ATM-depletion in breast cancer cells confers sensitivity to PARP inhibition

Human breast cancer cell lines, MCF-7 and ZR-75-1, and their transfected-derivatives were maintained in DMEM-Glutamax and RPMI-Glutamax, respectively, supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 U/ml streptomycin (all from Invitrogen). All cell lines were maintained in a 5% CO₂ atmosphere at 37°C. Cells were passaged once every 3–5 days (~90% confluence) and all experiments were performed within the first 10 passages from transfection. For drug treatment, doxorubicin (Sigma) and PARP inhibitors, olaparib and iniparib (Selleckchem), were prepared as stock solution in water or DMSO, respectively, aliquot and stored at -80°C until use.

Stable interference was obtained by retroviral-mediated expression of short-hairpin RNA (shRNA) using pRETRO-Super vector. Retroviruses were produced in HEK 293 T cells by cotransfecting pRETRO-Super together with plasmids encoding for gag-pol and VSV-G proteins. Viral supernatant was collected 48 hrs post-transfection, filtered through a 0.45 μm pore size filter and added to the cells in the presence of 2 μg/ml polybrene. After 48 hrs from infection, stable polyclonal populations of control and ATM-depleted cells were obtained by selection for two weeks with 2 μg/ml puromycin (Sigma).

The shATM construct (#1 position 912) in pRETRO-Super, generously provided by Y. Lerenthal and Y. Shiloh, has the following sequence: 5′-GAC TTT GGC TGT CAA CTT TCG-3′ [24]. Control shRNA, siR5, has the following sequence: 5′-GGA TAT CCC TCT AGA TTA-3′. Neither the ATM-targeting shRNA nor the control sequences have any homology with other human gene as tested by BLAST (http://​blast.​ncbi.​nlm.​nih.​gov/​Blast.​cgi).

Total cell extracts were prepared in lysis buffer [50 mM Tris–HCl pH 8, 300 mM NaCl, 1 mM EDTA, 0.5% sodium deoxycholate, 0.1% SDS, 1% Nonidet-P40, 1 mM EDTA] supplemented with protease-inhibitor mix (Roche), resolved on precast NuPAGE 4-12% gels (Invitrogen), and transferred onto nitrocellulose membranes (Bio-Rad). The following antibodies were employed for immunedetection: rabbit anti-ATM (Santa Cruz), mouse anti-α-tubulin (Immunological Sciences), HRP-conjugated goat anti-mouse and anti-rabbit (Cappel). Immunoreactivity was determined using the ECL-chemiluminescence reaction (Amersham Corp) following the manufacturer’s instructions.

When indicated, cells were irradiated using a ¹³⁷Cs source (IBL-437-C irradiator, CIS bio International) at a dose rate of 6.8 Gy/min.

Cells (5 × 10⁴/ml) were seeded in 96-well plates in growth medium and incubated 24 hrs at 37°C in 5% CO₂ atmosphere. Drugs were added at the indicated concentrations and for the indicated times before incubation with reagents of XTT, WST-1, and BrdU (all from Roche Applied Science), following the manufacturer’s instructions. The absorbance at 450 nm (XTT and WST-1) or at 370 nm (BrdU) were measured by the microplate reader Infinite F200 (Tecan). Each experiment was performed in triplicate. The survival fraction for a given dose was calculated as the plating efficiencies for that dose divided by the plating efficiencies of solvent-treated cells.

Treated and untreated cells (5 × 10⁵) were washed in PBS 1X and resuspended in 300 μl hypotonic fluorochrome solution [50 μg/ml propidium iodide, 0.1% sodium citrate, 0.1% Triton-X-100 (all from Sigma)] for 30 min at room temperature. DNA content was measured by a FACScan flow cytometer (Becton Dickinson).

Cells were treated with drugs at the indicated doses for 24 hrs, then plated at low density in 60 mm Petri dishes and grown for twelve days in the absence of drugs. Surviving colonies were fixed and stained with Cristal Violet (0.5% in methanol) (Sigma), air-dried, and counted.

