Tumor-Directed Therapeutic Targets in Cushing Disease

Abstract

Context

The most frequent cause of endogenous hypercortisolism is Cushing disease (CD), a devastating condition associated with severe comorbidities and high mortality. Effective tumor-targeting therapeutics are limited.

Design

Search in PubMed with key words “corticotroph” and “Cushing’s disease” plus the name of the mentioned therapeutic agent and in associated references of the obtained papers. Additionally, potential therapeutics were obtained from ClinicalTrials.gov with a search for “Cushing disease.”

Results

At present, the tumor-targeted pharmacological therapy of CD is concentrated on dopamine agonists (cabergoline) and somatostatin analogs (pasireotide) with varying efficacy, escape from response, and considerable side effects. Preclinical studies on corticotroph pathophysiology have brought forward potential drugs such as retinoic acid, silibinin, and roscovitine, whose efficacy and safety remain to be determined.

Conclusions

For many patients with CD, effective tumor-targeted pharmacological therapy is still lacking. Coordinated efforts are pivotal in establishing efficacy and safety of novel therapeutics in this rare but devastating disease.

The most frequent cause of endogenous hypercortisolism is ACTH hypersecretion due to a corticotroph tumor, known as Cushing disease (CD) (1). It is a devastating condition associated with severe comorbidities, including metabolic syndrome, musculoskeletal abnormalities, cardiovascular morbidity, increased risk of infections, fatigue and depression, and high mortality (2–5). The first line of therapy remains pituitary surgery, and although complication rates have decreased over the years, biochemical remission varies from 79%, in the case of microadenomas (<10 mm), down to 40% in macroadenomas (6–11). In severe disease, bilateral adrenalectomy is indicated, with the downside of life-long hydrocortisone replacement and the potential risk of progressive corticotroph tumor growth (Nelson syndrome) (12, 13).

Pharmacological therapy aims at managing the excess cortisol production centrally, at the adrenal level or the response at tissue targets (14, 15). At present, most approved pharmaceuticals target hypercortisolism at the periphery (i.e., extrapituitary level) either in the form of steroidogenesis inhibitors that inhibit cortisol production from the adrenals (ketoconazole, mitotane, metyrapone) or a blocker of glucocorticoid action at target tissues in the periphery [glucocorticoid receptor (GR) type II antagonist mifepristone] (15, 16). Investigational steroidogenesis inhibitors such as the 11β-hydroxylase inhibitor LCI699 (osilodrostat) and the ketoconazole enantiomer levoketoconazole (COR-003) are currently in clinical trials (ClinicalTrials.gov nos. NCT02468193, NCT02697734, and NCT03621280) (17). Use of ACTH neutralizing antibodies (ALD1613) to block the ACTH action at the adrenals is another promising therapeutic strategy, currently at the experimental stage (18).

These periphery-targeting drugs have demonstrated high efficacy in lowering and normalizing cortisol levels, but they cannot cure the cause of the disease, which is the corticotroph tumor. The effort to identify pharmaceuticals able to target the corticotroph tumor and suppress ACTH secretion as well as effectively shrink tumor mass in the case of macroadenomas is ongoing. Pharmaceuticals commonly used in pituitary disease, such as dopamine agonists and somatostatin analogs, have given mixed results that have been expertly covered in recent reviews and are therefore only briefly mentioned herein (15, 19–22). Instead, this review focuses on tumor-directed pharmacological agents, emphasizing the rationale that resulted in their testing in CD.

Cabergoline

Cabergoline is a D2 dopamine receptor (D2DR) agonist and a mainstay treatment of hyperprolactinemia (23). Corticotroph tumors express D2DR in almost 80% of cases (24), and dopamine agonists lower ACTH secretion in some cases in vitro (25). In contrast, earlier studies showed no suppressive effect of dopamine agonists on ACTH secretion from normal and tumoral corticotrophs in vitro (26, 27). With few exceptions, further studies on the effects of bromocriptine administration in patients with CD (including cases after bilateral adrenalectomy) reported lack of biochemical control in most cases (28–33). Cabergoline normalized 24-hour urinary free cortisol (UFC) in <40% of patients (34, 35) and 0% to 5% in a prospective study (36). Notably in several cases, the doses needed to achieve this effect are in the range of 1 to 7 mg/wk, which is higher than what is administered in patients with hyperprolactinemia (15). Furthermore, even in patients who initially showed a clinical and biochemical response, subsequent follow-up studies showed no sustained control of hypercortisolism (20, 25, 34). Nevertheless, individual reports have suggested that cabergoline could be a safe therapeutic option in patients with CD who are pregnant or desire to get pregnant (37–39).

