ENDOCRINE SOCIETY
DNA methylation is an independent prognostic marker of survival in adrenocortical cancer
Anne JOUINOT, Guillaume ASSIE, Rossella LIBE, Martin FASSNACHT, Thomas PAPATHOMAS, Olivia BARREAU, Bruno DE LA VILLEON, Simon FAILLOT, Nadim HAMZAOUI, Mario NEOU, Karine PERLEMOINE, Fernande RENE-CORAIL, Stéphanie RODRIGUEZ, Mathilde SIBONY, Frédérique TISSIER, Bertrand DOUSSET, Silviu SBIERA, Cristina RONCHI, Matthias KROISS, Esther KORPERSHOEK, Ronald DE KRIJGER, Jens WALDMANN, Detlef K BARTSCH, Marcus QUINKLER, Magalie HAISSAGUERRE, Antoine TABARIN, Olivier CHABRE, Nathalie STURM, Michaela LUCONI, Franco MANTERO, Massimo MANNELLI, Regis COHEN, Véronique KERLAN, Philippe TOURAINE, Gaelle BARRANDE, Lionel GROUSSIN, Xavier BERTAGNA, Eric BAUDIN, Laurence AMAR, Felix BEUSCHLEIN, Eric CLAUSER, Joel COSTE, Jérôme BERTHERAT
The Journal of Clinical Endocrinology & Metabolism Endocrine Society
Submitted: September 11, 2016 Accepted: December 08, 2016 First Online: December 14, 2016
Early Release articles are PDF versions of manuscripts that have been peer reviewed and accepted but not yet copyedited. The manuscripts are published online as soon as possible after acceptance and before the copyedited, typeset articles are published. They are posted “as is” (i.e., as submitted by the authors at the modification stage), and do not reflect editorial changes. No corrections/changes to the PDF manuscripts are accepted. Accordingly, there likely will be differences between the Early Release manuscripts and the final, typeset articles. The manuscripts remain listed on the Early Release page until the final, typeset articles are posted. At that point, the manuscripts are removed from the Early Release page.
DISCLAIMER: These manuscripts are provided “as is” without warranty of any kind, either express or particular purpose, or non-infringement. Changes will be made to these manuscripts before publication. Review and/or use or reliance on these materials is at the discretion and risk of the reader/user. In no event shall the Endocrine Society be liable for damages of any kind arising references to, products or publications do not imply endorsement of that product or publication.
THE JOURNAL OF CLINICAL ENDOCRINOLOGY & METABOLISM
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
DNA methylation is an independent prognostic marker of survival in adrenocortical cancer
Anne JOUINOT*1,2, Guillaume ASSIE*1,3, Rossella LIBE4,5, Thomas PAPATHOMAS’, Olivia BARREAU1,3, Bruno DE LA VILLEON1, Simon FAILLOT1, Nadim HAMZAOUI7, Mario NEOU1, Karine PERLEMOINE1, Fernande RENE-CORAIL1, Stéphanie RODRIGUEZ , Mathilde SIBONY1,8, Frédérique TISSIER1,8, Bertrand DOUSSET9, Silviu SBIERA4, Cristina RONCHI4, Matthias KROISS5, Esther KORPERSHOEK6, Ronald DE KRIJGER6,10, Jens WALDMANN11, Detlef K BARTSCH11, Marcus QUINKLER12, Magalie HAISSAGUERRE13, Antoine TABARIN13, Olivier CHABRE14, Nathalie STURM15, Michaela LUCONI16, Franco MANTERO17, Massimo MANNELLI16, Regis COHEN18, Véronique KERLAN19, Philippe TOURAINE20, Gaelle BARRANDE21, Lionel GROUSSIN1,3, Xavier BERTAGNA1,3, Eric BAUDIN22, Laurence AMAR23, Felix BEUSCHLEIN24, Eric CLAUSER7, Joel COSTE25, Jérôme BERTHERAT1,3
1- Institut Cochin, INSERM U1016, CNRS UMR8104, Paris Descartes University, Paris, France
2- Medical Oncology, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France
3- Department of Endocrinology, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France
4- Endocrinology and Diabetes Unit, University Hospital, University of Würzburg, Würzburg, Germany
5- Comprehensive Cancer Center Mainfranken, University of Würzburg, Würzburg, Germany
6- Department of Pathology, Erasmus MC University Medical Center, Rotterdam, the Netherlands
7- Department of Oncogenetics, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France
8- Department of Pathology, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France
9- Department of Digestive and Endocrine Surgery, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France
10- Department of Pathology, Reinier de Graaf Hospital, Delft, The Netherlands
11- Department of Surgery, University Hospital Giessen and Marburg, Campus Marburg, Germany
12- Department of Medicine, Charite University, Berlin, Germany
13- Department of Endocrinology, Diabetes and Metabolic Diseases, University Hospital of Bordeaux, Bordeaux, France
14- Department of Endocrinology, University Hospital of Grenoble, Grenoble, France
15- Department of biology and pathology, University Hospital of Grenoble, Grenoble, France
16- Department of Experimental and Clinical Biomedical Sciences, University of Florence, Florence, Italy
17- Department of Medicine, Endocrinology Unit, University of Padova, Padova, Italy
18- Department of Endocrinology, Saint Denis Hospital, Saint Denis, France
19- Department of Endocrinology, Brest University Hospital, Brest, France
20- Department of Endocrinology, Groupe Hospitalier Pitié-Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, France
21- Department of Endocrinology, Regional Hospital of Orléans, Orléans, France
22- Department of Nuclear Medicine and Endocrine Oncology, Institut Gustave Roussy, Villejuif, France
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
23- Hypertension Unit, Hôpital Européen Georges Pompidou, Assistance Publique-Hôpitaux de Paris, Paris, France
24- Medizinische Klinik und Poliklinik IV, Ludwig-Maximilians-Universität München, München, Germany
25-Biostatistics and Epidemiology Unit, Hôtel Dieu, Assistance Publique-Hôpitaux de Paris, Paris, France
* Anne JOUINOT and Guillaume ASSIE contributed equally to this work
$ Rosella LIBE and Martin FASSNACHT contributed equally to this work
Received 11 September 2016. Accepted 08 December 2016.
