Novel molecular signatures for adrenocortical carcinoma
Matthias J. Betz and Felix Beuschlein
The advent of high-throughput molecular analysis has appreciably improved the diagnostic and prognostic workup of several types of cancer. A French study now offers hope that benign and malignant adrenocortical tumors can be distinguished on the basis of distinct differences in their gene-expression profiles.
The ability to distinguish between benign and malignant adrenocortical tumors requires great expertise; as a consequence, improvements in the early diagnosis, risk stratification and selection of effective therapies are urgently needed. Over the past decade, high-throughput techniques have emerged as powerful tools for the molecu- lar and functional characterization of cancer cells. The researchers of a study published in the Journal of Clinical Oncology1 used gene- expression profiling to show that combined expression of DLG7 and PINK1 (Box 1) robustly predicts disease-free survival in patients with malignant adrenocortical tumors. Similarly, combined expression of BUB1B and PINK1 can be used to predict overall survival in this group of patients.
Adrenocortical tumors are not uncom- mon, and are increasingly detected incidentally by abdominal imaging proce- dures; the prevalence of these tumors in a
BoTCG A AGGGT
representative population aged >50 years is estimated to be around 3%. By contrast, primary adrenocortical carcinoma is a rare and highly malignant disease that affects both adults and children; these tumors have an incidence of 1-2 per million people. Retrospective studies of combined surgical and medical treatment indicate a very unfa- vorable 5-year survival rate of 15-35% for patients with malignant lesions.2
The diagnostic workup of a patient with suspected adrenocortical carcinoma consists of a thorough endocrine assessment before surgery, as well as imaging of the abdomen and thorax for preoperative evaluation and staging of the tumor. The size of an adrenal mass is currently considered one of the best indicators of malignancy, and specific imaging procedures that assess kinetics of contrast enhancement further aid the discrimination of adrenal adenomas from malignant adrenal carcinomas.1
The prognosis and choice of treatment modalities for patients with adrenocortical carcinoma are highly dependent upon the tumor stage. In 2004, the International Union Against Cancer (UICC) defined tumor-node-metastasis (TNM) criteria and published the first staging classifica- tion for adrenocortical carcinoma, which was based on the Sullivan modification of the Macfarlane system.3 A post hoc analysis revealed a limited prognostic value of the UICC TNM staging criteria; this analysis was endorsed by the European Network for the Study of Adrenal Tumors (ENS@T) and led to an alternative staging system that had improved prognostic accuracy.4
As a result of the lack of clear-cut criteria for malignancy, histopathological diagnosis of adrenal tumors can be complex. Such diagnosis is currently made on the basis
Box 1 | Prognostic markers
DLG7
· Discs, large homolog 7 (Drosophila)
· Implicated in cell-cycle regulation
· Expression downregulated during cellular differentiation
· Overexpressed in bladder, colon and hepatocellular carcinomas
PINK1
· PTEN-induced putative kinase 1
· Localized to mitochondria
· Potential serine/threonine kinase
· Expression decreased in ovarian cancer
BUB1B
· Budding uninhibited by benzimidazoles 1 homolog ß (yeast)
· Implicated in cell-cycle regulation (mitotic checkpoint) and genetic stability
· Genetic alterations linked to colon cancer
of a combination of several microscopic features and use of a scoring system. One of the most widely used scoring systems is the Weiss score, which combines a total of nine parameters related to tumor struc- ture, cytology and invasion. Some of these criteria are objective and reproducible (for example, mitotic count or nuclear grade), whereas others are quite subjective and depend upon the quality of the tumor sample (for example, diffuse architecture, pleomorphism, or tumor invasion), which makes the workup of tumors a challeng- ing and time-consuming task that requires expert knowledge. In addition to morpho- logical features, several studies have high- lighted the value of Ki-67 positivity as a marker of proliferative activity. The cut-off value of Ki-67 to discriminate between
Practice points
· Benign and malignant adrenocortical tumors can be distinguished by their unique gene-expression profiles
· Combined expression of DLG7 and PINK1 predicts disease-free survival in patients with malignant disease
· Combined expression of BUB1B and PINK1 predicts overall survival in patients with malignant disease
benign and malignant adrenocortical tumors varies from 1.5% to 4.0%; however, a high level of expression (>10%) has been associated with poor survival.2,5
A group of researchers led by Jérôme Bertherat (Hôpital Cochin, Paris, France)1 has attempted to overcome these diag- nostic limitations by using microarray and quantitative reverse-transcription poly- merase chain-reaction (qRT-PCR) assays to analyze 153 unilateral adrenocortical tumors and correlate gene-expression profiles with clinical, hormonal and histo- pathological data. Intriguingly, cluster analysis of the microarray data identified two gene-expression profiles that clearly differentiated between patients with malig- nant and benign clinical phenotypes. In addition, the cluster analysis revealed two robust subgroups within the malignant group, each with distinct clinical outcomes. The researchers were able to reduce the number of genes necessary for accurate dis- crimination of these two subgroups by using a training cohort of 47 randomly chosen samples to develop a two-gene signature of malignancy. The combination of expres- sion levels of DLG7 and PINK1 provided the best predictive rule for disease-free survival, as determined by qRT-PCR of a large, independent validation set of 104 tumor samples. All the obviously malignant (metastatic) tumors were correctly classified by this method. Furthermore, the molecular signature of malignancy was highly predic- tive of disease-free survival for tumors that were not obviously malignant before histo- pathological examination (Macfarlane stages 1 and 2), even when compared with, and adjusted for, the Weiss score.
