Abstract

   BRCA1 is one of the most frequently-mutated tumor suppressor genes in
   ovarian and breast cancers. Loss of BRCA1 triggers homologous
   recombination (HR) repair deficiency, consequently leading to genomic
   instability and PARP inhibitors (PARPi)-associated synthetic lethality.
   Although, the roles of BRCA1 in DNA repair and replication have been
   extensively investigated, its tumor suppressive functions beyond genome
   safeguard remain poorly understood. Here, we report that BRCA1 promotes
   ferroptosis susceptibility through catalyzing K6-linked
   polyubiquitination of GPX4 and subsequently accelerating GPX4
   degradation. Depletion of BRCA1 induces ferroptosis resistance in
   ovarian cancer cells due to elevated GPX4 protein, and silence of GPX4
   significantly suppresses the growth of BRCA1-deficient ovarian cancer
   xenografts. Importantly, we found that PARPi triggers ferroptosis in
   ovarian cancer cells, inhibition of GPX4 markedly increase
   PARPi-induced ferroptosis in BRCA1-deficient ovarian cancer cells.
   Combined treatment of GPX4 inhibitor and PARPi produces synergistic
   anti-tumor efficacy in BRCA1-deficient ovarian cancer cells, patient
   derived organoid (PDO) and xenografts. Thus, our study uncovers a novel
   mechanism via which BRCA1 exerts tumor suppressive function through
   regulating ferroptosis, and demonstrates the potential of GPX4 as a
   therapeutic target for BRCA1-mutant cancers.

   Keywords: BRCA1, Ferroptosis, PARP inhibitors, GPX4 ubiquitination,
   Ovarian cancer

Graphical abstract

   [39]Image 1
   [40]Open in a new tab

Highlights

     * •
       BRCA1 deficiency leads to increase of GPX4 and ferroptosis
       resistance.
     * •
       BRCA1 catalyzes GPX4 polyubiquitination and promotes its
       degradation.
     * •
       Ferroptosis represents a novel mechanism for BRCA1-mediated tumor
       suppression.
     * •
       GPX4i and PARPi synergistically suppresses BRCA1-deficient cancers.

1. Introduction

   As one of the most common mutated genes, loss of function of BRCA1
   plays critical roles in tumorigenesis and cancer progression in various
   cancers, especially in ovarian and breast cancers [[41]1].
   Approximately 20–30 % of ovarian or breast cancer patients carry BRCA1
   mutations, which are associated with poor overall survival [[42]2].
   BRCA1 gene is located on chromosome 17, consists of 22 exons and
   encodes a protein of 1863 amino acids [[43]3]. BRCA1 protein is the
   main controller in DNA double-strand break (DSB) repair pathway choice
   between homologous recombination (HR) and non-homologous end-joining
   (NHEJ) repair, it accumulates at DNA DSB sites and favors HR repair
   during S/G2 phase [[44][4], [45][5], [46][6]]. Meanwhile, BRCA1 is also
   involved in DNA replication fork protection during DNA replication
   stress [[47]7]. Thus, loss of BRCA1 leads to genomic instability
   characterized by increased DNA breaks and chromosomal aberrations. As
   an important cancer susceptibility gene, mutations in BRCA1 confers
   lifetime risks of breast and ovarian cancer up to 80 % and 50 %
   respectively [[48]2]. BRCA1-deficient cancers display an autonomous
   inflammatory state and are intrinsically resistant to immune checkpoint
   blockade (ICB) and MEK inhibitor [[49][8], [50][9], [51][10]]. Similar
   with BRCA1, the other breast cancer susceptibility gene, BRCA2 is also
   involved in regulation of DSBs repair and frequently mutated in ovarian
   cancer cells. However, only BRCA2 mutations, but not BRCA1 mutations,
   are associated with improved chemotherapy response compared with
   wild-type (WT) BRCA1/2 in patients with ovarian cancer [[52]11,[53]12].
   In addition, Brca1^185stop mutation predisposed mice to mammary tumors,
   albeit HR repair function is preserved in Brca1^185stop cells through
   alternative translation of RING domain-less BRCA1 protein
   [[54]13,[55]14]. These distinct features in BRCA1-deficient cancer
   cells suggest BRCA1 possess other functions beyond DNA repair
   regulation.

   PARP is a group of enzymes catalyzing the poly ADP-ribosylation
   (PARylation) on substrate proteins. PARP1 is the most abundant and
   well-investigated PARP family protein, it can rapidly recognize the DNA
   damage sites and directly participates in DNA single-strand break
   (SSBs) repair [[56]15]. Inhibition of PARP1 results in accumulation of
   unrepaired SSBs, which are subsequently converted into DNA DSBs. The
   DNA DSBs formed after treatment of PARP inhibitors (PARPi) highly rely
   on HR repair, as a result, PARP inhibitors selectively eliminate HR
   repair deficient cancer cells, such as the cancer cells with BRCA1/2
   mutations [[57]16]. Although, PARPi elicits synthetic lethality and
   have been approved for treatment of BRCA1/2-mutant cancers [[58]17],
   differential activity of PARP inhibitors has been observed in
   BRCA1-mutant versus BRCA2-mutant cancers and BRCA1-mutation carriers
   responded less favorably to PARP inhibitor than patients with
   BRCA2-mutations [[59]18]. Thus, the emergence of intrinsic PARP
   inhibitor resistance in BRCA1-mutant cancer cells prompts the novel
   strategy to optimize PARP inhibitor therapy in BRCA1-mutant cancers.

