EARLY RELEASE: The Journal of Clinical Endocrinology & Metabolism
The Endocrine Society
Molecular Screening for a Personalized Treatment Approach in Advanced Adrenocortical Cancer
Maria Cristina De Martino, Abir Al Ghuzlan, Sebastien Aubert, Guillaume Assié, Jean-Yves Scoazec, Sophie Leboulleux, Christine Do Cao, Rossella Libè, Cécile Nozières, Marc Lombès, François Pattou, Francoise Borson-Chazot, Ségolène Hescot, Clement Mazoyer, Jacques Young, Isabelle Borget, Annamaria Colao, Rosario Pivonello, Jean-Charles Soria, Jerome Bertherat, M. Schlumberger, Ludovic Lacroix, and Eric Baudin
Departments of Nuclear Medicine and Endocrine Oncology (M.C.D.M., S.L., I.B., M.S., E.B.), Medical Biology and Pathology (A.A.G., J .- Y.S., L.L.), and Medicine (J .- C.S.), and Translational Research Laboratory and Biobank (C.M., L.L.), Institut Gustave Roussy, F-94805 Villejuif, France; Institut de Pathologie (S.A.), Centre Hospitalier Régional Universitaire, F-59037 Lille, France; Institut Cochin (G.A., R.L., J.B.), Institut National de la Santé et de la Recherche Médicale Unité 1016, Centre National de la Recherche Scientifique Unité Mixte de Recherche 8104, Université Paris Descartes, and Department of Endocrinology Assistance Publique Hôpitaux de Paris, Hôpital Cochin, F-75014 Paris, France; Institut National de la Santé et de la Recherche Médicale Unité 1052 (J .- Y.S., C.N., F.B .- C.), Université Lyon 1, Hospices Civils de Lyon, Hôpital Edouard Herriot et Groupement Hospitalier Est, F-69002 Lyon, France; Service d’Endocrinologie (C.D.C.) and Service de Chirurgie Endocrine (F.P.), Hôpital Claude Huriez, Centre Hospitalier Régional Universitaire de F-69437 Lille, France; French Adrenal Cancer Network (R.L.), Institut National du Cancer, F-75014 Paris, France; Institut National de la Santé et de la Recherche Médicale Unité 693 (M.L., S.H.), Faculté de Médecine Paris-Sud, F-94276 Le Kremlin Bicêtre Cedex, France; Service d’Endocrinologie et des Maladies de la Reproduction (J.Y.), Hôpital Bicêtre, F-94275 Le Kremlin-Bicêtre, France; and Department of Clinical Medicine and Surgery (A.C., R.P.), Section of Endocrinology, “Federico II” University, 80131 Naples, Italy
Context: Adrenocortical cancer (ACC) is a rare cancer with poor prognosis and scant treatment options. In ACC, no personalized approach has emerged but no extensive molecular screening has been performed to date.
Objective: The objective of the study was to evaluate the presence of a large number of potentially targetable molecular events in a large cohort of advanced ACC.
Design, Setting, and Participants: We used hot spot gene sequencing (Ion Torrent, 40 patients) and comparative genomic hybridization (CGH; 28 patients; a subset of the entire cohort) in adult stage III-IV ACC samples to screen for mutations and copy number abnormalities of potential interest for therapeutic use in 46 and 130 genes, respectively.
Results: At least one copy number alteration or mutation was found in 19 patients (47.5%). The most frequent mutations were detected on TP53, ATM, and CTNNB1 [6 of 40 (15%), 5 of 40 (12.5%), and 4 of 40 (10%), respectively]. The most frequent copy number alterations identified were: amplification of the CDK4 oncogene (5 of 28; 17.9%) and deletion of the CDKN2A (4 of 28; 14.3%) and CDKN2B (3 of 28; 10.7%) tumor suppressor genes. Amplifications of FGFR1, FGF9, or FRS2 were discovered in three subjects (10.7%). Associated alterations were: deletions of CDKN2A, CDKN2B with ATM mutations, and TP53 mutations with CTNNB1 mutations.
Conclusions: No simple targetable molecular event emerged. Drugs targeting the cell cycle could be the most relevant new therapeutic approach for patients with advanced ACC. Inhibitors of the fibroblast growth factor receptor pathway could also be a therapeutic option in a subset of pa- tients, whereas other targeted therapies should be considered on a case-by-case basis.
A drenocortical cancer (ACC) is one of the most ag- gressive solid tumors in humans, as evidenced by a 5-year survival rate of below 15% at the metastatic stage (1,2). Its incidence is about one to two new cases per one million people per year, with an increased occurrence dur- ing childhood and the fourth to fifth decades of life. Sur- gery remains the only curative treatment in patients diag- nosed at an early stage, whereas treatment options for patients with advanced ACC are still scant (1-4).
