Adrenocortical tumors are frequently found in the general population, with an estimated prevalence of 3%–7% [1,2] and are most often discovered incidentally [3]. These tumors may be, more commonly, benign adrenocortical adenomas (ACA). Malignant adrenocortical carcinoma (ACC) is a rare malignancy with an estimated incidence of 0.7–2.0 cases per million per year [2,3,4].

Unfortunately the clinical, biochemical and histopathological distinction between ACA and ACC may be difficult in the absence of widely invasive or metastatic disease [5,6,7,8,9,10]. At present the Modified Weiss Score (a histopathologic scoring) is the most widely accepted means of classifying ACC. However, there is significant variation in inter-observer agreement and it does not identify molecular targets for diagnosis and therapeutic intervention. A better method of distinguishing benign from malignant ACTs is therefore needed, and attention has turned towards a search for molecular markers [11]. Current understanding of the development of adrenocortical carcinoma in the last decade originated, in large part, due to the study of rare genetic diseases that are associated with the development of ACC [12]. In addition, comprehensive genomic hybridization, methylation profiling, and genome wide mRNA and miRNA profiling have, in the more recent past, led to improvements in our understanding, as well as demonstrated several genes and pathways that may serve as diagnostic or prognostic markers [13,14,15,16,17].

Given the diagnostic challenges associated with the definitive diagnosis of ACC, especially in the absence of widely invasive or metastatic disease, identification of molecular markers has been the focus of much recent research. Several types of genomic studies such as gene expression analysis, comprehensive genomic hybridization (CGH), DNA methylation profiling, and miRNA profiling have helped elucidate the roles of several involved genes/pathways that could be used for classification, diagnostic or prognostic purposes. We recently demonstrated that somatic KCNJ5 mutations occur frequently in aldosterone producing ACA [50], but have not been identified in other ACTs.

A subject of much study, and debate, has been the possibility of stepwise progression of tumorigenesis, from benign to malignant ACTs. Arguments put forth are the epidemiology of ACTs and the lack of clear evolution from benign adenoma to ACC based on follow up of non-operated tumors [3,51]. However studies suggest that a significant number of distinct genetic alterations in adenomas are also seen in carcinomas, and cases have been reported which demonstrate both benign and malignant pathological areas of the same tumor [52]. Studies utilizing CGH [53] to study ACTs support the theory of progression from adenoma to carcinoma. Sidhu et al. suggested that activation of a protooncogene on chromosome 4 may be an early event, with progression from ACA to ACC involving activation of oncogenes on chromosomes 5 and 12 and inactivation of tumor suppressor genes on chromosome arms 1p and 17p [16]. Studies also suggest a positive correlation between the number of CGH alterations and tumor size [16,17]. Additionally studies of ACT clonality also demonstrate that most ACAs and all ACCs consist of a monoclonal population of cell, in contradistinction to hyperplastic adrenal tissue, which is usually polyclonal [54,55,56].

Several genome-wide expression studies (Table 1) have been performed over the last decade, which have identified a number of diagnostic markers [5,33,34,57,58,59,60,61,62]. IGF-2 has been consistently exhibited over-expression in a number of studies [33,34,57,58,62,63]. Soon et al. also demonstrate that IGF-2 can be used, in combination with Ki-67, with 96% sensitivity and 100% specificity in diagnosing ACCs [62]. IGF-2 is a growth factor that regulates growth and apoptosis through an interaction with the IGF-1 receptor (IGF-IR). Studies of ACT in the pediatric population have also demonstrated overexpression of IGF-2 in ACT compared to normal adrenal tissue, however to a lesser degree of over-expression compared to adult tumors. In addition, the majority of ACTs overexpressed immature forms of IGF-2 [63]. In another study evaluating the expression of IGF-2 in pediatric and adult ACT, IGF-2 was found to be overexpressed in adult ACC, whereas IGF-IR was found to be overexpressed in pediatric ACC. While IGF-IR expression was similar in benign and malignant adult ACT, IGF-IR over-expression was identified only in pediatric adrenocortical carcinomas, and was also found to be a predictor of metastases in children with ACT [64]. Several other growth factors have also been shown to be overexpressed in ACC, including TGFβ2 [34], TGF-β1 [65], insulin like growth factor-related genes; IGF2, IGF2R, IGFBP3 and IGFBP6 [58], and VEGF [66].

Szabo et al., in a meta-analysis of gene expression and CGH profiling data, revealed three major pathogenetic pathways to be dysregulated in ACC: cell cycle (c-MYC, CDC25B, CCNB2, CDC2, TOP2A CCNE1, CDK2, CDK7, UBC, MDM-2), retinoic acid signaling (RXRA, ALDH1A1, ALDH1A1-3), and complement system and antigen presentation SERPING1 and MHCII) [67].

Soon et al. demonstrate that a combination of IGF2 and Ki-67 exhibit 96% sensitivity and 100% specificity in diagnosing ACCs [62]. The combination of IGF2 and Ki-67 was also shown by Schmitt et al to accurately predict malignancy in ACT [68].

Several studies have evaluated prognostic markers in ACC. In a cluster analysis of the ACCs, Giordano et al. revealed two different prognostic subtypes that reflected tumor proliferation and survival differences [61]. In another study, combined expression of DLG7 (overexpressed) and PINK1 (under-expressed) was shown to be the best predictor of disease-free survival in ACT. Combined expression of BUB1B and PINK1 (decreased expression) was the best predictor of overall survival in ACC [5].

