Abstract
Objective
Adrenocortical adenomas are frequent in the general population and can be associated with autonomous cortisol excess, increasing morbidity and mortality. Altered cAMP/PKA signalling is common in sporadic cortisol-producing adenomas, typically due to somatic activating mutations in the catalytic subunit α of PKA (PRKACA) or the G-protein α subunit, Gαs (GNAS), which activate cAMP signalling. We previously identified a novel p.Lys58Gln GNAS somatic variant in a patient with a 5.3 cm adenoma and overt Cushing’s syndrome. This novel mutation was not charactersised before but provided enough evidence to warrant further investigation.
Design and methods
Using HEK293 cells depleted of GNAS, we established wild-type (WT) Gαs and Gαs-Lys58Gln stable cell lines and evaluated adrenocorticotropic hormone (ACTH) receptor signalling using a cAMP GloSensor assay, measured CREB transcription factor phosphorylation (pCREB) by AlphaLISA and assessed CRE luciferase reporter activity. Cell viability and apoptosis were also assessed over 5 days.
Results
The Gαs-Lys58Gln variant showed a significantly higher basal cAMP, pCREB and CRE luciferase reporter concentration and a greater response to ACTH (0–10 nM, P < 0.001) compared to WT Gαs. The variant had no effect on ligand potency. There was also significantly enhanced cell viability and apoptosis in cells with the Gαs-Lys58Gln variant.
Conclusions
In conclusion, our study demonstrated that the Gαs-Lys58Gln variant is associated with constitutive activation of GNAS signalling, similar to Arg201 mutations previously reported in adrenocortical adenomas, potentially representing a new pathogenic mechanism in a subset of patients with adrenal Cushing syndrome. This variant may also affect cell proliferation and requires further study.
Introduction
Adrenocortical adenomas (ACAs) are among the most frequent human neoplasias, with a prevalence of 2–3% in the general population. In 70% of the cases, they are endocrine inactive and usually incidentally discovered, while in the remainder of cases they are associated with autonomous cortisol or aldosterone secretion.
The genetic basis of autonomous cortisol secretion has been investigated by classical genetic approaches and next-generation sequencing. Among other findings, the cAMP/protein kinase A (PKA) pathway has been shown to play a central role in adrenocortical growth and steroidogenesis. Genetic alterations affecting the cAMP/PKA pathway, such as germline or somatic mutations in genes encoding the regulatory subunit 1 of PKA (PRKAR1A), the stimulatory G protein α subunit, Gαs (GNAS), and phosphodiesterases 11A and 8B (PDE11A and PDE8B) have been reported in cortisol-producing ACAs (CPAs) and bilateral micronodular adrenal hyperplasias (Bertherat et al. 2003, Fragoso et al. 2003, Horvath et al. 2006, Stratakis 2007, Rothenbuhler et al. 2012). Moreover, we and others have found somatic mutations in the gene encoding the catalytic subunit of PKA (PRKACA) in 35–70% of CPA associated with Cushing’s syndrome (CS) (Beuschlein et al. 2014, Cao et al. 2014, Di Dalmazi et al. 2014, Goh et al. 2014, Sato et al. 2014). These mutations translate into a constitutive activation of PKA by interfering with binding between its regulatory and catalytic subunits (Calebiro et al. 2014). Activating mutations in the gene encoding β-catenin (CTNNB1) have been reported in both ACAs and adrenocortical carcinomas (ACCs) with similar prevalence (10–30%) (Tissier et al. 2005, Tadjine et al. 2008, Assie et al. 2014) and are the most frequent lesions in non-cortisol-secreting tumours (Tissier et al. 2005).
In a previous large multicentre study, we defined the genetic landscape of sporadic unilateral ACAs by performing next-generation whole-exome sequencing (WES) in 99 ACAs, including 74 CPAs (Ronchi et al. 2016). Using this approach, we identified several somatic variants in genes of the cAMP/protein kinase A pathway, including three novel mutations in PRKACA (p.Trp197Arg, p.Glu32Val, and 731_745del) and a novel ‘probably damaging’ mutation in GNAS (p.Lys58Gln), associated with Cushing’s syndrome. The biological role of the three newly identified PRKACA mutations has been clarified in subsequent functional studies (Bathon et al. 2019, Walker et al. 2021, Weigand et al. 2021), but the role of the new Gαs mutation remains to be elucidated. The three-dimensional in silico analysis showed that lysine 58 is near the critical position Arg201, suggesting a functional significance for p.Lys58Gln substitution, similar to the known GNAS-activating mutations (Ronchi et al. 2016).
In the present study, we performed a functional in vitro characterisation of the GNAS variant Lys58Gln aiming to provide insights into its potential role in the pathogenesis of CPA associated with Cushing’s syndrome.
Materials and methods
Protein sequence alignment and prediction of pathogenicity
The GnomAD mutation database (v4.1.0, accessed June 14 2024) and Ensembl (human genome version GRCh38.p14, accessed June 14 2024) were searched for variants at amino acid position 58. Mutation tolerance was predicted using MutationTaster (GRCh37/Ensembl 69, accessed June 14 2024), SIFT (accessed June 14 2024) and PolyPhen2 (accessed June 14, 2024), while ensemble measures of pathogenicity were estimated using CADD+ (GRCh38, v1.4) and REVEL scores (GRCh38, v1.3). Multiple sequence alignment was assembled using Cobalt and ClustalW from reference sequences (Homo sapiens: NP_000507.1; Rattus norvegicus: NP_068617.6; Mus musculus: NP_034439.2; Canis lupus familiaris: XP_038288729.1; Xenopus tropicalis: XP_031750701.1; Danio rerio: XP_001335732.1).
Plasmids and cell culture
The SNAP-Gαs plasmid was used as a template to generate the SNAP-Gαs-Lys58Gln expression construct using site-directed mutagenesis (QuikChange Lightning kit, Agilent, UK) and oligonucleotides obtained from Merck (sequences listed in Supplementary Table 1 (see section on Supplementary materials given at the end of the article)). MC2R-Tango (a gift from Bryan Roth; Addgene plasmid #66428 (Kroeze et al. 2015)) and MRAP1_pcDNA6.2/EmGFP-Bsd (a gift from Roger Reeves; Addgene plasmid #176937 (Moyer et al. 2023)) plasmids were purchased from Addgene and used as templates to generate pEGFP-C1-MC2R and pmCherry-C1-MRAP1 expression constructs by standard cloning procedures and oligonucleotides obtained from Merck (sequences listed in Supplementary Table 1). All plasmids were sequence-verified by Source Bioscience (UK).
