Pericentromeric Satellite III transcripts induce etoposide resistance

Abstract

Non-coding RNA from pericentromeric satellite repeats are involved in stress-dependent splicing processes, maintenance of heterochromatin, and are required to protect genome stability. Here we show that the long non-coding satellite III RNA (SatIII) generates resistance against the topoisomerase IIa (TOP2A) inhibitor etoposide in lung cancer. Because heat shock conditions (HS) protect cells against the toxicity of etoposide, and SatIII is significantly induced under HS, we hypothesized that the protective effect could be traced back to SatIII. Using genome methylation profiles of patient-derived xenograft mouse models we show that the epigenetic modification of the SatIII DNA locus and the resulting SatIII expression predict chemotherapy resistance. In response to stress, SatIII recruits TOP2A to nuclear stress bodies, which protects TOP2A from a complex formation with etoposide and results in decreased DNA damage after treatment. We show that BRD4 inhibitors reduce the expression of SatIII, restoring etoposide sensitivity.

Fig. 1. Hypomethylation of pericentromeric satellite repeats correlates with etoposide resistance in non-small cell lung cancer PDX mouse models.

A Volcano plot shows the global methylation changes at repetitive elements between 22 patient-derived NSCLC xenograft tumor samples and their corresponding normal tissue (TvsN). Each dot represents one repetitive element, based on the RepeatMasker database, classified into subclasses (color code). The log fold change of methylation in tumor versus normal tissue is plotted on the x-axis, the y-axis shows the negative log 10 of the p-value. B Heatmap shows the Pearson’s correlation between the sensitivity of the PDXs (measured by the differential methylation between tumor and normal tissues) with the chemotherapeutics indicated by the column name. The adjacent bar indicates the repetitive region class. C Composition of the significantly correlated (p-value < 0.05) repeats classes: positive (left) and negative (right) correlations between response to Etoposide and differential methylation. D Methylation levels of two etoposide sensitive PDXs (7166 and 7298; green) and two etoposide resistant PDX models (7433 and 7466; red) at the Satellite III DNA locus on chromosome 9. Lighter colors represent the coverage of normal tissue (N), darker colors show the coverage of tumor samples (T). E RNA FISH staining of Satellite III transcripts (SatIII RNA) in FFPE tumor material of untreated etoposide resistant (7433) and etoposide sensitive (10395) PDX mice. The tissue was stained with SatIII RNA FISH probes (red) as well as Hoechst stain 33342 (blue). Scale bar, 10 µm. F Correlation of SatIII RNA foci and the relative tumor volume on native FFPE PDX tumor tissue. SatIII RNA foci were stained with SatIII RNA FISH, quantified, and set into relation to the response rate towards etoposide.

Fig. 2. Hypermethylation of the SatIII DNA locus by D-TALES diminishes SatIII RNA expression and increases sensitivity towards etoposide.

A Schematic representation of SatIII DNA locus with all CpG localizations (red) as well as the TALE-binding sites (green). B Representative images of the co-localization of TALE-GFP (green) and SatIII RNA (red). HeLa cells were transfected with the TALE-GFP construct and exposed to HS conditions (1 h at 44 °C) 24 h after transfection in order to induce SatIII RNA expression and foci accumulation. Cells were fixed, immunostained, and imaged. The histogram indicates co-localizations, represented by overlapping peaks of fluorescence intensities. C Methylation level of the SatIII DNA locus in D-TALEa (active) and D-TALEi (inactive vector control) transfected U2OS cells. At 24 h and 48 h post-transfection U2OS cells were exposed to 44 °C for 1 h (HS) conditions and harvested. Cells were FACS-sorted and mCherry positive cells used for DNA extraction and subsequent pyrosequencing of the SatIII DNA locus. Shown is the percent of methylation to a transfection control. D Quantitative PCR analyses for SatIII RNA expression of samples from (C). Samples incubated for 48 h with the transfection mix were used. Error bars represent standard deviation of the mean of two individual replicates. E Drug response of U2OS cells either transfected with D-TALEa or an inactive control plasmid (D-TALEi). Viabilities were examined by AlamarBlue. Error bars represent standard deviation of the mean of three replicates. Two-tailed paired Student’s t test significant P-values are marked: < 0.05 with (*), < 0.01 with (**), < 0.001 with (***). F Caspase-3/-7 assay of U2OS cells either transfected with the D-TALEa plasmid or an inactive control (D-TALEi). At 24 h post-transfection, cells were treated with different concentrations of etoposide or a DMSO control for 24 h. Depicted are the differences of blue fluorescence (BFU) signals between D-TALEa and D-TALEi. Error bars represent standard deviation of the mean of the three replicates. Two-tailed paired Student’s t-test significant P-values are marked: < 0.05 with (*), < 0.01 with (**), < 0.001 with (***). G Cell viability of HCC827 cells stably expressing either shRNA which targets SatIII (shSatIII, blue) or a shGFP-Control (red). Cells were exposed to 44 °C for 1 h (HS) and afterwards immediately treated with the indicated etoposide concentrations. After an additional 48 h cell viability was measured using AlamarBlue.

Fig. 3. SatIII recruits TOP2A to nSB foci thereby protecting cells from DNA double strand breaks.

