SNP Array Profiling of Childhood Adrenocortical Tumors Reveals Distinct Pathways of Tumorigenesis and Highlights Candidate Driver Genes
Eric Letouzé, Roberto Rosati, Heloisa Komechen, Mabrouka Doghman, Laetitia Marisa, Christa Flück, Ronald R. de Krijger, Max M. van Noesel, Jean-Christophe Mas, Mara A. D. Pianovski, Gerard P. Zambetti, Bonald C. Figueiredo, and Enzo Lalli
Programme Cartes d’identité des Tumeurs (E.Le., L.M.), Ligue Nationale Contre Le Cancer, 75013 Paris, France; Instituto de Pesquisa Pelé Pequeno Principe and Faculdades Pequeno Principe (R.R., H.K., B.C.F.), Curitiba PR 80250-060, Brazil; LIA Neogenex Centre National de la Recherche Scientifique (CNRS) (R.R., H.K., M.D., B.C.F., E.La.), Institut de Pharmacologie Moléculaire et Cellulaire CNRS Unité Mixte de Recherche 7275 (M.D., E.La.), and Université de Nice-Sophia Antipolis (M.D., E.La.), 06560 Valbonne, France; Department of Pediatric Endocrinology, Diabetology, and Metabolism (C.F.), University Children’s Hospital Bern, University of Bern, 3010 Bern, Switzerland; Departments of Pathology (R.R.d.K.) and Pediatric Oncology-Hematology (M.M.v.N.), Erasmus MC-University Medical Center, 3015 GJ Rotterdam, The Netherlands; Hôpital de l’Archet (J .- C.M.), Centre Hospitalier Universitaire, 06200 Nice, France; Center for Molecular Genetics and Cancer Research in Children (M.A.D.P.), Universidade Federal do Paraná, Curitiba PR 80030-130, Brazil; and Department of Biochemistry (G.P.Z.), St. Jude Children’s Research Hospital, Memphis, Tennessee 38105
Context: Childhood adrenocortical tumors (ACT) are rare malignancies, except in southern Brazil, where a higher incidence rate is associated to a high frequency of the founder R337H TP53 mu- tation. To date, copy number alterations in these tumors have only been analyzed by low-reso- lution comparative genomic hybridization.
Objective: We analyzed an international series of 25 childhood ACT using high-resolution single nucleotide polymorphism arrays to: 1) detect focal copy number alterations highlighting candidate driver genes; and 2) compare genetic alterations between Brazilian patients carrying the R337H TP53 mutation and non-Brazilian patients.
Results: We identified 16 significantly recurrent chromosomal alterations (q-value < 0.05), the most frequent being -4q34, +9q33-q34, +19p, loss of heterozygosity (LOH) of chromosome 17 and 11p15. Focal amplifications and homozygous deletions comprising well-known oncogenes (MYC, MDM2, PDGFRA, KIT, MCL1, BCL2L1) and tumor suppressors (TP53, RB1, RPH3AL) were identified. In addition, eight focal deletions were detected at 4q34, defining a sharp peak region around the noncoding RNA LINC00290 gene. Although non-Brazilian tumors with a mutated TP53 were similar to Brazilian tumors, those with a wild-type TP53 displayed distinct genomic profiles, with signif- icantly fewer rearrangements (P = 0.019). In particular, three alterations (LOH of chromosome 17, +9q33-q34, and -4q34) were significantly more frequent in TP53-mutated samples. Finally, two of four TP53 wild-type tumors displayed as sole rearrangement a copy-neutral LOH of the im- printed region at 11p15, supporting a major role for this region in ACT development.
Conclusions: Our findings highlight potential driver genes and cellular pathways implicated in childhood ACT and demonstrate the existence of different oncogenic routes in this pathology. (J Clin Endocrinol Metab 97: E1284-E1293, 2012)
First Published Online April 26, 2012
Abbreviations: ACT, Adrenocortical tumor; CGH, comparative genomic hybridization; GISTIC, genomic identification of significant targets in cancer; LOH, loss of heterozygosity; SF-1, steroidogenic factor-1; SNP, single nucleotide polymorphism.
E1284
T he origin of adrenocortical tumors (ACT) in children is believed to consist in a defect of the apoptotic pro- cess of the fetal adrenal, which normally takes place during the first months of life (1). The median age of ACT patients at diagnosis is around 3 yr, with a predominance of girls (2). Signs of virilization are present at the time of diagnosis in more than 90% of patients and may be associated with other signs of endocrine dysfunction, especially Cushing syndrome. Favorable prognostic factors in childhood ACT are stage I at diagnosis, tumor weight no greater than 200 g, age less than 4 yr, and the presence of virilization alone (2). mRNA and microRNA expression profiling studies have identified molecular markers of malignancy in childhood ACT and the IGF-mammalian target of rapa- mycin signaling pathway as a major potential therapeutic target (3, 4). These neoplasms are exceptionally prevalent in southern Brazil, where they are nearly invariably linked to a specific germline, low-penetrant TP53 mutation (R337H) (5). Although the existence of a founder effect for this mutation was initially excluded (5), subsequent stud- ies based on polymorphic markers (6) and single nucleo- tide polymorphisms (SNP) (7) within the TP53 gene con- cluded that all mutant R337H alleles in southern Brazil derived from a single founder. Additionally, this mutation is found associated with ACT in adults (8), choroid plexus carcinoma in children (9), breast cancer (10, 11), and Li- Fraumeni/Li-Fraumeni-like syndromes (11). Remarkably, TP53 mutations are also frequent in childhood ACT cases diagnosed in other geographical areas (12), whereas they are relatively rare in adult tumors (8).
In this study, we analyzed the genomic alterations pres- ent in 25 childhood ACT (13 from Brazil and 12 from other geographical areas) using high-resolution SNP ar- rays. We identified peak regions of gains and losses com- prising well-known oncogenes and tumor suppressors as well as a new tumor suppressor candidate, LINC00290. Tumors harboring TP53 mutations, both Brazilian and non-Brazilian, displayed a significantly higher number of genomic rearrangements than tumors with wild-type TP53. Remarkably, loss of heterozygosity (LOH) events at 11p15 were very common, and in two cases of TP53 wild-type tumors, they were the only genomic alteration detected. Our findings highlight potential driver genes and cellular pathways implicated in the development of child- hood ACT and demonstrate the existence of diverse on- cogenic pathways in this pathology.
