Signaling Under Stress: Targeting the Glucocorticoid Receptor in Cancer Authors: Melanie S. Flint, David M. O’Malley, Domenica Lorusso, Susan K. Lutgendorf, Alexander B. Olawaiye, Adrian M. Jubb, Premal H. Thaker Published: 2026 ABSTRACT AbstractGlucocorticoid receptor (GR) signaling is critical to our physiology but is dysregulated in cancer by the pain, psychologic distress, iatrogenic morbidity, and pathology associated with an advanced, invasive disease. Many tumors exploit glucocorticoid signaling to provide antiapoptotic survival signals, stimulate growth and metastasis, suppress immune surveillance, and drive treatment resistance. Endogenous cortisol and exogenous glucocorticoids are typically associated with poor responses to cytotoxic chemotherapy, targeted therapy, and immunotherapy in patients with solid tumors. Translational and nonclinical data show that selective GR antagonists (SGRA) synergize with anticancer agents, including taxanes, androgen receptor inhibitors, PARP inhibitors, and anti–programmed cell death ligand 1 (PDL-1) immune checkpoint inhibitors, driving improved anticancer activity. In a tumor-intrinsic manner, SGRAs downregulate the antiapoptotic proteins SGK1 and DUSP1, which induce resistance to cytotoxic chemotherapy. This synergy has been confirmed in several randomized controlled trials, which showed improved efficacy when SGRAs were added to standard-of-care taxane therapy in platinum-resistant ovarian cancer. In a distinct cellular context, SGRAs inhibit resistance to antiandrogen therapy mediated through the GR and have the potential for additive anticancer activity in prostate cancer. Herein, we summarize the role of glucocorticoid signaling in solid tumor biology, providing mechanistic insights into recent clinical advances and emphasizing key outstanding translational and clinical questions. Introduction Advanced cancer causes tissue damage, pain, and psychosocial distress related to diagnosis, recurrence, and associated mortality. Patients often have limited treatment options: invasive surgery, radiotherapy, chemotherapy, targeted therapy, and immunotherapy of modest efficacy and often substantial toxicity. The disease and the treatment journey often substantially diminish daily functioning and quality of life (1, 2). The illness and its treatment cause chronic stress, a complex physiologic process that results from the integration of sensory inputs, cognitive processing, physiologic signaling, and pattern recognition in the limbic system of the brain. In a landmark randomized controlled trial, Spiegel and colleagues (3) showed that 1 year of weekly supportive group therapy, with self-hypnosis for pain, prolonged survival in women with metastatic breast cancer. Similar findings were observed in several subsequent trials (4, 5) and meta-analyses with related interventions (6, 7). Furthermore, use of exogenous corticosteroids was associated with increased metastasis in observational studies on patients with breast cancer in the 1950s and 1960s (8, 9). However, these clinical observations outpaced our mechanistic understanding of how chronic stress, steroid use, and psychosocial interventions affect cancer outcomes. Herein, we make a case for a central role for cortisol, the principal stress hormone, and the glucocorticoid receptor (GR), a key mediator of cortisol signaling, in cancer progression and the efficacy of cancer treatment. GR agonists, referred to as glucocorticoids or corticosteroids, have an established role in the treatment of lymphoid malignancies (10) that is distinct from solid tumors and has been extensively reviewed elsewhere (11, 12). In brief, GR activation in lymphoid cancers triggers lineage-specific transcriptional programs that induce potent apoptotic responses. By contrast, in solid tumors, GR signaling exerts antiapoptotic and immunosuppressive effects. The functional heterogeneity of the GR pathway in lymphoid cancers versus solid tumors and its clinical implications are summarized in Supplementary Table S1. This review will focus on solid tumor biology and the potential for drugs that selectively target the GR to enhance the efficacy of agents targeted against related nuclear hormone receptors, cytotoxic chemotherapy, and immunotherapy. Physiologic GR Signaling and Function GR signaling is a central physiologic mechanism that integrates endocrine cues to regulate metabolism, immunity, and stress responses. Glucocorticoids are a class of natural and synthetic steroid molecules that agonize the GR encoded by NR3C1. In humans, cortisol is the endogenous glucocorticoid secreted by the adrenal gland, under the control of pituitary adrenocorticotropic hormone (ACTH) in the hypothalamic–pituitary–adrenal (HPA) axis. Under physiologic conditions, glucocorticoids regulate a wide variety of genes and cellular processes to maintain metabolic, immune, cardiovascular, renal, and bone homeostasis (13–15). In addition, cortisol influences the central nervous system, affecting mood, behavior, and sleep–wake cycles (14). The body’s circadian rhythm is controlled by the hypothalamic suprachiasmatic nucleus (SCN), which is influenced by light/dark cycles. In the morning, the SCN signals the hypothalamus to release corticotropin-releasing hormone (CRH), which in turn stimulates the release of ACTH from the anterior pituitary gland. ACTH acts on the adrenal cortex to stimulate the synthesis and release of cortisol. As cortisol rises, it exhibits negative feedback on CRH and ACTH release, and cortisol levels decline to reach their lowest point in the late evening before sleep (14, 16). Cortisol is also released in response to physical or psychologic stress (14). Neurocognitive networks work together to regulate cognitive executive functions and responses to stimuli (17). The prefrontal cortex mediates response inhibition, a core function governing impulse control. Human studies have shown that whereas acute increases in cortisol due to stress can temporarily enhance response inhibition (18), chronically elevated cortisol due to chronic stress impairs this critical cognitive process (19). The amygdala