Concurrent Mutations of Germline GPR101 and Somatic USP8 in a Pediatric Giant Pituitary ACTH Adenoma

Abstract

Background

Cushing’s disease (CD) is rare in pediatric patients. It is characterized by elevated plasma adrenocorticotropic hormone (ACTH) from pituitary adenomas, with damage to multiple systems and development. In recent years, genetic studies have shed light on the etiology and several mutations have been identified in patients with CD.

Case presentation

A girl presented at the age of 10 years and 9 months with facial plethora, hirsutism and acne. Her vision and eye movements were impaired. A quick weight gain and slow growth were also observed. Physical examination revealed central obesity, moon face, buffalo hump, supra-clavicular fat pads and bruising. Her plasma ACTH level ranged between 118 and 151 pg/ml, and sella enhanced MRI showed a giant pituitary tumor of 51.8 × 29.3 × 14.0 mm. Transsphenoidal pituitary debulk adenomectomy was performed and immunohistochemical staining confirmed an ACTH-secreting adenoma. Genetic analysis identified a novel germline GPR101 (p.G169R) and a somatic USP8 (p. S719del) mutation. They were hypothesized to impact tumor growth and function, respectively.

Conclusions

We reported a rare case of pediatric giant pituitary ACTH adenoma and pointed out that unusual concurrent mutations might contribute to its early onset and large volume.

Peer Review reports

Background

Cushing’s disease (CD) is caused by the overproduction of adrenocorticotropic hormone (ATCH) by pituitary adenomas (PAs). It is rare in children and accounts for approximately 75% of pediatric Cushing’s syndrome from 7 to 17 years of age [1]. Weight gain and facial changes are more common in children than in adults [2]. Growth retardation is also a characteristic of children with hypercortisolemia [3]. Genetic alterations such as somatic USP8RASD1TP53 mutations, and germline AIPMEN1, and CABLES1 mutations have been identified in CD patients [4]. Here we report a case of pediatric invasive pituitary ACTH macroadenoma associated with a novel germline GPR101 (p. G169R) and a somatic USP8 (p. S719del) mutation.

Case presentation

The girl was born at full term with a length of 48 cm and a weight of 2900 g. Her neuromotor and cognitive development was comparable to those of children of the same age. At the age of 9 years and 4 months she developed plethora, hirsutism, facial acne, rapid weight gain, and increased abdominal circumference. Her skin darkened, and purple striae appeared on thighs and in the armpits. She became dull and less talkative, as indicated by her parents. At 10 years and 3 months, the patient complained of pain around the left orbit with an intensity of 4–5 points on a numerical rating scale (NRS). Five months later bilateral blepharoptosis appeared, with significantly impaired vision of the left eye. Soon both eyes failed to rotate in all directions.

On admission the patient was 10 years and 9 months, with a height of 144 cm (90–97th percentile) and a weight of 48 kg (25–50th percentile). Her weight gain was 20 kg, while the height increased by only 2–3 cm in 18 months. Her blood pressure was 115/76mmHg, and her heart rate was 80 bpm. Apart from the signs mentioned above, physical examination revealed central obesity (BMI 23.1 kg/m2), moon face, buffalo hump, supra-clavicular fat pads and bruising at the left fossa cubitalis. Her pupils were 7 mm in diameter and barely reacted to light. There was a fan-shaped visual field defect in the left eye. Her breasts were Tanner stage III and pubic hair was Tanner stage II, although menarche had not yet occurred. The parents and her younger brother at 6 years of age did not have symptoms related to Cushing syndrome, acromegaly or gigantism. There was no family history of pituitary tumor or other endocrine tumors.

She had increased midnight serum cortisol (24.35 µg/dL, normal range < 1.8 µg/mL) and 24-hour urine free cortisol (24hUFC) (308.0 µg, normal range 12.3–103.5). The plasma ACTH level ranged from 118 to 151 pg/mL (< 46pg/mL). The 24hUFC was not suppressed (79.2 µg) after 48 h low-dose dexamethasone suppression test (LDDST), but suppressed to 32.8 µg (suppression rate 89.4%) after 48 h high-dose dexamethasone. Sella enhanced MRI showed a giant pituitary tumor measured 51.8 × 29.3 × 14.0 mm with heterogeneous density (Fig. 1). The mass compressed the optic chiasma and surrounded the bilateral cavernous sinus (Knosp 4). Therefore, an invasive giant pituitary ACTH adenoma was clinically diagnosed. The morning growth hormone (GH) was 1.0ng/ml (< 2 ng/ml) and insulin-like growth factor 1 416 ng/ml (88–452 ng/ml). The prolactin (PRL), luteinizing hormone (LH), follicle-stimulating hormone (FSH) and thyroid stimulating hormone (TSH) were all in normal ranges, as well as serum sodium, potassium, blood glucose and urine osmolality. Abdominal ultrasonography revealed a fatty liver. Tests concerning type 1 multiple endocrine neoplasia included serum calcium, phosphate, parathyroid hormone, gastrin and glucagon, which were all unremarkable (Table 1).

Fig. 1

figure 1

Contrast-enhanced coronal (A) and sagittal (B) T1-weighted MRI on admission. The sellar mass measured 51.8 × 29.3 × 14.0 cm (TD × VD × APD) with a heterogeneous density in the enhanced scan. The diaphragma sellea was dramatically elevated, with optic chiasm compressed. The sellar floor was sunken and bilateral cavernous sinus was surrounded (Knosp 4)

Table 1 Laboratory data on admission

Transsphenoidal pituitary debulk adenomectomy was performed immediately due to multiple cranial nerve involvement and the negative results of Sandostatin loading test. A decompression resection was done. The plasma ACTH level declined to 77 pg/ml and serum cortisol 30.2 µg/dl three days after the operation. Vision, pupil dilation, eye movements and blepharoptosis also partially improved. Histopathology and immunohistochemical staining confirmed a densely–granulated corticotroph adenoma (Fig. 2, NanoZoomer S360 digital slide scanner and NDP.view 2.9.25 software, Hamamatsu, Japan). Neither necrosis nor mitotic activity was observed. The immunostaining for somatostatin receptor SSTR2A was positive with a cytoplasmic pattern, while GH, PRL, TSH, FSH, LH and PIT were all negative. The Ki 67 index was found to be 10%. One month after the operation the ACTH level increased to 132 pg/mL again, and the parents agreed to refer their child for radiotherapy to control the residual tumor.

Fig. 2

figure 2

Histopathology and immunohistochemistry staining results of the pituitary tumor. By light microscopy, the tumor cells were mostly basophilic and arranged in papillary architecture. Neither necrosis nor mitotic activity was observed (A hematoxylin-eosin, ×200). Immunohistochemistry staining was positive for ACTH (B immunoperoxidase, ×200) and transcription factor T-PIT (C immunoperoxidase, ×200). Cytoplasmic staining of SSTR2A was observed in around 1/3 tumor cells besides the strong staining of endothelial cells (D immunoperoxidase, ×200). The Ki-67 index was 10% (E immunoperoxidase, ×200). Cytokeratin CAM5.2 was diffusely positive in the cytoplasm (F immunoperoxidase, ×200). The positive control for ACTH and T-PIT was the human anterior pituitary gland, and for SSRT2, Ki-67 and CAM5.2 were cerebral cortex, tonsil and colonic mucosa, respectively

The early onset and invasive behavior of this tumor led to the consideration of whether there was a genetic defect. Genetic studies were recommended for the families and they all agreed and signed the written informed consent forms. Whole exome sequencing (WES) was performed on the patient’s blood sample using an Illumina HiSeq sequencer to an average read depth of at least 90 times per individual. Raw sequence files were mapped to the GRCH37 human reference genome and analyzed using the Sentieon software. The results revealed a germline heterozygous GPR101 gene mutation c.505G > C (p.Gly169Arg), which was subsequently confirmed to be of maternal origin by Sanger sequencing. Meanwhile WES of the tumor tissue identified an additional somatic heterozygous c.2155_2157delTCC (p.S719del) mutation of the USP8 gene .

Discussion and conclusions

In this report, we described an extremely giant and invasive pituitary ACTH adenoma in a 10-year-old girl. According to Trouillas et al., invasive and proliferative pituitary tumors have a poor prognosis [5]. CD is rare among children, and the fast-growing and invasive nature of the tumor in this case led to the investigation of genetic causes. The somatic USP8 gene mutation has been recently reported to be associated with the pathogenesis of CD [67]. This gene encodes ubiquitin-specific protease 8 (USP8). S718, S719 and P720 are hotspots in different studies [6,7,8,9,10,11,12,13,14]. They are located at the 14-3-3 binding motif, and the mutations disrupt the binding between USP8 and 14-3-3 protein, which leads to increased deubiquitination and EGFR signaling. High levels of EGFR consequently trigger proopiomelanocortin (POMC) transcription and ACTH secretion [67]. The p.S719del mutation has been previously reported and its pathogenicity has been confirmed [7]. Thus, we speculate the p.S719del mutation plays a role in this patient with CD.

It is noteworthy that in our case, the pituitary corticotrophin adenoma was extremely giant and bilaterally invasive. USP8 mutations have been found in 31% of pediatric CD patients [10]. It is well known that microadenomas are most common in adult and pediatric CD patients. Previously, the Chinese and Japanese cohorts observed smaller sizes of USP8-mutated PAs than wild-type PAs [79]. The Chinese cohort also reported a lower rate of invasive adenomas in USP8-mutated PAs [7]. This may be explained by the finding that UPS8 mutations did not significantly promote cell proliferation more than the wild-type ones [6]. Other cohorts suggested no difference in tumor size or invasiveness between USP8-mutated and wild-type PAs [81012,13,14], which may be partially explained by the differences in sample sizes and ethnic backgrounds. Owing to the lack of evidence of USP8 mutations significantly contributing to tumor growth and invasiveness, additional pathogenesis should be investigated in this case.

The p.Gly169Arg mutation of the GPR101 gene has not been reported in patients with pituitary tumors. In silico predictions were performed using Polyphen-2, Mutation Taster and PROVEAN, and all of the programs reported it to be pathogenic. The GPR101 gene encodes an orphan G protein-coupled receptor (GPCR) and microduplication encompassing the gene has been proven to be the cause of X-linked acrogigantism (XLAG) [15]. XLAG is characterized by the early onset of pituitary GH-secreting macroadenomas. Point mutations of GPR101 have been found in patients with PAs that are mostly GH-secreting [15,16,17]. Although their prevalence is very low, an in vitro study supported the pathogenic role of p.E308D, the most common mutation of GPR101. This led to increased cell proliferation and GH production in rat pituitary GH3 cells [15]. Rare cases of PRL, ACTH or TSH-secreting PAs with GPR101 variants were also documented [1618]. To date, there have been five cases of ACTH-secreting PAs with four different germline GPR101 mutations: two cases of p.E308D, p.I122T, p.T293I and p.G31S, although in silico predictions and in vitro evaluations using AtT-20 cells have respectively determined the latter two mutations to be non-pathogenic [1618]. These patients were mainly children and young adults. Unlike pituitary GH-secreting tumors, the role of GPR101 mutations in the pathophysiology of CD is still questionable. Trivellin et al. demonstrated no statistically significant difference in GPR101 expression between corticotropinomas and normal human pituitaries. No significant correlation between GPR101 and POMC expression levels was found neither [18].

Given the evidences above, we hypothesize that the somatic USP8 mutation is responsible for the overexpression of ACTH in this CD girl while the germline GPR101 mutation contributes to the early onset and fast-growing nature of the tumor. Similarly, a 27-year-old woman with Nelson’s syndrome originally considered to be associated with a germline AIP variant (p.Arg304Gln) was recently reported to have a somatic USP8 mutation. The patient progressed rapidly and underwent multiple transsphenoidal surgeries [19]. Since germline AIP mutations are more commonly seen in GH-secreting PAs [20], the authors proposed that the USP8 mutation might have shifted the tumor towards ACTH-secreting [19]. Further investigations into the pathogenicity of GPR101 p.Gly169Arg and AIP p.Arg304Gln mutations are required to support the hypothesis.

In summary, we report a novel germline GPR101 and somatic USP8 mutation in a girl with an extremely giant pituitary ACTH adenoma. The concurrent mutations may lead to the growth and function of the tumor, respectively. Further investigations should be carried out to verify the role of the concurrent mutations in the pathogenesis of pediatric CD.

Availability of data and materials

The WES data of the blood sample of the patient is available in the NGDC repository (https://ngdc.cncb.ac.cn/gsa-human/) and the accession number is HRA002396. Any additional information is available from the authors upon reasonable request.

Abbreviations

CD:
Cushing’s disease
ACTH:
adrenocorticotropic hormone
PA:
pituitary adenoma
NRS:
numerical rating scale
24hUFC:
24-hour urine free cortisol
LDDST:
low-dose dexamethasone suppression test
USP8:
ubiquitin-specific protease 8
POMC:
proopiomelanocortin
GPCR:
G protein-coupled receptor
XLAG:
X-linked acrogigantism

References

  1. Weber A, Trainer PJ, Grossman AB, Afshar F, Medbak S, Perry LA, et al. Investigation, management and therapeutic outcome in 12 cases of childhood and adolescent Cushing’s syndrome. Clin Endocrinol (Oxf). 1995;43(1):19–28.

    CAS Article Google Scholar

  2. Storr HL, Alexandraki KI, Martin L, Isidori AM, Kaltsas GA, Monson JP, et al. Comparisons in the epidemiology, diagnostic features and cure rate by transsphenoidal surgery between paediatric and adult-onset Cushing’s disease. Eur J Endocrinol. 2011;164(5):667–74.

    CAS Article Google Scholar

  3. Magiakou MA, Mastorakos G, Oldfield EH, Gomez MT, Doppman JL, Cutler GB Jr, et al. Cushing’s syndrome in children and adolescents. Presentation, diagnosis, and therapy. N Engl J Med. 1994;331(10):629–36.

    CAS Article Google Scholar

  4. Hernández-Ramírez LC, Stratakis CA. Genetics of Cushing’s Syndrome. Endocrinol Metab Clin North Am. 2018;47(2):275–97.

    Article Google Scholar

  5. Trouillas J, Roy P, Sturm N, Dantony E, Cortet-Rudelli C, Viennet G, et al. A new prognostic clinicopathological classification of pituitary adenomas: a multicentric case-control study of 410 patients with 8 years post-operative follow-up. Acta Neuropathol. 2013;126(1):123–35.

    Article Google Scholar

  6. Reincke M, Sbiera S, Hayakawa A, Theodoropoulou M, Osswald A, Beuschlein F, et al. Mutations in the deubiquitinase gene USP8 cause Cushing’s disease. Nat Genet. 2015;47(1):31–8.

    CAS Article Google Scholar

  7. Ma Z-Y, Song Z-J, Chen J-H, Wang Y-F, Li S-Q, Zhou L-F, et al. Recurrent gain-of-function USP8 mutations in Cushing’s disease. Cell Res. 2015;25(3):306–17.

