Novel therapeutic strategies targeting telomere maintenance mechanisms in high-risk neuroblastoma

Telomeres are regions of repetitive nucleotide sequences (TTAGGG) located at the ends of chromosomes that protect chromosomes from DNA damage, unnecessary DNA repair and fusion with other chromosomes. In normal dividing cells, with each cell replication telomeres gradually shorten, until a critical level is reached, the Hayflick limit, after which cells undergo senescence [7]. This gradual shortening of telomeres associated with cellular aging is believed to be a protective mechanism against uncontrolled growth, preventing cancer development in humans and other mammals [8]. In keeping with this, the activation of a TMM to prevent the shortening of telomeres is necessary for the continued sustained proliferation of cancer cells and hence a hallmark of cancer [9].

Telomerase, a functional ribonucleoprotein enzyme complex, maintains telomere length by adding telomeric DNA repeats at the 3′ ends of linear chromosomes. It is a reverse transcriptase that consists of a catalytic protein subunit TERT, and an essential RNA component known as human telomerase RNA (hTR), encoded by hTERC. hTR acts as a template for the synthesis of telomere DNA and is involved in the catalysis, localisation, and assembly of the telomerase holoenzyme [10, 11]. Telomerase is widely expressed in human embryos between 16 and 20 weeks gestation, however by the early neonatal period telomerase activity can no longer be detected in most somatic tissues [12]. In contrast, the majority of cancers overexpress telomerase: in a systematic analysis of 31 tumour types, over-expression of TERT was identified in 73% of all cancers [13]. This was most commonly associated with genetic alterations in either the TERT gene/promoter or TERT promoter methylation.

ALT is defined as maintenance of telomeres in the absence of telomerase activity [14]. It can be detected in 10–15% cancers overall but is particularly prevalent in tumours of mesenchymal origin [14, 15]. There is a strong association between ALT and loss of function (LoF) genetic alterations in ATRX (Alpha Thalassemia mental Retardation-X linked) in multiple malignancies, including neuroblastoma [13, 16‐18].

A number of different non-canonical homologous recombination (HR) based mechanisms have been proposed to play a role in ALT telomere maintenance [19‐22]. Furthermore, a number of studies have focused on the underlying basis for the relationship between ATRX LoF and the development of the non-canonical HR mechanisms implicated in ALT (summarised in Fig. 1). Firstly an established role of ATRX is the maintenance of genomic stability via the deposition of H3.3 into telomeric regions [24, 25]. In the absence of ATRX, G4 quadruplex structures may occur in guanine rich regions of DNA such as telomeres, resulting in stalling of replication forks, which provides a substrate for HR [26, 27]. Secondly, in the absence of ATRX, the MRN (Mre11-RAD50-Nbs1) complex is redistributed to ALT associated PML body sites where it is thought to also facilitate HR mechanisms [26]. Finally, it has been shown that the long non-coding RNA TElomeric Repeat-containing RNA (TERRA) is functionally antagonistic with ATRX [28], and therefore in the absence of ATRX, TERRA can form DNA-RNA hybrids known as R loops, that promote homology directed repair of telomeres [29]. Further confirming the role of ATRX as an ALT repressor, ATRX knockdown has been shown to induce ALT activity in cells of mesenchymal origin [30]. However, ATRX depletion does not promote ALT activity in all cell types [31, 32] suggesting that ATRX LoF alone is not sufficient to induce ALT and that additional, as yet unidentified mechanisms are also needed. Reinforcing the notion that ALT arises as a result a combination of ATRX loss and other factors, it has recently been shown that during the immortalisation process, ATRX loss results in a progressive chromatin de-compaction and a gradual induction of telomere replication dysfunction which triggers an adaptive response eventually resulting in ALT activation [33]. Furthermore the authors show that the telomere dysfunction induced by ATRX loss cannot be overcome by endogenous telomerase activity.

Genetic alterations in the histone chaperone and ATRX binding partner DAXX (death domain associated protein) have also been shown to result in ALT due to a failure to localise ATRX to PML bodies [34]. In addition genetic alterations in tumour protein 53 (TP53) and retinoblastoma 1 (RB1) have also been associated with ALT [13] however the mechanistic basis for these associations is currently unknown.

In large-scale studies telomerase expression data is often used as a surrogate biomarker for telomerase activity [13], however quantification of telomerase activity by standardised assays based on the telomere repeat amplification (TRAP) protocol is also feasible in clinical trial settings both as a predictive biomarker in tumour tissue and as pharmacodynamic biomarker in peripheral blood mononuclear cells [35].

