Inhibition of PI3K-AKT-mTOR pathway sensitizes endometrial cancer cell lines to PARP inhibitors

Four endometrial cancer cell lines were used in this study (Table 1). Two cell lines were PTEN mutated and two were PTEN wild-type; all were previously purchased and kindly gifted by Dr. Jennifer K. Richer (Anschutz Medical Campus, University of Colorado, USA). HEC-50 PTEN wild-type is a High Grade (grade 3), type II, non-endometrioid endometrial adenocarcinoma cell line derived from ascitic fluid of a patient with recurrence [38, 39] and showed the capacity to differentiate into a papillary serous phenotype in a mouse model [40]. HEC-1B is a PTEN wild-type cell line that has been derived from a localized high grade endometrioid adenocarcinoma [41]. Ishikawa is a moderately differentiated (grade 2) endometrioid endometrial cancer expressing estrogen receptor and is PTEN mutated [Codon 289 del 1 bp (A) and Codons 317–318 del 4 bp (ACTT)]. These deletions are responsible for a truncated and non-functional protein, or the degradation of PTEN [42‐44]. AN3CA is a high grade endometrioid endometrial cancer cell line derived from a lymph node metastasis displaying a PTEN mutation (homozygote for codon 130 del 1 bp (G)) [43‐46]. All the cells lines were authenticated by short tandem repeat (STR) profiling by the DNA sequencing and analysis core of the University of Colorado which has extensive experience in evaluation of gynecological cell lines [47]. A minimum of 85% of similarity between a reference profile and our cell lines was observed. AN3CA, HEC-1B and HEC-50 were cultured in Eagle’s Minimum Essential Medium (EMEM) associated to 10% Foetal bovine serum (FBS) and 0.02 mg/mL gentamycin. Ishikawa cells were cultured in EMEM +2 mM Glutamine +1% Non-essential Amino Acids (NEAA) + 5% FBS + 0.02 mg/mL gentamycin [48]. Each cell line was passage every 4 – 6 days. All cells were maintained at 37 °C in a 5% CO2, 95% air atmosphere incubator. All assays were performed in the respective cell medium.

Olaparib (AZD2281), Talazoparib (BMN-673) and BKM-120 (NVP-BKM120) were ordered from AdooQ Bioscience (Catalog number #A10111, #A11243 and #A11016 respectively) diluted in 10 mM stocks in DMSO and stored at −20 °C. During experiments, aliquots of 1000-fold the final concentration were prepared in DMSO for each concentration used and stored at −20 °C. New aliquots were prepared directly from stocks every 5 – 10 uses to minimize drug degradation. Drug concentrations were designed according to the available clinical trials literature. As reported in a phase 1 clinical trial, the maximal plasma concentration of olaparib was between 3 and 8 μg/mL, which correspond to 6.9–18 μM [49]. Thus, the concentrations used in the present study ranged between 0.01 – 10 μM of olaparib which is at the lower range of that used in the clinical trial. Similarly, the maximal plasma concentration of BKM-120 was reported to be between 500 – 1500 ng/mL, corresponding to 1.2–3.7 μM [50]. Accordingly, we have used overlapping concentrations ranging from 0.1 to 5 μM of BKM-120 in our in vitro assays. With regard to Talazoparib (BMN-673), since there is still no clinical trial reporting its plasmatic concentration, its inhibitory activity was first tested in similar range of concentrations to that employed for olaparib. Following our preliminary results, we realized that Talazoparib had a smaller inhibitory dose than olaparib and modified the dosage accordingly.

Four hundred to eight hundred cells were plated in 6-well in duplicates. Cells were washed and fresh medium was added in the presence or absence of increasing doses of PI3K- inhibitor (BKM-120) and PARP-inhibitor (olaparib or BMN-673) alone and in combination after 24 h. Media containing the drug was refreshed on day 3. DMSO was used as a control. The experiment was discontinued when the clones reached 50 cells/clone in the DMSO-vector wells (7 to 12 days) and colonies were fixed and stained with 1.5 ml of 6.0% glutaraldehyde and 0.5% crystal violet and colonies were counted using the GelCount, (Oxford optronix, UK). The surviving fraction (SF) of cells was calculated as follows: SF = \( \frac{Number of colonies formed after treatment\ }{Number of cells seeded\ x\ Plating Efficiency\ } \), where Plating Efficiency = \( \frac{Number of colonies formed in control}{Number of cells seeded} \) [51]. Chou and Talalay method was used to assess the interaction between two inhibitors [52]. This method quantitatively describes the interaction between two or more drugs, with combination index (CI) values less than 1 indicating synergistic interactions, values greater than 1 indicate antagonistic interactions, and values equal to 1 indicate additive interactions. Calculations of the CI values were performed with CompuSyn Software (ComboSyn, Inc., Paramus, NJ. 07652 USA).

