FDI-6 inhibits the expression and function of FOXM1 to sensitize BRCA-proficient triple-negative breast cancer cells to Olaparib by regulating cell cycle progression and DNA damage repair

Inducing homologous-recombination (HR) deficiency is an effective strategy to broaden the indications of PARP inhibitors in the treatment of triple-negative breast cancer (TNBC). Herein, we find that repression of the oncogenic transcription factor FOXM1 using FOXM1 shRNA or FOXM1 inhibitor FDI-6 can sensitize BRCA-proficient TNBC to PARP inhibitor Olaparib in vitro and in vivo. Mechanistic studies show that Olaparib causes adaptive resistance by arresting the cell cycle at S and G2/M phases for HR repair, increasing the expression of CDK6, CCND1, CDK1, CCNA1, CCNB1, and CDC25B to promote cell cycle progression, and inducing the overexpression of FOXM1, PARP1/2, BRCA1/2, and Rad51 to activate precise repair of damaged DNA. FDI-6 inhibits the expression of FOXM1, PARP1/2, and genes involved in cell cycle control and DNA damage repair to sensitize TNBC cells to Olaparib by blocking cell cycle progression and DNA damage repair. Simultaneously targeting FOXM1 and PARP1/2 is an innovative therapy for more patients with TNBC.


INTRODUCTION
Poly(ADP-ribose) polymerases (PARPs) regulate various kinds of cell responses by transferring chains of ADP-ribose subunits to themselves or other target proteins [1]. Among the PARPs, PARP1/ 2 are important signaling molecules that involve in the repair of DNA single-strand break (SSB) through base excision repair (BER) pathway [2][3][4][5]. PARP1 is overexpressed in many cancers and correlated with poor prognosis [4][5][6][7]. DNA damage occurs under various conditions and can be divided into SSBs and doublestrand breaks (DSBs) [8]. Numerous DNA damage repair pathways are triggered by different conditions [8]. However, only a few pathways could result in the precise repair of damaged DNA and maintain the stability of chromosomes [8,9]. Homologousrecombination (HR) repair pathway is the most effective repair pathway that triggers precise repair of DSBs using a sister chromatid template for recombination [9,10]. Multiple essential HR genes, including breast cancer susceptibility gene 1/2 (BRCA1/ 2) and recombinase Rad51, are recruited to damaged DNA to facilitate repair [9][10][11][12]. When the enzymatic activities of PARP1/2 are inhibited, DNA SSBs translate to DNA DSBs [13]. When HR repair pathway is deficient, DNA DSBs are unrepaired or misrepaired [13,14]. Unrepaired DSBs induce cell death, and misrepaired DSBs promote unwanted chromosome rearrangements and genome instability [13]. Therefore, inhibition of PARP1 can effectively inhibit the growth of cancer cells with HR deficiency [14]. Several PARP inhibitors have been approved for the treatment cancers with HR deficiency [15]. Inhibiting PARP1 is a promising strategy for cancers with HR deficiency [15,16].
Triple-negative breast cancer (TNBC), which accounted for 10-15% of breast cancers, is the most refractory subtype of breast cancer with a 5-year survival 8-16% [17]. Because lacking estrogen, progesterone receptor and human epidermal growthfactor receptor-2 (HER2) expression, estrogen therapy and HER2targeted treatment do not work for TNBC [17,18]. Chemotherapy remains the standard therapy for TNBC, despite its limited benefit [17][18][19]. Recently, advances with novel agents have been made for a specific subgroup with germline BRCA-mutated tumors [17]. In particular, the success of Olaparib and other PARP inhibitors in the treatment of TNBC with BRCA1/2 mutation brings hope to patients suffering from this refractory cancer [20]. However, the majority of TNBC still does not benefit from PARP1 inhibitors [20]. Meanwhile, PARP inhibitor-induced increasing of HR repair capacity, decreasing of replication stress, and diminishing of PARP1 trapping and other changes could induce the generation of drug resistance [21,22]. Therefore, regulating HR capacity and/or cell cycle progression to sensitize more TNBC to PARP inhibitors is meaningful.
Recently, various new targets, reagents, and combination strategies have been explored to solve the limitation of PARP inhibitors, including the narrow clinical indications and adaptive resistance [20][21][22]. Inhibiting the enzymatic activity of kinases to induce HR deficiency is the mainly strategy to expand the indication of PARP inhibitors [20][21][22]. In addition to some kinases, transcription-factor forkhead box protein M1 (FOXM1) has been reported to involve in cell proliferation, cell cycle progression, DNA damage repair, apoptosis, and other biological processes [23,24]. FOXM1 is characterized by an evolutionarily conserved "forkhead" DNA-binding domain and overexpressed in many tumors, including TNBC [23,24]. Repression of FOXM1 has been reported to induce the death of cancer cells by inhibiting DNA damage repair and cell cycle progression [25,26]. FOXM1 transcriptionally activates cyclin B1 (CCNB1), cell-division cycle 25 A (CDC25A), and cell-division cycle 25B (CDC25B) to regulate cell cycle [27]. BRCA2 and X-ray cross-completing group (XRCC1), the critical genes involved in HR repair, are essential downstream targets of FOXM1 [28]. FDI-6 has been reported to block FOXM1 binding to DNA and suppress the transcription of genes under FOXM1 control [29,30]. Therefore, repressing FOXM1 to sensitize HR-proficient TNBC cells to Olaparib and other PARP inhibitors is a promising strategy. Herein, we selected the PARP inhibitor Olaparib and FOXM1 inhibitor FDI-6 to investigate their synergistic effect on the proliferation of BRCA-proficient TNBC cells in vitro and in vivo. Data in the Cancer Genome Atlas (TCGA) project showed that FOXM1 is generally overexpressed in tumor tissues but not in normal tissues. FDI-6 and Olaparib, at the ratio of 1:1 (FDI-6/ Olaparib), synergistically inhibited the proliferation of BRCAproficient TNBC cells in vitro and in vivo by blocking the progression of cell cycle and the precise repair of damaged DNA. The RNA sequencing results showed that Olaparib initiated adaptive resistance in MDA-MB-231 cells by inducing the recovery of HR repair capacity and acceleration of cell cycle progression. Unexpectedly, FDI-6 reversed Olaparib-induced adaptive resistance by inhibiting the expression of FOXM1 and PARP1/2 and regulating genes involved in cell cycle control and DNA damage repair. Simultaneously targeting PARP1/2 and FOXM1 is a promising strategy for more patients with TNBC.

