Corticotroph pituitary neuroendocrine tumors (corticotroph PitNETs) are the subtype of tumors of pituitary gland (1). They develop as a result of neoplastic transformation of corticotropic cells of the anterior pituitary (1). Most of corticotroph PitNETs cause Cushing’s disease (CD) which is characterized by a set of symptoms caused by increased ACTH secretion and consequent serum cortisol excess. Approximately 20% of corticotroph tumors are clinically hormonally inactive (silent), i.e. do not cause rise in cortisol in patients’ circulation (2). The reason why a proportion of ACTH-secreting tumors remain clinically silent and do not cause Cushing’s disease remains unclear. Intriguingly, the molecular profiles of these two groups of corticotroph PitNETs are similar (3, 4).

Hypothalamic–pituitary–adrenal (HPA) axis is a neurohormonal system that provides a negative feedback between pituitary and adrenal secretion (5). Glucocorticoids inhibit ACTH secretion by reducing the level of hypothalamic CRH hormone as well as by direct inhibition of pituitary corticotroph cells activity (5). Partial resistance to glucocorticoids, a hallmark of Cushing’s disease, results from the impairment in HPA axis regulation (6).

Glucocorticoids signal through two steroid hormone receptors: glucocorticoid receptor (GR) and mineralocorticoid receptor (MR) (5, 7). Both receptors are intracellular proteins that upon ligand-dependent activation translocate to the nucleus, where they exert their role of transcription factors and regulate the expression of glucocorticoid-responsive genes. GR and MR have sequence homology and high similarity in DNA binding domains that are responsible for receptor interaction with glucocorticoid response elements (GREs) in DNA (5, 7). The receptors directly regulate the transcription of target genes and/or influence the expression indirectly by interacting with other transcription factors (5, 7, 8). In cells expressing both receptors, MR and GR bind GREs as homo- or heterodimers and the role of dimer composition is not fully understood (7).

The role of corticosteroid receptors in pituitary corticotroph tumors is relatively poorly understood (6, 9). In CD-causing tumors, loss of heterozygosity (LOH) in the NR3C1 (encoding GR) locus was observed (10). More recently, somatic point mutations in NR3C1 were found in 6.4% of patients with CD (9). NR3C1 expression may be also downregulated by hsa-miR-124-3p miRNA in pituitary corticotroph tumors (11). Considering the growing interest in the role of GR and nearly no information on the role of MR in corticotroph PitNETs, we aimed to investigate the expression levels of both receptors in a relatively large cohort of functioning and silent corticotroph tumors.

Corticotroph PitNETs can be divided into functioning and silent tumors. They substantially differ in tumor clinical manifestation and results of biochemical tests. These differences are the basis for discriminating functioning corticotroph, CD-causing tumors and silent tumors. Functioning corticotroph PitNETs occur more frequently in women, while this sex-related prevalence is not observed in SCAs (14–16). Concordant difference in proportions of males and females in groups of CD and SCA patients was also observed in our patients cohort. Functioning corticotroph PitNETs are commonly small in size and are diagnosed due to excessive hormone secretion while SCAs are larger and commonly diagnosed following neurological symptoms caused by tumor volume. Accordingly, in our patients cohort SCAs are significantly larger than functioning tumors. In general, SCAs are considered as more aggressive tumors having higher rate of invasive growth and commonly they show sparse granulation pattern that is related to worse prognosis (17–19). In our study SG tumors were more prevalent in SCA patients while DG tumors were more frequent among CD patients that is also in line with previous reports (19). However, in our patient group we did not observe a difference in invasive growth status determined with Knosp’s grading or proliferation rate assessed based on Ki-67 staining. Majority of the tumors in each group were non-invasive ones with Ki-67 score below 3%. Low rate of invasive tumors may indicate that our group of SCA patients is not fully representative that should be taken into account with generalization of our results.

Glucocorticoids play a pivotal role in the regulation of hypothalamic–pituitary–adrenal (HPA) axis in a negative feedback mechanism. They inhibit ACTH secretion both indirectly via hypothalamus (by reducing CRH secretion) and directly acting on pituitary corticotroph cells (5). Glucocorticoids can be bound by two intracellular receptors GR and MR that regulate transcription of target genes (5, 7). It was shown in corticotroph cells that upon stimulation with glucocorticoids GR regulates POMC expression though various mechanisms including binding of GR to negative GREs (nGREs) in gene promoter (20), chromatin remodeling (21) or interaction with other nuclear receptor (NR4A1) involved in POMC regulation (22). Upon stimulation, GR may also act in nongenomic way by regulating membrane permeability for calcium influx (23). The function of GR and MR receptors in regulation of corticotroph cells activity suggests that they may play an important role in pituitary corticotroph tumors. According to our best knowledge, only few studies on limited number of patients were performed to address the role of GR expression and nearly no research on the role on MR expression in corticotroph PitNETs were published.