Patients and Methods
Patients
Clinical data of the 25 ACT patients included in this study are shown in Table 1. In all cases, one of the parents or legal repre-
sentatives signed an informed consent form approved by the local ethics committee. Samples from The Netherlands were anonymously used according to the code for adequate secondary use of tissue, code of conduct: “Proper Secondary Use of Human Tissue” established by the Dutch Federation of Medical Scien- tific Societies (http://www.federa.org). DNA extraction from tu- mor and peripheral blood DNA was performed as described in Ref. 13.
SNP arrays
Tumor samples were analyzed with Illumina Human CNV370- Duo v1.0 or Illumina Human 610-Quad v1.0 BeadChips (Illumina, Inc., San Diego, CA). Blood samples from six patients were ana- lyzed with Illumina Human CNV370-Quad v3.0 chips. Hybrid- ization was performed by IntegraGen (Evry, France), according to the instructions provided by the array manufacturer. Copy num- ber alterations and LOH were identified using the genome al- teration print method (14), and the genomic identification of significant targets in cancer (GISTIC) algorithm (15) was used to detect significantly recurrent copy number changes across the data set. A detailed description of the methods used to analyze SNP array data is provided in the Supplemental Data (published on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org).
Results
Twenty-five patients with childhood ACT were studied, including 13 patients from a Brazilian cohort (13) and 12 non-Brazilian patients coming from four different hospi- tals in France, Switzerland, The Netherlands, and the United States. Most subjects presented with virilization, and nine had symptoms consistent with Cushing syn- drome. TP53 mutations were screened in all patients. All Brazilian patients carried the germline R337H TP53 mu- tation. Among non-Brazilian tumors, seven of 12 (58%) harbored different mutations of this gene. Table 1 sum- marizes the clinical profiles of all patients, including sex, age, tumor stage and histological type, TP53 mutation, treatment regimen, and follow-up when available. No pa- tient presented clinical evidence for genetically inherited syndromes, like Li-Fraumeni syndrome or Beckwith-Wie- demann. The 25 tumors and six matched blood DNA sam- ples were hybridized to high-resolution Illumina SNP ar- rays, and genome-wide chromosome aberrations were detected using the genome alteration print method (14).
Haplotype analysis confirms the existence of a founder effect for the R337H TP53 mutation
Previous studies (5-7) have reached divergent conclu- sions about the hypothesis of a founder effect for the R337H TP53 mutation in southern Brazil. To test this hypothesis, we analyzed the genotypes of SNPs in and around TP53 in Brazilian tumors. Because the wild-type allele of TP53 has been lost by LOH in these tumors, the
| Sample ID | Origin | Gender | Age (months) | Histological type | TNM | Stage | TP53 mutation | Virilization | Cushing syndrome | Treatment | Follow-up (months) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| HS_01ª,b | Brazil | F | 20 | Ad | T2N0M0 | II | R337H | Y | Y | S | Alive (45) |
| HS_02b | Brazil | M | 21 | Ca | T1N0M0 | I | R337H | Y | N | S | Alive (90) |
| HS_03ª | Brazil | F | 39 | Ca | T3N0M0 | III | R337H | Y | N | SC | DAI (11) |
| HS_04ª,b | Brazil | M | 22 | Ca | T3N1M0 | III | R337H | Y | N | SC | DFP (13) |
| HS_05 | Brazil | F | 52 | Ca | T2N0M0 | II | R337H | Y | Y | SC | DAI (11) |
| HS_06b | Brazil | M | 110 | Ca | T3N0M0 | III | R337H | Y | Y | SC | DFP (21) |
| HS_07b | Brazil | F | 9 | Ad | T2N0M0 | II | R337H | Y | Y | S | Alive (92) |
| HS_08 | Brazil | F | 59 | Ca | T2N0M0 | II | R337H | Y | N | SC | DFP (25) |
| HS_09ª | Brazil | F | 11 | Ad | T1N0M0 | I | R337H | Y | N | S | Alive (64) |
| HS_10 | Brazil | F | 26 | Ca | T2N0M0 | II | R337H | Y | N | S | Alive (102) |
| HS_11ª | Brazil | F | 1 | Ad | T1N0M0 | I | R337H | N | N | S | Alive (49) |
| HS_13b | Brazil | F | 9 | Ca | T2N0M0 | II | R337H | Y | Y | S | Alive (29) |
| HS_14ª,b | Brazil | F | 43 | Ca | T3N0M0 | III | R337H | Y | Y | S | Alive (45) |
| HS_15 | France | M | 61 | Ca | T2N0M0 | II | wt | Y | N | S | Alive (9) |
| HS_46 | Switzerland | M | 8 | Ca | T2N0Mx | ND | R273L | ND | ND | S | DFP (5) |
| HS_67 | The Netherlands | F | 18 | Ad | T1N0M0 | I | Q331X | Y | N | S | Alive (26) |
| HS_68 | The Netherlands | M | 38 | Ca | T1N0M0 | I | E343X | Y | N | S | Alive (68) |
| HS_70 | The Netherlands | F | 82 | Ad | T1N0M0 | I | wt | N | N | S | Alive (38) |
| HS_71 | The Netherlands | F | 120 | Ca | T3N0M1 | IV | R342X | Y | Y | SC | DFP (38) |
| HS_72 | United States | M | 17 | Ca | T3NxM0 | III | E285V | N | N | SC | Alive (3) |
| HS_73 | United States | F | 47 | Ca | T1N0M0 | I | wt | Y | N | S | Alive (0) |
| HS_74 | United States | M | 41 | Ca | T2N0M0 | II | wt | Y | Y | SC | Alive (2) |
| HS_75 | United States | M | 13 | Ca | T3N1M0 | III | R196X | Y | N | S | Alive (18) |
| HS_76 | United States | M | 157 | Ca | T3N1M1 | IV | hdel | N | Y | SC | Alive (15) |
| HS_77 | United States | F | 34 | ND | T1N0M0 | I | IVS10-2A>G | Y | Y | S | Alive (21) |
F, Female; M, male; Ad, adenoma; Ca, carcinoma; Nx, node status not determined; Mx, in this case a second tumor was diagnosed in the brain, whose nature (primary or metastasis from ACC) is unknown; wt, wild-type; hdel, homozygous deletion; Y, yes; N, no; S, surgical resection only; SC, surgical resection plus chemotherapy; DAI, died from adrenal insufficiency; DFP, died from disease progression; ND, not determined.
