β-catenin in adrenal zonation and disease

Abstract

The Wnt signaling pathway is a critical mediator of the development and maintenance of several tissues. The adrenal cortex is highly dependent upon Wnt/β-catenin signaling for proper zonation and endocrine function. Adrenocortical cells emerge in the peripheral capsule and subcapsular cortex of the gland as progenitor cells that centripetally differentiate into steroid hormone-producing cells of three functionally distinct concentric zones that respond robustly to various endocrine stimuli. Wnt/β-catenin signaling mediates adrenocortical progenitor cell fate and tissue renewal to maintain the gland throughout life. Aberrant Wnt/β-catenin signaling contributes to various adrenal disorders of steroid production and growth that range from hypofunction and hypoplasia to hyperfunction, hyperplasia, benign adrenocortical adenomas, and malignant adrenocortical carcinomas. Great strides have been made in defining the molecular underpinnings of adrenocortical homeostasis and disease, including the interplay between the capsule and cortex, critical components involved in maintaining the adrenocortical Wnt/β-catenin signaling gradient, and new targets in adrenal cancer. This review seeks to examine these and other recent advancements in understanding adrenocortical Wnt/β-catenin signaling and how this knowledge can inform therapeutic options for adrenal disease.

1. Introduction

Wnt signaling was first discovered in Drosophila in the 1970s with the field rapidly expanding with several landmark studies in mice and humans. Loss of the wingless gene, as its name suggests, was responsible for a wingless phenotype due to altered polarity and wing morphogenesis in Drosophila (Sharma and Chopra, 1976). Within a decade, its mouse homolog, called integration site 1 (Int1), was discovered in a murine virus that resulted in sporadic mammary gland tumorigenesis (Nusse and Varmus, 1982). Mouse int1 was found to be nearly identical to its human homolog, a mammary gland proto-oncogene deemed WNT1 (wingless/Int1) (van Ooyen and Nusse, 1984; Nusse et al., 1990). Wnt1 and its associated pathway became the focal point of a vast array of biological research, spanning from Drosophila development to human cancer, over the next five decades.

WNT1 was first to be discovered among the WNT ligands, a family of 19 paralogs that code for secreted proteins that bind to Frizzled (FZD) receptors and activate downstream intracellular signaling (Nusse et al., 1991). WNT ligands activate two primary intracellular signaling cascades defined by β-catenin-independent and β-catenin-dependent transcriptional activation (reviewed in Wiese et al., 2018 and Zhan et al., 2017). Wnt signaling independent of β-catenin, termed non-canonical, comprises both the planar cell polarity (Wnt/PCP) and calcium (Wnt/Ca²⁺) pathways. The canonical Wnt pathway, hereafter referred to as Wnt/β-catenin, involves activated β-catenin-induced target gene expression (Fig. 1). Wnt/β-catenin signaling is co-regulated by a group of proteins known as the destruction complex, containing GSK3β, APC, and AXIN2, among others. In the absence of WNT ligands, cytoplasmic β-catenin is targeted for degradation through phosphorylation by GSK3β of the destruction complex. Upon FZD activation by WNT ligands, components of the destruction complex are recruited to the cell membrane, stabilizing β-catenin and allowing it to translocate to the nucleus where it serves as a coactivator of TCF/LEF-mediated transcriptional activation. An additional level of Wnt/β-catenin signaling modulation is provided by ZNRF3, an E3 ubiquitin ligase that targets FZD receptors for internalization and degradation. Moreover, R-spondin (RSPO) ligands bind to LGR receptors, leading to ZNRF3 turnover and enhancing Wnt/β-catenin signaling activation through FZD receptor bioavailability.

Figure 1. Canonical Wnt signaling activates β-catenin-dependent transcriptional activation.

(Top) WNT ligands are palmitoylated by Porcupine (PORCN) in the endoplasmic reticulum (ER) and joined by Wntless in the Golgi apparatus for secretion from WNT-producing cells. R-spondin (RSPO) ligands are also produced and secreted alongside WNT ligands. (Bottom) WNT ligands bind to Frizzled (FZD) and LRP receptors, which are marked for degradation and turnover by ZNRF3. RSPO ligands bind to LGR to inhibit ZNRF3 and potentiate WNT-activated signaling. Activation of FZD/LRP receptors recruits components of the β-catenin destruction complex to the cell membrane. In turn, β-catenin is stabilized and translocates to the nucleus where is coactivates target gene expression with TCF/LEF transcription factors. Created with BioRender.com.

Numerous studies have implicated Wnt/β-catenin signaling as a critical regulator of organ development and homeostasis through specific regulation of tissue-specific stem/progenitor cell populations, including those in the adrenal cortex. Such a role in the adrenal was first supported by observations in patients with familial adenomatous polyposis (FAP) who harbored germline mutations in the APC gene (Smith et al. 2000; Marchesa et al. 1997). In addition to early onset colonic polyps, patients with APC mutations had an increased incidence of benign and malignant adrenocortical tumors (ACTs), suggesting that, in addition to contributing to carcinogenesis, β-catenin may be crucial for proper adrenal homeostasis (Gaujoux et al., 2010).

The adrenal glands are bilateral organs which lie atop the kidneys and produce steroid hormones essential for various physiological functions, including salt balance, glucose metabolism, and sexual maturity. The adrenal is composed of an outer mesenchymal capsule and underlying cortex comprised of histologically distinct concentric zones responsible for the synthesis of three classes of steroid hormones (Fig. 2). The peripheral subcapsular zona glomerulosa (zG) is marked by small, compact cell clusters responsible for producing aldosterone, a mineralocorticoid essential for maintaining salt and water balance (Arnold, 1866). Centripetal to the zG lies the zona fasciculata (zF), comprised of fascicules of larger cells that produce cortisol in primates (corticosterone in rodents). Primate but not rodent adrenals also have a zona reticularis (zR) between the zF and central medulla that produces sex steroids. The centrally-located medulla is a functionally and histologically distinct coalescence of catecholamine-producing cells derived from the neural crest (Le Douarin and Teillet, 1974; Furlan et al., 2017).

Figure 2. The adrenal cortex is maintained through centripetal differentiation.

(Left) The adrenal gland contains an outer mesenchymal capsule, middle steroidogenic cortex, and inner medulla. The cortex is composed of concentric zones that are histologically and functionally distinct. Below the capsule lies the zona glomerulosa (zG), a progenitor cell compartment organized into rosettes (gold outline) of undifferentiated SHH-expressing/CYP11B2-negative (green) and aldosterone-producing SHH-negative/CYP11B2-expressing (dark purple) cells. Cells of the zG differentiate and centripetally replenish corticosteroid-producing CYP11B1-expressing cells of the zona fasciculata (zF; light purple). Adrenocortical cells eventually die at the corticomedullary boundary. (Middle) Wnt/β-catenin signaling exists in a centripetal gradient in the adrenal cortex. The adrenal zG is marked by Wnt-high activity, indicated by red nuclei. As zG cells differentiate, β-catenin is inhibited by PKA signaling (purple nuclei), allowing for zG-to-zF transdifferentiation and corticosteroid production. Created with BioRender.com.

Multiple groundbreaking studies have provided key insights into the importance of Wnt/β-catenin signaling not only in tumorigenesis but also in the development and maintenance of the adult adrenal cortex where Wnt-responsive progenitor cells of the zG play a critical role in renewal of the cortex throughout life (Heikkilä et al., 2002; Kim et al., 2008; Berthon et al., 2010; Ingle and Higgins, 1938; King et al., 2009; Freedman et al., 2013). In this review, we will highlight exciting new studies that have added tremendous depth to our understanding of Wnt/β-catenin signaling in adrenocortical zonation, progenitor cell biology, and adrenal cancer. The roles of Wnt pathway components and antagonists, including RSPO3, ZNRF3, PKA, and β-catenin itself, in adrenocortical zonation, architecture, and steroidogenesis will be discussed. The role of Wnt/β-catenin signaling in adrenocortical development, homeostasis, and regeneration will also be reviewed. Emphases on the molecular interplay between the capsule and cortex, the newly discovered adrenocortical Wnt/β-catenin signaling gradient, and continued questions regarding zG cell-specific transcriptional programs and potential cell(s) of origin in adrenal disease will be made throughout. Finally, we will outline the great advancements made in understanding the contribution of Wnt/β-catenin signaling to adrenal tumorigenesis, especially adrenocortical carcinoma (ACC), with implications for new targeted therapeutic options for patients.