The generation of HEK293 cells depleted of GNAS by CRISPR-Cas9 has previously been described (Stallaert et al. 2017). These cells were used as a template to generate cells stably expressing either SNAP-Gαs or SNAP-Gαs-Lys58Gln. Cells were maintained in high-glucose DMEM (Sigma, UK) supplemented with 10% calf serum (Thermo Fisher Scientific, UK) and 500 μg/mL geneticin (G418-sulphate; Thermo Fisher Scientific) at 37°C and 5% CO2. To assess stable expression of SNAP-Gαs and SNAP-Gαs-Lys58Gln, RNA was extracted from each cell line using an RNeasy Mini Kit (Qiagen, UK), 1 μg RNA was used for reverse transcription using the QuantiTect Reverse Transcription Kit, and then RT-PCR was performed using a 2X Taq polymerase master mix (Promega, USA) and primers specified in Supplementary Table 1.
For functional studies, cells were plated at 3.0 × 105 cells/well in a 6-well plate and transfections performed with 2 μL/well of Lipofectamine 2000 (Invitrogen, Thermo Fisher Scientific).
Western blot analysis
To confirm overexpression of the Gαs protein in the stable cell lines, cells were lysed in NP40 lysis buffer (1 mM EDTA, 150 mM NaCl, 1% IGEPAL CA-360 in 50 mM Tris–HCl, pH 7.4, supplemented with 1X Roche Complete Protease Inhibitor Cocktail tablet). As a positive control, SNAP-Gαs was transiently transfected at 1 μg DNA per well in a 6-well plate and proteins collected and lysed 48 h later. All samples were prepared with Laemmli buffer (2x Laemmli Sample Buffer; Bio-Rad, UK) and separated by 10% SDS-PAGE gel electrophoresis before transfer to polyvinylidene difluoride membrane. Blots were blocked in 5% milk in TBS-T (Tris-buffered saline; 15 mM NaCl, 5 mM Tris, pH 7.5 with 0.01% Tween). Blots were probed with a mouse monoclonal anti-Gαs primary antibody (1:1,000 Gαs antibody (12) sc-135914, Santa Cruz, USA) and an anti-mouse HRP-conjugated secondary antibody (1:3,000; Bio-Rad) and detected using Clarity™ Western ECL Substrate (Bio-Rad) on a BioRad Chemidoc XRS+ system. The blot was subsequently stripped using the Restore™ Western Blot Stripping Buffer (Thermo Fisher Scientific) and re-probed with an anti-calnexin polyclonal primary antibody (AB2301, EMDmillipore, SLS, UK) and anti-rabbit HRP-conjugated secondary antibody (Cell Signalling Technologies, Netherlands).
ApoTox-Glo™ triplex assay
Cell viability and apoptosis were measured using an ApoTox-Glo triplex assay kit, according to the manufacturer’s instructions (Promega). Briefly, 5,000 cells were seeded per well in five 96-well plates on day 1 and transfected 6 h later with 33 ng of MC2R (Melanocortin 2 Receptor) and 17 ng of MRAP1 per well. On day 1 (before transfection) and then 24 h later until 96 h, the viability reagent was added directly to the wells and incubated for 30 min at 37°C. Live-cell fluorescence was measured at 405Ex/495-505 Em using a GloMax Discover Microplate Reader (Promega). After reading fluorescence, excess media was removed to leave 25 μL, which was mixed 1:1 with Caspase-Glo 3/7 Buffer and then incubated for 30 min. Luminescence was measured on a GloMax Discover Microplate Reader. All media was then removed from wells, cells were washed once in PBS, and then 30 μL Coomassie reagent (Thermo Fisher Scientific) was added. Following a 10-minute incubation, absorbance was read on a GloMax Discover Microplate Reader. Data were analysed by dividing the fluorescence or luminescence values by the total protein values, then normalising to time 0, set as 1. Each point plotted represents five independent replicates. Statistical analyses were performed by two-way ANOVA with Sidak’s multiple comparisons test.
cAMP signalling assay
Wild-type Gαs or Gαs-Lys58Gln stably expressing HEK293 cells were transfected with 1 μg of MC2R, 500 ng of MRAP1 and 200 ng of pGloSensor-22F cAMP biosensor (Promega) per well. Transfected cells were replated 48 h later into white 96-well plates in 100 μL/well of FluoroBrite DMEM (Thermo Fisher Scientific) supplemented with 10% (v/v) serum and 1% (v/v) L-glutamine (to make complete FluoroBrite media). Four hours after replating, media was replaced with complete FluoroBrite media supplemented with 2% (v/v) GloSensor™ substrate (Promega) and 100 μM 3-isobutyl-1-methylxanthine (IBMX; Merck Life Science UK, UK) and incubated for 2 h at 37°C. Basal luminescence was read for 5 min (GloMax® Discover Microplate Reader, Promega) before cells were stimulated with ACTH (Tocris, Bio-Techne, UK) diluted in HBSS and 1 μM IBMX, or Forskolin (Tocris, Bio-Techne) at a concentration of 10 μM as a positive control. Plates were read for an additional 30 min. The cAMP response was corrected against baseline (pre-stimulation) luminescence and dose–response AUC was fitted with a three-parameter log agonist curve fit in the GraphPad Prism 9.4.1.