a Samples for which a blood sample was also analyzed: HS_01 (normal blood sample HS_039), HS_03 (HS_040), HS_04 (HS_041), HS_09 (HS_043), HS_11 (HS_045), and HS_14 (HS_23).
b Samples for which gene expression data were published by West et al. (3): HS_01 (West sample ID ACA4), HS_02 (ACC11), HS_04 (ACC3), HS_06 (ACC8), HS_07 (ACA2), HS_13 (ACC10), and HS_14 (ACA5).
genotype of a SNP in a tumor directly indicates the geno- type of this SNP in the mutant allele of the patient. A blood sample was also available for six of 13 patients. We could thus infer the genotypes of SNPs in the wild-type alleles of these patients by comparing the peripheral blood and tu- mor genotypes (see Patients and Methods). This analysis revealed a conserved haplotype spanning 522 kb around TP53 that was common to the 13 mutant alleles (Supple- mental Table 1). By contrast, the diversity of haplotypes encountered in wild-type alleles for the same region showed that this wide conserved haplotype could not be the result of a regional pattern of linkage disequilibrium. Instead, this segment represents the chromosome frag- ment carrying the R337H TP53 mutation that was inher- ited by all Brazilian patients from a common founder and that was not disrupted by crossovers along the lineage.
Recurrent chromosome rearrangements in childhood ACT
No genomic aberration (constitutive gain, deletion, or LOH) was encountered in the germline DNA of the six patients with matched blood samples, apart from the pres- ence of copy number variants referenced in the Database of Genomic Variants (http://projects.tcag.ca/variation/). By contrast, most tumors exhibited complex copy number
profiles, with numerous rearrangements. The fraction of altered arms, i.e. the proportion of chromosome arms with an aberrant copy number on more than 40% of their length, was above 0.5 in 19 of 25 samples (76%), with a median value of 0.7. In addition, 16 of 25 tumors (74%) were hyperdiploid (10 triploid and six tetraploid samples) with a large number of LOH (Supplemental Fig. 1, A and B), which may be explained by an accumulation of dele- tions followed by cell fusion or endoreplication. Using the GISTIC algorithm (15), we identified 16 copy number al- terations significantly recurrent after correction for mul- tiple testing (q-value < 0.05) (Table 2), including eight gains (at chromosome 1, 5q, 7p, 8q, 9q, 13q, 19, and 21q) and six deletions (at 3q, 4q, 11q, 17, 18q, and 22q). The most striking events were the highly significant gain at 9q33-q34 (q = 9.76e-11) and the sharp peak region of loss at 4q34 (q = 5.91e-10) (Fig. 1). In addition, two regions of LOH were extremely frequent. First, LOH of chromo- some 17 was detected in 22 of 25 tumors (12 LOH with deletion, 10 copy-neutral LOH). Notably, all tumors from TP53-mutation carriers harbored an LOH of chromo- some 17, such that only the mutant copy of this gene re- mained in the tumor cells. Second, a recurrent LOH was found on chromosome arm 11p in 21 of 25 tumors (12
| Index | Type | Broad/focal | Cytoband | q-value | Frequency (%) | Peak region start (bp) | Peak region end (bp) |
|---|---|---|---|---|---|---|---|
| 1 | Gain | Broad | 1p13-1q23 | 0,00163 | 48 | 114,900,679 | 156,824,800 |
| 2 | Gain | Broad | 5q31-5q35 | 1,76E-04 | 60 | 131,322,495 | 180,915,260 |
| 3 | Gain | Broad | 7p21 | 0,043 | 44 | 9,474,206 | 15,209,678 |
| 4 | Gain | Focal | 8q24 | 0,012 | 36 | 122,047,021 | 129,567,180 |
| 5 | Gain | Broad | 9q33-9q34 | 9,76E-11 | 64 | 126,303,372 | 140,964,757 |
| 6 | Gain | Focal | 13q33-13q34 | 0,00382 | 56 | 113,324,325 | 115,169,878 |
| 7 | Gain | Broad | 19p13-19q13 | 2,74E-06 | 60 | 1,572,370 | 32,952,087 |
| 8 | Gain | Focal | 21q11-21q21 | 0,039 | 36 | 15,350,027 | 17,685,245 |
| 9 | Loss | Broad | 3q26 | 0,00327 | 56 | 163,096,849 | 165,778,329 |
| 10 | Loss | Both | 4q34 | 5,91E-10 | 68 | 181,941,643 | 183,026,789 |
| 11 | Loss | Broad | 11q14-11q23 | 0,00413 | 52 | 83,634,984 | 112,218,096 |
| 12 | Loss | Broad | 17p11-17q12 | 0,00173 | 56 | 19,575,722 | 37,357,033 |
| 13 | Loss | Broad | 18q21-18q22 | 0,0156 | 52 | 46,558,290 | 70,493,686 |
| 14 | Loss | Broad | 22q13 | 0,0272 | 48 | 38,241,433 | 48,385,223 |
| 15 | LOH | Both | 11p15 | ND | 84 | 1 | 3,568,044 |
| 16 | LOH | Broad | 17 | ND | 98 | 1 | 81,195,210 |
ND, Not determined. The bp locations refer to the human genome map GRCh37. GISTIC was only used to identify recurrent copy number alterations; hence, no q-value was calculated for LOH.
LOH with deletion, nine copy-neutral LOH). Focal LOH in samples HS_07 and HS_15 (Supplemental Fig. 2) al- lowed the identification of a minimal 3.6-Mb LOH region at 11p15 comprising 103 genes (Supplemental Table 2). Among these, the imprinted genes IGF2, KCNQ1, and CDKN1C were, respectively, the first, third, and fifth most significantly deregulated genes in the seven tumors of our data set, whose gene expression profiles were previ- ously studied by West et al. (3) (Supplemental Table 3).