   Ferroptosis is a form of iron-dependent cell death, which is resulted
   from lethal lipid peroxidation and characterized by cell shrinkage and
   increased mitochondrial membrane density [[60]19]. Ferroptosis is
   controlled by intracellular oxidation and antioxidant systems and the
   selenium-containing anti-oxidant enzyme glutathione peroxidase 4 (GPX4)
   is considered as the central repressor of ferroptosis [[61]20]. GPX4
   catalyzes the reduction of hydroperoxides of polyunsaturated fatty
   acids in membranes and inhibits lipid peroxidation [[62]21].
   Ferroptosis plays critical roles in various of physiological and
   pathological processes including neurodegenerative diseases, acute
   renal failure, heart injury, cancer, and so on [[63]19]. As a
   regulated-cell death, ferroptosis susceptibility is highly associated
   with cancer progression and therapeutic resistance [[64]22,[65]23]. In
   our present study, we showed that BRCA1 ubiquitinates and degrades GPX4
   to dictate ferroptosis sensitivity in ovarian cancer cells, which is
   associated with its tumor suppressive function in vivo. We further
   found PARP inhibitors, the premier treatment for BRCA1/2-mutant
   cancers, triggered ferroptosis in ovarian cancer cells, and inhibition
   of GPX4 greatly promoted PARP inhibitor-induced ferroptosis in
   BRCA1-deficient ovarian cancer cells. Therefore, our study uncovers
   that GPX4-mediated ferroptosis susceptibility represents a novel tumor
   suppressive function of BRCA1, as well as a therapeutic target to
   optimize PARP inhibitor therapy in BRCA1-deficient cancers.

2. Materials and methods

2.1. Human samples

   Ovarian cancer tissues were collected at Nanjing Maternity and Child
   Health Care Hospital and informed consent was obtained from all
   patients. The use of all patients’ specimens was approved by the
   Institutional review board (IRB) of Nanjing Maternity and Child Health
   Care Hospital.

2.2. Reagents, antibodies and plasmids

   Anti-BRCA1 (sc-6954), anti-HA (sc-57592), anti-β-Actin (sc-47778) and
   anti-Flag (sc-807) antibodies were purchased from Santa Cruz
   Biotechnology (Santa Cruz, CA, USA). Anti-GPX4 (67763-1-Ig),
   anti-SLC7A11 (26864-1-AP) anti-myc (60003-2-Ig) antibody were purchased
   from Proteintech Group Inc. (Chicago, IL, USA). Anti–HO–1 (A1346) and
   anti-SLC40a1 (A14884）were purchased from ABclonal technology (Wuhan,
   China). anti-TFR1 (AF8136) were purchased from Beyotime
   biotechnology（Beijing, China). Anti-Ubiquitin (#3936) were obtained
   from Cell signaling technology (Danvers, MA). Anti-4-HNE(MAB3249) was
   purchased from R&D Systems (Minneapolis, MN, USA). Human BRCA1 siRNA
   (target sequence: #1, 5′-CCA CAC GAT TTG ACG GAA A-3’; #2, 5′-CTA CTC
   ATG TTG TTA TGA A-3′) were obtained from Ribobio (Guangzhou, China), in
   vivo cholesterol-modified GPX4 siRNA (target sequence: 5′-GGA GTA ACG
   AAG AGA TCA A-3′) were purchased from Ribobio (Guangzhou, China).
   Recombinant GPX4 protein (#26906) was purchased from Cayman Chemical
   Company (Ann Arbor, USA). Ubiquitin Activating Enzyme E1/UBA1 protein
   (11990-H20B-50) was purchased from Sino Biological Inc. (Beijing,
   China). Recombinant human ubiquitin-conjugating enzyme E2
   UBCH5C([66]P00347) was purchased from Solarbio Life Sciences (Beijing,
   China). Recombinant ubiquitin (20487) was purchased from Cayman
   Chemical (MI, USA). Flag-GPX4, HA-BRCA1, Myc-BARD1, Myc-BRCA1 plasmids
   were obtained from Jinbeijin biotechnology (Nanjing, China). GPX4
   shRNAs were obtained from MiaoLingBio (Wuhan, China). HA-WT, K6-, K11-,
   K27-, K29-, K33-, K48-, K63- Ubiquitin plasmids were kindly gifted by
   Dr. Chen, Yan (Shandong Normal University). Myc-BRCA1 truncates
   plasmids were generated by Nanjing Genebay Biotech Inc. (Nanjing,
   China) using Myc tagged full length BRCA1 as a templet.

