RAD51 and BRCA2 Enhance Oncolytic Adenovirus Type 5 Activity in Ovarian Cancer

Abstract

Homologous recombination (HR) function is critically important in high-grade serous ovarian cancer (HGSOC). HGSOC with intact HR has a worse prognosis and is less likely to respond to platinum chemotherapy and PARP inhibitors. Oncolytic adenovirus, a novel therapy for human malignancies, stimulates a potent DNA damage response that influences overall antitumor activity. Here, the importance of HR was investigated by determining the efficacy of adenovirus type 5 (Ad5) vectors in ovarian cancer. Using matched BRCA2-mutant and wild-type HGSOC cells, it was demonstrated that intact HR function promotes viral DNA replication and augments overall efficacy, without influencing viral DNA processing. These data were confirmed in a wider panel of HR competent and defective ovarian cancer lines. Mechanistically, both BRCA2 and RAD51 localize to viral replication centers within the infected cell nucleus and that RAD51 localization occurs independently of BRCA2. In addition, a direct interaction was identified between RAD51 and adenovirus E2 DNA binding protein. Finally, using functional assays of HR competence, despite inducing degradation of MRE11, Ad5 infection does not alter cellular ability to repair DNA double-strand break damage via HR. These data reveal that Ad5 redistributes critical HR components to viral replication centers and enhances cytotoxicity.

Implications: Oncolytic adenoviral therapy may be most clinically relevant in tumors with intact HR function.

Figure 1. Greater efficacy and viral DNA replication in HR competent than HR defective ovarian cancer cells

A. Competence of homologous recombination was assessed in PEO1 and PEO4 cells. Cells were treated with rucaparib (10 μM, 24h), permeabilised, fixed in 4% PFA and stained for RAD51 and γH2AX. RAD51 foci were counted in at least 30 nuclei per treatment condition. Bars represent mean (+/− s.d.) number of RAD51 foci per cell. Dotted line represents 2x number of foci in untreated cells. B. 10⁴ PEO1 and PEO4 cells were infected in triplicate with dl922-947. Cell survival was measured 120h post-infection by MTT assay. Mean +/− s.d. IC₅₀ for four experiments are shown: *; p=0.011 C. 10⁴ PEO1 and PEO4 cells were infected in triplicate with Ad5 WT (left) or dl309 (right) (MOI 0.001 – 1000 pfu/cell). Cell survival was measured 120h post-infection by MTT assay. D. PEO1 and PEO4 cells were infected with Ad CMV-GFP (left) at MOI 60 and 100 (PEO1) and 6 and 10 (PEO4). GFP positivity was assessed 24h post-infection by flow cytometry. PEO1 and PEO4 cells were also infected with dl922-947 (right) at MOI 60 and 100 (PEO1) and 6 and 10 (PEO4). Cell survival was assessed 120h post-infection by MTT assay. Data represent mean +/− s.d., n=3. ***: p<0.001. E. PEO1 and PEO4 cells were infected with dl922-947 MOI 100 (PEO1) or 10 (PEO4). Protein was harvested up to 72h post-infection. Expression of E1A and adenovirus 5 structural proteins was assessed by immunoblot. E1A band density was assessed from three separate exposures – at 24h, 48 and 72h, mean (+/− s.d.) E1A:KU70 ratio was 1.3 +/− 0.2, 1.4 +/− 0.1 and 1.4 +/− 0.2 for PEO1, and 0.7 +/− 0.4, 1.2 +/− 0.3 and 1.6 +/− 0.2 for PEO4. F and G. PEO1 and PEO4 cells were infected with dl922-947 MOI 100 (PEO1) or 10 (PEO4) for up to 72h. Virus replication was assessed by TCID50 (1F) or quantitative PCR (1G). ***: p<0.001. H. PEO1 and PEO4 cells were infected with dl922-947 MOI 100 (PEO1) or 10 (PEO4) for up to 72h. DNA was extracted and subjected to neutral pulsed-field gel electrophoresis, probed with HRP-labelled adenovirus type 5 probe. 100 ng purified dl922-947 DNA was run as positive control (+)

Figure 2. dl922-947 replication induces genomic DNA damage; RAD51 and BRCA2 co-localise with sites of adenovirus replication

A. PEO1 and PEO4 cells were harvested 48h following infection with dl922-947 (MOI 100 and 10 respectively) or mock infection, fixed in 70% cold ethanol, incubated with an anti-γH2AX Ab, counter-stained with PI and analysed by flow cytometry (left). Increase in γH2AX-positive cells and cells with >4N DNA are plotted (right) – bars represent mean +/− s.d., n=3. ***: p<0.001. B and C. PEO1 and PEO4 cells were harvested following infection with dl922-947 (MOI 100 and 10 respectively). Expression of E1A, MRE11 (B) and RAD51 (C) was detected by immunoblot. D and E. PEO1 and PEO4 cells were fixed in 4% PFA following infection with dl922-947 (MOI 300 and 30 respectively). Expression of adenovirus E2 DNA binding protein, BRCA2 (D) and RAD51 (E) was assessed by confocal microscopy.

