Choline metabolism drives metastasis in BRCA1-deficient ovarian cancers by activating FAM3C Authors: Cong Wang, Xuexia Xie, Jiahuang Li, Jiawei Zhu, Binghe Sun, Yue Wu, Wanying Zhang, Dake Li, Guo Chen Published: 2026 ABSTRACT BRCA1 mutations are a hallmark of hereditary ovarian cancer, strongly linked to deficiencies in homologous recombination (HR) DNA repair and impaired DNA replication fork protection. However, its roles in cancer progression beyond maintaining genomic integrity remain poorly understood. Here, through metabolomics approaches, we find that BRCA1-deficiency strikingly increases choline metabolism. Loss of BRCA1 promotes choline uptake through upregulating choline transporter-like protein 4 (CTL4). BRCA1 directly binds and recruits EZH2-mediated H3K27Me3 deposition to the CTL4 promoter. CTL4 is therefore overexpressed in ovarian cancer tissues with BRCA1 mutations. Furthermore, BRCA1-deficiency significantly promotes ovarian cancer invasion, while inhibition of CTL4 reverses the high metastatic potential of BRCA1-deficient ovarian cancer cells, suggesting the functionality and specificity of CTL4 as a therapeutic target. Additionally, we discover that phosphocholine, the choline metabolite increased by CTL4 overexpression, interacts with and stabilizes the epithelial-to-mesenchymal transition inducer FAM3C in BRCA1-deficient ovarian cancer cells. Importantly, we identify a potent CTL4 inhibitor, DT-13, which significantly reduces choline metabolism and effectively suppresses metastasis in BRCA1-deficient ovarian cancers. Therefore, our study uncovers a mechanism underlying metastasis in BRCA1-deficient cancers and identifies CTL4 as a therapeutic target for metastatic ovarian cancer patients with BRCA1 mutations. Introduction Metabolic reprogramming represents a hallmark of cancer cells and plays critical roles in cancer prognosis1. Among the various metabolic alterations observed in cancer cells, alteration of choline metabolism, characterized by an elevation of choline-containing metabolites such as phosphocholine (Pho-Cho), phosphatidylcholine (PtdCho), and sphingomyelin (SM), has emerged as a crucial feature with profound implications for cancer development and progression2–4. As an essential nutrient, choline is physiologically involved in various biological processes, such as maintaining the structural integrity of the cell membrane, providing neurotransmitter acetylcholine, and methyl donors5. Key enzymes in choline metabolism, such as choline kinase (CHK), phosphocholine cytidylyltransferase (CCT), and PtdCho-specific phospholipase D (PC-PLD), are frequently overexpressed or hyperactivated in cancer cells6–8. These enzymes not only contribute to the altered metabolic landscape of cancer but also hold potential as prognostic biomarkers and therapeutic targets. Although the activation of choline metabolism is strongly linked to oncogenic transformation and progression, the mechanisms through which it drives cancer progression remain largely unknown. Metabolic rewiring in cancer cells is driven by oncogenes and tumor suppressors to support their proliferation and metastasis potential9–12. BRCA1 is a well-documented cancer predisposition gene, which is frequently mutated in a variety of cancers, particularly in breast and ovarian cancers13. Pathogenic variants in BRCA1 are strongly linked to elevated risks of these cancers. The BRCA1 gene comprises 22 exons that encode a protein consisting of 1863 amino acids. Loss function of BRCA1 leads to genomic instability, accompanied by a defect in homologous-recombination (HR) mediated-DNA double strand break (DSB) repair and DNA replication fork protection14. BRCA1 directs DSB repair pathways toward high-fidelity HR by promoting DNA-end resection. As a result, cancer cells with deleterious BRCA1 mutations exhibit HR repair deficiencies and synthetic lethality with poly-ADP ribose polymerase inhibitors (PARPi)15. Beyond its role in safeguarding genomic integrity, BRCA1 participates in various physiological processes through transcriptional and post-translational regulation of gene expression. BRCA1 directly binds to gene promoter regions and can either stimulate or repress transcription16,17. In addition, BRCA1 protein possesses a RING domain at its N-terminus that exerts E3 ligase activity when heterodimerized with BRCA1-associated RING domain 1(BARD1)13. BRCA1/BARD1 ubiquitinates a number of oncogenic proteins, including phosphorylated RNA polymerase II, activated