TOP3A amplification and ATRX inactivation are mutually exclusive events in pediatric osteosarcomas using ALT

Abstract

In some types of cancer, telomere length is maintained by the alternative lengthening of telomeres (ALT) mechanism. In many ALT cancers, the α-thalassemia/mental retardation syndrome X-linked (ATRX) gene is mutated leading to the conclusion that the ATRX complex represses ALT. Here, we report that most high-grade pediatric osteosarcomas maintain their telomeres by ALT, and that the majority of these ALT tumors are ATRX wild-type (wt) and instead carry an amplified 17p11.2 chromosomal region containing TOP3A. We found that TOP3A was overexpressed in the ALT-positive ATRX-wt tumors consistent with its amplification. We demonstrated the functional significance of these results by showing that TOP3A overexpression in ALT cancer cells countered ATRX-mediated ALT inhibition and that TOP3A knockdown disrupted the ALT phenotype in ATRX-wt cells. Moreover, we report that TOP3A is required for proper BLM localization and promotes ALT DNA synthesis in ALT cell lines. Collectively, our results identify TOP3A as a major ALT player and potential therapeutic target.

Keywords: ATRX; TOP3A; alternative lengthening of telomeres; osteosarcomas; telomeres.

Figure 1. High frequency of alternative lengthening of telomeres (ALT) in high‐grade osteosarcomas

1. C‐circle assay for tumor samples. The presence of C‐circles was tested in 22 osteosarcomas by Rolling Circle Amplification (RCA) assay using Φ29 DNA polymerase. Φ29 DNA polymerase negative controls and Φ29 DNA polymerase‐based reactions starting, respectively, with 75, 37.5, and 18.75 ng of DNA were applied to dot blots and C‐circles were detected by hybridization with a ³²P‐(CCCTAA)₄ telomeric probe. U2OS and HeLa cells correspond to positive and negative control samples, respectively.

2. Representative images of telomere fluorescence in‐situ hybridization (FISH; red) and promyelocytic leukemia (PML) immunofluorescence IF (green) colocalizations (ALT‐associated PML body [APBs]) in osteosarcoma tumor samples. DNA is counterstained blue with DAPI. APBs are indicated by white arrows. Scale bars are 5 μm.

3. Representative image of telomere shortest length assay (TeSLA) Southern Blot of two representative samples of ALT‐positive (left) and telomerase‐positive (right) osteosarcomas. Nine TeSLA polymerase chain reactions (PCRs; 30 pg each reaction) were done for each DNA sample. DIG‐labeled MW ladder has been added a posteriori.

4. Plot of cumulative TeSLA fragment sizes (in kb) in ALT‐positive (n = 16) and ALT‐negative (n = 6) osteosarcomas. Points and error bars (±SEM) represent cumulative percentage of TeSLA fragments from all the TeSLA PCRs (n = 9 per tumor sample) from all the patients in both groups (n = 16 in ALT‐positive group, and n = 6 in ALT‐negative group). The Kolmogorov–Smirnov test was applied to identify statistical differences in TeSLA fragments distributions (P = 0.0016).

5. Individual TeSLa fragment lengths in ALT‐positive (n = 16) and ALT‐negative (n = 6) osteosarcomas. Each dot represents a TeSLA fragment, bars represent the median value for individual patient's samples.

Figure 2. α‐thalassemia/mental retardation syndrome X‐linked wild‐type (ATRX‐wt) in expressed in alternative lengthening of telomeres (ALT)‐positive high‐grade osteosarcomas

1. Oncoprint graph showing the distribution of mutations (green rectangles) and copy number variations (red rectangle, amplification; blue rectangle, copy loss) in all samples. Tumors are distributed according to ALT positivity (ALT⁺) or ALT negativity (ALT⁻) and ATRX status (ATRX‐mutated or ATRX‐wt).

2. Tissue sections of two representative high‐grade osteosarcomas with hematoxylin eosin saffron (HES) staining (left panel), anti‐ATRX (HPA064684) immunochemistry (middle panel), and anti‐DAXX (HPA008736) immunochemistry (right panel), showing ATRX protein expression in two ALT‐positive samples. In the top panel, intratumoral ATRX expression is negative (positive control osteoclasts are shown). In the bottom panel, intratumoral ATRX expression is high. Scale bars are 250 μm.

3. STRING network showing mutated gene‐clusters with related functions.

Figure EV1. Survival analysis according to ALT

1. Disease‐free survival curves of patients with ALT‐positive and ALT‐negative tumors.

2. Disease‐free survival curves of good and poor responders to neoadjuvant chemotherapy (left panel), then dichotomized according to ALT status of tumors (right panel).