synthesizes inputs from external senses and internal physiologic signals (e.g., hormone levels), identifying a state of stress through a complex process of pattern recognition. Signaling through the stria terminalis, the amygdala initiates CRH release by activating the hypothalamic paraventricular nucleus. The stria terminalis and noradrenaline also stimulate CRH release by activating neurons in the basal nucleus of the stria terminalis (16, 20). The consequent increase in cortisol triggers a wide variety of physiologic effects to help the body cope with stress, including increasing blood glucose levels, mobilizing fat and protein to provide additional energy, suppressing the immune system and inflammation, increasing cardiovascular responsiveness to epinephrine and norepinephrine, and increasing cognitive alertness (16, 20). Cortisol may become dysregulated with increased endogenous secretion in states of chronic psychologic stress and chronic disease, adversely affecting multiple organ systems (13, 21, 22). Psychologically this may manifest with increased anxiety, depressed mood, irritability, emotional lability, reduced motivation, and anhedonia (21, 23). High cortisol levels cause metabolic dysregulation resulting in insulin-resistant diabetes, visceral fat deposition/redistribution, and muscle atrophy (24, 25). Patients may exhibit hypertension, water/electrolyte imbalance, osteoporosis, cataracts, peptic ulcers, and thinning of the skin (22). These effects are also known toxicities of prolonged high doses of systemically administered synthetic glucocorticoids (e.g., dexamethasone, betamethasone, and prednisone), widely used to treat inflammatory and immune diseases (26, 27). The GR belongs to the nuclear receptor superfamily (14, 28) and is a ubiquitously expressed ligand-activated transcription factor (14, 26). In the absence of a ligand, the GR resides in the cell’s cytoplasm as a monomer, part of a large multiprotein complex that includes HSP90, immunophilins (e.g., FKBP51 and FKBP52), and other chaperone proteins (Fig. 1A; refs. 14, 26, 29). This complex maintains the GR in a conformation that has a high affinity for ligand binding. Cortisol, a lipophilic compound, freely diffuses across the cell membrane and binds a hydrophobic pocket in the ligand-binding domain (LBD) near the c-terminus of the GR protein. Binding of cortisol introduces a significant conformational change in the LBD, trapping the ligand and creating an optimal surface for coactivator protein binding, an essential change for the receptor’s transcriptional activation function (13, 14, 26, 29). The conformational change also dissociates the HSP90 chaperone complex, allowing homodimerization of the GR and exposing nuclear localization signals. The GR is then rapidly and actively transported into the nucleus through nuclear pores. Potential mechanism of GR-mediated effects on the apoptotic cascade. A, Ligand activation of GR leads to its nuclear translocation, where it interacts with GREs and coregulators and upregulates antiapoptotic genes SGK1 and DUSP1 (30). SGK1 phosphorylates the Forkhead family transcription factor FOXO3a, resulting in its exit from the nucleus and degradation, and reduced expression of proapoptotic genes BIM, BID, and PUMA (31, 32). This can reduce the activation of proapoptotic effectors BAK/BAX and inhibit apoptosis. DUSP1-mediated dephosphorylation of phospho-JNK, a member of the MAP-kinase superfamily, blocks inactivation of antiapoptotic proteins BCL-2 and BCL-XL (31, 33). This may inhibit BAK/BAX activation and apoptosis. Microtubule-targeting agents induce apoptosis through disruption of microtubule dynamics and prolonged mitotic arrest (34–36). B, Relacorilant, through GR antagonism, downregulates SGK1 and DUSP1, potentially improving MTA-induced activation of the intrinsic apoptotic cascade (37). Relacorilant (CORT125134, Corcept Therapeutics, Inc.) is an investigational, orally administered SGRA that can reverse the effects of glucocorticoids by restoring chemotherapy activity and antagonizing the antiapoptotic effects of glucocorticoids (37, 38). BAK, BCL-2 antagonist/killer; BAX, BCL-2–associated X; BCL-2, apoptosis regulator; BCL-XL, BCL-2–like 1; BID, BH3 interacting domain death agonist; BIM, BCL-2–like 11; FOXO3a, forkhead box O3; JNK, c-Jun N-terminal kinase; MOA, mechanism of action; MTA, microtubule-targeting agent; P, phosphorylated; PUMA, p53-upregulated modulator of apoptosis; U, ubiquitinated. The GR homodimer binds a general consensus 15-base pair DNA sequence 5′-GGTACAnnnTGTTCT-3′ (in which nnn is a 3-base pair spacer) referred to as a glucocorticoid response element (GRE; ref. 14). To activate or suppress transcription, ligand- and DNA-bound GR forms a transcriptional complex with multiple coactivator or corepressor proteins. The nature of this transcriptional complex varies and determines the magnitude of the transcriptional response. Factors that influence the makeup of the transcriptional complex include the ligand bound to the GR, the specific GRE, accessibility within the local chromatin environment, the cellular context (cell type, differentiation state, cellular signaling activity, posttranslational modifications of the GR and its regulators, and other transcription factors), and the concentration/availability of coregulators (29, 39, 40). As such, the effect of GR activation is both ligand- and context-specific, allowing this canonical signaling pathway to have pleiotropic effects involving multiple distinct physiologic processes. Dysregulation of GR Signaling in Cancer In solid tumors, dysregulated GR signaling drives prosurvival transcriptional programs that promote tumor growth, invasion, and resistance to therapy. The NR3C1 gene encoding the GR shows little evidence for functional somatic driver mutations or epigenetic changes in human tumors. Seven percent of endometrial cancers have inactivating nonsense or missense mutations in NR3C1, and 6% of renal clear cell carcinomas show evidence for amplifications involving the NR3C1 in The Cancer Genome Atlas, as analyzed using cBioPortal (RRID: SCR 014555), but the