Pasireotide

Pasireotide is a multiligand somatostatin analog able to bind somatostatin receptors (SSTRs) 1, 2, 3, and 5 (40). The high affinity for SSTR5, which is the predominant SSTR in corticotroph tumors cells and the preclinical evidence of a potent anti-ACTH action in vitro and in vivo, led the foundations for pasireotide to become the first approved tumor-targeted pharmacological treatment of CD (41–43).

Subcutaneous pasireotide treatment of 12 months was shown to normalize UFC in up to 25% of patients and decrease tumor volume (44–48). Additionally, it improved clinical signs such as body weight, blood pressure, and total cholesterol levels (48, 49). The first reports from the long-acting release pasireotide showed efficacy (i.e., 24-hour UFC normalization) in ∼40% of patients (50). Note that patients with moderate UFC elevation (UFC 1.5- to 5-fold the upper limit of normal) were selected as part of this study, and therefore long-acting pasireotide treatment may be more efficacious in patients with moderate CD. Accordingly, UFC normalization was highest in the group of patients with very mildly elevated UFC (1.5- to 2.0-fold the upper limit of normal). Pasireotide was also effective in reducing ACTH and hyperpigmentation in a patient with Nelson syndrome (51). Long-term (24-month) studies show that successful response is evident within the first few months of treatment and, when present, the beneficial effects of pasireotide usually persist and escape from response is unlikely (46, 48).

A very important and serious side effect of pasireotide treatment is the occurrence of hyperglycemia of various degrees in most patients, which demands attention and concomitant diabetes management (46, 48, 50).

Combined Cabergoline/Pasireotide Treatment and Chimeric Dopamine/Somatostatin Compounds

A recent meta-analysis showed that pasireotide achieved biochemical control in 41% of the patients, closely followed by cabergoline (35.7%) (52). The concomitant expression of SSTR5 and D2DR in corticotroph tumor cells suggested that combined somatostatin/dopamine agonist treatment might be more efficacious. Addition of cabergoline to pasireotide normalized UFC in a subset of patients; however, almost half of the cases required additional treatment with ketoconazole to achieve biochemical remission (45). Dopastatin, a chimeric compound able to bind to both SSTRs and D2DR, demonstrated similar antisecretory efficacy in human corticotroph tumors in vitro (53). However, in human subjects repeated dopastatin administration causes the accumulation of highly active dopaminergic metabolites that eventually block its action (54). Dopastatin’s clinical development has been halted, but second-generation chimeras are currently under development (54).

Retinoic Acid

Retinoic acid is a vitamin A derivative that binds to receptors of the nuclear receptor subfamily 1 (RAR) and 2 (RXR). Attention to retinoic acid was brought from in vitro studies on the transcriptional regulation of the ACTH precursor proopiomelanocortin (POMC). Retinoic acid potently suppressed ACTH, corticosterone, and tumor size in vitro and in vivo and was able to reverse experimental CD in xenograft animal models (55). Furthermore, treatment of 6 months effectively reduced cortisol levels and ameliorated symptoms of CD in inflicted dogs, where it successfully caused tumor shrinkage and prolonged survival, contrary to the control group that was treated with ketoconazole (56).

