The Gene of the Ubiquitin-Specific Protease 8 Is Frequently Mutated in Adenomas Causing Cushing's Disease

Abstract

Context:

We have recently reported somatic mutations in the ubiquitin-specific protease USP8 gene in a small series of adenomas of patients with Cushing's disease.

Objective:

To determine the prevalence of USP8 mutations and the genotype-phenotype correlation in a large series of patients diagnosed with Cushing's disease.

Design:

We performed a retrospective, multicentric, genetic analysis of 134 functioning and 11 silent corticotroph adenomas using Sanger sequencing. Biochemical and clinical features were collected and examined within the context of the mutational status of USP8, and new mutations were characterized by functional studies.

Patients:

A total of 145 patients who underwent surgery for an ACTH-producing pituitary adenoma.

Main Outcomes Measures:

Mutational status of USP8. Biochemical and clinical features included sex, age at diagnosis, tumor size, preoperative and postoperative hormonal levels, and comorbidities.

Results:

We found somatic mutations in USP8 in 48 (36%) pituitary adenomas from patients with Cushing's disease but in none of 11 silent corticotropinomas. The prevalence was higher in adults than in pediatric cases (41 vs 17%) and in females than in males (43 vs 17%). Adults having USP8-mutated adenomas were diagnosed at an earlier age than those with wild-type lesions (36 vs 44 y). Mutations were primarily found in adenomas of 10 ± 7 mm and were inversely associated with the development of postoperative adrenal insufficiency. All the mutations affected the residues Ser718 or Pro720, including five new identified alterations. Mutations reduced the interaction between USP8 and 14-3-3 and enhanced USP8 activity. USP8 mutants diminished epidermal growth factor receptor ubiquitination and induced Pomc promoter activity in immortalized AtT-20 corticotropinoma cells.

Conclusions:

USP8 is frequently mutated in adenomas causing Cushing's disease, especially in those from female adult patients diagnosed at a younger age.

Cushing's disease results from uncontrolled ACTH secretion by corticotroph adenomas of the pituitary, resulting in excess cortisol secretion. Thus, patients with Cushing's disease are characterized by clinical features of chronic hypercortisolism, such as central obesity, moon face, diabetes, hypertension, fatigue, easy bruising, depression, and reproductive disorders (1). Cushing's disease is associated with increased morbidity and mortality, mainly due to cardiovascular or cerebrovascular disease and infections (2).

Numerous previous studies attempted to gain insight into the molecular mechanisms underlying the development of Cushing's disease, but only a few rare mutations have been reported (3–8). Recently, an exhaustive exome-wide screening has led us to identify somatic mutations in the ubiquitin-specific protease 8 (USP8; Ensembl: ENSG00000138592) in six of 17 patients (9). This gene codes for a protein with deubiquitinase (DUB) activity that inhibits the lysosomal degradation of epidermal growth factor receptor (EGFR) (10). USP8 is tightly regulated by 14-3-3 proteins (11). Mutated USP8 overrides 14-3-3 control and displays higher DUB activity than the wild type, therefore increasing EGFR stability and enhancing EGFR-induced pro-opiomelacortin (POMC) transcription and ACTH secretion (9).

Here we aim to analyze the prevalence of USP8 mutations in a large series of 145 ACTH-positive pituitary adenomas, and we investigate the genotype-phenotype correlation.

Patients and Methods

Patients and samples

Pituitary tumors from 145 patients were collected in seven different centers in Europe, Brazil, and the United States during 1998–2013. Written informed consent was obtained from all the patients or, when needed, from their parents. All clinical investigations were conducted according to the principles expressed in the Declaration of Helsinki, and the study was approved by the ethics committee of each institution.

The diagnosis of Cushing's disease was made based on the combination of clinical signs and symptoms and results of biochemical tests suggesting hypercortisolism. Clinical features included central obesity, moon face, buffalo hump, muscle weakness, easy bruising, striae, acne, low-impact bone fractures, mood changes, irregular menstruation, infertility, and impotency. Biochemical diagnosis of Cushing's syndrome was based on increased levels of urinary free cortisol, late-night serum or salivary cortisol, and nonsuppressible serum cortisol after 1 mg overnight or 2 mg/d (48 h) dexamethasone test. ACTH dependency was confirmed by levels of basal plasma ACTH >2.2 pmol/L (10 pg/mL), >50% suppression of serum cortisol during an 8-mg high-dose dexamethasone test, and ACTH and cortisol response to CRH. All patients underwent transsphenoidal adenomectomy. The presence of an ACTH-producing tumor was confirmed histologically after resection. Fresh adenoma tissue was immediately frozen under nitrogen. Data collected at the time of surgery included adenoma size, postoperative plasma ACTH, serum cortisol, and 24-hour urinary free cortisol. Adrenal insufficiency was defined as concentrations of morning serum cortisol <5 μg/dL (138 nmol/L). Silent ACTH adenomas were defined as clinically nonfunctional pituitary adenomas with positive immunoreactivity for ACTH on histological examination.