In our previous study comparing silent and functioning pituitary corticotroph tumors we found a difference in the expression level of genes encoding GR and MR (NR3C1 and NR3C2, respectively) (11). In the current investigation we included a notably higher number of patients to validate previous results. The obtained data confirmed a higher NR3C1 expression in silent tumors as compared to tumors causing CD, but we did not observe the difference in NR3C2 level. Our results show that, although statistically significant, the difference is slight. This observation supports our previous suggestion that higher NR3C1 expression has rather modulatory contribution to silent nature of SCAs than a strong causative role. The general model of GR activity indicates that its higher expression may directly decrease ACTH secretion (5) and contribute to non-secreting nature of SCAs. Importantly, secretory activity of corticotroph cells is regulated in a complex way. Previous studies comparing silent and functioning corticotroph tumors showed that they differ in several elements of this regulation including POMC expression, prohormone convertases level, dopamine or somatostatin receptor level and expression of specific ion channel (4, 24–28). We suppose that changes of few regulatory elements can occur simultaneously with additive effect on functioning of corticotroph tumors and GR/MR expression can be one of these changes. The possibly distinct expression of NR3C1 in functioning and silent corticotroph tumors was previously investigated in smaller series of patients but no significant difference was found in these studies (27, 28). This was probably due to lower patients’ numbers, i.e. n = 50 (36 CD, 14 SCA) and n = 20 (12 CD, 8 SCA), respectively and a small difference in gene expression levels. Previous analysis of GR protein expression in SCA and CD-causing tumors also did not reveal difference between the tumors of diverse functional status (29).

The change of functional status of corticotroph tumor can be observed in some patients. Cases of spontaneous conversion of silent adenoma into CD-causing corticotroph tumor as well as functioning to non-functioning tumor were reported (30, 31). Periodic changes of the secretory activity of the tumor are also considered as probable cause of cyclic Cushing's syndrome in patient with ACTH-related hypercortisolemia who experience temporal switch between hypercortisolemia and normocortisolemia status (32). The observed different NR3C1 expression between silent and functioning corticotroph tumors suggest that changes in GR expression may also contribute to temporary changes of tumor activity. Obviously, our speculation require further investigation. Recently, the role of GR in CD pathogenesis attracted the attention since new somatic point mutations in NR3C1 gene in functioning corticotroph tumors were identified (33). It has been estimated that these mutations occur in 6.2% of CD patients (34) although the functional role of most of identified somatic variants is unclear and was not investigated in vitro. Three NR3C1 mutations that were found in study by Miao et al. were shown to produce truncated GR protein or to reduce GR expression level (35).

The role of GR expression in patients with Cushing’s disease was not investigated extensively and the results of only a few studies were published. In general, these studies did not reveal any relationship of NR3C1 gene expression level and patients’ clinical features (27, 28, 36, 37) or GR protein expression with patients’ characteristics (29). One study indicated the relationship between low NR3C1 expression and the response to high dose dexamethasone suppression test, however, only six functioning corticotroph tumors were used in this evaluation (38). Of note, most of these previous studies included relatively low number of CD patients: n=17 (36), n=6 (38), n=12 (27), n=8 (29), n=36 (28) with only one study including more than 50 persons (37). Interestingly, although GR and MR are both involved in the response to glucocorticoids, to our knowledge, there is no data on NR3C2 or MR protein expression in CD-causing tumor available to date. Initially, the expression of MR was considered more tissue-specific as compared to GR, with highest levels in kidney and adipose tissue but its expression was identified in multiple human tissues (39). According to publicly available databases on tissue specific gene/protein expression in human MR is widely expressed in most tissue types/organs including normal pituitary gland (Geno-type-Tissue Expression project (https://gtexportal.org) or Protein atlas (https://www.proteinatlas.org), accessed on November, 2022). The expression of both MR and GR was documented in PVN neurons, adipocytes, osteoblasts, immune cells and kidney cells (40–44).

In our study we found some interesting relationships between NR3C1 and NR3C2 expression and clinical/histological features in CD patients. The expression levels of both receptors were negatively correlated with plasma ACTH level and tumor size. The expression of NR3C2 was also higher in patients with post-surgery clinical remission than in patients who did not experience biochemical remission after surgery and in patients with DG tumors as compared to SG tumors. DG histology is considered as related to favorable clinical outcome (19). Expression levels of both genes were also higher in tumors with USP8 mutations, which are the most common driver mutations in corticotroph PitNETs. These mutations are generally related to smaller tumor size and less aggressive growth (28, 34) USP8-mutatated and wild type tumors are also characterized by different biological features (34). The expression of NR3C1 turned out to be the highest in USP8-mutated tumors from patients with clinical remission after surgery indicating that CD patients with this mutation and high NR3C1 level have the best chance for clinical remission.

