The Genomics of Adrenocortical Tumors
Simon Faillot1,2, Guillaume Assie1,3
1Institut Cochin, INSERM U1016, CNRS 8104, Paris Descartes University, Paris, France
2SIRIC (Site de Recherche Intégré sur le Cancer) CARPEM (CAncer Research for PErsonalized Medicine), Assistance Publique Hôpitaux de Paris, Paris, France
3Department of Endocrinology, Reference Center for Rare Adrenal Diseases, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris (France)
Address for correspondence: Pr Guillaume Assié, Dpt of Endocrinology, 27 rue du Faubourg Saint Jacques 75014 Paris (Tel: +33 1 58 41 18 20 Fax: +33 1 58 41 18 05, e-mail: guillaume.assie@aphp.fr)
Short title: Genomics of Adrenocortical tumors
Key words: genomic, adrenocortical tumors, next generation sequencing, precision medicine
Word count: 5029 words
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Abstract
The last decade witnessed the emergence of genomics, a set of high-throughput molecular measurements in biological samples. These pan-genomic and agnostic approaches have revolutionized the molecular biology and genetics of malignant and benign tumors. These techniques have been applied successfully to adrenocortical tumors.
29 30 31
Exome sequencing identified new major drivers in all tumor types, including KCNJ5, ATP1A1, ATP2B3 and CACNA1D mutations in aldosterone producing adenomas (APA), PRKACA mutations in cortisol producing adenomas (CPA), ARMC5 mutations in primary bilateral macronodular adrenocortical hyperplasia (PBMAH), and ZNRF3 mutations in adrenocortical carcinomas (ACC).
Moreover, the various genomic approaches -including exome sequencing, transcriptome, miRNome, genome, and methylome-, converge into a single molecular classification of adrenocortical tumors. Especially for ACC, two main molecular groups have emerged, showing major differences in outcomes. These ACC groups differ by their gene expression profiles, but also by recurrent mutations and specific DNA hypermethylation patterns in the subgroup of poor outcome.
42 43
44
45
The clinical impact of these findings is just starting. The main altered signaling pathways now become therapeutic targets. The molecular groups of diseases individualize robust subtypes within diseases such as APA, CPA, PBMAH and ACC. A revised nosology of adrenocortical tumors should impact the clinical research. Obvious consequences also include genetic counseling for the new genetic diseases such as ARMC5 mutations in PBMAH, and a better prognostication of ACC based on targeted measurements of a few discriminant molecular alterations.
46
47 Identifying the main molecular groups of adrenocortical tumors by extensively gathering the molecular variations is a significant step forward towards precision medicine.
48 49
24 25 26 27 28
32 33 34 35 36 37 38 39 40 41
52
Adrenocortical tumors gather a large panel of diseases, mainly differing by their pathology -e.g. benign or malignant, focal tumor or diffuse hyperplasia-, and by their secretion types and levels.
53 54 55
Adrenocortical adenomas (ACA) are common, and differ mainly by their secretion profiles. Indeed, the aldosterone producing adenomas (APA) are responsible for primary aldosteronism, the cortisol 56 producing adenomas (CPA) are responsible for Cushing syndrome, while a majority of ACAs are non- secreting adenomas (NSA). In-between are the subclinical Cushing adenomas (SCA) responsible for 57 58 59 60
subtle hormonal alterations but no overt clinical Cushing syndrome1. The adrenocortical hyperplasia includes several pathological forms. The primary bilateral macronodular adrenocortical hyperplasia (PBMAH) is probably the most common, potentially responsible for Cushing syndrome. The primary
61 pigmented nodular adrenal dysplasia (PPNAD) is much rarer, but quite well individualized in terms of pathology, and responsible for overt Cushing. Other forms exist, with more intermediate features2. 62
63 Finally adrenocortical carcinomas (ACC) are rare but overall aggressive tumors. However, the behavior of ACC differs between patients3. Indeed a few carcinomas are of limited aggressiveness and can be cured by surgery, while others won’t be cured by surgery, despite complete resection, even when localized to the adrenal gland.
64 65 66 67 A general hypothesis is that the variable presentations of adrenocortical tumors are related to tumor 68 69 variability, at the molecular level. To test this hypothesis, new generations of tools have been recently used, able to catch extensively molecular variability. These tools, also called “genomics”, 70 were mostly developed during the last decade. Genomics is a set of ever improving high throughput
71 measurement techniques, measuring variability at different levels: at the gene expression level
72 (transcriptome), at the miRNA expression level (miRNome), at the DNA methylation level 73 (methylome), at the chromosomal structure level (copy-number alterations, loss of heterozygosity
74 caught by SNP and CGH arrays), and at the DNA sequence level (mutations). All these genomic
75 approaches have been recently applied to adrenocortical tumors, providing an agnostic molecular
76 classification, leading to better classification and better understanding of pathophysiology.
77 The aim of this review is to show how genomics changed our knowledge of adrenocortical tumors,
78 and to present the clinical perspectives of this new knowledge.
79 Methods
80
The published literature on the genomics of adrenocortical tumor was screened using the following search terms on PubMed between 2009 and October 2015: “(adrenal OR adrenocortical) AND
81
82 (hyperplasia OR adenoma OR dysplasia OR cancer OR carcinoma)“. Review articles were also
83 considered, and their reference lists were screened for additional relevant publications.
84 Exome sequencing of adrenocortical tumors: it’s raining genes !
85
86 87 88 89
Before the genomic era, after two decades of candidate gene approaches, half a dozen of driver genes were identified for different adrenocortical tumor types. These drivers include mainly TP53 inactivating mutations in adrenocortical carcinomas (ACC), CTNNB1 activating mutations in ACA and ACC, and mutations activating the cyclic-AMP / Protein kinase A pathway in ACA and adrenocortical dysplasia/hyperplasia -including PRKAR1A inactivating mutations, PDE8B and PDE11A inactivating mutations, GNAS activating mutations4 - (Table 1, Figure 1.B). Since the recent rise of next generation
90 91 sequencing, and in particular exome sequencing, it took only four years to identify a new dozen of
92 driver genes in these tumors (Table 1).
93 94 95 96 hyperplasia.
Aldosterone Producing Adenomas (APA): mutations activating the calcium signaling pathway
Primary aldosteronism is either related to an APA, or to a bilateral disease referred to as adrenal
97 The first exome sequencing reported for an adrenal tumor was in APA, identifying recurrent 98 mutations of KCNJ5 in 20115. Mutations in this gene, encoding a potassium channel GIRK4, have been subsequently confirmed in larger series, occurring in about 26 to 40% of APA, with higher prevalence in Japanese population6. Subsequently, other mutations were found in APA, implying active ion transporters regulating the plasma membrane potential (ATP2B37, ATP1A17,8), and the voltage dependant calcium channels CACNA1D8,9 and CACNA1H10.
All these mutations share a common consequence of increasing intra-cellular calcium. Intra-cellular calcium is the main second messenger regulating aldosterone production in physiology. Indeed, both the Angiotensin and the raise of extracellular potassium induce calcium signaling pathways in glomerulosa cells, through the activation of type 2 Angiotensin Receptor -with subsequent Gq / Phospholipase C signaling-, and the plasma membrane depolarization -with subsequent opening of calcium voltage-dependant channels- respectively6. In APA, the new mutated genes also demonstrate the central role of intra-cellular calcium for autonomous aldosterone production. This intra-cellular calcium raise is either related to mutations in ion transporter genes (KCNJ5, ATP2B3 and ATP1A1) inducing a plasma membrane depolarization with subsequent opening of voltage- dependant calcium channels5,7,8, or to mutations in voltage-dependant calcium channels (CACNA1D and CACNA1H) inducing the opening at lower thresholds of membrane depolarization8,9 (Figure 1.D).
The majority of these mutations were identified in the somatic compartment, by comparing the APA (tumor tissue DNA) exome to the germline (leucocyte DNA) exome. In the germline compartment, rare instances of mutations were also identified for KCNJ5, CACNA1D and CACNA1H, mainly in families with primary aldosteronism, and as de novo mutations5,9,10.
The molecular alterations in the zona glomarulosa may be more complex. Indeed recently Fernandes-Rosa et al. identified different somatic mutations in the adrenals of patients presenting an APA with multiple nodules11. In addition, Nishimoto et al12 reported somatic mutations in CACNA1D, ATP1A1 and ATP2B3 in “normal adrenals” -e.g. from patients not included as presenting a primary
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121
aldosteronism -. These mutations were identified in clusters of glomerulosa cells with elevated expression of CYP11B2, and different mutations could be found in different such clusters within the same adrenal12.
Beyond the calcium signaling, the Wnt/beta-catenin seems to be another major player in APA (Figure 1). Berthon and colleagues recently reported Wnt/beta-catenin activation in a majority of APA, related to the down-regulation of a negative regulator of this pathway, named SFRP213-this downregulation seems to be due to a high methylation of SFRP2 promoter13 -. Another recent publication identified CTNNB1 mutations -encoding the beta-catenin- in Conn adenomas in pregnant women14. The relation between the calcium signaling activation and the Wnt/beta-catenin pathway remains to be elucidated.