Proliferation assays were used to determine the inhibitory effect of drugs on the studied cell lines. Control plates were created for each cell line using 6 wells of a 24-wells plate. Ten thousand cells in 1 mL were plated in 24 well plates for drug assessment. After 24 h of standard culture at 37 °C (D0), control plates were fixed using a 4% paraformaldehyde (PFA) solution for 30 min and then stored in 0.4% PFA at 4 °C. At the same time, plates were treated with olaparib (0.01 μM, 0.1 μM, 1 μM, 5 μM and 10 μM) and BKM-120 (0.1 μM, 0.5 μM, 1 μM, 2.5 μM, 5 μM). Each concentration was tested in triplicate. DMSO was used as control. Cells were fixed using a similar procedure at day 3 (D3) and 5 (D5). All drugs and vector-controls were refreshed at Day 3. After removal of PFA, a 0.1% crystal violet/10% Ethanol solution was used to stain the fixed cells and quantify proliferation (250 μL per well during 30 min at room temperature with shaking). The wells were then aspirated and allowed to air-dry at least 2 h. A 10% acetic acid was used to dissolve the staining dye (500 μL/well). At least, the 200 μL of each well were transferred into a 96-wells plate, before the absorbance was measured at 590 nm by spectrophotometry, as it is assumed that the level of absorbance is proportional to the number of cells in the well at the time of the fixation.

Cells were harvested (2 mL 0.25% Trypsin-EDTA 1×, Wisen Bio Products) and then lysed in 500 μL of radioimmunoprecipitation assay (RIPA) buffer (25 mM/L Tris-HCl pH 7.6, 150 mM/L NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS and 1 mM/L EDTA). Protein concentration was determined using bicinchoninic acid assay (BCA) kit (Ref 23,227, Pierce) using a spectrophotometer at 570 nm. Protein lysates (10–25 μg) were separated electrophoretically on a 7.5 – 12% denaturing SDS-polyacrylamide gels and transferred to 0.2 μm nitrocellulose membranes. Primary antibodies specific for PTEN (#9552; Cell Signaling, Beverly, MA, USA. 1:1000), PI3K (#4238; Cell Signaling; 1:500), phospho-PI3K (#4284; Cell Signaling; 1:500), AKT (#9272; Cell Signaling; 1:1000), phospho-AKT (Ser473, #9271S; Cell Signaling; 1:1000), S6 Ribosomal Protein (#2217; Cell Signaling; 1:1000), phospho-S6 (Ser240/244, #2215; Cell Signaling; 1:1000), and β-actin (#4967, Cell Signaling; 1:2000) were diluted in 0.1% Tween-PBS/5% Milk and put in presence of the membrane overnight at 4 °C. After 3 washing (0.1%Tween-PBS1X), membranes were exposed to secondary anti-rabbit-horseradish peroxidase (HRP; L170–6515; Bio-Rad, USA; 1:10,000) or anti-mouse HRP (L170–6516; Bio-Rad; 1:10,000) for 1 h at room temperature. Immunoreactive proteins were detected by chemiluminescence (WBKLS0500; Immobilon Western, Millipore) and autoradiography [53].

PTEN specific small hairpin RNA (shRNA) containing the following sequence: CCGGCCACAAATGAAGGGATATAAACTCGAGTTTATATCCCTTCATTTGTGGTTTTT were ordered in Bacterial Glycerol Stock (#TRCN0000002749, Sigma-Aldrich, Saint-Louis, MO, USA). shRNA were annealed 4 min at 95 °C in a PCR machine, inserted into pLKO.1 cloning vector (gift from Bob Weinberg, Addgene plasmid # 8453) and amplified in DH5-alpha bacterial cells before antibiotic selection by 100 μg/mL of ampicillin. PTEN wild type cell lines (HEC-50 and HEC-1B) were plated at approximately 30% confluence in 100-mm plates and incubated for 24 h before transfection with 2 μg of pLKO.1 PTEN shRNA plasmid. An empty plasmid was used as a control (pLKO.1 puro, Addgene plasmid #8453). Successfully transfected cells were selected using puromycin [20].

1 × 10⁵ cells / well were seeded in 6-well plates on a sterile cover slip. After Twenty-four hours, when cells reached ~60% confluency, then cells were washed and the medium was replaced with medium containing 500 nM doxorubicin for 1 h and allowed to recover for ~6 h. In another setting, the cells were treated with medium containing 1 μM BKM-120 for 24 h, followed by 500 nM doxorubicin for 1 h and allowed to recover for ~6 h. Fixation, permeablization and antibodies staining were performed as described earlier [20]. Images were analyzed and quantified using ImageJ software (NIH) [20].

