Mycoplasma-free MCF10A cells (cat. no. CRL-10317, freshly obtained from ATCC) were cultured in monolayers using equal volumes of DMEM-glutamax and F12-glutamax (Life Technologies; Thermo Fisher Scientific) supplemented with 5% fetal calf serum (Invitrogen; Thermo Fisher Scientific), antibiotics (50 U/ml penicillin and 50 µg/ml streptomycin, Invitrogen; Thermo Fisher Scientific), 10 µg/ml insulin (Sigma-Aldrich), 0.5 µg/ml hydrocortison (Sigma-Aldrich) and 20 ng/ml epidermal growth factor (Peprotech). Experiments were performed on cells in which BRCA1 and BRCA2 were knocked down (these cells are referred to as BRCA1i and BRCA2i cells). RNA interference (RNAi) was achieved by stable transduction with lentiviral vectors harboring DNA sequences encoding short hairpin RNAs specific for BRCA1 or BRCA2. In brief, lentiviral particles were constructed using pLKO.1-puro vectors (Addgene). The RNAi sequences for BRCA1 and BRCA2 were 5'-GCCCACCTAATTGTACTGAAT-3' and 5'-TACAATGTACACATGTAACAC-3', respectively. A negative control cell line, transduced with an empty pLKO.1-puro lentiviral vector (hereafter referred to as the control) was also established. This is an acknowledged limit of this study, as scrambled sequence transduction would have been preferable, though according to previous experience, this procedure may target unintended mRNAs. Moreover, empty vectors allow for the determination of the effects of transduction on cell response and gene expression (20). The transduction of the MCF10A cells was achieved by the addition of 1 µg/ml DNA, TurboFect (1.5 µg/ml, Thermo Fisher Scientific) and polybrene (1 µg/ml, Sigma-Aldrich) to a 30% confluent culture. Cells were grown in puromycin-supplemented DMEM medium (2 µg/ml, Life Technologies; Thermo Fisher Scientific) for 15 days to obtain stably transduced cell lines. Notably, stable BRCA1 knockdown by retrovirus-mediated RNAi does not alter the non-tumorigenic phenotype of MCF10A cells (21). According to our personal experience, BRCA1/2 knockdown neither induces transformation in vitro, nor tumorigenesis in vivo.

BRCA1 and BRCA2 mRNA knockdown was evaluated by RT-qPCR analysis. Total RNA was extracted using RNeasy Mini kits (Qiagen Benelux), following the manufacturer's instructions, without optional DNase treatment. The RNA concentration and quality were determined using a DropSense96 kit (Trinean) before removing contaminating DNA with the Heat&Run gDNA removal kit (Articzymes). Reverse transcription was achieved with the iScript cDNA synthesis kit, following the manufacturer's instructions. A total of 5 µl of qPCR reaction contained 10 ng cDNA (total RNA equivalents), 2.5 µl SsoAdvanced universal SYBR-Green supermix (Bio-Rad) and 2.5 µM forward and reverse primers. Cycling and detection was performed on a Lightcycler LC480 (Roche) with 2 min denaturation at 95˚C followed by 45 cycles with 5 sec at 95˚C, 30 sec at 60˚C and 1 sec at 72˚C. Melting curve analysis was performed to test for non-specific amplification. Three biological repeats were performed for each amplification. The analysis of relative gene expression was performed using a qBase+ platform (Biogazelle), developed by Hellemans et al (22), based on the ΔΔCq method by Livak and Schmittgen (23).