Cortisol producing adenomas (CPA): mutations of the PRKACA
In 2014, four independent research teams identified activating mutations of PRKACA -the alpha subunit of Protein Kinase A (PKA)- in CPA15-18. As for APA, CPA exome was compared to the germline exome, leading to the identification of a hotspot mutation transforming a Lysine into an Arginine at position 205 (L205R). These studies and subsequent ones identified this mutation in 22 to 50% of CPA associated with overt Cushing15-18.
In physiology, activation of the cAMP/PKA pathway is central in zona fasciculata cells for the cortisol production. This pathway is activated by the binding of ACTH to its receptor MC2R, subsequently activating Gs, then adenylyl cyclase, and finally PKA. In CPA with the L205R mutation, the autonomous cortisol production results from direct activation of PKA19. Indeed, the PRKACA L205R mutation was shown to increase PKA activity15,17. These PRKACA mutations add to previously reported GNAS20 and PRKAR1A21 mutations, identified in rare instances of CPA, and also activating the cAMP/PKA pathway (Figure 1).
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144
In addition to these mutations activating the cAMP/PKA pathway, mutations of CTNNB1 -activating the Wnt/beta-catenin pathway-, have also been reported in adrenocortical adenomas, found in about 25% of cases22. Adenomas with CTNNB1 mutations are bigger and secrete less cortisol than their wild-type counterparts23. Of note, the CPA with PRKACA mutations reported recently do not often harbor CTNNB1 mutation15-18. Are these mutations somehow mutually exclusive? Is there any interplay between the Wnt/beta-catenin and the AMPc/PKA pathway? These points remain to be clarified.
ARMC5, a new gene predisposing to Primary Bilateral Macronodular Adrenal Hyperplasia (PBMAH)
Though being bilateral and a global adrenal cortex disease, adrenal hyperplasia has not clearly been deemed as genetic diseases until recently, at least in its most common macronodular subtype. Indeed only rare families have been reported, and mutations in several genes (PRKAR1A, PDE11A, PDE8B, GNAS, PRKACB, MEN1 or FH among others; review in 4,24,25) have been identified only in rare instances and or specific subtypes of these diseases -such as PRKAR1A mutations in PPNAD.
In 2013, we identified ARMC5 mutations in 50% of PBMAH patients requiring an operation26. This gene was identified by comparing the macronodules to germline DNA. A SNP array approach first identified a somatic copy-neutral loss of heterozygosity in chromosome 16p in 25% of patients. An additional whole-genome sequencing identified mutations of ARMC5, located in 16p. ARMC5 is a typical tumor suppressor gene, with two hits identified in each macronodule: a first germline hit - found in all the tissues and all the adrenal macronodules of each patient-, and a second somatic hit - specific of each adrenal macronodule26.
Subsequent studies identified a prevalence of 25 to 50% of ARMC5 mutations in PBMAH27-29. Patients with ARMC5 mutations seem to present a more severe Cushing syndrome, and a more nodular adrenal shape than their wild type counterparts2. Alencar et al also reported a large family with ARMC5 mutations, giving important insight on the penetrance of the disease27.
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
ARMC5 is a “new gene”, poorly characterized in terms of function. ARMC5 inactivating mutations were shown to decrease steroidogenesis and to reduce apoptosis in adrenocortical cells24. Its precise role remains to be established. Does ARMC5 play a role in the recently reported intra-adrenal ACTH production30, or with the aberrant GPCR expression profiles31? Further studies are needed.
The mutation landscape of Adrenocortical carcinomas (ACC): a reduced number of genes with recurrent mutations
In 2014, we reported the first ACC genomic landscape based on exome sequencing and SNP arrays - the latter technique is convenient for detecting homozygous deletions and high-level amplifications-, comparing the tumor to germline DNA32. The Cancer Genome Atlas (TCGA) consortium has also recently presented the mutational landscape on a large cohort of ACC, also comparing the tumor and the germline exome33. Juhlin et al34 also reported a few months ago another analysis of the ACC exome, using a similar strategy. These studies converge on a reduced list of recurrent ACC drivers (Table 1). Two other studies used next generation sequencing to characterize the mutations in ACC, analyzing the genetic variation of a limited number of cancer-related genes, and restricting the analysis to the tumor DNA35,36
Among these drivers, ZNRF3 appeared as the most commonly altered gene in ACC, found in more than 20% of ACCs32-34. This is also the first report of ZNRF3 alterations in Cancer. ZNRF3 is an E3 ubiquitin ligase, with a negative regulator function of the Wnt/beta-Catenin pathway37. In ACC, ZNRF3 is inactivated, mainly by homozygous deletions, but also by mutations32-34. This suggested that ZNRF3 inactivation induces an activation of the Wnt/beta-catenin pathway (Figure 1). The Wnt/beta- catenin pathway can also be activated directly in ACC by CTNNB1 activating mutations22 -encoding the beta-Catenin-, also commonly found in ACC - ~ 15% of ACC32-34. Interestingly and logically, ZNRF3 and CTNNB1 mutations are mutually exclusive in ACC32-34.
Other recurrently mutated genes are related to the cell-cycle regulation. Among these, TP53 mutations are found in ~15%32-34. Other genes regulating the cell-cycle and commonly altered in
170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
cancer are altered in ACC, including the tumor suppressors CDKN2A and RB1, and the oncogenes MDM2 and CDK4 32-34,36,38,39 (Table 1). These alterations globally unbrake the cell cycle, one of the key elements of cancer proliferation (Figure 1A).
Other pathways emerging from the genetic landscape of ACC include the chromatin remodeling - with mutations of MEN1, DAXX, ATRX-, and the chromosome maintenance - with TERT and TERF2 amplifications 32-34,36 .
A few other recurrently mutated genes have been identified, including PRKAR1A and RPL2233. PRKAR1A mutations had been previously identified in PPNADs and in some CPA (see previous paragraph). These instances of ACC with PRKAR1A mutations raise the question of the potential role of the cAMP/PKA pathway in some ACC, and subsequently of the potential predisposition to ACC in the benign adrenal tumors with activated cAMP/PKA. Further studies are needed to clarify this point. Finally other genes are altered, with no clear recurrence among ACC (“private” mutations). The global importance of these genes in ACC pathophysiology remains to be clarified.
204 205 206 207 208 209 commonly available and used in other cancers.
Of note, none of these alterations oriented obviously towards the use of any targeted cancer therapy
210 2) Integrated genomics of ACC identify distinct molecular groups with
211 different outcomes.
212 Several teams focused on the genomics of ACC (Table 2), showing major differences between ACC 213 and ACA, and identifying molecular subgroups of ACC, with different outcomes.
214 Common molecular features of ACC
ACCs share common molecular characteristics (Table 2). At the transcriptome level, thousands of
215 216 genes are deregulated, in comparison to ACA. A large proportion of these genes is implicated in the
195 196 197 198 199 200 201 202 203
cell-cycle regulation, chromosomal maintenance, survival, inflammation, immunity, and the global modulation of gene expression, as part of a global “malignancy signature”40. This type of gene expression signature is shared by many different cancer types, independently of the tissue of origin. In addition to these unspecific malignancy features, the ACC specifically show a global decrease in expression of genes related to steroidogenesis, reflecting the dedifferentiation of ACC cells41-49. Finally massive IGF2 over-expression is one of the most striking difference between ACC and ACA, observed in >85% of ACC41. This was known before genomics, first observed in congenital Beckwith- Wiedemann syndrome30. IGF2 over-expression is considered as an early tumorigenesis event, activating the IGF1 receptor51.
MiRNA are small untranslated transcripts, regulating the gene expression level by RNA-interferrence. Several hundreds of miRNA have been identified in humans. The miRnome of ACC show globally several miRNA differentially expressed compared to ACA. The most consistent one is the overexpression miR-483-5p52-54, which gene is located in IGF2 locus, and probably deregulated simultaneously with IGF2. Other deregulated miRNA include downregulation of miR335 and miR19547,52-56 among others. There is currently no precise determination of the global impact of these miRNA deregulations on gene expression in ACC.
In terms of DNA methylation, ACC are globally hypomethylated compared with ACA57. This hypomethylation is mainly related to hypomethylation in intergenic regions of DNA, probably contributing to the genome instability, as seen in cancer in general58. Beyond this global hypomethylation, a subset of ACC shows specific patterns of methylation, with hypermethylation in CpG islands of gene promoter regions59. This point will be detailed in the next paragraph.
Finally the ACC genome tends to be altered compared to ACA, with extended chromosomal gains, losses and losses of heterozygosity in a majority of ACC60-63.
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
Genomics identify different molecular groups of ACC
The major finding of the integrated genomics of ACC is the existence of distinct molecular subgroups.
Remarkably, these distinct groups are strongly associated with the patient’s outcome.
At the transcriptome level, two main distinct subgroups have been defined, strongly associated to prognosis 33,42,64
. At the mIR level, 2 to 3 subgroups are reported, also associated with different outcomes 32,33,54,55 .