The identification of robust biomarkers of ALT has proved more challenging. The gold standard assay for ALT is to confirm the maintenance of telomeres, in the absence of telomerase activity through successive population doublings. However, this is only practical in cell lines, non quantitative, and requires long-term cell culture [14]. Although telomere length and telomere heterogeneity are often used as biomarkers of ALT the lack of specificity of these assays for ALT is becoming increasingly apparent [14, 36, 37].

PML bodies are ubiquitous throughout the genome and responsible for diverse functions including DNA repair. In ALT cancer cells PML bodies specifically co-localise to telomeric DNA, and are thought to facilitate HR [14, 26]. ALT associated PML bodies are now a well-established biomarker of ALT activity and have the advantage that they can be visualised in formalin fixed paraffin embedded (FFPE) material [14, 38].

ALT cells are also characterised by the presence of c-circles: single stranded, telomeric circular DNA strands which are thought to provide a template for ALT telomere synthesis [14]. The c-circle assay is a rolling circle PCR amplification assay, based on the self-priming nature of c-circles. The c-circle assay requires fresh-frozen tissue but is advantageous as c-circles are quantifiable, specific for ALT and c-circle levels can be used to evaluate response to ALT targeted agents [39, 40].

Amplification of the MYCN oncogene is found in almost 40% of clinical high-risk neuroblastomas [4, 41], and is associated with up-regulation of TERT expression and telomere dysfunction [4, 42]. In an additional 23–31% of high-risk neuroblastomas, TERT is activated through chromosomal rearrangements involving 5p15.33, proximal to the TERT gene, which induces transcriptional up-regulation of TERT by juxtaposing the TERT coding sequence with strong enhancer elements [4, 5]. The third distinct sub-group, accounting for 24% of high-risk neuroblastoma cases are those with ALT [36]. Approximately half of ALT neuroblastomas are associated with somatic alterations in ATRX [4, 5, 36]. In neuroblastoma, genetic alterations in ATRX are associated with a distinct clinical phenotype including older age at diagnosis and a chronic progressive disease course [43].

Neuroblastoma harbouring a TMM is associated with a poor prognosis regardless of clinical stage [4, 5, 44‐46]. More recently it has been shown that in the presence of a TMM a concurrent mutation in a TP53 or RAS pathway associated gene is associated with an even worse prognosis [6]. Conversely, neuroblastoma with TP53 or RAS pathway mutations (including canonical ALK mutations) in the absence of a concurrent TMM are not associated with worse survival and can spontaneously regress [6]. This molecular risk stratification of neuroblastoma, as described by Ackermann et al. [6] is summarised in Fig. 2.

Although the somatic alterations driving TMM’s in the majority of neuroblastoma are well defined, high telomerase expression can occur in the absence of either a TERT translocation or MYCN amplification [6]. Also in the absence of an ATRX alteration the underlying drivers of ALT neuroblastoma are currently unknown, and somatic alterations in other genes, known to be associated with ALT in other malignancies are rarely found in neuroblastoma [47]. Numerous additional mechansims for telomere maintenance in neuroblastoma have been suggested. A study has identified interstitial telomeric sequences at sites of unbalanced translocations in neuroblastoma cell lines and postulated that these may contribute to a defective telomere maintenance pathway [48]. Another possibility is that genetic predisposition may contribute to the development of TMMs. In keeping with this, six common single nucleotide polymorphisms known to be associated with telomere lengthening have been found to be associated with an increased risk of neuroblastoma [49]. In keeping with the hypothesis that additional underlying predisposing factors may contribute to TMMs, a substantial increase in telomeric DNA damage and active telomere trimming has been described consistently across all high-risk neuroblastomas, regardless of telomerase or ALT status [50].

In summary, data is accumulating to support the hypothesis that TMMs are the common defining molecular feature of aggressive disease in the majority of children with clinical high-risk neuroblastoma. Therefore, both telomerase and ALT represent attractive targets for the development of novel therapeutic strategies with the potential to benefit a significant proportion of high-risk neuroblastoma patients. Here, we review the current pre-clinical and clinical research focused on targeting telomere maintenance with a specific focus on the potential relevance to patients with neuroblastoma.

Imetelstat (GRN163L) is a competitive telomerase inhibitor with a complementary structure to the template region of the RNA component of telomerase that binds to and blocks the active site of the enzyme [51]. Confirmation of target inhibition and pre-clinical efficacy of imetelstat has been demonstrated in multiple cancer subtypes [52‐54].