The Wilcoxon test for paired samples has been used for repeated measurements. A p-value less than 0.10 (*) and less than 0.05 (**) were considered statistical significant.

To assess the influence of ATM in breast cancer susceptibility to PARP inhibitors, we genetically repressed ATM expression by RNA interference in MCF-7 cells. We chose the MCF-7 breast cancer cell line because it is ER positive, HER2 negative, and wild-type for the BRCA1, BRCA2, and TP53 genes [25], features we observed in breast tumors arising in our A-T heterozygotes [23]. Stable interference of ATM was obtained by MCF-7 transfection with shATM-carrying vectors (MCF7-ATMi) and its siR5 negative control (MCF7-ctr) (see Materials and methods). Stable-transfected cells were selected in the presence of puromycin for ten days and maintained as polyclonal populations. As shown in Figure 1A, a strong repression of ATM expression was obtained in the MCF7-ATMi cells compared to the MCF7-ctr ones. To verify whether ATM-depletion has a functional impact on MCF-7 cells, we assessed the sensitivity of ATM-depleted and control cells to IR and doxorubicin treatment, that are known to induce different outcomes in ATM proficient and defective cells. In particular, radiosensitivity is a defining feature of ATM-defective cells [26] whereas, in a wild-type p53 context, doxorubicin-resistance was shown to characterize ATM-deficient cells in vitro [27] and in breast cancer patients [28]. As shown in Figure 1B and 1C, MCF7-ATMi cells were more sensitive to IR and more resistant to doxorubicin than MCF7-ctr cells. The contribution of ATM in the latter result was confirmed in MCF-7 parental cells by KU 55933-induced ATM inactivation (Figure 1D). These results were further confirmed by evaluating the cell cycle profiles (Figure 1E). After 24 hrs from irradiation, both MCF7-ctr and MCF7-ATMi cells show the expected enrichment into the G2/M phase. After 48 hrs from irradiation, MCF7-ctr cells repair the damage and re-enter into the cell cycle; in contrast, MCF7-ATMi cells, which are known to have defects in sensing and repairing DNA double strand breaks [26], show a delay in re-entering into the cell cycle. In contrast, as expected from the data reported by Jiang and co-workers [27], the ATMi cells were more resistant to doxorubicin and a lower proportion of cells underwent cell death.

Altogether, these results show that MCF-7 transduction with shATM-carrying vectors interferes with ATM expression and elicits some aspects of a phenotype compatible with ATM-deficient cells.

To evaluate whether ATM-depletion modifies MCF-7 response to PARP inhibitors, we first used olaparib (AZD2281, Ku-0059436), an orally bioavailable compound whose effectiveness in BRCA1/2 mutated breast and ovarian cancers was studied in phase II clinical trials and, for ovarian cancers is under further evaluation in phase III clinical studies [12]. MCF7-ATMi and MCF7-ctr cells were incubated with increasing concentrations of olaparib or its solvent (DMSO) for 72 hrs and their viability assessed by XTT or WST-1, with comparable results. As shown in Figure 2A, ATM-depleted cells were mildly but significantly more sensitive than MCF7-ctr cells to olaparib. However, MCF7-ctr cells, as well as the parental MCF-7 cells (data not shown) were not completely resistant to olaparib and their viability declined with time (Figure 2B) and at the highest doses we employed (Figure 2A, 10 μM dose).

To further characterize the effect induced by olaparib, MCF7-ATMi and MCF7-ctr cells were treated for 48 hrs with 2.5 and 5 μM olaparib and their DNA content assessed by propidium iodide staining and FACS analysis. Consistently with the viability assays described above, cell death, measured by the appearance of hypodiploid cells, was detected only in the olaparib-treated MCF7-ATMi cells (Figure 2C). However, both ATM-depleted and control MCF-7 cells arrested in the G₂/M phase of the cell cycle, in a dose-dependent manner, as previously described [2]. The similarity in the cell cycle behavior between MCF7-ATMi and MCF7-ctr cells after olaparib treatment was confirmed by BrdU assay that showed a comparable reduction in the two cell populations (Figure 2D). These data indicate that MCF-7 sensitivity to olaparib is increased by ATM-depletion, but these cells are partially responsive to this compound, as also recently reported by others [29].