Drug-combination assay
The following combination ratios of FDI-6/Olaparib were selected: 1:0.25, 1:0.5, 1:1, 1:2, and 1:4. After TNBC cells were treated with FDI-6 and Olaparib at different combination ratios for 7 days, the viabilities of MDA-MB-231 cells, MDA231-LM2 cells, MDA-MB-468 cells and HCC1937 cells were measured by MTT assay. Untreated cells served as the control. The combination-index (CI) values were calculated by CompuSyn software using the equation CI = C A,X /IC X,A + C B,X /IC X,B [31] . Herein, C A,X and C B,X represent the concentrations of FDI-6 and Olaparib, respectively, that achieve an X% growth-inhibition ratio. IC X, A and IC X, B are the concentrations of single drugs (FDI-6 or Olaparib) that achieve the same growth-inhibition ratio. A CI value <1.0 represents synergy, CI = 1.0 represents additivity, and CI > 1.0 represents antagonism [31].

Gene expression and correlation analysis
The differential expression of FOXM1, PARP1, and PARP2 between cancer tissues and adjacent normal tissues in TCGA project was analyzed by Gene Expression Profiling Interactive Analysis version 2 (GEPIA2). Box plots were performed to compare the differential expression of PARP1, PARP2, and FOXM1 between breast-invasive carcinoma tissues and adjacent normal tissues of the Genotype Tissue Expression (GTEx) database with the following settings: P-value cutoff = 0.01, log 2 FC cutoff = 1.0, and "match TCGA normal and GTEx data". The correlation between PARP1 and FOXM1 in breast-invasive carcinoma from the TCGA project was also analyzed by GEPIA2.

Colony-formation assay
All cells were seeded in 6-well plates at a concentration of 500 cells/well. For the silencing groups, MDA-MB-231 cells stably expressing FOXM1 shRNA and NC shRNA were cultured in medium containing 0.5 μM puromycin with 4.0 μM Olaparib for 14 days. For the drug-treated groups, TNBC cells were exposed to various concentrations of FDI-6 and/or Olaparib for 14 days. The medium was changed every 2-3 days to allow colony formation. After 14 days of treatment, the colonies were fixed with cold methanol for 10 min, stained with 0.1% crystal violet for 10 min, and imaged using an upright biological microscope (Olympus BX53, Tokyo, Japan). Experiments were repeated at least three times.

