Suppressing proteasome mediated processing of topoisomerase II DNA-protein complexes preserves genome integrity

Abstract

Topoisomerase II (TOP2) relieves topological stress in DNA by introducing double-strand breaks (DSBs) via a transient, covalently linked TOP2 DNA-protein intermediate, termed TOP2 cleavage complex (TOP2cc). TOP2ccs are normally rapidly reversible, but can be stabilized by TOP2 poisons, such as the chemotherapeutic agent etoposide (ETO). TOP2 poisons have shown significant variability in their therapeutic effectiveness across different cancers for reasons that remain to be determined. One potential explanation for the differential cellular response to these drugs is in the manner by which cells process TOP2ccs. Cells are thought to remove TOP2ccs primarily by proteolytic degradation followed by DNA DSB repair. Here, we show that proteasome-mediated repair of TOP2cc is highly error-prone. Pre-treating primary splenic mouse B-cells with proteasome inhibitors prevented the proteolytic processing of trapped TOP2ccs, suppressed the DNA damage response (DDR) and completely protected cells from ETO-induced genome instability, thereby preserving cellular viability. When degradation of TOP2cc was suppressed, the TOP2 enzyme uncoupled itself from the DNA following ETO washout, in an error-free manner. This suggests a potential mechanism of developing resistance to topoisomerase poisons by ensuring rapid TOP2cc reversal.

Keywords: DNA damage; DNA repair; cancer; genetics; genome instability; genomics; human; mouse; topoisomerases.

Figure 1.. Proteasome machinery is essential for triggering DNA damage response to ETO treatment.

(A) As indicated in the schematic, primary splenic B-cells were isolated from mouse spleens and activated with a cocktail of Il-4, LPS, and RP105. 12 hr post-activation, while the B-cells were still in the G1 phase of the cell cycle, they were pre-treated with proteasome inhibitor (10 µM MG132 or 5 µM Bortezomib) for 1 hr prior to an additional 2 hr co-treatment with 50 µM ETO. Following the ETO pulse, the cells were washed with ice cold drug-free media (3 × 5 min spin at 1500 rpm and 4°C to pellet B-cells between washes; 15–20 min total time to complete washout), then returned to fresh drug free media and allowed to recover at 37°C for up to 48 hr. In the case of post-treatment, proteasome inhibitor was added to the wash media and then the cells were incubated for an additional 2 hr at 37°C with proteasome inhibitor only. For western blot and ICE assay analysis, cells were harvested immediately following drug treatment or washout (0 hr washout). For metaphase and SKY analysis, cells were harvested 24 hr following washout. For CellTiter-Glo viability, cells were harvested 48 hr following washout. For END-seq analysis, cells were harvested before drug washout, immediately following washout (0 hr washout), or after a 2 hr recovery in drug free media (2 hr washout). (B) γ-H2AX western blot in G1 WT primary splenic B-cells following 2 hr 50 µM ETO treatment ±1 hr proteasome inhibitor pre-treatment or 30 min following 5Gy IR ±1 hr proteasome inhibitor pre-treatment. Proteasome inhibitors tested: 10 µM MG132, 10 µM BTZ (Bortezomib), 2 µM EPN (Epoxomicin), 10 µM IXA (Ixazomib). (C) γ-H2AX immunofluorescence staining (red) in G1 arrested WT pre B cells following 2 hr, 50 µM ETO treatment ±1 hr 10 µM MG132 pre-treatment or 30 min after 5 Gy IR treatment ±1 hr 10 µM MG132 pre-treatment. (D) WT primary splenic B-cells (N = 3) in G1 were treated for 2 hr with 50 µM ETO ±1 hr, 5 µM BTZ pre-treatment. DNA was then isolated from cells following the ICE assay protocol and probed with anti-TOP2β antibody to quantify levels of TOP2βcc. Relative band intensity was measured for each sample and averaged for all three mice (ETO vs. pre-BTZ + ETO p=0.0168; ETO vs. ETO + 0 hr washout p=0.0045; pre-BTZ + ETO vs. pre-BTZ + ETO 0 hr washout p=0.0046; statistical significance calculated using student T-test). (E) eHAP cells were transfected for 48 hr with a YFP-tagged degron construct, which is degraded in a proteasome-dependent manner (Bence et al., 2001). Cells were then treated with 5 µM MG132 or 1 µM BTZ for 2 hr and then the drugs were washed out (3X wash in cold drug free media). Cells were fixed before washout with drugs still present or at 2 hr, 6 hr, and 18 hr post-washout. YFP signal intensity was quantified by FACS. (BTZ (before washout) vs. NT p=0.025; 2 hr post-washout MG132 vs NT p=0.049; statistical significance calculated using student T-test. MG132 vs BTZ p=0.038; statistical significance calculated using two column ANOVA). (F) pan-p53 levels in primary splenic B-cells treated for 3 hr with 10 µM MG132 or 5 µM BTZ. Drugs were then washed out and samples collected before washout or at 2 hr, 6 hr and 18 hr post-washout. (G) γ-H2AX levels in primary splenic B-cells as determined by Western blotting. Cells were treated with 50 µM ETO for 2 hr ±1 hr or 30 min pre-treatment, co-treatment and post-treatment with 5 µM BTZ at three different timepoints: before washout (treated), 2 hr and 6 hr post-washout.

