The Solo Play of TERT Promoter Mutations

The reactivation of telomerase reverse transcriptase (TERT) protein is the principal mechanism of telomere maintenance in cancer cells. Mutations in the TERT promoter (TERTp) are a common mechanism of TERT reactivation in many solid cancers, particularly those originating from slow-replicating tissues. They are associated with increased TERT levels, telomere stabilization, and cell immortalization and proliferation. Much effort has been invested in recent years in characterizing their prevalence in different cancers and their potential as biomarkers for tumor stratification, as well as assessing their molecular mechanism of action, but much remains to be understood. Notably, they appear late in cell transformation and are mutually exclusive with each other as well as with other telomere maintenance mechanisms, indicative of overlapping selective advantages and of a strict regulation of TERT expression levels. In this review, we summarized the latest literature on the role and prevalence of TERTp mutations across different cancer types, highlighting their biased distribution. We then discussed the need to maintain TERT levels at sufficient levels to immortalize cells and promote proliferation while remaining within cell sustainability levels. A better understanding of TERT regulation is crucial when considering its use as a possible target in antitumor strategies.


Telomerase Reverse Transcriptase Promoter (TERTp) Mutations
TERTp mutations were first described in congenital and sporadic melanoma in 2013 [49,50]. Subsequent large-scale cohort studies together with seminal mechanistic studies both ascertained the TERTp mutation prevalence in many other forms of cancer and characterized their mode of action.
All of these TERTp mutations (at positions −146, −124, −57, and −139/−138) create novel Ets/TCF transcription factor binding sites. The Ets/TCF transcription factors bind to GGAA motifs (or TTCC on the opposite strand). The 30 members of the Ets/TCF-family transcription factors are important contributors to oncogenesis and include Ets-1, Ets-2, and GA binding protein (GABP) [68]. So far, GABP has been reported to selectively bind the −124 C>T and −146 C>T mutations in GBM,

Telomerase Reverse Transcriptase Promoter (TERTp) Mutations
TERTp mutations were first described in congenital and sporadic melanoma in 2013 [49,50]. Subsequent large-scale cohort studies together with seminal mechanistic studies both ascertained the TERTp mutation prevalence in many other forms of cancer and characterized their mode of action.

Cancer Distribution of TERTp Mutations
The clinicopathological association of TERTp mutations is cancer-dependent. It is a consideration for fine tumor stratification and orientation of patients towards personalized treatments, and provides insight into the process of cellular transformation.
TERTp mutations are independently associated with older age, late clinical stage, poor prognosis, and shorter overall survival (OS) in GBM/glioma and IDH-wt astrocytoma patients. The presence of TERTp mutations alone is associated with a worse prognosis than TERTp mutations together with IDH-mutations [4,60,64,65,77,[79][80][81]84,85,112]. Conversely, GBM patients with ALT and no TERTp mutations have longer OS than patients with TERTp mutations only [77,112,113]. In terms of treatment, Grade II and III IDH-wt CNS tumors generally respond to adjuvant radiation and chemotherapy with temozolomide (TMZ). However, the presence of TERTp mutations decreases sensitivity to genotoxic therapies. It has therefore been proposed to use TERTp mutations to further stratify IDH-wt Grade II and III gliomas into subgroups to orient treatment [60,81,114].
Consistent with their UV-induced origin in skin cancers, TERTp mutations are also highly prevalent at sun-exposed sites in non-melanoma squamous cell (50%) and basal cell carcinomas (46.2%, range 38-74%), the most common skin tumor [55,89,90]. TERTp mutations display unique features in melanoma and non-melanoma cancers. First, −146 C>T and −124 C>T occur with similar frequencies in contrast to all other cancers, where −124 C>T is by far the most prevalent mutation (Table 1). Second, −139/−138 CC>TT and −125/−124 CC>TT tandem mutations are often reported. Third, TERTp mutations were detected in 9/10 melanomas with ALT in one study [117] and together (−124 C>T + −146 C>T) in two patients with BCC in another study [89], indicating that more than one telomere maintenance mechanism can, unusually, coexist in skin cancers.