2. Adrenal cortex: zonation and renewal

The adrenal cortex contains concentric zones of steroidogenic cell populations that are histologically and functionally distinct (Figure 2). Steroidogenic cells of the adrenal cortex are marked by expression of steroidogenic factor 1 (Sf1/Nr5a1/Ad4bp), an orphan nuclear receptor that acts as a master regulator of steroidogenic enzyme gene expression and cell fate (Morohashi et al., 1992; Lala et al., 1992). SF1 is essential for gland development and function throughout life as evidenced by adrenal hypoplasia and hypofunction in the setting of mice and human patients with loss-of-function (LOF) mutations in SF1 (Morohashi et al., 1992; Ingraham et al., 1994; Sadovsky et al., 1995; Luo et al., 1994; Hammer et al., 1999; Beuschlein et al., 2002; Achermann et al., 1999). Immediately below the overlying capsule lies the zG, consisting of dense cell clusters historically referred to as glomeruli due to a similar appearance to the glomeruli of the kidney nephron (Zelander, 1957; Cater and Lever, 1954). Each glomerulus is a biological rosette, a round grouping of differentiated steroid-producing cells and undifferentiated non-steroid-producing progenitor cells that is surrounded by laminin β1 (Leng et al., 2020). Steroid-producing cells within zG rosettes are marked by aldosterone synthase (AS, or CYP11B2), the terminal enzyme in aldosterone production that is regulated by plasma angiotensin II (AngII), sodium (Na²⁺), potassium (K⁺), and calcium (Ca²⁺) in primates and rodents (Ogishima et al., 1992; Imai et al., 1992; Mitani et al., 1994; Kakiki et al., 1997; Ye et al., 2003; Clyne et al., 1997; Fakunding et al., 1979; Guagliardo et al., 2020). The zG makes up only a small portion of the cortex, which largely consists of zF. The zF is organized in long fascicules of larger Cyp11b1-expressing cells responsible for producing corticosteroids in response to ACTH as part of the hypothalamus-pituitary-adrenal (HPA) axis.

Foundational experiments in rats showed that enucleated adrenals with a remaining zG and capsule renew over time, providing evidence that the capsule and zG together are a crucial progenitor cell compartment responsible for complete cortical renewal over time (Greep and Deane, 1949). Lineage tracing studies extended these findings by showing that the underlying zF is replenished by zG cells, which differentiate from aldosterone-producing to corticosteroid-producing cells that are centripetally displaced inward toward the medulla to eventually die at the corticomedullary boundary (Wright et al., 1973; Zajicek et al., 1986; King et al., 2009; Freedman et al., 2013). When aldosterone-producing zG cells were permanently labeled with GFP utilizing a Cre recombinase lineage tracing system under the control of the Cyp11b2 promoter (AS-Cre; mTmG), zG descendants were observed within the cortex 5 weeks after birth and completely replenished the cortex by 12 weeks (Freedman et al., 2013). In addition to a steroid-producing lineage in the zG, a non-steroid-producing SF1-expressing progenitor cell population marked by SHH also resides in the zG. These self-renewing SHH-expressing progenitor cells serve to replenish steroid-producing cells of the cortex by differentiating into CYP11B2-expressing or CYP11B1-expressing cells of the zG or zF, respectively (King et al., 2009; Walczak et al., 2014). Interestingly, SHH ligands secreted by zG progenitor cells signal to the overlying capsule, where they bind to Smoothened (SMO) receptors, inhibiting Patched (PTCH) and activating intracellular GLI1-mediated gene transcription (Nuesslein-Volhard and Wieschaus, 1980; King et al., 2009). Finally, the adrenal capsular Gli1-expressing cell lineage also contributes (albeit at a lower frequency) to cortical cell replenishment over time, particularly during establishment of the adult cortex following encapsulation of the fetal gland and under various injury paradigms (King et al., 2009; Finco et al., 2018; Grabek et al., 2019). These data collectively define the capsular-cortical unit that makes up the adrenocortical progenitor cell niche.

The adrenocortical progenitor cell niche involving the capsule and zG is regulated by both paracrine and endocrine signals that govern the balance between progenitor cell self-renewal and differentiation into steroid-producing cells to consistently maintain the adrenal cortex over time. Wnt signaling is an essential paracrine pathway involved in zG maintenance and adrenocortical homeostasis. β-catenin, the main intracellular effector of canonical Wnt signaling, is highly expressed in the zG in mice, although recent evidence has shown that β-catenin activity exists in a centripetal zG-to-zF gradient (Kim, A.C. et al., 2008; Basham et al., 2019). β-catenin regulates both the proliferation and steroidogenic activity of zG cells to maintain the progenitor cell population while also activating Cyp11b2 expression and aldosterone production in cells that eventually replenish the zF (Kim et al., 2008; Berthon et al., 2010; Freedman et al., 2013; Pignatti et al., 2020). Wnt/β-catenin signaling activates the expression of downstream target genes Lef1 and Axin2 in the zG, the latter of which also marks cells whose descendants repopulate the inner cortex over time (Filali et al., 2002; Jho et al., 2002; Finco et al., 2018; Grabek et al., 2019). Furthermore, fluorescence-activated cell sorting (FACS) experiments on GFP-labeled Wnt-responsive zG cells identified a population of SHH-expressing/CYP11B2-negative progenitor cells (King et al., 2009; Walczak et al., 2014). Notably, developmental SF1-Cre-mediated ablation of Shh in mice resulted in adrenocortical hypoplasia that continued into adulthood (Ching and Vilain, 2009). Together, these studies demonstrate that the zG is a heterogeneous progenitor cell compartment that is maintained by active Wnt/β-catenin signaling. However, the underlying cell-specific β-catenin-dependent mechanisms that regulate zG cell fate, as well as zonal identity and morphology, are not yet fully understood.

3. Wnt/β-catenin signaling in adrenal zonation

Following the discovery of adrenal tumors in FAP patients harboring APC mutations, numerous studies focused on the role of β-catenin in adrenal biology. Mice harboring adrenal-specific GOF β-catenin presented with adrenocortical hyperplasia and tumorigenesis later in life, while reciprocal studies showed that loss of β-catenin in the adrenal resulted in adrenocortical hypoplasia and late-stage failure of the gland (Berthon et al., 2010; Heaton et al., 2012; Kim, A.C. et al., 2008). These seminal findings led to further implementation of mouse models to study the underlying role of Wnt/β-catenin signaling in normal adrenocortical development and homeostasis. The following sections will outline Wnt/β-catenin-regulated adrenocortical zonation and homeostasis in the adult gland. The oncogenic role of β-catenin and other Wnt pathway components will be discussed in more detail in subsequent sections.

3.1. Capsular signaling in adrenocortical β-catenin activation

Adrenocortical Wnt/β-catenin activity is highest in the subcapsular zG, which contains both steroid-producing CYP11B2-expressing/SHH-negative and non-steroid-producing CYP11B2-negative/SHH-expressing progenitor cells (Kim, A.C. et al., 2008; Freedman et al., 2013; Walczak et al., 2014). Importantly, constitutive β-catenin activity results in zG-restricted hyperplasia in mice, suggesting that β-catenin induces proliferation in a zone-specific manner (Berthon et al., 2010; Pignatti et al., 2020). In addition to active β-catenin, Wnt-responsive cells are detected by expression of Lef1, the main nuclear β-catenin binding partner in the adrenal cortex, and the classic Wnt/β-catenin target genes Axin2 and Apc, whose protein products are essential for negative regulation of β-catenin through the destruction complex (Filali et al., 2002; Ikeda et al., 1998; Amit et al., 2002; Lee et al., 2003; Xing et al., 2004; reviewed in Kimelman & Xu, 2006). WNT ligands, a family of 19 secreted glycoproteins, activate β-catenin by recruiting members of the destruction complex to the cell membrane (Nusse et al., 1991; reviewed in MacDonald et al., 2009). Of the WNT ligand family, Wnt4 was detected in the Wnt-responsive adrenal zG by in situ hybridization (ISH), suggesting that it at least in part regulates Wnt/β-catenin signaling in the cortex (Heikkilä et al., 2002). However, loss of Wnt4 in mice resulted only in decreased aldosterone production and thus a physiological but not a morphological defect at 12 weeks of age (Heikkilä et al., 2002; Vidal et al., 2016). These observations led to the prediction that other ligands expressed in the capsule and/or cortex may be regulating β-catenin activity in the adrenal zG (Walczak et al., 2014). Indeed, ISH studies have identified the expression of Wnt5a and Wnt11 in the cortical periphery and Wnt2b in the capsule, but their roles in Wnt/β-catenin signaling-dependent adrenocortical zonation have not yet been characterized (Lin et al., 2001; Pietila et al., 2016; Basham et al., 2019; Lako et al. 1998).