Figure 2.. Proteasome inhibition decreases TOP2-mediated DSB persistence following ETO washout.

(A) UCSC genome browser snapshot of a TOP2-mediated DSB captured within the Nkx2-2 gene locus by END-seq (ExoVII+ExoT) at three timepoints: in primary splenic B-cells treated for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment, immediately following drug washout (0 hr washout), and at 2 hr post-washout (2 hr washout). (B) Venn diagram depicting the overlap between TOP2-mediated DSBs generated by ETO genome-wide in the 50 µM ETO only sample (green) and the 5 µM BTZ pre-treatment sample (orange). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment and processed by END-seq using ExoVII+ExoT. (C) Scatterplot depicting initial intensity of TOP2-mediated DSBs induced by ETO genome-wide in the 50 µM ETO only sample vs. the 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶; statistical significance calculated by Pearson’s correlation coefficient). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment. (D) Bar graph measuring the number of TOP2-mediated DSBs (END-seq with ExoVII+ExoT) at 0 hr and 2 hr post-washout timepoints, represented as a fraction of the initial number of TOP2-mediated DSBs generated genome-wide (% initial breaks after ETO treatment determined by peak calling). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment. (E) Box-plot of TOP2-mediated DSB persistence (relative intensity of initial breakage) genome-wide at 0 hr and 2 hr post-washout timepoints (left and right panels respectively). Green box 50 µM ETO only sample and orange box 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶, statistical significance calculated by student T-test). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment prior to washout. (F) Bar graph depicting the percentages of total TOP2-mediated DSBs (measured by ExoVII+ExoT), protein-free DSBs (measured by ExoT), and TOP2ccs (inferred by subtracting protein free DSBs from the levels of total TOP2-mediated DSBs). Reversible TOP2ccs (light blue), irreversible TOP2ccs (dark blue), and protein-free DSBs (red). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment. (G) Bar graph depicting the number of TOP2-mediated DSBs at 2 hr post-washout (END-seq with ExoVII+ExoT), represented as a fraction of the initial number of TOP2-mediated DSBs measured genome-wide (% initial breaks after ETO treatment determined by peak calling). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±pre treatment, co-treatment, and post-treatment with 5 µM BTZ. Post treatment with BTZ was for 2 hr. (H) Box-plot depicting TOP2-mediated DSB persistence genome-wide at 2 hr post-washout (p<2.2e⁻¹⁶ for ETO vs. pre-BTZ + ETO, ETO vs co-BTZ + ETO, and ETO vs post-BTZ + ETO; statistical significance calculated by student T-test). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±pre treatment, co-treatment, and post-treatment with 5 µM BTZ.

Figure 2—figure supplement 1.. TOP2-Medated DSBs are reproducibly captured by END-seq.

