Introduction

Programmed loss of telomeric DNA (telomere shortening) in somatic cells leads to the activation of DNA damage responses and subsequent cell cycle arrest and senescence [1]. Telomerase reverse transcriptase, which is encoded by the TERT gene, forms an important component of the telomerase complex and TERT mediated de novo telomeric DNA synthesis in rapidly proliferating cancer cells prevents chromosomal ends from being recognized as sites of DNA damage [1]. This prevents the initiation of repair pathways and allows neoplastic cells to escape telomere crisis [1]. Alterations of the TERT gene such as “hotspot” promoter mutations, amplifications and promoter rearrangements have been associated with increased TERT expression and these alterations have been frequently seen in multiple cancer types [2,3,4,5,6].

Adrenocortical carcinomas are rare endocrine malignancies with an estimated annual incidence of 0.7 to 2 per million and advanced stage neoplasms are associated with extremely poor outcomes [7,8,9,10,11,12]. Whole genome doubling associated with decreased telomere length and increased TERT expression has been associated with disease progression in adrenocortical carcinomas in The Cancer Genome Atlas (TCGA) datasets [12]. A high incidence of genomic amplifications at the TERT locus (5p15.33) has also been identified in at least two separate studies (TCGA≈15%, 13 of 89; Assie et al. ≈ 6%, 7 of 122) [12, 13]. In comparison, “hotspot” TERT promoter alterations are relatively rare in this tumor type along with TERT promoter rearrangements, and the latter are poorly characterized and under-recognized [3, 4, 12, 13]. Due to an association between whole genome doubling and TERT expression in a prior TCGA study, it has been suggested that TERT is required in a subset of adrenocortical carcinomas for telomere maintenance [12].

As there is a paucity of prognostic biomarkers in adrenocortical carcinomas, we interrogated a large cohort of 169 cases (institutional cohort: 78, TCGA: 91) for TERT alterations and correlated the presence of these alterations with various clinicopathologic parameters and outcomes.

Materials and Methods

Patient Specimens

This study was approved by our Institutional Review Board. Adrenocortical carcinomas were diagnosed using modified Weiss criteria and some of the features of malignant tumors are illustrated in Fig. 1. In addition, the Lin-Weiss-Bisceglia system was used for oncocytic adrenocortical tumors [14,15,16,17,18]. Tumors with a score of 3 under the Weiss system, and those fulfilling 1 to 4 minor criteria under the Lin-Weiss-Bisceglia system were classified as “borderline” in this study. Eighty-seven adrenocortical neoplasia (76 adrenocortical carcinomas, 4 adrenocortical borderline tumors and 7 adrenocortical adenomas) were analyzed by a next generation sequencing (NGS)-based assay, Memorial Sloan Kettering Cancer Center Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT), as part of an institutional clinical cancer genomics initiative [5, 19, 20]. A subset of 9 adrenocortical carcinomas and 1 borderline adrenocortical tumor were analyzed using fluorescence in situ hybridization (FISH). Cases profiled by FISH included 3 cases that were only profiled using this testing modality (adrenocortical carcinoma: 2; adrenocortical borderline tumor: 1). Archived H&E and immunohistochemical-stained slides were reviewed and medical records were accessed for relevant clinicopathologic features including outcomes on follow up.

Fig. 1
figure 1

Histopathologic features of adrenocortical carcinomas including high mitotic index (a), necrosis (b), extra-capsular extension (c), lymphovascular invasion (d), and lymph node involvement (e) have been depicted. An adrenocortical carcinoma with myxoid features (f) that harbored a TERT amplification has been shown

Full size image

Next Generation Sequencing-Based Molecular Profiling

Specifics regarding the MSK-IMPACT assay have been previously reported [5, 19, 20]. This assay involves hybridization capture-based library preparation followed by deep sequencing of select noncoding regions and 6614 protein-coding exons of 468 genes (previous versions interrogated 341 and 410 genes) including TERT [6]. Sequencing 500 bases upstream of the TERT promoter allows for the identification of “hotspot” promoter mutations as well as structural variants that may occur in this region [6]. In addition, the homogenous distribution of single nucleotide polymorphism tiling probes across the genome allows for assessment of copy number alterations. Based on previously reported criteria, gains were defined as a fold change ≥ 1.5 and < 2.0, while amplifications were defined as a fold change ≥ 2.0 [21,22,23].