    CAS Article Google Scholar

  8. Perez-Rivas LG, Theodoropoulou M, Ferrau F, Nusser C, Kawaguchi K, Stratakis CA, et al. The Gene of the Ubiquitin-Specific Protease 8 Is Frequently Mutated in Adenomas Causing Cushing’s Disease. J Clin Endocrinol Metab. 2015;100(7):E997–1004.

    CAS Article Google Scholar

  9. Hayashi K, Inoshita N, Kawaguchi K, Ibrahim Ardisasmita A, Suzuki H, Fukuhara N, et al. The USP8 mutational status may predict drug susceptibility in corticotroph adenomas of Cushing’s disease. Eur J Endocrinol. 2016;174(2):213–26.

    CAS Article Google Scholar

  10. Faucz FR, Tirosh A, Tatsi C, Berthon A, Hernandez-Ramirez LC, Settas N, et al. Somatic USP8 Gene Mutations Are a Common Cause of Pediatric Cushing Disease. J Clin Endocrinol Metab. 2017;102(8):2836–43.

    Article Google Scholar

  11. Albani A, Perez-Rivas LG, Dimopoulou C, Zopp S, Colon-Bolea P, Roeber S, et al. The USP8 mutational status may predict long-term remission in patients with Cushing’s disease. Clin Endocrinol (Oxf). 2018;89:454–8.

    CAS Article Google Scholar

  12. Ballmann C, Thiel A, Korah HE, Reis AC, Saeger W, Stepanow S, et al. USP8 Mutations in Pituitary Cushing Adenomas-Targeted Analysis by Next-Generation Sequencing. J Endocr Soc. 2018;2(3):266–78.

    CAS Article Google Scholar

  13. Losa M, Mortini P, Pagnano A, Detomas M, Cassarino MF, Pecori Giraldi F. Clinical characteristics and surgical outcome in USP8-mutated human adrenocorticotropic hormone-secreting pituitary adenomas. Endocrine. 2019;63(2):240–6.

    CAS Article Google Scholar

  14. Weigand I, Knobloch L, Flitsch J, Saeger W, Monoranu CM, Hofner K, et al. Impact of USP8 Gene Mutations on Protein Deregulation in Cushing Disease. J Clin Endocrinol Metab. 2019;104(7):2535–46.

    Article Google Scholar

  15. Trivellin G, Daly AF, Faucz FR, Yuan B, Rostomyan L, Larco DO, et al. Gigantism and acromegaly due to Xq26 microduplications and GPR101 mutation. N Engl J Med. 2014;371(25):2363–74.

    CAS Article Google Scholar

  16. Lecoq AL, Bouligand J, Hage M, Cazabat L, Salenave S, Linglart A, et al. Very low frequency of germline GPR101 genetic variation and no biallelic defects with AIP in a large cohort of patients with sporadic pituitary adenomas. Eur J Endocrinol. 2016;174(4):523–30.

    CAS Article Google Scholar

  17. Iacovazzo D, Caswell R, Bunce B, Jose S, Yuan B, Hernández-Ramírez LC, et al. Germline or somatic GPR101 duplication leads to X-linked acrogigantism: a clinico-pathological and genetic study. Acta Neuropathol Commun. 2016;4(1):56.

    Article Google Scholar

  18. Trivellin G, Correa RR, Batsis M, Faucz FR, Chittiboina P, Bjelobaba I, et al. Screening for GPR101 defects in pediatric pituitary corticotropinomas. Endocr Relat Cancer. 2016;23(5):357–65.

    CAS Article Google Scholar

  19. Perez-Rivas LG, Theodoropoulou M, Puar TH, Fazel J, Stieg MR, Ferrau F, et al. Somatic USP8 mutations are frequent events in corticotroph tumor progression causing Nelson’s tumor. Eur J Endocrinol. 2018;178(1):57–63.

    CAS Article Google Scholar

  20. Tatsi C, Stratakis CA. The Genetics of Pituitary Adenomas. J Clin Med. 2019;9(1).

Download references

Acknowledgements

We thanked Dr. Xiaohua Shi and Dr. Yu Xiao from the Department of Pathology, Peking Union Medical College Hospital for their expertise in pituitary pathology and critical help in accomplishment of our manuscript.

Funding

This research was supported by “The National Key Research and Development Program of China” (No. 2016YFC0901501), “CAMS Innovation Fund for Medical Science” (CAMS-2017-I2M–1–011). They mainly covered the fees for genetic analysis and publications.

Author information

Authors and Affiliations

  1. Department of Pediatrics, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, 100730, China

    Xu-dong Bao

  2. Department of Endocrinology, Key Laboratory of Endocrinology of National Health Commission, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, 100730, China

    Lin Lu, Hui-juan Zhu, Xiao Zhai, Yong Fu, Feng-ying Gong & Zhao-lin Lu

  3. Department of Neurosurgery, Peking Union Medical College Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, 100730, China

    Yong Yao, Ming Feng & Ren-zhi Wang

Contributions

XB and LL contributed to the study design and manuscript writing. HZ and FG performed genetic analysis. XZ and YF collected the clinical data. YY, MF and RW provided the tumor tissue and histopathology data. ZL revised the manuscript. All authors have read and approved the final manuscript.

Corresponding author

Correspondence to Lin Lu.

Ethics declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Peking Union Medical College Hospital. The parents of the patient provided written informed consent for research participation.

Consent for publication

The parents of the patient provided written informed consent for the publication of indirectly identifiable data in this research.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The Genomic Landscape of Corticotroph Tumors: From Silent Adenomas to ACTH-Secreting Carcinomas

Abstract

Corticotroph cells give rise to aggressive and rare pituitary neoplasms comprising ACTH-producing adenomas resulting in Cushing disease (CD), clinically silent ACTH adenomas (SCA), Crooke cell adenomas (CCA) and ACTH-producing carcinomas (CA). The molecular pathogenesis of these tumors is still poorly understood. To better understand the genomic landscape of all the lesions of the corticotroph lineage, we sequenced the whole exome of three SCA, one CCA, four ACTH-secreting PA causing CD, one corticotrophinoma occurring in a CD patient who developed Nelson syndrome after adrenalectomy and one patient with an ACTH-producing CA. The ACTH-producing CA was the lesion with the highest number of single nucleotide variants (SNV) in genes such as USP8, TP53, AURKA, EGFR, HSD3B1 and CDKN1A. The USP8 variant was found only in the ACTH-CA and in the corticotrophinoma occurring in a patient with Nelson syndrome. In CCA, SNV in TP53, EGFR, HSD3B1 and CDKN1A SNV were present. HSD3B1 and CDKN1A SNVs were present in all three SCA, whereas in two of these tumors SNV in TP53, AURKA and EGFR were found. None of the analyzed tumors showed SNV in USP48, BRAF, BRG1 or CABLES1. The amplification of 17q12 was found in all tumors, except for the ACTH-producing carcinoma. The four clinically functioning ACTH adenomas and the ACTH-CA shared the amplification of 10q11.22 and showed more copy-number variation (CNV) gains and single-nucleotide variations than the nonfunctioning tumors.

1. Introduction

The pathological spectrum of the corticotroph includes ACTH (adrenocorticotropic hormone)-secreting pituitary adenomas (PA), causing Cushing disease (CD), silent corticotroph adenomas (SCA), Crooke cell adenomas (CCA) and the rare ACTH-secreting carcinoma (ACTH-CA). Pituitary carcinomas account for 0.1 to 0.2% of all pituitary tumors and are defined by the presence of craniospinal or distant metastasis [1,2,3]. Most pituitary carcinomas are of corticotroph or lactotrope differentiation [3]. Although a few cases present initially as CA, the majority develop over the course of several months or years from apparently benign lesions [3,4]. CCA are characterized by the presence of hyaline material in more than 50% of the cells of the lesion, and most of them arise from silent corticotroph adenomas (SCA) or CD-provoking ACTH-secreting adenomas [5]. SCA are pituitary tumors with positive immunostaining for ACTH but are not associated with clinical or biochemical evidence of cortisol excess; they are frequently invasive lesions and represent up to 19% of clinically non-functioning pituitary adenomas (NFPA) [6]. ACTH-secreting PA represents up to 6% of all pituitary tumors and causes eloquent Cushing disease (CD), which is characterized by symptoms and signs of cortisol hypersecretion, including a two- to fivefold increase in mortality [7,8]. The 2017 World Health Organization (WHO) classification of PA considers not only the hormones these tumors synthesize but also the transcription factors that determine their cell lineage [9]. TBX19 is the transcription factor responsible for the terminal differentiation of corticotrophs [9]. All tumor lesions of corticotroph differentiation are positive for both ACTH and TBX19.
ACTH-secreting PA causing CD are among the best genetically characterized pituitary tumors, with USP8 somatic variants occurring in up to 25–35% of sporadic cases [9]. Yet, information regarding the molecular pathogenesis of the lesions conforming to the whole pathological spectrum of the corticotroph is scarce. The aim of the present study is to characterize the genomic landscape of pituitary tumors of corticotroph lineage. For this purpose, we performed whole exome sequencing to uncover the mutational burden (single-nucleotide variants, SNV) and copy-number variations (CNVs) of these lesions.

2. Results

2.1. Clinical and Demographic Characteristics of the Patients

A total of 10 tumor samples from 10 patients were evaluated: 4 ACTH-secreting adenomas causing clinically evident CD, three non-functioning adenomas that proved to be SCA upon immunohistochemistry (IHC), one ACTH-secreting CA with a prepontine metastasis, one rapidly growing ACTH-secreting adenoma after bilateral adrenalectomy (Nelson syndrome) in a patient with CD and one non-functioning, ACTH-producing CCA (Table 1). All except one patient were female; the mean age was 38.8 ± 16.5 years (range 17–61) (Table 1). They all harbored macroadenomas with a mean maximum diameter of 31.9 ± 13 mm (range 18–51). Cavernous sinus invasion was evident on MRI in all but one of the patients (Table 1). Homonymous hemianopia was present in seven patients, whereas right optic nerve atrophy and amaurosis were evident in patient with the ACTH-CA, and in patient with CD and pituitary apoplexy (Table 1). Detailed clinical data are included in Supplementary Table S1. Death was documented in only the patient with pituitary apoplexy, and one patient was lost during follow-up, as of October 2018.
Table 1. Clinical features of the tumors analyzed and SNV present in each tumor.
Table

2.2. General Genomic Characteristics of Neoplasms of Corticotrophic Lineage

Overall, approximately 18,000 variants were found, including missense, nonsense and splice-site variants as well as frameshift insertions and deletions. Of these alterations, the majority corresponded to single-nucleotide variants, followed by insertions and deletions. The three most common base changes were transitions C > T, T > C and C > G; most of the genetic changes were base transitions rather than transversions (Figure 1). There were several genes across the whole genome affected in more than one way, meaning that the same gene presented missense and nonsense variants, insertions, deletions and splice-site variants (Figure 2). Many of these variants are of unknown pathogenicity and require further investigation. Gains in genetic material were found in 44 cytogenetic regions, whereas 72 cytogenetic regions showed loss of genetic material in all corticotroph tumors.
Ijms 23 04861 g001 550
Figure 1. Panel (A) shows the gadolinium-enhanced magnetic resonance imaging of the patient with ACTH-CA, highlighting in red the metastatic lesion in the prepontine area. Panel (B) shows the hematoxylin and eosin staining displaying the hyaline structures in the perinuclear areas denoting a Crooke cell adenoma. Panel (C,D) depict a representative corticotroph tumor with positive ACTH and TBX19 immunohistochemistry, respectively. Panel (E) shows four graphics: variant classification, variant type, SNV class and transition (ti) or transversion (tv) describing the general results of exome sequencing of the corticotroph tumors.
Ijms 23 04861 g002 550
Figure 2. Representative rainfall plots showing the SNV alterations throughout the whole genome of corticotroph tumors (A) CCA, (B) SCA, (C) CD and (D) ACTH-CA, displaying all base changes, including transversions and transitions. No kataegis events were found. Alterations across the genome were seen in all corticotroph tumors.

2.3. ACTH-Secreting Carcinoma (Tumor 1)

SNV missense variants were found in the genes encoding TP53 (c.215G > C [rs1042522], p.Pro72Arg); AURKA (c.91T > A [rs2273535], p.Phe31Ile); EGFR (epidermal growth factor receptor, c.1562G > A [rs2227983], p.Arg521Lys); HSD3B1 (3-ß-hydroxisteroid dehydrogenase, c.1100C > A [rs1047303], p.Thr367Asn); CDKN1A (cyclin-dependent kinase inhibitor 1A or p21, c.93C > A [rs1801270], p.Ser31Arg); and USP8 (c.2159C > G [rs672601311], p.Pro720Arg). Interestingly, the previously reported USP48, BRAF, BRG1 and CABLES1 variants in pituitary CA cases were not found in this patient’s tumor (Figure 3). All SNV detected in WES experiments were validated by Sanger sequencing. The variants described were selected due to their potential pathogenic participation in other tumors and the allelic-risk association with tumorigenesis. Hereafter, all the mentioned variants in other corticotroph tumors are referred to by these aforementioned variants. Even though these same genes presented other variants, currently the significance of those variants is unknown.
Ijms 23 04861 g003 550
Figure 3. Panel (A) shows the oncoplot from the missense variants of the selected genes and their clinical–pathological features. Panels (BG) depict USP8, EGFR, TP53, AURKA, CDKN1A and HSD3B1 proteins, respectively, with the changes found in DNA impacting aminoacidic changes.
In general, the pituitary CA presented more CNV alterations than the benign tumors, with 27 and 32 cytogenetic regions showing gains and losses of genetic material, respectively. The cytogenetic regions showing gains were 10q11.22, 15q11.2, 16p12.3, 1p13.2 and 20p, where genes SYT15, POTEB, ARL6IP1, HIPK1 and CJD6 are coded, respectively. By contrast, 8p21.2 was the cytogenetic region showing loss of genetic material. The previously reported amplification of 1p13.2 was also detected in this tumor (Figure 4) [10].
Ijms 23 04861 g004 550
Figure 4. Hierarchical clustering of corticotroph tumors according to their gains and losses across the whole genome (somatic chromosomes only). High contrast was used to enhance potential CNV alterations; nevertheless, there were only 44 unique cytogenetic regions that showed gains in genetic material with statistical significance, whereas only 72 unique cytogenetic regions showed loss of genetic material with statistical significance.

2.4. Crooke Cell Adenoma (Tumor 2)

The CCA showed SNV in the genes encoding TP53, EGFR, HSD3B1 and CDKN1A. However, neither the genes encoding AURKA and USP8 nor those encoding USP48, BRAF, BRG1 and CABLES were affected in this tumor. In CCA, only two and fifteen gains and losses were observed in copy-number variation, respectively. CNVs only showed gains in cytogenetic regions 17q12 and 10q11.22, harboring genes CCL3L1 and NPY4R, respectively, whereas losses were found in cytogenetic regions 18q21.1, 15q12 and 2q11.2, harboring genes KATNAL2, TUBGCP5 and ANKRD36.