(A) Venn diagram depicting the overlap in initial TOP2-mediated DSBs generated by ETO genome-wide between two separate END-seq experiments (blue and green). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO. (B) Scatterplot comparing the initial intensity in TOP2-mediated DSBs generated genome-wide by ETO across two separate END-seq experiments. (Replicate one vs. Replicate 2 p<2.2e⁻¹⁶; statistical significance calculated by Pearson’s r). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO. (C) Venn diagram depicting the overlap between TOP2-mediated DSBs generated by ETO genome-wide in the 50 µM ETO only sample (green) and the 5 µM BTZ pre-treatment sample (orange). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment and processed by END-seq using ExoVII+ExoT. Dataset from END-seq replicate 2. (D) Scatterplot depicting initial intensity of TOP2-mediated DSBs induced by ETO genome-wide in the 50 µM ETO only sample vs. the 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶; statistical significance calculated by Pearson’s correlation coefficient). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment. Dataset from END-seq replicate 2. (E) Bar graph measuring the number of TOP2-mediated DSBs (END-seq with ExoVII+ExoT) at 0 hr and 2 hr post-washout timepoints in the break dataset from END-seq replicate 2, represented as a fraction of the initial number of TOP2-mediated DSBs generated genome-wide (% initial breaks after ETO treatment determined by peak calling). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment. (F) Box-plot of TOP2-mediated DSB persistence (relative intensity of initial breakage) genome-wide at 0 hr and 2 hr post-washout timepoints (left and right panels respectively). Green box 50 µM ETO only sample and orange box 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶, statistical significance calculated by student T-test). WT primary splenic B-cells were treated for 2 hr with 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment prior to washout. Dataset from END-seq replicate 2.

Figure 3.. Proteasomal degradation of the TOP2cc is associated with DSB resection.

(A) Zoomed in UCSC genome browser snapshot of the same TOP2-mediated DSB as in Figure 2A at three timepoints: 2 hr with 50 µM ETO (ETO), immediately following drug washout (0 hr WO), and at 2 hr post-washout (2 hr WO) (left panel). In right panel, WT primary splenic B-cells were pre-treated for 1 hr with 5 µM BTZ pre-treatment prior to washout. Red box indicates areas with resection signal. (B) Bar graph depicting the number of TOP2-mediated DSBs that had a resection signal immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ±5 µM BTZ pre-treatment. Green bar- 50 µM ETO only sample and orange bar- 5 µM BTZ pre-treatment sample. (C) Box-plot quantifying maximum resection distance for all persistent TOP2-mediated DSBs genome-wide immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ±5 µM BTZ pre-treatment. Green box 50 µM ETO only sample and orange box 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶; statistical significance calculated by student T-test). (D) Box-plot of resected (grey) and non-resected (green) TOP2-mediated DSBs genome-wide immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ± 5 µM BTZ pre-treatment. TOP2-mediated DSB resection values are plotted as a function of their persistence at the 0 hr washout timepoint.

Figure 3—figure supplement 1.. TOP2-Medated DSB resection is reproducibly captured by END-seq.

(A) Bar graph depicting the number of TOP2-mediated DSBs that had a resection signal immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ±5 µM BTZ pre-treatment. Green bar- 50 µM ETO only sample and orange bar- 5 µM BTZ pre-treatment sample. Dataset from END-seq replicate 2. (B) Box-plot quantifying maximum resection distance for all persistent TOP2-mediated DSBs genome-wide immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ±5 µM BTZ pre-treatment. Green box 50 µM ETO only sample and orange box 5 µM BTZ pre-treatment sample (ETO vs. pre-BTZ + ETO p<2.2e⁻¹⁶; statistical significance calculated by student T-test). Dataset from END-seq replicate 2. (C) Box-plot of resected (grey) and non-resected (green) TOP2-mediated DSBs genome-wide immediately following drug washout (0 hr washout) in WT primary splenic B-cells treated for 2 hr with 50 µM ETO ± 5 µM BTZ pre-treatment. TOP2-mediated DSB resection values are plotted as a function of their persistence at the 0 hr washout timepoint. Dataset from END-seq replicate 2.