Figure 3. RAD51 and BRCA2 co-localise with sites of adenovirus replication in multiple malignant cell lines

A. Homologous recombination competence was assessed in TOV21G, HeLa and IGROV1 cells as for Figure 1A. Bars represent mean (+/− s.d.) number of RAD51 foci per cell. Dotted line represents 2x number of foci in untreated cells. TOV21G and HeLa demonstrate HR competence, whilst IGROV1 are HR defective. B and C. Cells were fixed in 4% PFA following infection with dl922-947 (MOI 10). Expression of adenovirus E2 DNA binding protein, BRCA2 (B) and RAD51 (C) was assessed by confocal microscopy. D. RAD51 was immunoprecipitated from TOV21G infected with dl922-947 (MOI 10) and the presence of E2 DNA binding protein was detected by immunoblotting.

Figure 4. RAD51 knockdown decreases adenovirus efficacy and replication

A. Using two different siRNA pools, RAD51 was knocked down in both PEO1 and PEO4 cells. B and C. 24h following siRNA-mediated RAD51 knockdown, PEO1 (MOI 300) and PEO4 cells (MOI 30) were infected with dl922-947 (left) and dl309 (right, MOI 500 and 50). Survival was assessed 96h post-infection by MTT assay (B). Viral replication was also assessed 48h post-infection by quantitative PCR (C). *; p<0.05. ***: p<0.001. D – F. 24h following siRNA-mediated RAD51 knockdown, TOV21G (D), HeLa (E) and IGROV1 (F) cells were infected with dl922-947 (MOI 1, 8 and 5 respectively). Survival was assessed 96h post-infection by MTT assay. RAD51 knockdown was confirmed by immunoblot. Viral replication was also assessed in TOV21G 48h post-infection. **; p<0.01. ***: p<0.001.

Figure 5. BRCA2 knockdown also decreases adenovirus efficacy and replication

A. 24h following siRNA-mediated BRCA2 knockdown, PEO4 cells were infected with dl922-947 (MOI 30). Survival was assessed 96h post-infection by MTT assay. BRCA2 knockdown was confirmed by quantitative RT-PCR, normalised to 18S RNA. *; p<0.05. ***: p<0.001. B and C. 24h following siRNA-mediated BRCA2 knockdown, TOV21G (B) and HeLa (C) cells were infected with dl922-947 (MOI 1 and 2 for TOV21G; MOI 8 for HeLa). Survival was assessed 96h post-infection by MTT assay. BRCA2 knockdown was confirmed by quantitative RT-PCR in TOV21G cells, normalised to 18S RNA. *; p<0.05. ***: p<0.001.

Figure 6. Adenovirus infection does not inhibit homology-mediated DNA repair

A. PEO4 cells were infected with dl922-957 (MOI 10). 24h post-infection, cells were fixed. Expression of γH2AX and RAD51 was assessed by confocal microscopy. B. PEO1 and PEO4 cells stably expressing DR-GFP plasmid were transfected with a plasmid encoding the rare cutting endonuclease (pI-SceI) or control plasmid. 24h thereafter, expression of GFP was assessed by flow cytometry. Data show fold change in GFP positive events following pI-SceI transfection relative to control plasmid (dotted line). Bars represent mean +/− s.d. *: p<0.05 compared to control plasmid transfection. C. PEO1 and PEO4 cells stably expressing DR-GFP plasmid were infected with dl922-947 (MOI 50 and 500 respectively). 24h later, they were transfected with pI-SceI or control plasmid. 24h thereafter, expression of GFP was assessed by flow cytometry. Data again show fold change in GFP positive events following pI-SceI transfection relative to control plasmid (dotted line). Bars represent mean +/− s.d. ***: p<0.001 compared to control plasmid transfection. D. PEO4 cells were infected with dl922-947 MOI 50 and 24 hours later treated with rucaparib (0.01 – 300 μM) for 72h.