the steroidogenic ability of tumor cells, which suggests that a Wnt pathway inhibitor may be a potential, actionable therapeutic option for the treatment of hyperaldosteronism. Being younger, female and without a family history of hypertension predicts a better outcome after adrenalectomy in primary aldosteronism patients. However, further studies are required to elucidate the prognostic significance of immunostaining on surgical outcome of unilateral PA.