Morimoto et al. analyzed the Ki67 labeling index, and demonstrated that an index of 7% or more was associated with significantly shortened disease-free survival in ACC [69]. Sbiera et al. demonstrated that the SF-1 gene and protein expression significantly correlated with poor clinical outcome, including tumor stage-adjusted hazard ratio for death and recurrence [70]. SF-1 is amplified and overexpressed in childhood ACC [71,72], increases proliferation, and decreases apoptosis of human adrenocortical cells, and induces adrenocortical tumors in transgenic mice [71,72,73].

Two recent studies that classified ACCs according to their transcriptome identified two distinct prognostic groups of ACCs [61,74]. Ragazzon et al. describe two prognostic clusters with genes involved in transcription and the cell cycle overexpressed in the poor-outcome group, and genes involving cell metabolism and intracellular transport overexpressed in the good-outcome group [74]. Cluster analysis of the ACCs performed by Giordano et al revealed two subtypes that reflected tumor proliferation, as measured by mitotic counts and cell cycle genes [61].

ACCs are associated with significant chromosomal gains and losses compared to ACAs, and these changes have been demonstrated to correlate with tumor size [53].

Chromosomal gains in 6q, 7q, 12q, and 19p, and chromosomal losses in 3, 8, 10p, 16q, 17q, and 19q are associated with worse prognosis [53]. Mateo et al., in a study on pediatric and adolescent ACTs, also demonstrated a relationship between tumor size and number of chromosomal changes. The most frequent chromosomal changes were gains were of the 1p21-p31.2 and 2p12-p21 regions, and the loss of region 20p11.2-p12, more frequently in ACC compared with ACA, thus supporting the concept of a progressive pattern in adrenocortical tumorigenesis, previously mentioned in other studies on adult ACTs [16,43]. A diagnostic tool developed by Barreau et al., by combining DNA copy number estimates at six loci (5q, 7p, 11p, 13q, 16q, and 22q) distinguished ACC from ACA in an independent validation cohort with a sensitivity of 100% and specificity of 83% [75].

There is a growing body of evidence to suggest that epigenetic abnormalities including DNA methylation and histone modification play a key role, in conjunction with genetic modifications, to cause altered patterns of gene expression, resulting in tumorigenesis. DNA hypermethylation of the promoter cytosine in cytosine phosphate guanine islands (CpG islands) causing downstream gene silencing has been shown to be intimately involved in tumorigenesis [76,77,78]. Altered DNA methylation of the H19 promoter has been shown to be involved in the abnormal expression of both H19 and IGF-2 genes in adrenocortical carcinomas [79]. Promoter methylation of TP53 however, has been demonstrated not to be a significant event in the development of adrenocortical carcinomas [80]. Two recent genome-wide DNA methylation profiling studies have shed additional light on the role of methylation in ACTs. We described six genes (CDKN2A, GATA4, DLEC1, HDAC10, PYCARD, SCGB3A1/HIN1) that were demonstrated to be significantly hypermethylated in ACC when compared with ACA and normal tissue. This finding was also associated with a corresponding decrease in gene expression. In addition, treatment of H295R cells with a demethylating agent restored expression of the hypermethylated genes, indicating that this is a reversible event [13]. Rechache et al. demonstrate that ACC samples were globally hypomethylated and describe distinct methylation patterns that could distinguish normal, benign, primary malignant, and metastatic tissue samples [81].

MicroRNAs (miRNAs) are short non-coding RNAs that are involved in post-transcriptional regulation of gene expression. miRNAs regulate approximately one third of coding genes, therefore changes in miRNA expression may be associated with cancer development and progression.

miR-483-5p is one of the most investigated miRNAs involved in ACC. It has been shown to be upregulated in ACC compared to ACA, and has been identified as a predictor of poor prognosis in ACC [14,15]. miR-483-5p and IGF2 were also demonstrated to be statistically significantly co-expressed in ACC [14]. Given that miR-483-5p is located at 11p15.5 within the second intron of IGF2, it is likely that increased expression of miR-483-5p observed in ACC may be an indirect consequence of IGF2 over-expression. miR-483-3p was also found to be upregulated in ACC compared to ACA, which significantly correlated with miR-483-5p levels [82]. This upregulation of miR-483-3p was also seen in the study by Soon et al. [15], as well as in pediatric ACT [83]. miR-503 was also found to be upregulated in ACCs compared to normal adenocortical tissue and ACA in both pediatric and adult ACTs [15,83].

miR-184 has also been shown to be over-expressed in ACCs compared to ACAs [84]. miR-195 is another miRNA that has been the focus of much study in ACC. It is significantly down-regulated in ACCs compared to ACAs, and its low expression in ACCs is significantly associated with poor overall disease specific survival [14,15], It is also down-regulated in childhood ACTs [83].

miR-7 and miR-335 are also significantly down-regulated in ACCs compared to ACAs and normal adrenocortical tissue [15]. miR-214 and miR-375 are similarly significantly down-regulated in ACCs compared to ACAs and normal adrenocortical tissue in both adult and pediatric ACT [83,84].

Given the lack of a definitive histopathologic method of diagnosis of ACC, especially in the absence of widely invasive or metastatic disease, the identification of molecular markers has been the focus of much recent research. Mutations found in familial cancer syndromes have contributed significantly to our understanding of the pathogenesis of sporadic ACTs. In addition, genome-wide gene expression and DNA methylation studies, along with miRNA profiling have helped identify certain diagnostic and prognostic markers. There are however, substantial differences in the candidate molecular markers identified, indicating a heterogenous genetic and epigenetic basis. This, together with the relative rarity of ACC and the resultant small sample sizes analyzed, signify that additional studies are still required to further our understanding of this disease process.