Focal copy number alterations reveal genes potentially involved in adrenocortical tumorigenesis
High-level amplifications and homozygous deletions are particularly useful for mapping the key genes driving
1e-10
Gains
1e-8
1e-6
1e-4
q-value
1e-2
0
1e-2
0.05
1e-4
1e-6
1e-8
Deletions
1e-10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Genomic position
carcinogenesis. We identified 43 amplicons (Supplemental Table 4) and 13 homozygous deletions (Table 3) in the data set. Amplicons contained six oncogenes known to be amplified in cancers (16) (Cancer Gene Census: http:// www.sanger.ac.uk/genetics/CGP/Census): MYC (ampli- fied in two tumors); MDM2, PDGFRA, and KIT (coam- plified); MCL1 and BCL2L1. Interestingly, these two latter genes are implicated in Bcl2-mediated apoptosis (16), suggesting an important role for this pathway in childhood ACT. Homozygous deletions comprised two known tumor suppressors (RB1 and TP53), as well as the RPH3AL gene, already shown to represent a potential tumor suppressor (17). Strikingly, three homozygous de- letions were located at 4q34, two of which encompassed the noncoding gene LINC00290. Focal hemizygous dele- tions were also frequent in this area in other samples, defining a sharp peak region of loss comprising only LINC00290, which is then a prime tu- mor suppressor candidate for this re- 0.05 gion (Fig. 2).
Effect of chromosomal alterations on gene expression
To evaluate whether copy number alterations have a significant impact on gene expression in childhood ACT, we analyzed the expression profiles of seven tumors of our data set included in a previous gene expression study (3) us- ing the analysis of CNAs by expression data (ACE) strategy (18). In brief, this algorithm computes the differential ex-
| TABLE 3. Homozygous deletions in the series of 25 childhood ACT | |||||||
|---|---|---|---|---|---|---|---|
| Sample | Chromosome | Start SNP | End SNP | Start position (bp) | End position (bp) | Cytoband | Genes |
| HS_03 | 4 | rs7659915 | rs7672826 | 181941643 | 182407405 | 4q34.3 | LINC00290 |
| HS_09 | 4 | rs12498419 | rs2597125 | 181959925 | 182998269 | 4q34.3 | LINC00290 |
| HS_05 | 4 | rs2138743 | rs6829893 | 182800851 | 183661672 | 4q34.3 | MGC45800, ODZ3, MIR1305 |
| HS_08 | 5 | rs1482357 | rs4293904 | 113149125 | 113179089 | 5q22.3 | [YTHDC2] |
| HS_06 | 7 | rs757657 | rs7810597 | 49542314 | 49592382 | 7p12.2 | [VWC2] |
| HS_076 | 12 | rs17034107 | rs1593806 | 103959185 | 103999400 | 12q23.3 | STAB2 |
| HS_076 | 13 | rs1045668 | rs7324752 | 48902122 | 49089107 | 13q14.2 | LPAR6, RB1, RCBTB2 |
| HS_076 | 17 | cnvi0019153 | rs8078273 | 1 | 398106 | 17p13.3 | DOC2B, RPH3AL, FAM101B, C17orf97 |
| HS_076 | 17 | rs11552708 | rs8077824 | 7459063 | 7672047 | 17p13.1 | EFNB3, TP53, ATP1B2, WRAP53, TNFSF13, FXR2, SNORA48, MPDU1, SENP3, SOX15, SAT2, EIF4A1, SHBG, SNORA67, SNORD10, RPL29P2, CD68 |
| HS_076 | 17 | rs7219148 | rs4791334 | 8899138 | 9026076 | 17p13.1 | NTN1 |
| HS_075 | 17 | rs9913683 | rs7225084 | 52640842 | 52687499 | 17q22 | [TOM1L1] |
| HS_06 | 21 | rs1125036 | rs2256797 | 34999184 | 35161946 | 21q22.11 | ITSN1, CRYZL1 |
| HS_06 | 22 | rs6005907 | rs 10854810 | 29213012 | 29390048 | 22q12.1 | ZNRF3 |
pression between a set of tumors and normal reference samples along the genome. This signal is then smoothed by genomic region to produce a “virtual comparative genomic hybridization (CGH)” profile. In our case, this profile is remarkably similar to the real copy number pro- file (Supplemental Fig. 3). Regions frequently lost in child- hood ACT (e.g. chromosomes 3, 4, or 11) display de- creased regional expression in tumors, whereas regions with recurrent gains (e.g. chromosome 5 or 9q) display increased regional expression in tumors with respect to normal samples. Besides, a genome-wide comparison of gene expression in these seven tumors shows that genes are significantly down-regulated in samples in which they are deleted (P < 2.2e-16; Wilcoxon rank sum test) and sig- nificantly up-regulated in samples in which they are gained (P = 6.38e-9). Altogether, these results indicate that gene expression is significantly affected by copy number alter- ations in childhood ACT.
Association with clinical parameters
We also assessed whether the occurrence of the 16 re- current chromosome rearrangements correlated with var- ious clinical parameters. No aberration was significantly associated to metastatic progression or survival. Besides, the overall genomic instability, evaluated as the fraction of aberrant chromosome arms, was not significantly higher in the tumors of patients having undergone metastasis or died from disease progression. No significant association with tumor stage (TNM.T), lymph node invasion (TNM.N) or metastasis (TNM.N) was encountered, but three aberrations showed a trend toward significance
when comparing adenomas with carcinomas: +21q11-21 (none of six adenomas, eight of 18 carcinomas; P = 0.066), -11q14-23 (five of six adenomas, six of 18 car- cinomas; P = 0.061), and -17p11-q12 (five of six ade- nomas, six of 18 carcinomas; P = 0.061). Notably, three rearrangements were significantly more frequent in sam- ples harboring a mutation of TP53: +9q33-q34 (P = 0.0047), -4q34 (P = 0.023), and LOH of chromosome 17 (P= 0.0043). Similarly, the gain at 9q33-q34 and focal deletions at 4q34 were significantly more frequent in Bra- zilian than in non-Brazilian tumors (P = 0.015 and 0.030, respectively), as well as the LOH at 11p15 (P = 0.039). We then compared the genomic profiles of Brazilian tumors with those of non-Brazilian tumors with or without a mu- tation or homozygous deletion of TP53. Although the ge- nome-wide frequencies of gains and losses in non-Brazil- ian tumors with altered TP53 were similar to those of Brazilian tumors, TP53 wild-type tumors showed drasti- cally different profiles (Fig. 3), with significantly fewer alterations (P = 0.019; Wilcoxon test). Notably, two of these four tumors displayed as sole event a copy-neutral LOH at 11p15 (Supplemental Fig. 1C).