The involvement of capsule-derived WNT ligands in adrenocortical Wnt/β-catenin signaling is supported by studies on R-spondin (RSPO) proteins, a family of secreted ligands (RSPO1-4) that potentiate Wnt/β-catenin signaling in both human and mouse (Chen et al., 2002; Kamata et al., 2004; Kim, K.A. et al., 2008). In the absence of RSPO ligands, FZD receptors are marked for internalization and degradation by ZNRF3/RNF43 transmembrane ubiquitin ligases (Chen et al., 2013; reviewed in Lau et al., 2014). When present, RSPO ligands bind to one of three leucine-rich repeat-containing G-protein coupled receptors (LGR4/5/6), coupling them with and subsequently neutralizing ZNRF3/RNF43. Thus, RSPOs inhibit negative regulators of Wnt/β-catenin signaling, increasing Wnt receptor availability and ligand binding on the cell membrane (Fig. 1). The potentiation of Wnt/β-catenin signaling by RSPO ligands and the resultant increase of FZD receptor availability on the cell membrane through ZNRF3/RNF43 inhibition has proven to be essential for proper development and maintenance of several tissues. Recently, in vitro and in vivo studies have revealed that RSPO2/3, but not RSPO1/4, can potentiate Wnt/β-catenin signaling independent of LGR receptors through binding of heparin sulfate proteoglycans (HSPGs), a mechanism previously thought to act outside the scope of Wnt signaling (Lebensohn and Rohatgi 2018). Wnt/β-catenin signaling in the developing limb was unperturbed in mice lacking all three LGR receptors (Lgr4/5/6 triple KO), suggesting that RSPO2/3 but not RSPO1/4 ligands continue to mediate ZNRF3/RNF43 turnover even in the absence of their receptors (Szenker-Ravi et al. 2018). Although Lgr4 is expressed in the mouse adrenal cortex and would be predicted to elicit downstream RSPO-dependent turnover of ZNRF3/RNF43, this receptor-dependent activity of in the adrenal cortex has not yet been confirmed (Yi et al., 2013; Mazerbourg et al., 2004).

The necessity of Rspo3 in adrenocortical zonation and zG identity was anticipated by single molecule in situ hybridization (smISH) that revealed the expression of Rspo3 exclusively in the adrenal capsule (Vidal et al., 2016). Inducible capsule-restricted Gli1-CreERT2 and ubiquitous CAG-CreERT drivers were used to delete Rspo3 in adrenal capsule in mice, which exhibited decreased adrenocortical Wnt/β-catenin activity and zG loss. Interestingly, Wnt4 was downregulated in the adrenal cortices of Rspo3 KO mice, suggesting that a capsular WNT ligand is required for zG maintenance. Indeed, Sf1-Cre-driven Wnt4 loss recapitulated Cyp11b2 downregulation that has been previously shown but not the zG loss observed in Rspo3 KO mice (Heikkilä et al., 2002; Vidal et al., 2016). Together, these data suggest that adrenocortical β-catenin activity is primarily dependent upon capsular RSPO3 and WNT ligands to maintain both the structure and function of the adrenal zG—a critical progenitor cell compartment—and overall zonation of the adrenal cortex. The hypothesis that capsule-derived WNT ligands act alongside RSPO3 is supported by the secretion of RSPO and WNT ligands from underlying mesenchymal cells of the intestine that activate Wnt/β-catenin signaling and proliferation in neighboring stem cells to maintain the intestinal stem cell crypt (Valenta et al., 2016). Future studies to elucidate WNT ligands that concurrently activate adrenocortical Wnt/β-catenin signaling alongside RSPO3 to govern zG cell fate and identity would be fruitful.

3.2. Adrenocortical Wnt/β-catenin signaling gradient

Historically, adrenocortical Wnt/β-catenin activity was restricted to the zG as observed in reporter mice (Walczak et al., 2014). However, newer, more sensitive molecular techniques have revealed a centripetal Wnt/β-catenin signaling gradient in the zG and upper zF that is guided by ZNRF3 (Basham et al., 2019). Recent work through both the European Network for the Study of Adrenal Tumors (ENSAT) and The Cancer Genome Atlas (TCGA) identified ZNRF3 as a commonly deleted gene in ACC (detailed in 6.2; Assié et al., 2014; Zheng et al., 2016). ZNRF3 is an E3 ubiquitin ligase that labels FZD receptors for degradation to negatively regulate Wnt/β-catenin signaling (Hao et al., 2012; Koo et al., 2012). Deletion of Znrf3 in the mouse adrenal cortex resulted in remarkable adrenocortical hyperplasia of cells in the upper zF, demonstrating for the first time the importance of this Wnt/β-catenin antagonist in mediating proper adrenocortical homeostasis and organization (Basham et al., 2019). Together with observations in the context of Rspo3 loss by Vidal et al., Basham et al. provided corroborating evidence for an RSPO3/LGR4/ZNRF3 signaling mechanism from the adrenal capsule to the cortex that mediates Wnt/β-catenin signaling and cortical zonation (Vidal et al., 2016; Basham et al., 2019).

Intriguingly, although Znrf3 was expressed throughout the entire mouse adrenal cortex, the expansion of cells of the upper zF in Znrf3-null mice suggested that ZNRF3 is not primarily active in the zG, which remained normal (Basham et al., 2019). ZNRF3 activity thus contrasted that of β-catenin, which is most highly active in the zG and moderately active in the upper zF. Commensurate with the active β-catenin gradient was a Wnt4 expression gradient, suggesting that high WNT4 levels may reinforce Wnt/β-catenin activity in the zG. Furthermore, Znrf3-null mice exhibited remarkably larger cortices than their control littermates due to expansion of the Wnt-moderate cell population of the upper zF that failed to inhibit β-catenin activity, although corticosterone levels were normal (Basham et al., 2019). The dependence of the observed hyperplasia on Wnt/β-catenin activity was demonstrated by β-catenin heterozygosity or Porcn loss, each of which partially rescued the effects of Znrf3 loss. However, it is important to note that Porcn loss was restricted to the cortex, allowing for any potential intact WNT ligand processing and secretion from the capsule to blunt the rescue in Znrf3/Porcn dKO mice. These results support an antagonistic relationship between cortical ZNRF3 and capsular RSPO3 in mediating adrenocortical Wnt/β-catenin signaling: capsule-derived RSPO3 inhibits ZNRF3 to potentiate Wnt/β-catenin activity in zG cells immediately juxtaposed to the capsule, whereas ZNRF3 centripetally inhibits Wnt/β-catenin-dependent zF cell proliferation (Basham et al., 2019; Vidal et al., 2016). Furthermore, these data suggest a transient, proliferative Wnt-moderate cell population under the zG-zF boundary that, in light of the importance of cell proliferation in adrenal tumorigenesis and recurrent ZNRF3 deletions in ACC, has probable implications in ACC to be described later in this review.

3.3. Negative regulation of β-catenin in transdifferentiation

Proper transdifferentiation of zG cells to zF cells is essential for functional maintenance of the adrenal cortex throughout life (Pignatti et al., 2020). Multipotent cortical cell types of the zG, including both non-steroid-producing progenitor (SHH-secreting) cells and aldosterone-producing (CYP11B2-expressing) cells, expectedly undergo various transcriptional and epigenetic changes during their transition to corticosteroid-producing zF cells. Indeed, Wnt/β-catenin activity is diminished as cortical cells are centripetally displaced deep into the zF, implicating the necessity of Wnt/β-catenin inhibition during zG-to-zF transdifferentiation (Basham et al., 2019). In addition to regulation by ZNRF3 in the upper zF, Wnt/β-catenin signaling is also potently inhibited by cyclic AMP/protein kinase A (cAMP/PKA) signaling in the zF (Drelon et al., 2016a; Novoselova et al. 2018). Adrenocortical cells express melanocortin receptor 2 (MC2R), the receptor for adrenocorticotropin hormone (ACTH) that is necessary for zF development and stimulation of corticosteroid production (Chida et al. 2007; Gorrigan et al. 2011; Cooray et al. 2008). Differentiated zF cells also express melanocortin receptor accessory protein (MRAP), an MC2R binding partner that is essential for intracellular cAMP/PKA signaling induction and is frequently mutated in familial glucocorticoid deficiency (FGD) (Cooray et al. 2008; Gorrigan et al. 2011; Metherell et al. 2005). Corticotropin releasing hormone (CRH) from the hypothalamus induces ACTH secretion from the pituitary gland. ACTH binds to the MC2R/MRAP complex on the surface of zF cells and activates an intracellular Gs-protein response. Consequently, increased intracellular cAMP binds to the regulatory domain of PKA, freeing its catalytic domain and allowing it to translocate to the nucleus. PKA phosphorylates cAMP response element binding protein (CREB), a transcription factor that, together with MAPK-phosphorylated SF1, regulates the ACTH-dependent expression of multiple steroidogenic enzymes (Hammer et al., 1999; Winnay et al., 2006). Thus, upon ACTH binding in the adrenal cortex, zF cells are induced to proliferate and increase corticosteroid production. In turn, plasma corticosteroids inhibit further CRH and ACTH secretion from the hypothalamus and pituitary, respectively, in a typical endocrine negative feedback loop.