DNA extraction and Sanger sequencing

Genomic DNA was extracted from 122 fresh-frozen adenomas as described previously (9). RNA was extracted from 23 fresh-frozen adenomas using the RNeasy Mini kit (QIAGEN) and converted to cDNA by means of the Moloney murine leukemia virus reverse transcriptase (Invitrogen). DNA was amplified using a GoTaq DNA polymerase (Promega). The primers used for PCR and sequencing are listed in Supplemental Tables 1 and 2. Sanger sequencing of PCR products was performed using the ABI Prism Big Dye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems) on an ABI Prism 3700 DNA Analyzer (Applied Biosystems).

Plasmids, cell culture, and transfection

Plasmids for expressing human USP8 and pRC/CMV-hEGFR were described before (9). Site-directed mutagenesis was performed on wild-type pME-Flag-USP8 using the QuikChange II kit (Agilent Technologies) and the primers listed in Supplemental Table 1. HeLa, COS-7 and mouse AtT-20/D16vF2 corticotropinoma cells were cultured as previously described (9) and transfected using FuGENE6 (Promega), jetPRIME (Polyplus-transfection) or SuperFect (QIAGEN), following the manufacturer's protocol. For stimulation with human epidermal growth factor (EGF) (100 ng/mL; PeproTech), cells were previously grown during 24 hours in serum-deprived medium (0.5% fetal bovine serum).

Interaction and activity assays

COS-7 cell lysates were prepared in 20 mm Tris-HCl (pH 7.4), 100 mm NaCl, 50 mm NaF, 0.5% Nonidet P-40, 1 mm EDTA, 1 mm phenylmethylsulfonyl fluoride 1 μg/mL aprotinin, 1 μg/mL leupeptin, and 1 μg/mL pepstatin A, and the supernatants were collected after centrifugation. Immunoprecipitation and immunoblotting were performed using standard procedures. Anti-Flag (1 μg, clone M2; Sigma-Aldrich) was used for immunoprecipitation. Primary antibodies for immunoblotting were: anti-14-3-3 (2 μg/mL, clone H-8; Santa Cruz Biotechnology), anti-Ubiquitin (5 μg/mL, clone P4D1; Covance), and anti-Flag (4 μg/mL) antibodies. Secondary antibodies were peroxidase-conjugated anti-mouse IgG and anti-rabbit IgG antibodies (GE Healthcare). Blots were detected using ECL Western Blotting Detection Reagents (GE Healthcare) and ImageQuant LAS 4000mini (GE Healthcare). For DUB assays, Flag-tagged USP8 proteins were expressed in COS-7 cells, immunoprecipitated from their lysates using agarose beads conjugated with anti-Flag antibody (anti-FLAG M2 affinity gel; Sigma-Aldrich), and eluted by incubation with 100 μL of PBS containing the Flag peptide (150 μg/mL; Sigma-Aldrich). Purity and concentration of eluted USP8 proteins were assessed by Coomassie staining using purified bovine serum albumin as a standard. Immunopurified USP8 proteins (∼100 ng; ∼40 nm) were incubated in 20 μL of PBS containing 5 mm MgCl₂ and 2 mm dithiothreitol for 1 hour at 37°C with 0.5 μg (∼700 nm) of Lys63-linked ubiquitin oligomers (dimer-heptamer; Boston Biochem). Reaction products were separated by SDS-PAGE and detected by immunoblotting.

Immunofluorescence

HeLa cells were fixed with 4% paraformaldehyde in PBS for 10 minutes on ice, permeabilized in 0.2% Triton X-100 in PBS, and blocked in 5% fetal bovine serum in PBS. Cells were then incubated with rabbit anti-Flag (1 μg/mL, no. F7425; Sigma-Aldrich) and mouse anti-EGFR (5 μg/mL, clone 6F1; MBL) antibodies. Secondary antibodies were Alexa Fluor 488- and 594-conjugated antimouse IgG and antirabbit IgG antibodies (1:1000; Invitrogen). Nuclei were stained with 4′,6-diamidino-2-phenylindole (1 μg/mL; Nacalai Tesque) during incubation with secondary antibodies. Fluorescence images were captured with a laser-scanning confocal microscope (LSM 780; Carl Zeiss). For quantitative measurement, the fluorescence intensity, the area of fluorescent regions in cells and the area of whole cells in confocal images were measured using ImageJ (National Institutes of Health). Images were subjected to the threshold function using the same threshold for all images.