Discussion
This study is the first to describe the genomic alterations present in childhood ACT using high-density SNP arrays. Even if the number of cases investigated in our study is limited, due to the rarity of the disease, our results allowed for the identification of highly significant alterations by
GISTIC deletion q-value
1e-9
1e-6
1e-3
HS_067 HS_14 HS_10 HS_09 HS_05 HS_04 HS_03 HS_01
4
1 43210
0
3
2
1
0
3
2
1
0
3
2
1
0
3
2
1
0
4
3
2
1
0
3
2
1
0
3
2
1
0
LINC00290
MIR1305
DCTD H
H
1
I
H
I
MGC45800
ODZ3 FAM92A3
181
182
183
184
Position on chromosome 4 (in Mb)
the GISTIC method (15). In particular, the identification of focal copy number alterations harboring known onco- genes and tumor suppressor genes suggests that oncogenic pathways activated in other types of cancers (16), but whose role was previously unsuspected in childhood ACT, are relevant for its pathogenesis.
Potential driver genes identified in this study are mostly implicated in two hallmarks of cancer: sustaining prolif- erative signaling (IGF2, CDKN1C, PDGFRA, KIT, MYC, RB1), and resisting cell death (TP53, MDM2,
MCL1, BCL2L1) (19). The imprinted gene IGF2 (for which only the paternal allele is ex- pressed) lies in the minimal region of LOH that we identified at 11p15. IGF2 is highly overex- pressed in childhood ACT, as compared with the normal adrenal gland (1, 3), and has a ma- jor role in regulating growth of fetal and tumor adrenocortical cells. Simple gains of 11p15 without LOH are never observed in our series, suggesting that the deregulation of other im- printed genes of the region may also provide a growth advantage. An important role may be thus attributable to the reduced expression of the Cdk inhibitor p57Kip2, encoded by the CDKN1C gene (for which only the maternal allele is expressed). CDKN1C mutations are indeed associated with the Beckwith-Wiede- mann syndrome (20), and mice lacking p57Kip2 display adrenocortical hyperplasia together with other characteristics of the Beck- with-Wiedemann syndrome (21). In addition to the deregulation of the imprinted region at 11p15, we report, for the first time in child- hood ACT, high-level amplifications of PDG- FRA, KIT, and MYC, as well as a homozygous deletion of RB1. These events participate in the activation of cell cycle at various levels of the signaling cascade: PDGFRA and KIT are ty- rosine kinase receptors for growth factors im- plicated in the pathogenesis of several cancer types (22), MYC is a key transcription factor activating cell growth (23), and RB1 is a neg- ative regulator of E2F1, a transcription factor with a crucial role in cell cycle progression (24). Overall, an alteration of at least one driver gene implicated in sustained proliferative signaling was encountered in 22 of 25 tumors of our series.
The regulator of apoptosis most frequently altered in childhood ACT is TP53, which is frequently mutated in cases from Brazil and other geographical areas (5, 12). However, most of these mutations confer relatively low penetrance for predisposition to cancer and result in a partial inactivation of the p53 protein. Consistently, we detected an amplification of MDM2 (which encodes an inhibitor of p53) in tumor HS_05, which already harbored a R337H mutation of TP53 with loss of the wild-type counterpart by LOH. Additional events altering the TP53 pathway may thus still be selected in ACT cells in the presence of a TP53 mutation. Amplifications of MCL1 and BCL2L1 were also detected in TP53-mutated sam-
Brazilian tumors
100%
Frequency
50%
0%
50%
100%
1
2
”
+
5
0
”
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Genomic position
Non-Brazilian tumors with altered TP53
100%
Frequency
50%
0%
50%
100%
1
2
”
+
5
0
-
00
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Genomic position
Non-Brazilian tumors with wild-type TP53
100%
Frequency
50%
0%
50%
100%
1
2
”
+
5
0
N
8
®
10
11
12
13
14
15
16
17
18
19
20
21
22
Genomic position
ples. These genes, belonging to the BCL-2 family, have antiapoptotic activities, and their amplifications were al- ready detected in other cancers (16). Interestingly, the BCL2 gene is located within the peak region of loss at 18q21-22 (Supplemental Table 2), in agreement with an important role of this pathway in adrenocortical tumor- igenesis. Thus, despite the presence of a TP53 mutation in most ACT, alterations of other genes regulating apoptosis may still be required for the development of these tumors.
Apart from driver genes identified in focal high-ampli- tude events, the most striking peak regions identified by GISTIC were +9q33-34 and -4q34. Recurrent gains/am-
plifications of 9q33-34 have been previously described in both childhood (25, 26) and adult ACT (27-29). Despite the use of a high-resolution SNP array, the peak region delineated in our study is quite large (14.7 Mb, 262 genes) because most events affecting this region, including high- level amplifications, were broad aberrations (Supplemen- tal Table 4). A prime candidate driver gene for this region is NR5A1. This gene encodes steroidogenic factor-1 (SF- 1), a nuclear receptor that plays a pivotal role in adreno- gonadal development and function (reviewed in Ref. 30). NR5A1 was previously found to be amplified in eight of nine Brazilian childhood ACT cases studied (31), and the
SF-1 protein was overexpressed in 10 of 10 cases, even if two of them lacked gene amplification (32). Further stud- ies indicated a pivotal role for SF-1 in the control of pro- liferation of human adrenocortical cells and in adrenal cortex tumorigenesis in a transgenic mouse model (33). Despite strong evidence for the implication of NR5A1 in adrenocortical tumorigenesis, the extent of the peak re- gion suggests that other genes lying in a telomeric position with respect to NR5A1 may also play an important patho- genetic role in ACT. Potential candidates in the peak re- gion include ABL1, encoding a protein tyrosine kinase involved in the pathogenesis of chronic myeloid and acute lymphoid leukemias (34); NOTCH1, a member of a fam- ily of transmembrane receptors involved in the pathogen- esis of T-cell leukemias and lymphomas (35); EDF1/ MBF1, which encodes a highly conserved transcriptional cofactor for SF-1 and other nuclear receptors (36); and several other genes with a relationship to cancer (Supple- mental Table 2).