functional consequences of these changes have not been studied (41–43). Genes encoding the synthesis of endogenous GR ligands, or GR coregulators, are not the target of recurrent somatic alterations in human tumors (cBioPortal, RRID: SCR 014555). Outside of functional endocrine tumors, there is limited evidence for autocrine or paracrine signaling through the GR in the tumor microenvironment through direct synthesis by CYP11B1 or recycling from inactive metabolites by 11β-HSD1 (44–48). Yet, the concentrations of glucocorticoid produced in the tumor microenvironment are 10- to 200-fold lower than circulating cortisol and may be of limited physiologic relevance (49). Multiple splice variants of GR have been reported, as have alternate transcripts originating from different start sites, including the GRβ splice variant that lacks a ligand-binding domain (29). However, the physiologic significance of these forms of the GR is uncertain given their limited expression pattern and ligand/DNA-binding properties (50). There is no evidence for alternate splicing or transcript forms having a role in tumor initiation, promotion, or metastasis. The GR is widely expressed in normal tissues, and its expression can be upregulated by hormones, cytokines, and chronic stress (51). Consequently, numerous tumor cell types consistently exhibit GR (52) expression, even without genetic amplification or active oncogenic signaling mechanisms. Preclinical work using synthetic glucocorticoids suggest that normal endogenous glucocorticoid levels are sufficient to activate the GR, as the morning serum cortisol range of 276 to 552 nmol/L is higher than the concentration at which half-maximal GR activation occurs (9.5 nmol/L; refs. 53–55). The GR provides prosurvival signals in malignant solid tumors (56, 57), promotes tumor growth (58) and invasion (59), drives metastasis (60), and confers resistance to chemotherapy (61–64). Consistent with these functional observations, higher tumor cell GR expression is associated with poor outcomes in patients with ovarian and endometrial cancers (65–67). A Mendelian randomization study investigated the causal association between 3 genetic variants associated with morning plasma cortisol levels at the genome-wide significance level (P < 5 × 10−8) in the Cortisol Network consortium and the risk of various cancers in datasets from large-scale cancer consortia, the UK Biobank, and the FinnGen consortium (68). No significant associations were observed for cancer overall or breast, colorectal, lung, ovarian, or prostate cancer. The notable exception was an increased risk of endometrial cancer associated with higher genetically predicted cortisol levels [odds ratio, 1.5; 95% confidence interval (CI), 1.13–1.99; P = 0.005] in a meta-analysis of the endometrial cancer consortium and the FinnGen datasets (68). This association may be driven by an increase in estrogen or testosterone, known risk factors for endometrial cancer, which are associated with endogenous hypercortisolism. Alternatively, the association between plasma cortisol and endometrial cancer might be mediated by obesity, which is strongly associated with an increased risk of endometrial cancer and may be caused by elevated cortisol. Additional mediators of the observed effect might be hyperglycemia and hyperinsulinemia, which have been implicated in endometrial cancer pathogenesis and are a direct effect of cortisol dysregulation. However, the observed association did not materially change after multivariate analyses controlling genetic predictions for these factors. Further research is necessary to elucidate the potential mechanism linking hypercortisolism to endometrial cancer development. A small proportion of patients with cancer have tumors that produce ectopic CRH, ACTH, or cortisol, causing a paraneoplastic syndrome due to extremely high levels of cortisol (and/or mineralocorticoids). With a few exceptions, such as lung carcinoids (69), these functioning endocrine tumors have consistently worse outcomes than nonfunctioning endocrine tumors of the same histology and cell of origin. Adrenal cortical cancers that secrete cortisol recur faster (relative risk 1.43-fold), and patients with these secreting adrenal cortical cancers have a shorter survival (relative risk 1.71-fold) than patients with nonsecreting adrenal cortical cancers (70). Pituitary adenomas that secrete ACTH have double the mortality risk versus nonsecretors [hazard ratio (HR) for age-adjusted mortality 2.35; ref. 71]. Ectopic ACTH secretion by other tumor types (e.g., small-cell lung, thyroid, and pancreatic cancers) is often associated with advanced, aggressive disease with short overall survival (OS), estimated at <4 months for small-cell lung cancer (SCLC) in one study (72–74). The impact may be due to immune suppression, metabolic derangements, resistance to chemotherapy, and diagnostic delay secondary to the cortisol excess. As such, hormone-secreting tumors often require more aggressive management due to their poorer prognosis, but they are rare. Beyond patients with functional endocrine tumors, there is evidence for disruption of the HPA in patients with cancer due to physical stress, persistent pain, and debilitating fatigue and the accompanying psychologic stress from the diagnosis, fear of recurrence, associated morbidity, cost of care, and impending mortality (75, 76). Treatment-related factors also play a significant role in altering cortisol levels and GR signaling in patients with cancer. Synthetic glucocorticoids are frequently used to support cancer therapy, treating chemotherapy-induced nausea and vomiting and inflammation (77, 78). The administration of these exogenous steroids can lead to iatrogenic Cushing syndrome. Clinically, a meta-analysis of 83,614 patients with a wide range of solid cancers has shown that those who receive supportive care with exogenous glucocorticoids have a significantly shorter OS and progression-free survival (PFS; ref. 79). Possible mechanisms for these adverse outcomes include a tumor-intrinsic role for corticosteroids, the impact of corticosteroids on the immune system, or