Pomc promoter activity and hormone secretion

Pomc promoter activity was determined using a luciferase reporter plasmid as previously described (9). The transfection efficacy was determined by cotransfection with the RSV-β-gal construct, and results are presented as luciferase:β-galactosidase activity ratio. The empty vectors pME-Flag and pRC/CMV were used as negative controls.

Mouse ACTH was determined by a specific RIA as previously described (9).

Statistical analysis

Continuous variables were compared between groups using Student's t test, the Mann-Whitney U test, or the Kruskal-Wallis test, as appropriate. Measurements are reported as means with SD or median with interquartile range (IQR). Categorical variables were compared using Fisher's exact test. Binary logistic regression in a backward-stepwise fashion was used for multivariate analysis. An exact, two-tailed significance level of P < .05 was considered to be statistically significant. Analysis was performed using the statistical software package SPSS, version 22.0 (IBM SPSS Statistics).

Results

We performed a retrospective study including 134 patients diagnosed with Cushing's disease—105 adults (>18 y old) and 29 pediatric cases—and 11 adult patients with silent corticotroph macroadenomas recruited from seven participating centers. Clinical and hormonal data of the study population are summarized in Supplemental Table 3. Females were more represented than males (98 vs 36 cases, respectively), as widely reported for Cushing's disease (1). We included a similar number of microadenomas and macroadenomas in this study (69 and 65, respectively).

Mutations in USP8 are commonly found in adenomas causing Cushing's disease

We have searched for somatic mutations in the complete coding region of USP8 in 19 cases and analyzed the exon 14 of USP8 in another 126 samples, including the 11 silent adenomas. The prevalence of mutations in the overall cohort was 35.8%, but it varied depending on age and sex (Table 1), being much more frequent in adults than in pediatric cases (41 vs 17%; P = .027) and in females than in males (43 vs 17%; P = .005). When accounting only for adult patients, mutations were associated with a younger age at diagnosis (36 ± 10 vs 44 ± 13 y old in those with wild-type tumors; P = .001). Sex-specific prevalence was also related to the age (Figure 1A). USP8 mutations were uncommon in female pediatric patients (12%) but were detected in 49% of adult women (40 of 82 cases), who were also younger than those with wild-type adenomas (36 ± 10 vs 42 ± 14 y old; P = .01). No alterations in USP8 were found in any of the silent corticotropinomas.

Table 1.

Clinical Features in Patients With Wild-Type Versus USP8-Mutated Adenomas

	Wild Type	Mutated	P Value
Patients, n (%)	86 (63.4)	48 (35.8)
Pediatric cases	24 (82.8)	5 (17.2)	.03
Adult cases	62 (59.0)	43 (41.0)
Age (mean, SD), y
Age at diagnosis	35.7, 17.4	33.75, 1.7	.49
Age of pediatric cases	14.6, 2.2	15.2, 2.6	.62
Age of adult cases	43.8, 13.3	35.9, 10.3	.001
Sex, n (%)
Males	30 (83.3)	6 (16.7)	.005
Females	56 (57.1)	42 (42.9)
Maximum adenoma size (median, IQR), mm	8.5, 12.0	10.0, 8.0	.32
Microadenomas, n (%)	46 (65.2)	23 (34.8)
Size (median, IQR), mm	6.0, 3.0	8.0, 3.0	.01
Macroadenomas, n (%)	40 (61.5)	25 (38.5)
Size (median, IQR), mm	18.0, 13.0	16.0, 9	.09
Body mass index (mean, SD), kg/m2	30, 6.6	32.7, 6.6	.04

Figure 1.

Age, sex, and tumor size are influenced by USP8 mutations. A, Histograms representing the time and sex distribution of cases with wild-type (gray) or mutated (white) USP8. The gray dashed line represents the age of 18. Each bar represents 5 years. B, Dot plot representing the distribution of tumor size in male (left panel) and female (right panel) patients without and with USP8 mutations. Bars represent the median tumor size with the interquartile range, and the asterisk indicates significant difference (P < .05).