A striking finding of our study was the identification of a deletion at 4q34 present at a very high frequency (11 of 13) in Brazilian ACT. Deletions on chromosome 4 had been identified in previous CGH studies (25, 26), but they usually comprised most of chromosome arm 4q. The higher definition of the SNP array used in this study al- lowed the detection of eight focal deletions, including three homozygous deletions, defining a sharp peak region (1.1 Mb), comprising only the LINC00290 noncoding gene. Interestingly, SNPs in the vicinity of LINC00290 (also known as NCRNA00290) were recently found to predict radiation response in a genome-wide association study performed on lymphoblastoid cell lines, showing a potential relationship with the process of DNA repair (37). However, the eight focal deletions did not all overlap with the peak region determined by GISTIC, and some of them comprised no gene at all, suggesting that cis-regula- tory regions might be affected by these losses. We also cannot rule out a potential role for the protein-coding gene ODZ3 and the micro-RNA MIR1305, both located less than 50 Kb from the peak region, and comprised in one of the homozygous deletions affecting the locus. Further studies are thus required to characterize the role of these candidate genes in adrenocortical tumorigenesis.
Because of the high prevalence of childhood ACT in southern Brazil, most cases studied so far were from Bra- zilian R337H TP53 mutation carriers (3-5, 13, 25, 31, 32). In this study, we also characterized the genomic pro- files of 12 samples from other geographic areas, with and without TP53 mutations. Most childhood ACT displayed numerous chromosome rearrangements, in agreement with previous studies (25, 26), and hyperdiploidy was fre- quent in these tumors. We have shown that a direct cor-
relation exists between the presence of TP53 mutations and the extent of genomic abnormalities in ACT (Fig. 3). Furthermore, three alterations were significantly enriched in TP53 mutated tumors: LOH of chromosome 17, +9q33-q34, and -4q34. LOH of chromosome 17 is ob- viously enriched in TP53-mutated tumors because it im- plies the loss of the wild-type TP53 counterpart in these samples. The increased frequencies of +9q33-q34 and -4q34 may be the result of increased genomic instability in these samples and/or indicate that driver genes in these regions provide a particular growth advantage in tumor cells with altered TP53.
Our study reveals that a subset of genomic alterations is shared by childhood and adult ACT. Chromosome 5, 9q, and 19 gains were also frequently described in adult adrenocortical adenomas and carcinomas using conven- tional or array CGH (27-29, 38). Additionally, LOH at chromosome 17 and 11p15 have also been described in adult tumors, especially in relationship with malignancy (39). On the other hand, chromosome 1p gain and 4q losses, which were seldom detected in adult ACT, are fre- quent in childhood tumors. These data suggest that patho- genetic mechanisms may only partially overlap in child- hood and adult ACT. One probable explanation for these discrepancies can be found in the different occurrence of TP53 mutations in childhood vs. adult ACT (8). Addi- tionally, germline alterations other than TP53 mutations may explain in part the pattern of somatic alterations in childhood ACT. Because no germline chromosome rear- rangement was detected in our study, future analyses using high-throughput sequencing technologies will be of great interest to elucidate the link between somatic chromosome aberrations and predisposing germline mutations.
Our results suggest a clear series of events for the de- velopment of childhood ACT in southern Brazil: 1) inher- itance of the R337H TP53 germline mutation; 2) loss of the wild-type TP53 allele in the tumor by LOH; 3) con- sequent genomic instability due to defective TP53 function (40); 4) occurrence of secondary genomic aberrations, some of which (+9q33-q34, -4q34, LOH of the 11p15 region) are likely to drive the early oncogenic transforma- tion of adrenocortical cells but not malignancy; 5) in some cases, further genomic aberrations may occur, enhancing tumor aggressiveness and establishing malignancy. By contrast, non-Brazilian ACT are heterogeneous and prob- ably arise through distinct oncogenic routes. Non-Brazil- ian tumors with a mutation of TP53 are highly rearranged and display a pattern of gains and losses similar to Bra- zilian tumors. The development of these ACT thus re- quires genomic instability and probably involves driver genes similar to Brazilian ACT. Conversely, non-Brazilian ACT with wild-type TP53 develop in the absence of mas-
sive genomic rearrangements. Strikingly, two of these tu- mors only displayed a copy-neutral LOH of the 11p15 region, suggesting that LOH of this imprinted region can be the sole chromosomal rearrangement required to trig- ger oncogenesis. However, other genetic alterations, un- detectable with SNP arrays, like point mutations or epi- genetic mechanisms may play an important pathogenic role in these tumors. Further characterization of child- hood ACT with high-throughput sequencing techniques will be of great help to gain a complete vision of the genes and pathways involved in this pathology.
Acknowledgments
We thank Jacqueline Godet and Jacqueline Metral for support through the Cartes d’Identité des Tumeurs program.
Address all correspondence and requests for reprints to: Eric Letouzé, Ph.D., Program Cartes d’Identité des Tumeurs, Ligue Nationale Contre Le Cancer, 14 rue Corvisart, 75013 Paris, France. E-mail: LetouzeE@ligue-cancer.net. Or, Enzo Lalli, M.D., Institut de Pharmacologie Moléculaire et Cellulaire, 660 route des Lucioles, 06560 Valbonne, France. E-mail: ninino@ipmc.cnrs.fr.
This work was supported by grants from La Ligue contre le Cancer (Cartes d’Identité des Tumeurs program), Institut Na- tional du Cancer and Centre National de la Recherche Scienti- fique (LIA Neogenex) (to E.La.); FP7 ENS@T-CANCER (to R.R.d.K. and E.La.); State of Paraná (SETI, 2005 and 2008), Conselho Nacional de Desenvolvimento Cientifico e Tec- nológico (2009) and Coordenadoria de Aperfeiçomento de Pes- soal de Ensino Superior (CAPES/Nanobiotechnology Grant 024/ 2009) (to B.C.F.).
Data accession numbers: GEO Series GSE35066.
Disclosure Summary: The authors have nothing to disclose.