side effects of excess corticosteroids use on normal body systems. Defining elevated cortisol in patients with cancer is a challenge due to its diurnal variation. In general, patients with cancer do not have dramatically elevated plasma cortisol levels. However, increases in cortisol, a flattening of the diurnal slope of cortisol, and/or elevated evening cortisol are reported in cohorts of patients with many different solid tumor types (80–83). Dysregulation of cortisol has been associated with feelings of hopelessness, depression, fatigue, disability (83), more advanced disease, and shorter survival in patients with many different tumor types (83, 84), including breast cancer (80), renal cell carcinoma (85), non-SCLC (NSCLC; ref. 86), and ovarian cancer (87). These are retrospective exploratory subset analyses, often using different methodologies and cut-offs to define cortisol dysregulation, making it challenging to confirm the findings or exclude publication bias. However, mechanistic animal models show evidence for a stress–cortisol–cancer axis, whereby stress increases endogenous glucocorticoid production, which in turn drives the growth and/or metastasis of cancer cells (88–94) through tumor-intrinsic and -immune mechanisms (60, 82, 95, 96). Thus, the totality of the data indicates a direct link between stress, GR signaling, and poor outcomes for patients with cancer. The Role of the GR in Resistance to Cytotoxic Chemotherapy GR signaling plays a key role in chemoresistance by inducing antiapoptotic signaling pathways that enable cancer cells to evade cell death. GR signaling reduces tumor sensitivity to chemotherapy by increasing the expression of antiapoptotic proteins, such as SGK1 and DUSP1 (Fig. 1A; refs. 54, 63, 65, 94, 97, 98). Upon ligand binding, the GR translocates into the nucleus, where it binds glucocorticoid response elements and recruits specific coregulatory complexes. This interaction reshapes the local chromatin architecture and regulates the transcription of downstream target genes, including SGK1 and DUSP1 (Fig. 1A; ref. 33). Increased expression of SGK1 and DUSP1 inactivates the FOXO3a and JNK pathways, which are necessary for apoptosis, and modulates the activity of antiapoptotic BCL-2 family members (Fig. 1A; refs. 31–33). These molecular changes are consequences of GR activation and allow cancer cells to survive cytotoxic damage without undergoing apoptosis and induce resistance to chemotherapy. In support of this, dexamethasone treatment of GR-expressing ovarian cancer cell lines suppressed paclitaxel-induced cell death and upregulated SGK1 and DUSP1 (30). In patients with ovarian cancer, dexamethasone infusion elevated SGK1 and DUSP1 gene expression in tumor specimens (30). Data derived from experimental systems with physiologic levels of glucocorticoids suggest that the role of GR signaling in developing chemotherapy resistance may apply broadly to all patients with solid tumors that express the GR. However, it is likely that the effects are exacerbated in patients with cancer with dysregulated cortisol signaling and in patients receiving exogenous glucocorticoids to support the use of anticancer therapies. Due to the widespread use of exogenous glucocorticoids in this setting, delineating the precise contribution of physiologic cortisol and exogenous glucocorticoids is challenging with the current evidence base. Two main classes of synthetic pharmacologic GR antagonists have entered clinical development: steroidal antiprogestins (mifepristone and its analogs) and selective nonsteroidal glucocorticoid receptor antagonists (“-corilants,” e.g., relacorilant and exicorilant; Supplementary Table S2). These structurally distinct drug classes have a similar potency and activity at the GR but have distinct coregulator interactomes (99) and distinct selectivity, which likely account for their differential clinical activity and therapeutic window. Among the selective nonsteroidal glucocorticoid receptor antagonists, relacorilant (CORT125134) has the most mature dataset to date; little data have been published on other compounds in this class, and generalizability across agents remains uncertain. As such, most of the discussion in this article will use relacorilant as an example. The potent and highly selective GR antagonist, relacorilant [Ki 0.15 nmol/L for GR binding; IC50 5.6–7.2 nmol/L for GR antagonism, and Ki >1,000 nmol/L for binding of androgen receptor (AR), progesterone receptor (PR), and mineralocorticoid receptor], was developed from a nonsteroidal scaffold (99, 100). Relacorilant reversibly binds the ligand-binding domain of the GR, inhibiting the effects of GR agonists but allowing dimerization of the GR, nuclear entry, and recruitment of a distinct interactome of coregulators to GR agonists or mifepristone (99). When bound to relacorilant, GR does not activate transcription at canonical GR response elements, demonstrating that relacorilant is a functional antagonist (99). Selective GR antagonists (SGRA) such as relacorilant, and less selective mifepristone analogs, such as ORIC-101, show synergy with several types of cytotoxic chemotherapy, including gemcitabine, taxanes, and platinum agents (54, 61, 101, 102). In vitro and in vivo GR inhibition suppresses SGK1 and DUSP1 expression (Fig. 1B; refs. 54, 63, 93) and reverses the chemotherapy-resistant epithelial-to-mesenchymal transition phenotype (103–105). This leads to greater chemotherapy-induced apoptosis when SGRAs are combined with chemotherapy (54) and significant tumor growth inhibition, including regression in some nonclinical models (54, 101, 105). In a phase I clinical trial, the addition of relacorilant to nab-paclitaxel in 73 patients with advanced solid tumors (75% of whom had received a prior taxane) achieved a 16% objective response rate (ORR) in all tumor types with a 4.6 months median PFS in patients with ovarian cancer (61). The duration of benefit was longer for relacorilant plus nab-paclitaxel than the prior taxane-containing regimen in 29% of patients (61). In a phase II trial (CORT125134-552), the addition of an intermittent schedule of