At the DNA methylation level, an important variability occurs when focusing on the methylation of gene promoter regions, in regions enriched in “CG” bases (called CpG islands)57,59. Indeed some ACC are hypermethylated in these regions, evoking a CpG Island Methylator Phenotype (CIMP), as described in other malignancies such as colon cancer65. In ACC, CIMP is associated with a poor outcome59.
As stated in the previous paragraph, recurrent somatic mutations have been found in ~50% of ACC32-
34. These ACC are associated with a poorer outcome.
254 255 256 257 258 259 260 261
These omics classifications have been extensively generated in two independent international cohorts of ACC: one from Europe (European Network for the Study of Adrenal tumors)32,33 and one from the TCGA ACC Group from America, Europe and Australia32,33. A global molecular classification has emerged from these studies, highly consensual32,33. Indeed, a strong agreement was found between the different omics classifications, defining two main molecular groups. Interestingly these two groups are associated with a quite different outcome. Finally a more granular classification can isolate a third molecular subgroup, associated with intermediate survival (Figure 2). The TCGA cohort should be published soon, warranting a more detailed comparison with the ENSAT cohort.
The molecular group of ACC with better outcome
262 263
In terms of transcriptome, the molecular group of ACC with better outcome shows a higher expression of genes intervening in cell metabolism, intracellular transport, cell differentiation and in
240 241 242 243 244 245 246 247 248 249 250 251 252 253
264 265
apoptosis42. These tumors have a low mutation rate, and scarcely harbor mutations in the recurrently mutated genes described in the previous paragraph32,33. Instead private mutations can be observed. 266 Finally in terms of DNA methylation, these tumors do not show any CIMP32,33 (Figure 2).
Though being of better outcome, with few recurrences after surgery, these tumors do correspond to malignant tumors in terms of molecular profiling, and not to benign tumors. Indeed the transcriptome and miRnome profiles of these tumors are much closer to the transcriptome of other ACCs -the aggressive ones-, compared to ACAs. In addition the chromosomes of these tumors are extensively altered in a majority of tumors, a feature not observed in ACAs60,62,63. Finally, in our study, a majority of these tumors have an elevated number of malignancy features as defined by pathology -scored by the Weiss score-, whereas almost all the tumors defined as benign by the transcriptome show evident pathological features of ACA -with Weiss scores of 0 or 142. Interestingly in this study seven tumors show borderline pathology -corresponding to Weiss scores of 2 or 3 -. Six out of seven are classified as malignant by the transcriptome, including four in the molecular group of better outcome42. This suggests that adrenal tumors with borderline pathology are probably malignant in terms of molecular profile. Larger cohorts are needed to confirm this finding.
The molecular group of Aggressive ACC
In terms of transcriptome, the molecular group of aggressive ACC shows a higher expression of Cell- cycle related genes42,64. In terms of mutations, these tumors accumulate alterations on the recurrent ACC drivers reported in the previous paragraph32-34 (Table 1).
286 287 288
In terms of pathology, Giordano et al suggested that this group of ACC is of “higher grade”, as defined by increased proliferation index - e.g. high Ki67 immunopositivity and high mitotic count-64. The transcriptome signature is indeed in agreement with the notion of increased proliferation. However, despite a significant difference in terms of proliferation index between the two main molecular types of ACC, this criterion is not sufficient on its own to discriminate. Indeed in this study, some low grade ACC are observed in the transcriptome group of poor outcome, and some high-grade
267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
ACC are observed in the transcriptome group of better outcome64. This relation between proliferation index and molecular classification deserves further studies on larger cohorts. If these discrepancies are confirmed, it will be of major interest to specifically determine the prognosis of ACC with discrepant proliferation index and molecular group. Finally Ki67 reproducibility and mitotic count are not absolute66. Of note, the reproducibility of the discriminant molecular measures within a single ACC has not been tested yet.
Within the group of aggressive ACC, a subset of tumors shows hypermethylation on CpG Islands in genes promoters (CIMP). This hypermethylation is associated with a poorer prognosis compared with tumors with no hypermethylation32,33. Therefore methylation probably further divides the group of aggressive ACC into one subgroup of poorer outcome, and another subgroup of intermediate outcome. Further studies are needed to confirm this observation.
None of the driver genes is known to impact directly DNA methylation, and the specific methylation pattern associated with a poor outcome -e.g. CIMP- is not explained yet in terms of underlying mechanisms. The link between driver mutations and transcriptome can probably be partially explained: mutations in drivers such as TP53 or CDKN2A probably contribute to the cell-cycle signature identified in the subgroup of poor outcome, as observed in other cancer types, though the precise transcriptome signature induced by each driver mutation in adrenal cells is not precisely established. Finally, concerning the link between methylation and transcriptome, the CIMP is potentially responsible for modulating the expression level of 15% of genes59.
309
3. Clinical implications of the genomic knowledge
310
The genomics findings have clinical implications for adrenocortical tumors.
311 312 313
Towards an improved Nosology of adrenocortical tumors?
Before genomics, gold standards for classifying the tumors were based on a combination of pathology, hormonology and clinical outcome. Genomics provide agnostic molecular classifications.
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
Can these molecular classifications help to better define the different diseases and their subtypes?
314 315 Possibly.
316 For primary aldosteronism, can we properly discriminate a true APA, from an adrenal incidentaloma associated with a diffuse or multifocal zona glomerulosa disease? The adrenal vein sampling performed before surgery warrants somehow this problem by demonstrating the asymmetry of aldosterone secretion between the two adrenal glands. However the thresholds are often discussed and asymmetric forms of hyperplasia have been reported67. “Functional pathology” has emerged as a potential way to confirm the diagnosis of APA. The strategy is to ascertain the expression of aldosterone synthase (CYP11B2) -the key aldosterone producing enzyme- in the adenoma. This was first achieved by in situ hybridization and more recently by immunohistochemistry (see6 for review). Genomics has brought up the knowledge of somatic mutations in APA (affecting KCNJ5, ATP2B3, ATP1A1, CACNA1D), and finding such a mutation may also be a way to ascertain the true nature of an APA. The potential benefit of this genotyping in clinical practice remains to be established. For instance will it be possible to use molecular markers for preoperative diagnosis, such as detecting the somatic mutations in the patient’s plasma from the cell-free tumor DNA, or detecting circulating miRNA signing the presence of an APA?
Bilateral adrenal hyperplasia include a wide spectrum of pathological aspects - including mainly macronodular and micronodular hyperplasia, hyperplasia with no nodule, and PPNAD-, and are associated with different levels of cortisol secretion2. Are these diseases different, or are they variations of a single disease? Genomics finding help to clarify this question, especially with the discovery of driver genes defining specific diseases. The most obvious example is the PRKAR1A mutation in PPNAD. In PPNAD there is globally a good agreement between the genotype and the pathological phenotype -micronodules, adrenal atrophy between the nodules, and pigmented spots -. However some PPNAD do show macronodules68 (often harboring a somatic CTNNB1 mutation), and reversely, no PRKAR1A mutation is found in ~30% of PPNAD. Are these PPNAD a different disease?
317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338
With genomics another gene, ARMC5, was recently discovered. ARMC5 mutations seem to be related to a phenotype with several macronodules, and in the index cases more often an overt Cushing syndrome29. Meningiomas may also be associated to the disease69. However more than 50% of these hyperplasias are not related to ARMC5. Are these different diseases? How many diseases? Associated with which phenotype? With which secretion level? Identifying the molecular drivers of these diseases could help to better classify these diseases.
Genetic counseling for the new genetic diseases
Some adrenal tumors are hereditary genetic diseases. From the genomic studies, drivers have been identified, opening new possibilities for genetic counseling. In PBMAH patients with ARMC5 mutations, a genetic screening can be proposed to the relatives. A first estimation of the penetrance arises from the large Brazilian family reported by Alencar et al27. This genetic screening is important because the long term consequences of hypercortisolism, progressively starting subclinical, should be prevented in ARMC5 mutations carriers, through early and aggressive correction of mild complications -hypertension, diabetes, osteoporosis-, and early medical treatment or adrenalectomy if required.
In primary aldosteronism, familial forms with KNCJ5 mutations have been reported6. Here again,
screening the relatives should help to identify the patients at risk, and diagnose and treat early their hypertension. However these familial forms seem very rare compared to the vast majority of sporadic diseases related to somatic mutations.
A better diagnosis and a better determination of prognosis for ACC patients
360 361 362
Genomic studies clearly discriminate benign from malignant carcinomas. There is probably no need for sophisticated molecular tool when the pathological diagnosis is obvious, either for benign tumors with Weiss scores of 0 or 1, or for malignant tumors with Weiss scores of 4 or more. However in
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
several situations, molecular diagnosis may be relevant. First, tumors with borderline pathology - especially those with Weiss scores of 2 or 3- may be properly classified with molecular tools. As stated previously, in our study a majority of these tumors are most often malignant in terms of molecular profile42. Confirmation is necessary on larger series. Another question is about the inter- and intra-observer variability of the pathological criteria from the Weiss scoring system66. Can molecular diagnosis help to limit the impact of this variability? To which extent is a molecular assessment reproducible? This remains to be determined. Finally diagnosing malignancy may not be trivial in specific subtypes such as the oncocytic variants of adrenocortical tumors. A specific scoring has been proposed by Biseglia et al70. Would a molecular diagnostic tool be relevant for such tumors? This remains to be determined.