References
1. Wilkin F, Gagné N, Paquette J, Oligny LL, Deal C 2000 Pediatric adrenocortical tumors: molecular events leading to insulin-like growth factor II gene overexpression. J Clin Endocrinol Metab 85: 2048-2056
2. Michalkiewicz E, Sandrini R, Figueiredo B, Miranda EC, Caran E, Oliveira-Filho AG, Marques R, Pianovski MA, Lacerda L, Cristo- fani LM, Jenkins J, Rodriguez-Galindo C, Ribeiro RC 2004 Clinical and outcome characteristics of childhood with adrenocortical tu- mors. An analysis of 254 cases from the International Pediatric Ad- renocortical Tumor Registry. J Clin Oncol 22:838-845
3. West AN, Neale GA, Pounds S, Figueredo BC, Rodriguez Galindo C, Pianovski MA, Oliveira Filho AG, Malkin D, Lalli E, Ribeiro R, Zambetti GP 2007 Gene expression profiling of childhood adreno- cortical tumors. Cancer Res 67:600-608
4. Doghman M, El Wakil A, Cardinaud B, Thomas E, Wang J, Zhao W, Peralta-Del Valle MH, Figueiredo BC, Zambetti GP, Lalli E 2010 Regulation of IGF-mTOR signalling by miRNA in childhood adre- nocortical tumors. Cancer Res 70:4666-4675
5. Ribeiro RC, Sandrini F, Figueiredo B, Zambetti GP, Michalkiewicz E, Lafferty AR, DeLacerda L, Rabin M, Cadwell C, Sampaio G, Cat
I, Stratakis CA, Sandrini R 2001 An inherited p53 mutation that contributes in a tissue-specific manner to pediatric adrenal cortical carcinoma. Proc Natl Acad Sci USA 98:9330-9335
6. Pinto EM, Billerbeck AE, Villares MC, Domenice S, Mendonça BB, Latronico AC 2004 Founder effect for the highly prevalent R337H mutation of tumor suppressor p53 in Brazilian patients with adre- nocortical tumors. Arq Bras Endocrinol Metabol 48:647-650
7. Garritano S, Gemignani F, Palmero EI, Olivier M, Martel-Planche G, Le Calvez-Kelm F, Brugiéres L, Vargas FR, Brentani RR, Ashton- Prolla P, Landi S, Tavtigian SV, Hainaut P, Achatz MI 2010 Detailed haplotype analysis at the TP53 locus in p.R337H mutation carriers in the population of Southern Brazil: evidence for a founder effect. Hum Mutat 31:143-150
8. Herrmann LJ, Heinze B, Fassnacht M, Willenberg HS, Quinkler M, Reisch N, Zink M, Allolio B, Hahner S 2012 TP53 germline muta- tions in adult patients with adrenocortical carcinoma. J Clin Endo- crinol Metab 97:E476-E485
9. Custodio G, Taques GR, Figueiredo BC, Gugelmin ES, Oliveira Figueiredo MM, Watanabe F, Pontarolo R, Lalli E, Torres LF 2011 Increased incidence of choroid plexus carcinoma due to the germline TP53 R337H mutation in Southern Brazil. PLoS ONE 6:e18015
10. Assumpção JG, Seidinger AL, Mastellaro MJ, Ribeiro RC, Zambetti GP, Ganti R, Srivastava K, Shurtleff S, Pei D, Zeferino LC, Dufloth RM, Brandalise SR, Yunes JA 2008 Association of the germline TP53 R337H mutation with breast cancer in southern Brazil. BMC Cancer 8:357
11. Achatz MI, Olivier M, Le Calvez F, Martel-Planche G, Lopes A, Rossi BM, Ashton-Prolla P, Giugliani R, Palmero EI, Vargas FR, Da Rocha JC, Vettore AL, Hainaut P 2007 The TP53 mutation, R337H, is associated with Li-Fraumeni and Li-Fraumeni-like syn- dromes in Brazilian families. Cancer Lett 245:96-102
12. Varley JM, McGown G, Thorncroft M, James LA, Margison GP, Forster G, Evans DG, Harris M, Kelsey AM, Birch JM 1999 Are there low-penetrance TP53 alleles? Evidence from childhood adre- nocortical tumors. Am J Hum Genet 65:995-1006
13. Rosati R, Cerrato F, Doghman M, Pianovski MA, Parise GA, Custó- dio G, Zambetti GP, Ribeiro RC, Riccio A, Figueiredo BC, Lalli E 2008 High frequency of loss of heterozygosity at 11p15 and IGF2 overexpression is not associated with clinical outcome in childhood adrenocortical tumors positive for the R337H TP53 mutation. Can- cer Genet Cytogenet 186:19-24
14. Popova T, Manié E, Stoppa-Lyonnet D, Rigaill G, Barillot E, Stern MH 2009 Genome alteration print (GAP): a tool to visualize and mine complex cancer genomic profiles obtained by SNP arrays. Ge- nome Biol 10:R128
15. Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D, Vivanco I, Lee JC, Huang JH, Alexander S, Du J, Kau T, Thomas RK, Shah K, Soto H, Perner S, Prensner J, Debiasi RM, Demichelis F, Hatton C, Rubin MA, Garraway LA, Nelson SF, Liau L, Mischel PS, Cloughesy TF, Meyerson M, Golub TA, Lander ES, Mellinghoff IK, Sellers WR 2007 Assessing the significance of chromosomal aberrations in cancer: methodology and application to glioma. Proc Natl Acad Sci USA 104:20007-20012
16. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Don- ovan J, Barretina J, Boehm JS, Dobson J, Urashima M, Mc Henry KT, Pinchback RM, Ligon AH, Cho YJ, Haery L, Greulich H, Reich M, Winckler W, Lawrence MS, Weir BA, Tanaka KE, Chiang DY, Bass AJ, Loo A, Hoffman C, Prensner J, Liefeld T, Gao Q, Yecies D, Signoretti S, Maher E, Kaye FJ, Sasaki H, Tepper JE, Fletcher JA, Tabernero J, Baselga J, Tsao MS, Demichelis F, Rubin MA, Janne PA, Daly MJ, Nucera C, Levine RL, Ebert BL, Gabriel S, Rustgi AK, Antonescu CR, Ladanyi M, Letai A, Garraway LA, Loda M, Beer DG, True LD, Okamoto A, Pomeroy SL, Singer S, Golub TR, Lander ES, Getz G, Sellers WR, Meyerson M 2010 The landscape of somatic copy-number alteration across human cancers. Nature 463:899- 905