Fluorescence In Situ Hybridization

Fluorescence in situ hybridization (FISH) analysis was performed on paraffin sections using a 3-color probe set. The probe mix consisted of BAC clones spanning the TERT gene (5p15), while probes for the 5p12 and 5q11 loci served as controls. Probe labeling, tissue processing, hybridization, post-hybridization washing, and fluorescence detection were performed according to standard laboratory procedures. Slides were scanned using a Zeiss Axioplan 2i epifluorescence microscope equipped with MetaSystems (Waltham, MA) imaging system. Metafer and VSlide modules within the system were used to generate virtual image of H&E and DAPI-stained sections. For all cases, the corresponding H&E sections assisted in localizing regions of interest for downstream analysis. Signal counts were performed on a minimum of 20 discrete nuclei and only nuclei with at least 2 signals for TERT and control probes were selected. TERT amplification was defined as ≥ 10 copies of TERT or a TERT: 5q11 probe ratio > 2. Cells with large clusters of TERT signal (high-level HSR type amplification), were interpreted as being amplified.

Literature Review and Data Extraction from the Cancer Genome Atlas Datasets

The publicly available cBioPortal.32e34 platform was used to analyze data from The Cancer Genome Atlas (TCGA) pertaining to pan-genomic characterization of adrenocortical carcinomas [24, 25].

Statistical Analysis

Continuous variables were evaluated with frequency counts and percentages and all tests to assess for statistical significance were two-sided and p-values < 0.05 were considered significant.

Results

Prevalence of TERT Alterations in Adrenal Cortical Carcinomas

76 adrenocortical carcinomas were evaluated for “hotspot” TERT promoter mutations as well as for structural variants involving the first 500 bases upstream of the transcriptional start site (Table 1). In addition, using a combination of NGS (n = 76) and FISH (n = 9), 78 cases of adrenocortical carcinomas were evaluated for copy number alterations (Table 1).

Table 1 Prevalence of TERT alterations in adrenal cortical tumors
Full size table

The most common alteration involving the TERT gene involved genomic gains/amplifications which were seen in 30.7% of cases (24/78; Fig. 2A–B; Table 1). After taking into consideration similar events identified in TCGA datasets, the combined incidence for gains/amplifications was 22.2% (37 of 167 cases) [25]. This was significantly higher compared to both TERT promoter rearrangements (2/165, 1.2% cases; Fig. 2C; Table 1) and “hotspot” C > T promoter mutations at position -124 relative to the transcription start site (also referred to as the C228T alteration, 7/167, 4.2% cases; Fig. 2D; Table 1) [25]. Interestingly, consistent with a prior TCGA study, no C250T promoter alterations were identified in our institutional dataset [12]. In addition, these TERT alterations (gains/amplifications, “hotspot” promoter mutations and promoter rearrangements) were mutually exclusive events.

Fig. 2
figure 2

Copy number alterations for an adrenocortical carcinoma, including TERT amplification, determined using next generation sequencing and fluorescence in situ hybridization has been depicted (a, b). Relative (Log2) tumor/normal ratios (y axis) and corresponding chromosomes (x axis) are displayed, with each blue dot representing an individual probe region and amplified regions are shown in red (a). Corresponding fluorescence in situ hybridization results using probes for TERT (5p12, red) and control probes (5p12, yellow; 5q11, green) is shown (b). A structural rearrangement of the TERT promoter region with CLPTM1L (c), and a somatic TERT promoter mutation (g.1295228C > T) (d) identified using hybridization-capture based next generation sequencing has been depicted

Full size image

Of note, 2 of 5 borderline adrenocortical tumors showed similar TERT gains/amplifications (Table 1), while no TERT alterations were identified for 7 adrenocortical adenomas.

Clinicopathologic Features of Adrenocortical Carcinomas with TERT Alterations

After evaluating abstracted data from institutional cases (n = 78) and TCGA cases (n = 89), no significant differences were identified regarding mean age at presentation, tumor laterality, size and weight, irrespective of TERT alteration status (Table 2). Similarly, while cases with TERT alterations tended to have a lower mitotic index and Ki67 labeling (proliferative index), neither of these metrics were statistically significant (Table 3). Interestingly, adrenocortical carcinomas in both the institutional and TCGA cohorts showed a female predilection, irrespective of TERT status and the significance of this observation is unclear (Table 2).