relacorilant to nab-paclitaxel in patients with platinum-resistant ovarian cancer extended PFS (HR, 0.66; 95% CI, 0.44–0.98; P = 0.038) with a trend to improved OS (P = 0.066; ref. 37). In a recently disclosed phase III trial, ROSELLA, patients receiving relacorilant plus nab-paclitaxel had a statistically significant improvement in PFS by blinded independent central review (BICR) compared with nab-paclitaxel monotherapy (HR, 0.70; 95% CI, 0.54–0.91; P = 0.008; ref. 106). At an interim analysis, there was a clinically meaningful improvement in OS with the addition of relacorilant to nab-paclitaxel (HR, 0.69; 95% CI, 0.52–0.92; median 16 vs. 11.5 months; P = 0.012; ref. 106). At the final analysis, the combination resulted in a statistically and clinically significant improvement in OS compared with nab-paclitaxel alone (HR, 0.65; 95% CI, 0.51–0.83; median 16 vs. 11.9 months; P = 0.0004; ref. 107). In the phase II relacorilant trial, reductions in SGK1 and DUSP1 were observed in whole blood, providing evidence of a pharmacodynamic link to the proposed antiapoptotic mechanism (37, 108). ORIC-101 was also evaluated in combination with nab-paclitaxel in a phase I study but showed a more modest benefit in patients who had progressed on a prior taxane, with a 3.2% ORR across all tumor types and a 2.4 months median PFS in patients with ovarian cancer (101), reflecting the more modest pharmacodynamic effects seen for ORIC-101 (109) and the reduced selectivity when compared with relacorilant (ORIC-101 has an IC50 of 21 nmol/L for the PR whereas relacorilant has no affinity at 1,000 nmol/L; refs. 100, 109). These nonclinical, translational, and clinical observations support a role for SGRAs in sensitizing tumors to chemotherapy and overcoming chemotherapy resistance in ovarian cancer. Similar nonclinical data have been reported in triple-negative breast cancer (56, 58, 64, 94, 110), and patients with SCLC brain metastases are reported to have worse outcomes to teniposide with corticosteroid use (111), suggesting that these observations may apply more broadly to advanced solid tumors. In addition to chemoresistance being mediated through SGK1 and DUSP1 in triple-negative breast cancer (93, 94), the GR may also act through the YAP oncogene to promote survival and reduce sensitivity to cytotoxic agents (112). The reported synergistic effects of PARP inhibitors with mifepristone in patient-derived xenografts also suggest that SGRAs may have further potential in maintenance therapy for patients with ovarian cancer (113). GR Signaling and Resistance to AR Inhibition The AR is the lineage-defining transcription factor for the luminal prostate epithelium (114, 115). The GR and AR are both nuclear hormone receptors with near-identical consensus DNA-binding sites. They regulate the transcriptional activity of an overlapping group of genes (116, 117). Androgen signaling is required for the initiation and progression of early prostate cancer (118). Chemical castration (GnRH modulators), antiandrogens (abiraterone), and AR inhibitors (enzalutamide and apalutamide) are mainstays of prostate cancer therapy (119). Eventually, though, resistance to antiandrogen therapy develops through multiple mechanisms, including mutations that reduce the affinity of AR inhibitors, ligand-independent AR signaling (120), and parallel oncogenic signaling pathways (116). The AR directly represses GR expression (121), and GR expression increases in prostate cancers following exposure to AR inhibitors in cell lines, human tumor xenograft models, and in the clinic (116). Therefore, there is a strong biological rationale that GR activity allows prostate cancers to bypass their dependence on AR signaling and thereby develop resistance to AR inhibitors. Activation of GR reduces the efficacy of AR inhibitors and accelerates the development of resistance, whereas GR antagonism has the opposite effect in nonclinical models (116). Nevertheless, this is a complex context-dependent biology; corticosteroids can also have modest anticancer activity in combination with chemotherapy (122, 123) and in patients progressing on abiraterone (124), secondary to lower adrenal androgen production, resulting in stabilization or modest reductions in prostate-specific antigen (PSA) levels (125). In other contexts, GR agonists may directly activate mutant forms of the AR (such as L701H) that can arise secondary to androgen deprivation therapy (126). Inhibition of the GR has been evaluated in clinical trials of patients with prostate cancer. Initially, the nonselective GR antagonist mifepristone was used. Mifepristone antagonizes GR in the presence of physiologic ligand concentrations, with an IC50 value of 3.3 nmol/L in a GR luciferase antagonist assay (109), but can also show context-dependent partial agonism, with a maximal effect of up to 25% of dexamethasone (127). Mifepristone has off-target activity against several other nuclear hormone receptors, with antagonism (IC50 7.8 nmol/L) of the AR (128). In the absence of androgens, high concentrations of mifepristone (100 nmol/L) have weak partial agonist activity against the AR (15%–20% of the response of the potent agonist R1881 at 0.01 nmol/L; ref. 129), but in other contexts, mifepristone has no agonistic activity (128), and functional assays suggest minimal agonist activity at therapeutically relevant concentrations (130). Monotherapy treatment with mifepristone in 19 men with castration-resistant prostate cancer showed no clinical activity, but a significant increase in circulating androgens (testosterone and dihydrotestosterone increased by 80%–90%) was observed secondary to a negative feedback loop driving increased ACTH production (131). When mifepristone was combined with enzalutamide versus enzalutamide alone in a randomized phase II trial of 66 men with metastatic castration-resistant prostate cancer, no difference in the rate of radiographic or PSA progression was observed (132). Like the monotherapy mifepristone study, a 2.5-fold increase in circulating testosterone was observed in patients that received mifepristone, with the potential to drive tumor