Loss of the PKA regulatory domain Prkar1a, leading to constitutive PKA signaling, results in downregulation of the Wnt/β-catenin pathway in ACTH-sensitive zF cells (Drelon et al., 2016a; Dumontet et al., 2018). In the nuclei of ACTH-bound cells, CREB inhibits expression of Wnt4. Consequently, levels of active β-catenin are decreased and Wnt/β-catenin target genes are downregulated, promoting a zF phenotype. This was further supported by ACTH treatment of Sf1-Cre; Wnt4^fl/fl mice that resulted in significantly reduced expression of Axin2 compared to untreated animals. Conversely, β-catenin activation inhibited MC2R expression and corticosterone production (Walczak et al., 2014). These studies were further supported by MRAP-deficient mice, the adrenals of which had a diminished zF and an enhanced zG marked by β-catenin and LEF1 (Novoselova et al. 2018). Together, these data suggest that while high Wnt/β-catenin activity promotes a zG-like phenotype by inhibiting zF-specific gene expression, it is then inhibited along a centripetal gradient as PKA signaling is activated and zG cells concomitantly differentiate into cells of the zF.

In addition to Wnt/β-catenin and other paracrine signaling mechanisms, transitioning zG cells have been hypothesized to undergo epigenetic modifications, which play an essential role in the differentiation of multi-/pluripotent stem/progenitor cells of various organs (Margueron and Reinberg, 2011). Supporting this hypothesis is the recent finding that enhancer of zeste homolog 2 (EZH2), a histone methyltransferase involved in polycomb repressive complex 2 (PRC2)-dependent transcription repression, was widely overexpressed in an ACC patient cohort, suggesting a critical regulatory role in normal adrenal homeostasis (Drelon et al., 2016b). Work by Mathieu et al. described an Ezh2 KO mouse model which displayed reduced adrenal size, corticosteroid insufficiency, and marked disruption of zF differentiation (Mathieu et al., 2018). Perhaps most striking was the reduction of PKA-dependent gene expression, suggesting that EZH2 is necessary for PKA activation in steroid-producing cells. These data further complexify Wnt/β-catenin-mediated transition of cellular identity between zG and zF. It was concluded that EZH2 regulates zG-to-zF transdifferentiation through PKA-dependent Wnt/β-catenin inhibition (Mathieu et al., 2018). In light of the common dysregulation of Wnt/β-catenin signaling and EZH2 in ACC, it is intriguing to speculate whether EZH2 facilitates the silencing of β-catenin-dependent transcriptional programming and/or the activity of ZNRF3 in cells transdifferentiating from zG to zF (Drelon et al., 2016b; Pignatti et al., 2020).

3.4. β-catenin-mediated adherens junctions and zG architecture

Target gene activation by nuclear β-catenin, as discussed so far, must be tightly regulated to maintain proper homeostasis and zonation of the adrenal cortex. However, the scope of Wnt/β-catenin signaling reaches far beyond gene regulation and plays a crucial role in cell adhesion and tissue organization. β-catenin is an essential component of cadherin complexes that act to maintain cell-cell adhesion and reorganize intracellular actin filaments upon tissue reorganization (Aberle et al., 1994; Jou et al., 1995; Vasioukhin et al., 2000). Stabilized cytoplasmic β-catenin interacts with E-cadherin in epithelial tissues, for example, and binds to its ortholog α-catenin, linking the cadherin-catenin complex to the actin cytoskeleton to induce morphological changes (Buckley et al., 2014). The focal point of these critical cell-cell interactions is the adherens junction (AJ), the rendezvous of extracellular cadherin domains and extracellular matrix (ECM) components of multiple cells organized to maintain the structural integrity of the tissue.

The adrenal zona glomerulosa was named such in light of its composition of dense cell clusters called glomeruli (Zelander, 1957; Cater and Lever, 1954). It was further shown that these glomeruli are heterogeneous bundles of SHH-expressing/CYP11B2-negative and CYP11B2-expressing/SHH-negative cells, whereas virtually all are marked by high β-catenin activity (Walczak et al. 2014; Basham et al., 2019). Recent work from Leng et al. has provided crucial insight into the molecular and cellular characteristics of these glomeruli and the role of β-catenin in mediating zG morphology. Glomeruli of the zG were found to be surrounded by laminin β1, a basement membrane marker that clearly outlines each glomerulus (Leng et al., 2020). Furthermore, β-catenin was observed in punctate clusters at the center of the glomerulus along with F-actin and N- and K-cadherin. These results suggested to the authors the co-identity of glomeruli with rosettes—round, multicellular structures that are prominent during embryogenesis of many organs but generally disappear by adulthood. Rosettes are characterized by cadherin and β-catenin interactions bundled in the center AJ that physically links all the clustered cells together. Evidence was provided for the postnatal development of rosettes in the adrenal zG by 6 weeks of age in mice.

Because β-catenin is a critical regulator of adrenocortical zonation, it was hypothesized that its activity was essential for proper AJ structure and, subsequently, zG morphology. Indeed, β-catenin GOF and LOF experiments resulted in an overall increase in zG thickness and rosette frequency and a thinner zG with rosette disorganization, respectively (Leng et al., 2020). Further experiments identified downstream target genes Fgfr2IIIb and Shroom3 as critical regulators of AJ maintenance in the adrenal zG (Leng et al. 2020). Additional independent studies characterized rosettes as crucial components of zG-specific AngII-induced Ca²⁺ signaling, a significant contributor to both aldosterone production and extracellular AJ biology (Guagliardo et al., 2020). These studies provide a greater understanding of the Wnt/β-catenin-dependent mechanisms that govern extracellular zG identity and morphology. Wnt/β-catenin activity is central to activating downstream target genes in response to both WNT ligands and cell-cell adhesion signaling components. How this dual role of β-catenin is itself regulated in various contexts is an important question to consider. First, while β-catenin-mediated AJs are important for proper zG maintenance, their role in zG-to-zF cell transdifferentiation remains unknown (Leng et al., 2020; Mathieu et al., 2018). Since constitutive β-catenin activity was recently shown to block this transdifferentiation, leading to an expanded zG, it may be predicted that β-catenin inhibition is necessary for cellular delamination and centripetal movement out of the zG (Pignatti et al. 2020; Berthon et al. 2010). Cellular markers of nascent delaminated zG cells would provide further insight into this transient cell population predicted to be at least in part regulated by ZNRF3, PKA, and EZH2 (Basham et al. 2019; Drelon et al., 2016a; Mathieu et al. 2018). Finally, these studies beg the question if the zG is a progenitor cell compartment at least in part due to its structural morphology (King et al., 2009; Freedman et al., 2013; Walczak et al., 2014). Rosette structures, particularly in the brain, have been proposed to be stem/progenitor cell niches that maintain the unique environment necessary for multipotent cells to undergo long-term self-renewal (Fuentealba et al., 2012). In addition to the various paracrine signaling components that have been shown to regulate SHH-expressing adrenocortical progenitor cells, how β-catenin-regulated rosette architecture in the zG additionally maintains the progenitor cell niche must be further explored.