BRCA1 and BRCA2 protein knockdown was evaluated by western blot analysis. Protein extraction was performed in subconfluent cultures of the control, BRCA1i and BRCA2i cells using a tris-EDTA lysis buffer containing 1% NP-40 and 1% protease inhibitor (Sigma-Aldrich). For each sample, 50 µg protein were loaded together with 25% SDS sample buffer (Thermo Fisher Scientific) and 10% dithiothrietol (DDT, Sigma-Aldrich) on a 3-8% Tris-Acetate gel (Novex; Thermo Fisher Scientific), and run for 5 h at 25 mA. Proteins were transferred to a methanol-pretreated PVDF membrane in Tris/glycine blotting buffer enriched with 10% methanol for 16 h at 30 V. Following standard blocking, the membrane was incubated overnight at 4˚C with a primary antibody (rabbit polyclonal anti-BRCA1, cat. no. 07-434, diluted 1:1,000, or monoclonal anti-BRCA2, cat. no. OP95, diluted 1:500; Millipore), together with mouse monoclonal α-actinin (cat. no. 05-384, diluted 1:30,000 Millipore). The membranes were washed and incubated for 2 h at room temperature with horseradish peroxidase (HRP)-conjugated secondary antibodies (goat anti-rabbit-HRP, Perbio, cat. no. 31460, diluted 1:1,000, or goat anti-mouse-HRP, cat. no. 31432, Thermo Fisher Scientific, diluted 1:1,000). Visualization was achieved with a chemoluminescence kit (Thermo Fisher Scientific), and analysis of relative expression at the protein level was performed using a shareware ImageJ software (version 1.52q, bundled with Java 1.8.0_172; developer: NIH).

The HR pathway is active in the S and G2 phases of the cell cycle to repair DSBs induced by IR (2,24-26), and relies on RAD51, whose recruitment to a DSB site is mediated by BRCA1 and BRCA2. RAD51 recruitment can be typically detected in the form of nuclear foci upon immunostaining (25,27). The RAD51 irradiation-induced foci assay has been widely used to detect HR defects both in cancer biopsy samples and in non-tumor cells of BRCA1/2 breast cancer patients (28-30).

The cells were switched to puromycin-free culture DMEM medium (Life Technologies) at the beginning of each experiment. Cells (n=200,000) were seeded in 2 ml culture medium in 6-well plates. Approximately 24 h after seeding, the cells were examined for subconfluency. To achieve a maximum number of cells in the S and G2 phases of the cell cycle, the cells were synchronized by the addition of the DNA polymerase inhibitor, aphidicolin (1 µg/ml, Sigma-Aldrich) to the culture medium for 24 h. The cells were subsequently washed with PBS (1.78 g/l Na₂HPO₄; 0.42 g/l KH₂PO₄; 7.2 g NaCl, VWR) and incubated at 37˚C for 8 h with fresh culture medium.

The MCF10A control cells (empty vector-transduced) were harvested at various time points following synchronization to evaluate the percentage of cells in each phase of the cell cycle. Cell permeabilization was achieved by fixation in 95% ethanol at -20˚C and the DNA was subsequently stained with propidium iodide (PI, Sigma-Aldrich) in a hypotonic staining buffer containing 0.1% Sodium citrate, 0.3% Triton X-100, 0.01% PI and 0.002% ribonuclease A (all reagents were from Sigma-Aldrich). The PI cell content was analyzed on a FACSCanto^TM (BD Biosciences). Cells of interest were selected based on forward and side scatter area patterns. A non-synchronized sample was used as reference.

At 3 h following aphidicolin removal, the cells were irradiated with 2 Gy 220 kV-13 mA X-rays to generate DSBs (SARRP unit, XSTRAHL Ltd.). The cells were subsequently incubated at 37˚C for 5 h for optimal RAD51 foci formation. To determine this optimal time point, different time points varying between 2 and 8 h were previously tested (data not shown).

Since cells defective in HR are sensitive to poly(ADP-ribose) polymerase (PARP) inhibitor (PARPi) drugs (31-33), the PARPi, olaparib, was used to enhance the number of DSBs in S-phase cells. To part of the cultures, 5 µM olaparib (Bio-Connect) was added to the medium 1 h prior to irradiation, in order to block the repair of radiation-induced single-strand breaks (SSBs). RAD51 foci formation was also evaluated in untreated cell cultures and in cultures exposed to olaparib alone. In total, 8 repeats were performed for each of the 4 experimental conditions (irradiated or sham-irradiated cells, with or without olaparib treatment).

The induction of DNA DSBs by IR was assessed by gamma-H2AX staining as previously described (34) (Fig. S1). Inhibition by olaparib was confirmed by the PARP activity following exposure to H₂O₂ in the presence of olaparib, as previously described (35) (Fig. S2).