2.5. Silent Corticotroph Adenomas (Tumors 3–5)

The three SCA shared SNVs in the genes encoding HSD3B1 and CDKN1A. SCA 4 and 5 showed SNV in the genes encoding EGFR, whereas SNV in the genes encoding AURKA and TP53 were present in SCA 3 and 5. None of the SCA were found to have SNV in the genes encoding USP8, USP48, BRAF, BRG1 or CABLES1.
The SCA presented only two and eighteen gains and losses (CNV), respectively. In regard to CNV, the these clinically silent tumors presented gains of genetic material in cytogenetic regions 17q22 and 10q11.22, which harbor genes encoding CCL3L1 and NPY4R. Eighteen losses were found distributed in cytogenetic regions 18q21.1, 15q12 and 2q11.2, encompassing the genes encoding KATNAL2, TUBGCP5 and ANKRD36. This CNV pattern closely resembles the one found in the CCA, which is somewhat expected if we consider that both neoplasms are clinically non-functioning

2.6. ACTH-Secreting Adenomas (Cushing Disease) (Tumors 6–9)

SNV of the genes encoding TP53 and HSD3B1 were present in tumor samples from all four CD patients, whereas none of these patients harbored adenomas with SNV in the genes encoding USP8 or CDKN1A. An SNV in the gene encoding AURKA was identified in only one of these tumors (tumor 8). EGFR SNV were found in tumors 7 and 9. None of the CD-causing ACTH-secreting adenomas showed the previously reported SNV in the genes encoding USP48, BRAF, BRG1 and CABLES1.
CNV analysis in this group of eloquent-area corticotroph tumors revealed 25 gains and 55 losses of genetic material. The gains occurred in cytogenetic regions 17q12, 2p12, 9p24 and 10q11.22, where genes CCL3L1, CTNNA2, FOXD4 and NPY4R are coded, respectively. The losses were localized in cytogenetic regions 21p12, 15q11.2, and 8p23, harboring genes USP16, KLF13 and DEF130A, respectively. We also detected the previously reported 20p13 amplification [10].

2.7. ACTH-Secreting Adenoma Causing Nelson Syndrome (Tumor 10)

This patient’s tumor showed SNV in the genes encoding USP8, TP53, HSD3B1 and CDKN1A but no alterations were found in the genes encoding EGFR and AURKA. This tumor and the ACTH-CA were the only two neoplasms that harbored a USP8 variant. No SNV were identified in the genes encoding USP48, BRAF, BRG1 and CABLES1. Interestingly, CNV analysis revealed the same gains and losses of genetic material found in tumors from other patients with CD.

2.8. Tumor Phylogenic Analysis

We performed a phylogenetic inference analysis to unravel a hypothetical sequential step transformation from an SCA to a functioning ACTH-secreting adenoma and finally to an ACTH-CA. The theoretical evolutive development of the ACTH CA, departing from the SCA, shows two main clades, with the smallest one comprising two of the three SCA and two of the five ACTH-adenomas causing CD. Since these four tumors have the same SNV profile, we can assume that they harbor the genes that must be altered to make possible the transition from a silent to a clinically eloquent adenoma; the gene encoding ATF7IP (c.1589A > G [rs3213764], p.K529R) characterizes this clade. The second and largest clade includes the CCA, the ACTH-CA, one of the three SCA and three of the five most aggressive ACTH adenomas causing CD, including the adenoma of the patient with Nelson syndrome. This clade represents the molecular alterations required to evolve from a CD-causing ACTH-adenoma to a more aggressive tumor, or even to a CA and is characterized by the gene encoding MSH3 (c.235A > G [rs1650697], p.I79V) (Figure 5).
Ijms 23 04861 g005 550
Figure 5. Phylogenetic analysis of the corticotroph tumors. The theoretical evolutive development of the ACTH-CA, departing from the SCA shows two main clades. The first clade, characterized by ATF7IP gene, comprises 2 of the 3 SCA and 2 of the 5 ACTH-adenomas causing CD. The second clade is characterized by the gene encoding MSH3 and includes the CCA, the ACTH-CA, one of the 3 SCA and 3 of the 5 most aggressive ACTH adenomas causing CD, including the adenoma of the patient with Nelson syndrome. Red dots represent the Cushing Disease provoking adenomas, green dots represent the silent corticotroph tumors, brown dot represent the Crooke cell adenoma and the blue dot represent the corticotroph carcinoma.

2.9. Correlation between Gene Variants and Clinicopathological Features

The USP8 variant positively correlated with increased tumor mass (p = 0.019). The CDKN1A variant was significantly associated with silent tumors (p = 0.036). The rest of the genetic variants did not correlate with any of the clinicopathological features tested. The presence of the EGFR variant was not distinctly associated with any of the clinical parameters and was equally present in functional as well as non-functional tumors (p = 0.392). AURKA SNV did not correlate with any of the features, including recurrence (p = 0.524). Detailed statistical results are presented in Supplementary Table S2.

3. Discussion

Corticotrophs are highly specialized cells of the anterior pituitary that synthesize and secrete hormones that are essential for the maintenance of homeostasis. In this study, we sequenced the exome of 10 corticotroph tumors, including three SCA, four ACTH adenomas causing CD, an ACTH adenoma in a patient with Nelson syndrome, a CCA and an ACTH-CA in total, representing the broad pathological spectrum of this cell. Our results portray the genomic landscape of all the neoplasms that are known to affect the corticotroph.
The neoplasm with the highest number of genomic abnormalities, including SNV and CNV, was the ACTH-CA, followed by the CCA and the CD tissues. Of all the genes harboring SNVs, six were found to be present in at least two of our tumor samples: HSD3B1, TP53, CDKN1A, EGFR, AURKA and USP8.
The HSD3B1 gene encodes a rate-limiting enzyme required for all pathways of dihydrotestosterone synthesis and is abundantly expressed in adrenal tumors. Gain of function of this HSD3B1 variant, which has a global allelic prevalence of 0.69678 [11], results in resistance to proteasomal degradation with the consequent accumulation of the enzyme and has been associated with a poor prognosis in patients with prostate cancer [12]. Nine of the ten corticotroph tumors in our cohort harbored an SNV of the tumor suppressor gene TP53. The TP53 variant described in our cohort has been reported to be present in 80% of non-functioning pituitary adenomas and is apparently associated with a younger age at presentation and with cavernous sinus invasion [13]. Furthermore, this TP53 variant results in a reduced expression of CDKN1A and an increased expression of vascular endothelial growth factor (VEGF) as well as an increased cellular proliferation rate [13]. CDKN1A (also known as p21) is a cyclin-dependent kinase inhibitor regulating cell cycle progression. The SNV described in our study was reported to alter DNA binding ability and expression and has a global allelic frequency of 0.086945 [14]. This cyclin-dependent kinase inhibitor SNV was found to be associated with breast carcinoma [15] and lung cancer [16]. The presence of this SNV has not been previously explored in pituitary adenomas, although CDKN1A is downregulated in clinically non-functioning pituitary adenomas of gonadotrophic lineage but not in hormone-secreting tumors [17]. EGFR encodes a transmembrane tyrosine kinase receptor, activation of which leads to mitogenic signaling [18]. This gene is upregulated in several cancers and represents a target for molecular therapies [19]. The EGFR SNV described in our corticotroph tumor series was found to be associated with the response to neoadjuvant chemotherapy in patients with breast and lung cancer [18]. EGFR is normally expressed in corticotrophs, where it participates in the regulation of POMC (proopiomelanocortin) gene transcription and cellular proliferation [20]. The EGFR rs2227983 has a 0.264334 global allelic frequency [21]. AURKA is a cell-cycle regulatory serine/threonine kinase that promotes cell cycle progression by the establishment of the mitotic spindle and centrosome separation [22]. Alterations of these gene are related to centrosomal amplification, dysfunction of cytokinesis and aneuploidy [22]; it has a global allelic frequency of 0.18078 [23]. This same SNV has been associated with overall cancer risk, particularly breast, gastric, colorectal, liver and endometrial carcinomas, but it has never been formally studied in pituitary tumors [22]. Activating somatic variants of the gene encoding USP8 were recently found in 25–40% of ACTH-secreting adenomas causing CD [24,25]. Patients harboring these variants are usually younger, more frequently females and were found to have higher long-term recurrence rates in some but not all studies [26,27]. USP8 mediates the deubiquitination of EGFR by inhibiting its interaction with protein 14-3-3, which in turn prevents its proteosomal degradation. Signaling through the recycled deubiquitinated EGFR is increased, leading to increased POMC transcription and cellular proliferation. Most activating USP8 variants are located within its 14-3-3 binding motif [24,25]. Recently, USP8 and TP53 SNV were described in corticotroph tumors as drivers of aggressive lesions [28]. To our knowledge, USP8 variants have not been evaluated in patients with pituitary carcinomas, and none of the previously mentioned studies have included patients with Nelson syndrome. In our cohort, neither the CCA nor the SCA showed variants in USP8, in concordance with previously published studies [25,29], or in the genes USP48, BRAF, BRG1 and CABLES1 [9], and none of them were present in our cohort.
Genetic structural variations in the human genome can be present in many forms, from SNV to large chromosomal aberrance [30]. CNV are structurally variant regions, including unbalanced deletions, duplications and amplifications of DNA segments ranging from a dozen to several hundred base pairs, in which copy-number differences have been observed between two or more genomes [31,32]. CNV are involved in the development and progression of many tumors and occur frequently in PA [30,33]. Hormone-secreting pituitary tumors show more CNV than non-functioning tumors [34]. Accordingly, our non-functioning SCA and CCA had considerably fewer chromosomal gains and losses than the CD-causing adenomas and the ACTH-CA. Expectedly, the ACTH-CA had significantly more cytogenetic abnormalities than any other tumor in our series. Interestingly, the ACTH-adenomas causing CD, the SCA and the CCA shared the gain of genetic material in 17q12, highlighting their benign nature. The 17q12 amplification has been described in gastric neoplasms [35]. The only cytogenetic abnormality shared by all types of corticotroph tumors was the gain of genetic material in 10q11.22. Amplification of 10q11.22 was previously described in Li–Fraumeni cancer predisposition syndrome [36]. The ACTH-CA, the CCA and one SCA clustered together showing a related CNV pattern; this CNV profile could be reflective of the aggressive nature of these neoplasms, since both CCA and SCA can follow a clinically aggressive course [5,6].
Our results show that all lesions conforming to the pathological spectrum of the corticotroph share some of the SNV and CNV profiles. These genomic changes are consistent with the potential existence of a continuum, whereby silent tumors can transform into a clinically eloquent tumor and finally to carcinoma, or at least a more aggressive tumor. It can also be interpreted as the common SNV shared by aggressive tumors. It is known that silent corticotroph adenomas may switch into a hormone-secreting tumor [37] and are considered a marker for aggressiveness and a risk factor for malignancy since most of the carcinomas are derived from functioning hormone-secreting adenomas. Our phylogenetic inference analysis showed that the genes ATF7IP and MSH3 could participate in a tumor transition ending in aggressive entities or even carcinomas. ATF7IP is a multifunctional nuclear protein mediating heterochromatin formation and gene regulation in several contexts [38], while MSH3 is a mismatch-repair gene [39]. Events related to heterochromatin remodeling and maintenance have been related to aggressive pituitary adenomas and carcinomas [40]. Additionally, alterations in mismatch-repair genes are related to pituitary tumor aggressiveness and resistance to pharmacologic treatment [41,42]. The variants described in ATF7IP and MSH3 are related to prostate and colorectal cancer, respectively [43,44]. There is evidence suggesting that the ATF7IP variant could be deleterious because it leads to a negative regulation of transcription [45]. Thus, these events could be biologically relevant to corticotroph tumorigenesis, although more research is needed.

4. Conclusions

We have shown genomic evidence that within the tumoral spectrum of the corticotroph, functioning ACTH-secreting lesions harbor more SNV and CNV than non-functioning ACTH adenomas. The ACTH-secreting CA shows more genomic abnormalities than the other lesions, underscoring its more aggressive biological behavior. Phylogenetic inference analysis of our data reveals that silent corticotroph lesions may transform into functioning tumors, or at least potentially, into more aggressive lesions. Alterations in genes ATF7IP and MSH3, related to heterochromatin formation and mismatch repair, could be important in corticotroph tumorigenesis. The main drawback of our study is the limited sample size. We are currently increasing the number of samples to corroborate our findings and to be able to perform a more comprehensive complementary phylogenetic analysis of our data. Finally, further research is needed to uncover the roles of these variants in corticotroph tumorigenesis.

5. Materials and Methods

5.1. Patients and Tumor Tissue Samples

Ten pituitary tissues were collected: one ACTH-CA, one CCA, three SCA, and five ACTH-secreting PA causing CD, including the tumor of a patient who developed Nelson syndrome after bilateral adrenalectomy. All tumors included in the study were sporadic and were collected from patients diagnosed, treated and followed at the Endocrinology Service and the Neurosurgical department of Hospital de Especialidades, Centro Médico Nacional Siglo XXI of the Instituto Mexicano del Seguro Social, Hospital General de Mexico “Dr. Eduardo Liceaga” and Instituto Nacional de Neurologia y Neurocirugia “Manuel Velazquez”. All participating patients were recruited with signed informed consent and ethical approval from the Comisión Nacional de Ética e Investigación Científica of the Instituto Mexicano del Seguro Social, in accordance with the Helsinki declaration.
CD was diagnosed according to our standard protocol. Briefly, the presence of hypercortisolism was documented based on two screening tests, namely a 24 h urinary free-cortisol level above 130 µg and the lack of suppression of morning (7:00–8:00) cortisol after administration of 1 mg dexamethasone the night before (23:00) to less than 1.8 µg/dL, followed by a normal or elevated plasma ACTH to ascertain ACTH-dependence. Finally, an overnight, high-dose (8 mg) dexamethasone test, considered indicative of a pituitary source, and a cortisol suppression > 69%, provided that a pituitary adenoma was clearly present on magnetic resonance imaging (MRI) of the sellar region. In none of the 10 patients included in the study was inferior petrosal venous sampling necessary to confirm the pituitary origin of the ACTH excess. Invasiveness was defined by the presence of tumor within the cavernous sinuses (CS).
DNA was extracted from paraffin-embedded tumor tissues using the QIAamp DNA FFPE tissue kit. From frozen tumors, DNA was obtained using the Proteinase K-ammonium acetate protocol.

5.2. Construction and Sequencing of Whole Exome Libraries

Exome libraries were prepared according to the Agilent SureSelect XT HS Human All exon v7 instructions. Briefly, 200 ng of DNA was enzymatically fragmented with Agilent SureSelect Enzymatic Fragmentation Kit. Fragmented DNA was end-repaired and dA-tail was added at DNA ends; then, molecular barcode adaptors were added, followed by AMPure XP bead purification. The adaptor-ligated library was amplified by PCR and purified by AMPure XP beads. DNA libraries were hybridized with targeting exon probes and purified with streptavidin-coated magnetic beads. The retrieved libraries were amplified by PCR and purified by AMPure XP beads and pooled for sequencing in NextSeq 500 using Illumina flow cell High Output 300 cycles chemistry. All quality controls of the libraries were carried out using Screen tape assays and quantified by Qubit fluorometer. Quality parameters included a DNA integrity number above 8 and a 100X sequencing depth aimed with at least 85% of coverage.