2.3. Cell culture and transfection

   UWB1.289, UWB1.289-BRCA1, A2780, SKOV3, OVCAR3 and HCC1937 cells were
   obtained from American Type Culture Collection (ATCC). UWB1.289 and
   UWB1.289-BRCA1 cells were cultured in 50 % RPMI (Roswell Park Memorial
   Institute)-1640 (Gibco, USA) plus 50 % MEGM (Mammary Epithelial Growth
   Medium) (#CC-3150, Lonza, NJ) supplemented with 3 % fetal bovine serum
   (FBS, ExCell Bio, Shanghai, China) and penicillin-streptomycin (100
   U/mL, Gibco, USA). A2780, SKOV3, HCC1937 and OVCAR3 were cultured in
   RPMI-1640 medium supplemented with 10 % FBS and
   penicillin-streptomycin. Cells were regularly tested for mycoplasma
   contamination using MycoBlue Mycoplasma Detector (D101-01, Vazyme,
   Nanjing, China). Cells were transfected with plasmids using ExFect
   transfection reagent (Vazyme, NJ, China) according to manufacturer's
   instruction.

2.4. Organoid culture

   Patient derived ovarian cancer organoid (PDO) (BRCA1^c.3756_3759del,
   PDO-ID: KOOA-009) model was established by Beijing K2 Oncology Inc.
   (Beijing, China). This ovarian cancer tissue was collected at People's
   Hospital of Yixing City and was approved by the Medical Ethical
   Committee of the People's Hospital of Yixing City. Informed consent was
   obtained from the patient. The organoids were cultured in ovarian
   cancer organoids culture medium (#K2O-M-OA, K2 Oncology, Beijing,
   China). After seeding in 96 wells plates, organoids were treated with
   serial concentrations of Olaparib, ML210 or their combination for 5
   days. Then, organoid viability was determined by the Cell Titer-Glo
   assay (#G9683, Promega, Madison, WI) and synergistic effects were
   calculated by ZIP synergy score.

2.5. Ovarian cancer xenograft

   All the animal experiments were approved by the Institutional Animal
   Care and Use Committee (IACUC) of China Pharmaceutical University. The
   6 weeks-old female BALB/c athymic nude mice were purchased from the
   GemPharmatech (Nanjing, China) and housed in a specific pathogen-free
   (SPF) facility. 1 × 10^7 of A2780 parental or BRCA1-KO cells in 100 μl
   PBS were inoculated subcutaneously into the right flank of mice. Tumor
   width (W) and length (L) were measured and the tumor volume (V) was
   calculated by formula V = L × W^2/2. When tumor volume reached around
   50–100 mm^3, mice were randomly grouped and treated respectively. in
   vivo cholesterol-modified GPX4 siRNA were administrated intratumorally
   three time per week. For evaluating therapeutic efficacy of
   ML210/Olaparib combination, mice with BRCA1-KO xenografts were treated
   with PBS (control), ML210 (intraperitoneal, 30 mg/kg), Olaparib (oral,
   100 mg/kg) or their combination. Tumor volumes were measured every
   three days and tumor growth were monitored.

2.6. Immunoprecipitation (IP)

   Cells were collected in EBC buffer (50 mM Tris-HCl pH = 7.6–8.0, 120 mm
   NaCl, 0.5 % NP-40, 1 mM EDTA, 50 mm NaF, 1 mmNa[3]VO[4] and 1 mM
   β-mercaptoethanol) supplemented with cocktail protease inhibitor (EMD
   Millipore, Burlington, MA) and lysed by sonication. After
   centrifugation, cell lysates were incubated with indicated antibodies
   and protein A/G agarose beads (sc-2003, Santa Cruz biotechnology, USA)
   at 4 °C overnight with rotation. For the flag immunoprecipitation, cell
   lysates were incubated with anti-flag M2 affinity gel (A2220,
   Sigma-Aldrich, USA) at 4 °C overnight with rotation. After washing 5
   times with EBC buffer, beads were boiled in 40 μl of 2 × SDS-PAGE
   sample buffer for 6min and the proteins in the immunoprecipitation
   complex were analyzed by Western blot.

2.7. Immunohistochemical (IHC) staining

   The ovarian cancer tissue microarray (ZL-OVA961) was purchased from
   Shanghai Wellbio Biotechnology (Shanghai, China). Firstly, the sections
   were deparaffinized in 100 % xylene and rehydrated with gradient
   ethanol (100 %, 90 %, 70 %, 30 % and 0 %). After inactivation of
   endogenous peroxidase by 3 % hydrogen peroxide and antigen retrieval in
   citrate buffer by heat, the IHC staining was performed using VECTASTAIN
   universal Quick kit (PK-8800, Vector Laboratories, USA) according to
   the manufacturer's instruction. The following primary antibody
   dilutions were used: anti-GPX4 (67763-1-IG,1:2000), anti-BRCA1
   (ab16780,1:1000), anti-4-HNE (MAB3249, 1:500). The semiquantitative
   determination of GPX4, 4-HNE and BRCA1 were carried out using
   immunoscore based on the percentage and intensity of positive staining
   as previous described.