Figure EV2. No difference was observed in tumor mutational burden or percentage of altered genome according to ALT

1. Tumor mutational burden in ALT‐negative (n = 6) and ALT‐positive (n = 16) tumors. Box‐and‐whisker plot were defined with default parameters by median value (central band at the 50^th percentile), interquartile ranges (IQR, box limited by 25^th and 75^th percentile) and whisker boundaries defined by minimum and maximum value. Wilcoxon test was used to compare modalities.

2. Percentage of altered genome using array‐comparative genomic hybridization (aCGH) in ALT‐negative (n = 6) and ALT‐positive (n = 16) tumors. Box‐and‐whisker plot were defined with default parameters by median value (central band at the 50^th percentile), interquartile ranges (IQR, box limited by 25^th and 75^th percentile) and whisker boundaries defined by minimum and maximum value. Wilcoxon test was used to compare modalities.

Figure 3. TOP3A amplification and α‐thalassemia/mental retardation syndrome X‐linked (ATRX) inactivation are mutually exclusive in alternative lengthening of telomeres (ALT) pediatric osteosarcoma

1. Copy number alteration (CNA) profiles of osteosarcomas. Representation of the GISTIC analysis with false discovery rate (FDR (q‐value)) for all cases (n = 22; left panel), ALT‐positive/ATRX wild‐type (wt) cases (n = 11; middle panel), and ALT‐positive/ATRX‐mutated cases (n = 5; right panel).

2. Gain/amplification of TOP3A gene region (17p11) is a genomic signature of ATRX‐wt osteosarcomas. For each ATRX‐mutated (top part) and ATRX‐wt (bottom part) case, from the left to the right, the chromosome 17 and regional genomic profiles were established, with CGH Analytics software. A focus was made within the genomic interval (17.25–19.11 Mb) of the short arm of chromosome 17 (hg38 human genome mapping; Build 38 from NCBI, December 2013 version) including TOP3. Color profiles corresponding to the different tumors are defined at the top of each group. Most ATRX‐wt cases show gain or amplification of this region (bottom part), whereas no ATRX‐mutated cases display this alteration (top part).

3. Chromosome 17 copy number variation analysis showing amplification/gain of the TOP3A region that do not extend into the TP53 gene in ALT ATRX‐wt tumors (left panel), and amplification of the TOP3A region for which TP53 region is at the edge of the amplification/gain in ALT‐positive ATRX‐wt tumors (right panel).

Figure EV3. Histone genes in patient's osteosarcoma tumors

1. Lollipop plots of H1.4, H2A, H3.1, and H4 genes showing histone mutations detected by tNGS.

2. Expression of histone genes according to ALT and ATRX status: ALT⁺/ATRX⁺ (n = 7), ALT⁺/ATRX⁻ (n = 2) et ALT⁻ (n = 3). Box‐and‐whisker plot were defined with default parameters by median value (central band at the 50^th percentile), and interquartile ranges (IQR, box limited by 25^th and 75^th percentile) and whisker boundaries were defined by minimum and maximum value. An ANOVA statistical test, and Tukey's range test were used to compare modalities.

Figure 4. TOP3A is overexpressed in alternative lengthening of telomeres (ALT)‐positive/α‐thalassemia/mental retardation syndrome X‐linked wild‐type (ATRX‐wt) osteosarcomas

1. Volcano plot of the supervised analysis of mRNA expression profiles of osteosarcomas according to ALT‐positive vs. ALT‐negative samples, showing upregulated (in red) and downregulated (in green) genes. In all, 12 tumors had sufficient RNA quality to be analyzed: seven ALT‐positive ATRX‐wt, two ALT‐positive/ATRX‐mutated and three ALT‐negative.

2. TOP3A gene expression according to ALT and ATRX status: ALT⁺/ATRX‐wt (n = 7), ALT⁺/ATRX‐mut (n = 2) and ALT⁻ (n = 3). Box‐and‐whisker plot was defined with default parameters by median value (central band at the 50^th percentile), interquartile ranges (IQR, box limited by 25^th and 75^th percentiles) and whisker boundaries defined by minimum and maximum value. An ANOVA statistical test and Tukey's range test were used to compare modalities.

3. Normalized enrichment score bar plot of top 15 significant REACTOME gene sets according to ALT‐positive vs. ALT‐negative GSEA analysis. Ontologies associated with ALT positivity are in red and those associated with ALT negativity are in blue.

Figure EV4. TOP3A expression in osteosarcoma cell lines

1. TOP3A expression (by RT‐qPCR) in human osteosarcoma cell lines. Mean expression level of three technical replicates; error bars represent the mean ± SEM.

