ELSEVIER
Disponible en ligne sur ScienceDirect www.sciencedirect.com
Elsevier Masson France EM|consulte www.em-consulte.com
Annales d’Endocrinologie Annals of Endocrinology
Annales d’Endocrinologie xxx (2017) xxx-xxx
Klotz Communications 2018: Cortisol and all its disorders Genomic insights into Cushing syndrome
L’apport de la génomique dans le syndrome de Cushing Guillaume Assié a, b,*
a Service d’endocrinologie, centre de référence des maladies rares de la surrénale, Assistance publique-Hôpitaux de Paris, hôpital Cochin, 75014 Paris, France b Inserm U1016, CNRS 8104, institut Cochin, université Paris Descartes, 75014 Paris, France
Abstract
In the setting of Cushing syndrome, genomic analyses can be performed either in tumors responsible for endogenous Cushing, or in patients exposed to glucocorticoid excess. Genomics of tumors identified several new genes - including ZNRF3 in adrenocortical carcinomas, PRKACA in cortisol-producing adrenal adenomas, ARMC5 in primary macronodular adrenal hyperplasia and USP8 in pituitary corticotroph adenomas. These genes shed new lights on the mechanisms responsible for these tumors. Integrated genomic studies of adrenal carcinomas identified distinct molecular classes, with remarkably different prognostic outcome. Beyond the mechanistic novelties, a new generation of prognostic markers emerges, with potentially important impact on patients care. For the future, genomic efforts should be pursued, focusing on poorly characterized tumors responsible for Cushing syndrome - including endocrine tumors secreting ACTH. In addition, epigenomics is emerging as an outstanding set of tools for characterizing tumors, unraveling unprecedented aspects of tumorigenesis. Applying these tools to endocrine tumors responsible for Cushing syndrome may also lead to important discoveries. Genomics of patients exposed to glucocorticoid excess is an emerging research field. Proof of principle studies have been performed, identifying molecular markers of glucocorticoid excess in blood. Research efforts should now concentrate on markers of mild glucocorticoid excesses - endogenous or exogenous -, owing to their high prevalence in general population. In addition, markers of individual susceptibility to each type of glucocorticoid complication are needed. It remains to be determined whether genomics can identify such markers.
@ 2018 Elsevier Masson SAS. All rights reserved.
Keywords: Genomic; Cushing syndrome; Glucocorticoid
Résumé
Dans le cadre du syndrome de Cushing, des études basées sur la génomique ont été réalisées pour caractériser soit les tumeurs responsables du syndrome de Cushing endogène, soit les patients exposés à un excès de glucocorticoïdes. La génomique des tumeurs a identifié plusieurs nouveaux gènes - ZNRF3 dans les carcinomes surrénaliens, PRKACA dans les adénomes surrénaliens produisant du cortisol, ARMC5 dans l’hyperplasie macronodulaire primitive des surrénales, et USP8 dans les adénomes hypophysaires corticotropes. Ces gènes apportent un nouvel éclairage sur les mécanismes responsables de ces tumeurs. Les études génomiques intégrées des carcinomes surrénaliens ont identifié des classes moléculaires distinctes, associées à un pronostic remarquablement différent. Au-delà des nouveautés en termes de mécanismes, une nouvelle génération de marqueurs pronostiques en ressort, avec un impact potentiellement important sur la prise en charge des patients. Dans l’avenir, les efforts génomiques doivent être poursuivis, en se concentrant sur les tumeurs responsables de syndrome de Cushing mal caractérisées - notamment les tumeurs endocrines sécrétant de l’ACTH. De plus, l’épigénomique apparaît comme un ensemble d’outils remarquables pour caractériser les tumeurs, révélant des aspects originaux de la tumorigenèse. L’application de ces outils aux tumeurs endocrines responsables du syndrome de Cushing peut également conduire à des découvertes importantes. La génomique des patients exposés à un excès de glucocorticoïdes est un domaine
* Correspondence. Service d’endocrinologie, centre de référence des maladies rares de la surrénale, Assistance publique-Hôpitaux de Paris, hôpital Cochin, 75014 Paris, France.
E-mail address: guillaume.assie@aphp.fr
G. Assié / Annales d’Endocrinologie xxx (2017) xxx-xxx
de recherche émergent. Des études préliminaires ont été réalisées, identifiant des marqueurs moléculaires de l’excès de glucocorticoïdes dans le sang. Les efforts de recherche devraient maintenant se concentrer sur les marqueurs d’excès de glucocorticoïdes a minima - endogènes ou exogènes - en raison de leur prévalence élevée dans la population générale. En outre, des marqueurs de susceptibilité individuelle à chaque type de complications secondaires à l’exposition aux glucocorticoïde sont nécessaires. Il reste à déterminer si la génomique peut identifier de tels marqueurs.