2.8. Generation of knockout (KO) cells

   BRCA1 was depleted in A2780 and OVCAR3 cells using CRISPR/Cas9
   technique as previously described [[67]24,[68]25]. Briefly, cells were
   transduced with lentiCRISPRV2-puro carrying BRCA1 sgRNA and then
   selected with puromycin (2 μg/ml) and seeded for growing single clones.
   Then, single clones were picked and BRCA1 expression was examined by
   Western blot. BRCA1 KO A2780 cells were transduced with
   lentiCRISPRV2-neo-53BP1 sgRNA, selected with G418 (800 μg/ml), and
   53BP1 KO was examined by Western blot. The sgRNA sequences targeting
   BRCA1 or 53BP1 were in [69]Supplemental Table 2.

2.9. Lipid peroxidation assay

   Lipid peroxidation was measured by BODIPY 581/591 C11 (Thermo Fisher
   Scientific, Waltham, MA, USA) staining. Briefly, cells were seeded in
   6-well plates at a density of 1 × 106 cells per well and treated with
   or without indicated agents for 24 h. After treatment, cells were
   incubated with fresh complete medium containing 10 μM BODIPY 581/591
   C11 dye (D3861, Invitrogen) at 37 °C for 1 h. After washing with PBS,
   lipid peroxidation was then analyzed by flow cytometry (Accuri C6
   cytometer, BD Biosciences) using the 488-nm laser for excitation. The
   relative lipid peroxidation level was indicated by the percentage of
   cells gated by the black dotted line based on the fluorescence
   intensity in FL1 channel, and the data were analyzed using the FlowJo
   10 software.

2.10. In situ proximity ligation assay (PLA)

   Proximity Ligation Assay (PLA) was performed using the Duolink In Situ
   Red Starter Kit Mouse/Rabbit (DUO92101, MilliporeSigma) according to
   the manufacturer's protocol (MilliporeSigma, Burlington, MA). Briefly,
   A2780 cells were fixed with 4 % paraformaldehyde for 15 min and
   permeabilized with 0.2 % TritonX-100. After blocking with Duolinke
   blocking solution for 1 h at room temperature, cells were then
   incubated with primary antibodies overnight at 4 °C. After incubation
   with PLA probe solution at 37 °C for 1 h, cells were then incubated
   with amplification solution for 100 min at 37 °C. After mounting with
   Prolong Gold antifade reagent containing DAPI (Invitrogen, CA), PLA
   signals were captured by confocal microscope (Lecia SP8, Leica
   Microsystems Inc.).

2.11. Ubiquitination assay

   For the in vivo ubiquitination assay, cells were co-transfected with
   indicated combination of constructs and treated with MG132 (5 μM) for
   6 h before harvesting. Cells were then collected in EBC buffer (50 mM
   Tris-HCl pH = 7.6–8.0, 120 mM NaCl, 0.5 % NP-40, 1 mM EDTA, 2 mM NaF,
   1 μM Na[3]VO[4] and 1 mM β-mercaptoethanol) supplemented with protease
   inhibitor and lysed by sonication. After centrifugation, the
   supernatant was incubated with anti-flag affinity gel at 4 °C overnight
   with rotation. After extensive washing with EBC buffer, the beads were
   suspended with 2 × SDS sample buffer and boiled at 100 °C for 5 min.
   The bound proteins were separated on SDS-PAGE and the GPX4
   ubiquitination was analyzed by Western blot. For the in vitro ubiquitin
   ligation assay, the heterodimer of WT or I26A flag-BRCA1^1−303 and
   BARD1 were purified from transfected 293T cells by anti-flag affinity
   chromatography and 3 × FLAG peptide elution. His-GPX4 (1 μg) was
   incubated in the reaction mixture (50 mM Tris-HCl pH7.5, 120 mM NaCl,
   0.5 mM DTT, 5 mM ATP, 5 mM MgCl[2], 2 mM NaF and 5 μM ZnCl[2])
   containing 0.1 μg E1, 0.5 μg UBCH5C, 2 μg Ub, 1 μg each of
   flag-BRCA1^1−303 and BARD1. After incubation at 37 °C for 3 h, SDS-PAGE
   sample buffer was added into the reaction and boiled at 100 °C for
   5 min. The ubiquitinated GPX4 was detected by Western blot with anti-Ub
   antibody.