A new prognostic marker based on DNA methylation in ACC
Context: Adrenocortical cancer (ACC) is an aggressive tumor with heterogeneous outcome. Prognostic stratification is difficult even based on tumor stage and Ki67 index. Recently integrated genomics studies have demonstrated that CpG islands hypermethylation is correlated with poor survival.
Objective: To confirm the prognostic value of CpG islands methylation on an independent cohort with a single commonly available methylation assay.
Design: CpG islands methylation was measured by methylation-specific-multiplex-ligation- dependent-probe-amplification (MS-MLPA) using the ME002 kit (MRC-Holland).
Setting: MS-MLPA was performed in a training cohort of 50 ACC to identify the best set of probes correlating with disease-free (DFS) and overall survival (OS). These were then validated in an independent cohort from 21 ENSAT centers.
Patients: The validation cohort included 203 ACC: 64% females, median age 50 years, 80% localized tumors.
Intervention: None.
Main Outcome Measures: DFS and OS (Cox models).
Results: In the training cohort, mean methylation in CpG islands of 4 genes (PAX5,GSTP1,PYCARD,PAX6) was the strongest methylation marker. In the validation cohort, methylation was a significant prognostic factor of DFS (p<0.0001) and OS (p<0.0001). Methylation, Ki67 and ENSAT stage were combined in multivariate models. For DFS, methylation (p=0.0005) and ENSAT stage (p<0.0001) but not Ki67 (p=0.19) remained highly significant. For OS, methylation (p=0.0006), ENSAT stage (p<0.0001) and Ki67 (p=0.024) were independent prognostic factors.
Conclusions: Tumor DNA methylation emerges as an independent prognostic factor in ACC. MS-MLPA is readily compatible with clinical routine, and above clinical and pathological features, should enhance our ability for prognostication and precision medicine.
Précis: We studied tumor DNA methylation by MS-MLPA in 203 adrenocortical cancers and found that methylation is a strong prognostic factor, independent of tumor stage and Ki67.
Introduction
Despite a poor overall survival (OS), the prognosis of adrenocortical carcinoma (ACC) is heterogeneous. The 5-year survival rate remains below 40% in most series(1) but variability in clinical presentations and outcome is observed(2). Prognostic stratification is important to discuss and evaluate adjuvant or curative therapies and to individualize patient follow-up. The tumor stage at diagnosis, as defined by the European Network for the Study of Adrenal Tumours (ENSAT), represents the most powerful prognostic factor(1,3). The OS of patients with the same tumor burden is yet highly variable(4). In localized ACC, a recent study has identified Ki67
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
proliferation index as a major prognostic factor to predict recurrence after complete surgical resection(5). However, a recent study reported important interobserver variability for Ki67(6), which is a limitation for clinical routine. In recent years, genomic and epigenomic analyses have drawn a new classification of ACC, based on molecular alterations(7-9). Gene and miRNA expression, somatic mutations, chromosomal alterations and DNA methylation could identify subgroups of ACC with markedly different outcomes(7-9).
DNA methylation might be involved in the pathogenesis of various diseases, including cancer. In comparison with normal cells, tumor cells exhibit a different methylome where hypomethylation and hypermethylation patterns can be observed. Hypermethylation of CpG islands - that are located in the regulatory regions of genes - contributes to the inactivation of tumor suppressor genes(10). In colorectal cancer, a subset of tumors are characterized by hypermethylation of several hundreds of CpG islands, a pattern called CpG island methylator phenotype (CIMP)(11). These tumors are associated with specific molecular and clinico- pathological features(12), including poor clinical outcome(13), and CIMP was thought to represent a distinct pathway of colorectal carcinogenesis. Subsequently, a similar phenotype has been reported in other cancers(14).
Three studies have investigated genome-wide DNA methylation profiles in ACC(15-17) and found a hypermethylation of CpG islands, evoking a CIMP phenotype in a subset of ACC(15- 17). Based on a pangenomic study, our team has recently reported that the levels of methylation at CpG islands are related to survival and that the CpG island hypermethylation / CIMP phenotype is associated with a poor outcome in ACC(17). The poor prognosis of hypermethylation in ACC was recently confirmed by an independent molecular classification from the Cancer Genome Atlas (TCGA) project(9).
Therefore down-grading from global profiling to focus assessments of methylation is a necessary step to use methylation as a molecular prognostic factor in clinical routine. The methylation of targeted genes can be determined by methylation specific multiplex ligation- dependent probe amplification (MS-MLPA), a PCR-based technique using a methylation- sensitive restriction enzyme. Due to its excellent reliability, it is now a gold standard for targeted methylation oncogenetics analysis, such as MLH1 promoter methylation assay in the screening for Lynch syndrome(18). In ACC, we previously reported the feasibility of MS-MLPA to determine the CIMP status(17).
The aim of the current study was to set up a simplified and optimized molecular tool measuring methylation, based on MS-MLPA, in a first cohort of ACC. The prognostic value of this DNA marker was then tested on a large independent cohort and compared to other prognostic factors.
Patients and Methods
Patients
A total of 253 adult ACC patients were included: a training cohort of 50 ACC patients for setting up the MS-MLPA marker -in this cohort, all tumors have been previously analyzed by methylation array-, and an independent validation cohort of 203 patients. A cohort of 15 adrenocortical adenomas (ACA) served as control for MS-MLPA.
The 50 ACC of the training cohort were collected as previously reported(17). The 203 ACC of the validation cohort were collected in the ENSAT network by 21 centers from France, Germany, The Nederlands and Italy(19).
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
Tumor samples were frozen and stored as previously reported(7,20). Diagnosis of malignancy was confirmed by expert pathologists according to the Weiss criteria(21).