@ 2018 Elsevier Masson SAS. Tous droits réservés.
Mots clés : Génomique ; Syndrome de Cushing ; Glucocorticoïde
Genomics include a set of large-scale molecular measure- ments, applicable to any biological tissue. These techniques provide unbiased molecular profiles, and can be applied to Cush- ing syndrome in two main ways: either to tumors responsible for Cushing syndrome, or to patients exposed to glucocorti- coid excess. Such studies provide insights on pathophysiological mechanisms, and markers potentially relevant for diagnosis, prognosis and prediction to treatment response.
1. Genomic profiling of tumors responsible for endogenous Cushing
Tumors responsible for Cushing syndrome include adrenal tumors - carcinomas, adenomas, dysplasia and hyperplasia -, pituitary adenomas secreting ACTH, and endocrine tumors with paraneoplasic ACTH secretion. Genomic analyses have been applied to some of these tumor types.
1.1. Exome sequencing in tumors: it’s raining genes!
This is probably the most spectacular outcome of genomics. For almost all tumors responsible for Cushing syndrome, exome sequencing revealed new driver genes.
In adrenal cancer, exome sequencing identified ZNRF3, encoding for an ubiquitin ligase, regulating negatively Wnt/beta- catenin pathway [1,2]. These mutations are mutually exclusive from CTNNB1 mutations -encoding beta-catenin, which is com- monly mutated in these tumors as well -, and underline the importance of this pathway in adrenal cancer.
In adrenocortical adenomas, exome sequencing identified PRKACA mutations in up to 30% of patients, encoding the catalytic subunit of Protein-Kinase A [3-6]. In primary macron- odular hyperplasia, PRKACA locus duplications were also identified, and seem rare [6]. PRKACA mutations enlarge the list of mutations affecting PKA/cAMP pathway genes-GNAS, PRKAR1A, PDE8B and PDE11A- previously identified in rare forms of benign adrenal diseases associated with overt Cushing.
In primary macronodular adrenal hyperplasia, exome sequencing identified ARMC5 mutations in 25 to 50% of cases with overt Cushing and large multinodular adrenals, encoding for a protein of unknown function [7]. Mutations are germline, along with a somatic second hit specific to each adrenal nodule.
In corticotroph adenomas, exome sequencing identified USP8 mutations [8,9], encoding for a deubiquitinase, in up to 30% of cases. The precise link between USP8 and tumorigenesis remains to be elucidated.
1.2. Integrated genomics of tumors
The molecular landscape of adrenocortical carcinomas has been extensively characterized by studies combining genomic approaches [1,2]. Distinct molecular classes have emerged, based on important differences in terms of gene expression sig- nature, DNA methylation, chromosomal alteration profiles and mutations. Remarkably, these molecular classes are associated with major differences in terms of outcome.
We recently achieved an integrated genomic analysis of benign adrenal tumors, including all types of adenomas, hyperplasias and dysplasias (unpublished). Different genomic approaches converged into molecular classification showing three distinct molecular signatures associated with cortisol secretion:
· one associated with PKA/cAMP pathway alterations, includ- ing tumors responsible for overt Cushing, irrespective of their tumor types;
· one associated with CTNNB1 mutations, including adenomas either non-secreting or with subclinical Cushing;
. and a third class associated with ARMC5-mutated macron- odular adrenal hyperplasia.
1.3. Genomic alterations in tumors: and so what?
Genomics should be considered as a molecular microscope. Genomic characterization of tumors provides an ultimate unbi- ased molecular classification of tumors. This comes along with optical microscopes from pathologists, which for a long time was the gold standard for classifying tumors. In addition, in Endocrinology, the hormonal secretion phenotype adds to tumor characterization. In addition to pathology and hormonology, genomic classifications add valuable additional insight in the characterization of endocrine tumors responsible for Cushing syndrome.