2.12. Mass spectrometry analysis of GPX4 ubiquitination sites

   The sites of poly-ubiquitination on GPX4 were identified by mass
   spectrometry (LC-MS/MS) in PTM-Biolabs Inc. (Hangzhou, China). Briefly,
   HEK293T cells were co-transfected with flag-GPX4 and HA-ubiquitin, and
   poly-ubiquitinated flag-GPX4 was purified from cells after MG132
   treatment (20 μM, 6 h) by immunoaffinity purification using
   anti-DYKDDDDK G1 affinity resin (GenScript, Nanjing, China) and 3X flag
   peptide elution (Beyotime, Shanghai, China). The poly-ubiquitinated
   GPX4 were then resolved in SDS-PAGE and in-gel tryptic digestion was
   performed. After extraction and dry, the tryptic peptides were then
   resuspended in 2 % acetonitrile/0.1 % formic acid and subjected to
   tandem mass spectrometry (MS/MS) in Q-Exactive plus (Thermo Fisher
   Scientific, Waltham, MA, USA). The resulting mass spectra data were
   processed using Proteome Discoverer 2.4. software and searched against
   human protein database.

2.13. Apoptosis analysis

   Cells were treated with indicated concentrations of etoposide for 48 h
   and cell apoptosis was measured using FITC-Annexin V apoptosis
   detection kit (556547, BD Biosciences) according to manufacturer's
   instruction. Briefly, after washing with PBS, cells were then incubated
   with 200 μl 1xbinding buffer containing annexin V-FITC and propidium
   iodide (PI) at room temperature in dark for 25 min. Then, 200 μl
   1Xbinding buffer was added and cell apoptosis was analyzed by flow
   cytometry.

2.14. Cell death and viability assay

   Cell death was measured by PI (Propidium Iodide) staining as previously
   descried. After treatment, cells were collected and washed once with
   PBS, followed by incubation with 50 μg/ml PI in PBS for 15 min. Then,
   dead cell with PI positive were analyzed by flow cytometry. Cell
   viability was measured using Cell Counting Kit-8 (CCK-8) (Doijindo,
   Japan). Cells were seeded into 96-well plate at 5 × 10^3 cells per
   well, and were then treated as indicated for 72 h. After treatment,
   cells were incubated with CCK-8 reagent for 1 h at 37 °C, followed by
   measurement of the OD value at 450 nm with a microplate reader.

2.15. Colony formation assay

   10^3 of cells were seeded into 6-well plate and grown in presence of
   indicated concentrations of agents for 7–14 days until colonies were
   visible. Then, colonies were washed with PBS, and stained with 0.5 %
   crystal violet in 20 % methanol. Colonies were then counted and the
   relative colony formation was normalized to un-treated group.

2.16. Quantitative real time PCR (qRT-PCR)

   Total mRNA was extracted using TRIzol reagent (Solarbio，Beijing, China)
   according to the manufacturer's instruction, followed by synthesis of
   complementary DNA (cDNA) using oligo (dT) primers by a reverse
   transcriptase kit (Vazyme, Nanjing, China). Then, qPCR was performed
   using SYBR Green Master Mix wiper (Vazyme, Nanjing, China) on Bio-Rad
   Real-Time PCR instruments. The relative mRNA levels were calculated by
   the 2^–ΔΔCT method and normalized to β-Actin. The specific primers
   were: BRCA1 p1: 5′-GAA ACC GTG CCA AAA GAC TTC-3’; BRCA1 p2: 5′-CCA AGG
   TTA GAG AGT TGG ACA C-3’; GPX4 p1: 5′-GAG GCA AGA CCG AAG TAA ACT
   AC-3’; GPX4 p2：5′-CCG AAC TGG TTA CAC GGG AA-3’; GAPDH p1: 5′-GGA GCG
   AGA TCC CTC CAA AAT-3’; GAPDH p2: 5′-GGC TGT TGT CAT ACT TCT CAT GG-3’.

2.17. Subcellular fractionation

   Nuclear and cytosolic fractionation assay was performed using
   cytoplasmic extraction Kit (P0028, Beyotime, Beijing, China) according
   to manufacturer's instructions. Briefly, 5 × 10^6 of cells were
   suspended in cytoplasmic extraction reagent. After incubation in ice
   for 15 min, followed by centrifugation at 14000 rpm × 10 min, the
   supernatant cytoplasmic protein was collected. And the pellet was then
   suspended in nuclear extraction reagent and incubated in ice for 30 min
   to extract the nuclear proteins after centrifugation.

2.18. Label free quantitative proteomics analysis

   Quantitative proteomics was performed by Shanghai Applied Protein
   Technology Co., Ltd (Shanghai, China). Briefly, UWB1.289 and
   UWB1.289-BRCA1 cells were collected and lysed in SDT (4%SDS，100 mM
   Tris-HCl pH7.6) buffer. 20 μg of protein for each sample was resolved
   in SDS-PAGE and followed by trypsin digestion according to filter-aided
   sample preparation (FASP) procedure. The digested peptides were then
   desalted and reconstituted in 40 μl of 0.1 % (v/v) formic acid.
   LC-MS/MS analysis was performed on a Q Exactive mass spectrometer
   (Thermo Scientific) that was coupled to Easy nLC (Proxeon Biosystems,
   now Thermo Fisher Scientific). The MS raw data for each sample were
   combined and searched using the MaxQuant 1.6.14 software for
   identification and quantitation analysis.