A phase I trial of imetelstat in 20 children with refractory or recurrent solid tumours demonstrated telomerase inhibition, which was sustained through to day 8 of cycle 1 of therapy [55]. Two of 16 patients had a partial response, however no responses were seen in any of the 6 patients with neuroblastoma enrolled on the trial. The main dose limiting toxicities were neutropenia, thrombocytopenia, and lymphopenia. A Phase II study on imetelstat in children with recurrent or refractory central nervous system malignancies was also associated with significant haematological toxicity and was discontinued after two children died of intra-tumoral haemorrhage [35], thus paediatric development for this compound has subsequently been discontinued.

BIBR1532 is a potent synthetic, non-nucleoside telomerase inhibitor [56]. However like imetelstat there are significant concerns regarding toxicity [57], and this agent has not yet been evaluated in clinical trials.

Sodium metaarsenite (KML001) binds to telomeric sequences, displacing hTERT from the nucleus into the cytoplasm [58] and has been shown to be cytotoxic in neuroblastoma cells in vitro [59]. In a phase I trial in adults with advanced solid tumours, objective responses to KML001 and cisplatin were seen in four out of 18 patients, however this trial was also discontinued due to toxicity [60].

Telomestatin stabilises G-quadruplexes, which in turn, inhibits telomerase activity [61]. This compound has been shown to induce apoptosis in-vitro in telomerase expressing neuroblastoma cell lines [62] but is not yet in clinical development.

6-thio-2′-deoxyguanosine (6-thio-dG) is a nucleoside analogue, which in telomerase active cells is recognised by telomerase and incorporated into telomeres resulting in telomere dysfunction [63]. Pre-clinical efficacy has been demonstrated in melanoma, non-small cell lung cancer and paediatric brain tumour models [64‐66]. Although 6-thio-dG has not yet been evaluated in clinical trials, due to the novel mechanism of action of the compound it is thought to be less toxic than traditional telomerase inhibitors [67]. 6-thio-dG has been shown to be effective in-vivo in neuroblastoma models with TERT activation. However, the response in MYCN amplified xenografts was mixed, likely reflective of the additional oncogenic pathways activated by MYCN [47]. This is a priority compound for further development for neuroblastoma (Table 1).

XAV939 is an inhibitor of tankyrase, a positive regulator or telomerase. It has been shown to induce apoptosis in the telomerase expressing neuroblastoma cell line SH-SY5Y [70].

In some other paediatric malignancies, ALT (identified by the presence of ALT associated PML bodies) and telomerase activation have been shown to co-exist in the same tumour [88, 89]. In neuroblastoma, intra-tumoural diversity of telomere length in individual tumours has been identified using quantitative telomere fluorescence in-situ hybridisation (FISH) [90], however this is not a sufficiently sensitive or specific biomarker of ALT activity [14, 36]. In fact, one study has identified a subset of neuroblastomas with extremely long telomeres in the absence of either telomerase activity or ALT (detected by c-circle assay and ALT associated PML bodies) [36]. Further studies have shown that although TERT alterations and MYCN amplification do co-exist in a small proportion of cases, there is no overlap between the telomerase expressing and ALT positive neuroblastoma [47]. This is in keeping with a recent publication also showing that ATRX mutations and MYCN amplification are synthetically lethal in neuroblastoma [91]. Taken together, the evidence in neuroblastoma to date is that the presence of either ALT or telomerase activation is associated with differing distinct genetic drivers and occurs in mutually exclusive nature [4, 5].

Although data thus far indicates that subgroups of neuroblastoma are driven by either telomerase or ALT activation, it is highly likely that a selective pressure targeting one TMM will support the emergence of an alternative mechanism. In multiple other cancer subtypes it has been shown that long-term telomerase inhibition results in the emergence of features consistent with ALT [92‐94]. Conversely, with the development of ALT targeted therapeutics it is probable that an up-regulation of telomerase will be seen. However, encouragingly, following ATRX LoF, it appears that ALT is a necessary adaption for cancer cell survival, and that reactivation of telomerase activity cannot overcome endogenous telomere dysfunction [33].

Finally although it has been shown that activation of a TMM is the key determinant of poor outcome in neuroblastoma, the key drivers of TMM’s in neuroblastoma; MYCN amplification and genetic alterations in ATRX are also associated with distinct patterns of wider transcriptional activation which drive malignant transformation [91, 95, 96]. Furthermore it is known that the co-occurrence of RAS/TP53 pathway alterations with a TMM is associated with a particularly dismal outcome [6] and accordingly, alterations in the RAS and TP53 pathways are enriched at the time of neuroblastoma relapse [97, 98]. Therefore, TMM targeting strategies will only be beneficial when given in combination with agents targeting these key pathways and the evaluation of combination therapies is urgently needed.