Prior to RAD51 foci staining, the cells were harvested, cytospinned on polylysine-coated slides (VWR) and fixed in 3% paraformaldehyde (Sigma-Aldrich) for 20 min. The slides were washed twice in PBS and antigen retrieval was achieved by incubation for 20 min in heated (95˚C) citrate buffer (0.02% citric acid, pH 6, Sigma-Aldrich). The slides were subsequently washed and incubated with a blocking serum containing 1% BSA (Roche), 5% goat serum (Dako) and 0.2% Tween-20 (Sigma-Aldrich) in PBS. The slides were then incubated overnight at 4˚C with a RAD51 H-92 rabbit primary antibody (dilution: 1:2,000, Santa Cruz Biotechnology, cat. no. sc-8349), washed with PBS containing 3% Tween-20 (Sigma-Aldrich), and incubated for 30 min at room temperature with a secondary antibody (goat anti-rabbit; dilution: 1:1,000, Thermo Fisher Scientific; cat. no. A32731). Finally, the slides were washed in PBS/3% Tween-20 and mounted with 200 ng/ml DAPI in fluoromount (Sigma-Aldrich).

The slides were scanned using the Metacyte software module on a Metafer4 scanning platform (Axio Imager, Metasystems) at a x63 magnification. This software module enables automatic cell detection and foci counting according to set parameters, resulting in an unbiased data acquisition. The number of RAD51 foci was automatically scored in at least 500 cells for each experimental condition, and expressed as the number of RAD51 foci per cell (RAD51 foci/cell).

Chromosomal damage was assessed with the MN assay as previously described (36). Briefly, the cells were seeded in 2 ml culture medium in 6-well plates (200,000 cells/well) 1 day prior irradiation. Subconfluent cultures of exponentially dividing cells were irradiated with doses of 0.2, 0.5, 1, 2 and 4 Gy, and cytochalasin B (2.25 µg/ml; Sigma-Aldrich) was immediately added to block cytokinesis. The cells were maintained at 37˚C in a humidified 5% CO₂ atmosphere incubator for 16 h. Sham-irradiated cultures were included in each experiment. The cells were then harvested and subjected to a cold hypotonic shock with 0.075 M KCl, followed by overnight fixation in 3/1/4 methanol/acetic acid/ringer solution (ringer: 9 g/l NaCl, 0.42 g/l Kcl and 0.24 g/l CaCl₂). Subsequently, the cells were fixed in a 3:1 methanol/acetic acid solution. For further analysis, the cells were stained with DAPI (200 ng/ml; Sigma-Aldrich). The slides were scanned at 10 X magnification with the MSearch software module of the Metafer 4 scanning system and the MNScore software (Metasystems). The automated image analysis system selects BN cells and determines the number of MN for each BN cell. BN cells and MN were manually examined for false positives and negatives. Two slides for each culture were automatically scanned and approximately 400 BN cells were scored in each slide. All experiments were performed in duplicate. Each experiment was repeated thrice.

A protocol described previously was used for this assay (37,38). Briefly, following irradiation (doses of 0.5, 1, 2, 3, 4, 6 and 8 Gy), the cultures were further incubated for 4 days at 37˚C until sham-irradiated plates nearly reached confluence. The cells were fixed for 10 min in a solution of buffered formalin (3.7%), washed with PBS (pH 7.3) and stained with a 0.01% crystal violet solution (Sigma-Aldrich). The stain was dissolved overnight in 1 ml 10% sodium dodecyl sulfate (SDS). The optical density of the samples was measured with a spectrophotometer at 590 nm. All cell viability assays were performed in quadruplicate. Each experiment was repeated 3 times.

Differences in mean RAD51 foci, in micronucleus counts and in cell viability data were analyzed by one-way analysis of variance (ANOVA). The Tukey's range test was applied to perform post-hoc analysis of the significance of comparisons. A 5% alpha error threshold (P-value <0.05) was applied to all analyses. Statistical inference was performed using the R software environment. For the Tukey post-hoc test, the multcompView package in R was used.