2.19. Statistical analysis

   Data were shown from one representative experiment of at least three
   independent replicates and presented as mean ± standard deviation (SD).
   A two-sided t-test was used to analyze the statistical significance of
   differences between groups. Multiple comparisons for more than two
   groups were analyzed by two-way ANOVA with Tukey's test. p < 0.05 was
   considered statistically significant.

3. Results

3.1. BRCA1-deficiency confers ferroptosis resistance in ovarian cancer cells

   As a commonly mutated tumor suppressor, BRCA1 maintains genomic
   integrity to restrain cancer development [[70]26]. The specific
   mechanism of tumor suppression beyond DNA repair remains unclear. To
   explore the anti-tumor functions of BRCA1, BRCA1-null cell line
   UWB1.289 and UWB1.289 transduced with BRCA1 (UWB-BRCA1) were subjected
   to quantitative proteomics analysis ([71]Fig. 1A). We found 214
   proteins were significantly increased and 187 proteins were decreased
   in UWB.1289 cells compared with UWB-BRCA1 ([72]Supplemental Table 1).
   Notably, among these differentially expressed proteins, GPX4 was
   remarkedly increased in BRCA1-deficient UWB1.289 cells ([73]Fig. 1B),
   and gene ontology (GO) enrichment revealed that cell response to lipid
   pathway was significantly affected ([74]Fig. 1C). Thus, we speculated
   that BRCA1 might affect cell vulnerability to lipid peroxidation. To
   test our hypothesis, we treated cells with ferroptosis inducer RSL3 and
   Erastin, we interestingly found that BRCA1 expression rendered
   UWB1.289 cells susceptible to RSL3 or Erastin induced-cell death
   ([75]Supplemental Fig. 1A). Meanwhile, iron chelating agent
   deferoxamine (DFO) reversed RSL3-induced cell death in UWB-BRCA1 cells
   ([76]Fig. 1D), suggesting BRCA1 regulates ferroptosis sensitivity. In
   contrast, silence of BRCA1 protected BRCA1-proficient A2780 cells from
   RSL3-induced cell death ([77]Fig. 1E–G). Meanwhile, we did not detect
   significant change of etoposide induced-apoptotic cell death in
   BRCA1-expressing UWB1.289 or BRCA1-silenced A2780 cells ([78]Figs.
   S1B–D). And, ferroptosis inhibitor ferrostatin-1 (fer-1) completely
   reversed RSL3-induced cell death in both control or BRCA1-knockdown
   cells ([79]Fig. S1E).

Fig. 1.

   [80]Fig. 1
   [81]Open in a new tab

   BRCA1 promotes ferroptosis susceptibility in ovarian cancers. (A)
   Western blot analysis of BRCA1 protein expression in UWB1.289 and
   UWB1.289+BRCA1(UWB-BRCA1) cells. (B) Volcano plot of quantitative
   proteomics displays the differentially expressed proteins between
   UWB1.289 and UWB-BRCA1 cells. (C) Gene Ontology (GO) pathway enrichment
   analysis of significantly changed proteins. (D) Indicated cells were
   treated with increasing concentrations of RSL3 for 24 h in presence or
   absence of 100 μM DFO (desferoxamine), dead cells were measured by PI
   staining. Data were presented as mean ± standard deviation (SD), n = 3.
   (E) Western blot analysis of BRCA1 expression in A2780 cells
   transfected with control (Ctrl) or BRCA1 siRNA. (F–G) Control (Ctrl) or
   BRCA1 siRNA transfected A2780 cells were treated with indicated dosages
   of RSL3 for 24 h, followed by analysis of cell viability (F) and PI
   staining (G). (H–J) UWB1.289/UWB-BRCA1 cells (H), Ctrl/BRCA1 siRNA
   transfected A2780 cells (I), BRCA1 small-guide RNA (sgRNA) mediated-
   A2780 knockout (KO) cells (J) were treated with or without 0.5 μM RSL3
   for 24 h, lipid peroxidation was detected by BODIY 581/591 C11
   labeling. The representative flow cytometry profile (left) and
   quantification of lipid peroxidation (right) was shown. The depletion
   of BRCA1 in A2780 BRCA1 KO cells were confirmed by Western blot (upper
   right, J). Data were presented as mean ± SD from three replicates. (K)
   Immunohistochemistry (IHC) analysis of 4-HNE(4-Hydroxynonenal) in
   wild-type (WT) or BRCA1-mutant ovarian cancer tissues. The presentative
   staining (left) or quantification of staining score (right) was shown.
   Data were presented as mean ± standard deviation (SD), n = 9.
   **p < 0.01 and ***p < 0.001 by 2-tailed t-test.