AlphaLISA phosphorylation assays
Wild-type Gαs or Gαs-Lys58Gln stably expressing HEK293 cells were plated in 24-well plates and transiently transfected with 1 μg of MC2R and 500 ng MRAP1. After 48 h, cells were stimulated with ACTH diluted in HBSS for 30 min. Cells were lysed in AlphaLISA buffer supplied with the pCREB Ser133 kit (Catalog# ALSU-PCREB-A500, Revvity, UK) and pCREB and GAPDH (Catalog# ALSU-TGAPD-A500, Revvity, Oxford) assays performed according to the manufacturer’s instructions. AlphaLISA readings were made on a Pherastar FS (BMG Labtech, UK) plate reader in duplicates and values for phosphorylated proteins normalized to GAPDH values. Ratios of pCREB:GAPDH were expressed as a percentage of the wild-type Gαs Emax and dose–response curves fitted with a three-parameter log agonist curve fit in the GraphPad Prism 9.4.1. Statistical analyses were performed by two-way ANOVA with Bonferroni’s multiple comparisons test for curves and by unpaired t-test for pEC50.
Luciferase reporter assays
Wild-type Gαs or Gαs-Lys58Gln stably expressing HEK293 cells were plated in 24-well plates and transiently co-transfected with 1 μg MC2R and 500 ng MRAP1, 100 ng pGL4.10-CRE luciferase reporter plasmid (Promega) and 10 ng pRL (Renilla) control vector. After 24 h, cells were exposed to ACTH for 48 h in cell culture media. Cells were lysed in the manufacturer’s supplied buffer and luciferase assays performed according to the manufacturer’s instructions using the dual-luciferase reporter assay system (Promega), with readings made on a GloMax Discover plate reader. The firefly luciferase activity was normalised to luciferase activity (Firefly/Renilla ratio) and ratios expressed as a percentage of the wild-type Gαs Emax and dose–response curves fitted with a three-parameter log agonist curve fit in the GraphPad Prism 9.4.1. Statistical analyses were performed by two-way ANOVA with Sidak’s multiple comparisons test for curves and by unpaired t-test for pEC50.
Structural modelling
Protein structures of Gαs were retrieved from the RCSB Protein Data Bank (PDB) structure repository. The inactive Gαs structure (6AU6) bound to GDP (Hu & Shokat 2018) and the active Gαs structure (1AZS) bound to adenylyl cyclase and GTP (Tesmer et al. 1997) were visualised in PyMOL (The PyMOL Molecular Graphics System, Version 3.0, Schrödinger LLC, USA). In silico mutagenesis was performed on these models using the PyMOL mutagenesis tool to change lysine to glutamine at position 58, and polar interactions (within 3.0A) between amino acids of interest were predicted using the PyMOL.
Statistical analyses
Statistical tests used for each experiment are indicated in the legends of each figure and the number of independent biological replicates denoted by N. Data were plotted and statistical analyses performed in the GraphPad Prism 9. Comparisons of concentration–response curves used two-way ANOVA with Bonferroni’s or Sidak’s multiple comparisons test. pEC50 values were obtained from each independent repeat and presented as the mean ± SEM. Statistical analysis for pEC50 values was performed using an unpaired Student’s t-test. A P value of <0.05 was considered statistically significant.
Results
The Lys58Gln Gαs somatic variant is predicted pathogenic and disrupts interactions within the interdomain interface
The heterozygous p.Lys58Gln somatic variant (c.172A>C) was previously identified in the GNAS gene in a single patient with CPA associated with overt Cushing’s syndrome (reported as COSM5966204) (Ronchi et al. 2016) and this mutation is not present in the genome aggregation database (GnomAD). We used five pathogenicity prediction tools to assess this variant, including CADD and REVEL, which assess multiple predictive features including amino acid or nucleotide conservation and biochemical properties of amino acid substitutions (Ioannidis et al. 2016). All five tools predicted the variant to be pathogenic or not tolerated (Table 1). The Lys58 residue is highly conserved across vertebrates (Fig. 1A) with a genomic evolutionary rate profiling (GERP) score of 3.87 (a positive score represents high conservation) (Table 1). In addition, the Lys58 residue is conserved in representative members of each G protein subfamily (Gαq, Gαi1, and Gα12) (Fig. 1A). In combination, these findings suggest that the Lys58Gln somatic variant is likely pathogenic.
Pathogenicity prediction for the GNAS Lys58Gln variant.
| Nucleotide change | Amino acid change | CADD score* | PolyPhen-2** | SIFT # | Mutationtaster2 † | REVEL score ‡ |
|---|---|---|---|---|---|---|
| c.172A>C | Lys58Gln | 32.0 | Probably damaging | Not tolerated | Disease causing | 0.921 |
CADD (version GRCh38-v1.7) integrates multiple annotations (e.g. conservation, missense changes) into one metric and provides a measure of deleteriousness score, with a high score (>20) representing variants that are not stabilised by selection (i.e. more likely to be disease-causing) (Schubach et al. 2024).
PolyPhen-2 integrates sequence and structure-based predictive features to predict damaging mutations (probabilistic score above 0.85 considered probably damaging).
SIFT makes predictions based on a combination of homology and amino acid properties (Vaser et al. 2016).
MutationTaster2 makes predictions based on SNP frequency and ClinVar data (Schwarz et al. 2014).
REVEL integrates scores from thirteen predictors to produce a number between 0 and 1 that aims to distinguish between rare neutral mutations (low score) and pathogenic mutations (high score; >0.75) (Ioannidis et al. 2016).
Predicted effects of the Lys58Gln variant on the Gαs protein. (A) Multiple protein sequence alignment of Gαs orthologs (top) and G protein paralogs (bottom) showing that the Lys58 residue and surrounding residues are highly conserved across five orthologs and representative members of each G protein subfamily. The wild-type Lys58 and mutant Gln58 (mut) are shown in bold blue text. Conserved residues are shaded grey. Numbering shows amino acid number in Gαs. (B) Predicted three-dimensional structure of Gαs in the inactive conformation (PDB ID: 6AU6) showing the helical domain in light blue and the GTPase domain in green, with the bound GDP in black. The wild-type Lys58 residue is shown in dark blue and forms a contact with Leu197 (yellow) across the interdomain interface. The Arg201 residue that is frequently mutated in adrenocortical adenoma is also shown adjacent to the switch 1 region. (C) The Gln58 mutant disrupts the contact with Leu197. (D and E) Predicted three-dimensional structure of Gαs in the active conformation (PDB: 1AZS, pink) overlaid over the structure of inactive Gαs (green) showing the consequences of activation on the positioning of the amino acid residue at position 58 (blue/light blue: inactive, red: active) on both the (D) wild-type Lys58 residue and (E) Gln58 variant. The Arg201 residue that is frequently mutated in adrenocortical tumours is shown in orange.