growth and adversely affect efficacy (132). Mifepristone analogs, such as ORIC-101 (109), have also failed to show meaningful anticancer activity in prostate cancer (133). When ORIC-101 was combined with enzalutamide in 41 patients with metastatic castration-resistant prostate cancer who had progressed on enzalutamide, no meaningful antitumor effect was observed (133). In healthy subjects, relacorilant was able to completely suppress increases in circulating FKBP5 and GILZ secondary to a physiologic dose of prednisone (134). In a phase Ib/II study, relacorilant was combined with enzalutamide in heavily pretreated patients with metastatic castration-resistant prostate cancer that had progressed following treatment with an antiandrogen (135). Thirty-five patients were enrolled; 94% had received prior abiraterone, 89% at least one AR inhibitor, 9% both enzalutamide and apalutamide, 69% docetaxel, and 49% cabazitaxel. All patients remained castrate, with no increases in testosterone levels observed on treatment. Twelve patients were evaluable for a PSA response at 12 weeks, of whom 4 (33%) achieved at least a 50% reduction in PSA. Twenty-two patients were evaluable by RECIST, of whom 13 (59%) had the best radiographic response of stable disease; there were no objective responses. The median PSA-defined PFS was short at 2.7 months; radiographic median PFS was 1.8 months. However, 4 patients stayed on treatment for >6 months, with time on therapy of 11, 17.5, 22.1 (ongoing), and 38.1 (ongoing) months at the time of the primary analysis, respectively. These data are not sufficient to support further development of relacorilant in an unselected population of patients with heavily pretreated metastatic castration-resistant prostate cancer. However, the data suggest that combined inhibition of the AR and GR may have a role in a selected group of patients and/or in a less treatment-refractory setting. Building on these initial findings, a randomized phase II trial evaluating the addition of relacorilant to enzalutamide compared with enzalutamide alone was initiated in patients with high-risk localized prostate cancer (NCT05726292). Patients received 6 months of therapy prior to surgical resection with the aim of improving pathologic response and reducing recurrence rates with the addition of relacorilant. The nature of the study facilitates the collection of tissue and blood biomarkers before and after treatment to help identify predictive biomarkers of clinical benefit. If positive, these results may inform the design of a pivotal phase III study in patients with advanced prostate cancer. GR Signaling in Estrogen-Driven Tumors The GR shares structural homology, chromatin binding sites, and coregulatory networks with the estrogen receptor (ER). Consequently, the functional impact of GR signaling in estrogen-responsive malignancies, particularly breast and endometrial tumors, is dependent on ER expression and activity. In endometrial cancer, expression of the GR is associated with a worse prognosis but only in tumors that express high levels of the ER (136). More than one-third of DNA binding by the GR in endometrial cancer cells is altered by ER activation, suggesting a complex transcriptional regulatory interplay between the receptors (136). Interactions between activated GR and ER, with cross-receptor influences on transcriptional programs, have also been reported in breast cancer (137–139). In patients with ER+ breast cancers, high GR expression and a signature of GR activation are both associated with improved outcomes (97, 140), unlike triple-negative breast cancers in which high GR expression is a poor prognostic factor (141). GR agonists can reduce circulating estradiol in mice bearing MCF7 xenografts by increasing the expression of estrogen sulfotransferase in tumor cells and the liver (142). Moreover, the GR can tether and displace the ER from estrogen DNA response elements in the promoters of canonical target genes such as CCND1, leading to reduced expression of proliferative genes (139, 143). The GR can also reduce the expression of wild-type and mutant ESR1, reducing the proliferation and metastasis of ESR1-mutant breast cancer xenografts (140). Finally, the GR can modulate the ER transcriptional program, leading to an increase in ER-driven transcription of tumor suppressors such as ZBTB16 (144). These findings were recently confirmed by Padrão and colleagues (145), who also showed that GR activation can enhance the efficacy of tamoxifen in ER+ breast cancer models. Intriguingly ligand binding of the GR by synthetic GR modulators with either antagonist or partial agonist activity can also suppress tumor growth, replicating the effects of GR liganded by a glucocorticoid agonist in ER+ breast cancer models (146). However, in PI3K-mutant PTEN-null breast cancer cells, dexamethasone reduces tumor growth and mifepristone rescues this effect (147). In contrast to luminal A breast cancers, luminal B tumors do not seem to be sensitive to GR activity (144). The role of the GR in lobular breast cancers seems to mirror ER+ ductal carcinomas, yet GR activation results in an enhanced metastatic phenotype while slowing the tumor growth rate. GR overexpression in SUM44 xenografts suppresses primary mammary tumor growth and paradoxically increases bone metastasis (148). Together, these data show that the directionality of GR modulation is highly context-dependent in ER+ breast cancer; therapeutic development should proceed cautiously with a clear understanding of tumor biology to ensure an appropriate benefit–risk profile. GR Signaling in Immuno-oncology Immune surveillance is an important mechanism to control tumor growth and metastasis through coordinated innate and adaptive immune responses (149). However, tumors frequently develop mechanisms to evade or suppress antitumor immunity. Glucocorticoids suppress immune responses (51), including T-cell proliferation and function, and interferon gamma (IFNγ) and tumor necrosis factor alfa (TNFα) release (150–154). Increased glucocorticoid levels alter the cytokine milieu from a proinflammatory to an anti-inflammatory