Alkaline comet assay
An alkaline comet assay was performed using a comet assay kit (KeyGEN Biotech, Nanjing, Jiangsu, China) [31]. For the silencing groups, MDA-MB-231 S.-P. Wang et al. cells stably expressing FOXM1 shRNA and NC shRNA were treated with 4.0 μM Olaparib for 7 days. For the drug-treated groups, MDA-MB-231 cells and MDA-MB-468 cells were treated with different concentrations of FDI-6, Olaparib, or their combination for 7 days. An alkaline comet assay was performed following the manufacturer's protocol. In brief, cells (1 × 10 4 /ml) were mixed with low-melting-point agarose at a ratio of 1:10 (v/v), layered onto slides, lysed by lysis buffer at 4°C for 2 h, and then unwound with alkaline-unwinding solution for another 30 min at room temperature. Following electrophoresis at 21 V for 30 min, the cells were stained with propidium iodide (PI) and observed with an inverted biological microscope (Olympus BX53, Tokyo, Japan). Five images were randomly captured per slide. After treatment, the cells were collected and stained with Annexin V-FITC and PI following the manufacturer's protocol to analyze cell apoptosis. Cellular DNA was stained with PI following the manufacturer's protocol. Cell apoptosis and the cell cycle were detected by BD FACSCelesta flow cytometry (New York, USA). Apoptosis data were analyzed by FlowJo 7.6, and the cell cycle data were analyzed by ModFit LT [31].

Immunofluorescence assay
For the silencing groups, MDA-MB-231 cells stably expressing FOXM1 shRNA and NC shRNA were treated with 4.0 μM Olaparib for 7 days. For the drug-treated group, MDA-MB-231 cells and MDA-MB-468 cells were treated with different concentrations of FDI-6, Olaparib or their combination for 7 days. After treatment, cells were fixed with 4% formaldehyde, permeabilized with 0.2% (v/v) Triton X-100 in PBS, blocked with 1% (w/v) bovine serum albumin (BSA) in PBS for 1.0 h, and stained with anti-p-Histone 2AX (γH2AX) antibody labeled with Alexa Fluor 488. After staining, cellular DNA was counterstained with 4,6-diamidino-2phenylindole (DAPI). Fluorescence signals were detected using a Carl Zeiss LSM700 laser confocal microscope (Jena, Germany). Five fields per sample were quantified.

Animal care and treatment
All animal procedures were carried out in accordance with the institutional guidelines for the care and use of laboratory animals and approved by the committee on the ethics of Animal Experiments of China Pharmaceutic University. Every effort was made to ensure the comfort and safety of the animals. BALB/C nude mice (female, 5-6 weeks of age, weighting 18-20 g) and Swiss ICR mice (female/male, 5-6 weeks of age, weighting 18-20 g) were provided from the Animal House in Model Animal Institute of Nanjing University (Nanjing, Jiangsu, China). BALB/C nude mice were used to detect the in vivo antitumor activity of FDI-6, Olaparib, and their combination on the growth of MDA-MB-231 cells. Swiss ICR mice were used to detect the acute toxicity of FDI-6, Olaparib and their combination. For in vivo antitumor assay, MDA-MB-231 cells (5×10 6 ) were injected into the right flank of each BALB/C nude mouse. After the average tumor volume (mm 3 ) reached 100 mm 3 , the tumorbearing mice were randomly divided into four groups (n = 5/group). FDI-6 (60 mg/kg), Olaparib (60 mg/kg), and the combination of FDI-6 (30 mg/kg) and Olaparib (30 mg/kg) were administered by intraperitoneal injection for 28 consecutive days. Tumor volume was recorded every two days. After 28 days of treatment, the mice were killed to detect the weight of tumors in each group. Tumor volume was measured with a Vernier caliper and calculated using the formula (ab 2 )/2, where a and b represent the length and width of the tumor, respectively. For acute-toxicity assay, Swiss ICR mice (female/male, 5-6 weeks of age, weighting 18-20 g) were divided into six groups (n = 10/group, female = 5, male = 5). Menstruum, FDI-6 (100 mg/kg, 200 mg/kg), Olaparib (200 mg/kg), and their combination (FDI-6 + Olaparib: 100 + 100 mg/kg, 200 + 200 mg/kg) were administered by intraperitoneal injection on the first day. After treatment, the weight and life status of ICR mice were recorded for seven consecutive days.

