Telomerase Expression by Aberrant Methylation of the TERT Promoter in Melanoma Arising in Giant Congenital Nevi

expression by aberrant methylation of the TERT promoter in melanoma arising in giant congenital nevi.

telomerase-expressing pediatric brain tumors but not in telomerase-negative normal brain tissues or low-grade tumors (Castelo-Branco et al., 2013).
We previously showed that an aggressive form of pediatric melanoma developing within giant congenital nevi (GCN) retains the wild-type TERT promoter (Lu et al., 2015). To determine whether epigenetic modifications may play a role in telomerase expression in this melanoma subtype, we analyzed the DNA methylation profile of a CpG-rich region of the TERT promoter, shown previously to be differentially methylated between normal and malignant tissues (Castelo-Branco et al., 2013), in 13 melanomas (3 arising in GCN; 7 conventional; and 3 spitzoid) and 10 benign or borderline melanocytic tumors (1 GCN; 3 GCN with nodular proliferation; and 6 borderline spitzoid melanocytic neoplasms) from 23 pediatric and adult patients. The human investigations were performed after approval by the local institutional review boards. Written, informed patient consent was waived because the research involved no more than minimal risk to the subjects. The status of the TERT promoter, BRAF and NRAS mutations, and kinase fusions was available for the spitzoid tumors and a subset of melanoma samples from our prior studies (Lee et al., 2015;Lu et al., 2015).
Next, we evaluated the association of TERT promoter CpG methylation with telomerase expression by TERT mRNA in situ hybridization (ISH) and by gene expression analysis (Methods, Supplementary Material online). TERT mRNA ISH revealed distinct, intracellular punctate signals in melanomas arising in GCN (Figure 2c and 2f) but not in the proliferative nodules in GCN (Figures 2i and 2l). The TERT promoter methylation level was calculated as the log2 ratio of the total number of methylated Cs versus the total number of unmethylated Cs in the sequenced region [logit (B-value)]. The TERT expression level was measured by using RNA sequencing data available for a subset of samples. An association analysis revealed a strong correlation between TERT promoter methylation and TERT expression level (P = 0.0422, adjusted r 2 = 0.5145; Supplementary Figure S2 online).
Our data demonstrate that epigenetic modification through TERT promoter CpG methylation is an alternative pathway for TERT reactivation in melanoma. Although epigenetic remodeling by promoter methylation is generally considered a signature of gene silencing, TERT expression is paradoxically increased by promoter methylation (Guilleret and Benhattar, 2004). Although the exact mechanism underlying CpG DNA methylation in TERT upregulation is not known, one possible mechanism is by inhibiting transcriptional repressors such as CTCF (Renaud et al., 2007), SIN3A, or MAZ (Xu et al., 2013) from binding to the target DNA-binding sites in the region ( Figure S1). Also, even when the promoter is largely methylated, a small region of the core promoter upstream of the transcription start site remains unmethylated to allow for the continued transcriptional activity of TERT (Renaud et al., 2007;Zinn et al., 2007).
Individuals with GCN are at increased risk for developing melanoma (Figures 2a-b and 2de) that occurs most frequently in the first decade of life (Bittencourt et al., 2000). A much more frequent change in these nevi than melanoma is the development of clonal proliferations often in the form of nodules (Figures 2g-h and 2j-k), which may suggest or mimic melanoma on clinical or histological grounds (Yelamos et al., 2014). The differential pattern for TERT promoter methylation and telomerase expression between melanomas in GCN and proliferative nodules demonstrated in our study is consistent with the benign clinical course of proliferative nodules compared with the invariably aggressive behavior of melanoma arising in GCN. Further studies in a larger number of patients are needed to determine the potential diagnostic value of TERT promoter methylation assays for ambiguous proliferative lesions within GCN.
In our cohort, TERT promoter hypermethylation or promoter mutations occurred in all melanoma samples but in none of the the benign or low-grade melanocytic lesions, suggesting that a panel incorporating TERT promoter methylation and mutation assays may help discriminate between benign/borderline and overtly malignant melanocytic neoplasms. Future studies need to assess the potential use of these assays for diagnostic or prognostic purposes in the clinic.
In summary, we demonstrate that in subsets of malignant melanoma, TERT is upregulated epigenetically by a methylation-dependent mechanism. These findings have potential therapeutic implications. TERT promoter CpG hypermethylation is a reversible phenomenon. Treatment with DNA demethylating agents reduced TERT expression and telomerase activity in telomerase-positive cell lines (Guilleret and Benhattar, 2003;Renaud et al., 2007). Together, these findings provide a rationale for developing a therapeutic strategy in preclinical studies through epigenetic modifications at TERT promoter regulatory sites in melanomas with the CpG island methylator phenotype. Association of the mutational status and the methylation profile of the TERT promoter with disease characteristics and outcome data for 23 patients with melanocytic tumors. The 26 CpG sites were aberrantly methylated in wild-type TERT promoter melanomas (the last 4 rows) but were predominantly unmethylated in low-grade or benign melanocytic tumors (atypical spitzoid neoplasms and GCN with proliferative nodules) and in mutant TERT promoter melanomas. Methylation panel color code: M, methylated; PM, partially methylated; PUM, partially unmethylated; UM, unmethylated; NA, not available. Photomicrographs (H&E-stained) and TERT mRNA in situ hybridization (ISH) for 2 melanomas in GCN (2 top panels) and 2 proliferative nodules in GCN (2 bottom panels). mRNA ISH shows numerous high-resolution red intracellular punctate signals in malignant melanocytes (c and f) and no signals above the background level in melanocytes of proliferative nodules (i and l). Scale bars= 1000 μm (a) and 50 μm (b).