Telomeres are nucleoprotein structures at the ends of chromosomes (1) that contain 5-10 kilobases of the canonical hexanucleotide (5’-TTAGGG-3’) repeat sequence encased in sheltering proteins (2). This complex protects genomic DNA from replicative erosion (3, 4), shields the ends of chromosomes from aberrant fusion (5), and prevents DNA damage response (DDR) elements from errantly recognizing genomic DNA (2, 6, 7). Approximately 85-90% of all cancers achieve replicative immortality by utilizing the telomere maintenance mechanism (TMM) telomerase (TA), a ribonucleotide reverse transcriptase (8, 9). The remaining 10-15% of cancer cases (~250,000 U.S. patients annually) use a non-telomerase TMM called alternative lengthening of telomeres (ALT) (10, 11).

Incidence of ALT varies amongst sarcomas ( Tables 1 , 2 ), with the majority of cases arising from tissues of mesenchymal or neuroepithelial origin (10, 22). Cancers with an estimated ALT frequency >40% include osteosarcoma (OS), diffuse and anaplastic astrocytomas, undifferentiated pleomorphic sarcomas, and pediatric grade 4 glioblastoma multiforme (10, 14). Previously reported patient sample screenings have demonstrated a broad range of ALT frequency amongst pediatric cancers, from 0% in Ewing’s Family Sarcoma (EFS), up to 85% in OS (10, 12, 14). Recently, there have been calls for assessing patient samples with currently available ALT biomarkers to confirm historically reported ALT frequencies, especially for OS (23).

The hallmarks of the ALT phenotype include absence of TA activity (TERT mRNA expression provides a suitable surrogate for TA activity) (24) with the presence of high telomere content and heterogenous telomere length (25, 26), non-canonical telomere variant repeats (27), extra-chromosomal telomeric repeats (28), ALT-associated PML bodies (APBs) (14, 29–31), ultrabright telomere foci by FISH (10), and telomeric DNA C-circles ( Tables 1 , 2 ) (32). These characteristic markers have been used to screen tumor sample cohorts to determine the frequency of ALT among various tumor histologies (10, 13, 14, 18, 33, 34). Each of the methods has advantages and disadvantages ( Table 3 ).

Historically, ALT has been identified by the telomerase repeated amplification protocol (TRAP) assay to demonstrate low TA activity (35) and/or low TERT mRNA expression, since TA is mutually exclusive to ALT (36). Telomere content and heterogeneity have been evaluated by telomere restriction fragment (TRF) analysis (37) and telomere fluorescence in situ hybridization for ultra-bright telomeric foci (UTF). UTF was combined with immunofluorescence (IF) of the PML protein, which was discovered to co-localize with telomeres in ALT samples, to detect APBs, yielding an additional ALT feature (38). Recently C-circles, circular self-primed telomeric DNA repeats, have been shown to be a sensitive and specific biomarker for ALT in tumors (32) that also circulate in patient plasma, potentially increasing the clinical utility of C-circles as a biomarker (39, 40).

After genomic DNA is extracted ( Figure 1 ) from fresh frozen tumor, or plasma, C-circles can be enriched and subsequently detected by blot or real-time PCR (41, 42). First, C-circles are selectively amplified by ϕ-29 DNA polymerase via isothermal rolling-circle amplification ( Figure 1 ) (41, 43, 44), which enriches the partially double-stranded telomeric DNA, termed C-circles, when compared to a reaction without ϕ-29. Subsequently, the telomeric signals can be compared by real-time PCR ( Figure 1 ) for the ϕ-29 and no ϕ-29 reactions ( Supplementary Figures S1A, C ), which is then normalized to a single copy gene (e.g. VAV2) for the same ϕ-29 and no ϕ-29 reactions ( Supplementary Figures S1B, D ) (41, 42). This unique molecular diagnostic assay allows for high-throughput screening of DNA from fresh frozen tumor and plasma samples with as little as 1 ng of template input (14, 41). Herein, we sought to assess the frequency of ALT in pediatric sarcomas using the real-time PCR CCA on DNA samples extracted from fresh frozen tumor.

The prognostic value of ALT, and other TMM, is gaining traction (23, 33, 53), including in veterinary care (54). Studies have linked high telomerase expression with exceptionally aggressive tumors that can result in rapid progression and poor clinical outcomes (15, 33). By contrast, ALT has been associated with indolent disease progression; yet, patients with various tumor types have been observed to have a worse overall survival (23, 28, 33, 55). Recently, we observed high amounts of ATM kinase activation (which promotes chemotherapy resistance) in patient-derived neuroblastoma cell lines (PDCLs) and patient-derived xenografts (PDXs) (48), and also in PDCLs of other histologies (rhabdomyosarcoma, osteogenic sarcoma, triple negative breast cancer, and colorectal cancer) that have the ALT phenotype (47). We have also observed that certain clinical stage drugs (an ATM kinase inhibitor (48) and a p53 reactivator (47)) are active in reversing chemotherapy resistance in ALT PDCLs and PDXs. Thus, robust identification of ALT has the potential to be a prognostic biomarker and a companion diagnostic for ALT-targeted therapies.

Generally, ALT is activated by loss-of-function (LOF) genetic alterations in the chromatin remodelers α thalassemia-mental retardation, X linked (ATRX) (33) and death domain-associated protein 6 (DAXX) (56). ATRX inactivating mutations are commonly observed among different tumor types, while DAXX mutations are primarily associated with pancreatic neuroendocrine tumors (PanNETs) (56). ALT is less frequently associated with LOF alterations in H3F3A (57, 58) and SMARCAL1 mutated tumors (59, 60). Previous studies have used these genomic alterations as proxies to identify ALT, but depending on histology, as many as ½ ALT cancers can be wild-type for ATRX or DAXX (41, 61).

C-circles, TERT expression, high telomere content with heterogenous telomere length, and APBs have been used to screen sample sets to establish ALT frequencies amongst sarcomas; however, each of these techniques have their own advantages and disadvantages. Relatively fragile, C-circles can be degraded by excess freeze-thaw cycles, prolonged vortexing, and formalin-fixing; thus, proper sample handling and storage are required (62). Recently, ALT tumors have been shown to protect C-circles from nuclease degradation in the blood by releasing C-circles within exosomes, which may provide a non-invasive blood-based biomarker for the detection and monitoring of ALT tumors in vivo (40). Although there is no standardized method for determining ALT status (28), C-circles are the only known molecule specific to ALT (40), and the molecularly based real-time PCR C-circle assay can utilize DNA that has been isolated for sequencing; thus, it is readily translatable to the clinical laboratory, and it’s for these reasons that we selected this approach (28, 32, 42, 61).

We observed no CC+ EFS cases, which is likely due to the activation of TERT by EFS fusion proteins (39). The ALT phenotype is known to be essentially exclusive to TERT activation (32, 39, 40). In OS patients, expression of TERT has been shown to portend an unfavorable clinical prognosis (15); however, stage and clinical outcomes of ALT cases were shown to be equivalent to TA cases (16), but, the ALT phenotype provides a potentially targetable mechanism present in the majority of OS patients, some of which have poor clinical outcomes (47, 63, 64).

ALT is also known to occur in RMS (14), the most common pediatric soft tissue sarcoma (65). Classically, pediatric RMS cases were generally categorized histologically as ERMS, which was linked with better prognoses, or ARMS, which was associated with poor clinical outcomes (66). Further, molecular identification of PAX3, or PAX7, fusions with forkhead box protein O1 (FOXO1), is currently considered the preferred method of distinguishing the latter from the former (67). Instead of histologic criteria, which are inexact, the fusion status identifies ARMS and ERMS, which are FP and FN, respectivly (68).

The tested RMS samples were from banked DNA extracted from fresh frozen tissue; thus, it is possible that the age of the samples, or excess freeze-thaw cycles could have contributed to the lower ALT frequency, due to the degradation of C-circles (41). APB analysis from a previous study (14) has the advantage of using FFPE material, which enables distinguishing of tumor cells from stromal tissue; however, the APB assay is very labor intensive, not all ALT samples have APBs (33), and C-circles have been postulated to be more specific than other ALT markers (32). Ideally, future studies should evaluate the various methods for detecting ALT in the same histology within the same patient sample cohort, since each ALT marker is not necessarily present in every ALT sample or tumor model (14, 33). However, our data suggests that the real-time PCR CCA can identify ALT in sarcomas, and it has potential as a companion diagnostic assay for ALT targeted therapies in RMS, and especially OS, patient populations.