In addition to assessment of disease-free survival, de Reyniès et al.1 used a subset of 14 overtly or potentially malignant tumors (Weiss score ≥2) and nine randomly selected tumors as a training cohort to identify 28 genes that could predict overall survival.
The combined expression levels of BUB1B and PINK1 provided the best prediction rule for this outcome and was highly pre- dictive of overall survival when tested on an independent validation cohort of overtly or potentially malignant tumors. This signa- ture performed better than the Macfarlane tumor stage, even after subgroup analysis of Macfarlane stages 1 to 2 and stages 1 to 3, respectively. Notably, the findings of another study (which used a similar approach) were in-line with those reported by de Reyniès and colleagues. Giordano et al.6 used principal-component analysis of microarray data that clearly separated adrenocortical carcinomas from normal adrenal glands and adrenocortical adenomas. Although another set of genes was identified in their analysis, these researchers were also able to define two subgroups of adrenocortical carcinomas that reflected different mitotic grades (high-grade versus low-grade), and differed appreciably in terms of survival.
The study of de Reyniès et al.1 makes an important translational step, in that the researchers correlated disease-free survival and overall survival with molecular data, and thereby gained valuable information about the diagnosis of malignancy and the predic- tion of survival. Although confirmation of the malignant state is often straightforward for adrenocortical carcinoma, especially in cases with local invasion or distant meta- stasis, the malignant potential of some local- ized tumors remains uncertain, even after histopathological workup. The molecular prediction tool developed by de Reyniès et al.1 could potentially improve diagnosis of these difficult cases and support diagnosis in all other patients, as the techniques required are widely available, easy to standardize and relatively inexpensive.
Despite this progress in diagnosis and risk stratification, identification of additional molecular markers that predict treatment responses will be required before a persona- lized treatment concept for patients with adrenocortical cancer can be introduced into clinical practice. Complete removal of the tumor by a specialized endocrine surgeon is paramount in the management of tumor stages 1 to 3; however, the majority of patients have locally advanced or metastatic disease at initial presentation. The main- stay of medical therapy in stage 4 disease is mitotane, an adrenolytic and adrenostatic compound, which promotes tumor regres- sion in 25% of patients and controls hormone
excess in the majority of cases. Although experience with cytotoxic chemotherapy in patients with adrenocortical carcinomas is still limited, the most commonly used treatment regimens are the combination of mitotane plus etoposide, doxorubicin and cisplatin, or mitotane plus streptozoto- cin. The efficacy and safety of these treat- ment options is currently being assessed in the First International Randomized Trial in Locally Advanced and Metastatic Adrenocortical Carcinoma Treatment (FIRM-ACT) study.7 In the context of this study, Ronchi et al.8 were able to demon- strate that high expression of ERCC1 strongly correlated with overall survival after platinum-based chemotherapy.
Future studies that combine carefully col- lected clinical data with molecular markers could provide the opportunity to tailor individual cancer treatment regimens that aim to minimize toxic effects but increase therapeutic efficacy. Current progress in the field provides great hope that this challenge can also be met for a rare tumoral entity, such as adrenocortical carcinoma.
Division of Endocrinology, Campus Innenstadt University Hospital Munich, Germany (M. J. Betz, F. Beuschlein).
Correspondence: F. Beuschlein, Division of Endocrinology, Campus Innenstadt University Hospital Munich, Ziemssenstrasse 1, Munich, DE-80336, Germany
felix.beuschlein@med.uni-muenchen.de
doi:10.1038/nrendo.2009.93
Competing interests
The authors declare no competing interests.
1. de Reyniès, A. et al. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J. Clin. Oncol. 27, 1108-1115 (2009).
2. Allolio, B. & Fassnacht, M. Clinical review: adrenocortical carcinoma: clinical update. J. Clin. Endocrinol. Metab. 91, 2027-2037 (2006).
3. DeLellis, R. A., Lloyd, R. V., Heitz, P. U. & Eng, C. (eds) Pathology and Genetics of Tumours of Endocrine Organs (IARC WHO Classification of Tumours) (IARC Press, Lyon, 2004).