Another question is about the potential malignant evolution of an ACA. Does it exist? Is there an ACA to ACC sequence? Epidemiology does not plead in favour of this sequence, considering the contrast between the high prevalence of adrenal ACA and low prevalence of ACC. However, rare cases of ACC in patients with ACA or hyperplasia have been reported (discussed in 63). Genomics also brings some information. The most remarkable is the high proportion of CTNNB1 mutations in both ACA and ACC. In addition, CTNNB1-mutated ACA are bigger, and secrete less cortisol23: is it a premalignant state? In addition somatic mutations of PRKAR1A have been reported in ACC33, a gene mutated in PPNAD and in rare instances of adrenal adenomas. No extensive exome sequencing of ACA has been performed so far. Such data would help to better define the full spectrum of mutated genes shared between ACC and ACA. Concerning other omics, it is hard to draw any conclusion about any potential ACA to ACC sequence from transcriptome, miRnome and methylome. Finally a few chromosomal alterations common to ACA and ACC have been identified63, but no definitive conclusion can be drawn either.
Among ACC, distinct molecular subgroups have emerged from genomics, associated with different outcomes. This information is of major importance, especially if it is possible to discriminate ACC
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
potentially curable by a complete surgery from ACCs with later recurrences despite a complete (R0) surgery. There is here a unique opportunity to classify the ACC patients in distinct mechanistic and prognostic groups (Figure 2). We proposed a tool to discriminate the two transcriptome groups of ACC, and could recapitulate the whole transcriptome information with the expression of 2 genes, BUB1B and PINK1. This tool was further validated by Fragoso et al71, and is currently prospectively tested in a routine clinical laboratory at our institution. As stated in the previous paragraphs, other omics help to classify the ACC, such as DNA methylation and recurrent drivers sequencing. This information could also be important for optimal oncogenetic classification. However it will be necessary to promote additional studies to confirm the predictive value of each of these molecular markers, precisely determine the interactions between these markers, and assess the benefits compared to the already existing prognostic markers -especially the proliferation index -.
The success of molecular tools in clinical practice will also depend on the simplicity, the robustness and the price of the molecular tools. As stated before, gene expression, but also miRNAs can be conveniently assayed by quantitative-RTPCR. Concerning methylation, the CIMP can be recapitulated by focal measurements of methylation, using methylation-specific Multiplex Ligation Probe Assay (MS-MLPA), a technique widely available in oncogenetic departments59. Finally targeted sequencing of driver genes can be performed by next generation sequencing35,36. In theory all these techniques are easily applicable.
Molecular findings from genomics should also open the way to liquid biopsies in a close future. These include a set of measurements in the blood reflecting the molecular features of the tumor. The most straightforward application is the detection of cell-free tumor DNA in patients’ plasma by catching in plasma a specific tumor mutation, used as a proxy of tumor DNA72. We could consider looking for CTNNB1 mutations, mutated in ~15% of ACC, and corresponding usually to a limited number of mutation hotspots. Such hotspots could be detected prior to surgery. After surgery, sequencing one patient’s tumor could help to identify specific mutations that could be later used as proxys for
388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
monitoring remaining tumor DNA, and therefore remaining tumor tissue. Another application is the detection of circulating miRNAs signatures, also reflecting the diagnosis, and potentially the prognosis of an ACC54.
Towards new treatments of adrenocortical tumors?
In APA, many patients present a somatic mutation increasing the intracellular calcium by depolarizing the plasma membrane implying the opening of voltage-dependent calcium channels. These patients may benefit from calcium channel blockers, beyond the antihypertensive effect of these treatments(see6 for review).
For ACC patients the medical therapies currently available are of limited efficiency, and there is a crucial need for new treatments. Genomics may help to identify new therapeutic targets. However the exome sequencing of ACC did not identify directly drugable molecular alterations, considering the drugs currently commonly available. Especially, no tyrosine kinase mutation could be recurrently found32-34. However the genomic analyses have identified and/or confirmed the implication of a few cardinal signaling pathways. These include mainly the Wnt/beta-catenin signaling and p53 / Rb signaling. These pathways are commonly deregulated in cancer, and important effort is now put to try and develop drugs specifically targeting these pathways in cancer, without altering too much these essential pathways in all other cells. For instance, several inhibitors of the Wnt/beta-catenin pathway have been identified73. In contrast, successes in targeting p53/Rb pathway have been limited so far, despite wide efforts. Hopefully this field of research will benefit to ACC patients in a close future.
Genomics may also help ACC patients through new generations of treatment. One example is the RNA interference therapy, a promising strategy for targeting molecular alterations. The team of Sidhu proposed to target these tumors using miR-774. The genomic classification may also help to propose a new algorithm for treating ACC patients, based on a better risk stratification. Indeed, ACC showing the molecular features associated with a better outcome could probably be proposed a
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
simple follow up as far as a complete surgery could be performed, and in case of recurrence the preferred use of repeated loco regional treatments over systemic treatments. On the opposite, ACC with molecular features associated with a poor outcome could be proposed systemic therapies in priority, even adjuvant systemic therapies after an apparently complete surgery. Such clinical recommendations based on the molecular classification, would need to be validated by prospective clinical trials.
438 439 440 441 442 443 444
445 446 447 Simon Faillot: None
Declaration of interest
Guillaume Assie: None
448 449 450 451
Funding
This work was supported by the Institut National de la Santé et de la Recherche Médicale (G.A.is
recieving a Contrat d’Interface), and the SIRC (Site de Recherche Intégré sur le Cancer) CARPEM (CAncer Research for PErsonalized Medicine) (S.F. is receiving a PhD grant).
452 Acknowledgements
The authors thank Pr Jérôme Bertherat, Dr Ludivine Drougat, Dr Bruno Ragazzon and Mr Mario Neou
for their thoughtful comments and the members of the Genomics and Signaling team of Cochin
Institute for helpful discussions
459
453 454 455 456 457 458
460 References
1. Terzolo, M., Pia, A. & Reimondo, G. Subclinical Cushing’s syndrome: definition and management. Clin. Endocrinol. (Oxf.) 76, 12-18 (2012).
2. Stratakis, C. A. & Boikos, S. A. Genetics of adrenal tumors associated with Cushing’s syndrome: a new classification for bilateral adrenocortical hyperplasias. Nat. Clin. Pract. Endocrinol. Metab. 3, 748-57 (2007).
3. Fassnacht, M., Kroiss, M. & Allolio, B. Update in adrenocortical carcinoma. J. Clin. Endocrinol. Metab. 98, 4551-4564 (2013).
4. Stratakis, C. A. New genes and/or molecular pathways associated with adrenal hyperplasias and related adrenocortical tumors. Mol. Cell. Endocrinol. 300, 152-157 (2009).
5. Choi, M. et al. K+ Channel Mutations in Adrenal Aldosterone-Producing Adenomas and Hereditary Hypertension. Science 331, 768-772 (2011).
6. Zennaro, M .- C., Boulkroun, S. & Fernandes-Rosa, F. An update on novel mechanisms of primary aldosteronism. J. Endocrinol. 224, R63-77 (2015).
7. Beuschlein, F. et al. Somatic mutations in ATP1A1 and ATP2B3 lead to aldosterone-producing adenomas and secondary hypertension. Nat. Genet. 45, 440-444, 444e1-2 (2013).
8. Azizan, E. A. B. et al. Somatic mutations in ATP1A1 and CACNA1D underlie a common subtype of adrenal hypertension. Nat. Genet. 45, 1055-1060 (2013).
9. Scholl, U. I. et al. Somatic and germline CACNA1D calcium channel mutations in aldosterone- producing adenomas and primary aldosteronism. Nat. Genet. 45, 1050-1054 (2013).
10. Scholl, U. I. et al. Recurrent gain of function mutation in calcium channel CACNA1H causes early- onset hypertension with primary aldosteronism. eLife 4, e06315 (2015).
11. Fernandes-Rosa, F. L. et al. Different Somatic Mutations in Multinodular Adrenals With Aldosterone-Producing Adenoma. Hypertension 66, 1014-1022 (2015).
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
12. Nishimoto, K. et al. Aldosterone-stimulating somatic gene mutations are common in normal adrenal glands. Proc. Natl. Acad. Sci. U. S. A. 112, E4591-4599 (2015).
13. Berthon, A. et al. WNT/B-catenin signalling is activated in aldosterone-producing adenomas and controls aldosterone production. Hum. Mol. Genet. 23, 889-905 (2014).
14. Teo, A. E. D. et al. Pregnancy, Primary Aldosteronism, and Adrenal CTNNB1 Mutations. N. Engl. J. Med. 373, 1429-1436 (2015).
15. Beuschlein, F. et al. Constitutive activation of PKA catalytic subunit in adrenal Cushing’s syndrome. N. Engl. J. Med. 370, 1019-1028 (2014).
