Combined inhibition of BADSer99 phosphorylation and PARP ablates models of recurrent ovarian carcinoma.

Background: Poly (ADP-ribose) polymerase inhibitors (PARPis) have been approved for the treatment of recurrent epithelial ovarian cancer (EOC), regardless of BRCA status or homologous recombination repair deficiency. However, the low response of platinum-resistant EOC, the emergence of resistance in BRCA-deficient cancer, and therapy-associated toxicities in patients limit the clinical utility of PARPis in recurrent EOC.
Methods: The association of phosphorylated (p) BADS99 with clinicopathological parameters and survival outcomes in an EOC cohort was assessed by immunohistochemistry. The therapeutic synergy, and mechanisms thereof, between a pBADS99 inhibitor and PARPis in EOC was determined in vitro and in vivo using cell line and patient-derived models.
Results: A positive correlation between pBADS99 in EOC with higher disease stage and poorer survival is observed. Increased pBADS99 in EOC cells is significantly associated with BRCA-deficiency and decreased Cisplatin or Olaparib sensitivity. Pharmacological inhibition of pBADS99 synergizes with PARPis to enhance PARPi IC50 and decreases survival, foci formation, and growth in ex vivo culture of EOC cells and patient-derived organoids (PDOs). Combined inhibition of pBADS99 and PARP in EOC cells or PDOs enhances DNA damage but impairs PARPi stimulated DNA repair with a consequent increase in apoptosis. Inhibition of BADS99 phosphorylation synergizes with Olaparib to suppress the xenograft growth of platinum-sensitive and resistant EOC. Combined pBADS99-PARP inhibition produces a complete response in a PDX derived from a patient with metastatic and chemoresistant EOC. Conclusions: A rational and efficacious combination strategy involving combined inhibition of pBADS99 and PARP for the treatment of recurrent EOC is presented.

Introduction

Recurrent epithelial ovarian cancer (EOC) remains the most lethal among all gynecological cancers[1]. Despite advances in early detection and treatment, the majority of EOC patients experience disease recurrence[2]. The median survival of patients with recurrent EOC ranges from 12–24 months[3]. Targeted therapies such as poly (ADP-ribose) polymerase (PARP) inhibitors have emerged as a treatment option for recurrent EOC, particularly in patients with deficient BRCA1/2 gene or a non-functional homologous recombination (HR) repair pathway[4,5]. Cancer cells with an impaired HR repair system rely on PARP-mediated DNA repair for survival[6]. Inhibition of the base excision repair (BER) pathway by PARP inhibitors (PARPis) induces double-strand breaks (DSB) from DNA single-strand (DSS) lesions, thus leading to “synthetic lethality”[7]. PARPis are also reported to be involved in “PARP trapping” on damaged DNA, defective recruitment of DNA repair proteins, and activation of error-prone non-homologous end-joining (NHEJ) in cancer cells[8]. Such features of PARPis advocate their clinical utility in patients with BRCA-proficient cancer. However, frequent transient responses to PARPis in BRCA competent or platinum-resistant EOC[9,10], low complete response (CR) to PARPi therapy (complete response (CR): 3%)[11-13], the emergence of resistance to therapy, and therapy-associated toxicities[14-17] (tabulated in Supplementary Table 5) limit the utility of PARPis in the clinic. To efficaciously improve PARPi-based treatment, novel targeted combination approaches that produce a more favorable prognosis are required.