Citation: Endocrine Oncology 5, 1; 10.1530/EO-25-0009
Most Gαs mutations occur in the Arg201 hotspot residue and disrupt the G protein-binding site. To determine whether the Lys58Gln variant may similarly affect Gαs activity, we performed analyses of existing Gαs structural models. We used two published crystal structures to predict the consequences of the Lys58Gln variant, Gαs in the inactive GDP-bound form (PDB ID: 6AU6) (Hu & Shokat 2018) and active Gαs bound to GTP and the catalytic domains of adenylyl cyclase (PDB ID: 1AZS) (Tesmer et al. 1997). The Lys58 residue is located in the GTPase domain of Gαs proximal to the switch 1 region of the GTP/GDP-binding pocket but does not directly participate in Gαs binding to GDP (Fig. 1B). Instead, the side chain of the Lys58 residue projects into the space between the GTPase and helical domains of Gαs (the interdomain interface) and forms a contact with Leu197 within the helical domain, which may increase stability of the GDP-bound state (Fig. 1B). Mutation to the adrenocortical adenoma-associated Gln58 residue abolishes this interdomain contact and may disrupt the stability of the GDP-bound state and increase the stability of the active state (Fig. 1C). The same intramolecular interaction is predicted in both the active and inactive structures and is abolished in both by mutation to glutamine (Fig. 1D and E). We predict that the Lys58Gln mutation may have consequences for the stability of the GDP-binding pocket that could result in a change in Gαs activity.
The Gαs Lys58Gln mutation enhances ACTH receptor constitutive activity
To assess the functional effect of the Gαs-Lys58Gln variant, we first generated cell lines stably overexpressing either the wild-type or Lys58Gln variant Gαs protein. We used the previously described HEK293 cell line depleted of Gαs/l (Stallaert et al. 2017) to establish cells stably overexpressing either SNAP-Gαs or SNAP-Gαs-Lys58Gln plasmids in the absence of endogenous Gαs protein. Overexpression of either the wild-type or Lys58Gln variant Gαs protein was confirmed in the cell lines by RT-PCR and western blot analysis (Fig. 2A and B).
Establishment of cell lines stably expressing Gαs wild-type or the Lys58Gln variant. (A) RT-PCR showing expression of SNAP-tagged GNAS in the wild-type and Lys58Gln stable cell lines. The parental HEK Gαs/l knockout cells were used as a negative control. GAPDH was used as a loading control. (B) Western blot showing stable overexpression of the SNAP-Gαs plasmids in the WT and Lys58Gln cell lines. The parental HEK Gαs/l knockout cells were used as a negative control. HEK293 cells were used to illustrate absence of the SNAP-Gαs and endogenous expression of Gαs/l. HEK293 transiently transfected with the SNAP-Gαs were used to illustrate the presence of the transfected SNAP-Gαs and endogenous Gαs/l. Calnexin was used as a loading control.
Citation: Endocrine Oncology 5, 1; 10.1530/EO-25-0009
Activating mutations in Gαs enhance constitutive signalling by the MC2R in adrenal cells. To assess the functional effects of the Gαs-Lys58Gln mutation on ACTH-mediated signalling, the Gαs-WT and Gαs-Lys58Gln stable cell lines were transfected with pEGFP-C1-MC2R and pmCherry-C1-MRAP1 (which is required for MC2R plasma membrane expression (Noon et al. 2002)) and real-time generation of cAMP measured by GloSensor assays. ACTH stimulation (10−5 to 10−12 M) elicited robust luminescence responses in both WT Gαs and Gαs-Lys58Gln cell lines (Fig. 3A). The Gαs-Lys58Gln variant had a significantly higher basal cAMP response and had enhanced activity at ACTH concentrations from 10−11 to 10−8 M when compared to Gαs-WT cells. The WT Gαs cells had a significantly higher maximal response to ACTH stimulation (Fig. 3B); however, there was no statistical difference in the potency of the ACTH receptor response between the WT Gαs and Gαs-Lys58Gln cell lines (Fig. 3C).
The Gαs-Lys58Gln variant enhances constitutive activity of the ACTH receptor. (A) ACTH-mediated cAMP responses measured by GloSensor in Gαs wild-type and Gαs-Lys58Gln variant cell lines transiently expressing MC2R and MRAP1. (B) Area under the curve was used to generate cAMP concentration–response curves. (C) Potency of ACTH in cells expressing either WT or Lys58Gln Gαs transfected with the ACTH receptor. n = 5 biological replicates. (D) ACTH-mediated phospho-CREB (pCREB) responses measured by AlphaLISA in Gαs wild-type and Gαs-Lys58Gln variant cell lines transiently expressing MC2R and MRAP1, with (E) pEC50. n = 4 biological replicates. (F) ACTH-mediated CRE luciferase activity in Gαs wild-type and Gαs-Lys58Gln variant cell lines transiently expressing MC2R and MRAP1, with (G) pEC50. n = 4 biological replicates. Statistical analyses were performed by two-way ANOVA with Bonferroni’s multiple-comparisons test in B and D, two-way ANOVA with Sidak’s multiple-comparisons test in F and unpaired t-test in C, E, G. ****P < 0.0001, ***P < 0.001, ns = not significant.