A prospective proof-of-concept study on seven human patients with CD showed that all-trans retinoic acid (tretinoin) treatment of 6 months decreased UFC by >50% in five patients. This effect persisted during the 12-month study period in three cases that achieved biochemical normalization (57). A subsequent study on a larger cohort of 16 patients with CD with persistent disease after transsphenoidal surgery using 13-cis-retinoic acid (isotretinoin) observed normalization of UFC in six patients after 6 months of treatment (58). In both studies, successful biochemical response was accompanied by improvement of clinical features, including reduction in body weight, blood pressure, and glycemic control. Side effects were observed as expected from retinoic acid treatment, namely photosensitivity, conjunctival irritation, mucositis, arthralgia, and dyslipidemia, and were usually mild and transient. Nevertheless, the teratogenic nature of retinoic acid demands special attention in the case of female patients of reproductive age.

Most responding patients showed rebound after halting the treatment, and in the two studies performed so far, loss of response was observed in two out of five and two out of six patients, respectively (49, 50). In vitro data showed that combining cis retinoic acid with bromocriptine suppresses ACTH synthesis from human CD tumors in primary cell culture more effectively than does each treatment alone, suggesting that concomitant retinoic acid and dopamine agonist administration may be more efficacious in CD (59).

It is noteworthy that retinoic acid does not affect ACTH synthesis in the normal corticotrophs, probably due to the presence of the COUP-TFI factor in the normal but not in the tumoral corticotroph cells (55). As COUP-TFI antagonizes retinoic acid action, its expression could be a putative biomarker of successful response in CD tumors (60).

Thiazolidinediones

Thiazolidinediones (TZDs) are ligands of the peroxisome proliferator–activated receptor γ (PPARγ) that, similar to RAR, belongs to the nuclear receptor subfamily 1 and regulates gene expression often in heterodimers with receptors for retinoic acid. The finding of high PPARγ expression in corticotroph tumors prompted the testing of TZDs in CD models in vitro and in vivo (61). Indeed, the TZD rosiglitazone potently suppressed ACTH synthesis in vitro and corticosterone and tumor volume in mice with corticotroph tumor xenografts in vivo.

In the clinic, acute rosiglitazone administration in 10 patients with CD had no effect on basal and CRH-induced ACTH and cortisol levels (62). Overall, administration of rosiglitazone and pioglitazone (45 mg/d, 30 days) in patients with CD failed to improve biochemical parameters and clinical features (63–65). Two investigations showed that short-term treatment (3 to 12 weeks) with rosiglitazone could reduce ACTH and/or normalize UFC in 2 and 6 (out of 14) patients with CD (63, 66). In the second study, although UFC normalization was not always accompanied by a significant reduction in serum ACTH levels, a sustainable effect was reported in two patients after 7 months of treatment (63). In patients with Nelson syndrome (16 in total), high-dose (8 to 12 mg/d) rosiglitazone administration was not able to normalize ACTH for the duration of the treatment (8 to 12 weeks, up to 12 months) (67–69).

The discrepancy between the preclinical and clinical studies can be attributed to the dosology: in human patients, rosiglitazone was used at the maximum doses accepted at the time (i.e., 8 to 12 mg/d), which are much lower than the 150 mg/kg/d used in the animal studies (68, 69).

Notably, both PPARγ and RAR/RXR are antagonized by testicular receptor 4 (TR4), a member of the nuclear receptor subfamily 2 that is overexpressed in CD tumors (70). TR4 compromises the inhibitory action of GRs on the POMC promoter, thereby contributing to glucocorticoid resistance and inducing ACTH synthesis (70, 71). The antagonistic action of TR4 on RAR/RXR and PPARγ may be responsible for the low efficacy of retinoic acid and especially TZD in patients with CD (72, 73).

Silibinin

Silibinin is a flavonolignan derived from milk thistle (Silybum marianum) that was recently shown to inhibit heat shock protein 90 (HSP90), a chaperone that regulates the GR (74). Attention to HSP90 was brought from research on targetable mechanisms responsible for the partial resistance to glucocorticoid feedback seen in CD. HSP90 may be essential for the maturation of the GR, but its overexpression compromises GR transcriptional activity (75–77). HSP90 was found to be overexpressed in CD tumors but not in the normal pituitary gland, placing it as a valuable pharmacological target (78).