17. Smith JS, Tachibana I, Allen C, Chiappa SA, Lee HK, McIver B, Jenkins RB, Raffel C 1999 Cloning of a human ortholog (RPH3AL)
of (RNO)Rph3al from a candidate 17p13.3 medulloblastoma tu- mor suppressor locus. Genomics 59:97-101
18. Hu G, Chong RA, Yang Q, Wei Y, Blanco MA, Li F, Reiss M, Au JL, Haffty BG, Kang Y 2009 MTDH activation by 8q22 genomic gain promotes chemoresistance and metastasis of poor-prognosis breast cancer. Cancer Cell 15:9-20
19. Hanahan D, Weinberg RA 2011 Hallmarks of cancer: the next gen- eration. Cell 144:646-674
20. Weksberg R, Smith AC, Squire J, Sadowski P 2003 Beckwith-Wie- demann syndrome demonstrates a role for epigenetic control of nor- mal development. Hum Mol Genet 12:R61-R68
21. Zhang P, Liégeois NJ, Wong C, Finegold M, Hou H, Thompson JC, Silverman A, Harper JW, DePinho RA, Elledge SJ 1997 Altered cell differentiation and proliferation in mice lacking p57(KIP2) indicates a role in Beckwith-Wiedemann syndrome. Nature 387:151-158
22. Fletcher JA 2004 Role of KIT and platelet-derived growth factor receptors as oncoproteins. Semin Oncol 31(2 Suppl 6):4-11
23. Adhikary S, Eilers M 2005 Transcriptional regulation and transfor- mation by Myc proteins. Nat Rev Mol Cell Biol 6:635-645
24. Yamasaki L 2003 Role of the RB tumor suppressor in cancer. Cancer Treat Res 115:209-239
25. Figueiredo BC, Stratakis CA, Sandrini R, DeLacerda L, Pianovsky MA, Giatzakis C, Young HM, Haddad BR 1999 Comparative genomic hybridization analysis of adrenocortical tumors of child- hood. J Clin Endocrinol Metab 84:1116-1121
26. James LA, Kelsey AM, Birch JM, Varley JM 1999 Highly consistent genetic alterations in childhood adrenocortical tumors detected by comparative genomic hybridization. Br J Cancer 81:300-304
27. Dohna M, Reincke M, Mincheva A, Allolio B, Solinas-Toldo S, Lichter P 2000 Adrenocortical carcinoma is characterized by a high frequency of chromosomal gains and high-level amplifications. Genes Chromosomes Cancer 28:145-152
28. Zhao J, Roth J, Bode-Lesniewska B, Pfaltz M, Heitz PU, Kommi- noth P 2002 Combined comparative genomic hybridization and genomic microarray for detection of gene amplifications in pulmo- nary artery intimal sarcomas and adrenocortical tumors. Genes Chromosomes Cancer 34:48-57
29. Barreau O, de Reynies A, Wilmot-Roussel H, Guillaud-Bataille M, Auzan C, René-Corail F, Tissier F, Dousset B, Bertagna X, Bertherat J, Clauser E, Assié G 2012 Clinical and pathophysiological Impli- cations of chromosomal alterations in adrenocortical tumors: an integrated genomic approach. J Clin Endocrinol Metab 97:E301- E311
30. Lalli E 2010 Adrenocortical development and cancer: focus on SF-1. J Mol Endocrinol 44:301-307
31. Figueiredo BC, Cavalli LR, Pianovski MA, Lalli E, Sandrini R, Ri- beiro RC, Zambetti G, DeLacerda L, Rodrigues GA, Haddad BR 2005 Amplification of the steroidogenic factor 1 gene in childhood adrenocortical tumors. J Clin Endocrinol Metab 90:615-619
32. Pianovski MA, Cavalli LR, Figueiredo BC, Santos SC, Doghman M, Ribeiro RC, Oliveira AG, Michalkiewicz E, Rodrigues GA, Zam- betti G, Haddad BR, Lalli E 2006 SF-1 overexpression in childhood adrenocortical tumors. Eur J Cancer 42:1040-1043
33. Doghman M, Karpova T, Rodrigues GA, Arhatte M, De Moura J, Cavalli LR, Virolle V, Barbry P, Zambetti GP, Figueiredo BC, Heck- ert LL, Lalli E 2007 Increased steroidogenic factor-1 dosage triggers adrenocortical cell proliferation and cancer. Mol Endocrinol 21: 2968-2987
34. Gotoh A, Broxmeyer HE 1997 The function of BCR/ABL and re- lated proto-oncogenes. Curr Opin Hematol 4:3-11
35. Palomero T, Ferrando A 2008 Oncogenic NOTCH1 control of MYC and PI3K: challenges and opportunities for anti-NOTCH1 therapy in T-cell acute lymphoblastic leukemias and lymphomas. Clin Cancer Res 14:5314-5317
36. Kabe Y, Goto M, Shima D, Imai T, Wada T, Morohashi K, Shi- rakawa M, Hirose S, Handa H 1999 The role of human MBF1 as a transcriptional coactivator. J Biol Chem 274:34196-34202
37. Niu N, Qin Y, Fridley BL, Hou J, Kalari KR, Zhu M, Wu TY, Jenkins GD, Batzler A, Wang L 2010 Radiation pharmacogenomics: a genome-wide association approach to identify radiation response biomarkers using human lymphoblastoid cell lines. Genome Res 20:1482-1492
38. Stephan EA, Chung TH, Grant CS, Kim S, Von Hoff DD, Trent JM, Demeure MJ 2008 Adrenocortical carcinoma survival rates corre- lated to genomic copy number variants. Mol Cancer Ther 7:425- 431
39. Gicquel C, Bertagna X, Gaston V, Coste J, Louvel A, Baudin E, Bertherat J, Chapuis Y, Duclos JM, Schlumberger M, Plouin PF, Luton JP, Le Bouc Y 2001 Molecular markers and long-term recur- rences in a large cohort of patients with sporadic adrenocortical tumors. Cancer Res 61:6762-6767
40. Shlien A, Tabori U, Marshall CR, Pienkowska M, Feuk L, Novok- met A, Nanda S, Druker H, Scherer SW, Malkin D 2008 Excessive genomic DNA copy number variation in the Li-Fraumeni cancer predisposition syndrome. Proc Natl Acad Sci USA 105:11264- 11269