Maximum adenoma size was not significantly different between USP8-mutant and wild-type adenomas (Table 1); however, when analyzing microadenomas separately, mutated samples were significantly larger (P = .013), whereas when analyzing macroadenomas only, mutated cases tended to be smaller (P = .089). In female patients, tumors with USP8 mutations were slightly larger than wild-type (maximum diameter median, 8 vs 10 mm, respectively; IQR, 9 vs 7, respectively; P = .048) (Figure 1B).

Concerning hormonal parameters, preoperative levels of cortisol after a high-dose (8 mg) dexamethasone test were significantly lower in patients with USP8 mutations (P = .01), whereas differences in basal plasma ACTH, basal serum cortisol, or cortisol levels after overnight 1 mg and low (2 mg) dose dexamethasone test did not reach significance (Table 2). Patients with wild-type adenomas were more likely to develop adrenal insufficiency postoperatively than those with USP8 mutations (71 vs 49%, respectively; P = .026), independently of other factors (P = .028 corrected age and tumor size). These differences in adrenal insufficiency were more evident in younger patients (below the age of 36 y, 82 vs 49%; P = .014); similarly, postoperative urinary free cortisol was significantly higher in patients with USP8-mutated adenomas (P = .007).

Table 2.

Hormonal Status in Patients With Wild-Type Versus USP8-Mutated Adenomas

	Wild Type	Mutated	P Value
Preoperative variables
Basal plasma ACTH, pg/mL	74.0, 67.7	67.0, 57.0	.76
Basal serum cortisol, μg/dL	24.1, 14.8	21.6, 9.2	.41
Urinary free cortisol, μg/24 h	379.6, 415.0	370.0, 490.1	.62
Serum cortisol after 1/2 mg DMX, μg/dL	14.7, 14.1	17.2, 16.1	.60
Serum cortisol after 8 mg DMX, μg/dL	5.2, 6.75	2.5, 2.5	.01
Postoperative variables
Basal levels of plasma ACTH, pg/mL	8.3, 12.3	14.0, 30	.12
Minimum serum cortisol, μg/dL	2.5, 7	3.3, 7.9	.72
Urinary free cortisol, μg/24 h	2.5, 6.0	22.5, 241.3	.007
Adrenal insufficiency, n (%)
No	19 (29.2)	21 (51.2)	.03
Yes	46 (70.8)	20 (48.8)

Abbreviation: DMX, dexamethasone. Data are expressed as median (IQR) unless described otherwise.

The mutations identified in this study are shown in Figure 2A and Supplemental Figure 1, whereas their relative prevalence is reported in Supplemental Table 4. In concordance with our previous results, all mutations were heterozygous, were located in the region of USP8 coding for the 14-3-3 binding motif, and were not found in the germline. In particular, 52% of all changes targeted the residue of Ser718, whereas 48% affected the residue Pro720. We identified five new mutations: the substitution Pro720Gln (P720Q) in two unrelated patients and four short in-frame deletions of 6–18 nucleotides (named DEL1-DEL4) in four patients. No somatic mutation was detected in any other exon of USP8.

Figure 2.

The new mutations in USP8 alter 14-3-3 interaction, USP8 cleavage, and DUB activity. A, Diagram summarizing the different somatic mutations. All the mutations found in corticotroph adenomas clustered into a hotspot region overlapping with the 14-3-3 binding motif. The position of the hotspot is shown in the context of the complete protein. MIT, microtubule-interacting and trafficking domain; RHOD, rhodanese-like domain; SBM, SH3-binding motif; DUB, deubiquitinase catalytic domain. B, The new mutations impair 14-3-3 binding. Immunoblots with anti-14-3-3 (top) and anti-Flag (middle) after anti-Flag immunoprecipitation of COS-7 cell lysates expressing the different Flag-tagged USP8 mutants. Bottom, endogenous 14-3-3 expression in the total lysates. C, The new mutations enhance USP8 cleavage. Flag-tagged USP8 proteins expressed in COS-7 cells were immunoprecipitated with anti-Flag antibody and eluted from the antibody with a Flag peptide. Empty arrow, full-length protein (130 kDa); black arrow, cleaved form (90 kDa). D, DUB activity is also enhanced in the new mutants. Immunopurified Flag-USP8 proteins were incubated with Lys63-linked ubiquitin oligomers (dimer-heptamer) for 1 hour at 37°C. Wild-type USP8 (WT) and an inactive form of USP8 (CA) were used as controls.