BCL2–associated death promoter (BAD) protein is a BH3-only member of the BCL-2 family of proteins. BAD was first identified in the rat ovary being an ovarian BCL2-interacting pro-apoptotic protein and with a functional contribution to follicular atresia[18]. Unphosphorylated BAD protein heterodimerizes with and sequesters BCL2, BCL2-XL, and BCL-W with consequent promotion of intrinsic apoptosis through disruption of mitochondrial membrane potential[19]. BAD acts as a common and core downstream effector protein of both the RAS/MEK/ERK and PI3K/AKT/mTOR pathways[20]. Human (h) BAD is primarily phosphorylated at Serine (S)75 through p44/42 MAP kinase pathway activation[21], and at S99 through activation of AKT/p70S6K[22,23]. The RAS/MEK/ERK and PI3K/AKT/mTOR signaling cascades converging at BAD participate in maintaining cell survival following both cisplatin and paclitaxel treatment in ovarian cancer. Consequent phosphorylation of BAD protein promotes cell survival and resistance to therapy[24-26]. Therefore, it has been suggested that inhibition of either the RAS/MEK/ERK or PI3K/AKT/mTOR cascades may sensitize ovarian cancer cells to paclitaxel or cisplatin[27]. Interestingly, inhibition of PARP activity by Olaparib in EOC cells has also been reported to stimulate the RAS/MEK/ERK and PI3K/AKT/mTOR signaling pathways[28-31]. A number of studies have demonstrated the potential for combination therapy of PARPis with PI3K/AKT/mTOR-[32-34] or MEK-inhibitors[31] in BRCA-deficient and BRCA-proficient preclinical models[35]. Indeed, clinical responses of Olaparib and PI3K/AKT/mTOR inhibitor combinations in BRCA wild-type or platinum-resistant EOC were superior to monotherapies (summarized in Supplementary Table 7). In an ongoing phase Ib trial (NCT02208375), the mTOR inhibitor Vistusertib (AZD2014) (37) and AKT inhibitor Capivasertib (AZD5363) (38) combined with Olaparib demonstrated a 20% objective response rate (ORR) in ROC and a 24% ORR in a cohort of recurrent breast, endometrial, and OC patients respectively. Of those with OC treated with the combination of Ola and AZD5363, several patients achieved durable responses (4/6 stable disease (SD) and 1/7 partial response (PR)). Similarly, clinical trials of Olaparib in combination with different PI3K inhibitors (NCT02338622 and NCT01623349)[36-38] or MEK inhibitor Selumetinib (NCT03162627)[39] also demonstrated clinical benefit in patients with advanced OC. Hence, combinations of PI3K/AKT/MEK inhibitors with PARPis may be efficacious beyond BRCA-associated or platinum-sensitive OC[40]. Despite effective responses of the combinations of PARPis and PI3K/AKT/MEK-inhibitors, the ORR of <30% indicates that possible feedback or compensatory bypass mechanisms were activated. To date, no combination with PARP inhibition has received regulatory approval.

Elevated BAD phosphorylation has been reported to be positively correlated with acquired cisplatin/paclitaxel resistance and poorer prognosis of EOC patients[26,41-43]. Based on bioinformatics, physio-chemical, and cell-based analyses previously described[44], a small molecule termed NPB interacts directly with human BAD protein and inhibits the phosphorylation of Serine99. NPB does not affect the phosphorylation of other signaling molecules (including AKT) nor the predominantly p44/42 MAP kinase phosphorylated Ser75 residue on BAD, demonstrated using multiple in vitro biochemical assays[44]. Thus, NPB has been proposed to hinder the phosphorylation of BADS99 independent of the upstream AKT-kinase or other phosphorylation activities[44]. Herein, it was demonstrated that the level of pBADS99 predicted a worse survival outcome for patients with EOC and that NPB was equally efficacious as PARPis in models of EOC. The combination of pBADS99 inhibition with PARPis (Olaparib, Rucaparib, or Talazoparib) synergized to enhance apoptosis of EOC cells by enhancing DNA damage by limiting PARPi-induced DNA repair. The use of EOC cell line and patient-derived organoids and xenografts further demonstrated the efficacy of the NPB-PARPi combination. Hence, a novel and efficacious regimen for the treatment of recurrent EOC is provided.

Methods

EOC patient specimens and histopathological analysis

The human tissue samples used herein consisted of 80 EOC and 20 non-cancer ovarian (NCO) tissues from patients that underwent surgery at the University of Hong Kong-Shenzhen Hospital (HKU-SZH, Shenzhen, Guangdong, China) between 2016 and 2018. The use of patient specimens in this study was approved by the Institutional Ethical Committee of the HKU-SZH (certificate No. hkuszh2019105, approval No.伦 [2019] 096, Date: 2019.03.26). Patient consent forms were obtained from all patients in accordance with the declaration of Helsinki. Clinical information and follow-up data were obtained from the hospital medical records (Supplementary Data 1). The NCO tissue was collected from ovaries of patients with uterine fibroids (non-ovarian pathology) who elected to undergo concomitant salpingo-oophorectomy. All cancer patients were staged according to the International Federation of Gynecology and Obstetrics (FIGO) classification. Ki67 and TP53 status was determined after surgery.