RNA sequencing and data processing of DEGs
MDA-MB-231 cells were treated with Olaparib (8.0 μM), FDI-6 (8.0 μM), or the combination of FDI-6 (4.0 μM) and Olaparib (4.0 μM). After treatment for 7 days, cells were collected to obtain total RNA using TRIzol reagent (Vazyme, Nanjing, China) according to the manufacturer's manual. A total of 500 ng of RNA was used to prepare libraries using the NEBNext Ultra RNA Library Perp Kit for Illumina. RNA quantity and quality were assessed on an Agilent 2100 Bioanalyzer. RNA library sequencing was performed on an Illumina HiSeqTM 2500/4000 by Gene Denovo Biotechnology Co., Ltd. (Guangzhou, Guangdong, China). Differentially expressed genes (DEGs) in the control group vs the Olaparib-treated group, the control group vs the FDI-6-treated group, and the control group vs the FDI-6/Olaparib cotreated group were identified based on a | log 2 FC | >1.0 and an adjusted P < 0.05. DEGs with a log 2 FC < 1.0 were considered downregulated genes, while DEGs with a log 2 FC > 1.0 were considered upregulated genes [32].

Gene ontology and KEGG pathway enrichment analysis
The characteristic biological attributes of the DEGs were identified by gene ontology (GO) enrichment analysis. The functional attributes of the DEGs were identified by KEGG pathway enrichment analysis. The GO enrichment and KEGG pathway enrichment analyses were performed using Omicsmart, a real-time interactive online platform for data analysis (http://www. omicsmart.com) [32].

PPI network construction and module analysis
Four DEGs were selected based on the GO and KEGG enrichment analyses. The Search Tool for the Retrieval of Interacting Genes (STRING) was used to evaluate and establish the protein-protein interaction (PPI) networks of PARP1, PARP2, FOXM1 and the four DEGs. The STRING application in Cytoscape was used to detect the potential correlations between these DEGs. The MCODE application in Cytoscape was utilized to examine the modules of the PPI network and identify the central genes among the DEGs (degree cutoff = 2, max. depth = 100, k-core = 2, and node-score cutoff = 0.2) [32].  [31]. cDNA was synthesized using a HiScript II one-step RT-PCR kit (P612-01, Vazyme, Nanjing, China) with 1.0 μg of total RNA in a 20 μl reaction system. The resulting cDNA was diluted 1:2 in nuclease-free water, and 1.0 μl was used per Q-PCR in triplicate. Q-PCR was carried out using ChamQ SYBR Q-PCR master mix (Q311-02, Vazyme, Nanjing, China) on a QuantStudio 3 real-time PCRdetection system (Life Tech, New York, USA) including a nontemplate negative control. GAPDH was used to normalize the level of mRNA expression. The sequences of the primers are listed in Supplementary  Table 2.

Western blot analysis
Total proteins in MDA-MB-231 cells stably expressing FOXM1 shRNA and NC shRNA, MDA-MB-231 cells, and MDA-MB-231 xenografts were extracted with RIPA cell-lysis buffer (Beyotime, Shanghai, China) with added protease/phosphatase-inhibitor cocktail. For nuclear-fraction assays, MDA-MB-231 cells were collected and lysed with cytoplasmic extraction buffer from the Nuclear/Cytoplasmic Protein Extraction Kit (KGP1100, KeyGEN Biotech, Nanjing, Jiangsu, China) and then centrifuged at 3000 rpm for 10 minutes. The supernatant was labeled as the cytoplasmic extract. The pellets were further lysed with nuclear-extraction buffer to obtain nuclear extracts. Protein concentration was determined by the BCA assay, and equal amounts of proteins were loaded for Western blot analysis. In brief, equal amounts of total proteins were loaded for SDS-PAGE and transferred onto a PVDF membrane (Millipore, Billerica, MA, USA). Membranes with protein were blocked with 5% (w/v) skim milk, incubated with primary antibody in Supplementary Materials, and then incubated with secondary antibodies (1:2000) for detection. β-Actin and Histone H3 were used to normalize the level of protein expression. Densitometric analysis was performed with ImageJ software.