In summary, mutations of USP8 were frequent in corticotroph adenomas, with a higher prevalence in young adult females, and their presence was inversely associated with the development of postoperative adrenal insufficiency.

Functional analysis of the new USP8 mutations

USP8 is catalytically repressed by members of the 14-3-3 family, universal regulators that participate in multiple cellular processes (12); 14-3-3 binds to USP8 through a residue of phosphoserine (in human, pSer718) included into the consensus motif RSYpSSP, and this interaction inhibits USP8 activity (11). We have reported that mutations in Ser718 or Pro720 disrupt the interaction of USP8 with14-3-3 and increase USP8 cleavage and its DUB activity. To determine the effect of the five new mutations on USP8 regulation, we generated the corresponding constructions by site-directed mutagenesis and examined their ability to bind endogenous 14-3-3 by coimmunoprecipitation from lysates of transfected COS-7 cells, showing that 14-3-3 binding is severely affected in all the new mutants (Figure 2B). This demonstrated increased susceptibility to protein cleavage as evidenced by the presence of a minor band of approximately 90 kDa corresponding to the N-terminal fragment of USP8 (N90) in immunopurified mutants but not in the wild-type control (Figure 2C and Supplemental Figure 2); this finding supported our previous observation that mutants unable to bind 14-3-3 are susceptible to be cleaved (9). Because this cleavage increases DUB activity, wild-type USP8 and the different mutants were immunopurified and incubated with Lys63-linked ubiquitin chains. As shown in Figure 2D, the mutants hydrolyzed ubiquitin chains more efficiently than the wild-type USP8.

USP8 deubiquitinates EGFR and counteracts its ubiquitination-mediated lysosomal degradation; thus, we postulated that the new mutants will increase EGFR stability. To evaluate this premise, HeLa cells were transfected with wild-type USP8 or any of the mutants, and the subcellular location of endogenous EGFR was examined by immunofluorescence (Figure 3, A and B). As expected, EGFR was internalized and transported to the endosome after 1 hour of EGF stimulation. Expression of wild-type USP8 had a mild effect, but transfection of any of the mutants significantly diminished EGFR lysosomal trafficking and elevated its cell surface level.

Figure 3.

The new mutations enhance EGFR stability and ACTH synthesis. A, USP8 mutants hampered EGFR transport to lysosomes. Immunofluorescent staining of HeLa cells transfected with the new Flag-tagged USP8 mutants and stimulated 1 hour with EGF. The cells were triple-stained with anti-Flag antibody (magenta) to detect USP8, anti-EGFR antibody (green), and 4′,6-diamidino-2-phenylindole (blue). Asterisks indicate cells not expressing Flag-USP8 proteins. Scale bars, 10 μm. B, Proportion of EGFR-positive area per cell in three cells in the confocal images from panel A, calculated by measuring the area of EGFR-positive region (green) in a cell and the area of the whole cell. C, Pomc promoter activity in AtT-20 cells cotransfected with EGFR and USP8 wild type or mutants. Data are ratio of luciferase/β-galactosidase activities, means of three experiments with each transfection condition in triplicates and presented as a percentage of EGFR + empty vector control (pME-Flag; mock). RLA, relative luciferase activity. *, P < .01; **, P < .001 compared to EGFR+mock; #, P < .05 to EGFR+WT (t test and Mann-Whitney U test). Error bars, standard deviation. D, ACTH secretion in AtT-20 cells cotransfected with EGFR and USP8 wild type or mutants as determined with a specific RIA. Data are represented as the ratios of ACTH (pg/mL) to cell viability values (WST-1 colorimetric assay; OD, 450 nm) and shown as percentage of EGFR+empty vector control (pME-Flag; mock). The graph shows the means of three measurements, with each transfection condition in triplicate. *, P < .01; **, P < .001 compared to EGFR+mock; #, P < .05 to EGFR+WT (t test and Mann-Whitney U test). Error bars, standard deviation.

In corticotroph cells, EGFR activation induces POMC transcription and ACTH secretion (13). In murine AtT-20 corticotropinoma cells overexpressing the human EGFR, all USP8 mutants increased Pomc promoter activity compared to empty vector control (P < .05; Figure 3C). Similarly, the USP8 mutants increased ACTH secretion compared to the wild-type protein (P < .05; Figure 3D).