5.3. Bioinformatics Analysis

The fastq files were subjected to quality control using FastQC v0.11.9, the adapters were removed using Cutadapt v3.4, the alignment was carried out with Burrows–Wheeler Alignment Tool v0.7.17 with the -M option to ensure compatibility with Picard and GRCh38 as a reference genome. The marking of duplicates as well as the sorting was carried out with Picard v2.26.4 with the AddOrReplaceReadGroups programs with the option SORT_ORDER = coordinate and MarkDuplicates, respectively. Variant calling was carried out using Genomic Analysis Toolkit (GATK) v4.2.2.0 following the Best Practices guide (available at https://gatk.broadinstitute.org/) [46] and with the parameters used by Genomic Data Commons (GDC), available at https://docs.gdc.cancer.gov/ [47]. The GATK tools used were CollectSequencingArtifactMetrics, GetPileupSummaries, CalculateContamination and Mutect2. Mutect2 was run with the latest filtering recommendations, including a Panel of Normal and a Germline Reference from the GATK database. Filtering was performed with the CalculateContamination, LearnReadOrientationModel and FilterMutectCalls tools with the default parameters. For the calculation of CNV GISTIC v2.0.23 was used with the parameters used by GDC. Catalog of Somatic Mutation in Cancer (COSMIC) was used to uncover pathogenic variants. For the analysis of variants and CNV, the maftool v2.10.0 and ComplexHeatmap 2.10.0 packages were used. All analyses were carried out on the GNU/Linux operating system under Ubuntu v20.01.3 or using the R v4.0.2 language in Rstudio v2021.09.0+351. A second bioinformatics pipeline was also used, SureCall software (Agilent) with the default parameters used for SNV variant calling. The variants found by both algorithms were taken as reliable SNV. Data were deposited in Sequence Read Archive hosted by National Center for Biotechnology Information under accession number PRJNA806516.
Phylogenetic tree inference (PTI) was run by means of the default parameters using matrices for each sample. These matrices contain an identifier for each variant, mutant read counts, counts of reference reads and the gene associated with the variant. The only PTI parameter was Allele Frequency of Mutation and was used to improve the speed of the algorithm. Briefly, PTI uses an iterative process on the variants shared between the samples. First, it builds the base of the tree using the variants shared by all the samples; second, it eliminates these variants and establishes a split node; and third, it eliminates the variants of the sample that produced the division (split). PTI iteratively performs these three steps for all division possibilities. Each tree is given a score based on an aggregated variant count, and the tree with the highest score is chosen as the optimal tree.

5.4. Sanger Sequencing forConfirmation of Exome Findings

Exome variant findings in exome sequencing were validated by Sanger sequencing using BigDye Terminator v3.1 Cycle Sequencing kit (ThermoFischer) in a 3500 Genetic Analyzer. Primers used for USP8 [48], TP53 [49], EGFR [50], AURKA [51], CDKN1A [52,53] and HSD3B1 sequencing have been previously reported.

5.5. Hormone and Transcription Factor Immunohistochemistry

Paraffin-embedded, formalin-fixed tissue blocks were stained with hematoxylin–eosin and reviewed by a pathologist. Tumors were represented with a 2-fold redundancy. Sections (3 μm) were cut and placed onto coated slides. Immunostaining was performed by means of the HiDef detection HRP polymer system (Cell Marque, CA, USA), using specific antibodies against each pituitary hormone (TSH, GH, PRL, FSH, LH and ACTH) and the lineage-specific transcription factors TBX19, POU1F1 and NR5A1, as previously described [54]. Two independent observers performed assessment of hormones and transcription factors expression at different times.

5.6. Statistical Analysis

Two-tailed Fisher exact tests and Student’s t tests were used to evaluate the relationship between the identified gene variants and clinicopathological features. A p value of <0.05 was considered statistically significant. Statistical software consisted of SPSS v28.0.1

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms23094861/s1.

Author Contributions

D.M.-R., K.T.-P. and M.M. conceived, designed and coordinated the project, performed experiments, analyzed, discussed data and prepared the manuscript. S.A.-E., G.S.-R., E.P.-M., S.V.-P., R.S., L.B.-A., C.G.-T., J.G.-C. and J.T.A.-S. performed DNA purification, library preparation, sequencing experiments, bioinformatics analysis and wrote the manuscript. A.-L.E.-d.-l.-M., I.R.-S., E.G.-A., L.A.P.-O., G.G., S.M.-J., L.C.-M., B.L.-F. and A.B.-L. provided biological samples and detailed patient information. All authors have read and agreed to the published version of the manuscript.

Funding

This work was partially supported by grants 289499 from Fondos Sectoriales Consejo Nacional de Ciencia y Tecnologia, Mexico, and R-2015-785-015 from Instituto Mexicano del Seguro Social (MM).

Institutional Review Board Statement

Protocol approved by the Comisión Nacional de Ética e Investigación Científica of the Instituto Mexicano del Seguro Social, in accordance with the Helsinki declaration (R-2019-785-052).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

Data were deposited in Sequence Read Archive hosted by National Center for Biotechnology Information under accession number PRJNA806516.

Acknowledgments

Sergio Andonegui-Elguera is a doctoral student from Programa de Doctorado en Ciencias Biomédicas, Universidad Nacional Autónoma de México (UNAM) and received fellowship 921084 from CONACYT. KTP is a recipient of Consejo Nacional de Ciencia y Tecnología (CONACyT) fellowship “Estáncias posdoctorales por Mexico 2021” program. DMR is a recipient of the National Council for Science and Technology Fellowship “Catedra CONACyT” program.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Heaney, A. Clinical review: Pituitary carcinoma: Difficult diagnosis and treatment. J. Clin. Endocrinol. Metab. 2011, 96, 3649–3660. [Google Scholar] [CrossRef] [PubMed]
  2. Raverot, G.; Burman, P.; McCormack, A.; Heaney, A.; Petersenn, S.; Popovic, V.; Trouillas, J.; Dekkers, O.M.; The European Society of Endocrinology. European Society of Endocrinology Clinical Practice Guidelines for the management of aggressive pituitary tumours and carcinomas. Eur. J. Endocrinol. 2018, 178, G1–G24. [Google Scholar] [CrossRef]
  3. Todeschini, A.B.; Beer-Furlan, A.; Montaser, A.S.; Jamshidi, A.O.; Ghalib, L.G.; Chavez, J.A.; Lehman, N.L.; Prevedello, D.M. Pituitary carcinomas: Review of the current literature and report of atypical case. Br. J. Neurosurg. 2020, 34, 528–533. [Google Scholar] [CrossRef]
  4. Dudziak, K.; Honegger, J.; Bornemann, A.; Horger, M.; Müssig, K. Pituitary carcinoma with malignant growth from first presentation and fulminant clinical course-case report and review of the literature. J. Clin. Endocrinol. Metab. 2011, 96, 2665–2669. [Google Scholar] [CrossRef] [PubMed]
  5. Di Ieva, A.; Davidson, J.; Syro, L.; Rotondo, F.; Montoya, J.; Horvath, E.; Cusimano, M.D.; Kovacs, K. Crooke’s cell tumors of the pituitary. Neurosurgery 2015, 76, 616–622. [Google Scholar] [CrossRef] [PubMed]
  6. Fountas, A.; Lavrentaki, A.; Subramanian, A.; Toulis, K.; Nirantharakumar k Karavitaki, N. Recurrence in silent corticotroph adenomas after primary treatment: A systematic review and meta-analysis. J. Clin. Endocrinol. Metab. 2019, 104, 1039–1048. [Google Scholar] [CrossRef] [PubMed]
  7. Molitch, M. Diagnosis and Treatment of Pituitary Adenomas: A Review. JAMA 2017, 317, 516–524. [Google Scholar] [CrossRef]
  8. Melmed, S. Pituitary-Tumor Endocrinopathies. N. Engl. J. Med. 2020, 382, 937–950. [Google Scholar] [CrossRef] [PubMed]
  9. Lopes, M.B.S. The 2017 World Health Organization classification of tumors of the pituitary gland: A sumary. Acta Neuropathol. 2017, 134, 521–535. [Google Scholar] [CrossRef] [PubMed]
  10. Song, Z.-J.; Reitman, Z.; Ma, Z.-Y.; Chen, J.-H.; Zhang, Q.-L.; Shou, X.-F.; Huang, C.X.; Wang, Y.F.; Li, S.Q.; Mao, Y.; et al. The genome-wide mutational landscape of pituitary adenomas. Cell Res. 2016, 26, 1255–1259. [Google Scholar] [CrossRef]
  11. NCBI. Available online: https://www.ncbi.nlm.nih.gov/snp/rs1047303#frequency_tab (accessed on 23 February 2022).
  12. Shiota, M.; Narita, S.; Akamatsu, S.; Fujimoto, N.; Sumiyoshi, T.; Fujiwara, M.; Uchiumi, T.; Habuchi, T.; Ogawa, O.; Eto, M. Association of Missense Polymorphism in HSD3B1 With Outcomes Among Men With Prostate Cancer Treated With Androgen-Deprivation Therapy or Abiraterone. JAMA Netw. Open 2019, 2, e190115. [Google Scholar] [CrossRef]
  13. Yagnik, G.; Jahangiri, A.; Chen, R.; Wagner, J.; Aghi, M. Role of a p53 polymorphism in the development of nonfunctional pituitary adenomas. Mol. Cell Endocrinol. 2017, 446, 81–90. [Google Scholar] [CrossRef]
  14. Heidari, Z.; Harati-Sadegh, M.; Arian, A.; Maruei-Milan, R.; Salimi, S. The effect of TP53 and P21 gene polymorphisms on papillary thyroid carcinoma susceptibility and clinical/pathological features. IUBMB Life 2020, 72, 922–930. [Google Scholar] [CrossRef]
  15. Akhter, N.; Dar, S.; Haque, S.; Wahid, M.; Jawed, A.; Akhtar, M.S.; A Alharbi, R.; A A Sindi, A.; Alruwetei, A.; Choudhry, H.M.Z.; et al. Crosstalk of Cyclin-dependent kinase inhibitor 1A (CDKN1A) gene polymorphism with p53 and CCND1 polymorphism in breast cancer. Eur. Rev. Med. Pharmacol. Sci. 2021, 25, 4258–4273. [Google Scholar]
  16. Wang, C.; Nie, H.; Li, Y.; Liu, G.; Wang, X.; Xing, S.; Zhang, L.; Chen, X.; Chen, Y.; Li, Y. The study of the relation of DNA repair pathway genes SNPs and the sensitivity to radiotherapy and chemotherapy of NSCLC. Sci. Rep. 2016, 6, 26526. [Google Scholar] [CrossRef] [PubMed]
  17. Taniguchi-Ponciano, K.; Portocarrero-Ortiz, L.A.; Guinto, G.; Moreno-Jimenez, S.; Gomez-Apo, E.; Chavez-Macias, L.; Peña-Martínez, E.; Silva-Román, G.; Vela-Patiño, S.; Ordoñez-García, J.; et al. The kinome, cyclins and cyclin-dependent kinases of pituitary adenomas, a look into the gene expression rofile among tumors different lineages. BMC Med. Genom. 2022, 15, 52. [Google Scholar] [CrossRef]
  18. Sobral-Leite, M.; Lips, E.; Vieira-Monteiro, H.; Giacomin, L.; Freitas-Alves, D.; Cornelissen, S.; Mulder, L.; Wesseling, J.; Schmidt, M.K.; Vianna-Jorge, R. Evaluation of the EGFR polymorphism R497K in two cohorts of neoadjuvantly treated breast cancer patients. PLoS ONE 2017, 12, e0189750. [Google Scholar] [CrossRef]
  19. Zhang, H.; Berezov, A.; Wang, Q.; Zhang, G.; Drebin, J.; Murali, R.; Greene, M. ErbB receptors: From oncogenes to targeted cancer therapies. J. Clin. Investig. 2007, 117, 2051–2058. [Google Scholar] [CrossRef] [PubMed]
  20. Liu, X.; Feng, M.; Dai, C.; Bao, X.; Deng, K.; Yao, Y.; Wang, R. Expression of EGFR in Pituitary Corticotroph Adenomas and Its Relationship With Tumor Behavior. Front. Endocrinol. 2019, 10, 785. [Google Scholar] [CrossRef]
  21. NCBI. Available online: https://www.ncbi.nlm.nih.gov/snp/rs2227983#frequency_tab (accessed on 23 February 2022).
  22. Wang, S.; Qi, J.; Zhu, M.; Wang, M.; Nie, J. AURKA rs2273535 T>A Polymorphism Associated With Cancer Risk: A Systematic Review With Meta-Analysis. Front. Oncol. 2020, 10, 1040. [Google Scholar] [CrossRef] [PubMed]
  23. NCBI. Available online: https://www.ncbi.nlm.nih.gov/snp/rs2273535#frequency_tab (accessed on 23 February 2022).
  24. Reincke, M.; Sbiera, S.; Hayakawa, A.; Theodoropoulou, M.; Osswald, A.; Beuschlein, F.; Meitinger, T.; Mizuno-Yamasaki, E.; Kawaguchi, K.; Saeki, Y.; et al. Mutations in the deubiquitinase gene USP8 cause Cushing’s disease. Nat. Genet. 2015, 47, 31–38. [Google Scholar]
  25. Perez-Rivas, L.; Theodoropoulou, M.; Ferraù, F.; Nusser, C.; Kawaguchi, K.; Faucz, F.; Nusser, C.; Kawaguchi, K.; Stratakis, C.A.; Faucz, F.R.; et al. The Gene of the Ubiquitin-Specific Protease 8 Is Frequently Mutated in Adenomas Causing Cushing’s Disease. J. Clin. Endocrinol. Metab 2015, 100, E997–E1004. [Google Scholar] [CrossRef] [PubMed]
  26. Albani, A.; Pérez-Rivas, L.G.; Dimopoulou, C.; Zopp, S.; Colón-Bolea, P.; Roeber, S.; Honegger, J.; Flitsch, J.; Rachinger, W.; Buchfelder, M.; et al. The USP8 mutational status may predict long-term remission in patients with Cushing’s disease. Clin. Endocrinol. 2018, 89, 454–458. [Google Scholar] [CrossRef] [PubMed]
  27. Wanichi, I.Q.; de Paula Mariani, B.M.; Frassetto, F.P.; Siqueira, S.A.C.; de Castro Musolino, N.R.; Cunha-Neto, M.B.C.; Ochman, G.; Cescato, V.A.S.; Machado, M.C.; Trarbach, E.B.; et al. Cushing’s disease due to somatic USP8 mutations: A systematic review and meta-analysis. Pituitary 2019, 22, 435–442. [Google Scholar] [CrossRef]
  28. Uzilov, A.; Taik, P.; Cheesman, K.; Javanmard, P.; Ying, K.; Roehnelt, A.; Wang, H.; Fink, M.Y.; Lau, C.Y.; Moe, A.S.; et al. USP8 and TP53 Drivers are Associated with CNV in a Corticotroph Adenoma Cohort Enriched for Aggressive Tumors. J. Clin. Endocrinol. Metab. 2021, 106, 826–842. [Google Scholar] [CrossRef]
  29. Hayashi, K.; Inoshita, N.; Kawaguchi, K.; Ibrahim, A.; Suzuki, H.; Fukuhara, N.; Okada, M.; Nishioka, H.; Takeuchi, Y.; Komada, M.; et al. The USP8 mutational status may predict drug susceptibility in corticotroph adenomas of Cushing’s disease. Eur. J. Endocrinol. 2016, 174, 213–226. [Google Scholar] [CrossRef] [PubMed]
  30. Shao, X.; Lv, N.; Liao, J.; Long, J.; Xue, R.; Ai, N.; Xu, D.; Fan, X. Copy number variation is highly correlated with differential gene expression: A pan-cancer study. BMC Med Genet. 2019, 20, 175. [Google Scholar] [CrossRef] [PubMed]
  31. Shlien, A.; Malkin, D. Copy number variations and cancer. Genome Med. 2009, 1, 62. [Google Scholar] [CrossRef]
  32. Pös, O.; Radvanszky, J.; Styk, J.; Pös, Z.; Buglyó, G.; Kajsik, M.; Budis, J.; Nagy, B.; Szemes, T. Copy Number Variation: Methods and Clinical Applications. Appl. Sci. 2021, 11, 819. [Google Scholar] [CrossRef]
  33. Cui, Y.; Li, C.; Jiang, Z.; Zhang, S.; Li, Q.; Liu, X.; Zhou, Y.; Li, R.; Wei, L.; Li, L.; et al. Single-cell transcriptome and genome analyses of pituitary neuroendocrine tumors. Neuro-Oncol. 2021, 23, 1859–1871. [Google Scholar] [CrossRef]
  34. Neou, M.; Villa, C.; Armignacco, R.; Jouinot, A.; Raffin-Sanson, M.-L.; Septier, A.; Letourneur, F.; Diry, S.; Diedisheim, M.; Izac, B.; et al. Pangenomic Classification of Pituitary Neuroendocrine Tumors. Cancer Cell 2020, 37, 123–134.e5. [Google Scholar] [CrossRef] [PubMed]
  35. Varis, A.; Wolf, M.; Monni, O.; Vakkari, M.-L.; Kokkola, A.; Moskaluk, C.; Frierson, H.; Powell, S.M.; Knuutila, S.; Kallioniemi, A.; et al. Targets of gene amplification and overexpression at 17q in gastric cancer. Cancer Res. 2002, 62, 2625–2629. [Google Scholar] [PubMed]
  36. Shlien, A.; Tabori, U.; Marshall, C.; Pienkowska, M.; Feuk, L.; Novokmet, A.; Nanda, S.; Druker, H.; Scherer, S.W.; Malkin, D. Excessive genomic DNA copy number variation in the Li-Fraumeni cancer predisposition syndrome. Proc. Natl. Acad. Sci. USA 2008, 105, 11264–11269. [Google Scholar] [CrossRef] [PubMed]
  37. McCormack, A.; Dekkers, O.; Petersenn, S.; Popovic, V.; Trouillas, J.; Raverot, G.; Burman, P. Treatment of aggressive pituitary tumours and carcinomas: Results of a European Society of Endocrinology (ESE) survey 2016. Eur. J. Endocrinol. 2018, 178, 265–276. [Google Scholar] [CrossRef]
  38. Trouillas, J.; Jaffrain-Rea, M.; Vasiljevic, A.; Raverot, G.; Roncaroli, F.; Villa, C. How to Classify the Pituitary Neuroendocrine Tumors (PitNET)s in 2020. Cancers 2020, 12, 514. [Google Scholar] [CrossRef]
  39. Hu, H.; Khodadadi-Jamayran, A.; Dolgalev, I.; Cho, H.; Badri, S.; Chiriboga, L.A.; Zeck, B.; Gregorio, M.L.D.R.; Dowling, C.M.; Labbe, K.; et al. Targeting the Atf7ip-Setdb1 Complex Augments Antitumor Immunity by Boosting Tumor Immunogenicity. Cancer Immunol. Res. 2021, 9, 1298–1315. [Google Scholar] [CrossRef]
  40. Park, J.; Huang, S.; Tougeron, D.; Sinicrope, F. MSH3 mismatch repair protein regulates sensitivity to cytotoxic drugs and a histone deacetylase inhibitor in human colon carcinoma cells. PLoS ONE 2013, 8, e65369. [Google Scholar] [CrossRef]
  41. Raverot, G.; Ilie, M.; Lasolle, H.; Amodru, V.; Trouillas, J.; Castinetti, F.; Brue, T. Aggressive pituitary tumours and pituitary carcinomas. Nat. Rev. Endocrinol. 2021, 17, 671–684. [Google Scholar] [CrossRef]
  42. Syro, L.; Rotondo, F.; Camargo, M.; Ortiz, L.; Serna, C.; Kovacs, K. Temozolomide and Pituitary Tumors: Current Understanding, Unresolved Issues, and Future Directions. Front. Endocrinol. 2018, 9, 318. [Google Scholar] [CrossRef]
  43. Mamidi, T.K.K.; Wu, J.; Hicks, C. Integrating germline and somatic variation information using genomic data for the discovery of biomarkers in prostate cancer. BMC Cancer 2019, 19, 229. [Google Scholar] [CrossRef]
  44. Caja, F.; Vodickova, L.; Kral, J.; Vymetalkova, V.; Naccarati, A.; Vodicka, P. Mismatch repair gene variant in sporadic solid cancers. Int. J. Mol. Sci. 2020, 21, 5561. [Google Scholar] [CrossRef]
  45. Song, G.G.; Kim, J.H.; Lee, H. Genome-wide pathway analysis in major depresive disorder. J. Mol. Neurosci. 2013, 51, 428–436. [Google Scholar] [CrossRef] [PubMed]
  46. GATK. Available online: https://gatk.broadinstitute.org/ (accessed on 6 October 2021).
  47. GDC. Available online: https://docs.gdc.cancer.gov/ (accessed on 6 October 2021).
  48. Chang, M.; Yang, C.; Bao, X.; Wang, R. Genetic and Epigenetic Causes of Pituitary Adenomas. Front. Endocrinol. 2021, 11, 596554. [Google Scholar] [CrossRef] [PubMed]
  49. Ballmann, C.; Thiel, A.; Korah, H.E.; Reis, A.C.; Saeger, W.; Stepanow, S.; Köhrer, K.; Reifenberger, G.; Knobbe-Thomsen, C.B.; Knappe, U.J.; et al. USP8 Mutations in Pituitary Cushing Adenomas-Targeted Analysis by Next-Generation Sequencing. J. Endocr. Soc. 2018, 2, 266–278. [Google Scholar] [CrossRef] [PubMed]
  50. Naidoo, P.; Naidoo, R.; Ramkaran, P.; Chuturgoon, A. Effect of maternal HIV infection, BMI and NOx air pollution exposure on birth outcomes in South African pregnant women genotyped for the p53 Pro72Arg (rs1042522). Int. J. Immunogenet. 2020, 47, 414–429. [Google Scholar] [CrossRef] [PubMed]
  51. Leite, M.; Giacomin, L.; Piranda, D.; Festa-Vasconcellos, J.; Indio-do-Brasil, V.; Koifman, S.; de Moura-Neto, R.S.; de Carvalho, M.A.; Vianna-Jorge, R. Epidermal growth factor receptor gene polymorphisms are associated with prognostic features of breast cancer. BMC Cancer 2014, 14, 190. [Google Scholar] [CrossRef]
  52. Baumann, A.; Buchberger, A.; Piontek, G.; Schüttler, D.; Rudelius, M.; Reiter, R.; Gebel, L.; Piendl, G.; Brockhoff, G.; Pickhard, A. The Aurora-Kinase A Phe31-Ile polymorphism as possible predictor of response to treatment in head and neck squamous cell carcinoma. Oncotarget 2018, 9, 12769–12780. [Google Scholar] [CrossRef]
  53. Vargas-Torres, S.L.; Portari, E.A.; Silva, A.L.; Klumb, E.M.; da Rocha Guillobel, H.C.; de Camargo, M.J.; Santos-Rebouças, C.B.; Russomano, F.B.; Macedo, J.M.B. Roles of CDKN1A gene polymorphisms (rs1801270 and rs1059234) in the development of cervical neoplasia. Tumour Biol. 2016, 37, 10469–10478. [Google Scholar] [CrossRef]
  54. Taniguchi-Ponciano, K.; Andonegui-Elguera, S.; Peña-Martínez, E.; Silva-Román, G.; Vela-Patiño, S.; Gomez-Apo, E.; Chavez-Macias, L.; Vargas-Ortega, G.; Espinosa-de-Los-Monteros, L.; Gonzalez-Virla, B.; et al. Transcriptome and methylome analysis reveals three cellular origins of pituitary tumors. Sci. Rep. 2020, 10, 19373. [Google Scholar] [CrossRef]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Copeptin Levels Before and After Transsphenoidal Surgery for Cushing Disease: A Potential Early Marker of Remission