profile, leading to the reprogramming of protective type 1 T-helper (Th1) cells toward anti-inflammatory type 2 T-helper (Th2) cells and immunosuppressive regulatory T cells (Treg; refs. 154–156). These effects of endogenous and synthetic glucocorticoids interfere with immune surveillance and the elimination of cells containing non–self-antigens, including cancer cells (27, 51). Immune checkpoint inhibitors target key immune regulators that tumors use to evade immunosurveillance and have approved indications for many tumor types (Fig. 2; refs. 157–161). Programmed cell death protein 1 (PD-1) is an immune checkpoint that binds to its ligands, PD-1 ligand 1 (PD-L1) and PD-1 ligand 2 (PD-L2), and thereby suppresses T cell–mediated antitumor responses, increasing cancer growth and metastases (162). Although blockade of PD-1 signaling has transformed the care of many patients, a majority of patients have a suboptimal response or no clinical benefit (163–165). Effects of glucocorticoids on antitumor immune responses. TCR-mediated activation of immune responses forms a crucial aspect of antitumor immunity. DCs are antigen-presenting cells that contribute to the activation of CD4+ and CD8+ T cells through interactions between MHC class molecules and TCR and costimulatory molecules CD28, CD80/86, CD40, and CD40L. Glucocorticoids inhibit DC maturation and costimulation of T cells via decreased expression of MHC classes I and II, CD28 and CD80/86 molecules, interference with TCR signaling, and reduced CD40-mediated stimulatory potential (51, 154, 166). DCs stimulate Th cell antitumor responses, whereas Treg cells have a protumor effect via release of IL10 and TGFβ. Glucocorticoids prevent Th1 differentiation and direct the T-cell response toward a Th2 phenotype (potentially through downregulation of IL12 and upregulation of IL4 and IL10 cytokines) and promote Treg activity (51, 154, 166). Glucocorticoids suppress T-cell function by inhibiting the production of proinflammatory cytokines such as IL2, TNFα, and IFNγ (166). Activated CD8+ T cells and NK cells contribute to tumor cell death via granzymes and perforin; glucocorticoids inhibit both pathways (166, 167). M1 macrophages are responsible for tumor cell clearance, whereas M2 macrophages exert a tumor-promoting effect. Glucocorticoids inhibit M1 polarization of macrophages and promote switching to the M2 phenotype, thereby decreasing proinflammatory cytokines (e.g., IL6, IL12, and TNFα) and increasing anti-inflammatory cytokines (e.g., IL10 and TGFβ; ref. 166). Immune checkpoint proteins like PD-1 and its ligand PD-L1 play a key role in evading antitumor immune mechanisms, and drugs inhibiting this interaction have approved indications for many tumor types (157–161). Glucocorticoids may induce PD-1 and PD-L1 expression, thereby inhibiting cytotoxic T-cell activation (166). CTLA-4, cytotoxic T cell–associated protein-4; DC, dendritic cell; MHC, major histocompatibility complex; NK, natural killer; TCR, T-cell receptor; TGFβ, transforming growth factor β; Th1/2, T helper 1/2. In syngeneic tumors grown in immunocompetent mice, dexamethasone accelerates tumor growth (168), and in an autochthonous pancreatic cancer mouse model, corticosterone (the murine equivalent of cortisol) abolishes immunotherapy responses (169). In a prospective observational clinical study, dysregulated endogenous cortisol was reported to blunt the response to anti–PD(L)-1 therapy (86). Higher peak doses of glucocorticoids used to manage emergent immune-related adverse events are associated with worse outcomes in patients with a range of solid tumor indications (170). In addition, the use of glucocorticoids prior to the start of immune checkpoint inhibitor therapy or early in the first cycle is associated with worse outcomes in NSCLC (171–173). In tumor cells, inhibition of GR increases expression of MHC-I, enhancing tumor antigen presentation and leading to increased infiltration of CD8+ T cells (174, 175). Thus, targeting the GR has the potential to turn “cold” (immune-desert) tumors “hot” (immune-inflamed), extending the efficacy of anticancer immunotherapies to additional tumor types that have shown suboptimal benefit to date. The selective knockout of NR3C1 in T-effector cells in both hot and cold syngeneic models leads to an increase in effector function, increased tumor growth inhibition, loss of T-cell exhaustion markers (LAG3, TIM3, TIGIT, and PD-1) but no change in T-cell infiltration (45). These data show that GR antagonism has the potential to reactivate T cells in tumors following the failure of immune checkpoint therapy. GR deletion in FOXP3+ Tregs reduces the suppressive phenotype of tumor-infiltrating Treg cells and leads to tumor growth inhibition in syngeneic models (46). Although not yet demonstrated in the context of solid tumors, GR expression is increased on myeloid-derived suppressor cells in experimental models of immune tolerance. Dexamethasone increases and mifepristone decreases the suppressive activity of myeloid-derived suppressor cells in mixed lymphocyte reactions (176). Dexamethasone also suppresses PD-L1 expression by neutrophils, monocytes, and dendritic cells in the tumor microenvironment of breast cancer allograft models (145). Moreover, in animal models, tumor-associated monocytes and macrophages have been shown to express CYP11A1 or 11β-HSD1, supporting their role as a source of glucocorticoids in the tumor microenvironment (45, 48). Thus, GR antagonism may also have utility in tumors that have a suppressive immune milieu. If agonism of the GR reduces the efficacy of anticancer immunotherapy, it is rational to consider that selective antagonism of the GR may enhance the efficacy of anti–PD(L)-1 antibodies. Stress-elevated endogenous glucocorticoids blocked IFN responses in antigen-presenting cells and IFNγ T-cell activation in nonclinical models (82). Administration of a GR antagonist reversed the adverse therapeutic impact of stress and restored therapeutic control. In nonclinical ex vivo experiments, SGRAs reversed the immunosuppressive effects of corticosteroids on proinflammatory cytokine release