Urothelial Bladder Cancer
TERTp mutations have been detected in 64.6% (range 29.5-100%) of urothelial bladder and upper urinary tract cancers. They are the most common somatic lesions in this cancer type [52,57,61,77,92,94,118,119]. They have been associated with reduced survival, disease recurrence, and distal metastases [61,118,119], although there appears to be no difference between early-and late-stage patients [52,94].
TERTp mutations appear to be more frequent in HCV-associated HCC [62,77,95,96,122] and less frequent or excluded from HBV-associated HCC [62,96,121,122], although this remains controversial [63,77,95]. HBV DNA insertion in the TERTp is a recurrent mechanism of TERT transcriptional reactivation in HBV-associated HCC [34,123,124], and a genetic screen of TERT in HCC found TERTp mutations to be mutually exclusive with HBV integration, TERT CNVs, and ATRX mutations [121].

The rs2853669 Polymorphism
Among TERT polymorphisms, a common polymorphism (rs2853669 A>G) which disrupts a pre-existing Ets/TCF binding site located 245 bp upstream of the TERT TSS has been reported to modify the effect of TERTp mutations. It decreases TERT transcription in vitro and reverses TERT upregulation by TERTp mutations [56,61,81,85,130]. Controversial clinical impacts have been reported, from a beneficial effect on OS and limited tumor recurrence in TERTp-mutated urothelial bladder cancer, renal clear cell carcinoma, melanoma, and GBM [56,61,81,85,116,131], to unchanged or worsened clinical outcome in GBM, melanoma, or differentiated thyroid carcinomas [64,65,84,91,102,103]. In HCC, the rs2853669 polymorphism in combination with TERTp mutations has been associated with decreased OS and DSF, and increased TERTp methylation and expression [47]. Possible reasons for these conflicting reports could be homozygosity versus heterozygosity of the variant, or its occurrence on the same allele as TERTp mutations. Further studies are needed to assess the relevance of screening for this polymorphism for prognostic and treatment purposes.

Cancer Bias of TERTp Mutations
TERTp mutations have been recorded in individuals of Caucasian, African, and Asian descent, with no race-related bias. The −124 C>T mutation has an overwhelmingly higher prevalence than the −146 C>T mutation in all cancers, with the exception of skin cancers, where both hotspots are mutated with comparable frequencies (Figure 2 and Table 1). Although both −124C>T and −146C>T mutations generate identical sequences, enable binding of GABPA, and are equally efficient in increasing TERT transcription in vitro [57,69], in vivo, the −124 C>T mutation was associated with higher TERT mRNA in GBM [57,112]. This would suggest that the Ets/TCF binding site at position −124 provides a more favorable or accessible hotspot for the transcriptional machinery [109]. The overrepresentation of the −146 C>T mutation in skin cancers hints at different etiologies of TERTp mutations. TERTp mutations in melanoma and non-melanoma skin cancers have been attributed to UV damage [49,51,55,[88][89][90][91]116], which triggers C→T transitions at CC dinucleotides [55,127]. Nevertheless, C→T transitions where C is preceded by C also conform to the preferred target of Apolipoprotein B mRNA Editing Catalytic Polypeptide-like (APOBEC)3A/B de-aminations and to aging mutations [127,133]. APOBEC3 mutations are highly prevalent in ovarian and HPV-associated cervical and oral SCC [125][126][127], as well as in HCC and in cirrhotic lesions [121,134]. A role for APOBEC and aging-associated de-aminations is consistent with potentially increased accessibility of the −124 position to DNA binding proteins and with the association of TERTp mutations with older age at diagnosis in GBM, melanoma, and PTC [52,57,60,63,64,77,79,80,82,86,88,98,[100][101][102]. These observations therefore raise the possibility that UV-driven lesions account for TERTp mutations in skin cancers, while APOBEC and age-driven de-aminations account for the −124 C>T mutation in other cancers. Further epidemiological and mechanistic studies are needed to shed light on this point.