Abstract

Context

Arginine-vasopressin and CRH act synergistically to stimulate secretion of ACTH. There is evidence that glucocorticoids act via negative feedback to suppress arginine-vasopressin secretion.

Objective

Our hypothesis was that a postoperative increase in plasma copeptin may serve as a marker of remission of Cushing disease (CD).

Design

Plasma copeptin was obtained in patients with CD before and daily on postoperative days 1 through 8 after transsphenoidal surgery. Peak postoperative copeptin levels and Δcopeptin values were compared among those in remission vs no remission.

Results

Forty-four patients (64% female, aged 7-55 years) were included, and 19 developed neither diabetes insipidus (DI) or syndrome of inappropriate anti-diuresis (SIADH). Thirty-three had follow-up at least 3 months postoperatively. There was no difference in peak postoperative copeptin in remission (6.1 pmol/L [4.3-12.1]) vs no remission (7.3 pmol/L [5.4-8.4], P = 0.88). Excluding those who developed DI or SIADH, there was no difference in peak postoperative copeptin in remission (10.2 pmol/L [6.9-21.0]) vs no remission (5.4 pmol/L [4.6-7.3], P = 0.20). However, a higher peak postoperative copeptin level was found in those in remission (14.6 pmol/L [±10.9] vs 5.8 (±1.4), P = 0.03]) with parametric testing. There was no difference in the Δcopeptin by remission status.

Conclusions

A difference in peak postoperative plasma copeptin as an early marker to predict remission of CD was not consistently present, although the data point to the need for a larger sample size to further evaluate this. However, the utility of this test may be limited to those who develop neither DI nor SIADH postoperatively.

Arginine vasopressin (AVP) and CRH act synergistically as the primary stimuli for secretion of ACTH, leading to release of cortisol [12]. The role of AVP in the hypothalamic-pituitary-adrenal (HPA) axis is via release from the parvocellular neurons of the paraventricular nuclei (and possibly also from the magnocellular neurons of the paraventricular and supraoptic nuclei), the secretion of which is stimulated by stress [3-6]. AVP release results in both independent stimulation of ACTH release and potentiation of the effects of CRH [37-9]. Additionally, there is evidence that glucocorticoids act by way of negative feedback to suppress AVP secretion [1011-20]. Further, parvocellular neurons of the hypothalamic paraventricular nuclei have been shown to increase AVP production and neurosecretory granule size after adrenalectomy, and inappropriately elevated plasma AVP has been reported in the setting of adrenal insufficiency with normalization of plasma AVP after glucocorticoid administration [21-24]. This relationship of AVP and its effect on the HPA axis has been used in the diagnostic evaluation of Cushing syndrome (CS) [14] and evaluation of remission after transsphenoidal surgery (TSS) in Cushing disease (CD) by administration of desmopressin [25].

Copeptin makes up the C-terminal portion of the AVP precursor pre-pro-AVP. Copeptin is released from the posterior pituitary in stoichiometric amounts with AVP, and because of its longer half-life in circulation, it is a stable surrogate marker of AVP secretion [26-28]. Plasma copeptin has been studied in various conditions of the anterior pituitary. In a study by Lewandowski et al, plasma copeptin was measured after administration of CRH in assessment of HPA-axis function in patients with a variety of pituitary diseases. An increase in plasma copeptin was observed only in healthy subjects but not in those with pituitary disease who had an appropriately stimulated serum cortisol, and the authors concluded that copeptin may be a sensitive marker to reveal subtle alterations in the regulation of pituitary function [7]. Although in this study and others, plasma copeptin was assessed after pituitary surgery, it has not, to the best of our knowledge, been studied as a marker of remission of CD before and after pituitary surgery [729].

In this study, plasma copeptin levels were assessed as a surrogate of AVP secretion before and after TSS for treatment of CD. Because there is evidence that glucocorticoids exert negative feedback on AVP, we hypothesized that there would be a greater postoperative increase in plasma copeptin in those with CD in remission after TSS resulting from resolution of hypercortisolemia and resultant hypocortisolemia compared with those not in remission with persistent hypercortisolemia and continued negative feedback. In other words, we hypothesized that an increase in copeptin could be an early marker of remission of CD after TSS. We aimed to complete this assessment by comparison of the peak postoperative copeptin and change in copeptin from preoperative to peak postoperative copeptin for those in remission vs not in remission postoperatively.

Subjects and Methods

Subjects

Adult and pediatric patients with CD who presented at the Eunice Kennedy Shriver National Institute of Child Health and Human Development under protocol 97-CH-0076 and underwent TSS between March 2016 and July 2019 were included in the study. Exclusion criteria included a prior TSS within 6 weeks of the preoperative plasma copeptin sample or a preoperative diagnosis of diabetes insipidus, renal disease, or cardiac failure. Written informed consent was provided by patients aged 18 years and older and by legal guardians for patients aged < 18 years to participate in this study. Written informed assent was provided by patients aged 7 years to < 18 years. The 97-CH-0076 study (Investigation of Pituitary Tumors and Related Hypothalamic Disorders) has been approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development institutional review board.

Clinical and Biochemical Data

Clinical data were extracted from electronic medical records. Age, sex, body weight, body mass index (BMI), pubertal stage (in pediatric patients only), and history of prior TSS were obtained preoperatively during the admission for TSS. Clinical data obtained postoperatively included TSS date, histology, development of central diabetes insipidus (DI) or (SIADH), time from TSS to most recent follow-up, and clinical remission status at postoperative follow-up.

Preoperatively, serum sodium, 24-hour urinary free cortisol (UFC), UFC times the upper limit of normal (UFC × ULN), midnight (MN) serum cortisol, MN plasma ACTH, and 8 AM plasma ACTH were collected. Postoperatively, serum sodium, serum and urine osmolality, urine specific gravity, serum cortisol, and plasma ACTH were collected. For serum cortisol values < 1 mcg/dL, a value of 0.5 mcg/dL was assigned for the analyses; for plasma ACTH levels < 5 pg/mL, a value of 2.5 pg/mL was assigned.