Akt, and Hippo signaling regulator NF218–20. Thus, while BRCA1 potently prevents tumorigenesis through a variety of mechanisms, whether BRCA1 could shape metabolic pathways to promote cancer development is still not fully understood. In light of its versatile functions, BRCA1 deficiency confers unique characteristics to cancer cells. Recent studies have revealed that tumors with BRCA1-mutations responded poorly to immune checkpoint blockade (ICB), despite alterations in the HR repair pathway represent benefit biomarker for ICB21. Moreover, an inflammatory microenvironment with increased microvascular density has been found in BRCA1-mutant cancer tissues due to the activation of STING signaling22. Loss of BRCA1 has also been associated with epithelial-to-mesenchymal transition (EMT) and increased invasion in breast cancer cells23. However, the mechanisms through which BRCA1 influences cancer cell motility remain unclear. Here, we found that BRCA1-deficiency enhanced choline uptake through upregulation of choline transporter CTL4, which consequently promotes metastasis in ovarian cancer cells. Mechanistically, the choline metabolite phosphocholine interacts with and inhibits the polyubiquitination of pro-metastatic proteins, stabilizing their expression and enhancing metastatic potential. We further employed RNA-seq and ChIP-seq to demonstrate that BRCA1 binds directly to the CTL4 promoter and silences its expression by recruiting the histone methyltransferase EZH2. Genetic manipulation and pharmacological targeting of CTL4 effectively suppresses metastasis in BRCA1-deficient ovarian cancers, identifying CTL4 as a promising therapeutic target for managing metastatic ovarian cancer in patients with BRCA1 mutations. Results BRCA1 loss enhances choline metabolism in ovarian cancer cells As a well-defined tumor suppressor, loss of function of BRCA1 promotes tumorigenesis and accelerates cancer progression24. To investigate whether BRCA1 deficiency drives carcinogenesis through remodeling of cellular metabolism pathways, we employed BRCA1-null human ovarian cancer cell UWB1.289 and its isogeneic BRCA1 re-expressing cell line UWB1.289 + BRCA1 (UWB-BRCA1) (Fig. 1a). Then, metabolomics was used to profile intracellular metabolites in UWB1.289 and UWB-BRCA1. The pathway enrichment of altered metabolites showed that many metabolic pathways were affected by BRCA1 expression, with the choline metabolism pathway most significantly regulated (Fig. 1b and Supplementary Data 1). Hierarchical clustering and heatmap analysis revealed that the level of several metabolites from the choline metabolism pathway was significantly increased in UWB1.289 cells compared with UWB-BRCA1 cells (Fig. 1c). Choline metabolism is critically important for cell membrane synthesis and malignant transformation2, we thus focus on this metabolic pathway for further mechanism study. Choline metabolism primarily facilitates sphingomyelin (SM) synthesis, a critical constituent of cellular membranes, through an enzyme cascade comprising choline kinase (CK), CTP: phosphocholine cytidylyltransferase (PCYT1), CDP-choline:1,2-diacylglycerol cholinephosphotransferase (CPT), and sphingomyelin synthase (SMS). As shown in Fig. 1d, the critical metabolic intermediates, including phosphocholine, CDP-choline, phosphatidylcholine (PC), and sphingomyelin (SM), were significantly upregulated in BRCA1-deficient UWB1.289 cells from metabolomics analysis.Fig. 1BRCA1-deficiency drives remodeling of choline metabolism.a Western blot analysis of BRCA1 protein expression in UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells.The samples derive from the same experiment but different gels; all gels were processed in parallel. b KEGG enrichment analysis of changed metabolites showing the significantly affected metabolic pathways. (Variable Importance for the Projection, VIP > 1, p value < 0.05). c Heatmap and cluster analysis of differentiated metabolites between UWB1.289 and UWB1.289 + BRCA1 (UWB-BRCA1) cells. (Variable Importance for the Projection, VIP > 1, p value < 0.05). d Illustration of the choline metabolism pathway and the quantitative results of the changed choline-containing metabolites (n = 5). e Western blot analysis of BRCA1 protein level in A2780 cells transfected with control (ctrl) or BRCA1 siRNA. f Western blot analysis of indicated proteins in parental or BRCA1 knockout (KO) KGN cells. g Quantitative measurement of the intracellular level of choline or sphingomyelin (SM) in A2780 cells transfected with ctrl or BRCA1 siRNA (n = 3). h Measurement of the intracellular level of choline or SM in parental or BRCA1-KO KGN cells (n = 3). i Quantitative analysis of choline or SM level in wild-type (WT) or BRCA1-mutant human ovarian cancer tissues (BRCA1 WT, n = 5 patients, BRCA1 Mutant, n = 9 patients). The value of n indicates biologically independent samples. Data are presented as mean ± SD (d, g, h) or mean ± SEM (i). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (g, h) or two -tailed unpaired Student’s t-test (d, i). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, e, f). Source data are provided as a Source data file. BRCA1-deficiency drives remodeling of choline metabolism. a Western blot analysis of BRCA1 protein expression in UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells.The samples derive from the same experiment but different gels; all gels were processed in parallel. b KEGG enrichment analysis of changed metabolites showing the significantly affected metabolic pathways. (Variable Importance for the Projection, VIP > 1, p value < 0.05). c Heatmap and cluster analysis of differentiated metabolites between UWB1.289 and UWB1.289 + BRCA1 (UWB-BRCA1) cells. (Variable Importance for the Projection, VIP > 1, p value < 0.05). d Illustration of the choline metabolism pathway and the quantitative results of the changed choline-containing metabolites (n = 5). e Western blot analysis of BRCA1 protein level in A2780 cells transfected with control (ctrl) or BRCA1 siRNA. f Western blot analysis of indicated proteins in parental or BRCA1 knockout (KO) KGN cells. g Quantitative measurement of the intracellular level of choline or sphingomyelin (SM) in A2780 cells transfected with ctrl or BRCA1 siRNA (n = 3). h Measurement of the intracellular level of choline or SM in parental or BRCA1-KO KGN cells (n = 3). i Quantitative analysis of choline or SM level in wild-type (WT) or BRCA1-mutant human ovarian cancer tissues (BRCA1 WT, n = 5 patients, BRCA1 Mutant, n = 9 patients). The value of n indicates biologically independent samples. Data are presented as mean ± SD (d, g, h) or mean ± SEM (i). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (g, h) or two -tailed unpaired Student’s t-test (d, i). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, e, f). Source data are provided as a Source data file. To further validate BRCA1-mediated regulation of choline metabolism, BRCA1 was silenced in A2780, SKOV-3, and CAOV3 cells (Fig. 1e and Supplementary Fig. 1a), and the intracellular levels of choline and its metabolic product SM were measured. Consistent with metabolomics results in UWB1.289 cells, significant increases of choline and SM were detected in BRCA1-silenced A2780 cells, SKOV3, and CAOV3 cells (Fig. 1g and Supplementary Fig. 1b). Similar results were obtained in BRCA1-sgRNA-expressed KGN cells (Fig. 1f, h). The elevation of choline and its downstream metabolite phosphocholine (Pho-Cho) in BRCA1-depleted A2780 and KGN cells were further confirmed by liquid chromatography with tandem mass spectrometry (LC-MS/MS) (Supplementary Fig. 1c, d). In addition, a remarkable increase of choline and SM concentrations was also observed in BRCA1-mutant human ovarian cancer tissues compared with ovarian cancer tissues carrying wild-type (WT) BRCA1 (Fig. 1i). However, we did not observe an obvious change of choline and SM in BRCA2-depleted ovarian cancer cells (Supplementary Fig. 1e). These results demonstrate that BRCA1 regulates choline metabolism in ovarian cancer. BRCA1 suppresses choline uptake via choline transporter-like protein 4 (CTL4) Given that intracellular choline level was reduced by BRCA1 expression (Fig. 1), we thus evaluated whether BRCA1 affects choline uptake through in situ fluorescence labeling of propargyl choline in cells (Supplementary Fig. 2a). Uptake of choline analog propargyl choline can be detected using click chemistry reaction with an azide-modified fluorophore25 (Supplementary Fig. 2a). By this approach, we observed strikingly higher rate of propargyl choline uptake in BRCA1-deficient UWB1.289 cells (Fig. 2a). Consistently, increased propargyl choline uptake was also detected in BRCA1-depleted A2780, SKOV-3 and KGN cells (Fig. 2b and Supplementary Fig. 2b, c), notably, this metabolic alteration occurred without compromising cellular viability and cell cycle progression (Supplementary Fig. 2d and Supplementary Fig. 3a–e), suggesting BRCA1 negatively regulates