A personalized approach to cancer treatment is based on the use of drugs able to target specific molecular alter- ations playing pivotal roles in oncogenesis that are tar- getable in a given patient. The study of ACC-associated syndromes has suggested that the IGF-II signaling path- way (Beckwith-Wiedemann syndrome), p53 (Li-Frau- meni syndrome), or Wnt/ß-catenin signaling (familial ad- enomatous polyposis) are currently the most attractive targets for ACC (2). Among these targets, only IGF-I re- ceptor antagonists are currently under investigation in ACC. The results of phase I studies evaluating the effects ofIGF-I receptor antagonists have already been published, demonstrating that drugs targeting the IGF-I receptor can induce tumor stabilization in a subset of patients, but ob- jective tumor responses are rare (5). Regarding p53 or Wnt/B-catenin signaling, drugs targeting these pathways are still in early developmental phases. Preliminary studies using antiangiogenics or epithelial growth factor (EGFR)- targeted drugs in patients with ACC have yielded disap- pointing results (2, 6-11). However, both the lack of screening for the relevant targeted events and the full dem- onstration of the relevance of such targets in ACC may explain these results.
In sporadic ACC, several studies have used gene ex- pression microarrays or comparative genomic hybridiza- tion (CGH), mainly to establish criteria aimed at differ- entiating benign from malignant tumors or to identify prognostic markers (12, 13). Few studies have evaluated the presence of putative biomarkers for new targeted agents in ACC. Indeed, small numbers of molecular alter- ations (up to five per series) were screened in 8-35 ACC patients, among whom only 11 exhibited indisputable cri- teria of malignancy, as evidenced by the presence of ex- traadrenal disease (6, 14-16).
We hypothesize that testing a large number of poten- tially targetable molecular events in patients with indis- putably malignant ACC could accelerate the drug devel- opment process in this rare and aggressive solid tumor. To achieve this goal, we collected a large number of primary
or locally recurrent well-characterized malignant ACCs. We used hot spot gene sequencing and CGH to evaluate the presence of mutations and copy number abnormalities in, respectively, 46 and 130 genes of potential interest.
Materials and Methods
Patient population and sample acquisition
Samples from different ACC patients were collected in four centers of the French Cortico- et Medullo-Surrénale, les Tumeurs Endocrines network (Institut National du Cancer): the Institut Gustave Roussy, the Centre Hospitalier Régional et Universita- ire-Tumorothèque-CRRC de Lille; the Hospices Civils de Lyon; and the Hôpital Cochin, Paris. Samples were snap frozen and stored in tumor biobanks according to national ethics recom- mendations and local procedures. The 40 samples selected for the present study and processed for DNA extraction had to meet the following inclusion criteria: ACC diagnosis confirmed by a local expert pathologist (17), indisputable malignancy (stage III- IV) based on the European Network for the Study of Adrenal Tumors staging definition (18), the presence of more than 50% of tumor cells based on histological examination, and age older than 17 years. The following clinical parameters were recorded: date of diagnosis, age, gender, Weiss score (17), mitotic count (available in 39 patients and calculated in 50 or 10 high-power fields for 29 or 10 patients, respectively), hormonal status based on hormonal measurements, tumor node metastasis stage, and previous systemic treatments. Informed consent was obtained from each patient.
DNA isolation
Genomic DNA was extracted from several 10- to 20-um sec- tions of each tumor specimen after digestion with proteinase K (3 h), using the DNeasy tissue kit (QIAGEN), according to the manufacturer’s protocol. DNA concentrations were assessed with the Qubit fluorometer (Invitrogen).
Mutational analysis
DNA sequencing was performed with the Ion Torrent tech- nique (PGM sequencer; Life Technologies). Ten nanograms of each sample were amplified with multiplex PCR based on the cancer panel primers pool followed by library preparation ac- cording to recommendations of the Ion AmpliSeq library kit 2.0 protocol (Life Technologies), using 17 cycles for multiplex PCR and adding Ion Xpress bar code adapters during the ligation step to allow for subsequent pooling of the samples. The list of se- quences covered by multiplex PCR (.bed file) is available (www.ampliseq.com). Each individual library was quantified us- ing the Qubit fluorometer (Invitrogen) and controlled using a bioanalyzer (Agilent Technologies). Libraries were diluted to ob- tain a final dilution of 3 ng/ML for each library, and 8 or 16 libraries were pooled together for amplification on spheres using the Ion OneTouch 200 template kit version 2 (Life Technolo- gies). Spheres obtained by eight different libraries were loaded
Abbreviations: ACC, adrenocortical cancer; CGH, comparative genomic hybridization; EGFR, epithelial growth factor; FGF, fibroblast growth factor; FGFR, FGF receptor; MAS, McCune-Albright syndrome; PJS, Peutz-Jegher’s syndrome.
onto an Ion 316 chip, and spheres obtained by the remaining 32 libraries were loaded onto an Ion 318 chip for sequencing using the Ion PGM 200 sequencing kit for 520 flows.