Citation: Endocrine Oncology 5, 1; 10.1530/EO-25-0009
Following cAMP activation, a signalling cascade is initiated that involves PKA, phosphorylation of the transcription factor cAMP response element-binding protein (pCREB) at the Ser-133 residue, and modulation of the transcription of genes that contain cAMP response elements (CRE). To determine whether the Gαs-Lys58Gln variant also affects the generation of pCREB, AlphaLISA phosphorylation assays were performed in cells transfected with pEGFP-C1-MC2R and pmCherry-C1-MRAP1. ACTH stimulation significantly enhanced pCREB generation in both Gαs-WT and Gαs-Lys58Gln cell lines in a concentration-dependent manner (Fig. 3D). The Gαs-Lys58Gln variant had a significantly higher basal pCREB response and had enhanced activity at ACTH concentrations from 10−11 to 10−8 M when compared to Gαs-WT cells. The maximal response and pEC50 values were not significantly different (Fig. 3D and E). To determine whether this also affected gene transcription downstream of cAMP-CREB, luciferase reporter assays were performed with a plasmid containing a CRE-response element upstream of luciferase in cells transiently transfected with pEGFP-C1-MC2R and pmCherry-C1-MRAP1. ACTH stimulation again enhanced CRE luciferase activity in a concentration-dependent manner (Fig. 3F) in both Gαs-WT and Gαs-Lys58Gln cell lines. The Gαs-Lys58Gln variant had a significantly higher basal response and enhanced activity at ACTH concentrations from 10−11 to 10−7 M when compared to Gαs-WT cells. The maximal response and pEC50 values were not significantly different (Fig. 3F and G). Thus, the Lys58Gln variant enhances Gαs constitutive activity, similarly to other CPA-associated Gαs mutations.
The Gαs-Lys58Gln variant enhances cell growth and apoptosis
To characterise the effect of the Gαs Lys58Gln mutant on cell growth and cytotoxicity, assays were performed in each stable cell line with transient expression of MC2R-GFP and MRAP1-mCherry over 96 h. Assays were performed in the absence of ACTH to determine the effects of constitutive activity on cell growth and apoptosis. Cell growth was slow and similar over the first 48 h, then increased by 72 h. Cell lines expressing the Gαs-Lys58Gln variant had significantly higher cell viability than their counterparts that stably express WT Gαs by 96 h (Fig. 4A). Similarly, apoptosis rates were low over the first 48 h, then increased by 72 h. Gαs-Lys58Gln stable lines had a significantly higher level of apoptosis by 96 h (Fig. 4B).
The Lys58Gln Gαs variant enhances cell growth and apoptosis. (A) Cell viability in Gαs wild-type and Gαs-Lys58Gln variant cell lines transiently expressing MC2R and MRAP1 measured over 96 h. Data were expressed corrected to the total protein, then normalized to hour 0. (B) Apoptosis measured over 96 h was expressed corrected to the total protein, then normalised to hour 0. Statistical analyses were performed by unpaired t-test in n = 5 replicates. **P < 0.01, *P < 0.05.
Citation: Endocrine Oncology 5, 1; 10.1530/EO-25-0009
The Gαs-Lys58Gln variant in the clinical context
The probably damaging variant p.Lys58Gln was initially observed in the tumour tissue of a 44-years-old female patient (Ronchi et al. 2016) with pre-operative overt Cushing’s syndrome, as confirmed by the presence of signs and symptoms of cortisol excess, cortisol levels after dexamethasone suppression test of 21 μg/dL (579.3 nmol/L) with elevated salivary cortisol, suppressed ACTH and low DHEAS levels.
The tumour was a histologically confirmed left adrenocortical adenoma with a large size of 5.3 cm. At WES, the tumour carried overall a total of nine somatic variants, including ARMC9 (armadillo repeat-containing 9), CAPSN2 (calpain-2), ETV7 (ETS variant transcription factor 7), LCN1 (lipocain-1), KIDIN220 (ankyrin repeat-rich membrane-spanning protein), ZNFR354C, and most importantly a somatic mutation in CTNNB1 gene (p.Ser45Cys) (Ronchi et al. 2016).
Discussion
The cAMP/PKA pathway plays a key role in adrenocortical growth and steroidogenesis. Genetic alterations affecting the cAMP/PKA pathway have been reported in sporadic CPAs, but over one-third of cases remain unexplained (Ronchi 2019). In the present study, we performed a functional characterisation of a novel GNAS variant Lys58Gln previously observed in a case of CPA associated with overt Cushing syndrome (Ronchi et al. 2016). We provide insights into its potential role in the pathogenesis of autonomous cortisol secretion.
First, five prediction tools indicated the variant to be ‘pathogenic’ or ‘not tolerated’ and the Lys58 residue is highly conserved across vertebrates, suggesting that the Lys58Gln somatic variant is likely pathogenic. In addition, the Lys58 residue and surrounding amino acids are conserved in other G protein families, indicating it likely has an important functional role in G proteins. Three-dimensional modelling indicated that the Lys58 residue forms part of the interface between the helical and GTPase domains of the G protein. Residues within this interdomain interface are important for stabilising the GDP-bound state and for GDP/GTP exchange upon interaction with activated receptor (Van Eps et al. 2011). Previous studies of the Gαi1 structure showed that weakening of the interdomain interface promotes activation, while alanine mutagenesis of the Lys51 residue of Gαi1 (equivalent to Lys58 of Gαs) destabilised the GDP-bound state and increased the relative stability of the activated Gαi1 complex by 15–20% (Sun et al. 2015). Our in silico structural analyses showed that the Lys58 residue forms an interdomain contact with Leu197 within the helical domain. The Gln58 mutant identified in an adrenocortical tumour lost this contact with Leu197 (Fig. 1). Germline mutations that disrupt contacts across the interdomain interface have previously been described to cause human disease including Gα11 mutations that enhance receptor activation and cause hypocalcaemia in humans (Gorvin et al. 2020) and mice (Gorvin et al. 2017), while inactivating mutations in Gα11 cause hypercalcaemia (Gorvin et al. 2016) and in Gαs cause obesity (Mendes de Oliveira et al. 2021). Thus, we predict that the Gαs-Lys58Gln mutation decreases the stability of the GDP-bound state and increases the stability of the active state by disrupting interdomain interactions analogous to alanine mutagenesis of Lys51 in Gαi1 (Sun et al. 2015).