4. Fassnacht, M. et al. Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: proposal for a revised TNM classification. Cancer 115, 243-250 (2009).
5. Volante, M., Buttigliero, C., Greco, E., Berruti, A. & Papotti, M. Pathological and molecular features of adrenocortical carcinoma: an update. J. Clin. Pathol. 61, 787-793 (2008).
6. Giordano, T. J. et al. Molecular classification and prognostication of adrenocortical tumors by transcriptome profiling. Clin. Cancer Res. 15, 668-676 (2009).
7. . First International Randomized trial in Locally Advanced and Metastatic Adrenocortical Carcinoma Treatment
(FIRM-ACT), firm act study [online] www.firm-act.org (2009).
8. Ronchi, C. et al. Expression of excision repair cross complementing group 1 and prognosis in adrenocortical carcinoma patients treated with platinum-based chemotherapy. Endocr. Relat. Cancer doi:10.1677/ERC-08-0224 (2009).
DIABETES
The HEART2D trial-new questions beyond answers
Louis Monnier and Claude Colette
Persistent hyperglycemia increases the risk of cardiovascular events in patients with type 2 diabetes mellitus who have a history of acute myocardial infarction. Whether clinicians should target prandial glucose levels rather than basal glucose levels to reduce this excess risk is keenly debated.
Defining the most appropriate insulin regimen to manage glycemic disorders associated with type 2 diabetes mellitus has been a key concern for diabetologists over the past few years.1,2 A large body of evidence exists that favors initiation of insulin therapy to prevent progression of vascular diseases; however, whether control of basal hyperglycemia or post- prandial hyperglycemia is best in this setting remains an open question. The researchers of the Hyperglycemia and its Effect after Acute Myocardial Infarction on Cardiovascular Outcomes in Patients with Type 2 Diabetes Mellitus (HEART2D)3 trial set out to provide an answer. The primary
objective of HEART2D was to demon- strate the differential effect of two strate- gies of insulin treatment on the time to the first combined, adjudicated cardiovascular event in patients with type 2 diabetes mel- litus who had a history of acute myocardial infarction. The researchers expected that an intervention designed to target prandial hyperglycemia would be more beneficial than one that targeted basal hyperglycemia. Surprisingly, they found little difference between the two treatments.
Insulin regimens based on a once-daily injection of a long-acting preparation of basal insulin are generally preferred to prandial insulin regimens for the treat- ment of patients with progressive type 2 diabetes mel- litus, at least when insulin therapy is first implemented. Basal insulin regi- mens are associated with a lower risk of hypoglycemia than are prandial insulin regimens. Furthermore, basal insulin regimens are easier to implement than are prandial insulin regimens, which require mul- tiple daily injec- tions of rapid-acting
insulin analogs to be given before each meal. In terms of efficacy, HbA . goals can be achieved by either insulin regimen; never- theless, analysis of data from clinical trials suggests that the choice of basal or pran- dial insulin regimens remains a subject of debate.1,2 However, these trials were limited to the investigation of short-term effects and were never intended to test long-term outcomes, such as risk of cardiovascular events. By contrast, the HEART2D trial was designed to determine which glucose target-basal or prandial-should be preferentially selected to prevent additional cardiovascular events after acute myocardial infarction in patients with type 2 diabetes mellitus.
HEART2D was a well-controlled, multi- national, prospective, open-label, two-arm, parallel clinical trial. The investigators enrolled patients aged 30-75 years with poorly controlled type 2 diabetes mellitus who had experienced acute myocardial infarction within the previous 21 days. Eligible participants were randomly allo- cated to one of two treatment groups. The first group (558 patients) was assigned to a strategy that targeted basal glycemia (fasting and interprandial) by administration of either twice-daily neutral protamine Hagedorn insulin or once-daily insulin glargine. The target fasting or premeal blood glucose level was <6.7 mmol/l. In the second group (557 patients), the rapid-acting insulin analog lispro was administered thrice-daily at meal times to target prandial glycemia; the target postprandial blood glucose level was <7.5 mmol/l. Treatment in the prandial insulin group was intensified by the addi- tion of neutral protamine Hagedorn insulin at bedtime if satisfactory glycemic control was not achieved.
The target HbA level for both treatment groups was <7%. The two insulin regimens caused similar falls in the levels of HbA16 however, only 28% of patients in the pran- dial insulin group and 31% of patients in the basal insulin group attained the target level. The primary outcome measure was the time to the first combined, adjudicated cardio- vascular event (cardiovascular death, non- fatal myocardial infarction, nonfatal stroke, coronary revascularization, or hospitaliza- tion for acute coronary syndrome). No difference in cardiovascular events was detected between the two treatment groups, which led the investigators to halt the trial after a mean follow-up of 2.7 years. The null