Using 190 primer pairs, this approach allowed the simulta- neous study of hot spot regions of 46 critical oncogenes or tumor suppressor genes of potential interest to predict drug sensitivity (reported in Supplemental Table 1, published on The Endocrine Society’s Journals Online web site at http://jcem.endojournal- s.org). All reported somatic genetic variants were compared with the relative GRCh37 (h19) reference sequences using Torrent Suite version 2.2 software (variantCaller v2.2.3-31149; Life Technologies) and annotated using Alamut version 2.2 software (Interactive Biosoftware). The variants with a read frequency higher than 10%, none synonymous with and not known as common polymorphisms, were retained as interesting variants (mutations) and were confirmed by Sanger direct sequencing, as previously described (19). Moreover, AKT1 (exon 4), PI3KCA (exons 5-10-21), and CTNNB1 (exon 3) were sequenced in all samples to complete the information obtained with the Ampliseq cancer panel.
Oligonucleotide CGH microarrays
For microarray hybridizations, 400 ng of DNA from each DNA sample was digested and sample integrity was measured using an Agilent bioanalyzer. The test DNA samples were labeled with Cy5 fluorescent dye and the reference DNA samples were labeled with Cy3 fluorescent dye using the genomic DNA enzy- matic labeling kit (Agilent Technologies), following the manu- facturer’s protocol. Cy3-labeled and Cy5-labeled DNAs were hybridized to the SurePrint G3 Human CGH Microarray 4×180K (Agilent Technologies), prior to washing and scanning with the Agilent scanner G2565CA.
Oligonucleotide CGH microarray analysis
Oligonucleotide CGH array processing was performed as de- tailed in the manufacturer’s protocol (version 7.1, December 2011; http://www.agilent.com). Data were extracted from scanned images using the Feature Extraction software (version 10.7.3,1; Agilent Technologies), along with protocol CGH_107_Sep09. Acquired signals were normalized according to their dye and local GC percentage content using in-house scripts under the R statistical environment (http://cran.r-projec- t.org). The resulting log2 (ratio) values were segmented using the circular binary segmentation algorithm (20) implementation from the DNAcopy package for R. Aberration status calling was automatically performed for each profile according to its internal noise [absolute variation of log2 (ratio) values across consecutive probes on the genome]. All genomic coordinates were estab- lished on the University of California Santa Cruz build hg19 Homo sapiens genome (21). The analysis focused on 130 genes of potential interest to predict drug sensitivity, including the 46 genes studied by sequencing and additional genes involved in the IGF pathway as reported in Supplemental Table 1. We described all copy number alterations above zero as gains and all the al- terations below zero as losses. However, only copy number gains with log2 ratio values higher than 1 were considered as ampli- fied, and copy number losses with a log2 ratio value lower than 1 were considered as deletions.
Statistical analysis
Descriptive parameters were calculated using statistical soft- ware (SPSS version 15.0; SPSS Inc). Quantitative data were ex- pressed using means and SD and medians and ranges. Qualitative data were expressed using percentages.
Results
Study population
The study population included 40 adult patients. The main clinical characteristics of these patients are reported in Table 1. Malignancy was ascertained by the stage: 10 and 30 patients had stage III or IV disease, respectively. All samples were collected from primary tumors (34 samples) or local recurrences (six samples). Each sample corre- sponded to a different patient. In two patients with stage IV disease, a complete Weiss score could not be assigned because of insufficient available material.
Hot spot gene sequencing
Using the Cancer Panel primers pool for Ion Torrent sequencing (Life Technologies), all 40 samples were in- formative for the analysis of sequence variants of the eval- uated hot spot regions. The mean number of mapped reads per tumor was 167 680 ± 122 768 (mean ± SD); base coverage depth per tumor was 835X ± 598; mean read on-target was 90% ± 6% (median 93%; range 61%- 95%); the average coverage at 100 times was 93.4% ± 2% (median 93%; range 89%-99%) and the median of eight variants (range 5-16) was detected. Variants were re- ported per tumor. The interesting variants retained were all confirmed by Sanger sequencing.
More than one quarter (14 of 40, 35%) of the samples exhibited at least one mutation, as defined above (details reported in Supplemental Table 2). Single mutations in the TP53 gene were found in six samples (15%). All of these mutations are predicted to be associated with a disruption of p53 function (www.p53.iarc.fr). Single mutations in the ATM gene were found in five samples (12.5%). Single mutations of CTNNB1 were found in four samples (10%), all located in exon 3.
In individual cases (1 of 40; 2.5%), single mutations in the genes coding for ERBB4, FLT3, STK11, SMO, and GNAS were found (Supplemental Table 2). A GNAS-ac- tivating mutation was found in a patient presenting with an isolated Cushing’s syndrome (stage III ACC, WS 6) and no clinical evidence of fibrous dysplasia. A STK11 muta- tion was found in a patient presenting with an isolated nonsecreting ACC (stage III, WS 6).