Table 2 Clinical features of adrenal cortical carcinomas with TERT alterations
Full size table
Table 3 Histopathologic and genomic features of adrenal cortical carcinomas with TERT alterations
Full size table

Recurrent alterations that co-occurred with TERT alterations included 22q (NF2/SMARCB1) copy number losses in 24% of cases (9 of 38). Furthermore, copy number losses at 9p21 (CDKN2A/CDKN2B) and alterations involving TP53, CTNNB1 and NF1 were each seen in ≤ 16% of cases with TERT alterations in the institutional and TCGA cases combined (n = 38) (Table 3).

For patients with at least 3 months of clinical follow up (institutional cohort, n = 48; TCGA, n = 48, total, n = 96) the presence of TERT alterations correlated with adverse clinical outcomes. While TCGA data suggested a trend towards worse outcomes (69.2% vs 38.2%, p = 0.06), defined as documented metastatic disease either at diagnosis or on follow up, no such statistically significant differences were identified in the institutional cohort of cases (70.6% vs 70.4%, Table 2). However, at the same time for all cases combined, those with TERT alterations (28/40, 70%) had worse outcomes compared to those that lacked alterations of the TERT gene (65/127, 51.2%, p = 0.04; Table 2), including 6/35 (17.1%) patients with TERT alterations dying of disease related complications compared to 6/127 patients that lacked these alterations (4.7%, p = 0.02; Table 2). These findings suggest that TERT alterations do indeed adversely affect outcomes.

Discussion

This is a large study involving 181 cases of adrenocortical tumors. A combination of hybrid-capture based next generation sequencing/ fluorescence in situ hybridization (n = 78) and whole exome sequencing/targeted TERT promoter sequencing (The Cancer Genome Atlas, n = 91) was used to define both the prevalence of copy number alterations and “hotspot” promoter mutations affecting the TERT gene in adrenocortical carcinomas. Unique findings of this study include understanding the landscape of TERT alterations in these tumors which predominantly include gene amplifications, with a smaller subset harboring “hotspot” promoter mutations and rearrangements. The documentation of an increased incidence of metastatic disease (70% vs 51.2%, p = 0.04) and disease specific death (6 of 25, 17.1% vs 6 of 127, 4.7%) among patients with known TERT alterations in combined institutional/ TCGA cases of adrenocortical carcinomas suggests that these alterations may have significant impact on prognostication. This observation is further underscored by the fact that no TERT alterations were identified in 7 cases of adrenocortical adenomas. While TERT amplifications were identified in 2 of 5 borderline adrenocortical tumors, clinical follow up was available for only one of these patients and the prognostic significance of this finding in borderline tumors is therefore unclear.

The spectrum of alterations that have been shown to increase TERT expression across tumor types are broad [6]. While genomic amplification events for TERT at the 5p15.33 locus have been documented in multiple studies of adrenocortical carcinomas, other mechanisms of upregulation of TERT have been documented across various tumor types [6, 12, 13, 26,27,28,29,30]. For instance, G > A substitutions at − 124 and − 146 base pairs upstream of the transcription start site for the TERT gene (also referred to as “C228T” and “C250T” promoter mutations) are known to increase recruitment of the heterotetrameric GABP transcription factor complex; this complex binds both at this site and at an adjacent ETS binding site to consequently upregulate TERT gene expression [5, 6, 31,32,33].

Rearrangement events occurring within the TERT promoter region have been documented in select tumor types and are thought to upregulate TERT expression by juxtaposing the TERT coding sequence to strong enhancer elements [3, 4, 6]. While whole genome sequencing approaches have been successful in identifying such alterations, these events are likely under-recognized as most contemporary next generation sequencing assays only sequence a limited portion of the TERT promoter [3, 6]. For instance, the MSK-IMPACT assay which sequences the first 500 base pairs upstream of the TERT transcription start site was able to identify one such event and in the absence of gene expression data, the functional significance of such alterations remains to be better defined [6]. Only one such case was identified in the cohort of adrenocortical carcinoma profiled by TCGA (TCGA-OR-A5JS) and this case harbored a CCDC47-TERT fusion. Due to the rarity of these events in adrenocortical carcinomas there is insufficient data at present to comment on whether TERT structural variants in this tumor type exhibit a conserved breakpoint.

In addition, hypermethylation of specific CpG islands upstream of the TERT transcriptional start site have been associated with increased TERT expression and aggressive disease in both pediatric brain tumors and in adrenocortical carcinomas [30, 34]. However, a recent study suggests that these hypermethylation events are not mutually exclusive with other TERT alterations such as promoter mutations and/or copy number alterations [6].