H&E staining
The lung, liver, heart, kidney, and spleen-tissue samples of xenograft model mice were fixed with 4% paraformaldehyde, dehydrated with ethanol, immersed in xylene, embedded in paraffin, and cut into 4.0-μm longitudinal sections. The paraffin-embedded sections were stained with hematoxylin and eosin (H&E) according to the manufacturer's instructions (Beyotime, Shanghai, China). Each group of samples was observed with a DM6B-positive fluorescence microscope (Leica, Frankfurt, Germany). Five images were randomly captured per slide.

Immunohistochemical staining
The tumors from xenograft-model mice were embedded in paraffin and cut into longitudinal sections. The paraffin-embedded sections were incubated with 0.3% hydrogen peroxide for 30 min to block endogenous peroxidase and then incubated with 1.0% bovine serum albumin (BSA) for blocking. After blocking, the paraffin-embedded sections were incubated with the primary antibody overnight at 4°C, incubated with secondary antibody for another 1 h at room temperature and then counterstained for 1 min with hematoxylin. Each group was examined using a DM6B-positive fluorescence microscope (Leica, Frankfurt, Germany). Five images were randomly captured per slide. The percentage of stained dots was analyzed by ImageJ software.

Statistical analysis
The data were analyzed using SPSS 19.0. The results are expressed as means ± SD. Differences between treatment regimens were analyzed by two-tailed Student's t-test or one-way ANOVA. P < 0.05 was considered statistical significance.

FOXM1 repression inhibits PARP1 expression and sensitizes TNBC cells to Olaparib
Transcription factor FOXM1, involving in HR-mediated repair, is a promising target for solving the limitation of Olaparib and other PARP inhibitors [23,24]. To evaluate whether FOXM1 is a valuable target, we analyzed the differential expression of FOXM1, PARP1, and PARP2 between cancer tissues and adjacent normal tissues in 31 cancers from the TCGA projects (Fig. 1A, B and Supplementary  Fig. 1A, B). FOXM1 and PARP1 were generally overexpressed in various cancer tissues (Fig. 1A, B and Supplementary Fig. 1A, B). Compared with PARP1/2, the expression levels of FOXM1 in normal tissues were very low (Fig. 1A, B). Targeting FOXM1 would kill tumor cells in a selective way, which could avoid the toxicity to normal cells. Subsequently, we analyzed the correlation of FOXM1 vs PARP1, FOXM1 vs PARP2 and PARP1 vs PARP2 using the data from TCGA project (Fig. 1C-E). The results showed that FOXM1 expression was positively correlated with PARP1 (R = 0.3, P < 0.01) and PARP2 (R = 0.28, P < 0.01) expression, which suggested that FOXM1 and PARP1/2 could promote the expression of each other (Fig. 1C, D). The results of Q-PCR and Western blots showed that the expressions of FOXM1 and PARP1/2 were high in the four TNBC cells, and the nuclear accumulation of FOXM1 was the highest in MDA-MB-231 cells (Fig. 1F, G and Supplementary Fig. 2). Silencing FOXM1 repressed Olaparibinduced PARP1 expression, inhibited colony formation and promoted cell apoptosis (Fig. 1H-K and Supplementary Fig. 2). Moreover, silencing FOXM1 enhanced the effects of Olaparib on colony formation and cell apoptosis (Fig. 1J, K). Thus, FOXM1 knockdown increased the sensitivity of BRCA-proficient TNBC cells to Olaparib.