Next, we verified the long-term effect of olaparib by performing colony formation assays. MCF7-ATMi and MCF7-ctr cells were treated for 24 hrs with 0.5 and 1 μM olaparib, then plated at low density and grown for twelve days in the absence of drug. As shown in Figure 2E, a significant reduction in the colony forming capacity was observed in the ATM-depleted cells compared to the controls. Consistent with the results described above, a mild reduction in colony formation was also observed in the olaparib-treated MCF7-ctr cells compared with their DMSO-treated controls (Figure 2E, blue columns).

Overall, these data indicate that ATM-depletion increases sensitivity to olaparib in breast cancer MCF-7 cells; however, factors other than ATM might contribute to the response of this cell line to this PARP-inhibitor.

Next, we asked whether ATM-depletion can sensitize MCF-7 cells to iniparib (BSI-201, SAR240550), a compound originally described as an irreversible inhibitor of PARP-1 [30], but recently shown to act as a nonselective modifier of cysteine-containing proteins [31, 32]. MCF7-ATMi and MCF7-ctr cells were treated with iniparib or its solvent, DMSO, and analyzed for colony formation capacity, DNA content by FACS analysis, and BrdU assay. As shown in Figure 3A, ATM-depletion reduced the ability of MCF-7 cells to produce colonies after iniparib-treatment while no effect was observed in MCF7-ctr cells. At variance with olaparib-treatment, DNA content analysis did not reveal any significant difference between MCF7-ATMi and MCF7-ctr cells in the appearance of hypodiploid, death cells, whereas only the MCF7-ATMi population experienced an accumulation of cells in the G₂/M phase of the cell cycle (Figure 3B). This effect on the cell cycle was confirmed by BrdU assays (Figure 3C). Together, these results suggest that ATM-depletion can also influence MCF-7 cell response to iniparib.

To further assess the impact of ATM-depletion in breast cancer cell response to olaparib and iniparib, we selected the ZR-75-1 line, whose cells, like the MCF-7 ones, are ER positive, HER2 negative, and wild-type for BRCA1/2 and TP53 genes [25]. Stable interference of ATM in ZR-75-1 cells was obtained as described for MCF-7 cells. Polyclonal populations, ZR-ATMi and ZR-ctr, were obtained by puromycin selection and ATM-depletion confirmed by Western blot analysis (Figure 4A). Next, dose–response viability assays were performed on ZR-ATMi and ZR-ctr cells upon incubation with olaparib, iniparib, or their solvent, DMSO. As shown in Figures 4B, ZR-ctr cells were strongly resistant to olaparib whereas their ATM-depleted counterpart became considerably sensitive and showed a partial accumulation in the G2/M phase of the cell cycle (Figure 4D). These results, confirmed by colony formation assays (Figure 4E), sustain the observations made with MCF-7 cells and support a synthetic lethal relationship between ATM-depletion and olaparib-treatment in ER positive, wild-type BRCA 1/2 breast cancer cells.

In contrast with the sensitivity induced by ATM-depletion in MCF-7 cells, when treated with iniparib, both ZR-ATMi and ZR-ctr cells showed a substantial loss of viability that was independent of ATM, as indicated by the similarity of their survival curves (Figure 4C) and cell cycle distribution (Figure 4D). These results were confirmed by the complete inhibition of colony formation induced by iniparib in ZR-75-1 cells, independent of their ATM status (Figure 4F). In addition, the different response between MCF-7 and ZR-75-1 cells to this drug suggests that ER expression and the wild-type status of BRCA1/2 and TP53 are not involved in the sensitivity to iniparib. These results might be explained by the recent observations indicating that the primary mechanism of action for iniparib is a nonselective modification of cysteine-containing proteins, rather then inhibition of PARP activity [32].