We evaluated PTEN protein expression in endometrial cancer cells, as shown in Fig. 1a. Western blot analysis confirmed that HEC-50 and HEC-1B cells were PTEN wild-type, whereas Ishikawa and AN3CA cells carrying PTEN deletion mutations (resulting in PTEN truncated or degraded proteins), did not express wild type PTEN protein, confirming previously published data about their PTEN mutated status [43, 54]. In standard culture condition, Ishikawa had high levels of phospho-AKT (p-AKT) and phospho-S6 (p-S6), thus, as expected from PTEN mutated cell lines, it is expressing an activated PI3K-AKT-mTOR pathway. Conversely, AN3CA displayed a lower level of p-AKT but a high level of p-S6. HEC-50 also had a lower level of p-AKT but p-S6 was present similarly to the PTEN mutated cell lines. Therefore, to confirm the effect of PTEN on pAKT and pS6 expression level, HEC-50 and HEC-1B cells were transfected with shPTEN and were assessed by western blot. As shown in Fig. 1b, we found reduced levels of PTEN protein as well as increased p-AKT and p-S6 in the PTEN-depleted cells compared to the control cells.

We next evaluated the sensitivity of the endometrial cancer cells to PARP-inhibitors by clonogenic and cell proliferation assays. We observed that the mutated cell lines (AN3CA and Ishikawa) were more sensitive to the PARP-inhibitor olaparib compared to the PTEN wild-type cell lines (HEC-50 and HEC-1B) after 5 days (Fig. 2a). Similar data was observed after 3 days (data not shown). The median inhibitory concentration (IC50) was 0.5 μM for AN3CA cells and 3 μM for ishikawa cells compared to 6 μM for HEC-50 cells and 10 μM for HEC-1B cells. The same pattern was observed with a more potent PARP-inhibitor, BMN-673 (Fig. 2b), with IC50 at 0.005 μM for AN3CA cells and 0.008 μM for ishikawa cells. To confirm the difference in sensitivity to PARP-inhibitors, between the PTEN mutated to the PTEN non-mutated cells, we transfected the PTEN wild-type cells, HEC-50 and HEC-1B with shPTEN. As shown in Fig. 2c-d and Table 2, the transfected shPTEN cells were more sensitive to both olaparib and BMN-673 compared to their un-transfected counterparts. A proliferation assay showed overall similar results (data not shown).

We evaluated whether HR impairment provoked by PI3K inhibition conferred increased sensitivity to PARP inhibition in the context of PTEN mutation. We cultured these endometrial cancer cell lines in a fixed, low dose of olaparib (0.1 μM and 0.5 μM) and increasing drug concentrations of BKM-120 ranging from 0 to 10 μM. The idea was to mix a sublethal dose of olaparib (number of growing clones comprised between 0.8 – 1.2 fold the control number), with increasing concentrations of BKM-120 to see if it was enough to induce a stronger inhibitory effect on cell growth compared to that of BKM-120 by itself. As shown in Fig. 3a. we found a significant decrease in the clonogenic indices of PTEN-mutated cells with combination treatment as compared to PI3K and PARP-inhibitor alone. Further, to determine if the PI3K and PARP-inhibitors combination were additive or synergistic, we used the ‘multiple drug effect analysis’ method of Chou and Talalay (see Materials and Methods) [52]. Interestingly, in both PTEN mutated cell lines tested, we observed the combination treatment to be strongly synergistic, with a CI ˂1.0 (Fig. 3c). When we looked at the PTEN wild-type cells, we also found decreased survival with combination treatment as compared to PI3K and PARP-inhibitor alone (Fig. 3b). However, in both wild type cell lines tested, we observed the CI to be close to 1, suggesting an additive effect (Fig. 3c). The CI for HEC-50 was 1.5 (olaparib 0.1 μM and BKM-120 1 μM) and 1.02 for HEC-1B (olaparib 0.1 μM and BKM-120 1 μM). This effect was less pronounced in shPTEN transfected cells, with somewhat lower CI: 1.2 for HEC-50, shPTEN (olaparib 0.1 μM and BKM-120 1 μM) and 0.86 for HEC-1B, shPTEN (olaparib 0.1 μM and BKM-120 1 μM).

We then assessed the correlation of PTEN protein expression and HR functionality in these cells. To assess HR functionality, we evaluated the baseline levels of RAD51 protein expression and foci formation in the PTEN-mutated and wild-type endometrial cancer cells upon DNA damage, with 1 h treatment of 500 nM doxorubicin, by immunofluorescence, as an indication of the cells ability to repair DNA double strand break [51]. As shown in Fig. 4a only one of the two PTEN mutated cells (Ishikawa) did not express RAD51 protein at all. In addition, the two shPTEN transfected wild type cells expressed a decreased amount of RAD51 (Fig.4b). As shown in Fig. 4c and d, we observed a significant reduction in nuclear RAD51 foci formation after doxorubicin treatment in cells having low or absent PTEN protein expression, suggesting a lower capacity of DNA repair through HR. Next, using RAD51 foci formation assay, we found that the cells treated with the 1 μM PI3K- inhibitor (BKM-120), showed decreased formation of RAD51 foci after induction of DNA damage with 500 nM doxorubicin treatment, as compared to cells treated with 500 nM doxorubicin treatment alone (Fig. 4e and f). This reduction was probably due to a decrease in RAD51 protein levels in response to treatment with the PI3K-inhibitor (data not shown) suggesting that PI3K inhibition directly impacts HR functionality in our studied cancer cells.