16. Goh, G. et al. Recurrent activating mutation in PRKACA in cortisol-producing adrenal tumors. Nat. Genet. 46, 613-617 (2014).
17. Sato, Y. et al. Recurrent somatic mutations underlie corticotropin-independent Cushing’s syndrome. Science 344, 917-920 (2014).
18. Cao, Y. et al. Activating hotspot L205R mutation in PRKACA and adrenal Cushing’s syndrome. Science 344, 913-917 (2014).
19. Wilmot Roussel, H. et al. Identification of gene expression profiles associated with cortisol secretion in adrenocortical adenomas. J. Clin. Endocrinol. Metab. (2013). doi:10.1210/jc.2012- 4237
20. Kobayashi, H. et al. Mutation analysis of Gsalpha, adrenocorticotropin receptor and p53 genes in Japanese patients with adrenocortical neoplasms: including a case of Gsalpha mutation. Endocr. J. 47, 461-466 (2000).
21. Bertherat, J. et al. Molecular and functional analysis of PRKAR1A and its locus (17q22-24) in
sporadic adrenocortical tumors: 17q losses, somatic mutations, and protein kinase A expression and activity. Cancer Res. 63, 5308-19 (2003).
22. Tissier, F. et al. Mutations of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors.
485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 Cancer Res. 65, 7622-7 (2005).
23. Bonnet, S. et al. Wnt/B-catenin pathway activation in adrenocortical adenomas is frequently due to somatic CTNNB1-activating mutations, which are associated with larger and nonsecreting tumors: a study in cortisol-secreting and -nonsecreting tumors. J. Clin. Endocrinol. Metab. 96, E419-426 (2011).
24. Drougat, L., Espiard, S. & Bertherat, J. Genetics of primary bilateral macronodular adrenal hyperplasia: a model for early diagnosis of Cushing’s syndrome? Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 173, M121-131 (2015).
25. Correa, R., Salpea, P. & Stratakis, C. A. Carney complex: an update. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 173, M85-97 (2015).
26. Assié, G. et al. ARMC5 mutations in macronodular adrenal hyperplasia with Cushing’s syndrome. N. Engl. J. Med. 369, 2105-2114 (2013).
27. Alencar, G. A. et al. ARMC5 Mutations Are a Frequent Cause of Primary Macronodular Adrenal Hyperplasia. J. Clin. Endocrinol. Metab. 99, E1501-E1509 (2014).
28. Gagliardi, L. et al. ARMC5 mutations are common in familial bilateral macronodular adrenal hyperplasia. J. Clin. Endocrinol. Metab. 99, E1784-1792 (2014).
29. Espiard, S. et al. ARMC5 Mutations in a Large Cohort of Primary Macronodular Adrenal Hyperplasia: Clinical and Functional Consequences. J. Clin. Endocrinol. Metab. 100, E926-935 (2015).
30. Louiset, E. et al. Intraadrenal corticotropin in bilateral macronodular adrenal hyperplasia. N. Engl. J. Med. 369, 2115-2125 (2013).
31. El Ghorayeb, N., Bourdeau, I. & Lacroix, A. Multiple aberrant hormone receptors in Cushing’s syndrome. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 173, M45-60 (2015).
32. Assié, G. et al. Integrated genomic characterization of adrenocortical carcinoma. Nat. Genet. 46, 607-612 (2014).
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
33. Antonio M Lerario et al. in Late-Breaking Adrenal/HPA Axis // LBF-073-LBF-073 (Endocrine Society, 2015). at <http://press.endocrine.org/doi/abs/10.1210/endo-
meetings.2015.AHPAA.12.LBF-073>
34. Juhlin, C. C. et al. Whole-exome sequencing characterizes the landscape of somatic mutations and copy number alterations in adrenocortical carcinoma. J. Clin. Endocrinol. Metab. 100, E493- 502 (2015).
35. Ross, J. S. et al. Next-generation sequencing of adrenocortical carcinoma reveals new routes to targeted therapies. J. Clin. Pathol. 67, 968-973 (2014).
36. De Martino, M. C. et al. Molecular Screening for a Personalized Treatment Approach in Advanced Adrenocortical Cancer. J. Clin. Endocrinol. Metab. 98, 4080-4088 (2013).
37. Hao, H .- X. et al. ZNRF3 promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 485, 195-200 (2012).
38. Zhao, J. et al. Combined comparative genomic hybridization and genomic microarray for detection of gene amplifications in pulmonary artery intimal sarcomas and adrenocortical tumors. Genes. Chromosomes Cancer 34, 48-57 (2002).
39. Ragazzon, B. et al. Mass-array screening of frequent mutations in cancers reveals RB1 alterations in aggressive adrenocortical carcinomas. Eur. J. Endocrinol. 170, 385-391 (2014).
40. Ciriello, G. et al. Emerging landscape of oncogenic signatures across human cancers. Nat. Genet. 45, 1127-1133 (2013).
41. Giordano, T. J. et al. Distinct Transcriptional Profiles of Adrenocortical Tumors Uncovered by DNA Microarray Analysis. Am. J. Pathol. 162, 521-531 (2003).
42. Reyniès, A. de et al. Gene Expression Profiling Reveals a New Classification of Adrenocortical Tumors and Identifies Molecular Predictors of Malignancy and Survival. J. Clin. Oncol. 27, 1108- 1115 (2009).
535 536 537 538 539 540 541 542 543
544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
43. de Fraipont, F. et al. Gene expression profiling of human adrenocortical tumors using complementary deoxyribonucleic Acid microarrays identifies several candidate genes as markers of malignancy. J. Clin. Endocrinol. Metab. 90, 1819-29 (2005).
44. Velázquez-Fernández, D. et al. Expression profiling of adrenocortical neoplasms suggests a molecular signature of malignancy. Surgery 138, 1087-94 (2005).
45. Slater, E. P. et al. Analysis by cDNA microarrays of gene expression patterns of human adrenocortical tumors. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 154, 587-98 (2006).
46. Fernandez-Ranvier, G. G. et al. Identification of biomarkers of adrenocortical carcinoma using genomewide gene expression profiling. Arch. Surg. Chic. III 1960 143, 841-846; discussion 846 (2008).
47. Tömböl, Z. et al. Integrative molecular bioinformatics study of human adrenocortical tumors: microRNA, tissue-specific target prediction, and pathway analysis. Endocr. Relat. Cancer 16, 895- 906 (2009).
48. Soon, P. S. H. et al. Microarray gene expression and immunohistochemistry analyses of adrenocortical tumors identify IGF2 and Ki-67 as useful in differentiating carcinomas from adenomas. Endocr. Relat. Cancer 16, 573-583 (2009).
49. Laurell, C. et al. Transcriptional profiling enables molecular classification of adrenocortical tumours. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 161, 141-152 (2009).
576 577
50. Gicquel, C. et al. Rearrangements at the 11p15 locus and overexpression of insulin-like growth factor-II gene in sporadic adrenocortical tumors. J. Clin. Endocrinol. Metab. 78, 1444-53 (1994).
51. Barlaskar, F. M. et al. Preclinical Targeting of the Type 1 Insulin-like Growth Factor Receptor in Adrenocortical Carcinoma. J. Clin. Endocrinol. Metab. (2008). doi:jc.2008-1456
578 579 580 581 582 583
52. Patterson, E. E., Holloway, A. K., Weng, J., Fojo, T. & Kebebew, E. MicroRNA profiling of adrenocortical tumors reveals miR-483 as a marker of malignancy. Cancer 117, 1630-1639 (2011).
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
53. Soon, P. S. H. et al. miR-195 and miR-483-5p Identified as Predictors of Poor Prognosis in Adrenocortical Cancer. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 15, 7684-7692 (2009).
54. Chabre, O. et al. Serum miR-483-5p and miR-195 are predictive of recurrence risk in adrenocortical cancer patients. Endocr. Relat. Cancer 20, 579-594 (2013).
55. Özata, D. M. et al. The role of microRNA deregulation in the pathogenesis of adrenocortical carcinoma. Endocr. Relat. Cancer 18, 643-655 (2011).
56. Schmitz, K. J. et al. Differential expression of microRNA-675, microRNA-139-3p and microRNA- 335 in benign and malignant adrenocortical tumours. J. Clin. Pathol. 64, 529-535 (2011).
57. Rechache, N. S. et al. DNA methylation profiling identifies global methylation differences and markers of adrenocortical tumors. J. Clin. Endocrinol. Metab. 97, E1004-1013 (2012).
58. Kulis, M. & Esteller, M. DNA methylation and cancer. Adv. Genet. 70, 27-56 (2010).
59. Barreau, O. et al. Identification of a CpG Island Methylator Phenotype in Adrenocortical Carcinomas. J. Clin. Endocrinol. Metab. (2012). doi:10.1210/jc.2012-2993
60. Stephan, E. A. et al. Adrenocortical carcinoma survival rates correlated to genomic copy number variants. Mol. Cancer Ther. 7, 425-431 (2008).