For the training cohort, clinical information was collected as previously reported(20). Briefly, cortisol excess was ascertained by clinical and hormonal evaluation, the latter including 24 hours urinary cortisol, 1mg Dexamethasone suppression test, midnight cortisol, and early morning plasma ACTH. Other hormone excesses were based on clinical signs and appropriate blood hormone assays. For the validation cohort, clinical information was obtained from the ENSAT database(22) including age, sex, hormone secretion, Ki67 staining, ENSAT tumor stage, disease- free survival (DFS) and overall survival (OS). All patients were naive of chemotherapy and radiation therapy at the time of surgery or biopsy. Three patients received mitotane before biopsy. For 15 patients, no sample was available from the primary tumor, and a sample from a local recurrence or a metastasis was considered instead; these tumors showed similar molecular profiles and survival features compared to the rest of the cohort, and were therefore not excluded (data not shown). In the absence of synchronous metastases, patients were operated. Metastatic patients were either operated or treated by mitotane and/or cytotoxic chemotherapy.
Patients were followed until the date of their death, their last examination or the end of their follow-up period.
Signed informed consent for the genetic testing on tissue samples, and the collection and use of clinical data was obtained from all participants, and the study was approved by the local institutional review board of each clinical center.
Tumor DNA preparation
Tumor samples (10-50 mg) were powdered under liquid nitrogen and DNA was extracted as previously reported(7).DNA quality assessment was controlled with the Nanodrop ND-1000 spectrophotometer (Nyxor Biotech).
Methylome analysis
Bisulfite-converted genomic DNA of the 50 ACC from the training cohort were analyzed using the Infinium HumanMethylation27 Beadchip (Illumina) as previously described(17). Methylation level for each tumor was calculated as the mean M-value for CpG within CpG islands of all chromosomes, excluding chromosomes X and Y.
Methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA)
MS-MLPA is a PCR-based assay detecting changes in methylation status of up to 50 selected sequences in a single experiment(23).
Tumor DNA of the 253 ACC and 15 ACA were assessed with the MLPA ME002 tumor suppressor-2 probe mix, combined with the MLPA EK1-Cy5 or EK1-FAM reagent kits (MRC- Holland). The probe mix contains 14 reference probes and 27 MS-MLPA probes that detect the methylation status of CpG islands in the promoter regions of 25 different tumor suppressor genes (Supplementary Table S1). This kit was chosen for its enrichment in genes that were hypermethylated in the CIMP tumors from the pangenomic methylome analysis(17). Information regarding the probe sequences and restriction sites can be found at www.mlpa.com.
The kit was used according to the manufacturer’s protocol, starting from 500 ng of genomic DNA. In brief, DNA was denatured (10 min at 98℃) and cooled at 25℃, after which the probe mix was added to the samples and hybridized by incubation at 60℃ for 16 hours. Each sample was divided in two tubes, in which one half was ligated, and the other was ligated and digested using the methylation-sensitive restriction enzyme HhaI (Promega). Both samples were subsequently subjected to a PCR reaction, and fragment analysis was performed on a capillary sequencer (Beckman Ceq 8800, Beckman-Coulter or ABI 3130xl, Applied Biosystems).
JCEM
ENDOCRINE SOCIETY
Electrophoregrams obtained were analyzed using Ceq 8800 Genetic Analysis System (Beckman- Coulter) or Gene Mapper software (Applied Biosystem). The peak heights of each probe were considered for the calculations. Following the manufacturer’s recommendations, an “intra- sample data normalization” was done, by dividing the peak height of each probe by the peak height of every reference probe in the sample, thus creating as many ratios per probe as there were reference probes. We obtained a normalization constant by calculating the median value of all probe ratios per probe. Finally, the percentage of methylation of each probe was calculated by dividing the normalization constant of a probe in the digested sample by the normalization constant of the same probe in the undigested sample, and by multiplying this ratio by 100.
Since each MS-MLPA probe showed a specific dynamic range, and prior to averaging the values of different probes, we performed for each probe a normalization with a linear transformation of the measured percentage of methylation to a value ranging from 0 to 100%. The 0% was defined as the median value measured in the 15 ACA. The 100% value was defined as the median value measured in the 24 ACC previously identified as hypermethylated (CIMP) by methylation array(17). After transformation, any values lower than 0% or upper than 100% were replaced by 0 and 100% respectively.
For each one of the 27 MS-MLPA probes, we tested the relation with the CpG islands methylation status (CIMP or non-CIMP) as defined by methylation array, and with survival. The probes showing significant association with methylation status, DFS and OS were selected for calculating the percentage of methylation for each tumor.
Statistical analysis
Analyses were performed using R statistical software (R Stats Package) and SAS (Statistical Analysis System) software.
Comparison between groups was performed with the Student’s t test for quantitative variables and with the Fischer’s exact test for qualitative variables.
Correlation between quantitative variables was assessed by the Pearson and Spearman coefficients.
OS was defined as time elapsed between surgery of the primary tumor and death or last follow-up visit. DFS was defined as time elapsed between surgery and the first evidence of relapse or last follow-up without evidence for disease.
Survival curves were obtained with Kaplan-Meier estimates and compared with the log-rank test. Cox proportional hazards regression was used to determine univariate and multivariate hazard ratios for selected potential predictors of disease-free survival (DFS) and overall survival (OS).
Univariate survival analyses were performed for quantitative -age, tumor size, Ki67 and methylation- and qualitative -sex, cortisol secretion and ENSAT stage- variables. These variables were either selected for their a priori prognostic value according to previous publications(1,5,24), or for their being part of standard demographic and oncologic features.
In line with the main objective of this study, namely finding out whether or not methylation would add to ACC prognosis information in clinical practice, methylation was tested in multivariate models including primarily ENSAT stage and Ki67, the only well established and routinely used prognostic factors(1,5). Multivariate models were performed in a subset of 146 patients with complete data for the selected variables.