3.5. Interplay between endocrine signaling and β-catenin activity

Cells of the adrenal cortex are regulated by many paracrine signaling components, such as WNT4, RSPO3, ZNRF3, and SHH, to maintain both structural and functional integrity of the tissue (Heikkilä et al., 2002; Vidal et al., 2016; Basham et al., 2019; King et al., 2009). In addition to these paracrine mechanisms, major endocrine pathways, namely the renin-angiotensin-aldosterone system (RAAS) and the HPA axis, must also be encumbered in any model of adrenocortical homeostasis. Endocrine signals have been shown not only to regulate steroid hormone production in the adrenal cortex through physiological negative feedback loops, but also to induce cortical cell proliferation and regulate the thickness of concentric cortical zones (Nishimoto et al., 2014; McEwan et al., 1999; Tian et al., 1995). Low intravascular volume induces the RAAS and AngII-mediated activation of differentiated cells of the adrenal zG to proliferate and produce aldosterone, which is activated by β-catenin and WNT4 (Heikkilä et al. 2002; Berthon et al. 2010; Pignatti et al. 2020). However, the delicate interplay between endocrine AngII and paracrine Wnt/β-catenin signaling in maintaining appropriate progenitor cell numbers and aldosterone production in the RAAS is not completely understood.

Another important consideration in Wnt/β-catenin-regulated adrenocortical zonation is the role of ACTH in promoting zG-to-zF transdifferentiation through PKA signaling activation. While the role of cAMP/PKA signaling in inhibiting β-catenin and promoting a zF phenotype in cells has already been discussed, whether ACTH only promotes corticosteroid production in existing zF cells or is also actively engaged in the transition of a zG cell to an ACTH-responsive zF cell is unknown (Drelon et al., 2016a; Mathieu et al., 2018). Moreover, it is unclear if the ACTH-dependent proliferation observed during adrenal regeneration occurs in residual zF cells or in transdifferentiating SHH-expressing or CYP11B2-expressing cells of the zG (Finco et al., 2018).

4. Wnt/β-catenin signaling in adrenal development

In addition to Wnt/β-catenin signaling being an important regulator of homeostasis of the adult adrenal cortex, it is also critical for the establishment and zonation of the nascent adult cortex. Adrenocortical development begins early in embryogenesis when Sf1 expression defines the adrenogonadal primordium (AGP), which delaminates from the coelomic epithelium at E9.0 (Hatano et al. 1996; Zubair et al., 2008). Cells of the AGP further differentiate, and by E10.5 in the mouse, cells expressing higher levels of Sf1 under the control of the fetal adrenal enhancer (FAdE) branch off from the developing gonads to become the fetal adrenal anlagen (Zubair et al. 2006). The medulla is then formed at E12.5 by neural crest cells invading and coalescing within the developing cortex, and the developing adrenal is fully encapsulated by mesenchymal cells of the intermediate mesoderm at E13.5. β-catenin expression has been observed by ISH around E12.5 in the peripheral cortex, but it is still unclear how it contributes to the establishment of the definitive progenitor cell niche (Kim, A.C., et al., 2008).

Several Wnt/β-catenin signaling components contribute significantly to the growth of the definitive cortex during prenatal development in mice. Formation of the adult cortex was drastically attenuated when β-catenin was conditionally deleted using an Sf1-Cre driver (Kim et al., 2008). Additionally, the loss of capsular Rspo3 at encapsulation (E13.5) initiated by Gli1-CreERT2 resulted in zG loss by E18.5 (Vidal et al., 2016). The use of ubiquitous CAG-CreERT-activated Rspo3 deletion at E11.5 prior to encapsulation exacerbated this phenotype by the same end point. These data suggest capsular signals begin activating cortical Wnt/β-catenin signaling almost immediately upon encapsulation, providing essential molecular cues for definitive zonation of the adult cortex. While studies on the loss of β-catenin activation in the developing adrenal cortex have resulted in zonation defects and progenitor cell loss, mouse models with overactivation of β-catenin, such as constitutively activated β-catenin (ΔCat) and Znrf3 KO mice, have not reported developmental defects, although collectively the data would support an early phenotype due to the necessity of β-catenin regulation for normal development (Berthon et al., 2010; Basham et al., 2019). Future studies are expected to shed light on the role of Wnt/β-catenin in the early establishment of the progenitor cell niche during specification of the adult adrenal cortex.

5. Wnt/β-catenin signaling in adrenal regeneration

The adrenal cortex regularly undergoes rapid molecular, cellular, and morphological changes in response to endocrine stimuli and paracrine signals between the capsule and cortex, including a Shh-Wnt signaling relay, that together regulate the cortical progenitor cell pool to establish and maintain adrenocortical zonation. Adrenal regeneration experiments have provided insights into homeostatic signaling pathways that both maintain adrenocortical zonation and mediate cell fate decisions. Using three different markers of cell populations in the zG, lineage tracing experiments have revealed that WNT-responsive (Axin2-CreERT2; RGFP) cells, undifferentiated Shh-expressing progenitors (Shh-CreERT2; mTmG), and differentiated Cyp11b2 (AS)-expressing (AS-Cre; mTmG) cells all contribute to cortical cell renewal throughout life (Walczak et al., 2014; Finco et al., 2018; King et al., 2009; Freedman et al., 2013). Finco et al. demonstrated the importance of paracrine Shh and Wnt/β-catenin signaling pathways between the capsule and cortex, respectively, in mediating adrenocortical regeneration. Using dexamethasone therapy to suppress ACTH levels and induce zF atrophy in mice, the concomitant robust increase in Shh expression and activation of Wnt/β-catenin signaling in the zG was observed during regeneration after dexamethasone treatment ceased (Finco et al., 2018). Expectedly, cortical recovery was stunted by both genetic ablation of β-catenin in WNT-responsive cortical cells (Axin2-CreERT2; β-catenin^fl/fl) and pharmacological inhibition of capsular Shh signaling activation. Moreover, constitutively activated Shh signaling in the capsule through genetic manipulation of the SMO receptor led to an increase in Wnt/β-catenin signaling and rate of recovery, suggesting that SHH activation of capsular GLI1-expressing cells accentuates cortical Wnt/β-catenin signaling, which in turn either directly or indirectly upregulates Shh expression in the zG.

The interplay between Wnt/β-catenin and Shh signaling pathways in adrenal regeneration further support the importance of the adrenal capsule in maintenance of cortical homeostasis. Interestingly, Rspo3 has been implicated as a direct GLI1 target gene, although this has not yet been verified in the adrenal capsule (Lewandowski et al., 2015). It is also unknown whether GLI1 activates WNT ligand expression in the capsule (Lako et al. 1998; Lin et al., 2001; Pietila et al., 2016; Basham et al., 2019). Furthermore, although Shh is downregulated in Rspo3 KO mice, it is unknown whether Shh is a bona fide β-catenin target gene (Vidal et al., 2016). Finally, while all zG cells are WNT-responsive, how Wnt/β-catenin signaling mediates the fate of both differentiated and undifferentiated cell populations and their contributions to homeostatic renewal of the adrenal cortex, especially under endocrine system manipulations as performed in regeneration experiments, will continue to be a challenging yet exciting avenue of research in adrenal biology. Future studies employing ChIP and single cell RNA sequencing (scRNA-seq), among others, can help define the underlying mechanisms of the adrenocortical Shh-Wnt relay that maintains zG identity and cortical zonation over time.

6. Wnt/β-catenin signaling in adrenal disease

Homeostatic molecular and cellular mechanisms of the adrenal cortex must be tightly regulated throughout life as dysregulated processes lead to a variety of disorders. Adrenal diseases range from hormonal hypo- and hyperfunction to hypo- and hyperplasia, together with benign and malignant tumors of the cortex and medulla. Pathophysiological excess of cortisol causing Cushing syndrome, affecting thousands of patients every year, can present as familial or sporadic disease in the setting of adrenocortical hyperplasia or adrenocortical tumors (ACTs), including benign adrenocortical adenoma (ACA) and carcinoma (ACC). Molecular mediators of ACTs have recently been characterized. ACA is a common benign tumor of the adrenal cortex whereas ACC is a rare, highly metastatic malignancy with poor overall survival (reviewed in Else et al., 2014). Many adult ACCs harbor somatic inactivating mutations in TP53 and overexpression of TERT, whereas the majority of pediatric ACC cases are driven by germline TP53 mutations associated with Li-Fraumeni syndrome (Tissier et al., 2005; Else et al., 2008; Svahn et al., 2018; Ribeiro et al., 2001; Wasserman et al., 2015; Pinto et al., 2015). Early tumor microarray data also defined IGF2 overexpression in more than 90% of ACCs but not ACAs (Giordano et al., 2003; Giordano et al., 2009). On the other hand, active nuclear and cytoplasmic β-catenin has been observed in both sporadic ACA and ACC, consistent with previous observations in FAP patients harboring inactivating alterations of APC (Heaton et al., 2012; Smith et al., 2000; Marchesa et al., 1997). Alterations in Wnt/β-catenin signaling components have since been widely observed in aldosterone-producing adenoma (APA) of the adrenal cortex, the leading cause of primary aldosteronism (PA), as well as ACC (Fagugli et al., 2011).