A Representative images of the SatIII RNA and TOP2A co-localization in HeLa cells exposed to HS conditions (1 h at 44 °C) or HS plus 24 h recovery at 37 °C and DMSO or etoposide 10 µM treatment. SatIII RNA was stained using smFISH (red), TOP2A was stained using a protein-specific antibody (green). Scale bar, 10 µm. B Quantification of (A) by counting the number of foci per cell. Quantification was performed using an automated ImageJ pipeline, n = 5. C Cell cycle assay utilizing Hoechst staining was performed to clarify SatIII expression patterns during the cell cycle. HeLa cells were subjected to HS (1 h at 44 °C) or control conditions, fixed and stained with Hoechst staining dye. The cells were monitored over the course of the cell cycle in an automated HCS microscope. A minimum of 3000 cells, separated in n = 6 replicates, were quantified for each condition. D Effects of SatIII RNA knockdown on DNA damage was investigated by immunofluorescence staining for 53BP1. HeLa cells were transfected with siSatIII and scramble RNA (control), respectively. Cells were then exposed to HS conditions (1 h at 44 °C) or constant 37 °C and treated with 20 µM etoposide or DMSO. After 24 h, cells were fixed and stained with a protein-specific 53BP1antibody (green). Counterstaining of nuclei was performed with Hoechst stain. Imaging and analyses were performed utilizing HCS microscope-based quantifications of the staining. Error bars represent the standard deviation of the mean of three replicates. Two-tailed paired Student’s t test significant P-values are marked: < 0.05 with (*), < 0.01 with (**), < 0.001 with (***). Scale bar, 10 µm. E RNA immunoprecipitation in HeLa cells subjected to three different treatment conditions: HS (1 h at 44 °C), HS with a 24 h recovery time at 37 °C (HS + rec), and non-HS conditions (nHS, 37 °C). Chromatin was sheared by sonication and precipitated using an antibody against human TOP2A or HSF1. Binding to SatIII was analyzed using qPCR. HSF1 was used as a positive control. Figure shows a typical result for two biological replicates, each with three technical replicates. F Caspase-3/-7 assay of HeLa cells either transfected with the siRNA targeting SatIII (siSatIII) or control siRNA (siCo). After transfection, cells were treated as indicated with etoposide or a DMSO control. Error bars represent standard deviation of the mean of three replicates. Two-tailed paired Student’s t-test significant P-values are marked: < 0.05 with (*), < 0.01 with (**), < 0.001 with (***).

Fig. 4. BET protein inhibitors revert etoposide resistance through SatIII regulation.

A Representative images of co-localization of SatIII RNA and BRD4 after exposure to HS conditions (1 h at 44 °C). Immediately following HS, cells were fixed, immunostained, and imaged. SatIII was stained with smFISH (red), BRD4 with a protein-specific antibody (green). Scale bar, 10 µm. B Binding of BRD4 to the SatIII DNA locus after HS (1 h at 44 °C). For the ChIP-Seq experiment cells were subjected to HS or control conditions. ChIP was performed with a BRD4-specific antibody. C Effect of BRD4 inhibition on SatIII expression. HeLa cells were treated with various BRD4 inhibitors and exposed to HS conditions (1 h at 44 °C). Two read-out methods were applied: RNA expression was measured by qPCR and RNA FISH was used to quantify SatIII RNA foci. Values of non-treated cells (DMSO control) were set to 100%. Error bars represent SD of the mean of n = 2 independent replicates. P-values < 0.05 marked with (*). Significance was determined using two-tailed paired Student’s t-test. D Cell viability of HeLa cells treated with etoposide and the BRD4 inhibitor JQ1 (1 µM) or DMSO as control. Directly after exposure to HS (1 h at 44 °C), cells were treated with the indicated etoposide concentrations in combination with JQ1. After an additional 48 h, cell viabilities were measured using AlamarBlue. Error bars represent SD of the mean of n = 2 replicates. P-values <0.05 are marked with (*), Significance was determined using two-tailed unpaired Student’s t-test. E Same experiment as in (D) but with the BRD4 inhibitor CPI203 (1 µM). F Cell proliferation of HeLa cells either stably overexpressing SatIII RNA or an empty vector control. Cells were treated with 20 µM etoposide and cell proliferation was measured by acquisition of images every 30 min over a time course of 48 h. Confluency was analyzed with the cell profiler software. G Cell proliferation assay performed as described in (F) but with a combination treatment of 20 µM etoposide and 5 µM JQ1.

Fig. 5. Model of SatIII regulation in tumorigenesis and chemoresistance.

Loss of BRCA1 results in an increased SatIII RNA expression through reduced ubiquitination of H2A and a relaxation of pericentromeric heterochromatin, reflected by a loss of H3K9me2 (left side, Zhu et al., 2018, Padeken et al.). SatIII RNA interacts with the BRCA1-associated protein network and destabilizes replication forks which in turn enhances DNA damage and genomic instability, ultimately promoting tumor growth. Etoposide also drives SatIII expression, but in this case, SatIII RNA facilitates the recruitment of TOP2A to TOP2ccs located at nSBs (our data). This leads to less DNA damage and subsequent downstream mechanisms that decrease genomic instability and therefore cells are more resistant against etoposide.