Mutations in GNAS occur in a wide variety of tumours (including in the pituitary, thyroid, pancreas, colon and parathyroid) and the majority affect two hotspot residues, Arg201 and Gln227 (O'Hayre et al. 2013). Mutations of the Arg201 residue are particularly common in CPAs and adrenal hyperplasia (Fragoso et al. 2003, Ronchi et al. 2016). Early studies showed that mutations in the Arg201 and Gln227 residue enhance the constitutive activity of Gαs, resulting in elevations in cAMP and its downstream signalling pathways (Graziano & Gilman 1989, Landis et al. 1989). Such constitutive activity may arise from inhibition of the intrinsic GTPase activity of Gαs by Arg201 mutations, resulting in longer association with GTP and a prolonged active state (Landis et al. 1989), or Arg201 mutations may instead bypass the need for GTP binding by directly activating GDP-bound Gαs to G protein activity (Hu & Shokat 2018). Further support for the latter mechanism has been provided by saturation transfer difference NMR (STD-NMR) spectroscopy, which showed that the Gαs-Arg201Cys mutant enhanced stability and increased the hydrogen bond network for GDP-bound-Gαs (Anazia et al. 2024). Similar to mutations in the Arg201 and Gln227 residues, our studies of the Lys58Gln Gαs variant in HEK293 cells showed enhanced constitutive cAMP, pCREB and CRE luciferase reporter production, consistent with a gain-of-function. The cAMP, pCREB and CRE luciferase activity were higher at most ACTH concentrations, and similar enhanced activity at multiple ligand concentrations has been observed for other adrenal tumour-associated mutations within the cAMP-PKA signalling pathway (Almeida et al. 2012). However, we observed no difference in the EC50, indicating that MC2R is still able to respond to elevations in ACTH. The Lys58Gln Gαs variant reduced the maximal response of MC2R to the highest concentration of ACTH in cAMP assays. This may be due to desensitisation of the receptor or depletion of signalling components due to prolonged constitutive activation. However, we did not detect this change in maximal response in other assays.
Expression of the Gαs-Lys58Gln variant in HEK293 cells resulted in enhanced cell viability in the absence of ACTH (Fig. 4). Previous studies of the common constitutively active Arg201 and Gln227 Gαs mutations in various tissues have also shown increased viability and/or proliferation. The Gln227Leu mutation enhances cell proliferation in thyroid cells (Muca & Vallar 1994); mice with conditional expression of the Arg201Cys mutation in intestinal cells have enhanced proliferation and intestinal tumourigenesis (Wilson et al. 2010), while the Arg201Cys mutation is associated with cellular hyperplasia in human adrenal adenoma (Weinstein et al. 1991). Although enhanced expression of MAPK signalling pathways has been shown to contribute to this enhanced cell growth in some cell types (Wilson et al. 2010), increased stimulation of the cAMP pathway has been shown in other cells (Muca & Vallar 1994), consistent with the known regulatory role of cAMP in cell proliferation and adrenal hyperplasia (Goh et al. 2014, Calebiro et al. 2015). Moreover, enhanced proliferation is associated with Gαs-mediated constitutive activation of cAMP in the absence of hormone stimulation in thyroid cells, consistent with the increased cell viability that we observed for the Gαs-Lys58Gln variant. This indicates that it is likely that the Gαs-Lys58Gln variant contributes to adrenal cell tumourigenesis, which may be further enhanced by co-expression of the activating β-catenin (CTNNB1) mutation identified in this patient.
We also showed that cell lines expressing the Gαs-Lys58Gln variant had a higher level of apoptosis than their counterparts that express WT Gαs. Although this finding of increased cell viability and apoptosis may seem contradictory, cAMP is known to stimulate both pathways, and there is evidence that growth inhibitory and proliferative signals coexist in some cells (Stork & Schmitt 2002, Murray & Insel 2013). The role of apoptosis in adrenal tumours is incompletely defined, and some studies suggest that there may be a correlation between apoptosis and poor prognosis in adrenal tumours. In a study of childhood adrenocortical tumours, a positive association was identified between CASP9 expression (which cleaves and activates the effector caspases 3 and 7) and a higher chance of unfavourable events (defined as relapse or death) in ACT (Lorea et al. 2012). Similarly, in a study of ACC, there was an enrichment of genes involved in the caspase cascade (enhanced CASP2 and CASP9 expression) in patients with poor outcomes compared to those with more favourable outcomes (de Reynies et al. 2009). Moreover, enhanced CASP9 expression was associated with metastatic potential in non-small-cell lung cancer. However, the childhood adrenocortical tumour study also identified lower expression of CASP3 with unfavourable events, and apoptosis likely needs to be explored in larger cohorts of patients.
From a clinical perspective, the p.Lys58Gln variant was initially observed in the adrenal tumour of a 44-years-old female patient with pre-operative overt Cushing’s syndrome and a large adrenocortical adenoma. The tumour carried overall a total of nine somatic variants, including the p.Lys58Gln GNAS variant and a p.Ser45Cys CTNNB1 variant. Of note, the concomitant presence of GNAS and CTNNB1 mutations is unusual for a CPA (see Fig. 3A) (Ronchi et al. 2016), suggesting a potentially additive effect of Gαs and β-catenin alterations on adrenocortical cell proliferation. This, however, remains to be demonstrated.
In conclusion, our study shows that the Gαs-Lys58Gln variant is associated with constitutive activation of GNAS signalling, similar to Arg201 mutations, potentially representing a new pathogenic mechanism in a subset of patients with adrenal Cushing’s syndrome. This variant may also affect cell proliferation and requires further study.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EO-25-0009.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.
Funding
An Academy of Medical Sciences Springboard Award supported by the British Heart Foundation, Diabetes UK, the Global Challenges Research Fund, the Government Department of Business, Energy and Industrial Strategy and the Wellcome Trust. Reference: SBF004|1034 (CMG). A Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society. Grant number 224155/Z/21/Z (CMG). A Wellcome Trust Senior Research Fellowship. Grant number 212313/Z/18/Z (DC). CR is the recipient of a research grant from HRA Pharma Rare Disease, which is not related to this study.
Author contribution statement
Conceptualisation was done by CMG and CLR. Methodology was given by AJ, RW, DC, CMG and CLR. Investigation was done by AJ, RW, AP, OR and CMG. Writing of the original draft was done by AJ, RW, CMG and CLR. Writing was the review and editing was done by all authors.