The HSP90 inhibitors known at the time also caused the degradation of client proteins, an action that would be undesirable for GR (79). The discovery of a novel biochemical mechanism through which an HSP90 inhibitor could enhance glucocorticoid binding to the GR and potentiate its transcriptional activity when it targets specifically the C terminus of the protein paved the way for specific targeted inhibitors (78). Silibinin was identified as a C-terminal inhibitor and effectively sensitized murine and human corticotroph tumor cells to dexamethasone in vitro, suppressed ACTH synthesis and tumor volume, and ameliorated biochemical and clinical signs of CD in xenograft mice in vivo (78). Its favorable safety profile [currently used for the treatment of liver disease and poisoning (80)] and its potential chemotherapeutic action (prostate cancer; see ClinicalTrials.gov no. NCT00487721) suggest silibinin as a promising drug for the management of CD.

Gefitinib

Gefitinib is an oral tyrosine kinase inhibitor with high affinity for epidermal growth factor receptor (EGFR). The EGFR system is of special relevance to corticotroph pathophysiology with EGF being a potent ACTH secretagogue and its receptor abundantly expressed in the most CD tumors (reviewed in Refs. 81–83). Recently, its importance has been consolidated by the discovery that the genetic basis of almost half of CD tumors is activating somatic mutations in the USP8 gene, which encodes a deubiquitinase that rescues EGFR and triggers ACTH synthesis (84; reviewed in Ref. 85).

Gefitinib treatment reduced ACTH synthesis from human corticotroph tumors in vitro and decreased tumor volume and biochemical markers in xenograft mice in vivo (83). Gefitinib has been approved for the treatment of metastatic non–small cell lung cancer with moderate side effects and no chemotherapy-associated effects such as alopecia and myelosuppression; the most severe side effect is increased risk of interstitial lung disease (86). Currently, there is one registered clinical trial for gefitinib (250 mg/d) in patients with CD with USP8 mutated tumors, but with no known outcome (ClinicalTrials.gov no. NCT02484755).

Roscovitine

Roscovitine is an oral competitive inhibitor of cyclin-dependent kinases that triggers cell cycle progression and is specific for CDK1/cyclin B, CDK2/cyclin A and CDK2/cyclin E, and CDK5/p35 (87). The rationale for using CDK inhibitors in CD was provided by the observation that CDK2/cyclin E is specifically deregulated in corticotroph tumors (88, 89).

Roscovitine decreased corticotroph cell viability and tumor growth, and it had a profound inhibitory action on POMC in zebrafish, murine and human corticotroph cells, and ACTH levels in xenograft mice (90, 91). This inhibitory effect on ACTH synthesis is dissociated from the antiproliferative action, but it is in fact due to a direct regulatory role of cyclin E on POMC expression (91). Interestingly, at a slightly higher micromolar range, roscovitine also inhibits extracellular regulated kinases ERK1/2 that mediate the effects of EGFR on POMC (83, 84, 87). Roscovitine is currently in a phase 2 clinical trial in patients with CD (R-roscovitine 400 mg orally twice daily, 4 d/wk for 4 weeks; ClinicalTrials.gov no. NCT02160730).

Other Treatments

The histone deacetylase inhibitor vorinostat (suberoylanilide hydroxamic acid) was shown to induce apoptosis and inhibit ACTH synthesis in murine and human corticotroph tumors in vitro and in xenografts in vivo (92). Vorinostat is approved for the treatment of cutaneous T-cell lymphoma, but its efficacy in CD in a clinical setting remains to be determined.

Studies on the neuroimmunomodulatory regulation of ACTH synthesis provided the rationale of testing INFγ. Recombinant INFγ displayed potent antiproliferative action in corticotroph tumor cells in vitro and inhibited tumoral ACTH biosynthesis without affecting ACTH secretion from dispersed healthy corticotroph cells; the study highlighted a novel inhibitory mechanism on POMC that involves STAT1 and nuclear factor κB (93). Following up, natural compounds with anti-inflammatory properties, such as curcumin and triptolide, were found to suppress cell viability and ACTH synthesis in murine and human corticotroph tumors in vitro (94, 95).