Second order polynomial terms were tested to evaluate nonlinear relationships between methylation and survival. The proportional hazards assumption was tested for each model constructed.
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
Results
Patient characteristics
The main baseline clinical, hormonal and pathological characteristics of 15 ACA patients, the 50 ACC of the training cohort and the 203 ACC patients of the validation cohort are summarized in Table 1. Age, sex, hormonal secretion, Weiss score and ENSAT tumor stage were not different between the groups. The training cohort included more cortisol secreting (66% vs 49%, p=0.039) and larger tumors (mean tumor size: 114.5 vs 90 mm, p=0.008) than the validation cohort. However Ki67 was lower (2.3 vs 15.7, p<10-16) in the training cohort. Despite these few differences, recurrence (62% vs 57%) and death (50% vs 44%) rates were not significantly different between the training and the validation cohorts respectively (Table 1). In the validation cohort, 9 patients (4%) were lost to follow-up. For patients alive at the end of the study, follow- up period ranged from 8.4 to 272.5 months and was higher than two years for 92% of patients. Outcome was not related to center size (data not shown).
Targeted measurements of methylation by MS-MLPA reflect the whole genome CpG methylation status
In the training cohort, the methylation level of 27 probes was assayed by MS-MLPA.
Four probes (GSTP1, PYCARD, PAX6, PAX5) were positively correlated with the CpG islands methylation status (CIMP or non-CIMP) as defined by methylation array(17) (Pearson correlation, p<0.05), and associated with DFS and OS (cox regression, p<0.05; Supplementary Table S2). Subsequently, for each tumor, a global CpG islands methylation level was calculated by averaging the percentage of methylation of these four MS-MLPA probes. This global CpG islands methylation level correlated well with the CpG islands methylation measured by the methylome beadchip (Pearson correlation coefficient = 0.81, p<10-12, Figure 1).
MS-MLPA results for each patient are provided as Supplementary Table S3 and Supplementary Figure S1.
CpG islands methylation measured by MS-MLPA predicts survival
In the validation cohort, age, cortisol secretion, tumor size, ENSAT stage, Ki67 and the methylation measured by four MS-MLPA probes showed a significant prognostic value on DFS and OS (Table 2).
CpG islands methylation measured by MS-MLPA was a significant prognostic factor of DFS, with a hazard ratio (HR) for recurrence of 1.013 (p<10-6) for all patients, and 1.017 (p<10 7) for ENSAT stages I-III patients. Methylation was also a significant prognostic factor of OS, with a HR for death of 1.012 (p<10-4) per 1% increase of methylation (Table 2). Death was related to ACC recurrence for 86%, to treatment toxicity for 3% and to other cause for 11% of patients.
In agreement with previous studies, survival dropped in stage III patients, and even more in stage IV patients. Indeed, HR for death was 3.265 (p<104) for stage III vs I-II, and 9.490 and 2.507 (p<10-3) in ENSAT stage IV versus stages I-II and versus stage III respectively.
Proliferation index, reflected by the Ki67 appeared also as a significant prognostic factor of survival, with a HR for recurrence of 1.022 (p<10-3), and a HR for death of 1.028 (p<10-7) per 1% increase of Ki67 index (Table 2).
Correlations between these variables are provided in Supplementary Table S4. Methylation was not strongly correlated with the other prognostic factors. In contrast, Ki67 was more correlated to ENSAT stage (r=0.43).
JCEM
ENDOCRINE SOCIETY
The prognostic value of CpG island methylation is independent from Ki67 and tumor stage In a multivariate model, including Ki67 and ENSAT stage, methylation remained of prognostic significance with a HR of 1.012 for recurrence (p=0.0005) and a HR for death of 1.014 (p=0.0006) per 1% increase of methylation (Table 3).
The ENSAT stage also remained significant, whether excluding stage IV (metastatic patients) or not (for DFS: p= 0.0032 and <10-13; for OS: p=0.001 and <10-11 respectively). The Ki67 did not reach significance for DFS prediction (p=0.19) but was significant for OS (p=0.024).
CpG islands methylation as a prognostic molecular marker for clinical care
For a convenient use of methylation in clinical practice, splitting the methylation values into categories was considered. A methylation level of 25% was determined as the best cutoff in the training cohort (Supplementary Figure S2). In the validation cohort, CpG islands hypermethylation ≥25% was associated with a poor prognosis, both on DFS (HR of 2.80 [1.92- 4.10]; p<10-6) and on OS (HR of 2.39 [1.56-3.67]; p<10-4). Median DFS and OS were 12 and 39 months for hypermethylated tumors and 106 and 180 months for non-hypermethylated tumors (Figure 2).
Methylation, Ki67 and ENSAT stage were then combined into a global prognostic score predicting survival (Figure 3). The score was obtained by the following calculation: methylation ≥25% (+1 point), Ki67 ≥20% (+1 point), tumor stage III (+1 point) and tumor stage IV (+2 points, to take into account the worse prognosis of stage IV vs stage III patients). For DFS, the prognostic score was a significant prognostic factor (log-rank p<10-16; Supplementary Figure S3), and remained significant when restricted to stage I to III patients (log-rank p<10-8; Figure 3A). For OS the prognostic score was also a significant prognostic factor (log-rank p<10-16; Figure 3B). Of particular interest, in patients with stage I to III and Ki67<20%, heterogeneous outcome is observed, well stratified by the methylation level (DFS HR of 3.59 [1.75-7.36], p=0.0005; OS HR of 4.15 [1.71-10.07], p=0.0016, Supplementary Figure S4).
Discussion
This study describes and validates the CpG islands methylation analysis by MS-MLPA as an innovative and important prognostic factor in ACC. In multivariate analyses, methylation assay is established as a strong prognostic factor, independent of the best-established prognostic factors, including tumor stage and Ki67.
Recently genomics have highlighted the importance of methylation analysis for ACC classification opening new perspective both for oncogenesis and precision medicine. This study illustrates how this recent gain in knowledge is useful to develop new molecular tools for ACC prognostic. Targeted measurements by MS-MLPA represent a cost-effective approach for clinical management compared with methylation microarrays. The reliability of MS-MLPA could be confirmed in this study, considering the methylation arrays as the gold standard.