   Lipid peroxidation is a hallmark of ferroptosis [[82]19]. Consistently,
   we clearly detected lipid peroxidation in UWB-BRCA1 cells, but not in
   UWB1.289 cells after RSL3 treatment ([83]Fig. 1H). In contrast, silence
   of BRCA1 attenuated RSL3-induced lipid peroxidation in A2780 cells
   ([84]Fig. 1I). Similar observation of decreased lipid peroxidation was
   also detected in A2780-BRCA1 knockout (KO) cells after RSL3 treatment
   ([85]Fig. 1J). Treatment of fer-1 completely abolished RSL3-induced
   lipid peroxidation in both BRCA1-profienct and deficient cells
   ([86]Figs. S1F–G).To further confirm the ferroptosis regulation by
   BRCA1 in vivo, we analyzed the level of 4-hydroxynonenal (4-HNE), a
   reactive aldehyde product and a biomarker of lipid peroxidation
   [[87]27], in BRCA1 wild-type (WT) or mutant ovarian cancer tissues. As
   expected, significant decrease of 4-HNE were detected in BRCA1-mutant
   cancer tissues ([88]Fig. 1K). These results reveal that
   BRCA1-deficiency in ovarian cancer cells promotes ferroptosis
   resistance.

3.2. BRCA1 promotes ferroptosis through degrading GPX4

   To validate the decrease of GPX4 in BRCA1-expressing UWB1.289 cells, we
   then examined the levels of GPX4 protein along with several common
   ferroptosis regulators in UWB1.289 and UWB-BRCA1 cells, as well as
   A2780 cells with or without BRCA1 knockdown by Western blot. As
   expected, we observed that BRCA1-expression reduced GPX4 expression in
   UWB1.289 cells, while BRCA1-silence increased GPX4 protein level in
   A2780 cells ([89]Fig. 2A). Although, the levels of SLC40A1, HO-1 and
   SLC7A11 were changed in UWB-BRCA1 cells, we failed to detect similar
   changes in BRCA1-silenced A2780 cells ([90]Fig. 2A). Similar
   observations of increased GPX4 were also detected in A2780 BRCA1-KO
   cells and BRCA1-silenced SKOV3 or OVCAR-3 cells. ([91]Fig. 2B–C).
   Whereas, we did not detect significant changes of GPX4 mRNA levels in
   BRCA1-expressed UWB1.289 cells or BRCA1-silenced ovarian cancer cells
   ([92]Fig. 2D–E), suggesting that BRCA1 post-transcriptionally regulates
   GPX4 expression.

Fig. 2.

   [93]Fig. 2
   [94]Open in a new tab

   BRCA1 negatively regulates GPX4 protein stability. (A) Western blot
   analysis of indicated protein levels in UWB1.289, UWB1.289
   re-expressing BRCA1(UWB-BRCA1), control (ctrl) or BRCA1 siRNA
   transfected-A2780 cells. (B–C) Western blot analysis of GPX4 protein
   levels in Ctrl/BRCA1-KO A2780 cells (B), SKOV3 and OVCAR-3 cells
   transfected with ctrl or BRCA1 siRNA(C). (D–E) Relative mRNA levels of
   BRCA1 and GPX4 were quantified in UWB1.289/UWB-BRCA1 cells (D), or
   BRCA1-silenced A2780, SKOV3 and OVCAR-3 cells (E). (F) UWB1.289 and
   UWB-BRCA1 cells were treated with 100 μg/ml cycloheximide (CHX) for
   0–8 h, followed by Western blot analysis of indicated proteins. (G)
   A2780 cells were transfected with ctrl or BRCA1 siRNA, 48 h after
   transfection, cells were treated with 100 μg/ml cycloheximide (CHX) for
   0–8 h, then indicated proteins were assessed by Western blot. (H–I)
   IHC(H) and Western blot (I) analysis of GPX4 expression in WT or
   BRCA1-mutant ovarian cancer tissues. (J–K) UWB1.289, UWB-BRCA1 or
   UWB-BRCA1 cells stably expressing flag-GPX4 were treated with or
   without 0.5 μM RSL3 for 24 h, followed by analysis of lipid
   peroxidation using BODIY 581/591 C11 labeling (J) and cell death using
   PI staining (K). The GPX4 expressions in above cells were confirmed by
   Western blot. Data were presented as mean ± standard deviation (SD).
   *p < 0.05, **p < 0.01 and ***p < 0.001 by two-way ANOVA test. n.s.
   stands for no significance.

   We then examined whether BRCA1 affects GPX4 degradation through
   cycloheximide (CHX) chase assay, As shown in [95]Fig. 2F,
   BRCA1-restoration markedly accelerated GPX4 degradation. Similarly,
   BRCA1-silence delayed GPX4 degradation in A2780 cells ([96]Fig. 2G).
   Besides, treatment of proteasome inhibitor MG-132 significantly
   increased GPX4 protein level in UWB-BRCA1 cells ([97]Fig. S2A). To
   further confirm the regulation of GPX4 by BRCA1 in vivo, we analyzed
   GPX4 protein level in BRCA1-WT or mutant ovarian cancer tissues.
   Consistently, we detected significant higher level of GPX4 in
   BRCA1-mutant cancer tissues compared with BRCA1-WT cancer ([98]Fig.
   2H–I). These results demonstrate that BRCA1 promoting GPX4 degradation.
   Notably, forced expression of GPX4 significantly reversed RSL3
   induced-cell death and lipid peroxidation in UWB-BRCA1 cells ([99]Fig.
   2J–K). Thus, these results suggest BRCA1 promotes ferroptosis through
   the down-regulation of GPX4.