DNA copy number changes
Twenty-eight of the 40 evaluated samples (a subset of the entire cohort) generated informative profiles by CGH.
| Parameters | Class | Number | Frequency, % | |
|---|---|---|---|---|
| Gender | 40 | |||
| Female | 29 | 72,5 | ||
| Male | 11 | 27,5 | ||
| Age, y, median 55 (19-77) y, mean 52 ± 16 y | ||||
| 19-29 | 3 | 7.5 | ||
| 30-39 | 9 | 22.5 | ||
| 40-49 | 4 | 10 | ||
| 50-59 | 8 | 20 | ||
| 60-69 | 12 | 30 | ||
| ≥70 | 4 | 10 | ||
| Weiss score, median 7 (4-9), mean 6.9 + 1.4 | ||||
| 4 | 2 | 5 | ||
| 5 | 4 | 10 | ||
| 6 | 10 | 25 | ||
| 7 | 10 | 25 | ||
| 8 | 4 | 10 | ||
| 9 | 8 | 20 | ||
| Unknown | 2 | 5 | ||
| Mitotic index | ||||
| ≤5 | 3 | 7.7 | ||
| >5 but <10 | 9 | 23.1 | ||
| ≥10 | 27 | 69.2 | ||
| Unknown | 1 | 2.5 | ||
| Hormonal | ||||
| secretion | ||||
| Present | 26 | 65 | ||
| Cortisol | 13 | 32.5 | ||
| Cortisol and | 9 | 22.5 | ||
| androgens | ||||
| Androgens | 3 | 7.5 | ||
| Androgens | 1 | 2.5 | ||
| and | ||||
| estrogens | ||||
| Absent | 14 | 35 | ||
| Stage | III | 10 | 25 | |
| IV | 30 | 75 | ||
| Origin of sampling | ||||
| Primary | 34 | 85 | ||
| tumors | ||||
| Recurrence | 6 | 15 | ||
| Previous systemic treatment | ||||
| Yes | 14 | 35 | ||
| Mitotane | 9 | 22.5 | ||
| Mitotane and | 5 | 12.5 | ||
| chemotherapy | ||||
| No | 26 | 65 | ||
The average profile of the 28 samples is reported in Figure 1. Most samples exhibited gains of oncogenes or loss of tumor suppressor genes, as reported in Supplemental Ta- ble 3.
Twelve of these profiles (42.9%) contained at least one deletion or amplification in the expression of the 130 eval- uated genes. Among these genes, the recurrent retained abnormality was the amplification of the CDK4 oncogene (Chr12q14) observed in 5 of the 28 evaluated samples (17.9%); the deletion of the CDKN2A tumor suppressor
gene (Chr9p21) observed in four other samples (14.3%) and the deletion of the tumor suppressor gene CDKN2B (Chr9p21) observed in three samples. All three of these patients with deletion of CDKN2B were also deleted for CDKN2A (Figure 2 and Supplemental Table 3). Overall, 32.1% of the samples exhibited an amplification/deletion (Figures 2 and 3) of one of these genes. Amplification of three different components of the fibroblast growth factor (FGF) pathway [FGF receptor (FGFR)-1; FGF9 and FRS2 [was found in 3 different samples (3/28; 10.7%) (Figure 2).
Integrated results of hot spot gene sequencing and DNA copy number variation
Taken together, the results of DNA copy number vari- ation and hot spot gene sequencing data (Figure 2) showed that almost half of the evaluated samples exhibited at least one molecular event (19 of 40; 47.5%).
Some alterations were recurrently associated, such as the ATM mutation and loss of CDKN2A observed in three of the five samples harboring ATM mutations. Three of the four patients with a CTNNB1 mutation were also carrying a TP53 mutation.
Discussion
To the best of our knowledge, this is the first study to screen a large series of indisputably malignant ACC for the presence of a large number (more than 40) of structural DNA changes that could help to select new targeted drugs. The strengths of this study include a selective use of ma- lignant ACC samples based on documented local or dis- tant invasion and the use of an innovative sequencing sys- tem, the Ion Torrent, which allows one to generate results concerning a large set of genes in a time span compatible with clinical needs. None of the previous studies evaluated more than five potential targets by DNA sequencing, and none of them were able to integrate the results of DNA sequencing with the study of DNA copy number altera- tions in this type of malignancy. Thirty-five percent of the samples exhibited at least one mutation and 42.9% had at least one deletion or amplification in the evaluated genes. Overall, almost half of the samples (47.5%) had at least one molecular abnormality. The present study provides evidence that drugs targeting the cell cycle represent the most relevant potential new therapeutic strategy for pa- tients with advanced ACC. Inhibitors of the FGFR path- way could be a potential target for treatment in a subset of ACC patients, whereas treatment with other targeted ther- apies could be considered exclusively on a case-by-case basis. The rarity of the molecular alterations usually used
in personalized oncology suggests that no short-term re- sults can be expected from the use of new available li- censed agents in most ACC patients.