References
Almeida MQ , Azevedo MF , Xekouki P , et al. 2012 Activation of cyclic AMP signaling leads to different pathway alterations in lesions of the adrenal cortex caused by germline PRKAR1A defects versus those due to somatic GNAS mutations. J Clin Endocrinol Metab 97 E687–E693. (https://doi.org/10.1210/jc.2011-3000)
Anazia K , Koenekoop L , Ferré G , et al. 2024 Interaction networks within disease-associated GαS variants characterized by an integrative biophysical approach. J Biol Chem 300 107497. (https://doi.org/10.1016/j.jbc.2024.107497)
Assie G , Letouze E , Fassnacht M , et al. 2014 Integrated genomic characterization of adrenocortical carcinoma. Nat Genet 46 607–612. (https://doi.org/10.1038/ng.2953)
Bathon K , Weigand I , Vanselow JT , et al. 2019 Alterations in protein kinase A substrate specificity as a potential cause of cushing syndrome. Endocrinology 160 447–459. (https://doi.org/10.1210/en.2018-00775)
Bertherat J , Groussin L , Sandrini F , et al. 2003 Molecular and functional analysis of PRKAR1A and its locus (17q22-24) in sporadic adrenocortical tumors: 17q losses, somatic mutations, and protein kinase A expression and activity. Cancer Res 63 5308–5319.
Beuschlein F , Fassnacht M , Assie G , et al. 2014 Constitutive activation of PKA catalytic subunit in adrenal Cushing's syndrome. N Engl J Med 370 1019–1028. (https://doi.org/10.1056/nejmoa1310359)
Calebiro D , Di Dalmazi G , Bathon K , et al. 2015 cAMP signaling in cortisol-producing adrenal adenoma. Eur J Endocrinol 173 M99–M106. (https://doi.org/10.1530/eje-15-0353)
Calebiro D , Hannawacker A , Lyga S , et al. 2014 PKA catalytic subunit mutations in adrenocortical Cushing's adenoma impair association with the regulatory subunit. Nat Commun 5 5680. (https://doi.org/10.1038/ncomms6680)
Cao Y , He M , Gao Z , et al. 2014 Activating hotspot L205R mutation in PRKACA and adrenal Cushing's syndrome. Science 344 913–917. (https://doi.org/10.1126/science.1249480)
de Reynies A , Assie G , Rickman DS , et al. 2009 Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol 27 1108–1115. (https://doi.org/10.1200/jco.2008.18.5678)
Di Dalmazi G , Kisker C , Calebiro D , et al. 2014 Novel somatic mutations in the catalytic subunit of the protein kinase A as a cause of adrenal Cushing's syndrome: a European multicentric study. J Clin Endocrinol Metab 99 E2093–E2100. (https://doi.org/10.1210/jc.2014-2152)
Fragoso MC , Domenice S , Latronico AC , et al. 2003 Cushing's syndrome secondary to adrenocorticotropin-independent macronodular adrenocortical hyperplasia due to activating mutations of GNAS1 gene. J Clin Endocrinol Metab 88 2147–2151. (https://doi.org/10.1210/jc.2002-021362)
Goh G , Scholl UI , Healy JM , et al. 2014 Recurrent activating mutation in PRKACA in cortisol-producing adrenal tumors. Nat Genet 46 613–617. (https://doi.org/10.1038/ng.2956)
Gorvin CM , Cranston T , Hannan FM , et al. 2016 A G-protein subunit-alpha11 loss-of-function mutation, Thr54Met, causes familial hypocalciuric hypercalcemia type 2 (FHH2). J Bone Miner Res 31 1200–1206. (https://doi.org/10.1002/jbmr.2778)
Gorvin CM , Hannan FM , Howles SA , et al. 2017 Galpha(11) mutation in mice causes hypocalcemia rectifiable by calcilytic therapy. JCI Insight 2 e91103. (https://doi.org/10.1172/jci.insight.91103)
Gorvin CM , Stokes VJ , Boon H , et al. 2020 Activating mutations of the G-protein subunit alpha 11 interdomain interface cause autosomal dominant hypocalcemia type 2. J Clin Endocrinol Metab 105 952–963. (https://doi.org/10.1210/clinem/dgz251)
Graziano MP & Gilman AG 1989 Synthesis in Escherichia coli of GTPase-deficient mutants of Gs alpha. J Biol Chem 264 15475–15482. (https://doi.org/10.1016/s0021-9258(19)84854-8)
Horvath A , Boikos S , Giatzakis C , et al. 2006 A genome-wide scan identifies mutations in the gene encoding phosphodiesterase 11A4 (PDE11A) in individuals with adrenocortical hyperplasia. Nat Genet 38 794–800. (https://doi.org/10.1038/ng1809)
Hu Q & Shokat KM 2018 Disease-causing mutations in the G protein Gαs subvert the roles of GDP and GTP. Cell 173 1254–1264.e11. (https://doi.org/10.1016/j.cell.2018.03.018)
Ioannidis NM , Rothstein JH , Pejaver V , et al. 2016 REVEL: an ensemble method for predicting the pathogenicity of rare missense variants. Am J Hum Genet 99 877–885. (https://doi.org/10.1016/j.ajhg.2016.08.016)
Kroeze WK , Sassano MF , Huang XP , et al. 2015 PRESTO-Tango as an open-source resource for interrogation of the druggable human GPCRome. Nat Struct Mol Biol 22 362–369. (https://doi.org/10.1038/nsmb.3014)
Landis CA , Masters SB , Spada A , et al. 1989 GTPase inhibiting mutations activate the alpha chain of Gs and stimulate adenylyl cyclase in human pituitary tumours. Nature 340 692–696. (https://doi.org/10.1038/340692a0)
Lorea CF , Moreno DA , Borges KS , et al. 2012 Expression profile of apoptosis-related genes in childhood adrenocortical tumors: low level of expression of BCL2 and TNF genes suggests a poor prognosis. Eur J Endocrinol 167 199–208. (https://doi.org/10.1530/eje-12-0183)
Mendes de Oliveira E , Keogh JM , Talbot F , et al. 2021 Obesity-associated GNAS mutations and the melanocortin pathway. N Engl J Med 385 1581–1592. (https://doi.org/10.1056/nejmoa2103329)
Moyer AJ , Fernandez FX , Li Y , et al. 2023 Overexpression screen of chromosome 21 genes reveals modulators of Sonic hedgehog signaling relevant to Down syndrome. Dis Model Mech 16 dmm049712. (https://doi.org/10.1242/dmm.049712)
Muca C & Vallar L 1994 Expression of mutationally activated G alpha s stimulates growth and differentiation of thyroid FRTL5 cells. Oncogene 9 3647–3653.