3.3. BRCA1 interacts with GPX4 via RING domain

   To investigate the underlying mechanism by which BRCA1 destabilizes
   GPX4, we performed immunoprecipitation (IP) analysis to examine whether
   BRCA1 physically interacts GPX4, and BRCA1/GPX4 interaction was
   detected in A2780 cells ([100]Fig. 3A). BRCA1/GPX4 interaction was also
   observed in A2780 cells exogenously co-expressed with Flag-GPX4 and
   HA-BRCA1 ([101]Fig. 3B and [102]Supplemental Fig. 2B). Meanwhile, we
   also found GPX4 interacted with BARD1, the heterodimeric binding
   partner of BRCA1 ([103]Fig. 3C). Then, we employed proximity ligation
   assay (PLA) to further confirm the BRCA1/GPX4 association.
   Consistently, PLA signals of BRCA1/GPX4 interaction was clearly
   detected, and their interaction was observed in both cytoplasm and
   nuclear of A2780 cells ([104]Fig. 3D), which is consistent with
   previous reports that both GPX4 and BRCA1 have been observed in
   cytoplasm and nucleus [[105]28,[106]29]. Next, we prepared purified
   his[6]-GPX4 to determine whether BRCA1/BARD1 directly interacts with
   GPX4 ([107]Fig. 3E). The in vitro IP assay confirmed that recombinant
   his[6]-GPX4 directly interacts with flag-BRCA1/BARD1 ([108]Fig. 3F).
   BRCA1 is a large protein comprises multiple functional domains
   including a RING finger domain, a pair of nuclear localization signals
   (NLS), a coiled-coil (CC) domain and a pair of BRCT domain from N- to
   C-terminus [[109]26]. To map the GPX4 binding site on BRCA1 protein, we
   generated five distinct BRCA1 deletion mutants ([110]Fig. 3G).
   Flag-GPX4 was co-expressed with full length or distinct truncated
   Myc-BRCA1 in A2780 cells, and the IP assay was carried out with
   anti-flag affinity gel and the result showed that removal of N-terminal
   303 amino acids abolished the interaction between BRCA1 and GPX4, while
   the BRCA1 truncations with N-terminal 303 amino acids maintained GPX4
   interaction ([111]Fig. 3H). This result indicating BRCA1 interacts with
   GPX4 through its N-terminus that contains RING domain. Given that GPX4
   is localized in cytoplasm and nucleus [[112]30]. We thus performed
   subcellular fractionation assay to examine BRCA1-mediated GPX4
   regulation. As shown in [113]Fig. 3I–J, elevation of both cytosolic and
   nuclear GPX4 were observed in BRCA1-deficient cells, which is
   consistent with BRCA1/GPX4 association in cytoplasm and nucleus.

Fig. 3.

   [114]Fig. 3
   [115]Open in a new tab

   BRCA1 interacts with GPX4 via RING domain. (A) Co-immunoprecipitation
   (co-IP) analysis of BRCA1/GPX4 interaction from lysates derived from
   A2780 cells. (B–C) A2780 cells were co-transfected with indicated
   plasmids, followed by IP analysis of BRCA1/GPX4 association using
   anti-flag affinity resin. (D) Proximity ligation assay (PLA) for the
   interaction between BRCA1 and GPX4 in A2780 cells. (E) The Coomassie
   blue staining of purified His-GPX4 proteins. (F) 100 ng of
   recombinant(r) his-GPX4 was incubated with or without purified
   flag-BRCA1/BARD1 protein in EBC buffer (50 mm Tris-HCl pH 7.6–8.0,
   120 mm NaCl, 0.5 % NP-40, 1 mm EDTA, 1 mm β-mercaptoethanol, 50 mm NaF,
   and 1 mmNa[3]VO[4]) for 20 min, followed by IP with anti-flag affinity
   resin. (G) Schematic diagram of BRCA1 protein and various myc tagged
   BRCA1 deletion mutants. (H) A2780 cells were co-transfected with
   Flag-GPX4 along with indicated myc tagged BRCA1 truncations, followed
   by IP with anti-flag affinity resin and Western blot with indicated
   antibodies. (I–J) Nuclear(N) and cytoplasmic (C) Fractions were
   isolated from UWB1.289/UWB-BRCA1 cells (I), parental or BRCA1-KO A2780
   (J) cells, and subjected to Western blot analysis with indicated
   antibodies. (For interpretation of the references to colour in this