FDI-6 and Olaparib synergistically inhibit the growth of TNBC cells in vitro and in vivo
FDI-6 is a small-molecular inhibitor of FOXM1 selected by highthroughput screening [30]. To confirm whether it is suitable to be a tool molecular, we tested the selectivity of FDI-6 against 361 kinases at the concentration of 1.0 μM in the Reaction Biology Corporation kinase panel ( Fig. 2A, B and Supplementary Table 3).
The inhibition ratios of FDI-6 against most kinases were under 20%, which proved that FDI-6 is a selective inhibitor of FOXM1. After 7 days of treatment, the proliferation of the four TNBC cell lines was significantly inhibited by FDI-6 and Olaparib (Fig. 2C). The synergistic effects of FDI-6, Olaparib, and their combination at fixed dose ratios on the proliferation of the four TNBC cells were investigated (Fig. 2D, E). Combining FDI-6 with Olaparib not only inhibited the growth of BRCA-deficient cells but also significantly inhibited the growth of BRCA-proficient MDA-MB-231, MDA-MB-468 and MDA 231-LM2 cells. FDI-6 and Olaparib synergistically inhibited the growth of the four TNBC cell lines at the dose ratio of 1:1 (CI < 1.0). Therefore, we selected the dose ratio of 1:1 for further study. FDI-6 and Olaparib synergistically inhibited the formation of colonies and promoted the apoptosis of MDA-MB-231 cells and MDA-MB-468 cells (Fig. 2F-H and Supplementary Fig.  3A).  Supplementary Fig. 3B). After analysis of H&E-stained organs, we found that FDI-6, Olaparib and their combination did not cause significant damage to the kidney, lung, spleen, liver or heart (Fig. 3E). The results of acute-toxicity experiment showed that FDI-6, Olaparib and their combination did not induce the death of mice and significantly affect their life status. FDI-6 (200 mg/kg) and the combination of FDI-6 (200 mg/ kg) and Olaparib (200 mg/kg) induced the decreasing of femalemice weight but did not affect the weight of male mice (Fig. 3F). The results of Q-PCR and Western blots showed that Olaparib increased the expression of FOXM1 and PARP1/2 in vitro and in vivo (Fig. 3G-J, Supplementary Fig. 4A-C, Supplementary Fig. 5 and Supplementary Fig. 6). FDI-6 inhibited the expression of FOXM1 and PARP1 and impaired Olaparib-induced overexpression of FOXM1 and PARP1 in vitro and in vivo (Fig. 3G-J and Supplementary Fig. 4A-C). Further investigation of immunohistochemistry confirmed that FDI-6 could reverse Olaparib-induced accumulation of FOXM1 and PARP1 in vivo (Fig. 3K, L).  Table 4, Supplementary Table 5 and  Supplementary Table 6) After GO functional enrichment, we found that DEGs induced by Olaparib were mostly associated with cell proliferation, DEGs induced by FDI-6 were mostly associated with DNA replication, and DEGs induced by the combination of FDI-6 and Olaparib were mostly associated with DNA replication (Fig.  4B). The results of KEGG pathway enrichment showed that the main pathways associated with the DEGs were the cell cycle and DNA replication (Fig. 4C). By a pairwise comparison of control group vs Olaparib-treated group, control group vs FDI-6-treated group and control group vs the FDI-6/Olaparib cotreated group, we selected five DEGs to analyze their expression. Olaparib significantly promoted the expression of cyclin-dependent kinase 6 (CDK6), FOXM1, double-cortin-like kinase 1 (DCLK1), and cyclin A1 (CCNA1). FDI-6 inhibited the expression of CDK6, FOXM1, DCLK1, CCNA1, and polo-like kinase 1 (PLK1) (Fig. 4D). Subsequently, we constructed a PPI network using PARP1, PARP2, and the five genes as the centers by STRING. As shown in Fig. 4E, 23 genes were selected for further mechanistic investigation.  