Based on DNA sequencing of 46 potentially actionable oncogenes or tumor suppressor genes, we identified alter- ations including mutations in ERBB4 and FLT3 genes
(Figure 2). No mutations were identified on EGFR, BRAF, KIT, PIK3CA, RET, or PDGFR-A. Previously, two stud- ies evaluated the results of EGFR (exons 18-21) sequenc- ing in ACC and reported a mutation frequency of, respec- tively, 0% (0 of 30) and 11% (4 of 35) in the cases (6, 14). A previous study conducted by our group found no mu-
Average profile of the population
1.0
MDM2
CDK4
0.5
Frequency of losses | gains
0.0
-0.5
CDK2NA/B
ATM
-1.0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 202122
X
Genomic position
Patient
Mutations retained
Copy number alterations retained
2
BRCA1
3
ATM
CDKN2A
CDKN2B
24
ATM
CDKN2A
CDKN2B
RPTOR
14
CDKN2A
11
ATM
CDKN2A
5
ATM
CDK4
19
CDK4
20
ERBB4
SMO
FLT3
CDK4
12
CDK4
TP53
9
CDK4
MDM2
FRS2
23
ATM
CTNNB1
TP53
7
TP53
22
CTNNB1
TP53
FGFR1
MYC
ABL1
PTK2
29
CTNNB1
TP53
8
CTNNB1
6
TP53
FGF9
FLT3
26
TP53
10
STK11
25
GNAS
tation in EGFR (exon 18, 19, and 21), BRAF, PIK3CA, and JAK2 in 18 ACCs, which is in line with the present results (15). Because some target gene regions were not well explored by Ampliseq Cancer Panel primers, we fur- ther validated our results by direct sequencing, and we confirmed the absence of AKT1 (exon 4) mutations and PI3KCA mutations (exons 5, 10, and 21) in all of our samples (data not shown).
Using CGH, different components of the FGFR path- way were found to be amplified in three different samples. A growing number of studies indicate that inhibition of the FGFR pathway may be an effective therapeutic option against cancer and several drugs targeting the FGFR path- way are under clinical development (phases I-III) (22). A microarray gene expression analysis performed in 11 ACC samples evidenced frequent up-regulation of FGFR1 and FGFR4 in one study (23). A recent study suggested that FGFR4 overexpression and gene amplification have a prognostic value in ACC (24). In the present study, FGFR1-FGFR were not frequently amplified, but gains of FGFR1-FGFR4 were frequently observed [22 of 28 (78.6%) samples exhibited gains in at least one of these
receptors; Supplemental Table 3]. Conversely, the study of copy number variations suggests that HER2 (ERBB2) am- plification is not present in ACC, in accordance with pre- vious reports (25-27). In addition, many other putative biomarkers of targeted treatments such as PTEN, AKT, ALK, c-MET/HGFR, EGFR, PDGFR, and ESR1 did not exhibit amplification/deletion, which completes the neg- ative results of the mutation screening approach.
CTNNB1 and TP53 mutations are genetic alterations well known to play a role in the pathogenesis of ACC (2). Based on DNA sequencing, we identified CTNNB1 mu- tations in four samples (10%) and TP53 mutations in six samples (15%). Interestingly, both CTNNB1 and TP53 mutations (Figure 2) were found in three samples. More- over, CTNNB1 gains were observed in all abnormal (mu- tated or deleted) TP53 samples (Supplemental Table 3), whereas they were rare (9%) in wild-type TP53 samples. This observation, which is very intriguing from the patho- genic point of view, and hitherto never clearly described in ACC, suggests an association between the status of the TP53 and CTNN1B genes. CTNNB1 mutations are early events in adrenocortical oncogenesis (16, 28). Previously
G1 --- >S-phase in ACC
APOPTOSIS
p53 (TP53) 15%; 3.6%
ATR
p14ARF (CDKN2B) 14.3%
p16 (CDKN2B) 14.3%
ATM 12.5%
DNA DAMAGE
MITOGENIC SIGNALS
p15 (CDKN2A) 10.7%
p21 (CDKN1)
p57KIP2 (CDKN1C)
MDM2 3.6%
PD0332991
RO5503781 RO5045337
cMYC 3.6%
CDK4 17.9%
CDK6
CDK2
CyclinD1
CyclinD3
CyclinE
B-catenin (CTNNB1) 15%
PRB
PRB
E2Fs
E2Fs
TRASCRIPTION of genes for S-phase
At least 40% of patients have abnormalities in this pathway
published studies reported CTNNB1 mutations in 20%- 30% of the evaluated samples (16, 28). The frequency of CTNNB1 mutations in our population was in the low range of previously published reports. Because some rare CTNNB1 exon 3 mutations previously described in ACC were outside the hot spot regions investigated in our study, we performed direct sequencing of exon 3 of the CTNNB1 gene in all ACC samples. This technique allowed us to detect CTNNB1 mutations in two additional patients. This led to a whole CTNNB1 mutation rate of 15%, which is still in the low range of previously published reports.