In recent years, attention has been focused on the biology and clinical relevance of the hypermethylated phenotype (CIMP) in different tumor types, especially in colorectal cancer. However, the targeted measurement reflecting the CIMP status of a tumor is challenging, because of the absence of clear definition of CIMP. Likewise, the mechanisms of altered methylation patterns remains unexplained in most tumor types(25). In addition, methylation is a dynamic process, with potential fluctuations over time(26). Despite these limitations, targeted methylation assays have identified a consistent CIMP pattern in a subset of colorectal cancers(27), with hypermethylation being associated with a poor prognosis. A majority of these studies are based on bisulphite treatment of tumor DNA prior to quantitative PCR which are
JCEM
EARLY RELEASE:
ENDOCRINE
SOCIETY
prone to potential biases(28). In the current study we used instead a technique based on differential enzymatic digestion of DNA depending on the methylation status. This technique is commonly used in oncogenetics routine laboratories.
The present study suffers from limitations inherent to retrospective investigations, in which potential confounding factors could impact prognosis. Specifically treatments proposed to patients were not standardized. However, except for complete surgery, these treatments showed a limited efficiency(29). Furthermore treatments were not chosen based on tumor methylation status. In addition, frequency and duration of follow-up was potentially variable among centers in this multicenter setting. These limitations are shared by the vast majority of studies that have addressed the prognosis of ACC (4,5,30-32). Of note, the cohort size is among the largest for molecular studies in the field. In addition, the study was designed in two independent cohorts, one for setting up the targeted methylation marker, and the other for validation.
In the present analysis, ENSAT tumor stage confirmed its strong prognostic power in univariate and multivariate analyses, as previously reported(1,3). The nuclear proliferation marker Ki67 is used as a prognostic marker in several types of neoplasms(33). A recent study in localized ACC highlights Ki67 as the most powerful prognostic factor of DFS after complete surgery. This study suggests a tumor grading based on three levels of Ki67 index <10%, 10-19%, and greater than 20%. The current investigation confirmed Ki67 index as a major prognostic factor in ACC. In the multivariate analysis including also the tumor stage and methylation, Ki67 index remained an independent prognostic factor of OS but not of DFS, and compared adversely to methylation. This better performance of methylation in multivariate analyses may be related to the lower correlation between methylation and ENSAT stage (r=0.22) compared to correlation between Ki67 and ENSAT stage (r=0.43). In addition the limited value of Ki67 could be favored by the limited reproducibility of Ki67, especially in our multicenter setting, lacking a centralized review of Ki67 indexes. However this observation just confirms the still existing problems with significant variability of Ki67 measurements in clinical routine. Indeed, recent studies show in breast and also adrenocortical cancers that the inter-observer variation for intermediate Ki67 indexes (10-30%) is quite high, related to samples preparation and quantification methods(6). This may be a concern for clinical practice since the suggested cuts-off for clinical decisions are in this range. This observation reinforces the need for additional robust markers for the prognostication of adrenocortical tumors.
Overall our study shows for the first time that targeted methylation measurement by MS- MLPA can be used in clinical routine as a prognostic marker, along with the ENSAT stage and the Ki67 index. We propose a prognostic score based on these three variables, readily compatible with clinical practice. However the prognostic model itself has not been tested in a different cohort and requires independent validation. Moreover we do not provide demonstration that this score is the best possible. Indeed potentially other markers, or other ways to combine these markers may prove more accurate in future studies. Especially in patients after “complete surgery”, the decision of an adjuvant treatment should be best based on a reliable prognostication. Patients with high Ki67 index are at high risk of recurrence, whereas the prognosis of patients with low Ki67 index remains heterogeneous. In this subgroup, methylation measured by MS-MLPA can importantly refine the prognosis. Adjuvant mitotane is usually proposed to patients with high-risk of recurrence, i.e. stage III or stage I-II high Ki67 tumors. Based on the results of this study, mitotane could also be offered to patients with stage I-II hypermethylated tumors. In addition, patients with two or three poor prognosis factors could also be proposed adjuvant chemotherapy. However these suggestions require specific validation by clinical trials. One of the advantage of such marker based on DNA analysis -by comparison with
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
previous very potent prognostic RT-qPCR based molecular marker derived from the transcriptome study28,47- relies on the fact that in clinical routine practice good quality tumor nucleic acid is easier to obtain for DNA than RNA. In the future, methylation will have to be confronted with any relevant molecular marker, including those derived from transcriptome, chromosomal alterations and mirnome studies(7,9,20,35-38).
In conclusion, this study demonstrates a strong value of targeted methylation assay by MS- MLPA for prognostication of ACC recurrence and death. This new DNA marker can be implemented in routine laboratory and complete the current clinical and pathological prognostic markers for precision medicine.
Acknowlegdments:
This work was supported by the ENSAT-CANCER Health-F2-2010-259735 program (FP7 program) to the ENSAT network, the Programme Hospitalier de Recherche Clinique to the COMETE network (COMETE-TACTIC), the Brou de Lauriere Foundation (to J Bertherat’s Laboratory), the Cony-Maeva Foundation (to J Bertherat’s Laboratory), the Institut National de la Santé et de la Recherche Médicale (G.A. is receiving a Contratd’Interface). We wish to thank the tumor bank of Cochin hospital (Prof B Terris), the “Centre de Ressources Biologiques” of Bordeaux university hospital, the Endocrinology departments of Ambroise Paré Hospital (Prof ML Raffin-Sanson), Le Mans Hospital (Dr N Saad), St Nazaire Hospital (Dr S Regnier-Le Coz), Lyon University Hospital (Prof F Borson-Chazot, Prof JL Peix), Toulouse University Hospital (Dr D Vezzosi), La Rochelle Hospital (Dr F Duengler), and the oncogenetic unit of Cochin Hospital (Prof E Clauser) for help in samples collection, the members of our laboratories and the COMETE and ENSAT networks for support and discussions and all the staffs of the clinical and pathology departments who were involved in patients care.