Targeted Treatments for Aggressive CD Tumors

CD tumors are usually benign, but their aggressive variants are common among the extremely rare pituitary carcinomas (96). In these rare cases, owing to their size and aggressive nature, treatment aims to control tumor growth in addition to biochemical normalization. Nevertheless, patients with aggressive corticotroph tumors and carcinomas are notoriously difficult to manage by surgery and other therapeutic means. Administration of chemotherapeutic agents such as 5-fluorouracil, carboplatin, and IFNα in rare cases of patients with corticotroph carcinomas had no effect on tumor volume or ACTH secretion and did not improve morbidities and mortality (97).

The alkylating chemotherapeutic agent temozolomide is able to control tumor growth and provide relief from headaches and other manifestations of tumor size effects in some cases (98–101). Temozolomide was reported to cause a peak in ACTH levels that was controlled by the addition of pasireotide (102). Concomitant or consecutive pasireotide treatment was reported to normalize UFC and, in one case, to cause tumor shrinkage in patients with aggressive CD tumors (103). However, subsequent reports showed that pasireotide treatment is accompanied by escape and even paradoxical responses and its effectiveness in patients with aggressive CD tumors is questioned (101, 104, 105).

Alternative pharmacological regimes tried in patients resistant to temozolomide included combined everolimus and octreotide therapy in one patient with corticotroph carcinoma, which, however, failed to control tumor growth and ACTH levels (106). A combined capecitabine and temozolomide scheme completely regressed or stabilized disease in four patients with aggressive CD tumors (107). A recent study reported a case of an ACTH-secreting carcinoma that acquired resistance to temozolomide and was managed with ipilimumab and nivolumab that suppressed ACTH levels and shrank the hepatic metastasis (108).

Critical View

More than a century after Harvey Cushing described the first patient with the namesake disease, hypophysectomy remains the primary tumor-targeted therapy. Approved tumor-targeted pharmaceuticals fall back in efficacy as shown in a recent meta-analysis in which the two clinically available, cabergoline and pasireotide, achieved biochemical normalization in 35.7% and 41% of patients with CD, respectively (52). Even when effective, loss of response is frequent and patients are burdened by side effects that require additional life-long management. A general word of caution regarding perceived disease control: studies have revealed a high intrapatient variability in UFC levels and lack of clear correlation between UFC and clinical features of CD (109). This hinders the interpretation of the results of the clinical studies and highlights the need for a better definition and improved biomarkers of disease control.

Preclinical research has provided a wealth of information on the mechanisms driving corticotroph tumorigenesis that have brought into the spotlight potential drugs such as retinoic acid, silibinin, and roscovitine. Although caution should be exercised when translating exciting preclinical findings with chemotherapeutics into the usually benign CD tumors, repurposing of drugs with a known safety profile into the population of patients with CD holds promise for the development of new tumor-targeted pharmaceutical approaches. These will need to prove their efficacy and safety profile in coordinated trials and benefit-risk analysis in patients with CD. Therefore, communication and collaboration are essential to translate into meaningful treatments for this rare but devastating disease.

Abbreviations:

Abbreviations:
 
• CD

  Cushing disease

 
• D2DR

  D2 dopamine receptor

 
• EGFR

  epidermal growth factor receptor

 
• GR

  glucocorticoid receptor

 
• HSP90

  heat shock protein 90

 
• POMC

  proopiomelanocortin

 
• PPARγ

  peroxisome proliferator–activated receptor γ

 
• SSTR

  somatostatin receptor

 
• TR4

  testicular receptor 4

 
• TZD

  thiazolidinediones

 
• UFC

  urinary free cortisol

Acknowledgments

The authors thank the anonymous reviewers for their useful suggestions.

Financial Support: This work was supported by Deutsche Forschungsgemeinschaft (DFG) Projektnummer: 314061271 – TRR 205. M.R. is also supported by Else Kröner-Fresenius Stiftung Grants 2012_A103 and 2015_A228.

Disclosure Summary: M.R. has received speaker’s fees from Ipsen, Novartis, and Pfizer and consultancy fees from Novartis. The remaining author has nothing to disclose.

References