TP53 mutations were suggested to be late events in adult adrenocortical oncogenesis (29). Previously pub- lished studies reported TP53 mutations in 10%-70% of the evaluated samples (30). The frequency of the TP53 mutation in our population was also in the low range of previously published reports. The difference in the fre- quency of CTNNB1 and TP53 in our study compared with previously published reports could be due to the low sample size and the heterogeneity of the ACC patient pop- ulation. Interestingly, based on CGH, one sample exhib- ited a TP53 deletion and another MDM2 amplification (Figure 2 and Supplemental Table 3). The MDM2 gene has been reported to be overexpressed in ACC (13) and is a potential target for treatment (www.clinicaltrials.gov). Regarding the IGF pathway, no major alteration was found with the methods used in our (Study Supplemental Table 3).
DNA-damage response and G1 cell cycle progression are new pathways whose exploration could be interesting in ACC patients. By sequencing, we demonstrated for the first time the presence of ATM mutations/interesting vari- ants in five ACC patients. ATM plays a role in cell response to DNA damage and genome stability. Mutations in this gene are associated with ataxia telangiectasia, a disorder associated with high frequency of cancer (31). All the de- tected ATM variants are considered of interest because they have been described as potentially involved in ma- lignancy (32-34). The frequencies of these ATM variants in our series is higher than that expected in the general population [relative database db single-nucleotide poly- morphism (build 137); Supplemental Table 2]. A recent study demonstrated the ATM gene copy number is re- duced in ACC compared with adrenal adenomas (35). We detected alterations of several key components of the cell cycle by CGH analysis: CDK4 amplification and CDKN2A and CDKN2B deletion. CDK4 encodes for a cyclin-dependent kinase that plays a crucial role in G1-S phase cell cycle progression. CDK4 has already been sug- gested to be overexpressed in ACC as compared with nor- mal adrenals (36). By alternative splicing, CDKN2A can
encode for two different gene products: the tumor sup- pressor protein p16 (a CDK4 inhibitor) and ARF (a sta- bilizer of p53) (Figure 3) (37). CDKN2B encodes for p15, another CDK4/6 inhibitor. Loss of nuclear immunostain- ing for p16 has been reported in three of seven ACCs (38). The integration of data obtained by sequencing and CGH allowed us to discover that three of the five ATM mutated samples also had a CDKN2A deletion and one had CDK4 amplification, suggesting a functional synergism between DNA damage checkpoints and G1 cell cycle progression pathways in the pathogenesis of ACC (Figures 2 and 3).
Finally, by sequencing, we detected two mutations de- scribed as part of well-characterized genetic disorders and hitherto never reported in adult ACC: mutations in the GNAS and STK11 genes. Activating mutations in the GNAS gene have been described in sporadic ACTH-inde- pendent macronodular adrenal hyperplasia and rarely in benign adrenal tumors. These mutations can cause the McCune-Albright syndrome (MAS), which is associated with ACTH-independent macronodular adrenal hyper- plasia (39). A single case of ACC with a somatic mutation of GNAS was recently reported in a child unaffected by MAS. Mutations in the STK11 tumor suppressor gene are associated with Peutz-Jegher’s syndrome (PJS). A single case of ACC was recently reported in a child with PJS (40). We did not find signs or symptoms, respectively, sugges- tive of MAS or PJS in either of these patients or in their family members. We suspect that both of these mutations could play a role in ACC oncogenesis in the affected patients.
Our study suffers from several limitations: 35% of the patients had received previous medical treatment and we were unable to discriminate between germline and somatic mutations. In addition, we did not explore mRNA and protein expression and protein function. Finally, we used a sequencing system that detects many, but not all, mu- tations in the evaluated genes. However, we decided to focus on the selected genomic events that are already widely used in oncology as predictors of drug responses.
Conclusions
In conclusion, we identified 47.5% mutations or CGH alterations in a large series of ACC patients. To our knowl- edge, this is the first time that ATM, STK11, and GNAS mutations have been reported in adult ACC patients. No relevant molecular alteration suggests the likelihood of a simple molecular-driven targeted approach in ACC pa- tients in the short term. However, our study predicts a potential future role for new compounds targeting DNA- damage responses, G1 cell cycle progression, and the FGFR pathway.
Acknowledgments
We thank the following for their help in this work: Dr L. Lacroix (Laboratoire de Recherche Translationnelle); Professor V. Lazar (Unité de Génomique Fonctionnelle and Centre de Ressources Biologiques and Dr P. Vielh (Institut Gustave Roussy); and Pro- fessor M. C. Copin (Tumorothèque CRRC de Lille; Tumor- orheque des Hospices Civils de Lyon) and Professor B Terris (Tumor Bank of Cochin Hospital). For technical assistance we also thank the following people: Ludovic Bigot; Marie Breckler; Isabelle Miran; Nelly Motte; Patrick Saulnier (Laboratoire de Recherche Translationnelle; Institut Gustave Roussy); Delphine Aarnaud (Centre de Ressources Biologiques; Institut Gustave Roussy); Catherine Richon (Unité de Génomique Fonctionnelle; Institut Gustave Roussy); Fernande Rene Corail (Cochin Hos- pital); and Géraldine Gouysse (Tumororheque des Hospices Civils de Lyon). We also thank Lorna Saint Ange for editing.