Cancer Bias of TERTp Mutations
TERTp mutations have been recorded in individuals of Caucasian, African, and Asian descent, with no race-related bias. The −124 C>T mutation has an overwhelmingly higher prevalence than the −146 C>T mutation in all cancers, with the exception of skin cancers, where both hotspots are mutated with comparable frequencies (Figure 2 and Table 1). Although both −124C>T and −146C>T mutations generate identical sequences, enable binding of GABPA, and are equally efficient in increasing TERT transcription in vitro [57,69], in vivo, the −124 C>T mutation was associated with higher TERT mRNA in GBM [57,112]. This would suggest that the Ets/TCF binding site at position −124 provides a more favorable or accessible hotspot for the transcriptional machinery [109]. The overrepresentation of the −146 C>T mutation in skin cancers hints at different etiologies of TERTp mutations. TERTp mutations in melanoma and non-melanoma skin cancers have been attributed to UV damage [49,51,55,[88][89][90][91]116], which triggers CT transitions at CC dinucleotides [55,127]. Nevertheless, CT transitions where C is preceded by C also conform to the preferred target of Apolipoprotein B mRNA Editing Catalytic Polypeptide-like (APOBEC)3A/B de-aminations and to aging mutations [127,133]. APOBEC3 mutations are highly prevalent in ovarian and HPV-associated cervical and oral SCC [125][126][127], as well as in HCC and in cirrhotic lesions [121,134]. A role for APOBEC and aging-associated de-aminations is consistent with potentially increased accessibility of the −124 position to DNA binding proteins and with the association of TERTp mutations with older age at diagnosis in GBM, melanoma, and PTC [52,57,60,63,64,77,79,80,82,86,88,98,[100][101][102]. These observations therefore raise the possibility that UV-driven lesions account for TERTp mutations in skin cancers, while APOBEC and age-driven de-aminations account for the −124 C>T mutation in other cancers. Further epidemiological and mechanistic studies are needed to shed light on this point.
The −139/−138 CC>TT tandem mutation is very infrequent, limited to skin cancers, and has been associated with lower DFS. This tandem mutation has been suggested to favor chromosomal instability [51].  The −139/−138 CC>TT tandem mutation is very infrequent, limited to skin cancers, and has been associated with lower DFS. This tandem mutation has been suggested to favor chromosomal instability [51].

Exclusiveness of TERTp Mutations
Aside from non-melanoma skin cancers [90], TERTp mutations are mostly monoallelic. This suggests that TERT reactivation on one allele is probably sufficient to ensure telomere maintenance or elongation in cancer cells [54]. In line with this observation, TERTp mutations appear to be mutually exclusive [50]. Likewise, TERTp mutations are generally absent from cancers where telomere elongation is ensured by ALT [77,79,80,98] or TERT copy-number duplications [38,121]. TERTp mutations are also less frequent in cancers where viral transformation or viral oncogenes reactivate TERT transcription, such as HBV-DNA or high-risk HPV16/18 E6 [30,32,33,36,37,62,95,96,121,122]. These observations reinforce the concept that, despite some exceptions [38,89,111,117], tumors generally rely on one mechanism for telomere maintenance. The reasons for such selectivity remain speculative to date. One possible explanation is that there is a threshold for TERT expression, above which the biological advantage is lost.
Consistent with this view, Phosphatidyl Inositol Kinase 3 (PIK3) CA and PIK3 Receptor 1 (PIK3R1) mutations are recorded in 50% of GBM with wt TERTp and tend to be mutually exclusive with TERTp mutations in ovarian clear cell carcinoma [79,86,132]. The PIK3CA/Akt signaling pathway is involved in cellular self-renewal in embryonic stem cells and cancer stem cells [135], as well as in TERT Ser227 and Ser824 phosphorylation, subsequent nuclear translocation, and cellular transformation [25][26][27][28]. Mutual exclusion of PIK3CA and TERTp mutations suggests that activation of the PIK3CA/Akt pathway or of TERT confer cells a similar growth and proliferative advantage. In the absence of TERT reactivation, other telomere maintenance mechanisms, such as ALT, can achieve immortalization [27]. Indeed, TERT also contributes to cell survival and proliferation through telomere-independent mechanisms; it facilitates Wnt/β-catenin-dependent [136,137], c-myc-dependent [138,139], and NF-κB-dependent gene transcription [140,141], thereby sustaining both oncogenic signaling pathways and its own transcription in a feedforward loop [29,142]. It also regulates methylation [48,143] and DNA damage responses [144,145], and protects cells from Endoplasmic Reticulum (ER) stress and apoptosis by buffering Reactive Oxygen Species (ROS) and modulating mitochondrial function [145][146][147][148][149][150][151]. It is highly likely that TERT homeostasis is also tuned by these functions within a given tumor type and microenvironment, and by related metabolic alterations that need to be preserved.