Corresponding author and person to whom reprint requests should be addressed:
Professor Jérôme Bertherat, Inserm U1016, 24 rue du Faubourg St Jacques, 75014 Paris, France, Phone: +33158411895, Fax: +33146338060, E-mail: jerome.bertherat@inserm.fr
Financial support: the ENSAT-CANCER Health-F2-2010-259735 program (FP7 program) to the ENSAT network, the Programme Hospitalier de Recherche Clinique to the COMETE network (COMETE-TACTIC), the Brou de Lauriere Foundation (to J Bertherat’s Laboratory), the Cony-Maeva Foundation (to J Bertherat’s Laboratory), the Institut National de la Santé et de la Recherche Médicale (G.A. is receiving a Contrat d’Interface).
Disclosure statement: The authors have nothing to disclose.
References
1. Fassnacht M, Johanssen S, Quinkler M, Bucsky P, Willenberg HS, Beuschlein F, et al. Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: proposal for a Revised TNM Classification. Cancer. 2009; 115:243-50.
2. Else T, Kim AC, Sabolch A, Raymond VM, Kandathil A, Caoili EM, et al. Adrenocortical carcinoma. Endocr Rev. 2014;35:282-326.
3. Lughezzani G, Sun M, Perrotte P, Jeldres C, Alasker A, Isbarn H, et al. The European Network for the Study of Adrenal Tumors staging system is prognostically superior to the international union against cancer-staging system: a North American validation. Eur J Cancer 1990. 2010;46:713-9.
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
4. Assié G, Antoni G, Tissier F, Caillou B, Abiven G, Gicquel C, et al. Prognostic parameters of metastatic adrenocortical carcinoma. J Clin Endocrinol Metab. 2007;92:148-54.
5. Beuschlein F, Weigel J, Saeger W, Kroiss M, Wild V, Daffara F, et al. Major prognostic role of Ki67 in localized adrenocortical carcinoma after complete resection. J Clin Endocrinol Metab. 2015;100:841-9.
6. Papathomas TG, Pucci E, Giordano TJ, Lu H, Duregon E, Volante M, et al. An International Ki67 Reproducibility Study in Adrenal Cortical Carcinoma. Am J Surg Pathol. 2016;40:569-76.
7. Assié G, Letouzé E, Fassnacht M, Jouinot A, Luscap W, Barreau O, et al. Integrated genomic characterization of adrenocortical carcinoma. Nat Genet. 2014;46:607-12.
8. Assié G, Jouinot A, Bertherat J. The “omics” of adrenocortical tumours for personalized medicine. Nat Rev Endocrinol. 2014;10:215-28.
9. Zheng S, Cherniack AD, Dewal N, Moffitt RA, Danilova L, Murray BA, et al. Comprehensive Pan-Genomic Characterization of Adrenocortical Carcinoma. Cancer Cell. 2016;29:723-36.
10. Kulis M, Esteller M. DNA methylation and cancer. Adv Genet. 2010;70:27-56.
11. Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB, Issa JP. CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A. 1999;96:8681-6.
12. Toyota M, Ohe-Toyota M, Ahuja N, Issa JP. Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype. Proc Natl Acad Sci U S A. 2000;97:710- 5.
13. Simons CCJM, Hughes L a. E, Smits KM, Khalid-de Bakker CA, de Bruïne AP, Carvalho B, et al. A novel classification of colorectal tumors based on microsatellite instability, the CpG island methylator phenotype and chromosomal instability: implications for prognosis. Ann Oncol. 2013;24:2048-56.
14. Teodoridis JM, Hardie C, Brown R. CpG island methylator phenotype (CIMP) in cancer: causes and implications. Cancer Lett. 2008;268:177-86.
15. Rechache NS, Wang Y, Stevenson HS, Killian JK, Edelman DC, Merino M, et al. DNA methylation profiling identifies global methylation differences and markers of adrenocortical tumors. J Clin Endocrinol Metab. 2012;97:E1004-1013.
16. Fonseca AL, Kugelberg J, Starker LF, Scholl U, Choi M, Hellman P, et al. Comprehensive DNA methylation analysis of benign and malignant adrenocortical tumors. Genes Chromosomes Cancer. 2012;51:949-60.
17. Barreau O, Assié G, Wilmot-Roussel H, Ragazzon B, Baudry C, Perlemoine K, et al. Identification of a CpG island methylator phenotype in adrenocortical carcinomas. J Clin Endocrinol Metab. 2013;98:E174-184.
18. Pérez-Carbonell L, Alenda C, Payá A, Castillejo A, Barberá VM, Guillén C, et al. Methylation analysis of MLH1 improves the selection of patients for genetic testing in Lynch syndrome. J Mol Diagn JMD. 2010;12:498-504.
19. Johnsen IK, Hahner S, Brière J-J, Ozimek A, Gimenez-Roqueplo AP, Hantel C, et al. Evaluation of a standardized protocol for processing adrenal tumor samples: preparation for a European adrenal tumor bank. Horm Metab Res. 2010;42:93-101.
20. de Reyniès A, Assié G, Rickman DS, Tissier F, Groussin L, René-Corail F, et al. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol. 2009;27:1108-15.
21. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol. 1984;8:163-9.
22. Stell A, Sinnott R. The ENSAT registry: a digital repository supporting adrenal cancer research. Stud Health Technol Inform. 2012;178:207-12.
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
23. Nygren AOH, Ameziane N, Duarte HMB, Vijzelaar RNCP, Waisfisz Q, Hess CJ, et al. Methylation-specific MLPA (MS-MLPA): simultaneous detection of CpG methylation and copy number changes of up to 40 sequences. Nucleic Acids Res. 2005;33:e128.