Formalin-fixed, paraffin-embedded (FFPE) tissues were accessed from HKU-SZH. For the immunohistochemistry procedure, dewaxing was executed in xylene at room temperature for 15 min twice, then rehydration was carried out using ethanol with a concentration gradient of 100, 90, 80, 70, 50, and 0%. Tissues were placed in ethanol, 10 min twice for 100% ethanol and 5 min once for the other concentrations of ethanol. Antigen retrieval was achieved with 0.1 M freshly made sodium citrate solution at boiling temperature for 20 min. After cooling, sections were immersed in a solution of 3% hydrogen peroxide for 10 min in order to reduce endogenous peroxidase activity. The sections were washed with phosphate-buffered saline (PBS) for 5 min before blocking with 5% bovine serum albumin (BSA) diluted in PBS. The sections were then incubated with a primary antibody at 4 °C for 16 h. Immunohistochemistry (IHC) analysis was performed using primary antibodies and secondary anti-antibody tabulated in Supplementary Table 9. The sections were washed three times with PBS and incubated with biotinylated secondary antibody for 1 h. Following three washes with PBS, 3,3-diaminobenzidine (ab64238, Abcam, USA) was applied for visualization, and slides were counterstained with hematoxylin. The sections were dehydrated through graded alcohols, immersed in xylene, mounted with coverslips, and analyzed under a light microscope (CX31, Olympus, Japan) with ×4, ×10, or ×20 magnifications. Pictures were obtained with a digital CCD camera system (JVC, Tokyo, Japan). The slides were evaluated by three independent examiners using immunoreactive score (IRS) assessment[45]. IRS consists of a positive cell proportion score (0–4) and staining intensity score (0–3). For the positive cell proportion score, no cell stained was 0, the area of cells stained <10% is 1, the area 10–50% is 2, the area 51–80% is 3, and the area 81–100% was scored 4. For the staining intensity score, no staining was 0, weak staining was 1, moderate staining was 2 and, strong staining was 3. The correlation of pBADS99 IRS with clinicopathological features (Age, FIGO stage, Lymph node metastasis, Ki67, and TP53) of the EOC patient cohort was determined using Spearman’s rank correlation coefficient. Survival analyses were performed on 80 EOC patients using SPSS 25 (IBMSPSS Statistics, IBM Corp., Armonk, NY, USA).

Cell culture and reagents

Five EOC cell lines from the Singapore ovarian cancer library (SGOCL) that were used in this study are previously described[44,46,47]. Anglne and OVCAR3 were purchased from Procell Life Science & Technology Co. Ltd (Wuhan, China). All cell lines have been tested for the absence of mycoplasma. All experiments were performed with 2% FBS in the respective media. Four fresh patient-derived EOC (~1 g) samples were obtained with written informed consent and approval from the patients and the Ethical Committee of the University of Hong Kong-Shenzhen Hospital (HKU-SZH, research No. hkuszh2019105, approval No.伦 [2019]096, Date: 2019.03.26). All pathologies were diagnosed by a pathologist. The detailed information of the patients is listed in Supplementary Fig. 6A. Tissue was minced with scissors (except for AFC cells), and the tissue fragments or AFC cells were incubated in serum-free Advanced Dulbecco’s modified Eagle’s medium (DMEM)/F12 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 1.5 mg/ml collagenase IV (Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Inc.) at 37 °C for 2–4 h with continuous slight shaking. The cells were filtered through a cell strainer after digestion followed by centrifugation for 5 min at a speed of 500 r/min. The supernatant was removed and the pellets were washed with PBS and centrifuged as described above. Single primary cells were resuspended in Advanced Dulbecco’s modified Eagle’s medium (DMEM)/F12 for use in subsequent 2D or organoid culture. For organoid culture, Matrigel (354262, Corning, US) was added to the plate and polymerized for 15 min at 37 °C, and then primary cells were mixed with growth medium supplemented with 2% FBS, 50% Matrigel (354262, Corning, USA), 1x N2 Supplement (Gibco; Thermo Fisher Scientific, Inc.), 1x B27 (Gibco; Thermo Fisher Scientific, Inc.), 50 ng/ml recombinant human EGF (PEPROTECH; USA), and 20 ng/ml basic fibroblast growth factor (Sigma-Aldrich; Merck). 1.25mM N-Acetylcysteine (Sigma-Aldrich; Merck), 50 ng/ml Rock inhibitor (Y-27632) (Sigma-Aldrich; Merck), 10 nM 17-β Estradiol (Sigma-Aldrich; Merck) and 5 nM A83-01 (Sigma-Aldrich; Merck) were added to the wells (200 μl/well in 48-well plates and 500 μl/well in 24-well plates). PDOs usually formed within 3 days of primary culture and were expanded for 9 days[48]. Treatment was begun on the third day after organoid formation and the medium with treatment was changed every 2 days. Organoids with diameters >100 µm were counted under an inverted microscope to determine organoid expansion. At termination, viability and apoptosis were determined using the ApoTox-Glo Triplex Assay Kit (G6320, Promega, China)[49-52]. The clinical characteristics, molecular profiling, and culture conditions for each cell line are tabulated in Supplementary Tables 3 and 4. The molecular profiling of patient-derived cells was performed at Kingmed Center for Clinical Laboratory, Guangzhou, China. Three PARP inhibitors, Olaparib (A ZD2281), Rucaparib (AG-014699), and Talazoparib (BMN 673) were purchased from Selleckchem (Houston, TX, USA).