Murray F & Insel PA 2013 Targeting cAMP in chronic lymphocytic leukemia: a pathway-dependent approach for the treatment of leukemia and lymphoma. Expert Opin Ther Targets 17 937–949. (https://doi.org/10.1517/14728222.2013.798304)
Noon LA , Franklin JM , King PJ , et al. 2002 Failed export of the adrenocorticotrophin receptor from the endoplasmic reticulum in non-adrenal cells: evidence in support of a requirement for a specific adrenal accessory factor. J Endocrinol 174 17–25. (https://doi.org/10.1677/joe.0.1740017)
O'Hayre M , Vazquez-Prado J , Kufareva I , et al. 2013 The emerging mutational landscape of G proteins and G-protein-coupled receptors in cancer. Nat Rev Cancer 13 412–424. (https://doi.org/10.1038/nrc3521)
Ronchi CL 2019 cAMP/protein kinase A signalling pathway and adrenocortical adenomas. Curr Opin Endocr Metab Res 8 15–21. (https://doi.org/10.1016/j.coemr.2019.06.003)
Ronchi CL , Di Dalmazi G , Faillot S , et al. 2016 Genetic landscape of sporadic unilateral adrenocortical adenomas without PRKACA p.Leu206Arg mutation. J Clin Endocrinol Metab 101 3526–3538. (https://doi.org/10.1210/jc.2016-1586)
Rothenbuhler A , Horvath A , Libe R , et al. 2012 Identification of novel genetic variants in phosphodiesterase 8B (PDE8B), a cAMP-specific phosphodiesterase highly expressed in the adrenal cortex, in a cohort of patients with adrenal tumours. Clin Endocrinol 77 195–199. (https://doi.org/10.1111/j.1365-2265.2012.04366.x)
Sato Y , Maekawa S , Ishii R , et al. 2014 Recurrent somatic mutations underlie corticotropin-independent Cushing's syndrome. Science 344 917–920. (https://doi.org/10.1126/science.1252328)
Schubach M , Maass T , Nazaretyan L , et al. 2024 CADD v1.7: using protein language models, regulatory CNNs and other nucleotide-level scores to improve genome-wide variant predictions. Nucleic Acids Res 52 D1143–D1154. (https://doi.org/10.1093/nar/gkad989)
Schwarz JM , Cooper DN , Schuelke M , et al. 2014 MutationTaster2: mutation prediction for the deep-sequencing age. Nat Methods 11 361–362. (https://doi.org/10.1038/nmeth.2890)
Stallaert W , van der Westhuizen ET , Schonegge AM , et al. 2017 Purinergic receptor transactivation by the beta(2)-adrenergic receptor increases intracellular Ca(2+) in nonexcitable cells. Mol Pharmacol 91 533–544. (https://doi.org/10.1124/mol.116.106419)
Stork PJ & Schmitt JM 2002 Crosstalk between cAMP and MAP kinase signaling in the regulation of cell proliferation. Trends Cell Biol 12 258–266. (https://doi.org/10.1016/s0962-8924(02)02294-8)
Stratakis CA 2007 Adrenocortical tumors, primary pigmented adrenocortical disease (PPNAD)/Carney complex, and other bilateral hyperplasias: the NIH studies. Horm Metab Res 39 467–473. (https://doi.org/10.1055/s-2007-981477)
Sun D , Flock T , Deupi X , et al. 2015 Probing Galphai1 protein activation at single-amino acid resolution. Nat Struct Mol Biol 22 686–694. (https://doi.org/10.1038/nsmb.3070)
Tadjine M , Lampron A , Ouadi L , et al. 2008 Frequent mutations of beta-catenin gene in sporadic secreting adrenocortical adenomas. Clin Endocrinol 68 264–270. (https://doi.org/10.1111/j.1365-2265.2007.03033.x)
Tesmer JJ , Sunahara RK , Gilman AG , et al. 1997 Crystal structure of the catalytic domains of adenylyl cyclase in a complex with Gsalpha.GTPgammaS. Science 278 1907–1916. (https://doi.org/10.1126/science.278.5345.1907)
Tissier F , Cavard C , Groussin L , et al. 2005 Mutations of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res 65 7622–7627. (https://doi.org/10.1158/0008-5472.can-05-0593)
Van Eps N , Preininger AM , Alexander N , et al. 2011 Interaction of a G protein with an activated receptor opens the interdomain interface in the alpha subunit. Proc Natl Acad Sci U S A 108 9420–9424. (https://doi.org/10.1073/pnas.1105810108)
Vaser R , Adusumalli S , Leng SN , et al. 2016 SIFT missense predictions for genomes. Nat Protoc 11 1–9. (https://doi.org/10.1038/nprot.2015.123)
Walker C , Wang Y , Olivieri C , et al. 2021 Is disrupted nucleotide-substrate cooperativity a common trait for Cushing's syndrome driving mutations of protein kinase A? J Mol Biol 433 167123. (https://doi.org/10.1016/j.jmb.2021.167123)
Weigand I , Ronchi CL , Vanselow JT , et al. 2021 PKA Calpha subunit mutation triggers caspase-dependent RIIbeta subunit degradation via Ser(114) phosphorylation. Sci Adv 7 eabd4176. (https://doi.org/10.1126/sciadv.abd4176)
Weinstein LS , Shenker A , Gejman PV , et al. 1991 Activating mutations of the stimulatory G protein in the McCune-Albright syndrome. N Engl J Med 325 1688–1695. (https://doi.org/10.1056/nejm199112123252403)
Wilson CH , McIntyre RE , Arends MJ , et al. 2010 The activating mutation R201C in GNAS promotes intestinal tumourigenesis in Apc(Min/+) mice through activation of Wnt and ERK1/2 MAPK pathways. Oncogene 29 4567–4575. (https://doi.org/10.1038/onc.2010.202)