In terms of mechanisms, major pathways emerged. In adrenocortical carcinomas, altered pathways may orient towards developing specific therapeutic strategies. Especially, ZNRF3 alterations have extended the importance of the Wnt/beta- catenin dysregulation in these tumors. Therefore targeting this pathway is a promising direction. In addition, the different molecular alterations in adrenal carcinoma should now be distin- guished, based on whether the alteration is common to almost all carcinomas - such as IGF2 overexpression -, or whether the alteration is specific to a molecular subgroup. For instance
G. Assié / Annales d’Endocrinologie xxx (2017) xxx-xxx
in the subgroup of poor outcome, cell-cycle alterations and DNA hypermethylation in gene promoter regions are prominent. Therefore agents targeting these molecular phenotypes should be tested specifically in this subtype.
In terms of prognosis, in adrenocortical carcinomas, the differences of outcome between molecular groups are major. Moreover, this prognostic value seems independent from tumor stage and tumor grade [10], the two strongest clinical prog- nostic markers. Therefore, a new generation of markers should appear soon in clinical routine. Stratification of adrenocortical carcinoma patients after complete surgery should have impor- tant consequences on orienting adjuvant treatment. Potentially, patients predicted of good outcome may be spared from unneces- sary adjuvant treatment, whereas those from the poorest outcome may be proposed an adjuvant therapy stronger than mitotane alone. Clinical trials will be needed to confirm the benefit of such strategies.
For benign tumors generating Cushing syndrome, genomic approaches help to better and deeper characterize and classify these diseases. For instance in the field of bilateral adrenal diseases, it now comes clear that primary pigmented nodular dysplasia and primary macronodular hyperplasias with distinct multiple nodules are specific entities associated with PRKAR1A and ARMC5 mutations. In-between these specific types of hyperplasias, the proper classification of diseases remain to be established, and a better molecular classification may help.
In terms of clinical benefits for patients, applications of genomic knowledge in benign tumors are probably less straight- forward than in cancer. Indeed, surgery will remain the first-line and most-efficient solution in a majority of cases of benign adrenal and pituitary situations. However there is certainly room for improvement when surgery has failed - for instance in pituitary diseases -, or would responsible for important side effects - such as definitive adrenal insufficiency after bilateral adrenalectomy for bilateral adrenal diseases. In such cases, a bet- ter understanding of molecular mechanisms would potentially orient towards new medical alternatives.
2. Genomic profiling of patients exposed to glucocorticoid excess
2.1. Rationale
Cushing syndrome is either related to endogenous abnor- mal secretion, or to exogenous glucocorticoid administration. Though endogenous overt Cushing syndromes are rare, endoge- nous mild or subclinical Cushing are common in general population. In addition, many humans have chronic alterations of their cortisol secretion pattern, related to chronic stress, depres- sion, or alcohol consumption for instance. Chronic alterations of sleep also impact cortisol secretion pattern, as observed in night-shift workers for instance. It is now established that these alterations, though being mild, have a health impact, through metabolic and cardiovascular alterations [11].
Glucocorticoid administration is common in general pop- ulation, at various doses [12]. Most often glucocorticoids are administered locally, but for a significant proportion of
patients, a systemic passage occurs, as ascertained by the altered endogenous cortisol production. Resulting from these glucocorticoid exposures, patients develop a Cushing syndrome to various extents.
Severity of impact is correlated both to the level and duration of glucocorticoid exposure. However, beyond these two obvious determinants, inter-individual susceptibility is major, for each type of complication associated to glucocorticoid excess. And today, predicting this susceptibility is almost impossible for a given individual.
Therefore the question can be raised, whether large-scale molecular measurements could identify markers of individ- ual susceptibility to glucocorticoids. In addition, owing to the pleiotropic effects of glucocorticoids, genomic studies of tis- sues exposed may provide relevant hints about the mechanisms implied into tissue-specific glucocorticoid effects.
2.2. Lessons from stress science
When looking for molecular markers, blood is a good tar- get tissue, since it is easily accessible and potentially rich in terms of information. Genomic markers of chronic stress have been searched in patients’ blood samples. A first study iden- tified a specific transcriptome signature [13]. Another recent study focused on whole-blood methylome, and identified spe- cific methylation marks associated with chronic stress [14]. These studies illustrate that molecular markers can potentially be used as surrogates in situations where no clear assay exist. This type of approach could potentially be extended to subclini- cal Cushing syndrome and chronic cortisol secretion alteration patterns. Such markers could help to objectify subtle changes in glucocorticoid secretion pattern with potential long-term car- diovascular and metabolic impact.
2.3. The model of glucocorticoid excess
Whole-blood methylome has been performed in patients undergoing glucocorticoid treatments [15], identifying methyla- tion marks specific of glucocorticoid exposure. In a pilot study, we compared the whole-blood methylome of patients before and after treatment of a Cushing syndrome, and could also identify a common glucocorticoid excess signature (unpublished).