Figure 4.. Proteasome inhibition suppresses ETO-induced genome instability and improves cell survival.

(A) Example of mitotic spreads from WT primary splenic B-cells treated for 2 hr in G1 with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment, followed by 24 hr recovery in drug free medium prior to harvesting metaphases. Colored arrows indicate the type of aberrations as quantified and described in B. (B) Analysis of chromosomal aberrations (Chromosome Breaks (yellow), Chromatid Breaks (orange), Fusions (blue), Radials (purple) and Fragments (green)). 50 metaphases were counted for each condition. (C) Example of mitotic spreads from WT primary splenic B-cells treated for 2 hr in G1 with 50 µM ETO ± 1 hr 10 µM MG132 pre-treatment, and then harvested 24 hr after washout for Spectral Karyotyping Analysis (SKY) (Left and right top panels). SKY reveals more complex fusions involving multiple chromosomes, compared to the fusions observable by DAPI staining only (Left and right bottom panels). (D) Analysis of chromosomal fusion determined by SKY analysis. Chromosome fusions were counted in 35 mitotic spreads per condition and were broken down by how many chromosomes were involved (1, two or greater than 3 (+)) per fusion event per cell. (E) Quantification of percent cells in S phase determined by FACS. WT primary splenic B-cells were treated in G1 for 2 hr with 50 µM ETO ± 1 hr 10 µM MG132 pre-treatment. 16 or 24 hr after washout, cells were pulsed for 30 min with EdU pulse an analyzed by FACS (16 hr NT vs MG132 p=0.012; 16 hr NT v ETO p=0.0485; 24 hr ETO vs pre-MG132 + ETO p=0.049; 24 hr ETO vs. post-MG132 p=0.46; statistical significance calculated by student T-test). (F) Viability of WT primary splenic B-cells as determined using the CellTiter-Glo luminescence assay 48 hr following 2 hr of 50 µM ETO treatment ±pre treatment, co-treatment, or post-treatment with 5 µM BTZ (BTZ v ETO p=0.0006, ETO v pre-BTZ + ETO p=0.015, statistical significance calculated by student T-test). During treatment cells were in the G1 phase of the cell cycle.

Figure 5.. Multiple DNA repair pathways contribute to ETO-induced genome instability.

(A) Mitotic spread analysis of WT (White), ATM^-/- (Orange), and H2AX^-/- (Green) primary splenic B-cells treated in G1 for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment. B-cells were fixed for mitotic spread analysis following a 24 hr recovery at 37°C in drug free media. Total chromosomal aberrations were counted in 50 metaphases (N = 3) and averaged for each genotype. (ETO vs. pre-BTZ + ETO: WT, p=0.006; ATM^-/-, p=0.0004; H2AX^-/-, p<0.0001; statistical significance calculated by student T-test). (B) Mitotic spread analysis of WT (White) and Lig4^-/- (Purple) primary splenic B-cells treated in G1 for 2 hr with 50 µM ETO ±1 hr 10 µM MG132 pre-treatment. B-cells were fixed for mitotic spread analysis following a 24 hr recovery at 37°C in drug free media. Total chromosomal aberrations were counted in 50 metaphases (N = 3) and averaged for each genotype. (ETO vs pre-MG132 + ETO: WT, p=0.0154; Lig4^-/-, p=0.0075; statistical significance calculated by student T-test). (C) Mitotic spread analysis of cycling (24 hr post-activation) WT (White) and BRCA1^Δ11 (Red) primary splenic B-cells treated for 2 hr with 50 µM ETO ± 1 hr 10 µM MG132 pre-treatment. B-cells were fixed for mitotic spread analysis following a 24 hr recovery at 37°C in drug free media. Total chromosomal aberrations were counted in 50 metaphases per condition. (D) Mitotic spread analysis of cycling primary splenic B-cells treated for 2 hr with 12.5 µM CPT ± 1 hr 10 µM MG132 pre-treatment. B-cells were fixed for mitotic spread analysis following a 24 hr recovery at 37°C in drug free media. Total chromosomal aberrations were counted in 50 metaphases per condition.