Address all correspondence and requests for reprints to: Lu- dovic Lacroix, PhD, Translational Research Laboratory and Biobank, Institut Gustave Roussy, 114 Rue Edouard Vaillant, 94805 Villejuif Cedex, France. E-mail: ludovic.lacroix@igr.fr; and Eric Baudin, MD, PHD, Department of Nuclear Medicine and Endocrine Oncology, Institut Gustave Roussy, 114 Rue Ed- ouard Vaillant, 94805 Villejuif Cedex, France. E-mail: eric.baudin@igr.fr.
This work was supported by the grant from the Diplôme Universitaire Européen de Recherche Translationnelle et Clin- ique en Cancérologie. This research project received a “Bourse de Recherche en Endocrinologie 2012” grant from the French So- ciety of Endocrinology.
Disclosure Summary: The authors have nothing to disclose.
References
1. Baudin E, Leboulleux S, Al Ghuzlan A, et al. Therapeutic manage- ment of advanced adrenocortical carcinoma: what do we know in 2011? Horm Cancer. 2011;2:363-371.
2. Fassnacht M, Libe R, Kroiss M, Allolio B. Adrenocortical carci- noma: a clinician’s update. Nat Rev Endocrinol. 2011;7:323-335.
3. Berruti A, Baudin E, Gelderblom H, et al. Adrenal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2012;23(suppl 7):vii131-138.
4. Fassnacht M, Terzolo M, Allolio B, et al. Combination chemother- apy in advanced adrenocortical carcinoma. N Engl J Med. 2012; 366:2189-2197.
5. Naing A, Lorusso P, Fu S, et al. Insulin growth factor receptor (IGF- 1R) antibody cixutumumab combined with the mTOR inhibitor temsirolimus in patients with metastatic adrenocortical carcinoma. Br J Cancer. 2013;108(4):826-830.
6. Adam P, Hahner S, Hartmann M, et al. Epidermal growth factor receptor in adrenocortical tumors: analysis of gene sequence, protein expression and correlation with clinical outcome. Mod Pathol. 2010;23:1596-1604.
7. Berruti A, Sperone P, Ferrero A, et al. Phase II study of weekly paclitaxel and sorafenib as second/third-line therapy in patients with adrenocortical carcinoma. Eur J Endocrinol. 2012;166:451-458.
8. Gross DJ, Munter G, Bitan M, et al. The role of imatinib mesylate (Glivec) for treatment of patients with malignant endocrine tumors positive for c-kit or PDGF-R. Endocr Relat Cancer. 2006;13:535- 540.
9. Kroiss M, Quinkler M, Johanssen S, et al. Sunitinib in refractory adrenocortical carcinoma: a phase II, single-arm, open-label trial. J Clin Endocrinol Metab. 2012;97:3495-3503.
10. Quinkler M, Hahner S, Wortmann S, et al. Treatment of advanced adrenocortical carcinoma with erlotinib plus gemcitabine. J Clin Endocrinol Metab. 2008;93:2057-2062.
11. Wortmann S, Quinkler M, Ritter C, et al. Bevacizumab plus cape- citabine as a salvage therapy in advanced adrenocortical carcinoma. Eur J Endocrinol. 2010;162:349-356.
12. de Reynies A, Assie G, Rickman DS, Tissier F, Groussin L, Rene- Corail F. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malig- nancy and survival. J Clin Oncol. 2009;27:1108-1115.
13. Szabo PM, Tamasi V, Molnar V, et al. Meta-analysis of adrenocor- tical tumour genomics data: novel pathogenic pathways revealed. Oncogene. 2010;29:3163-3172.
14. Kotoula V, Sozopoulos E, Litsiou H, Fanourakis G, et al. Mutational analysis of the BRAF, RAS, and EGFR genes in human adrenocor- tical carcinomas. Endocr Relat Cancer. 2009;16:565-572.
15. Ameur N, Lacroix L, Motte N, et al. Mutational status of EGFR, BRAF, PI3KCA and JAK2 genes in endocrine tumors. Int J Cancer. 2009;124:751-753.
16. Masi G, Lavezzo E, Iacobone M, Favia G, Palu G, Barzon L. Inves- tigation of BRAF and CTNNB1 activating mutations in adrenocor- tical tumors. J Endocrinol Invest. 2009;32:597-600.
17. Tissier F, Aubert S, Leteurtre E, et al. Adrenocortical tumors: im- proving the practice of the Weiss system through virtual microscopy: a National Program of the French Network INCa-COMETE. Am J Surg Pathol. 2012;36:1194-1201.