Such markers could help characterizing glucocorticoid excess in addition to current hormone assays. In addition, some of these markers may potentially be specifically associated with specific complications of glucocorticoid complications, and therefore could be specific markers of individual susceptibility to glucocorticoid complications.
3. Conclusions and future directions
Several tumor types responsible for Cushing syndrome have been extensively characterized using genomic approaches. These studies provided major advances in understanding the mechanisms responsible for tumorigenesis. In addition, in adrenal cancer, a new generation of strong prognostic mark- ers is emerging that may impact patients care. However
G. Assié / Annales d’Endocrinologie xxx (2017) xxx-xxx
several un-answered questions remain. First, some tumor types responsible for Cushing syndrome remain poorly explored by genomic approaches, including pituitary corticotroph adeno- mas, and endocrine tumors responsible for paraneoplasic ACTH secretion. In addition, new genomic techniques, focusing of epigenetic changes in tumors, are now emerging as major tools for digging deeper into the mechanisms responsible for tumorigenesis. Outstanding observations are reported in several tumor types. This type of approaches should now be applied to endocrine tumors as well. Considering the rarity of these tumors, collaborative research networks - such as the French COMETE network or the European ENSAT network for the study of adrenal tumors -, should pursue their efforts to include patients and coordinate research efforts.
More efforts should be put into finding markers of mild glucocorticoid excess, and of individual susceptibility to gluco- corticoid complications. Preliminary studies suggest that such markers exist. Overt endogenous Cushing is probably the best model for establishing such markers. The impact may poten- tially be important in terms of public health, owing to the high prevalence of mild endogenous and exogenous glucocorticoid excess, and to the high prevalence of multifactorial diseases such as hypertension, diabetes or osteoporosis - potentially related to unidentified glucocorticoid excess.
Disclosure of interest
The author declares that he has no competing interest.
References
[1] Assié G, et al. Integrated genomic characterization of adrenocortical car- cinoma. Nat Genet 2014;46:607-12.
[2] Zheng S, et al. Comprehensive pan-genomic characterization of adreno- cortical carcinoma. Cancer Cell 2016;29:723-36.
[3] Sato Y, et al. Recurrent somatic mutations underlie corticotropin- independent Cushing’s syndrome. Science 2014;344:917-20.
[4] Goh G, et al. Recurrent activating mutation in PRKACA in cortisol- producing adrenal tumors. Nat Genet 2014;46:613-7.
[5] Cao Y, et al. Activating hotspot L205R mutation in PRKACA and adrenal Cushing’s syndrome. Science 2014;344:913-7.
[6] Beuschlein F, et al. Constitutive activation of PKA catalytic subunit in adrenal Cushing’s syndrome. N Engl J Med 2014;370:1019-28.
[7] Assié G, et al. ARMC5 mutations in macronodular adrenal hyperplasia with Cushing’s syndrome. N Engl J Med 2013;369:2105-14.
[8] Ma Z-Y, et al. Recurrent gain-of-function USP8 mutations in Cushing’s disease. Cell Res 2015;25:306-17.
[9] Reincke M, et al. Mutations in the deubiquitinase gene USP8 cause Cush- ing’s disease. Nat Genet 2015;47:31-8, http://dx.doi.org/10.1038/ng.3166.
[10] Jouinot A, et al. DNA Methylation Is an Independent Prognostic Marker of Survival in Adrenocortical Cancer. J Clin Endocrinol Metab 2017;102:923-32.
[11] Di Dalmazi G, Pasquali R, Beuschlein F, Reincke M. Subclinical hypercortisolism: a state, a syndrome, or a disease? Eur J Endocrinol 2015;173:M61-71.
[12] Overman RA, Yeh J-Y, Deal CL. Prevalence of oral glucocorticoid usage in the United States: a general population perspective. Arthritis Care Res 2013;65:294-8.
[13] Nater UM, et al. Impact of acute psychosocial stress on peripheral blood gene expression pathways in healthy men. Biol Psychol 2009;82: 125-32.
[14] Rutten BPF, et al. Longitudinal analyses of the DNA methylome in deployed military servicemen identify susceptibility loci for post- traumatic stress disorder. Mol Psychiatry 00 2017, 1-12;00:1-12, http://dx.doi.org/10.1038/mp.2017.120.
[15] Wan ES, et al. Systemic steroid exposure is associated with differential methylation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012;186:1248-55.