in human peripheral blood mononuclear cells (53). The combination of an SGRA and anti–PD-1 reduced tumor growth compared with the anti–PD-1 antibody alone in a mouse syngeneic tumor model (53). SGRA effects were associated with increased cytotoxic CD8+ T-cell responses against tumor antigens and elevated expression of proinflammatory cytokines such as TNFα, concomitant with decreased levels of anti-inflammatory cytokines, such as interleukin 10 (IL10). In patients with cancer, GR antagonism decreases circulating IL8 (also known as C-X-C motif chemokine ligand 8), a neutrophil chemotactic factor that exerts direct protumorigenic effects, including the promotion of angiogenesis, epithelial-to-mesenchymal transition, and invasion and/or metastasis (177, 178). In prostate cancer models, dexamethasone is associated with an increase in infiltrating monocytes/macrophages that express IL8 (179). This effect can be reversed with mifepristone treatment. GR antagonism also decreases the circulating neutrophil-to-lymphocyte ratio (Data on file, Corcept Therapeutics Inc.; refs. 53, 95). Lower IL8 and a lower neutrophil-to-lymphocyte ratio are both associated with improved outcomes in patients treated with anti–PD(L)-1 therapy (180). Neutrophil-mediated changes in the tumor microenvironment secondary to chronic stress and increased cortisol signaling also enhance tumor metastasis in nonclinical models, reinforcing a direct link between IL8, neutrophil biology, cancer immunotherapy, and GR activation (95). These data suggest that GR antagonism can reverse the immunosuppressive and protumorigenic effects of glucocorticoids and potentiate anticancer responses to anti–PD(L)-1 immunotherapies. However, glucocorticoids are a mainstay of treatment for patients with immune-related adverse events, which occur in approximately half of all patients receiving an anti–PD(L)-1 agent, with ∼15% of patients requiring high-dose corticosteroid treatment (181). Therefore, close monitoring will be required when evaluating such combinations to ensure an appropriate benefit–risk profile. Considerations for the Use of Glucocorticoids in Supportive Care To date, no randomized controlled trial has assessed the effect of supportive glucocorticoid use on survival outcomes in patients with solid tumors. Maintaining >80% to 85% of planned dose intensity has been associated with improved outcomes in patients receiving nonadjuvant chemotherapy regimens (182). This indirect evidence supports the use of glucocorticoids to manage treatment-emergent toxicities, avoiding dose reductions and interruptions. However, prolonged or repeated use carries an increased risk of diabetes, dyslipidemia, hypertension, cardiac disease, adrenal insufficiency on withdrawal, muscle atrophy, central obesity, infection risk, autoimmune hematologic conditions, gastrointestinal bleeding, osteoporosis, and neuropsychiatric conditions (183). In addition, iatrogenic glucocorticoids are associated with reduced survival in certain cancer treatment settings (reviewed above). Therefore, wherever possible, steroid use should be minimized. Given their directly opposing mechanisms of action, concurrent administration of glucocorticoids and SGRAs should be avoided. Often glucocorticoids are introduced early in the clinical development of new drugs to manage hypersensitivity reactions as part of a cocktail of supportive care, but the need for steroids is rarely revisited. Steroid sparing chemotherapy, such as nab-paclitaxel, avoids these issues, and it is often possible to safely omit dexamethasone with repeat administration of drugs that are known to cause hypersensitivity (184). Given the benefit–risk implications, the use of glucocorticoids urgently requires further study in randomized controlled trials and appropriate updates to clinical guidelines. Limitations The findings presented here are derived largely from preclinical evidence and may not fully reflect the complexity of human tumor biology. There is limited clinical data on SGRAs, especially data supporting mechanisms other than the apoptotic effects described herein, and data on compounds other than relacorilant, mifepristone, and ORIC-101. More prospective clinical studies are needed to confirm their effects across tumor types and under different treatment contexts. Future Implications and Conclusions The impact of the human stress response on outcomes for patients with cancer has been known for more than 35 years (3), yet a mechanistic understanding of why this occurs and clinical data to support the therapeutic potential of this pathway have been wanting. The development of SGRAs (54, 100) has deepened our translational and clinical understanding of resistance to AR inhibition (135), chemotherapy (37), and targeted therapy (113), yielding therapeutic breakthroughs (106). Furthermore, improvements in OS and PFS for an SGRA in several randomized controlled trials (37, 106) in patients with advanced ovarian cancer have focused attention on the potential to improve patient outcomes more broadly. A greater appreciation for the potential of immunotherapy and the impact of glucocorticoids on the efficacy of immune checkpoint inhibitors (53, 82, 86, 95, 170, 172, 173, 185) begs further exploration. GR coregulators have a complex and diverse role, which has a differential impact on GR signaling depending on the cellular context. Recent studies in glioblastoma models have shown that this may be even more complex, with some tumors directly co-opting the body’s circadian cortisol rhythm to drive growth (186). Predicting the exact tissues and processes that will be affected by a SGRA is therefore challenging but could yield new mechanistic insights and therapeutic opportunities (187). The potential of this fundamental physiologic pathway is deserving of more study to refine and develop mechanistically differentiated SGRAs for future clinical research. Supplementary Material Opposing roles of the glucocorticoid receptor pathway in lymphoid cancers versus solid tumors and its clinical implications Glucocorticoid receptor antagonists assessed in patients with various solid tumors