Pan-genomic data from the Cochin-COMETE, European Network for the Study of Adrenal Tumors (ENSAT), The Cancer Genome Atlas ACC (TCGA-ACC) projects have recently detailed the genetic, epigenetic, and chromosomal landscape in adult ACC patient cohorts and uncovered alterations in various Wnt signaling components, such as β-catenin, APC, and ZNRF3 (Gaujoux et al., 2011; Assié et al., 2014; Zheng et al., 2016). The following sections will outline the contributions of each Wnt/β-catenin signaling component to adrenal pathophysiology, including what is known in both mouse models and phenotypes presented in the human diseases discussed above, as well as new and ongoing studies to target this pathway therapeutically.

6.1. Apc

As mentioned previously, the first correlation between activated Wnt/β-catenin signaling and adrenal disease was described in patients with familial adenomatous polyposis (FAP), a hereditary disease mainly affecting the colon that is caused by germline APC mutations (Devic and Bussy, 1912; Smith et al., 2000; Gaujoux et al., 2010). APC is a tumor suppressor gene that is critical for inhibition of β-catenin by the destruction complex (Amit et al., 2002; Lee et al., 2003; Xing et al., 2004;). Inactivating alterations in both APC alleles, effectively nullifying negative regulation of β-catenin by phosphorylation and rendering it constitutively active, are observed in 70% of colorectal cancers (CRCs) (Rowan et al., 2000; Schell et al., 2016). In addition to presenting with several colon polyps early in life followed by a high incidence of colorectal cancer within 30 years of age, FAP patients are also at a higher risk for several other tumor types, including ACTs. This was the first evidence that β-catenin played a critical role in adrenal biology and that perturbations in the Wnt/β-catenin pathway might contribute to adrenal cancer. Later studies showed that Apc mutant (Apc^min) mice, which recapitulate the APC loss in human patients and the resulting development of colon polyps and cancer, developed adrenocortical hyperplasia and tumorigenesis later in life that was exacerbated by Igf2 overexpression (Heaton et al., 2012; Guillaud-Bataille et al., 2014).

Interestingly, it was not until much later that the Cochin-COMETE, ENSAT, and TCGA-ACC studies all revealed a low percentage of APC alterations in ACC patients (Gaujoux et al., 2010; Gaujoux et al., 2011; Assié et al., 2014; Zheng et al., 2016). Only about 1-3% of these patients harbored a deactivating mutation in APC while nearly 40% of ACCs overall had Wnt/β-catenin pathway alterations. Together with studies in Apc^min mice, these data suggested that while FAP patients harboring a germline APC mutation have increased incidence of ACC, alterations in APC itself are not highly prevalent in primary ACC cases. This highlights an interesting difference between adrenal and colon pathophysiology. Whereas APC loss is present in only a small fraction of ACC cases, it drives disease in nearly 70% of sporadic CRCs (Rowan et al., 2000; Schell et al., 2016). Although the same pathway is highly activated in both cancers, the tissue-specific tumorigenic potential of particular signaling component alterations has provided clues for developing unique targeted therapies through future studies.

6.2. β-catenin

β-catenin is the main effector of Wnt signaling and is either indirectly (as in the case of APC loss) or directly activated in many adrenal disease cases. Soon after the implementation of mouse models to understand the underlying effects of Wnt/β-catenin dysregulations in adrenal disease, constitutive Wnt/β-catenin signaling activation marked by diffuse cytoplasmic and nuclear β-catenin staining was found in 70% of an APA patient cohort (Kim et al., 2008; Berthon et al., 2010; Berthon et al., 2013). While no β-catenin mutations were found in any Wnt-active APAs, SFRP2, an extracellular Wnt/β-catenin signaling inhibitor, was downregulated in most samples (Berthon et al., 2013). Indeed, Sfrp2 KO mice exhibited constitutive adrenocortical Wnt/β-catenin signaling and upregulation of Cyp11b2. This work demonstrated that aberrant Wnt/β-catenin signaling activation associated with APA may be caused by the downregulation of Wnt/β-catenin signaling inhibitors, such as SFRP2, upstream of β-catenin.

In addition to β-catenin activation by downregulation or deletion of Wnt signaling inhibitors, β-catenin itself can harbor activating mutations. CTNNB1, the gene encoding β-catenin, is comprised of 16 exons. The N-terminal exon 3 of CTNNB1 contains serine and threonine (Ser/Thr) residues that are phosphorylated by the destruction complex (Liu et al., 2002; Gao et al., 2002). β-catenin regulation by phosphorylation is a critical event that must be tightly regulated in development (Amit et al., 2002). Frequent hotspot mutations at these Ser/Thr residues of exon 3 render β-catenin constitutively active and are present in a large percentage of endometrial, pancreatic, and hepatocellular cancers (Harada et al., 1999; Machin et al., 2002; reviewed in Kim and Jeong, 2019). These mutations were then discovered in a cohort of 39 ACA and ACC patients (Tissier et al., 2005). Among several SNPs in exon 3 of CTNNB1 were those located at serine 45 (S45), which were observed in 7 or 39 (18%) ACTs as well as the human ACC NCI-H295R cell line. These mutations were verified in three large independent ACC cohorts, present in approximately 16% of cases, and correlated with poor survival (Gaujoux et al., 2011; Assié et al., 2014; Zheng et al., 2016). Interestingly, while β-catenin mutations are present in both adrenal adenomas and carcinomas, each tumor type harbors unique β-catenin-dependent transcriptional programs (Giordano et al., 2003). Future efforts are expected to define additional associated passenger mutations that contribute to β-catenin-mediate pathology.

Based on these early observations in FAP and ACT patients, mouse models harboring LOF or GOF β-catenin alterations were implemented to study the underlying role of β-catenin in normal and pathophysiologic adrenocortical biology. Whereas loss of β-catenin resulted in zonal disruption and eventual adrenal failure, constitutively active β-catenin resulting from adrenocortical Akr1b7-Cre-driven floxed exon 3 of Ctnnb1 (ΔCat) caused adrenocortical hyperplasia and late-stage tumorigenesis, although a low percentage of mice developed ACC (Kim, A.C. et al., 2008; Berthon et al., 2010). More recent studies have provided deeper insight into the consequences of constitutive β-catenin activation in the adrenal cortex. AS^Cre/Cre mice were used to simultaneously delete exon 3 of Ctnnb1 and activate the RAAS, which exacerbated zG-specific hyperplasia likely due to disrupted homeostatic zG-to-zF transdifferentiation (Berthon et al., 2010; Pignatti et al., 2020). While enhanced proliferation of differentiated (aldosterone-producing) cells was observed in this context, it remains unclear whether undifferentiated progenitor cells and/or differentiated cells of the adrenal zG serve as cells of origin for β-catenin-dependent adrenocortical tumorigenesis.

Sporadic tumors of the adrenal cortex, like many other tissues, often harbor several driver and passenger mutations that are needed for advanced malignancy. While β-catenin mutations were present in about 16% of ACC patients in three separate cohorts, they widely overlapped with TP53 loss and overexpression of TERT and IGF2 (Gaujoux et al., 2011; Assié et al., 2014; Zheng et al., 2016). Importantly, adult and pediatric Li-Fraumeni patients harboring germline heterozygosity of TP53 have an increased incidence of adrenal tumors, similar to FAP patients with APC loss (Li and Fraumeni, 1969a; Li and Fraumeni, 1969b; Li et al., 1988; Nichols et al., 2001; Raymond et al., 2013). Furthermore, it had previously been hypothesized that secondary mutations were needed in ΔCat mice to advance to metastatic ACC (Berthon et al., 2010). Indeed, human patients in one cohort with either Wnt/β-catenin activation or TP53 loss had a higher tumor-free survival rate than those with alterations in both pathways, and p53 loss in mice enhanced the metastatic potential of ACC in ΔCat mice (Borges et al., 2020).