2. TOP3A Western blots validating siRNA knockdowns in the indicated cell lines (U2OS, NY, IIIcf TC4, IIICf E6E7).

Figure 5. α‐thalassemia/mental retardation syndrome X‐linked (ATRX) is functional in alternative lengthening of telomeres (ALT)‐positive ATRX wild‐type (wt) cell lines and its ALT inhibitory function is counteracted by overexpression of TOP3A

1. Western blot showing ATRX protein expression in nine osteosarcoma cell lines and HTC116 as positive control.

2. Western blot showing TOP3A protein expression in nine osteosarcoma cell lines and NY siTOP3A as control.

3. Telomere fluorescence in‐situ hybridization (FISH; green), and promyelocytic leukemia (PML; red) and TOP3A (purple) immunofluorescence (IF), showing TOP3A localization at APBs in both ATRX‐wt and ATRX‐mutated ALT‐positive cells. TOP3A foci are indicated by white arrows. Scale bars are 5 μm.

4. C‐circle assay showing inverse effect of ATRX ectopic expression on C‐circle levels in ATRX‐mutated and ATRX‐wt cell lines. Error bars represent the mean ± SEM from n = 2 experiments, n.s. = non‐significant, *P < 0.05, Mann–Whitney test.

5. C‐circle assay showing rescue of ATRX ectopic expression by TOP3A overexpression in U2OS osteosarcoma cells. Error bars represent the mean ± SEM from n = 2 experiments, n.s. = non‐significant, *P < 0.05, Mann–Whitney test. Western blot and quantification of TOP3A expression is shown for U2OS cells overexpressing TOP3A.

6. ALT‐associated PML body (APB) frequency and mean telomeric DNA intensity in APBs according to TOP3A overexpression and ATRX transient expression; n = 150 cells scored per treatment, n.s., non‐significant, *P < 0.05; Mann–Whitney test.

Figure EV5. Effect of ectopic ATRX expression

Growth curve (Points and error bars (±SEM) represent extrapolation of phase‐contrast object confluence assessed by Incucyte optical system from n = 3 technical replicates) according to transfection condition (FuGENE only, pCMV6‐empty, −ATRX, ‐DAXX).

Figure 6. TOP3A inhibition affects alternative lengthening of telomeres (ALT) phenotype and leads to telomeric DNA damage

1. Relative C‐circle intensity according to TOP3A KD in ALT‐positive ATRX‐mutated or ATRX wild‐type (wt) cell lines. Error bars represent the mean ± SEM from n = 2 experiments.

2. Representative images of telomeric DNA (green) and promyelocytic leukemia (PML) protein (red) colocalizations (ALT‐associated PML body [APBs]) in osteosarcoma and in vitro‐immortalized ATRX‐mutated or ATRX‐wt cell lines according to TOP3A KD. Scale bars are 5 μm.

3. Quantification of APB frequency in osteosarcoma and in vitro‐immortalized ATRX‐mutated or ATRX‐wt cell lines according to TOP3A KD; Error bars represent the mean ± SEM from n = 3 experiments, n = 150 cells scored per treatment, n.s., non‐significant; Mann–Whitney test.

4. Effect of TOP3A KD on telomere dysfunction‐induced foci (TIFs) in osteosarcoma and in vitro‐immortalized cell lines; Error bars represent the mean ± SEM from n = 3 experiments, n = 150 cells scored per treatment, *P < 0.05, **P < 0.005; Mann–Whitney test.

5. Distributions of telomere shortest length assay (TeSLA) fragments in osteosarcoma and in vitro‐immortalized ATRX‐mutated or ATRX‐wt cell lines according to TOP3A KD (left panel; each dot represents a TeSLA fragment), and representative TeSLA Southern Blot image for NY (right panel).

Figure 7. TOP3A depletion affects BLM localization at alternative lengthening of telomeres (ALT)‐associated PML body (APBs) and disrupts telomeric DNA synthesis

1. Fluorescence in‐situ hybridization‐immunofluorescence (FISH‐IF) of telomeres (in red) and BLM (in green) in TOP3A‐depleted cells BLM foci at telomeres are indicated by white arrows. Error bars represent the mean ± SEM from n = 3 experiments, n = 150 cells scored per treatment, n.s. = non‐significant, **P < 0.01, Mann–Whitney test. Scale bars are 5 μm.

2. ATSA assay assessing ALT‐mediated telomeric DNA synthesis (telomere in red, EdU in purple) in promyelocytic leukemia (PML; in green) according to TOP3A KD. EdU signal at APB (telomere/PML) are indicated by white arrows. Approximately 150 cells were divided into five groups (0, 1–2, 3–4, 5–6n, and ≥ 7) based on the number of EdU + APBs. Scale bars are 5 μm.