Discussion

Several previous studies have addressed dysregulated methylation and gene expression patterns in different types of pituitary tumors (14), but the search for somatic mutations of candidate genes has been unsuccessful in most cases. Heterozygous germline mutations in certain genes, such as MEN1, AIP, PRKAR1A, X chromosome microduplication (15), and very rarely mutations in DICER1 and SDH predispose to the development of pituitary adenomas (16), whereas somatic mutations have been identified in the GNAS1 gene in somatotroph adenomas (17) and in USP8 mutation in a subset of patients with Cushing's disease (9).

Here we report the prevalence of somatic USP8 mutations in a representative cohort of patients with Cushing's disease. In our series, around one-third of the tumors harbored a USP8 mutation, but the prevalence rose to 40% in the adult patient population and to almost 50% among adult females. The absence of USP8 mutations in other types of pituitary tumors (9), including silent ACTH-producing adenomas, suggests that these alterations are specific traits of secretory corticotroph adenomas causing Cushing's disease.

In agreement with our initial report, mutations were principally found in female patients, and within the adult population, patients with mutated adenomas were younger. Although speculative, a potential growth-stimulating effect of estrogens on USP8 mutant corticotroph cells in the development of the disease could explain this observation (18). For example, a balanced sex ratio has been reported in unselected pediatric patients (19), with a male predominance under the age of 10 years (20) that turns into a female predominance during adolescence (21, 22). Interestingly, corticotroph cells express estrogen receptors (23, 24), and, at least in vitro, estradiol can stimulate corticotroph proliferation, an effect that is mediated through EGFR signaling (25). However, further efforts are needed to determine properly the factors causing age- and sex-related distributions of USP8 mutations in patients with Cushing's disease.

Tumor size is generally associated with clinical remission; patients with microadenomas usually exhibit higher rates of remission than those with larger tumors; on the other hand, those patients presenting with Cushing's disease but no visible lesions on imaging techniques have lower remission rates than patients with visible microadenomas (26–31). In our cohort, histologically proven corticotroph adenomas not visible on preoperative imaging and large macroadenomas were both primarily wild type, whereas >50% of USP8-mutated adenomas had a size ranging between 8 and 16 mm. In comparison to our previous data on 17 adenomas (all ≥5 mm), in this larger cohort, we found mutated adenomas from female patients being larger than their wild-type counterpart. In our previous study, males and females were analyzed together; when female cases were selected (Ref. 9, Supplemental Table 1), tumors containing USP8 mutations seemed to be larger than those wild type (median maximum diameter, 7.5 vs 6.0 mm, respectively; n = 11; P = .69).

A limitation of this multicentric study is the lack of a complete record of hormonal results as well as the lack of uniformity of the diagnostic protocols and postoperative management; thus, the exact impact of USP8 status on the outcome is difficult to assess. Postoperatively, urinary cortisol levels were higher in the mutated group, although the number of cases with available data was small (n = 30). Consistent with this finding, we have observed an inverse association between the presence of USP8 mutations and the development of adrenal insufficiency, which was more evident in younger patients. Postoperative hypocortisolism is an established parameter of long-term remission (31–33), although long-term follow-up is needed to detect late recurrences (34, 35). The current data, if confirmed by others, may indicate a worse outcome of transsphenoidal surgery in patients with USP8-mutant adenomas.

All USP8 somatic mutations detected in patients with Cushing's disease so far, including newly identified ones investigated in this study, are located at a single hotspot region of the USP8 gene and disrupt the inhibitory effect of 14-3-3. This results in the aberrant activation of USP8 and its subsequent cleavage via a still unknown mechanism. One of the consequences of this unregulated activity is the impairment of EGFR ubiquitination and turnover and its subsequent accumulation in the plasma membrane, where it remains active. EGFR is highly expressed in human corticotropinomas, where it triggers POMC transcription and ACTH synthesis (13, 36). The high prevalence of mutations on USP8 in patients with Cushing's disease, but not in any other pituitary adenomas, provides a plausible explanation for the dependence of corticotropinomas to EGFR signaling and suggests that the mutational status of USP8 can be used to stratify the patients for targeted therapies against EGFR (13, 37). This finding therefore highlights the possibility of individualized therapeutic approaches for patients with USP8-mutated corticotroph adenomas who remain clinically active despite surgical therapy.

Conclusion

We have found that USP8 mutations are common in adenomas causing Cushing's disease, more frequently in adult females and relatively less often in male patients and pediatric cases. Mutations associate with larger tumors in female patients and might be related to a worse postsurgical outcome in both sexes. They promote an abnormal USP8 activation that leads to a sustained ACTH synthesis. In summary, the present study emphasizes the role of USP8 in Cushing's disease.