Additionally, plasma copeptin levels were obtained preoperatively and on postoperative days (PODs) 1 through 8 after TSS at 8:00 AM. Peak postoperative copeptin was the highest plasma copeptin on PODs 1 through 8. The delta copeptin (Δcopeptin) was determined by subtracting the preoperative copeptin from the peak postoperative copeptin; hence, a positive change indicated a postoperative increase in plasma copeptin. Plasma copeptin was measured using an automated immunofluorescent sandwich assay on the BRAHMS Kryptor Compact PLUS Copeptin-proAVP. The limit of detection for the assay was 1.58 pmol/L, 5.7% intra-assay coefficient of variation, and 11.2% inter-assay coefficient of variation, with a lower limit of analytical measurement of 2.8 pmol/L. For those with multiple preoperative plasma copeptin values within days before surgery, an average of preoperative copeptin levels was used for analyses.

Diagnosis of CD was based on guidelines published by the Endocrine Society and as previously described for the adult and pediatric populations [3031]; diagnosis was further confirmed by either histologic identification of an ACTH-secreting pituitary adenoma in the resected tumor specimen, decrease in cortisol and ACTH levels postoperatively, and/or clinical remission after TSS at follow-up evaluation. All patients were treated with TSS at the National Institutes of Health Clinical Center by the same neurosurgeon. Remission after surgical therapy was based on serum cortisol of < 5 μg/dL during the immediate postoperative period, improvement of clinical signs and symptoms of cortisol excess at postoperative follow up, nonelevated 24-hour UFC at postoperative follow-up, nonelevated midnight serum cortisol at postoperative follow up when available, and continued requirement for glucocorticoid replacement at 3 to 6 months’ postoperative follow-up.

Diagnosis of SIADH was based on development of hyponatremia (serum sodium < 135 mmol/L) and oliguria (urine output < 0.5 mL/kg/h). Diagnosis of DI was determined by development of hypernatremia (serum sodium > 145 mmol/L), dilute polyuria (urine output > 4 mL/kg/h), elevated serum osmolality, and low urine osmolality.

Statistical Analyses

Results are presented as median (interquartile range [IQR], calculated as 25th percentile-75th percentile) or mean ± SD, as appropriate, and frequency (percentage). Where appropriate, we compared results using parametric or nonparametric testing; however, the median (IQR) and the mean ± SD were both reported to allow for comparisons with the appropriate testing noted. Subgroup analyses were completed comparing those who developed water balance disorders included patients who developed DI only (but not SIADH), those who developed SIADH only (but not DI), and those with no water balance disorder; hence, for these subgroup analyses, those who developed both DI and SIADH postoperatively (n = 4) were excluded. Preoperative copeptin, peak postoperative copeptin, and Δcopeptin were compared between those with and without remission at follow-up, using either t test or Wilcoxon rank-sum test, depending on the distribution of data. These were done in all patients combined, as well as within each subgroup. The same tests were used for comparing other continuous variables (eg, age, BMI SD score [SDS], cortisol excess measures) between those with and without remission. Categorical data (eg, sex, Tanner stage) were analyzed using the Fisher exact test. Comparisons of copeptin levels among the subgroups (DI, SIADH, neither) were carried out using mixed models and the Kruskal-Wallis test, as appropriate. Post hoc pairwise comparisons were adjusted for multiplicity using the Bonferroni correction, and as applicable, only corrected P values are reported. Mixed models for repeated measures also analyzed copeptin, serum sodium, and cortisol data for PODs 1 through 8. In addition, maximum likelihood estimation (GENMOD) procedures analyzed the effects of copeptin and serum sodium on the remission at follow-up. Correlation analyses were done with Spearman ρ. All analyses were tested for the potential confounding effects of age, sex, BMI SDS, and pubertal status, and were adjusted accordingly. For plasma copeptin reported as < 2.8 pmol/L, a value of 1.4 pmol/L (midpoint of 0 and 2.8 pmol/L) was used; sensitivity analyses repeated all relevant comparisons using the threshold limit of 2.8 pmol/L instead of 1.4 pmol/L. Odds ratios (OR) and 95% CIs, other magnitudes of the effect, data variability, and 2-sided P values provided the statistical evidence for the conclusions. Statistical analyses were performed in SAS version 9.4 software (SAS Institute, Inc, Cary, NC).

Results

Patient Characteristics

Forty-four adult and pediatric patients, aged 7 to 55 years (77.2% were < 18 years old), with CD were included in the study. The cohort included 28 female patients (64%), and the median BMI SDS was 2.2 (1.1-2.5). Thirty-four percent (15/44) had prior pituitary surgery (none within the prior 6 weeks). Seventy-five percent (33/44) had postoperative follow-up evaluations available, with median follow-up of 13.5 months (11.3-16.0). Of those 33 patients, 85% were determined to be in remission at follow-up. Comparing those in remission vs no remission, there was no difference in age, sex, BMI SDS, pubertal status (in pediatric ages only), preoperative measures of cortisol excess (UFC × ULN, PM serum cortisol, MN plasma ACTH, AM plasma ACTH), duration of follow-up, or development of DI or SIADH. There was a lower postoperative serum cortisol nadir in those in remission at follow-up compared with those not in remission at follow-up, as expected, because a postoperative serum cortisol < 5 μg/dL was included in defining remission status. Postoperatively, 8/44 (18%) developed DI, 13/44 (30%) developed SIADH, 4/44 (9%) developed both DI and SIADH, and 19/44 (43%) developed no water balance disorder (Table 1). There were no differences by remission status when assessing these subgroups (ie, DI, SIADH, and no water balance disorder) separately.

 

Table 1.

Demographic and clinical characteristics of subjects

All subjects, n = 44 All subjects by remission status, n = 33 All subjects by remission status, excluding those with DI or SIADH, n = 13
Remission, n = 28 No remission, n = 5 P Remission,
n = 10
No remission, n = 3 P
Age, median (range), y 14.5 (7-55) 17.4 ± 10.7
14.5 (12.5-17.5)
15.6 ± 13.2
11.0 (9.0-12.0)
0.11 13.7 ± 3.1
14.0 (13.0-15.0)
19.7 ± 16.8
11.0 (9.0-39.0)
0.60a
Sex
Female
28 (64%) 22 (78.6%) 3 (60.0%) 0.57 9 (90.0%) 2 (66.7%) 0.42
BMI SDS 2.2 (1.1-2.5) 1.7 ± 1.0
2.0 (0.9-2.5)
2.2 ± 0.4
2.2 (2.1-2.3)
0.70 1.7 ± 1.1
2.0 (0.7-2.5)
2.0 ± 0.4
2.1 (1.5-2.3)
0.65a
Pubertal status
Female (n = 19) (n = 15) (n = 2) 0.51 (n = 8) (n = 1) 0.44
  Tanner 1-2 6 4 (26.7%) 1 (50.0%) 3 (37.5%) 1 (25.0%)
  Tanner 3-5 13 11 (73.3%) 1 (50.0%) 5 (62.5%) 0
Male (n = 14) (n = 5) (n = 2) (n = 1) (n = 1)
Testicular volume < 12, mL 10 4 (80.0%) 2 (10.00%) 1 (100.0%) 1 (100.0%)
Testicular volume ≥ 12, mL 4 1 (20.0%) 0 1.0 0 0
Preoperative UFC ULN 3.3 (1.2-6.1) 4.9 ± 6.1
2.6 (1.0-7.6)
3.2 ± 1.3
3.7 (3.0-3.9)
0.70 7.2 ± 8.4
3.9 (1.8-9.1)
3.8 ± 0.7
3.9 (3.0-4.4)
0.93
Preoperative PM cortisol 11.9 (9.2-14.8) 13.3 ± 4.7
12.2 (9.2-16.8)
10.8 ± 2.1
11.5 (9.0-11.6)
0.30 13.3 ± 6.0
11.2 (8.4-16.5)
11.1 ± 2.6
11.6 (8.3-13.6)
0.57a
Preoperative MN ACTH 43.4 (29.3-51.6) 44.2 ± 25.5
46.1 (27.6-50.5)
40.9 ± 15.3
11.5 (9.0-11.6)
0.74 36.6 ± 16.6
37.4 (29.1-48.8)
34.0 ± 9.4
39.3 (23.1-39.5)
0.67
Preoperative AM ACTH 44.6 (31.4-60.5) 46.9 ± 28.9
44.0 (29.8-56.2)
48.6 ± 28.8
58.7 (21.7-60.5)
0.84 35.2 ± 16.2
40.3 (28.0-44.0)
45.4 ± 24.6
58.7 (17.0-60.5)
0.41a
Postoperative cortisol nadir 0.5 (0.5-0.5) 0.7 ± 0.7
0.5 (0.5-0.5)
7.8 ± 6.6
5.2 (2.2-12.3)
<0.001 0.6 ± 0.3
0.5 (0.5-0.5)
8.1 ± 7.9
5.2 (2.1-17.0)
0.003
Duration of follow-up 13.5 (11.3-16.0) 15.3 ± 7.9
14.0 (12.0-16.5)
14.0 ± 13.0
11.0 (6.0-14.0)
0.30 18.6 ± 11.2
15.5 (12.0-27.0)
16.7 ± 17.2
11.0 (3.0-36.0)
0.82a
DI only 8 (18%) 7/8 (87.5%) 1/8 (12.5%) 0.91
SIADH only 13 (30%) 8/9 (88.9%) 1/9 (11.1%)
Neither DI/SIADH 19 (43%) 10/13 (76.9%) 3/13 (23.1%)
Both DI and SIADH 4 (9%) 3/3 (100%) 0/3

Demographic and clinical characteristics of all subjects (n = 44) with Cushing disease. Data are also presented by remission status for all subjects with postoperative follow-up (n = 33) and by remission status after excluding those who developed DI or SIADH postoperatively with postoperative follow-up (n = 13). Both median (IQR) and mean ± SD reported to allow for comparisons, with P value provided using appropriate testing depending on distribution of data sets. Data are mean ± SD, median (25th-75th IQR), or frequency (percentage) are reported, except for age, which is presented as median (range).

Abbreviations: AM, 7:30-8 PM; BMI, body mass index; DI, diabetes insipidus; IQR, interquartile range; MN, midnight; N/A, not applicable; SDS, SD score; SIADH, syndrome of inappropriate antidiuresis; UFC, urinary free cortisol; ULN, upper limit of normal. p-values below the threshold of 0.05 are in bold.

aP value indicates comparison using parametric testing, as appropriate for normally distributed data.

Preoperative copeptin levels were higher in males (7.0 pmol/L [5.1-9.6]) than in females (4.0 pmol/L [1.4-5.8], P = 0.004) (Fig. 1). Age was inversely correlated with preoperative copeptin (rs = -0.35, P = 0.030) and BMI SDS was positively correlated with preoperative copeptin (rs = 0.54, P < 0.001) (Fig. 2).

 

Figure 1.

Preoperative plasma copeptin and sex. Preoperative plasma copeptin in all patients, comparing by sex. A higher preoperative plasma copeptin was found in males (7.0 pmol/L [5.1-9.6]) than in females (4.0 pmol/L [1.4-5.8], P = 0.004). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

Preoperative plasma copeptin and sex. Preoperative plasma copeptin in all patients, comparing by sex. A higher preoperative plasma copeptin was found in males (7.0 pmol/L [5.1-9.6]) than in females (4.0 pmol/L [1.4-5.8], P = 0.004). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

 

Figure 2.

Preoperative plasma copeptin and BMI SDS. Association of preoperative plasma copeptin and BMI SDS in all patients. A BMI SDS was positively associated with a preoperative plasma copeptin (rs = 0.54, P < 0.001). Shaded area = 95% confidence interval.

Preoperative plasma copeptin and BMI SDS. Association of preoperative plasma copeptin and BMI SDS in all patients. A BMI SDS was positively associated with a preoperative plasma copeptin (rs = 0.54, P < 0.001). Shaded area = 95% confidence interval.

Copeptin Before and After Transsphenoidal Surgery for CD

Among the 33 patients with postoperative follow-up, there was no difference in peak postoperative copeptin for patients in remission vs those not in remission (6.1 pmol/L [4.3-12.1] vs 7.3 pmol/L [5.4-8.4], P = 0.88). There was also no difference in the Δcopeptin for those in remission vs not in remission (2.3 pmol/L [-0.5 to 8.2] vs 0.1 pmol/L [-0.1 to 2.2], P = 0.46) (Fig. 3). Including all subjects, the mean preoperative copeptin was 5.6 pmol/L (±3.4). For patients with follow-up, there was no difference in preoperative copeptin for those in remission (4.8 pmol/L [±2.9]) vs no remission (6.0 pmol/L [±2.0], P = 0.47). POD 1 plasma copeptin ranged from < 2.8 to 11.3 pmol/L.

 

Figure 3.

(A) Peak postoperative plasma copeptin in all patients, comparing those in remission with no remission (6.1 pmol/L [4.3-12.1] vs 7.3 pmol/L [5.4-8.4], P = 0.88). (B) ΔCopeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin) in all patients, comparing those in remission with no remission (2.3 pmol/L [-0.5 to 8.2] vs 0.1 pmol/L [-0.1 to 2.2], P = 0.46). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

(A) Peak postoperative plasma copeptin in all patients, comparing those in remission with no remission (6.1 pmol/L [4.3-12.1] vs 7.3 pmol/L [5.4-8.4], P = 0.88). (B) ΔCopeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin) in all patients, comparing those in remission with no remission (2.3 pmol/L [-0.5 to 8.2] vs 0.1 pmol/L [-0.1 to 2.2], P = 0.46). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

When those who developed DI or SIADH were excluded, there was no difference in peak postoperative copeptin in those in remission vs no remission (10.2 pmol/L [6.9-21.0] vs 5.4 pmol/L [4.6-7.3], P = 0.20). However, because the distribution of the peak postoperative copeptins was borderline normally distributed, parametric testing was also completed for this analysis, which showed a higher peak postoperative copeptin in remission (14.6 pmol/L [±10.9]) vs no remission (5.8 [±1.4], P = 0.03). There was no difference in the Δcopeptin for those in remission vs not in remission (5.1 pmol/L [0.3-19.5] vs 1.1 pmol/L [-0.1 to 2.2], P = 0.39) (Fig. 4). Preoperative copeptin was not different for those in remission (4.7 pmol/L [±2.4]) vs no remission (4.9 pmol/L [±20.3], P = 0.91). There was no association between serum cortisol and plasma copeptin over time postoperatively (Fig. 5).

 

Figure 4.

(A) Peak postoperative plasma copeptin excluding those who developed DI or SIADH, comparing those in remission with no remission (10.2 pmol/L [6.9-21.0] vs 5.4 pmol/L [4.6-7.3], P = 0.20). (B) ΔCopeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin) excluding those who developed DI or SIADH, comparing those in remission with no remission (5.1 pmol/L [0.3-19.5] vs 1.1 pmol/L [-0.1 to 2.2], P = 0.39). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

(A) Peak postoperative plasma copeptin excluding those who developed DI or SIADH, comparing those in remission with no remission (10.2 pmol/L [6.9-21.0] vs 5.4 pmol/L [4.6-7.3], P = 0.20). (B) ΔCopeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin) excluding those who developed DI or SIADH, comparing those in remission with no remission (5.1 pmol/L [0.3-19.5] vs 1.1 pmol/L [-0.1 to 2.2], P = 0.39). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges.

 

Figure 5.

Plasma copeptin and serum cortisol vs postoperative day for patients who did not develop DI or SIADH. Plasma copeptin (indicated by closed circle) and serum cortisol (indicated by “x”). Results shown as (median, 95% CI).