Discussion
Hints for a model come from the observation that overall, TERTp mutations are associated with late-stage disease in GBM, melanoma, urothelial, and thyroid carcinoma [49,52,60,61,66,85,98,100,101,[103][104][105]112,118] and with the last steps of hepatocellular transformation [62,95]. They often occur with or after mutations in pathways associated with cell growth and proliferation. In GBM, TERTp mutations coexist with EGFR amplification [64,77,111], and in urothelial bladder carcinoma, they are associated with FGFR3 (Fibroblast Growth Factor Receptor 3) mutations [61,94]. In~50% of melanoma, urothelial, and thyroid cancers, TERTp mutations coexist with the common BRAF-V600E mutation [52,88,89,105,106,108,116,152]. GFR and BRAF/RAS kinases control the MAPK and PI3K-Akt pathways that lead to cell growth, survival, and angiogenesis. Constitutive activation of the GFR/FGFR-BRAF/RAS pathway leads to constitutive cell growth and division [153]. Mutations in these oncogenes are often detectable in low-grade tumors and probably precede TERTp mutations [22,61,77,112]. The picture is even more clear-cut in HCC, where mutations in β-catenin (CTNNB1) neatly precede TERTp mutations during the process of malignant transformation [62,95,120]. β-catenin is involved in cell adhesion and interacts with Wnt, promoting cell growth and division. The proliferative advantage conferred by driver mutations in these pathways leads to accelerated telomere erosion. Accordingly, most tumors display telomere dysfunction and shortened telomeres, which leads to chromosome instability [10,22,61,66,98,112,115]. In this scenario, TERT reactivation regenerates telomeres sufficiently to maintain them above the critical threshold and to stabilize the tumor genome [3,18,145]. This interpretation is consistent with the association of TERTp mutations with shortened telomeres and with age as in PTC, melanoma, and GBM/glioma, since cells from younger patients or with sufficiently long telomeres do not need to rely on telomerase reactivation to overcome telomeric crisis [10,29,57,77,85,98,101,115]. Partial telomere healing is coherent with a modest increase in TERT expression (2-to 4-fold) and with a single genetic mechanism of telomere elongation. It likely reflects an exquisite balance between escape from apoptosis resulting from telomere attrition and genomic instability, and cell sustainability in terms of oxygen and nutrient supplies.
Intriguingly, it was recently reported that GABPA controls the cell cycle and induces cell differentiation, thus acting as a tumor suppressor regulating cell proliferation, stemness, and adhesion. It decreased tumor invasiveness and distal metastases in PTC, HCC, and bladder carcinoma [154][155][156]. GABPA levels were decreased and even negatively associated with TERT expression in PTC [154][155][156]. One possible explanation is that other Ets/TCF family transcription factors bind TERTp mutations. Alternatively, the decrease in GABPA expression may follow rather than precede TERTp mutations. In this case, it would be a cellular adaptation which confers a selective advantage to TERTp-mutated (and GFR/BRAF/RAS-mutated) cells by containing TERT reactivation within sustainable limits. Decreased GABPA could also be an adaptation to the TERT-induced proliferation, stemness, and invasion to avoid contradictory signals. Further studies establishing the order of emergence of these mutations would be needed to shed light on this matter.
Taken together, these observations point to a fine tuning of TERT homeostasis and suggest that there is a narrow kinetic and quantitative window for TERT expression. Below that window, cells succumb to telomere crisis and DNA damage. Above that window, cells succumb to overwhelming genetic alterations or metabolic needs. This frailty could be exploited through strategies aiming to push cells either way beyond the threshold of TERT tolerability.

Concluding Remarks
TERTp mutations have only been described recently; however, they have prompted an impressive number of studies which draw a comprehensive picture of their prevalence across cancers, as well as providing clues on their mechanisms of action and their associated constraints. They have been proposed as potential biomarkers with predictive and treatment-orienting value. However, more structured studies are needed to validate their clinical potential, particularly since they appear at different stages in different malignancies, ranging from preneoplastic cirrhotic lesions to late stage GBM or melanoma with distal metastases. Cancer cells only require one mechanism of telomere maintenance. This underscores the key role of telomere stabilization in the process of transformation, as well as the necessity of maintaining an exquisitely balanced TERT homeostasis to achieve tumor cell selection, adaptation, and sustainability. TERT is a target of choice in antitumor strategies due to its reactivation in numerous cancers. A better understanding of TERT regulation, homeostasis, and functions could help to overcome the shortcomings of prior genetic and immunotherapy-based approaches targeting TERT.

Acknowledgments:
The authours are deeply grateful to Dr Jonathan D Turner for his thorough revision of this manuscript and for his insightful advise.

Conflicts of Interest:
The authors declare no conflict of interest.