In vivo models of β-catenin-dependent adrenal disease, particularly ACC, have proven useful in recapitulating human adrenal diseases and determining their unique characteristics and underlying mechanisms. They have provided a foundation for new and improved preclinical models for β-catenin inhibition in ACC treatment. Drugs targeting either β-catenin itself or its many nuclear binding partners, such as CREB-binding protein (CBP), are now being widely studied in basic research and clinical trials for several cancers (Takemaru and Moon, 2000; Mullighan et al., 2011; Yu et al., 2017). PKF-115-584 and many other compounds targeting the interaction between β-catenin and TCF/LEF coactivators have also been developed (Lepourcelet et al., 2004; Park et al., 2005; Chen et al., 2009). Treatment of the β-catenin-driven NCI-H295R ACC cell line with PKF-115-184 significantly induced cell death, supporting β-catenin as a potential therapeutic target in ACC (Doghman et al., 2008). Direct β-catenin antagonist BC2059 has shown efficacy against acute myeloid leukemia in mice (Fiskus et al., 2015). PRI-724, which inhibits the β-catenin/CBP interaction necessary for target gene activation, is currently being assessed in various Phase I and II clinical trials (Kimura et al., 2017; reviewed in Lenz and Kahn, 2014). Further developments of PRI-724-associated compounds ICG-001 and C-82 have also been undertaken to treat fibrosis and endometrial and colorectal cancers (Emami et al., 2004; Tokunaga et al., 2017; Okazaki et al., 2019; Hirakawa et al., 2019). These studies ultimately hope to inhibit the downstream effects of constitutively active β-catenin by either blocking β-catenin from binding to target gene promoters or by inhibiting crucial target gene products to alleviate tumorigenesis in patients.

6.3. Znrf3

ZNRF3 is a critical negative regulator of Wnt/β-catenin signaling that was found to be deleted in about 20% of ACC patients (Assié et al., 2014; Zheng et al., 2016). The ENSAT and TCGA studies were the first to show the loss of both ZNRF3 alleles in ACC, leading to the development of Znrf3 KO mouse models to determine its underlying biological role in ACC. Indeed, Znrf3 loss in mice resulted in remarkable adrenal enlargement caused by excessive proliferation of Wnt-moderate cells of the upper zF (Basham et al., 2019). It is worthy to note that, while alterations in CTNNB1 and ZNRF3 make up 36% of the TCGA-ACC patient cohort, these two alterations are mutually exclusive. Furthermore, Rnf43 loss does not result in an adrenal phenotype in the mouse and so is not predicted to play a significant role in ACC (Basham et al., 2019).

The lack of overlap of CTNNB1 and ZNRF3 alterations is suspected as alterations in both may prove lethal for a tumorigenic cell. However, these results further highlight the importance of defining the cell of origin in ACC, which may vary in different mutational profiles. Supporting this is the result that Znrf3 loss in mice caused expansion of Wnt-moderate cells of the upper zF, whereas tumorigenesis in ΔCat mice likely originated in the zG (Basham et al., 2019; Berthon et al., 2010; Pignatti et al., 2020). Although these observations may seemingly answer the cell of origin question, the etiology of each alteration must continue to be studied in the small but ever-growing ACC patient dataset. Additionally, whereas patients harboring CTNNB1 mutations may benefit from downstream inhibitors, ZNRF3-deficient tumors can be treated upstream of the pathway as they are highly sensitive to WNT and RSPO ligands, potentially even being dependent on these ligands to grow, transform, and metastasize. Clinical trials in a variety of cancers implementing inhibitors of PORCN and Wntless, cytoplasmic proteins necessary for WNT ligand post-translational modifications and secretion, are currently ongoing. Additionally, neutralizing antibodies of FZD and LGR receptors may prove effective in inhibiting β-catenin activation by WNT and RSPO ligands, respectively.

6.4. WNT ligands

Wnt4 is the main canonical WNT ligand expressed in the adrenal cortex (Heikkilä et al., 2002). In humans, WNT4 loss results in sex reversion, kidneys, adrenal, and lung dysgenesis (SERKAL) syndrome (Mandel et al., 2008). SERKAL syndrome is an autosomal recessive condition resulting in female to male sex reversion as well as defects in adrenal growth and function. This was modeled in female Wnt4 KO mice in which sex reversion was observed (Heikkilä et al., 2002). In addition, a functional but not structural adrenocortical defect that resulted in decreased aldosterone levels at 12 weeks of age was also noted in Wnt4 KO mice. It may be expected that Wnt4 overexpression would result in adrenocortical hyperplasia, but this model has not yet been generated. It is important to note that, while WNT ligands are upregulated in many cancers, they are rarely mutated themselves. For example, in breast and GI cancers, various WNT ligands are upregulated and associated with poor survival (reviewed in Zhan et al., 2017). This holds true in ACC, in which WNT4 is highly upregulated as a Wnt target gene itself in Wnt-active tumors (Zheng et al., 2016). WNT4 may thus play a critical role in ACC, especially WNT ligand-dependent ZNRF3-null cases. Studies in mice harboring combined loss of Wnt4 and Znrf3 and other Wnt signaling components would provide deeper insight into the importance of WNT4 and perhaps other WNT ligands in activating Wnt/β-catenin signaling in ACC.

6.5. Frizzled/Lrp receptors

WNT ligands bind to FZD and LRP receptors to activate intracellular signaling (Adler, 1992; Cadigan et al., 1998; Pinson et al., 2000). Ten FZD (1-10) and two LRP (5 and 6) receptors are expressed in mice and humans, several of which influence an array of human diseases. LRP5/6 have been implicated in human familial exudative vitreoretinopathy (FEVR) and bone loss and in various bone and Wnt-related developmental defects in mice (Toomes et al., 2004; Gong et al., 2001; Kato et al., 2002; Pinson et al., 2000). Interestingly, one somatic LRP6 mutation in ACC was noted (Zheng et al., 2016). FZD4 mutations are similarly found in human FEVR (Robitaille et al., 2002). Furthermore, various phenotypes due to FZD loss have been observed in mice, affecting bone, kidney, B cell, and brain development (Albers et al., 2011; Heilman et al., 2013; Ye et al., 2011; Ranheim et al., 2005; Stuebner et al., 2010). Redundancy of many FZD receptors has also been reported, as the loss of two FZDs of a distinct subfamily results in similar phenotypes in mice (Yu et al., 2012). Some pairs, such as FZDs 3 and 6, function to elicit β-catenin-independent Wnt/PCP signaling (Wang et al., 2006). Conversely, FZDs 4 and 8 redundantly activate canonical Wnt/β-catenin signaling in the developing kidney (Ye et al., 2009). The importance of FZD receptors in the mouse adrenal has not yet been studied. However, it can be hypothesized that FZD loss would result in a dramatic phenotype in the adrenal cortex and thus be therapeutically targetable in WNT ligand-dependent ACTs in patients. FZD inhibition was first studied with the use of niclosamide, an anti-helminthic drug used for tapeworm infection, which blocked activation of Wnt/β-catenin signaling in vitro (Chen et al., 2009). Separate studies then led to the development of both a peptide and an antibody targeting FZD receptors (Nile et al., 2018; Gurney et al., 2012). The latter therapy, OMP-18R5, or vantictumab, targets FZDs 1, 2, 5, 7, and 8 (Gurney et al., 2012). OMP-18R5, along with the FZD8 decoy receptor OMP-54F28. Both molecules are currently in Phase I clinical trials to assess efficacy against primary and metastatic pancreatic, breast, liver, and ovarian cancers, among others (Fischer et al., 2017; Jimeno et al., 2017). A Phase Ib clinical trial for vantictumab treatment of metastatic pancreatic cancer was stopped early due to excessive bone toxicities, but these effects seem to be reversible and manageable (Fischer et al., 2017; Smith et al., 2013). FZD inhibition would provide a viable treatment option for the subset of ACC patients with tumors harboring alterations upstream of β-catenin, such as ZNRF3 loss. While the adverse effects of FZD inhibition are significant, studies on the efficacy of FZD and LRP inhibition in mouse models of adrenocortical diseases are nonetheless urgently needed.