Western blot analysis

Western blot analysis was performed as previously described[44]. Briefly, cells were lysed in RIPA buffer, and proteins in the cell lysate were resolved using SDS polyacrylamide gel electrophoresis and visualized with Clarity™ and Clarity Max™ Western ECL Blotting Substrates (BIO-RAD, USA). The primary and secondary anti-rabbit and anti-mouse horseradish peroxidase (HRP)-conjugated antibodies used were tabulated in Supplementary Table 9.

Oncogenic and immunofluorescence (IF) analyses

Cell viability, apoptosis, and cytotoxicity assays were performed using ApoTox-Glo™ Triplex Assay Kit (G6320, Promega, China) as per the manufacturer’s instructions[44]. Fluorescence and luminescence were determined using a Tecan microplate reader. Phosphatidylserine exposure and cell death were assessed by CytoFLEX (Beckman Coulter, Inc. USA) using Annexin-V-FLUOS and PI-stained (Neobioscience, Shenzhen, China) cells as per the manufacturer’s instructions. Live/Dead cells visualization was performed as per the manufacturer’s instructions[44] using LIVE/DEAD™ Cell Imaging Kit (Thermo Fisher Scientific, USA). Combination index (CI) analysis was performed using the Chou-Talalay CI method[53].

IF analyses were performed as previously described using confocal microscopy (C2+, Nikon, Japan). Briefly, cells plated on chamber slides were fixed with 4% paraformaldehyde (PFA) for 30 min. After three washings with PBS, fixed cells were permeabilized with 0.2% Triton X-100/PBS for 5 min and incubated with 5% BSA /PBS for 15 min. Cells were incubated with primary antibodies overnight. On the second day, after three PBS washings, cells were subsequently incubated for 1 h with secondary antibodies. At least 100 cells were analyzed in three separate fields for each sample with NIS-Elements AR software. Information of primary/secondary antibodies were tabulated in Supplementary Table 9. Nuclei were stained with a mounting medium with DAPI (ab104193, Abcam). All functional assays were performed in medium with 2% FBS.

In vivo analyses

Xenograft studies were performed according to the Animal Research: Reporting In Vivo Experiments (ARRIVE) 2.0 guidelines. The care and use of laboratory animals were approved by the Laboratory Animal Ethics Committee (Certificate number: YW) at Peking University Shenzhen as previously described[54], and ethical approval was obtained from Tsinghua Shenzhen International Graduate School (Number:9, Year 2020). Mice were housed in a controlled atmosphere (25 ± 1 °C at 50% relative humidity) under a 12-h light/12-h dark cycle. Animals had free access to food and water at all times. Food cups were replenished with a fresh diet daily. Briefly, A2780, A2780cis, or AFC cells (5 ×  106 cells) were injected subcutaneously (s.c.) into the right flanks of five-week-old female BALB/c nude mice (Charles River, Beijing). Mice (n = 6) bearing similar xenograft sizes were randomly assigned to different treatment arms: control, NPB, Olaparib, or NPB-Olaparib as summarized in Supplementary Fig. 8A. The statistical differences between the treatment groups were compared using a one-way ANOVA followed by a Tukey’s multiple comparison test. Histological analyses were performed as previously described[44]. Xenograft volume = width × length × length/2. The response was determined by comparing tumor volume change to its baseline at time t: % tumor volume change = ΔVolt = 100% × ((V – VVinitial)/VVinitial). The Best Response was the minimum value of ΔVolt for t ≥ 10 d. For each time t, the average of ΔVolt from 0 to t was also calculated. Best average response was defined as the minimum value of this average for t ≥ 10 d. The criteria for response (mRECIST)[55,56] were defined as followed: complete response (mCR): best response < −95% and best avg response < −40%; partial response (mPR): best response < −50% and best avg response < −20%; stable disease (mSD), best response < 35% and best avg response < 30%; progressive disease (mPD), not otherwise categorized.

Statistics and reproducibility

SPSS 25 (IBMSPSS Statistics, IBM Corp., Armonk, NY, USA) and GraphPad Prism 7.0 (GraphPad Software, San Diego, CA, USA) were used to generate graphs and perform statistical analysis as previously described[44]. For in vitro assays, the statistical differences between the treatment groups were compared using a one-way ANOVA followed by a Tukey’s multiple comparison test. p-values < 0.05(*), p < 0.01(**) and p < 0.001(***) were considered statistically significant. Quantitative data are expressed as mean ± SD, unless otherwise stated (Supplementary Data 2). See individual “Methods” sections for specific statistical methods. Histopathological assays were replicated at least three times, with each orthogonal method confirming the same result. Western blot assays were replicated at least three times in each tested cell line, showing similar results. Immunoprecipitations followed by silver stains and immunoblots were replicated at least twice. IF assays were performed in triplicate. Different oncogenic assays were performed independently at least three times and showed similar results. Each group in in vivo experiments contained at least six mice.