24. Libé R, Borget I, Ronchi CL, Zaggia B, Kroiss M, Kerkhofs T, et al. Prognostic factors in stage III-IV adrenocortical carcinomas (ACC): an European Network for the Study of Adrenal Tumor (ENSAT) study. Ann Oncol. 2015;26:2119-25.
25. Issa J-P. CpG island methylator phenotype in cancer. Nat Rev Cancer. 2004;4:988-93.
26. Ogino S, Lochhead P, Chan AT, Nishihara R, Cho E, Wolpin BM, et al. Molecular pathological epidemiology of epigenetics: emerging integrative science to analyze environment, host, and disease. Mod Pathol. 2013;26:465-84.
27. Weisenberger DJ, Siegmund KD, Campan M, Young J, Long TI, Faasse MA, et al. CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat Genet. 2006;38:787-93.
28. Berg M, Hagland HR, Søreide K. Comparison of CpG island methylator phenotype (CIMP) frequency in colon cancer using different probe- and gene-specific scoring alternatives on recommended multi-gene panels. PloS One. 2014;9:e86657.
29. Fassnacht M, Terzolo M, Allolio B, Baudin E, Haak H, Berruti A, et al. Combination chemotherapy in advanced adrenocortical carcinoma. N Engl J Med. 2012;366:2189-97.
30. Terzolo M, Angeli A, Fassnacht M, Daffara F, Tauchmanova L, Conton PA, et al. Adjuvant mitotane treatment for adrenocortical carcinoma. N Engl J Med. 2007;356:2372-80.
31. Ayala-Ramirez M, Jasim S, Feng L, Ejaz S, Deniz F, Busaidy N, et al. Adrenocortical carcinoma: clinical outcomes and prognosis of 330 patients at a tertiary care center. Eur J Endocrinol Eur Fed Endocr Soc. 2013;169:891-9.
32. Else T, Williams AR, Sabolch A, Jolly S, Miller BS, Hammer GD. Adjuvant therapies and patient and tumor characteristics associated with survival of adult patients with adrenocortical carcinoma. J Clin Endocrinol Metab. 2014;99:455-61.
33. Rindi G, D’Adda T, Froio E, Fellegara G, Bordi C. Prognostic factors in gastrointestinal endocrine tumors. Endocr Pathol. 2007;18:145-9.
34. Fragoso MCBV, Almeida MQ, Mazzuco TL, Mariani BMP, Brito LP, Gonçalves TC, 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. 2012;166:61-7.
35. Giordano TJ, Kuick R, Else T, Gauger PG, Vinco M, Bauersfeld J, et al. Molecular classification and prognostication of adrenocortical tumors by transcriptome profiling. Clin Cancer Res. 2009;15:668-76.
36. Barreau O, de Reynies A, Wilmot-Roussel H, Guillaud-Bataille M, Auzan C, René-Corail F, et al. Clinical and pathophysiological implications of chromosomal alterations in adrenocortical tumors: an integrated genomic approach. J Clin Endocrinol Metab. 2012;97:E301-311.
37. Soon PSH, Tacon LJ, Gill AJ, Bambach CP, Sywak MS, Campbell PR, et al. miR-195 and miR-483-5p Identified as Predictors of Poor Prognosis in Adrenocortical Cancer. Clin Cancer Res. 2009;15:7684-92.
38. Chabre O, Libé R, Assie G, Barreau O, Bertherat J, Bertagna X, et al. Serum miR-483-5p and miR-195 are predictive of recurrence risk in adrenocortical cancer patients. Endocr Relat Cancer. 2013;20:579-94.
THE JOURNAL OF CLINICAL ENDOCRINOLOGY & METABOLISM
JCEM
EARLY RELEASE:
ENDOCRINE SOCIETY
Figure 1. Reliability of CpG islands methylation measurement by MS-MLPA. For 50 ACC, methylation was measured by MS-MLPA (X-axis) and by methylation array - the gold-standard technique- (Y-axis). The grey dots correspond to non-hypermethylated (non-CIMP) ACC, and the black dots to hypermethylated (CIMP) ACC, as defined by methylome array (average M- value(17)). The two measures strongly correlate (Pearson correlation coefficient = 0.81, p<10-12).
Figure 2. Kaplan-Meier estimates of CpG islands methylation status on disease-free survival and overall survival in the validation cohort. Methylation was measured by MS- MLPA using a cut-off of 25% to define hypermethylated tumors. A. Disease-free survival (DFS) according to methylation status in the validation cohort. B. Overall survival (OS) according to methylation status in the validation cohort
Figure 3. Prognostic value of CpG islands methylation, Ki67 and ENSAT stage combined into a single prognostic score in the validation cohort. The prognostic score is determined by adding the number of pejorative risk factors, including a methylation >25% (+1 point), Ki67 ≥20% (+1 point), ENSAT stage III (+1 point), ENSAT stage IV (+2 points). Kaplan-Meier estimates are presented for (A) disease-free survival for ENSAT stage I to III patients, and (B) overall survival of all ENSAT stages patients from the validation cohort.