61. Szabó, P. M. et al. Meta-analysis of adrenocortical tumour genomics data: novel pathogenic pathways revealed. Oncogene 29, 3163-3172 (2010).
62. Barreau, O. et al. Clinical and pathophysiological implications of chromosomal alterations in adrenocortical tumors: an integrated genomic approach. J. Clin. Endocrinol. Metab. 97, E301-311 (2012).
63. Ronchi, C. L. et al. Single nucleotide polymorphism array profiling of adrenocortical tumors — evidence for an adenoma carcinoma sequence? Plos One 8, e73959 (2013).
64. Giordano, T. J. et al. Molecular Classification and Prognostication of Adrenocortical Tumors by Transcriptome Profiling. Clin. Cancer Res. 15, 668-676 (2009).
65. Toyota, M. et al. CpG island methylator phenotype in colorectal cancer. Proc. Natl. Acad. Sci. U. S. A. 96, 8681-8686 (1999).
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
66. Tissier, F. et al. Adrenocortical tumors: improving the practice of the Weiss system through virtual microscopy: a National Program of the French Network INCa-COMETE. Am. J. Surg. Pathol. 36, 1194-1201 (2012).
67. Viola, A. et al. Diagnosis and treatment of unilateral forms of primary aldosteronism. Curr. Hypertens. Rev. 9, 156-165 (2013).
68. Gaujoux, S. et al. Wnt/beta-catenin and 3’,5’-cyclic adenosine 5’-monophosphate/protein kinase A signaling pathways alterations and somatic beta-catenin gene mutations in the progression of adrenocortical tumors. J. Clin. Endocrinol. Metab. 93, 4135-4140 (2008).
69. Elbelt, U. et al. Molecular and clinical evidence for an ARMC5 tumor syndrome: concurrent inactivating germline and somatic mutations are associated with both primary macronodular adrenal hyperplasia and meningioma. J. Clin. Endocrinol. Metab. 100, E119-128 (2015).
70. Bisceglia, M. et al. Adrenocortical oncocytic tumors: report of 10 cases and review of the literature. Int. J. Surg. Pathol. 12, 231-243 (2004).
71. Fragoso, M. C. B. V. et al. Combined expression of BUB1B, DLGAP5, and PINK1 as predictors of poor outcome in adrenocortical tumors: validation in a Brazilian cohort of adult and pediatric patients. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 166, 61-67 (2012).
72. Diaz, L. A. & Bardelli, A. Liquid biopsies: genotyping circulating tumor DNA. J. Clin. Oncol. Off. J. Am. Soc. Clin. Oncol. 32, 579-586 (2014).
73. Sebio, A., Kahn, M. & Lenz, H .- J. The potential of targeting Wnt/B-catenin in colon cancer. Expert Opin. Ther. Targets 18, 611-615 (2014).
74. Glover, A. R. et al. microRNA-7 as a tumor suppressor and novel therapeutic for adrenocortical carcinoma. Oncotarget (2015). doi:10.18632/oncotarget.5383
75. Lifton, R. P. et al. A chimaeric 11 beta-hydroxylase/aldosterone synthase gene causes glucocorticoid-remediable aldosteronism and human hypertension. Nature 355, 262-5 (1992).
76. Kirschner, L. S. et al. Mutations of the gene encoding the protein kinase A type I-alpha regulatory subunit in patients with the Carney complex. Nat. Genet. 26, 89-92 (2000).
610 611 612 613 614 615 616 617 618 619 620 621 622 623
624 625 626 627 628 629 630 631 632 633 634 635
77. Horvath, A., Mericq, V. & Stratakis, C. A. Mutation in PDE8B, a cyclic AMP-specific phosphodiesterase in adrenal hyperplasia. N. Engl. J. Med. 358, 750-2 (2008).
78. Horvath, A. et al. A genome-wide scan identifies mutations in the gene encoding phosphodiesterase 11A4 (PDE11A) in individuals with adrenocortical hyperplasia. Nat. Genet. 38, 794-800 (2006).
79. Weinstein, L. S. et al. Activating mutations of the stimulatory G protein in the McCune-Albright syndrome. N. Engl. J. Med. 325, 1688-1695 (1991).
80. Forlino, A. et al. PRKACB and Carney complex. N. Engl. J. Med. 370, 1065-1067 (2014).
81. Skogseid, B. et al. Clinical and genetic features of adrenocortical lesions in multiple endocrine neoplasia type 1. J. Clin. Endocrinol. Metab. 75, 76-81 (1992).
82. Shuch, B. et al. Adrenal nodular hyperplasia in hereditary leiomyomatosis and renal cell cancer. J. Urol. 189, 430-435 (2013).
83. Pilon, C. et al. Inactivation of the p16 tumor suppressor gene in adrenocortical tumors. J. Clin. Endocrinol. Metab. 84, 2776-2779 (1999).
84. Gaujoux, S. et al. Inactivation of the APC gene is constant in adrenocortical tumors from patients with familial adenomatous polyposis but not frequent in sporadic adrenocortical cancers. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 16, 5133-5141 (2010).
85. Lombardi, C. P. et al. Gene expression profiling of adrenal cortical tumors by cDNA macroarray analysis. Results of a preliminary study. Biomed. Pharmacother. Biomed. Pharmacothérapie 60, 186-90 (2006).
86. Fernandez-Ranvier, G. G. et al. Candidate diagnostic markers and tumor suppressor genes for adrenocortical carcinoma by expression profile of genes on chromosome 11q13. World J. Surg. 32, 873-881 (2008).
87. Lenzini, L. et al. Heterogeneity of aldosterone-producing adenomas revealed by a whole transcriptome analysis. Hypertension 50, 1106-13 (2007).
636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660
88. Wang, T. et al. Gene expression profiles in aldosterone-producing adenomas and adjacent adrenal glands. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 164, 613-619 (2011).
89. Azizan, E. A. B. et al. Microarray, qPCR, and KCNJ5 sequencing of aldosterone-producing adenomas reveal differences in genotype and phenotype between zona glomerulosa- and zona fasciculata-like tumors. J. Clin. Endocrinol. Metab. 97, E819-829 (2012).
90. Monticone, S. et al. Effect of KCNJ5 mutations on gene expression in aldosterone-producing adenomas and adrenocortical cells. J. Clin. Endocrinol. Metab. 97, E1567-1572 (2012).
91. Fonseca, A. L. et al. Comprehensive DNA methylation analysis of benign and malignant adrenocortical tumors. Genes. Chromosomes Cancer 51, 949-960 (2012).
92. Howard, B. et al. Integrated analysis of genome-wide methylation and gene expression shows epigenetic regulation of CYP11B2 in aldosteronomas. J. Clin. Endocrinol. Metab. 99, E536-543 (2014).
93. Murakami, M. et al. Integration of transcriptome and methylome analysis of aldosterone- producing adenomas. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 173, 185-195 (2015).
94. Glover, A. R. et al. Long noncoding RNA profiles of adrenocortical cancer can be used to predict recurrence. Endocr. Relat. Cancer 22, 99-109 (2015).
95. Szabó, D. R. et al. Analysis of circulating microRNAs in adrenocortical tumors. Lab. Investig. J. Tech. Methods Pathol. 94, 331-339 (2014).
96. Patel, D. et al. MiR-34a and miR-483-5p are candidate serum biomarkers for adrenocortical tumors. Surgery 154, 1224-1228; discussion 1229 (2013).
97. Robertson, S. et al. MicroRNA-24 is a novel regulator of aldosterone and cortisol production in the human adrenal cortex. Hypertension 62, 572-578 (2013).
98. Velázquez-Fernández, D. et al. MicroRNA expression patterns associated with hyperfunctioning and non-hyperfunctioning phenotypes in adrenocortical adenomas. Eur. J. Endocrinol. Eur. Fed. Endocr. Soc. 170, 583-591 (2014).
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685
99. Ronchi, C. L. et al. Single nucleotide polymorphism microarray analysis in cortisol-secreting adrenocortical adenomas identifies new candidate genes and pathways. Neoplasia N. Y. N 14, 206-218 (2012).
686 687 688 689 690
691
Figures legend
Figure 1: Principal pathways altered in adrenocortical tumors.
Abbreviations: ACC, adrenocortical carcinoma; ACA, adrenocortical adenoma; APA, aldosterone producing adenoma; MC2R, Melanocortin 2 receptor; AT1R, Angiotensin2 type 1 receptor; DVL: Dishevelled.
In red: activated proteins, in light-blue: inactivated proteins
Figure 2: Classification of adrenocortical tumors according to pathology, secretion, prognosis, genomics and pathways.
Abbreviations: ACC, adrenocortical carcinoma; ACA, adrenocortical adenoma; APA, aldosterone producing adenoma; CPA, cortisol producing adenoma; SCA, subclinical cortisol producing adenoma; NSA, non-secreting adenoma; PBMAH, primary bilateral macronodular adrenal hyperplasia; PPNAD,
primary pigmented nodular adrenal dysplasia; GPCR, G protein coupled receptors.