Plasma copeptin and serum cortisol vs postoperative day for patients who did not develop DI or SIADH. Plasma copeptin (indicated by closed circle) and serum cortisol (indicated by “x”). Results shown as (median, 95% CI).

All analyses here were repeated adjusting for serum sodium, and there were no differences by remission status for preoperative, peak postoperative, or Δcopeptin for all subjects or after excluding those who developed a water balance disorder (data not shown).

Copeptin and Water Balance Disorders

As expected, peak postoperative copeptin appeared to be different among patients who developed DI, SIADH, and those without any fluid balance disorder (P = 0.029), whereas patients with DI had lower median peak postoperative copeptin (4.4 pmol/L [2.4-6.9]) than those who developed no fluid abnormality (10.0 pmol/L [5.4-16.5], P = 0.04), the statistical difference was not present after correction for multiple comparisons (P = 0.13). Peak postoperative copeptin of patients with SIADH was 9.4 pmol/L (6.5-10.4) and did not differ from patients with DI (P = 0.32) or those with no fluid abnormality (P = 1.0). There was a difference in Δcopeptin levels among these subgroups (overall P = 0.043), which appeared to be driven by the lower Δcopeptin in those who developed DI (-1.2 pmol/L [-2.6 to 0.1]) vs in those with neither DI or SIADH (3.1 pmol/L [0-9.6], P = 0.05). However, this pairwise comparison did not reach statistical significance, even before correction for multiple comparisons (P = 0.16) (Fig. 6). Preoperative copeptin levels were also not different among the subgroups (P = 0.54).

 

Figure 6.

(A) Peak postoperative plasma copeptin, comparing those who developed DI, SIADH, or neither (P = 0.029 for comparison of all 3 groups). (B) ∆ Copeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin), comparing those who developed DI, SIADH, or neither (P = 0.043 for comparison of all 3 groups). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges. Top brackets = pairwise comparisons. P values presented are after Bonferroni correction for multiple comparisons.

(A) Peak postoperative plasma copeptin, comparing those who developed DI, SIADH, or neither (P = 0.029 for comparison of all 3 groups). (B) ∆ Copeptin (preoperative plasma copeptin subtracted from postoperative peak plasma copeptin), comparing those who developed DI, SIADH, or neither (P = 0.043 for comparison of all 3 groups). Horizontal lines = median. Whiskers = 25th and 75th interquartile ranges. Top brackets = pairwise comparisons. P values presented are after Bonferroni correction for multiple comparisons.

Association of Sodium and Copeptin

Longitudinal data, adjusting for subgroups (ie, DI, SIADH, neither), were analyzed. As expected, there was a group difference (P = 0.003) in serum sodium over time (all DI was missing preoperative serum sodium), with the difference being driven by DI vs SIADH (P = 0.007), and SIADH vs neither (P = 0.012). There was no group difference in plasma copeptin over POD by water balance status (P = 0.16) over time (Fig. 7). There was also no effect by remission status at 3 to 6 months for either serum sodium or plasma copeptin.

 

Figure 7.

(A) Serum sodium and (B) plasma copeptin by POD and water balance status longitudinal data, adjusting for subgroups (ie, DI, SIADH, neither). Data points at point 0 on the x-axis indicate preoperative values. As expected, there was a group difference (P = 0.003) in serum sodium over time (all with DI were missing preoperative serum sodium), with the difference being driven by DI vs SIADH (P = 0.007), and SIADH vs neither (P = 0.012). There was no group difference in plasma copeptin over POD by water balance status (P = 0.16) over time.

(A) Serum sodium and (B) plasma copeptin by POD and water balance status longitudinal data, adjusting for subgroups (ie, DI, SIADH, neither). Data points at point 0 on the x-axis indicate preoperative values. As expected, there was a group difference (P = 0.003) in serum sodium over time (all with DI were missing preoperative serum sodium), with the difference being driven by DI vs SIADH (P = 0.007), and SIADH vs neither (P = 0.012). There was no group difference in plasma copeptin over POD by water balance status (P = 0.16) over time.

Higher serum sodium levels from PODs 1 through 8 itself decreased the odds of remission (OR, 0.56; 95% CI, 0.42-0.73; P < 0.001) in all CD patients. Copeptin levels from these repeated measures adjusting for serum sodium did not correlate with remission status at 3 to 6 months’ follow-up (P = 0.38). There were no differences in preoperative, peak postoperative, or delta sodium levels by remission vs no remission in all patients and in those with no water balance disorders.

Discussion

AVP and CRH act synergistically to stimulate the secretion of ACTH and ultimately cortisol [12], and there is evidence that glucocorticoids act by way of negative feedback to suppress AVP secretion [1011-20]. Therefore, we hypothesized that a greater postoperative increase in plasma copeptin in those with CD in remission after TSS because of resolution of hypercortisolemia and resultant hypocortisolemia, compared with those not in remission with persistent hypercortisolemia and continued negative feedback, would be observed. Although a clear difference in peak postoperative and Δcopeptin was not observed in this study, a higher peak postoperative copeptin was found in those in remission after excluding those who developed DI/SIADH when analyzing this comparison with parametric testing, and it is possible that we did not have the power to detect a difference by nonparametric testing, given our small sample size. Therefore, postoperative plasma copeptin may be a useful early marker to predict remission of CD after TSS. The utility of this test may be limited to those who do not develop water balance disorders postoperatively. If a true increase in copeptin occurs for those in remission after treatment of CD, it is possible that this could be due to the removal of negative feedback from cortisol excess on pre-pro-AVP secretion, as hypothesized in this study. However, it is also possible that other factors may contribute to an increase in copeptin postoperatively, including from the stress response of surgery and postoperative hypocortisolism and resultant stimulation of pre-pro-AVP secretion from these physical stressors and/or from unrecognized SIADH.

It was anticipated that more severe hypercortisolism to be negatively correlated with preoperative plasma copeptin because of greater negative feedback on AVP. However, no association was found between preoperative plasma copeptin and markers of severity of hypercortisolism (MN cortisol, AM ACTH, UFC × ULN) in this study. Similarly, we would expect that the preoperative plasma copeptin would be lower compared with healthy individuals. However, comparisons of healthy individuals may be difficult because the fluid and osmolality status at the time of the sample could influence the plasma copeptin, and depending on those factors, copeptin could be appropriately low. A healthy control group with whom to compare the preoperative values was not available for this study, and the thirsted state was not standardized for the preoperative copeptin measurements. Future studies could be considered to determine if preoperative plasma copeptin is lower in patients with CD, or other forms of CS, compared with healthy subjects, with all subjects thirsted for an equivalent period. Further, if preoperative plasma copeptin is found to be lower in thirsted subjects with CS than a thirsted healthy control group, the plasma copeptin could potentially be a diagnostic test to lend support for or against the diagnosis of endogenous CS.

In the comparisons of those who developed DI, SIADH, or neither, no difference was found in the Δcopeptin. Peak copeptin was lower in DI compared with those without DI or SIADH (but not different from SIADH). Again, it is possible that there is a lower peak postoperative copeptin and change in copeptin in those with DI, but we may not have had the power to detect this in all of our analyses. These comparisons of copeptin among those with or without water balance disorders postoperatively are somewhat consistent with a prior study showing postoperative copeptin as a good predictor of development of DI, in which a plasma copeptin < 2.5 pmol/L measured on POD 0 accurately identified those who developed DI, and plasma copeptin > 30 pmol/L ruled out the development of DI postoperatively [29]. In the current study, 3 of 6 subjects with DI had a POD 1 plasma copeptin < 2.5 pmol/L, and none had a POD 1 plasma copeptin > 30 pmol/L. However, the study by Winzeler et al found that copeptin measured on POD 0 (within 12 hours after surgery) had the greatest predictive value, and POD 0 plasma copeptin was not available in our study. Further, we used the preoperative, peak, and delta plasma copeptin for analyses, so the early low copeptin levels may not have been captured in our data and analyses.

Additionally, this study revealed that increasing levels of serum sodium have lower odds of remission. Those who have an ACTH-producing adenoma that is not identified by magnetic resonance imaging and visual inspection intraoperatively have lower rates of remission and are more likely to have greater manipulation of the pituitary gland intraoperatively [32-36], and the latter may result in greater damage to the pituitary stalk or posterior pituitary, increasing the risk for development of DI and resultant hypernatremia.

A higher preoperative copeptin was associated with male sex and increasing BMI SDS. Increasing preoperative copeptin was also found in pubertal boys compared with pubertal girls, with no difference in copeptin between prepubertal boys and girls. It is particularly interesting to note that these associations were only in the preoperative plasma copeptin levels, but not the postoperative peak copeptin or Δcopeptin. Because the association of higher plasma in adult males and pubertal males in comparison to adult females and pubertal females, respectively, have been reported by others [2637-40], it raises the question of a change in the association of sex and BMI with plasma copeptin in the postoperative state. An effect of BMI or sex was not found by remission status, so it does not seem that the postoperative hypocortisolemic state for those in remission could explain this loss of association. However, this study may not have been powered to detect this.

Strengths of this study include the prospective nature of the study. Further, this is the first study assessing the utility of copeptin to predict remission after treatment of CD. Limitations of this study include the small sample size because of the rarity of the condition, difficulty in clinically diagnosing DI and SIADH, potential effect of post-TSS fluid balance disorders (particularly for those who may have developed transient partial DI or transient SIADH), lack of long-term follow-up, lack of any postoperative follow-up in 11 of the 44 total subjects, as well the observational nature of the study. Further, it is possible that pubertal status, sex, and BMI may have affected copeptin levels, which may have not been consistently detected because of lack of power. Lack of data on the timing of hydrocortisone replacement is an additional limitation of this study because postoperative glucocorticoid replacement could affect AVP secretion via negative feedback. Additional studies are needed to assess to further assess the role of vasopressin and measurement of copeptin in patients before and after treatment of CD.

A clear difference in peak postoperative plasma copeptin as an early marker to predict remission of CD after TSS was not found. Further studies with larger sample sizes are needed to further evaluate postoperative plasma copeptin as an early marker to predict remission of CD, though the utility of this test may be limited to those who do not develop water balance disorders postoperatively. Future studies comparing copeptin levels before and after treatment of adrenal CS would be of particular interest because this would minimize the risk of postoperative DI or SIADH which also influence copeptin levels. Additionally, comparison of thirsted preoperative plasma copeptin in those with endogenous CS and thirsted plasma copeptin in healthy controls could potentially provide evidence of whether or not preoperative plasma copeptin is lower in patients with CD, or other forms of CS, compared with healthy subjects. Further, if this is found to be true, it could potentially be a diagnostic test to lend support for or against endogenous CS.

Abbreviations

 

  • AVP

    arginine vasopressin

  • BMI

    body mass index

  • CD

    Cushing disease

  • CS

    Cushing syndrome

  • DI

    diabetes insipidus

  • HPA

    hypothalamic-pituitary-adrenal

  • IQR

    interquartile range

  • MN

    midnight

  • OR

    odds ratio

  • POD

    postoperative day

  • SDS

    SD score

  • SIADH

    syndrome of inappropriate antidiuresis

  • TSS

    transsphenoidal surgery

  • UFC

    urinary free cortisol

  • ULN

    upper limit of normal

Acknowledgments

The authors thank the patients and their families for participating in this study.

Funding

This work was supported by the Intramural Research Program, Eunice Kennedy Shriver National Institute of Child Health & Human Development (NICHD), National Institutes of Health.

Disclosures

C.A.S. holds patents on technologies involving PRKAR1A, PDE11A, GPR101, and related genes, and his laboratory has received research funding support by Pfizer Inc. for investigations unrelated to this project. C.A.S. is associated with the following pharmaceutical companies: ELPEN, Inc., H. Lunbeck A/S, and Sync. Inc.

Clinical Trial Information

ClinicalTrials.gov registration no. NCT00001595 (registered November 4, 1999).

Data Availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Published by Oxford University Press on behalf of the Endocrine Society 2022.
This work is written by (a) US Government employee(s) and is in the public domain in the US.

Sparsely Granulated Corticotroph Pituitary Macroadenoma Presenting with Pituitary Apoplexy Resulting in Remission of Hypercortisolism

https://doi.org/10.1016/j.aace.2022.04.003Get rights and content
Under a Creative Commons license
Open access

Highlights

• We describe a rare case of a patient with a sparsely granulated corticotroph pituitary macroadenoma with pituitary apoplexy who underwent transsphenoidal resection resulting in remission of hypercortisolism.
• Corticotroph adenomas are divided into densely granulated, sparsely granulated and Crooke’s cell tumors.
• macroadenomas account for 7-23% of patients with pituitary corticotroph adenomas
• Sparsely granulated corticotroph tumors are associated with longer duration of Cushing disease prior to diagnosis, larger tumor size at diagnosis, decreased immediate remission rate, increased proliferative marker Ki-67 and increased recovery time of hypothalamic-pituitary-adrenal axis after surgery.
• Granulation pattern is an important clinicopathological distinction impacting the behavior and treatment outcomes of pituitary corticotroph adenomas

Abstract

Background

/Objective: Pituitary corticotroph macroadenomas, which account for 7% to 23% of corticotroph adenomas, rarely present with apoplexy. The objective of this report is to describe a patient with a sparsely granulated corticotroph tumor (SGCT) presenting with apoplexy and remission of hypercortisolism.

Case Report

A 33-year-old male presented via ambulance with sudden onset of severe headache and nausea/vomiting. Physical exam revealed bitemporal hemianopsia, diplopia from right-sided third cranial nerve palsy, abdominal striae, facial plethora, dorsal and supraclavicular fat pad. Magnetic resonance imaging (MRI) demonstrated a 3.2 cm mass arising from the sella turcica with hemorrhage compressing the optic chiasm, extension into the sphenoid sinus and cavernous sinus. Initial investigations revealed plasma cortisol of 64.08 mcg/dL (Reference Range (RR), 2.36 – 17.05). He underwent emergent transsphenoidal surgery. Pathology was diagnostic of SGCT. Post-operatively, cortisol was <1.8ug/dL (RR, 2.4 – 17), adrenocorticotropic hormone (ACTH) 36 pg/mL (RR, 0 – 81), thyroid stimulating hormone (TSH) 0.07 uIU/mL (RR, 0.36 – 3.74), free thyroxine 1 ng/dL (RR, 0.8 – 1.5), luteinizing hormone (LH) <1 mIU/mL (RR, 1 – 12), follicle stimulating hormone (FSH) 1 mIU/mL (RR, 1 – 12) and testosterone 28.8 ng/dL (RR, 219.2 – 905.6) with ongoing requirement for hydrocortisone, levothyroxine, testosterone replacement and continued follow-up.

Discussion

Corticotroph adenomas are divided into densely granulated, sparsely granulated and Crooke’s cell tumors. Sparsely granulated pattern is associated with larger tumor size and decreased remission rate after surgery.

Conclusion

This report illustrates a rare case of hypercortisolism remission due to apoplexy of a SGCT with subsequent central adrenal insufficiency, hypothyroidism and hypogonadism.