| ACA Cohort (n=15) | ACC Training Cohort (n=50) | ACC Validation Cohort (n=203) | p-value* | ||||
|---|---|---|---|---|---|---|---|
| n | mean (sd) / n (%) | n | mean (sd) / n (%) | n | mean (sd) / n (%) | ||
| Age | 15 | 48.9 (18.7) | 50 | 45.9 (17.4) | 203 | 50.8 (15.8) | 0.077 |
| Sex (F/M) | 15 | 13/2 (87) | 50 | 38/12 (76) | 203 | 130/73 (64) | 0.133 |
| Hormonal secretion | 15 | 10 (67) | 50 | 41 (82) | 201 | 141 (70) | 0.112 |
| Cortisol secretion | 14 | 8 (57) | 50 | 33 (66) | 192 | 94 (49) | 0.039 |
| Weiss score | 15 | 2 (0) | 50 | 5.5 (2.1) | 186 | 5.6 (1.9) | 0.653 |
| Ki67 (%) | 7 | 0 (0) | 45 | 2.3 (4.8) | 147 | 15.7 (15.6) | <10-16 |
| ENSAT stage | 48 | 197 | |||||
| stage I | 4 (8) | 24 (12) | 0.092 | ||||
| stage II | 25 (52) | 91 (46) | |||||
| stage III | 4 (8) | 42 (21) | |||||
| stage IV | 15 (31) | 40 (20) | |||||
| Tumor size (cm) | 15 | 4.2 (2.1) | 48 | 11.45 (5.58) | 176 | 9.00 (5.19) | 0.008 |
| Post-operative mitotane | 15 | 0 (0) | 49 | 34 (69) | 144 | 70 (48) | 0.013 |
| Recurrence of ACC | 15 | 0 (0) | 50 | 31 (62) | 203 | 115 (57) | 0.526 |
| Death | 15 | 0 (0) | 50 | 25 (50) | 203 | 89 (44) | 0.526 |
| Follow-up (months) | 15 | 39.3 (25.4) | 50 | 66.6 (64.7) | 203 | 52.6 (51.6) | 0.161 |
* comparison of ACC Training Cohort and ACC Validation Cohort
JCEM
ENDOCRINE SOCIETY
| Disease Free Survival | Overall Survival | ||||
|---|---|---|---|---|---|
| N | HR [95% CI] | p-value | HR [95% CI] | p-value | |
| Age at diagnosis (/y) | 203 | 1.024 [1.011-1.036] | < 10-3 | 1.029 [1.015-1.044] | < 10-4 |
| Sex (F vs M) | 203 | 1.189 [0.816-1.731] | 0.367 | 1.087 [0.707-1.671] | 0.704 |
| Cortisol secretion | 192 | 1.969 [1.334-2.908] | < 10-3 | 2.137 [1.373-3.326] | < 10-3 |
| Tumor size (/cm) | 176 | 1.072 [1.034-1.113] | < 10-3 | 1.064 [1.021-1.110] | 0.004 |
| ENSAT stage | 197 | ||||
| II vs I | 1.422 [0.658-3.070] | 0.370 | 1.744 [0.6004-5.068] | 0.307 | |
| III vs I | 3.426 [1.553-7.558] | 0.002 | 5.070 [1.750-14.689] | 0.003 | |
| IV vs I | 13.960 [6.257-31.148] | < 10-9 | 13.678 [4.821-38.808] | < 10-6 | |
| IV vs I-III | 7.720 [4.94-12.065] | < 10-16 | 5.850 [3.738-9.158] | < 10-13 | |
| Ki67 (/1%) | 147 | 1.022 [1.013-1.032] | < 10-5 | 1.028 [1.018-1.038] | < 10-7 |
| Methylation measured by MS-MLPA (/1%) | 203 | 1.013 [1.008-1.018] | < 10-6 | 1.012 [1.006-1.018] | < 10-4 |
HR: Hazard Ratio.
| Disease Free Survival | Overall Survival | |||
|---|---|---|---|---|
| HR [95% CI] | p-value | HR [95% CI] | p-value | |
| ENSAT stage III vs I-II | 2.536 [1.366-4.707] | 0.0032 | 3.505 [1.656-7.418] | 0.0010 |
| ENSAT stage IV vs I-II | 11.252 [5.966-21.223] | <10-13 | 12.405 [6.080-25.309] | <10-11 |
| Ki67 (/1%) | 1.008 [0.996-1.020] | 0.1935 | 1.015 [1.002-1.029] | 0.0242 |
| Methylation measured by MS- MLPA (/1%) | 1.012 [1.005-1.018] | 0.0005 | 1.012 [1.006-1.021] | 0.0006 |
HR: Hazard Ratio
EARLY
JCEM
ENDOCRINE SOCIETY
EARLY RE
ENDOCRINE SOCIETY
· CIMP
· non-CIMP
0
20
40
60
80
100
Methylation measured by MS-MLPA (%)
EARLY RE
ENDOCRINE SOCIETY
D
100
Non-hypermethylated
Hypermethylated
Disease-free survival (%)
80
60
40
20
0
P = 1.2E-08
0
25
50
75
100
125
150
175
Time from surgery (months)
No. At Risk
| I-hypermethylated | 116 | 69 | 35 | 21 | 12 | 6 | 4 | 0 |
| Hypermethylated | 87 | 27 | 11 | 8 | 6 | 4 | 4 | 2 |
B
100
Non-hypermethylated
Hypermethylated
80
Overall survival (%)
60
40
20
0
P = 4.0E-05
0
25
50
75
100
125
150
175
Time from surgery (months)
No. At Risk
Non-hypermethylated 116
84
48
35
21
10
6
1
Hypermethylated 87
54
25
15
12
8
7
2
EARLY RE
ENDOCRINE SOCIETY
D
Disease-free survival (%)
100
80
60
0 point
1 point
2 points
40
3 points
20
0
P = 1.1E-09
0
25 50 75 100 125 150 17
B
100
Overall survival (%)
80
60
40
0 point
1 point
20
2 points
3-4 points
0
P < 1E-16
0
25
50
75
100
125
150
175
| Time from surgery (months) | Time from surgery (months) | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. At Risk | No. At Risk | ||||||||||||||||
| 0 point | 53 | 42 | 20 | 12 | 6 | 3 | 2 | 0 | 0 point | 53 | 47 | 27 | 21 | 11 | 3 | 2 | 0 |
| 1 point | 28 | 19 | 8 | 5 | 3 | 1 | 1 | 0 | 1 point | 28 | 23 | 14 | 12 | 11 | 6 | 6 | 1 |
| 2 points | 22 | 9 | 3 | 2 | 2 | 1 | 0 | 0 | 2 points | 28 | 21 | 9 | 2 | 2 | 1 | 0 | 0 |
| 3 points | 11 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 3-4 points | 37 | 9 | 3 | 1 | 0 | 0 | 0 | 0 |