The relationships between the expression levels of GR/MR-encoding genes and clinical parameters revealed in our study were not strong but statistically significant. This allows to cautiously generalize the results of this study. They clearly show a trend in which higher receptor expression corresponds to more favorable clinical profile in patients, including lower plasma ACTH level, smaller tumor size, clinical remission after surgery and favorable features of sparsely granulated, USP8-mutated tumors.

Due to a weak relationship with patients characteristics it is difficult to indicate any clinical usefulness of testing NR3C1 or NR3C2 expression but our data support the clinically relevant roles of the receptors expression levels in functioning corticotroph pituitary tumors.

Probably the clinically relevant results were not observed in previous studies on NR3C1 expression in CD due to lower number of patient and the fact that the observed relationship is not strong. We suppose that the small difference in gene expression affects the general lack of the finding of the significant associations in our GR and MR protein level analysis. While qRT-PCR results are numerical data by definition, immunohistochemistry is a qualitative method and based on subjective microscopic observations, thus, it is generally less accurate. In our setting, GR immunostaining evaluation produced a relatively small range of immunoreactivity scores and similarly for MR we observed mainly low or moderate immunoreactivity. This is probably also the reason why we did not observe a correlation between qRT-PCR results and protein quantification in our study, and any difference in GR expression when comparing SCAs and functioning tumors.

Targeting GR with receptor antagonist is one of the main pharmacological options in Cushing’s syndrome (CS) aimed to reduce cortisol-related complications (9). The efficacy of two GR modulators mifepristone and relacorilant were already tested and subjected to phase III clinical trials (9). In case of ACTH-related hypercortisolemia this therapeutical approach may provide both therapeutical benefit and adverse effect due to direct effect on corticotroph tumor. Blocking the receptors in tumor may decline the inhibitory effect of GR on ACTH secretion i.e. direct negative adrenal-pituitary feedback. In fact, the use of mifepristone in CD was shown to result in more than 2-fold increase of ACTH secretion in patients (45). More promising preliminary results were obtained in the evaluation of relacorilant (46), but phase III clinical trials are still in progress. Our observation of the negative correlation between the tumor expression of corticosteroid receptor genes and plasma ACTH level suggests that lower receptors levels contribute to less effective negative regulation and possibly could affect lower adverse effect of GR antagonist on direct negative adrenal-pituitary feedback. We suppose that evaluation of expression status of GR/MR receptors (at both gene and protein level) may provide some predictive value in CD patients treated with GR modulators. Perhaps, patients with low expression could benefit more from the treatment. This would be important because our data indicate that these patients are of worse prognosis. This issue requires further specific research. The roles of GR and MR are considered cell-type specific. These receptors may play distinct, even opposite functions such as in immune cells where MR mediates proinflammatory effects, while GR mediates immune suppression (47). Both receptors have pivotal roles in the regulation of HPA axis and reducing activity of pituitary corticotrophs. MR is considered to be ligand-activated throughout the entire circadian cycle, including periods of low circulating glucocorticoid concentrations (inactive phase), when GR receptor is inactive. In turn, GR receptor is activated at high glucocorticoid circadian phase (active phase) or during a stress response (5). These different roles probably result from notably higher affinity of MR to glucocorticosteroids and longer substrate binding time as compared to GR (5). However, more recent data indicate more dynamic activity of MR and circadian differences in its occupancy at particular target DNA loci, as well as cooperative action of both receptors forming MR : GR heterodimers (48, 49). Most of the published data on MR and GR role in regulation of HPA-axis concern their role in hypothalamic neurons while the mechanism of their interaction in corticotroph cells was not determined. Importantly, our observations of co-expression of NR3C1 and NR3C2 in corticotroph PitNETs and negative correlation between the expression of each gene and morning plasma ACTH level suggest that in pituitary corticotroph cells the receptors may synergistically regulate secretory activities of hypothalamic neurons. This hypothesis needs verification in functional research regarding the role of MR in pituitary corticotrophs and cooperation between GR and MR in this type of cells.

GR and MR are both expressed in corticotroph tumors and there is a correlation between mRNA levels of genes encoding these receptors. NR3C1 expression level is higher in silent than in functioning corticotroph tumors. In CD patients there is an evident trend towards more favorable clinical characteristics in tumors with higher receptors’ expression.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee of Maria Sklodowska-Curie National Research Institute of Oncology, Warsaw, Poland. The patients/participants provided their written informed consent to participate in this study.

PK, MB and MM contributed to conception and design. PK, NR, BM and MP contributed to investigation and acquiring laboratory results. GZ, JK, LD, PW and MM contributed to investigation and acquiring of resources (tissue samples and data for the analysis). PK, MM, MB contributed to visualization. PK, MB contributed to data analysis. PK, MB wrote the first draft of the manuscript. PW, MM contributed to writing—review and editing the manuscript. All authors contributed to the article and approved the submitted version.