Keywords

pituitary apoplexy
pituitary macroadenoma
pituitary tumor
sparsely granulated corticotroph tumor
Cushing disease

Introduction

The incidence of Cushing Disease (CD) is estimated to be between 0.12 to 0.24 cases per 100,00 persons per year1,2. Of these, 7-23% are macroadenomas (>1 cm)345. Pituitary apoplexy is a potentially life-threatening endocrine and neurosurgical emergency which occurs due to infarction or hemorrhage in the pituitary gland. Apoplexy occurs most commonly in non-functioning macroadenomas with an estimated prevalence of 6.2 cases per 100,000 persons and incidence of 0.17 cases per 100,00 persons per year6. Corticotroph macroadenoma presenting with apoplexy is uncommon with only a handful of reports in the literature7. We present a case of a sparsely granulated corticotroph (SGCT) which presented with apoplexy leading to remission of hypercortisolism and subsequent central adrenal insufficiency.

Case Presentation

A 33-year-old male who was otherwise healthy and not on any medications presented to a community hospital with sudden and severe headache accompanied by hypotension, nausea, vomiting, bitemporal hemianopsia and diplopia. Computed Tomography (CT) scan of the brain demonstrated a hyperattenuating 2.0 cm x 2.8 cm x 1.5 cm mass at the sella turcica with extension into the right cavernous sinus and encasement of the right internal carotid arteries (Figure 1A). He was transferred to a tertiary care center for neurosurgical management with endocrinology consultation post-operatively.

  1. Download : Download high-res image (404KB)
  2. Download : Download full-size image

Figure 1. hyperattenuating 2.0 cm x 2.8 cm x 1.5 cm mass at the sella turcica on unenhanced CT (A); MRI demonstrated a 1.9 cm x 3.2 cm x 2.4 cm heterogeneous mass on T1 (B) and T2-weighted imaging (C) showing small hyperintense areas in solid part of the sella mass with flattening of the optic chiasm, remodeling/dehiscence of the floor of the sella and extending into the right cavernous sinus with at least partial encasement of the ICA

In retrospect, he reported a 3-year history of ongoing symptoms of hypercortisolism including increased central obesity, dorsal and supraclavicular fat pad, facial plethora, abdominal purple striae, easy bruising, fatigue, decreased libido and erectile dysfunction. Notably, at the time of presentation he did not have a history of diabetes, hypertension, osteoporosis, fragility fractures or proximal muscle weakness. He fathered 2 children previously. His physical examination was significant for Cushingoid facies, facial plethora, dorsal and supraclavicular fat pads and central obesity with significant axillary and abdominal wide purple striae (Figure 2). Neurological examination revealed bitemporal hemianopsia, right third cranial nerve palsy with ptosis and impaired extraocular movement. The fourth and sixth cranial nerves were intact as was the rest of his neurological exam. These findings were corroborated by Ophthalmology.

  1. Download : Download high-res image (477KB)
  2. Download : Download full-size image

Figure 2. Representative images illustrating facial plethora (A); abdominal striae (B, C); supraclavicular fat pad (D); dorsal fat pad (E)

Initial laboratory data at time of presentation to the hospital included elevated plasma cortisol of 64.08ug/dL (RR, 2.36 – 17.05), ACTH was not drawn at the time of presentation, normal TSH 0.89 mIU/L (RR, 0.36 – 3.74), free thyroxine 0.91ng/dL (RR, 0.76 – 1.46), evidence of central hypogonadism with low total testosterone 28.8 ng/dL (RR, 219.2 – 905.6) and inappropriately normal luteinizing hormone (LH) 1mIU/mL (RR, 1 – 12) and follicle stimulating hormone (FSH) 3mIU/mL (RR, 1 – 12), low prolactin <1 ng/mL (RR, 3 – 20), and normal insulin growth factor – 1 (IGF–1) 179ng/mL (RR, 82 – 242).

A pituitary gland dedicated MRI was performed to further characterize the mass, which re-demonstrated a 1.9 cm x 3.2 cm x 2.4 cm heterogenous mass at the sella turcica extending superiorly and flattening the optic chiasm, remodeling of the floor of the sella and bulging into the sphenoid sinus and extending laterally into the cavernous sinus with encasement of the right internal carotid artery (ICA). As per the radiologist’s diagnostic impression, this appearance was most in keeping with a pituitary macroadenoma with apoplexy (Figure 1B – C).

The patient underwent urgent TSS and decompression with no acute complications. Pathological examination of the pituitary adenoma showed features characteristic of sparsely granulated corticotroph pituitary neuroendocrine tumor (adenoma)8, with regional hemorrhage and tumor necrosis (apoplexy). The viable tumor exhibited a solid growth pattern (Figure 3A), t-box transcription factor (T-pit) nuclear immunolabeling (Figure 3B), diffuse cytoplasmic CAM5.2 (low molecular weight cytokeratin) immunolabeling (Figure 3C), and regional weak to moderate intense granular cytoplasmic ACTH immuno-staining (Figure 3D). The tumor was immuno-negative for: pituitary-specific positive transcription factor 1 (Pit-1) and steroidogenic factor 1 (SF-1) transcription factors, growth hormone, prolactin, TSH, FSH, LH, estrogen receptor-alpha, and alpha-subunit. Crooke hyalinization was not identified in an adjacent compressed fragment of non-adenomatous anterior pituitary tissue. Ki-67 immunolabeling showed a 1.5% proliferative index (11 of 726 nuclei).

  1. Download : Download high-res image (2MB)
  2. Download : Download full-size image

Figure 3. Hematoxylin phloxine saffron staining showing adenoma with solid growth pattern (A); immunohistochemical staining showing T-pit reactivity of tumor nuclei (B); diffuse cytoplasmic staining for cytokeratin CAM5.2 (C); and regional moderately intense granular cytoplasmic staining for ACTH (D). Scale bar = 20 μm

Post-operatively, he developed transient central diabetes insipidus requiring desmopressin but resolved on discharge. His postoperative cortisol was undetectable, ACTH 36 pg/mL (RR, 0 – 81), TSH 0.07 mIU/mL (RR, 0.36 – 3.74), free thyroxine 1 ng/dL (RR, 0.8 – 1.5), LH <1mIU/mL (RR, 1 – 12), FSH 1 mIU/mL (RR, 1 – 12) and testosterone 28.8 ng/dL (RR, 219.2 – 905.6) (Table 1 and Figure 4). One month later, he reported 15 pounds of weight loss and a 5-inch decrease in waist circumference. He also noted a reduction in the dorsal and supraclavicular fat pads, facial plethora, and Cushingoid facies as well as fading of the abdominal stretch marks. His visual field defects and right third cranial nerve palsy resolved on follow up with ophthalmology post-operatively. Repeat MRI six months post-operatively showed minor residual soft tissue along the floor of the sella. He is being followed by Neurosurgery, Ophthalmology, and Endocrinology for monitoring of disease recurrence, visual defects, and management of hypopituitarism.

Table 1. Pre- and post-operative hormonal panel

POD -1 POD 0 POD1 POD2 POD3 POD16 6 -9 months Comments
Cortisol(2.4 – 17 ug/dL) 64↓ 32↓ 11↓ <1.8↓ <1.8↓ 1.8↓ HC started POD3 post bloodwork
ACTH(0 – 81 pg/mL) 41↓ 36↓ 28↓ 13↓
TSH(0.36 – 3.74 uIU/mL) 0.89 0.43 0.12↓ 0.07↓ 0.05↓ 0.73
Thyroxine, free(0.8 – 1.5 ng/dL) 0.9 0.9 1.1 1 2.1↑ 1 Levothyroxine started POD4
LH(1 – 12 miU/mL) 1↓ <1↓ 1↓ 3
FSH(1 – 12 mIU/mL) 3↓ 1↓ 1↓ 3
Testosterone(219.2 – 905.6 ng/dL) 28.8↓ <20↓ 175.9↓ Testosterone replacement started as outpatient
Testosterone, free(160 – 699 pmol/L) <5.8↓ 137↓
IGF-1(82 – 242 ng/mL) 179 79
GH(fasting < 6 mIU/L) 4.5 <0.3
Prolactin(3 – 20 ng/mL) <1↓ <1↓

POD, postoperative day; HC, hydrocortisone; ACTH, adrenocorticotropic hormone; TSH, thyroid stimulating hormone; LH, luteinizing Hormone; FSH, follicle stimulating hormone; IGF-1, insulin like growth factor – 1; GH, growth hormone

  1. Download : Download high-res image (259KB)
  2. Download : Download full-size image

Figure 4. Trend of select pituitary hormonal panel with key clinical events denoted by black arrows.

Discussion

Microadenomas account for the majority of corticotroph tumors, but 7% – 23% of patients are diagnosed with a macroadenoma345. It is even rarer for a corticotroph macroadenoma to present with apoplexy with only a handful of case reports or series in the literature7. Due to its rarity, appropriate biochemical workup on presentation, such as including an ACTH with the blood work, may be omitted especially if the patient is going for emergent surgery. In this case, the undetectable prolactin can reflect loss of anterior pituitary function and also suggest a functioning corticotroph adenoma due to the inhibitory effect of long term serum glucocorticoids on prolactin secretion9. After undergoing TSS, the patient developed central adrenal insufficiency, hypothyroidism and hypogonadism requiring hormone replacement. Presumably, the development of adrenal insufficiency demonstrated the remission of hypercortisolism as a result of apoplexy and/or TSS. The ACTH remains detectable likely representing residual tumor that was not obliterated by apoplexy nor excised by TSS given it location near the carotid artery and cavernous sinus. The presence of adrenal insufficiency in the setting of detectable ACTH is not contradictory as the physiological hypothalamic-pituitary-adrenal axis has been suppressed by the long-term pathological production of ACTH. IGF-1 and prolactin also failed to recover post-operatively. In CD where the production of IGF-1 and prolactin are attenuated by elevated cortisol, it would then be expected that IGF-1 and prolactin recover after hypercortisolism remission. However, the absence of this observation in our case is likely a sequalae of the apoplexy and extensive surgery leading to pituitary hypofunction.

We also want to highlight features of the pre-operative radiographical findings which can provide valuable insight into the subsequent histology. Previous literature has shown that, on T2-weight MRI, silent corticotroph adenomas are strongly correlated with characteristic a multimicrocystic appearance while nonfunctional gonadotroph macroadenomas are not correlated with this MRI finding10. The multimicrocystic appearance is described as small hyperintense areas with hyperintense striae in the solid part of the tumor (Figure 1C)10. This is an useful predictive tool for silent corticotroph adenomas with a sensitivity of 76%, specificity of 95% and a likelihood ratio of 15.310.

The ability to distinguish between silent corticotroph macroadenoma and other macroadenomas is important for assessing rate of remission and recurrence risk. In 2017, the WHO published updated classification for pituitary tumors. In this new classification, corticotroph adenomas are further divided into densely granulated, sparsely granulated and Crooke’s cell tumors11. DGCT are intensely Periodic Acid Schiff (PAS) stain positive and exhibit strong diffuse pattern of ACTH immunoreactivity, whereas SGCT exhibit faintly positive PAS alongside weak focal ACTH immunoreactivity4,12. Crooke’s cell tumors are characterized by Crooke’s hyaline changes in more than 50% of the tumor cells4. In the literature, SGCT account for an estimated 19-29% of corticotroph adenomas131415. The clinicopathological relevance of granulation pattern in corticotroph tumors was unclear until recently.

In multiple studies examining granulation pattern and tumor size, SGCT were statistically larger13,15,16. Hence, we suspect that many of the previously labelled silent corticotroph macroadenomas in the literature were SGCT. The traditional teaching of CD has been “small tumor, big Cushing and big tumor, small Cushing” which reflects the inverse relationship between tumor size and symptomatology17. This observation appears to hold true as Doğanşen et al. found a trend towards longer duration of CD in SGCT of 34 months compared to 26 months in DGCT based on patient history13,17. It has been postulated that the underlying mechanism of the inverse relationship between tumor size and symptomatology is impaired processing of proopiomelanocortin resulting in less effective secretion of ACTH in corticotroph macroadenomas3. Doğanşen et al. also found that the recurrence rate was doubled for SGCT, while Witek et al. showed that SGCT were less likely to achieve remission postoperatively13,16.

Similar to other cases of SGCT, the diagnosis was only arrived retrospective after pathological confirmation10. Interestingly, the characteristic Crooke’s hyaline change of surrounding non-adenomatous pituitary tissue was not observed as one would expect in a state of prolonged glucocorticoid excess in this case. Although classically described, the absence of this finding does not rule out CD. As evident in a recent retrospective study where 10 out of 144 patients with CD did not have Crooke’s hyaline change18. In patients without Crooke’s hyaline change, the authors found a lower remission rate of 44.4% compared to 73.5% in patients with Crooke’s hyaline change. Together with the detectable post-operative ACTH, sparsely granulated pattern and absence of Crooke’s hyaline change in surrounding pituitary tissue, the risk of recurrence is increased. These risk factors emphasize the importance of close monitoring to ensure early detection of recurrence.

Declaration of Interests

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Conclusion

We present a case of a sparsely granulated corticotroph macroadenoma presenting with apoplexy leading to remission of hypercortisolism and development of central adrenal insufficiency, hypothyroidism and hypogonadism requiring hormone replacement.

References

Possible Good News! Effects of Tubastatin A on Adrenocorticotropic Hormone Synthesis and Proliferation of Att-20 Corticotroph Tumor Cells

  • Rie HagiwaraDepartment of Endocrinology and Metabolism, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan
  • Kazunori KageyamaDepartment of Endocrinology and Metabolism, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan
  • Yasumasa IwasakiSuzuka University of Medical Science, Suzuka 510-0293, Japan
  • Kanako NiiokaDepartment of Endocrinology and Metabolism, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan
  • Makoto DaimonDepartment of Endocrinology and Metabolism, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan
Abstract

Cushing’s disease is an endocrine disorder characterized by hypercortisolism, mainly caused by autonomous production of ACTH from pituitary adenomas. Autonomous ACTH secretion results in excess cortisol production from the adrenal glands, and corticotroph adenoma cells disrupt the normal cortisol feedback mechanism. Pan-histone deacetylase (HDAC) inhibitors inhibit cell proliferation and ACTH production in AtT-20 corticotroph tumor cells. A selective HDAC6 inhibitor has been known to exert antitumor effects and reduce adverse effects related to the inhibition of other HDACs. The current study demonstrated that the potent and selective HDAC6 inhibitor tubastatin A has inhibitory effects on proopiomelanocortin (Pomc) and pituitary tumor-transforming gene 1 (Pttg1) mRNA expression, involved in cell proliferation. The phosphorylated Akt/Akt protein levels were increased after treatment with tubastatin A. Therefore, the proliferation of corticotroph cells may be regulated through the Akt-Pttg1 pathway. Dexamethasone treatment also decreased the Pomc mRNA level. Combined tubastatin A and dexamethasone treatment showed additive effects on the Pomc mRNA level. Thus, tubastatin A may have applications in the treatment of Cushing’s disease.

Access the PDF at https://www.jstage.jst.go.jp/article/endocrj/advpub/0/advpub_EJ21-0778/_pdf/-char/en

 

%d bloggers like this: