Identification of patient-derived glioblastoma stem cell (GSC) lines with the alternative lengthening of telomeres phenotype

Glioblastoma multiforme (GBM) is an aggressive brain tumor with a poor overall prognosis. Current standard of care involves surgical resection followed by adjuvant treatment with radiation (RT), temozolomide, and tumor treating fields (TTF) [13]. Despite this aggressive treatment modality, median overall survival is approximately 15 months. Telomeres are terminal DNA elements found at eukaryotic chromosomal ends consisting of hexagonal repeats of (TTAGGG)_n which are essential for maintaining genomic stability [1]. To maintain telomere length and circumvent the end-replication problem, most cancer cells express telomerase [8]. Telomerase is composed of two subunits: a catalytic component with reverse-transcriptase activity encoded by the gene TERT, and an 11 base-pair RNA template encoded by the gene TERC [11]. Mutations in the promoter region for TERT occur in approximately 60–80% of GBM, leading to increased telomerase activity and enabling replicative immortality [10]. A defining feature of anaplastic astrocytomas and a small fraction of secondary GBM, is activation of a telomerase-independent alternative lengthening of telomeres (ALT) mechanism, driven by homologous recombination (HR) machinery [7]. ALT tumors can readily be detected by assaying for the presence of extrachromosomal telomeric DNA C-Circles (CCs) via qPCR or ALT-associated telomere foci by FISH on pathological specimens [6]. ALT+ high grade glioma (HGG) are enriched in tumors with loss of function mutations in ATRX (alpha-thalassemia/mental retardation X-linked) and less commonly, SMARCAL1. When these chromatin remodeling genes are inactivated, the resultant replication stress and aberrant HR at telomeres is hypothesized to lead to ALT [2]. Mutations in both ATRX and SMARCAL1 are mutually exclusive with TERT promoter mutations suggesting functional redundancy between these two mechanistic pathways [3, 4].

Here, we sought to identify and characterize ALT+ GBM by screening through a panel of 24 patient-derived GBM stem cell lines (GSCs). We tested for ALT using a novel qPCR method that measures both telomere content (TC), which is indicative of overall telomere length, and DNA C-Circles (CCs), which are specific and quantifiable markers for ALT activity [9]. Telomerase expression was assessed by quantifying mRNA levels of TERT using whole transcriptome sequencing. ATRX protein expression was measured by immunoblotting.

Of the 24 GSCs that were tested, 2 were found to be ALT+ (8.3%), GS 5–22 and GS 8–18. These 2 cell lines have significantly elevated DNA CC content (P < 0.001, t-test) and telomere content (p < 0.001, t-test) relative to other GSCs (Fig. 1a and b). Furthermore, both GS 5–22 and GS 8–18 lack detectable full length ATRX protein upon immunoblot analysis (Fig. 1c). Whole transcriptome sequencing data (available for 22 of 24 GSCs) identified mRNA expression of TERT to be negligible in the two ALT+ GSCs, indicating absence of telomerase activity, whereas the remaining GSCs all had some quantifiable level of TERT expression (p = 0.0087, Mann-Whitney test) (Fig. 1d). Importantly, both GS 5–22 and GS 8–18 were derived from patients with secondary glioblastoma with concurrent IDH mutations. Also, p53 immunostaining was positive in both ALT+ GSCs (data not shown) corroborating p53 loss of function and mutant IDH along with ATRX loss as important in the development of ALT+ GBM. GS 5–22 and 8–18 display longer doubling times in vitro, 5 days and 8 days, respectively, relative to ATRX-intact TERT-positive GSCs which have a mean doubling time of ~ 3–4 days. We injected GS 522 cells intracranially into athymic mice to evaluate their ability to generate stable xenografts, and saw tumors form within 1 months’ time (Fig. 1e).

To date, only 2 ALT+ glioma cell lines have been documented (TG-20 and JHH-GBM14) [5, 12], however in these prior studies ALT was assayed for by immunofluorescent detection of telomere/PML body foci and lack of telomerase activity via the telomerase repeat amplification protocol (TRAP) assay. We report here that detection of DNA CCs via qPCR and mRNA quantification of TERT are also usable biomarkers that can reliably detect ALT and may be more applicable in a clinical setting as both assays require minute amounts of DNA and RNA. In conclusion, identification of these ALT+ GSCs will enable future explorations of the mechanisms and biology of the ALT phenotype, and will serve as pre-clinical models to test novel chemotherapeutic agents in an effort to improve outcomes in a subset of high-grade gliomas and secondary GBM.