The lentivirus-mediated RNA interference of BRCA1 and BRCA2 was confirmed at the RNA and protein levels by RT-qPCR analysis and western blot analysis. Compared to the control cells, 45 and 35% reductions in the BRCA1 and BRCA2 mRNA levels were achieved, respectively (Fig. S3). Notably, in experiments assessing knockout specificity, it was demonstrated that the knockdown of BRCA1 in the BRCA1i cells did not affect the mRNA levels of BRCA2 (97% of controls), and the knockdown of BRCA2 in the BRCA2i cells did not affect the mRNA levels of BRCA1 (97% of controls). The quantification of the protein knockdown using ImageJ software revealed an estimated 70% reduction of BRCA1 in BRCA1i cells, and an estimated 51% BRCA2 reduction in BRCA2i cells (Fig. 1).

The results of cell cycle analysis in non-synchronized cells, and in synchronized cells at various time points following aphidicolin removal are shown in Fig. 2. As an example, the histogram charts of all time points for one repeat are illustrated in Fig. 2A-E. The mean percentage (4 repeats) of cells in each phase of the cell cycle in a non-synchronized sample, and in synchronized cells at various time points (0, 2, 3 and 8 h) following aphidicolin removal is shown in Fig. 2F.

Synchronization with aphidicolin increased the number of cells in the S phase of the cell cycle. Immediately following aphidicolin synchronization, approximately 70% of the cells were at the beginning of the S phase, compared to 30% in non-synchonized cells. At 2 and 3 h following aphidicolin removal, the lowest number of cells in the G1 phase was achieved, while the number of cells in the S phase remained between 60 and 70%, compared to the non-synchronized cultures. At 8 h following aphidicolin synchronization, the cells shifted towards the G2 and M phases of the cell cycle. Since the number of cells in the late S phase peaked at 2 and 3 h following synchronization, these time points were selected for the addition of olaparib (2 h following the removal of aphidicolin) and irradiation (3 h following the removal of aphidicolin) to maximize the effects of these agents on RAD51 foci formation.

Representative examples of the induction of RAD51 foci by radiation in the nuclei of MCF10A cells, with or without a knockdown of BRCA1 or BRCA2 are illustrated in Fig. 3. The mean number of RAD51 foci per cell in synchronized cell cultures, assessed in the 4 tested conditions (irradiated or sham-irradiated, with or without olaparib) is shown in Fig. 4. The significance values for all comparisons are listed in Table SI.

Compared to the cells not irradiated and not exposed to olaparib, exposure to IR induced marked and significant increases in the mean RAD51 foci/cell, both in the presence (~2.3-fold) or absence (~1.9-fold) of olaparib (Fig. 4, blue bars and Table SI).

The exposure of sham-irradiated control cells to olaparib did not significantly increase the number of foci/cell. Conversely, in the irradiated control cells, exposure to olaparib caused a small, yet significant increase in RAD51 foci/cell (~1.2-fold; Fig. 4, blue bars and Table SI).

Compared to their respective controls, a significantly lower yield of RAD51 foci was observed in the irradiated BRCA1i and BRCA2i cells, exposed (BRCA1i cells, 49% reduction; BRCA2i cells, 60% reduction) or not (BRCA1i cells, 58% reduction; BRCA2i cells, 64% reduction) to olaparib (Fig. 4 and Table SI).

Compared to the BRCA1i or BRCA2i cells not irradiated and not exposed to olaparib, exposure to IR did not induce statistically significant increases in the mean RAD51 foci/cell, both in the presence or absence of olaparib (Fig. 4 and Table SI).

The spontaneous MN yield (mean number of micronuclei per 1,000 sham-irradiated binucleated cells ± SD) for the control, BRCA1i and BRCA2i cells was 28.1±2.5; 31.7±4.2 and 26.0±1.7, respectively. There was no significant difference in the spontaneous MN values between the BRCA1i or BRCA2i cell lines and control MCF10A cells.

The MN dose-response curves in the control cells, and in the BRCA1i and BRCA2i cells are shown in Fig. 5. Compared to the irradiated controls, significantly higher MN yields were obtained in both the BRCA1i and BRCA2i irradiated cell lines. The BRCA1i cells were the most radiosensitive, resulting in a steeper, quasi-linear dose-response curve. The P-values for all the comparisons are listed in Table SI.

Cell viability/survival curves of knockdown cell lines together with the control cell line are shown in Fig. 6. The BRCA1i and BRCA2i irradiated cells exhibited a dose-dependent decrease in survival, when compared to the control cells. The P-values for all the comparisons are listed in Table SI.