A prior study of pediatric adrenocortical carcinomas had documented a high incidence of TERT amplifications in a relatively small cohort of patients (13 of 19, 68.4%) without evaluating for “hotspot” promoter mutation status [26]. In the current study, only a minority of adrenocortical carcinomas occurred in patients less than 20 years of age (institutional cohort: 4 of 78 cases; The Cancer Genome Atlas: 3 of 91 cases) and no TERT alterations were documented in these cases.

After combining the results of our study with TERT alterations reported in other studies of adrenocortical carcinomas (Assie et al. 122 cases; Liu et al. 34 cases; Juhlin et al., 41 cases), the overall prevalence of combined amplification (n = 330) and “hotspot” promoter mutation events (n = 364) was approximately 18.2% (Fig. 3) [13, 27, 28]. Our results suggest that TERT gains/amplifications (mean 15.2%, range 5.7% to 30.8%) occur more frequently than “hotspot” promoter mutations (mean 3.0%, range 0 to 11.8%).

Fig. 3
figure 3

The prevalence of TERT alterations (genomic amplifications, black; “hotspot” promoter mutations, white) in the current study, compared to previously reported cases from representative studies in the literature have been shown. The total number of cases that were profiled for amplification events (Amp) and “hotspot” promoter mutations (PM) in each study has been depicted in the figure legend on the left. TCGA The Cancer Genome Atlas

Full size image

With regards to tumors with divergent histologic features, 3 cases of adrenocortical carcinoma with focal myxoid change were identified in a 45-year-old male, 64-year-old female and a 68-year-old female, respectively. All 3 patients had amplifications of the TERT gene and had metastatic disease at diagnosis or on follow up, and at least 2 patients were dead of cancer-related causes at last follow up. On histopathologic examination, the mitotic count for these cases was: 2, 5 and 56 per 50 high power fields, necrosis was identified in 2 cases, and capsular and/or vascular invasion was identified in all. None of the cases included in our series was a rare histologic variant. Additional studies are needed to address whether TERT alterations are enriched in rarer histologic variants of adrenocortical carcinomas such as the myxoid variant.

Due to the rarity of molecular profiling of adrenocortical carcinomas, data correlating TERT status with clinicopathologic variables and outcomes is limited, with at least 1 study suggesting that whole genome doubling associated with increased TERT expression has been associated with disease progression in adrenocortical carcinomas [12]. Specifically, abstracted data from institutional and TCGA cases were reviewed to identify potential associations between TERT alteration status and clinicopathologic features such as: age at presentation, sex, tumor laterality, size and weight. While no significant trend was identified for these variables in association with TERT alterations, a female predilection was identified for combined institutional and TCGA cases (female: 106 of 166 cases, 64%) and supports what has been reported in the literature [8]. Some parameters that have been used for prognostic stratification of adrenocortical carcinomas include: mitotic frequency and Ki67 labeling index [35,36,37,38,39,40]. However, no significant association between Ki67 labeling index and mitotic activity and TERT alteration status was identified in combined institutional / TCGA cases in our study. Finally, recurrent molecular alterations identified in cases with TERT alterations included 22q (NF2/SMARCB1) and 9p21 (CDKN2A/CDKN2B) copy number losses that were identified in 24% and 16% of cases, respectively, and the significance of this finding is unclear, as well.

In contrast to molecular analysis for TERT promoter mutation detection which yields a binary result, defining a cutoff for TERT amplification (using either FISH or NGS) is somewhat arbitrary. Based on previously reported criteria, gains were defined as a fold change ≥ 1.5 and < 2.0, while amplifications were defined as a fold change ≥ 2.0 for NGS assays. Regarding FISH: ≥ 10 copies of TERT, or a TERT: 5q11 probe ratio > 2, or large clusters of TERT signal (high-level HSR type amplification) met criteria for TERT amplification. For 6 cases of adrenocortical carcinomas with TERT gene amplifications profiled using both NGS and FISH, the results of both testing modalities were concordant. It must be noted that in contrast to NGS, FISH has the advantages of being a widely available and cost-effective test, with a shorter turnaround time that may be easier to implement into routine clinical practice. However, future studies are needed to further refine both NGS and FISH-based criteria to define amplification in adrenocortical carcinomas.

In summary, the results of our study suggest that TERT alterations occur frequently in adrenocortical carcinomas and not in adrenal adenomas, and that copy number gains/amplifications are much more frequent compared to “hotspot” promoter mutations. In addition, TERT alterations may be associated with adverse outcomes such as metastatic disease and death from disease. These findings need to be further validated in prospective studies.