Supplementary Fig. 7A). Similarly, silencing the expression of FOXM1 blocked the G 2 /M transition and arrested the cell cycle at the G 2 /M phase (Fig. 5B). Q-PCR and Western blots were used to analyze the expression of genes involved in cell cycle control based on the results of bioinformatic analysis in vitro and in vivo ( Fig. 5C-G, Supplementary Fig. 7B, Supplementary Fig. 8A-C,  Supplementary Fig. 9 and Supplementary Fig. 10). CDK6, cyclin D1 (CCND1), cyclin-dependent kinase 2 (CDK2), cyclin E2 (CCNE2), CDC25A and E2F transcription factor 2 (E2F2) regulate the G 1 /S transition, and cyclin-dependent kinase 1 (CDK1), CCNA1, CCNB1,  and CDC25B regulate the G 2 /M transition [33]. Olaparib significantly promoted the expression of CDC25B, CDK1, CCNA1, CCNB1, CDK6 and CCND1 to accelerate mitosis. FDI-6 significantly inhibited the expression of the ten genes-involved in cell cycle control and impaired Olaparib-induced expression of the six genes in MDA-MB-468 cells, MDA-MB-231 cells, and MDA-MB-231 tumor xenografts (Fig. 5C-G, Supplementary Fig. 7B and Supplementary  Fig. 8A-C).
FDI-6 promotes Olaparib-induced DNA damage by inhibiting precise repair of damaged DNA in vitro and in vivo Damaged DNA has a tail in the comet assay that resembles a comet [34]. FDI-6, Olaparib, and their combination significantly increased the number and extent of tailed DNA. FOXM1 knockdown also promoted DNA damage, and the number and extent of tailed DNA were significantly increased (Fig. 6A, B and Supplementary Fig. 11A). The accumulation of γH2AX also reflects the The results from three independent experiments were statistically analyzed using one-way ANOVA: * P < 0.05, ** P < 0.01 compared with the control; # P < 0.05, ## P < 0.01 compared with the FDI-6/Olaparib combined group (FDI-6/Olaparib: 30 + 30 mg/kg). extent of DNA damage [35]. The results of immunofluorescence showed that FDI-6, Olaparib, and their combination significantly increased nuclear γh2AX levels (Fig. 6C, D). Subsequently, we found that Olaparib significantly increased the expression of DCLK1, mediator of DNA damage checkpoint (MDC1), PLK1, BRCA1, BRCA2, and Rad51, and inhibited the expression of XRCC1 and X-ray cross-complementing group-2 protein (XRCC2) in vitro. FDI-6 and FOXM1 shRNA promoted MDC1 expression and inhibited the expression of DCLK1, PLK1, BRCA1, BRCA2, Rad51, XRCC1, and XRCC2 ( Fig. 6E-G, Supplementary Fig. 11B, C, Supplementary Fig. 12A, Supplementary Fig. 13 and Supplementary Fig. 14). Further in vivo investigation showed that FDI-6, Olaparib, and their combination induced DNA damage in vivo, and the expression of γH2AX was significantly increased (Fig. 6H, I). The regulation of FDI-6, Olaparib and their combination on the expression of genes involved in DNA repair was similar to their efficacy in vitro (Fig. 6H-K, Supplementary Fig. 12B and Supplementary Fig. 14).