Figure 5—figure supplement 1.. Multiple DNA repair mechanisms contribute to error-prone repair of ETO-induced damage.

Mitotic spread analysis of total chromosomal aberrations in PolQ^-/- and BRCA2^-/- primary splenic B-cells treated for 2 hr with 50 µM ETO, relative to their corresponding WT controls. ETO was administered in either G1 cells (PolQ^-/-) or cycling cells (BRCA2^-/-). Chromosomal fusions were counted in 50 metaphase spreads across at least two experiments for each genotype (N = 2) (statistical significance calculated by student T-test).

Figure 6.. The SUMO-targeted Ubiquitin-Ligase RNF4 functions in proteasome mediated processing of TOP2cc.

(A) Example γ-H2AX western blot in WT and RNF4^-/- MEFs treated with 10 µM ETO for 1 hr ±1 hr 1 µM BTZ pre-treatment (Left Panel). Bar graph quantifying the relative band intensity from γ-H2AX western blots (N = 3) of WT (White) and RNF4^-/- (Blue) MEFs treated with 10 µM ETO (Right Panel). (WT ETO vs RNF4^-/- ETO p=0.0018; statistical significance calculated by student T-test). (B) Scatterplots depicting initial intensity of all TOP2-mediated DSBs (ExoVII/ExoT, Top Left) and initial intensity of protein-free DSBs genome-wide (ExoT, Top Right) induced by 2 hr treatment of WT and RNF4^-/- B cells with 50 µM ETO. Scatterplots of all TOP2-mediated DSBs (ExoVII/ExoT, Bottom Left) and protein-free DSBs genome-wide (ExoT, Bottom Right) in WT primary splenic B-cells treated with 50 µM ETO only vs. 50 µM ETO + 5 µM BTZ pre-treatment. (C) Viability of WT (White) and RNF4^-/- (Blue) primary splenic B-cells, as determined by the CellTiter-Glo luminescence assay 48 hr following 2 hr 50 µM ETO treatment ±1 hr 5 µM BTZ pre-treatment. (D) Colony formation assay in WT (White) and RNF4^-/- (Blue) MEFs. Cells were seeded in 6 cm plates at 100, 1000 and 10000 cells/plate and pulsed with 5–20 µM ETO for 1 hr and then left to recover and form colonies for 7 days. The number of colonies from duplicate plates were averaged and plotted at each concentration (N = 3) (WT vs RNF4 p=0.0034; statistical significance calculated by ANOVA). (E) Mitotic spread analysis of WT (White) and RNF4^-/- (Blue) primary splenic B-cells treated in G1 for 2 hr with 50 µM ETO ± 1 hr 5 µM BTZ pre-treatment. B-cells were fixed for mitotic spread analysis following a 24 hr recovery at 37°C in drug free media. Total chromosomal aberrations were counted in 50 metaphases (N = 3) and averaged for each genotype. (WT vs. RNF4^-/- ETO: p=0.048, statistical significance calculated by student T-test).

Figure 7.. Model of proteasome mediated repair of TOP2ccs.

(A) ETO-stabilized TOP2ccs are targeted for proteasomal degradation by the 26S proteasome via both SUMOylation and Ubiquitination, mediated by the SUMO-targeting ubiquitin ligase RNF4 and possibly other enzymes. Upon degradation of the TOP2 protein within the TOP2cc, the remaining 5’-phosphotyrosyl bonds must be processed by TDP2, generating clean protein-free DSBs with overhangs that can be repaired in a manner that is not always error-free. (B) If the proteasome is not recruited due to loss of RNF4-mediated TOP2cc polyubiquitylation or its activity is inhibited chemically, the TOP2cc is not recognized by the DNA repair machinery and the TOP2 protein retains full enzymatic ability to reverse itself once ETO is washed out.

Author response image 1.