18. Fassnacht M, Johanssen S, Quinkler M, et al. Limited prognostic value of the 2004 International Union Against Cancer staging clas- sification for adrenocortical carcinoma: proposal for a Revised TNM Classification. Cancer. 2009;115:243-250.
19. Boichard A, Croux L, Al Ghuzlan A, et al. Somatic RAS mutations occur in a large proportion of sporadic RET-negative medullary thyroid carcinomas and extend to a previously unidentified exon. J Clin Endocrinol Metab. 2012;97(10):E2031-E2035
20. Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics (Oxford, England). 2004;5:557-572.
21. Karolchik D, Baertsch R, Diekhans M, et al. The UCSC Genome Browser Database. Nucleic Acids Res. 2003;31:51-54.
22. Dieci MV, Arnedos M, Andre F, Soria JC. Fibroblast growth factor receptor inhibitors as a cancer treatment: from a biologic rationale to medical perspectives. Cancer Discov. 2013;3:264-279.
23. Laurell C, Velazquez-Fernandez D, Lindsten K, et al. Transcrip- tional profiling enables molecular classification of adrenocortical tumours. Eur J Endocrinol. 2009;161:141-152.
24. Brito LP, Ribeiro TC, Almeida MQ, et al. The role of fibroblast growth factor receptor 4 overexpression and gene amplification as prognostic markers in pediatric and adult adrenocortical tumors. Endocr Relat Cancer. 2012;19:L11-L13.
25. Gupta D, Shidham V, Holden J, Layfield L. Value of topoisomerase IIa, MIB-1, p53, E-cadherin, retinoblastoma gene protein product, and HER-2/neu immunohistochemical expression for the prediction of biologic behavior in adrenocortical neoplasms. Appl Immuno- histochem Mol Morphol. 2001;9:215-221.
26. Nakamura M, Miki Y, Akahira J, et al. An analysis of potential surrogate markers of target-specific therapy in archival materials of adrenocortical carcinoma. Endocr Pathol. 2009;20:17-23.
27. Saeger W, Fassnacht M, Reincke M, Allolio B. Expression of HER- 2/neu receptor protein in adrenal tumors. Pathol Res Pract. 2002; 198:445-448.
28. Tissier F, Cavard C, Groussin L, et al. Mutations of ß-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res. 2005;65:7622-7627.
29. Libe R, Groussin L, Tissier F, et al. Somatic TP53 mutations are
doi: 10.1210/jc.2013-2165
relatively rare among adrenocortical cancers with the frequent 17p13 loss of heterozygosity. Clin Cancer Res. 2007;13:844-850.
30. Wasserman JD, Zambetti GP, Malkin D. Towards an understanding of the role of p53 in adrenocortical carcinogenesis. Mol Cell Endo- crinol. 2011;351:101-110.
31. Poehlmann A, Roessner A. Importance of DNA damage checkpoints in the pathogenesis of human cancers. Pathol Res Pract. 2010;206: 591-601.
32. Liberzon E, Avigad S, Stark B, et al. Germ-line ATM gene alterations are associated with susceptibility to sporadic T-cell acute lympho- blastic leukemia in children. Genes Chromosomes Cancer. 2004; 39:161-166.
33. Rudd MF, Sellick GS, Webb EL, Catovsky D, Houlston RS. Variants in the ATM-BRCA2-CHEK2 axis predispose to chronic lympho- cytic leukemia. Blood. 2006;108:638-644.
34. Lahdesmaki A, Kimby E, Duke V, Foroni L, Hammarstrom L. ATM mutations in B-cell chronic lymphocytic leukemia. Haematologica. 2004;89:109-110.
35. Ye J, Qi Y, Wang W, et al. Lower expression of ATM and gene deletion is more frequent in adrenocortical carcinomas than adre- nocortical adenomas. Endocrine. 2012;41:479-486.
36. Bourcigaux N, Gaston V, Logie A, Bertagna X, Le Bouc Y, Gicquel C. High expression of cyclin E and G1 CDK and loss of function of p57KIP2 are involved in proliferation of malignant sporadic adre- nocortical tumors. J Clin Endocrinol Metab. 2000;85:322-330.
37. Dominguez-Brauer C, Brauer PM, Chen YJ, Pimkina J, Raychaud- huri P. Tumor suppression by ARF: gatekeeper and caretaker. Cell Cycle (Georgetown, Tex.). 2010;9(1):86-89.
38. Pilon C, Pistorello M, Moscon A, et al. Inactivation of the p16 tumor suppressor gene in adrenocortical tumors. J Clin Endocrinol Metab. 1999;84:2776-2779.
39. Sidhu A, Debelenko L, Misra VK. Infantile adrenocortical tumor with an activating GNAS1 mutation. J Clin Endocrinol Metab. 2013;98:E115-E118.
40. Yalcin S, Kirli E, Ciftci AO, et al. The association of adrenocortical carcinoma and thyroid cancer in a child with Peutz-Jeghers syn- drome. J Pediatr Surg. 2011;46:570-573.