692 693 694 695 696 697 698 699 700 701 702 703
Table1: Mutations in adrenocortical tumors
Abbreviations : APA, aldosterone producing adenoma; CPA, cortisol producing adenoma; SCA, subclinical Cushing adenoma; NSA, non-secreting adenoma; ACC, adrenocortical carcinoma; HLRCC, heredidary leyomyomatosis renal clear cell carcinoma; PBMAH, primary bilateral macronodular adrenal hyperplasia; PPNAD, primary pigmented nodular adrenal dysplasia; MEN1, multiple endocrine neoplasia type 1.
| Gene Symbol | Gene function / Type of alteration | Specific phenotype | Compartment (% mutations) | ||
|---|---|---|---|---|---|
| APA | KCNJ5 | 5 | Potassium channel coupled with G protein. A hotspot mutation induces an abnormal sodium flux inward | More common in women | Somatic (30-50%) (germline) |
| ATP1A1 | 8 | Encodes Na+/K+ ATPase alpha1 subunit. Inactivating mutations induce a plasma membrane depolarization | Somatic (<10%) | ||
| ATP2B3 | 7,8 | Ca2+ ATPase, calcium transporter. Mutations decrease calcium efflux from cytosol | Somatic (<10%) | ||
| CACNA1D | 8,9 | Encodes Cav1.3, alpha subunit of Voltage-dependant Calcium channel. Mutations increase calcium entrance into glomerulosa cells | Somatic (<10% ) Germline | ||
| CACNA1H | 10 | Encodes Cav3.2, alpha subunit of Voltage-dependant Calcium channel. Mutations may increase calcium entrance into glomerulosa cells | Germline | ||
| CYP11B1 / CYP11B2 | 75 | Gene fusion of two steroidogenic enzymes, implicated in cortisol and aldosterone synthesis | Glucocorticoid- remediable aldosteronism | Germline | |
| CPA, SCA and NSA | PRKACA | 15-18 | Catalytic subunit of Protein Kinase A, key effector of the CAMP/PKA pathway. L205R activating hotspot mutation induces constitutive PKA activation. | Overt Cushing | Somatic (20-50% of CPA) |
| GNAS | 20 | Alpha subunit of protein Gs; hotspot gain of function activates adenylate cyclase, independently of any MC2R activation by ACTH | Overt Cushing | Somatic | |
| PRKAR1A | 21 | Negative Regulatory subunit of Protein Kinase A. Inactivating mutations activate PKA through derepression. | Overt Cushing | Somatic | |
| CTNNB1 | 22 | Encodes beta-catenin, key regulator of the Wnt/beta- catenin pathway. Activating mutations induce abnormal activation of this pathway | Low secretion (SCA and NSA), and larger size | Somatic(25%) | |
| Hyperplasia | ARMC5 | 26 | Encodes Armadillo-Repeat Containing 5. Inactivating mutations. | PBMAH ±Meningiomas | Germline, with a somatic second hit |
| PRKAR1A | 76 | Negative Regulatory subunit of Protein Kinase A. Inactivating mutations activate PKA through derepression. | PPNAD, + Carney Complex | Germline | |
| PDE8B, PDE11A | 77,78 | Encodes a phosphodiesterase, catalyzing the degradation of CAMP. Inactivating mutations activate the cAMP/PKA pathway by slowing cAMP degradation | Germline | ||
| GNAS | 79 | Alpha subunit of protein Gs; hotspot gain of function activates adenylate cyclase, independently of any MC2R activation by ACTH | ±McCune- Albright syndrome | Somatic mosaicism | |
| PRKACB | 80 | Catalytic subunit of Protein Kinase A, effector of the CAMP/PKA pathway. Activating mutation induces constitutive PKA activation. | Germline | ||
| MEN1 | 81 | Regulates transcription by coordinating the chromatin | MEN1 | Germline |
| remodeling. Inactivating mutations | |||||
|---|---|---|---|---|---|
| FH | 82 | Mitochondrial enzyme implicated in Krebs cycle. Catalyzes fumarate to malate transformation. Inactivating mutations. | ±HLRCC | Germline | |
| ACC | ZNRF3 | 32-34 | E3 ubiquitin ligase, negatively modulates the Wnt/beta- catenin pathway by promoting LRP5/Frizzled receptor turn- over. Inactivating mutations and homozygous deletions induce an activation of the Wnt/beta-catenin pathway. | Somatic (20%) | |
| CTNNB1 | 22 | Encodes beta-catenin, key regulator of the Wnt/beta- catenin pathway. Activating mutations induce abnormal activation of this pathway | Somatic (15-20%) | ||
| TP53 | 32- 34,38 | Encodes p53, key positive regulator of apoptose, cell-cycle arrest, DNA repair. Inactivating mutations | ±Li Fraumeni syndrome | Somatic (15-20%) (Germline) | |
| CDKN2A | 32- 34,36,8 3 | Tumor suppressor genes encoding two proteins, acting through the activation of p53 and pRB. Inactivating mutations and homozygous deletions | Somatic | ||
| CDK4 | 32- 34,36 | Oncogene, inhibating pRB (encoded by RB1) by phosphorylation. Activted by high level amplifications | Somatic | ||
| TERT | 32- 34,36 | Reverse transcriptase of telomerase complex. Need for maintaining the chromosome length in cancer cells. Activated by high level amplification | Somatic | ||
| RB1 | 32- 34,39 | Encodes pRB, negative regulator of cell-cycle. Inactivating mutations and homozygous deletions | Somatic | ||
| MEN1 | 32-34 | Regulates transcription by coordinating the chromatin remodeling. Inactivating mutations | (MEN1) | Somatic (Germline) | |
| DAXX | 32-34 | Implicated in chromatin remodelling, in alternative lengthening of telomere and in apoptosis. Inactivating mutations | Somatic | ||
| ATRX | 32-34 | Implicated in chromatin remodelling, and in alternative lengthening of telomere. Inactivating mutations | Somatic | ||
| MDM2 | 32- 34,38 | E3 ubiquitin ligase, negatively regulates p53 protein by leading it to its proteasomal degradation. High level amplification | Somatic | ||
| PRKAR1A | 33 | Negative Regulatory subunit of Protein Kinase A. Inactivating mutations activate PKA through derepression. | Somatic | ||
| RPL22 | 33 | Encodes 60S ribosomal protein L22. Inactivating mutations | Somatic | ||
| APC | 84 | Important negative regulator of Wnt/Beta-catenin pathway. Inactivating mutations induce abnormal activation of this pathway | Familial Adenomatous Polyposis | Germline |
| Reference | Disease | Summary |
|---|---|---|
| Transcriptome | ||
| Giordano et al, 2003 | ACC and ACA | ACC and ACA differs in their gene expressions. 91 genes are differently expressed between ACC and ACA (3-fold difference), including IGF2. |
| De Fraipont et al, 20054 | ACC and ACA | Two genes clusters, - IGF2 and steroidogenesis -, predict malignancy. The expression of 14 genes predicts recurrence. |
| Velazquez Fernandes et al, 200544 | ACC and ACA | Gene expression based unsupervised classification separates ACC from ACA. 571 genes are differentially expressed between ACC and ACA, included IGF2 and IGF related genes. |
| Slater et al, 20064 | ACC and ACA | 42 genes are differently expressed (>4-fold) between ACC and normal adrenal tissue, 21 genes between ACC and ACA. |
| Lombardi et al, 200685 | ACC and ACA | Four genes are differentially expressed (>1.5 fold) between ACC vs ACA |
| Fernandez- Ranvier et al, 200886 | ACC and ACA | 25/314 genes of 11q13 locus are differentially expressed between ACC and ACA (p<0.005, >2-fold). |
| Fernandez- Ranvier et al, 200846 | ACC and ACA | Transcriptome separates ACC from ACA, 37 genes are differentially expressed between ACC and ACA (>8-fold) |
| Tombol et al, 200947 | ACC and ACA | 614 genes show expression difference in ACC vs ACA and normal adrenal (2- fold, p>0.05), including IGF2 and TOP2A overexpressed in ACC. |
| Soon et al,200948 | ACC and ACA | Unsupervised classification separated ACC from ACA. 177 genes are diffentially expressed between ACC and ACA. |
| Giordano et al, 200964 | ACC and ACA | 1890 genes are differentially expressed between ACC and ACA (2-fold). Unsupervised classification of ACC according to their gene expression reveals two groups of distinct ACC, which differs in prognosis and mitotic count. |
| De Reyniès et al, 200942 | ACC and ACA | Unsupervised classification separates ACC from ACA, but also reveals two types of ACC with different outcome. Two RT-qPCR predictors are proposed : DLG7-PINK1 for disease-free survival and BUB1B-PINK1 for overall survival. |
| Laurell et al, 200949 | ACC and ACA | Clustering analysis discriminates ACC from ACA, and reveals two subtypes of ACC with different prognosis |