DISCUSSION
Currently, a few targeted therapies approved for the treatment of TNBC, the most challenging breast cancer subtype that has the poorest prognosis and median-survival rate after relapse [20]. Targeting PARP1/2 for synthetic lethality is an effective strategy for TNBCs with BRCA1/2 mutations [20]. However, the majority of TNBCs are BRCA1/2 wild type, and their prognosis remains poor, particularly after relapse [17][18][19][20]. Therefore, inducing HR deficiency is an effective strategy to broaden the indication of PARP inhibitors. Oncogenic transcription factor FOXM1 is overexpressed in breast cancer and plays important roles in DNA damage repair by regulating the expression of genes essential for DNA damage sensing, mediation, signaling, and repair, as well as for cell cycle and cell death control [23,24]. Targeting FOXM1 in monotherapy or combination therapy may have promising therapeutic benefits [14,28]. Herein, we inhibited the expression and function of FOXM1 to investigate its effects on the sensitivity of BRCA-proficient TNBC cells to Olaparib.
the ratio of 1:1 synergistically inhibited the growth of BRCAproficient TNBC cells in vitro and in vivo. Further mechanistic studies showed that FDI-6 also inhibited the expression of FOXM1 and PARP1/2 to sensitize BRCA-proficient TNBC cells to Olaparib. DCLK1 expression was significantly promoted by Olaparib. DCLK1 is a microtubule-associated oncoprotein that regulates cell apoptosis and DNA repair to promote tumorigenesis [36]. Therefore, the overexpression of DCLK1 contributes to the generation of adaptive resistance. Silencing FOXM1 expression or inhibiting FOXM1 function significantly inhibited the expression of DCLK1 to impair Olaparib-induced DCLK1 expression. The results from three independent experiments were statistically analyzed using one-way ANOVA: * P < 0.05, ** P < 0.01 compared with the control; # P < 0.05, ## P < 0.01 compared with the FDI-6/Olaparib-combined group (FDI-6/Olaparib: 4.0 + 4.0 μM, FDI-6/Olaparib: 30 + 30 mg/kg).
In response to DNA damage, cells trigger complex molecular reactions, such as detecting DNA damage, arresting cell cycle progression for DNA repair, and initiating DNA repair pathways [37]. Olaparib significantly increased the expression of CDK6, CCND1, CDK1, CCNA1, CCNB1, and CDC25B to accelerate cell cycle progression and arrested cell cycle at S and G 2 phases. Olaparib induces the accumulation of DSBs, which are critical DNA lesions [9,10]. HR-mediated repair triggers precise repair of DNA DSBs [9,10,20]. HR is activated only in the S and G 2 phases after DNA replication [33]. Thus, the arresting of cell cycle at the S and G 2 phases induced by Olaparib would activate HR repair, which is not conducive to killing cancer cells [37]. FOXM1 is a critical regulator of the G 1 /S and G 2 /M cell cycle transitions, as well as mitosis [27]. FDI-6 has been reported to inhibit the expression of downstream target genes of FOXM1, such as CDC25B and CCNB1 [30]. We found that FDI-6 and FOXM1 shRNA impaired Olaparib-induced expression of CDK1, CCNA1, CCNB1, and CDC25B to inhibit the mitosis of cancer cells. Meanwhile, the inhibition of FOXM1 repressed the expression of CDK6, CCND1, CDK2, CCNE2, E2F2, and CDC25A to arrest the cell cycle at G 2 phase, which blocks the promotion of cancer-cell mitosis [27].
In addition to cell cycle regulation, FOXM1 was implicated in the DNA damage response. XRCC1 and BRCA2 are the direct transcriptional targets of FOXM1 [28]. We found that inhibition of FOXM1 promoted Olaparib-induced DNA damage. Further mechanistic studies showed that Olaparib promoted the expression of PLK1, MDC1, BRCA1/2 and Rad51 and inhibited the expression of XRCC1/2. BRCA1/2, and Rad51 are directly involved in the HR repair pathway [10][11][12]. PLK1 plays a critical role in cell cycle progression and the HR repair [38]. PLK1 directly phosphorylates Rad51 to increase its interaction with Nbs1 to activate HR repair [38]. MDC1 is a tumor suppressor that interacts with Rad51 to facilitate HR repair [39]. XRCC1 is an essential scaffold protein for BER that interacts with multiple enzymatic components of DNA SSB repair to accelerate the repair of DNA SSBs [40]. PARP1 is required for the recruitment of XRCC1 to the correct position on DNA SSBs [40]. XRCC2 functions in HR-mediated DNA DSBs by enhancing the activity of Rad51 [41]. Olaparib-induced expression of PLK1, BRCA1/2, and Rad51 could promote HR repair. Inhibition of FOXM1 reverses Olaparib-induced promotion of HR repair and promotes Olaparib-induced XRCC1 inhibition to impair BER of SSBs.
In summary, we clarified the mechanism for the synergistic effect of FDI-6 and Olaparib on the growth of BRCA-proficient TNBC cells in vitro and in vivo. After a long period of treatment, Olaparib increases the expression of DCLK1, PARP1/2, FOXM1, and other genes to induce the recovery of DNA repair and acceleration of cell cycle progression, which causes the generation of adaptive resistance (Fig. 7A). For cell cycle, Olaparib increases the expression of CDK6 and CCND1 to promote the G 1 /S transition and causes the overexpression of CDK1, CCNA1, CCNB1, and CDC25B to accelerate cell mitosis. For DNA repair, Olaparib promotes the expression of PARP1/2 to recover the repair of DNA SSBs and induces the expression of PLK1, XRCC2, BRCA1/2, and Rad51 to promote the HR repair of DNA DSBs (Fig. 7A). FDI-6 inhibits the function and expression of FOXM1 and sensitizes TNBC cells to Olaparib. FDI-6 and Olaparib synergistically inhibit the expression of CDK6, CCND1, CDK1, CCNA2, CCNB1 and CDC25B to block cell cycle progression. Meanwhile, FDI-6 and Olaparib synergistically inhibit DNA SSB repair by inhibiting the expression of XRCC1 and PARP1/2. They also synergistically inhibit HR repair by decreasing the expression of DCLK1, PLK1, XRCC1, BRCA1/2, and Rad51 and increasing the expression of MDC1 (Fig.  7B). We believe that targeting FOXM1 and PARP1 in combination is a promising strategy for the treatment of TNBCs.

DATA AVAILABILITY
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.