6.6. Porcupine

Similar to FZD and LRP receptors, PORCN is essential for ligand-dependent Wnt/β-catenin signaling activation. PORCN is located on the endoplasmic reticulum where it palmitoylates WNT ligands for secretion, a process that is essential for mammalian embryonic development (Kadowaki et al., 1996; Cox et al., 2010; Biechele et al., 2011). Porcn loss in mice recapitulates the associated human disease focal dermal hypoplasia (FDH), an X-linked disorder that is lethal in males and causes various limb, skin, and other developmental defects in females (Barrott et al., 2011; Biechele et al., 2011; Grzeschik et al., 2007; Wang et al., 2007). PORCN is necessary for several Wnt-active cancers that are dependent on WNT ligand secretion. The role of PORCN in ACC was modeled in Znrf3; Porcn dKO mice in which the loss of adrenocortical Porcn partially rescued the hyperplastic phenotype cause by Znrf3 loss, suggesting that WNT ligands may play a significant role in inducing proliferation in ZNRF3-null ACC (Basham et al., 2019). Reversible PORCN inhibition by the compound IWP, a commercially available drug now widely used to deactivate Wnt/β-catenin signaling in the laboratory, was one of the first studies effectively targeting PORCN in cancer cells (Chen et al., 2009). Small molecule drugs C59 and ETC-159 were later developed and utilized in treating MMTV-WNT1-induced mammary tumors and colorectal cancer harboring an RSPO1 translocation, respectively (Proffitt et al., 2013; Madan et al., 2016). Several drugs, such as LGK974 (WNT974), are now in Phase I or II clinical trials for treating melanoma, colorectal, breast, GI, pancreatic, and head and neck cancers (Liu et al., 2013; reviewed in Jung and Park, 2020). The efficacy of targeting PORCN in these Wnt-active tumors provides support for its utility as a treatment for ACCs harboring deletion of ZNRF3 or overexpression of WNT ligands or receptors. It is therefore likely that PORCN inhibitors would pose therapeutically beneficial in such patients. Inhibition of PORCN would deplete the activating signal for Wnt-responsive cells to either manage disease burden, such as steroid hormone excess, or to effectively inhibit cancer cell proliferation, whether administered alone or in combination with other therapeutic strategies.

6.7. Rspo/Lgr

LGR receptors and their RSPO ligands are essential signaling components in the Wnt/β-catenin sphere. Embryonic loss of Rspo3, Lgr4, or Lgr5 results in pre- or postnatal lethality (Aoki et al., 2007; Mazerbourg et al., 2004; Morita et al., 2004). The Wnt-potentiating RSPO/LGR signaling module has also been found to be aberrantly activated in a variety of cancers. LGR5, a marker of intestinal stem cells, promotes CRC cell survival in mice and was found to be overexpressed in 64% of one patient cohort (Barker et al., 2007; Al-Kharusi et al., 2013; McClanahan et al., 2006). Upregulation of LGR5 has also been noted in several ovarian, breast, and hepatocellular cancers (McClanahan et al., 2006; Liu et al., 2018; Hou et al., 2018; Ko et al., 2019). Additionally, RSPO proteins, particularly RSPOs 2 and 3, have been implicated as potential drivers in CRC and liver and lung cancers (Starr et al., 2009; Conboy et al., 2019; Gong et al., 2016). Recurrent translocation events in CRC leading to overactive RSPO fusion proteins have also been widely observed (Seshagiri et al., 2013). However, conflicting studies have reported that overactive RSPO/LGR signaling may be therapeutically beneficial. Overexpression of Rspo1-3 correlated with increased survival in a lung patient cohort (Wu et al., 2019). Another study reported a tumor suppressive effect of the RSPO2/LGR5 module in CRC (Wu et al., 2014). These differences highlight the crucial need for more prospective and retrospective studies to determine the underlying effects of aberrant RSPO/LGR signaling in cancer patient cohorts. Surprisingly, no mutation, amplification, or translocation events involving Rspo1-4 or Lgr4/5/6 were reported in two ACC patient cohorts (Assié et al., 2014; Zheng et al., 2016). However, evidence suggests that inhibition of RSPO/LGR signaling could potentially dampen aberrant Wnt/β-catenin signaling activation in ACC, especially in ZNRF3-null tumors (Basham et al., 2019). Future basic and preclinical studies are needed to more fully understand the role of RSPO/LGR-dependent Wnt/β-catenin signaling potentiation in ACTs.

6.8. Other Wnt antagonists

In addition to APC and ZNRF3, various other Wnt signaling antagonists exist to help regulate the pathway. Axin2, a classic Wnt target gene that partakes in a negative feedback loop on the Wnt/β-catenin pathway, is enriched in the zG of the adrenal cortex. AXIN2 interacts with APC and GSK3β in the β-catenin destruction complex. Mutations in Axin1/2 affecting the destruction complex and β-catenin binding sites have been observed in a variety of cancers, including CRC (Liu et al., 2000; Webster et al., 2000). Adrenal tumors are mostly devoid of AXIN1/2 mutations, with deletions being observed in two adenomas and one ACC in two separate cohorts (Chapman et al., 2011; Guimier et al., 2013; Assié et al., 2014). These data suggest that AXIN2 does not play a significant role in ACC development and thus may not be a strategic therapeutic target.

As mentioned, PKA activation inhibits Wnt/β-catenin signaling and drives differentiation in the upper zF (Drelon et al., 2016a; Dumontet et al., 2018). Constitutively activating somatic mutations in PRKACA, encoding the catalytic subunit of PKA, have been observed in nearly half of adrenal adenoma patients and are associated with bilateral adrenal hyperplasia and resulting cortisol excess (Cushing’s syndrome) (Zilbermint and Stratakis, 2015; Weigand et al., 2017). Adrenocortical hyperplasia and Cushing’s syndrome was also observed in mice lacking the regulatory subunit of PKA encoded by Prkar1a, mutations of which result in constitutive PKA activation and occur frequently in primary pigmented nodular adrenocortical disease (PPNAD) (Sahut-Barnola et al., 2010; Dumontet et al., 2018). Additionally, alterations in PDE11A and GNAS, regulators of PKA signaling, have been observed in ACTH-independent macronodular adrenal hyperplasia (AIMAH) and McCune-Albright syndrome, respectively (Vezzosi et al., 2012; Weinstein et al., 1991). However, somatic mutations and deletions were observed in PRKAR1A but not PRKACA in human ACC (Assié et al., 2014; Zheng et al., 2016). Indeed, loss of Prkar1a induced adrenocortical tumorigenesis and corticosterone excess in mice, expected to be due to increased zF proliferation (Drelon et al., 2016a). Conversely, heterozygous loss of Prkaca resulted in β-catenin-induced tumorigenesis, which was partially abrogated by Wnt4 loss. These data suggest that PKA acts as a tumor suppressor in the adrenal cortex by inactivating Wnt/β-catenin signaling in the upper zF. Based on their roles, it can be proposed that PKA activation through recurrent PRKAR1A deletion is mutually exclusive with activating β-catenin mutations in ACC, leading to different clinical and prognostic outcomes. Aberrant PKA signaling has been described in PPNAD patients and mouse models, whether by alterations in PKA, GNAS, or various phosphodiesterases (Boikos et al., 2008; Mathieu et al., 2018; Pignatti et al., 2020; reviewed in Fragoso et al., 2015). Conversely, CTNNB1 mutations are associated with non-steroid-secreting tumors (Bonnet et al., 2011). However, further analyses must be done to understand the role of PKA signaling in human ACC.

In addition to intracellular regulators of Wnt/β-catenin signaling, secreted factors of the Dikkopf (DKK) family of proteins also negatively regulate the pathway at the cell surface by binding to LRP5/6 receptors (Glinka et al., 1998; Fedi et al., 1999). Dkk3 inactivation was observed in a Kcnk3 KO mouse model recapitulating human PA (El Wakil et al., 2012). Combined loss of Kcnk3 and Dkk3 resulted in further upregulation of Cyp11b2, suggesting that DKK3 acts to negatively regulate β-catenin-induced Cyp11b2 expression and aldosterone production. However, neither adrenocortical hyperplasia or tumorigenesis were observed, further supporting the need for alterations in crucial regulators and highlighting the complexity of Wnt/β-catenin signaling regulation in adrenocortical biology.

7. Concluding remarks

Wnt/β-catenin signaling governs the development and homeostasis of several tissues throughout the body. The remarkable history of Wnt signaling research has led us to a deeper understanding of organ maintenance and human diseases that occur due to aberrant pathway activation. The adrenal gland is one of many organs dependent upon Wnt/β-catenin signaling for lifelong renewal and function. Differentiated cells of the adrenal cortex rapidly produce crucial steroid hormones in response to various stressors. The peripheral zona glomerulosa (zG) contains aldosterone-producing cells and undifferentiated progenitor cells. Fate and function of both zG cell populations are guided in part by Wnt/β-catenin through communication with the overlying capsule. To maintain proper homeostatic renewal, this process is tightly regulated by antagonists like ZNRF3. Alterations of several Wnt/β-catenin signaling components drive aberrant aldosterone production and adrenocortical tumorigenesis. Recent historic advancements in adrenal research and vital developments of numerous Wnt/β-catenin inhibitors provide an exciting foundation for new preclinical models in treating Wnt-active diseases of the adrenal cortex.