| Lenzini et al, 200787 | APA | Transcriptome profiling of adenomas in the setting of primary aldosteronism identify different tumor types |
| Wang et al, 2011ªª | APA | 14 genes significantly overexpressed in APA vs normal adrenal. CYP11B2 is upregulated in APA, AKR1C3, CYB5 and CYP17 are downregulated. |
| Azizan et al, 201289 | APA | Gene expression separates APA from normal adrenal samples. CYP17A1 and CYP11B1 expression seem to define zona fasciculate-like APA. |
| Monticone et al, 201290 | APA | Gene expression shows 4 fold expression difference between KCNJ5 APA mutant vs normal adrenal for CYP11B2, which is upregulated in KCNJ5 compared to WT APA. |
| Wilmot-Roussel et al, 201319 | ACA | By unsupervised classification, transcriptome separates CPA from a mixed group of SCA, NSA and a few CPA. Steroidogenic genes, cholesterol metabolism genes and glutathione S transferases are positively correlated with cortisol production. |
| Methylome | ||
| Rechache et al57 | ACC and ACA | Methylome distinguishes normal adrenal, ACA, ACC and metastic tissue samples. ACC are hypomethylated in intergenic regions. Identication of 52 genes both hypermethylated and downregulated. |
| Fonseca et al91 | ACC and ACA | 212 CpG island in promoters genes are recurrently hypermethylated in ACC compared to ACA, including genes implicated in cell-cycle regulation and apoptosis. 6 of these genes, including CDKN2A and GATA4, are shown to be down-regulated in ACC |
| Barreau et al59 | ACC | Methylation of CpG island regions distinguishes two types of ACC according to the presence or absence of a hypermethylation (CIMP). CIMP tumors have a lower outcome. 1741 genes present a negative correlation between their expression and methylation, including H19. |
| Howard et al92 | APA | Global hypomethylation of Cpg island in APA compared to normal adrenal. CYP11B2 upregulated and hypomethylated, PRKCA and AVPR1A downregulated and hypermethylated |
| Murakami et al93 | APA | Global hypomethylation of CpG island in APA compaired to normal adrenal. 34 genes upregulated and hypomethylated, among them CYP11B2, MC2R and HPX implicated in aldosterone production |
| miRNome | ||
| Soon et al53 | ACC and ACA | 23 miRNA are differentially expressed between ACC and ACA; ACC with high miR-483-5p and low miR-195 show a lower survival. |
| Tombol et al47 | ACC and ACA | Six miRNAs are differentially expressed in ACC vs ACA and normal adrenal. Combining level expression measures of miR-511 and miR-503 predicts accurately malignancy. |
| Patterson et al52 | ACC and ACA | miRNA expression distinguishes ACC from ACA. 23 miR differentially expressed between ACC and ACA, including miR-483-5p, located within the parentally imprinted IGF2 locus. IGF2 and miR-483-5p expression are correlated. |
| Schmitz et al56 | ACC and ACA | miRNA expression distinguishes ACC from ACA. 248 miRNAs are differentially expressed between ACC and ACA, including miR-675 and miR-335. Low expression of these miRs is associated with malignancy. |
| Ozata et al55 | ACC and ACA | miRNA expression distinguishes ACC from ACA. 72 miRNAs are differentially expressed between ACC and ACA, including miR-503, miR-1202 and miR-1275. High expression of these miRNAs is associated with a poor prognosis. |
| Chabre et al54 | ACC and ACA | 12 miRNAs are differentially expressed between ACC and ACA. 29 miRNAs are differentially expressed between recurring/metastatic ACC and non-recurring ACC, including miR-483-5p and miR-195 which can be detected in patient serum. These circulating miRNAs have a diagnostic and prognostic value. |
| Glover et al94 | ACC and ACA | 85 long non-coding (lnc)RNA are differently expressed between ACC and ACA, 66 are associated with recurrence and the lncRNA PRINS is associated with metastastic disease at presentation |
| Szabo et al95 | ACC and ACA | Difference of expression of circulating miRNAs is found in ACC patients' plasma versus ACA patients' plasma. 5 miR, including miR-483-5p, are candidate circulating biomarkers for malignancy diagnosis. |
| Patel et al96 | ACC and ACA | Difference of expression of circulating miRNAs is found in ACC patients' plasma versus ACA patients' plasma. miR-483-5p et miR 34a are candidate biomarkers for diagnosis |
| Robertson et al97 | APA and CPA | miR24 is differentially expressed in APA than normal adrenal, presents binding sites on CYP11B1/CYP11B2 enzymes and regulate their activity |
| Velazquez Fernandes et al98 | ACA | miRNA discriminates CPA from APA. Specific miRNAs are associated respectively with CPA, APA and non-hyperfunctioning adenomas. |
| Chromosomal alterations | ||
| Zhao et al38 | ACC | MDM2 and CDK4 amplifications in ACC |
| Stephan et al60 | ACC | Several recurrent gains (5, 7, 12, 16q, 20) and losses (1, 3p, 10q, 11, 14q, 15q, 17, 22q) are identified in ACC. Several isolated CGH probes scattered among the genome are associated with survival. The combination of these probes identifies three groups of ACC with different survival. |
| Szabo et al61 | ACC | Several recurrent gains (5, 7, 12, 19q) are identified in ACC. |
| Barreau et al62 | ACC and ACA | A larger proportion of the genome is altered in ACC vs ACA. A qPCR-based diagnostic tool is proposed. The chromosome alterations also contain prognostic information. |
| Ronchi et al63 | ACC and ACA | A larger part of the genome is altered in ACC vs ACA in terms of copy number alterations, but also in terms of loss of heterozygosity. A specific pattern is associated with survival. |
| De Martino et al36 | ACC | Several recurrent gains (5, 7, 12, 19, 20) and losses (1, 22) are reported. Recurrent amplification of CDK4 and deletion of CDKN2A and CDKN2B are identified |
| Ronchi et al99 | CPA | SNP array analyses identify recurrent copy-numbers variations in CPA. 46 recurrent alterations affecting only one gene have been described, including CYP11B1 and CTNNB1 alterations. |
|---|---|---|
| Integrated genomics | ||
| Assie et al32 | ACC | Transcriptome, miRNome, methylome and exome are convergent and define molecular subgroups of ACC with different outcome. |
| Lerario et al33 | ACC | Unsupervised pan-genomics classification (RNA/miRNA sequencing, Copy-number alteration and methylation profiling identified 3 different subgroups of ACC with different outcome |
CAMP/PKA pathway (Cortisol secreting tumors)
A. Cell-cycle regulation
(ACC)
p16 (CDKN2A)
DNA damage
p19 (CDKN2A)
€
MDM2
Cyclin D
Cyclin E
CDK4
CDK2
p53 (TP53)
pRB (RB1)
p21
Caspases
E2F1
G1
S
Apoptosis
Cell-cycle
B.
ACTH
a
Plasma membrane
Y
α
MC2R
₿
Gs a-subunit (GNAS)
AdenylyÌ cyclase
Protein Gs
CAMP
F
CAMP
ATP AMP
Protein Kinase A
R1A Regulatory Subunit (PRKAR1A)
R
R
R
R
Phosphodiesterases (PDE8B/PDE11A)
Ca
Ca
Ca catalytic subunit (PRKACA)
Ca
Ca
Nucleus
CREB
pCREB
PKA targets transcription
(ACC and ACA)
WNT
LRP6
Plasma membrane
Frizzled receptor
ZNRF3
DVL
GSK3B
LRP/Frizzled ubiquitination/internalization
B-Catenin
6-Catenin destruction complex
APC
CKI
Phosphorylation & ubiquitination
AXIN
ub
B-Catenin
Degradation
Nucleus
K
B-Catenin
Degradation
Wnt/Beta- catenin targets transcription
LEF/TCF
(APA)
Mutated Na+/K+ ATPase (ATP1A1)
Mutated Ca2+ channel (CACNA1D)
Ca2+
Angiotensin II
A
Mutated Ca2+ ATPase (ATP2B3)
Na+
Ca2+
AT1R
Normal K channel
[Ca2+]/
K+
Depolarization
Na+
CYP11B2
Mutated K+ channel (KCNJ5)
Aldosterone synthesis
-
ACC
ACA
Pathology
Tumor type Proliferation index
Hyperplasia PPNAD
PBMAH
High
Low
Secretion
± Steroid hormones
APA
CPA
SCA
NSA
Aldo- sterone
Cortisol
Prognosis
Poor outcome
Intermediate outcome
Better outcome
Chromosomes
Numerous chromosomal alterations
Few chromosomal alterations
Few chromosomal alterations
Malignancy signature
Steroidogenic signature
Steroidogenic signature
Transcriptome & miRnome
Genomics
High proliferation signature
Low proliferation signature
ZNRF3, CTNNB1, TP53, RB1, CDKN2A, MEN1, DAXX, ATRX, TERT, TERF2, RPL22, PRKAR1A
KCNJ5 ATP1A1 ATP2B3 CACNAD3
PRKA -CA
Mutations
CTNNB1
PRKA- R1A
ARMC5
Methylome
CIMP- high
Non-CIMP
Pathways
IGF2
p53, pRB
Ca,
CAMP / PKA
Wnt / beta-catenin
CAMP / PKA
Aberrant GPCR
Wnt/beta-catenin