choline uptake in ovarian cancer cells.Fig. 2BRCA1 suppress choline-uptake via transporter CTL4.Choline uptake in UWB1.289/ UWB1.289 + BRCA1(UWB-BRCA1) cells (a), or control (ctrl)/BRCA1 siRNA-transfected A2780 cells (b) was analyzed as described in “Method” using choline analog propargyl-choline (n = 5). Scale bar, 10 μm. c Volcano scatterplot of differentially expressed genes (DEGs) between UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells (p value < 0.05). d Integrative Genomics Viewer (IGV) tracks of the CTL4 transcript peak by RNA-seq in UWB1.289 (UWB) or UWB1.289 + BRCA1 (UWB-BRCA1) cells. The SLC44A4(CTL4) genomic region is flanked by the NEU1 and EHMT2 genes on the right and left sides, respectively. e Protein levels of choline transporters in indicated BRCA1-proficient or deficient ovarian cancer cells were analyzed by western blot. The samples derive from the same experiment. CTL1, CTL2, CTL3, CHT1, and BRCA1 were each detected on separate gels, and CTL4, together with β-actin, was detected on one gel. All gels were processed in parallel. f Immunochemical (IHC) analysis of CTL4 protein levels in BRCA1-WT or mutant human ovarian cancer tissues (BRCA1 WT, n = 8 patients, BRCA1 Mutant, n = 8 patients). Scale bar, 50 μm. g Western blot analysis of CTL4 protein expression in BRCA1-WT or mutant human ovarian cancer tissues. h Western blot analysis of CTL4 expression in parental or BRCA1-KO KGN cells expressing ctrl/CTL4 shRNA.The samples derive from the same experiment, but different gels for BRCA1 and another for CTL4, together with β-actin. All gels were processed in parallel. i Choline uptake analysis in parental or BRCA1-KO KGN cells expressing ctrl/CTL4 shRNA (n = 5). Scale bar, 10 μm. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, b, f, i). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (b, i) or two -tailed unpaired Student’s t-test (a, f). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, b, e, g–i). Source data are provided as a Source data file. BRCA1 suppress choline-uptake via transporter CTL4. Choline uptake in UWB1.289/ UWB1.289 + BRCA1(UWB-BRCA1) cells (a), or control (ctrl)/BRCA1 siRNA-transfected A2780 cells (b) was analyzed as described in “Method” using choline analog propargyl-choline (n = 5). Scale bar, 10 μm. c Volcano scatterplot of differentially expressed genes (DEGs) between UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells (p value < 0.05). d Integrative Genomics Viewer (IGV) tracks of the CTL4 transcript peak by RNA-seq in UWB1.289 (UWB) or UWB1.289 + BRCA1 (UWB-BRCA1) cells. The SLC44A4(CTL4) genomic region is flanked by the NEU1 and EHMT2 genes on the right and left sides, respectively. e Protein levels of choline transporters in indicated BRCA1-proficient or deficient ovarian cancer cells were analyzed by western blot. The samples derive from the same experiment. CTL1, CTL2, CTL3, CHT1, and BRCA1 were each detected on separate gels, and CTL4, together with β-actin, was detected on one gel. All gels were processed in parallel. f Immunochemical (IHC) analysis of CTL4 protein levels in BRCA1-WT or mutant human ovarian cancer tissues (BRCA1 WT, n = 8 patients, BRCA1 Mutant, n = 8 patients). Scale bar, 50 μm. g Western blot analysis of CTL4 protein expression in BRCA1-WT or mutant human ovarian cancer tissues. h Western blot analysis of CTL4 expression in parental or BRCA1-KO KGN cells expressing ctrl/CTL4 shRNA.The samples derive from the same experiment, but different gels for BRCA1 and another for CTL4, together with β-actin. All gels were processed in parallel. i Choline uptake analysis in parental or BRCA1-KO KGN cells expressing ctrl/CTL4 shRNA (n = 5). Scale bar, 10 μm. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, b, f, i). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (b, i) or two -tailed unpaired Student’s t-test (a, f). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, b, e, g–i). Source data are provided as a Source data file. To explore the molecular mechanism by which BRCA1 regulates choline uptake, we performed RNA-seq to globally analyze BRCA1-regulated gene transcription, 839 downregulated and 967 upregulated genes were identified in UWB1.289 cells compared with UWB-BRCA1 cells (Fig. 2c and Supplementary Data 2). Of those, choline transporter-like protein 4 (CTL4) is one of the most upregulated genes in UWB1.289 cells (Fig. 2c, d). Transportation of extracellular choline into human cells is mainly mediated by high-affinity choline transporter 1 (CHT1) and choline transporter-like proteins (CTL1-4)2. As shown in Fig. 2e, we confirmed that the protein level of CTL4, but not other transporter proteins, was significantly increased in UWB1.289 cells, BRCA1-deficient A2780 or KGN cells. Remarkable increase of CTL4 mRNA was also detected in those BRCA1-deficient ovarian cancer cells (Supplementary Fig. 3f). Immunohistochemical staining revealed that CTL4 protein was predominantly localized to the cytoplasmic and membrane compartments of human ovarian cancer cells (Fig. 2f). In addition, higher levels of CTL4 protein were detected in BRCA1-mutant ovarian cancer tissues compared with BRCA1-WT cancer tissues (Fig. 2f, g). To test whether the enhanced choline-uptake in BRCA1-deficient cells resulted from elevation of CTL4, we knocked down CTL4 expression in KGN cells expressing BRCA1 sgRNA (Fig. 2h). The uptake assay revealed that silence of CTL4 reversed BRCA1-depletion induced increase of choline uptake without affecting cell viability in BRCA1-depleted cells (Fig. 2i and Supplementary Fig. 3g, h), demonstrating that BRCA1 regulates choline uptake through transporter protein CTL4. BRCA1 associates with the CTL4 promoter and regulates its expression BRCA1 could bind with DNA and directly regulate gene transcription16,17. We thus conducted chromatin immunoprecipitation sequencing (ChIP-seq) in A2780 cells to further identify the BRCA1-targeted genes. ChIP-seq profiling identified a total of 10921 BRCA1-binding peaks, approximately 24% of them located within gene promoter regions (Fig. 3a). Integrative analysis of RNA-seq and ChIP-seq data revealed that 110 genes were potential direct targets of BRCA1 (Fig. 3b), and the SLC44A4(CTL4) gene was one of them. Notably, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed they are enriched pathways including metabolic pathway (Fig. 3c). Peak analysis revealed that BRCA1 binding motif mainly located at 0 and −1210 bp upstream from CTL4 gene transcription start site (TSS) (Fig. 3d and Supplementary Fig. 4a). Meanwhile, we failed to detect BRCA1 binding to CTL1-3 and CHT1 genes (Supplementary Fig. 4b). To further verify the binding of BRCA1 to CTL4 promoter, we performed ChIP-PCR using a primer pair located between 78 and 538 bp upstream of the TSSs. As shown in Fig. 3e, f, the association of BRCA1 with CTL4 promoter was confirmed by ChIP-PCR against either endogenous or exogenously expressed BRCA1. Importantly, analysis of cancer genome atlas (TCGA) datasets confirmed that CTL4 mRNA levels were significantly higher in BRCA1-mutant ovarian cancer tissues compared with BRCA1-unaltered ovarian cancers (Fig. 3g).Fig. 3BRCA1 associates with the CTL4 promoter and negatively regulates CLT4 expression.a Genomic distribution of BRCA1-binding peaks in A2780 cells. b Venn analysis of the overlaps of DEGs (UWB1.289 vs. UWB-BRCA1) and BRCA1-targeting genes obtained from the ChIP result. c Pathway enrichment analysis of 110 DEGs. d The BRCA1-binding peaks in the SLC44A4 (CTL4) gene by ChIP-seq were shown. ChIP-PCR analysis of BRCA1 occupancy at CTL4 promoter (CTL4-p) in A2780 cells using antibody against endogenous BRCA1 (e) or exogenously expressed flag-BRCA1 (f) (n = 3). g CTL4 mRNA levels in BRCA1-mutant or unaltered ovarian cancer tissues were analyzed from the TCGA ovarian cancer dataset. (BRCA1 Mutations, n = 53 patients, max = 7.385, min = 0.1097, median = 4.022, 25% percentile = 1.755, 75% percentile = 5.642; BRCA1 unaltered, n = 938 patients, max = 8.892, min = 0, median = 1.202, 25% percentile = 0.292, 75% percentile = 4.930;). The value of n indicates biologically independent samples. Data are presented as mean ± SD (e, f). Statistical significance was assessed by a two-tailed unpaired Student’s t-test (e–g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (e, f). Source data are provided as a Source data file. BRCA1 associates with the CTL4 promoter and negatively regulates CLT4 expression. a Genomic distribution of BRCA1-binding peaks in A2780 cells. b Venn analysis of the overlaps of DEGs (UWB1.289 vs. UWB-BRCA1) and BRCA1-targeting genes obtained from the ChIP result. c Pathway enrichment analysis of 110 DEGs. d The BRCA1-binding peaks in the SLC44A4 (CTL4) gene by ChIP-seq were shown. ChIP-PCR analysis of BRCA1 occupancy at CTL4 promoter (CTL4-p) in A2780 cells using antibody against endogenous BRCA1 (e) or exogenously expressed flag-BRCA1 (f) (n = 3). g CTL4 mRNA levels in BRCA1-mutant or unaltered ovarian cancer tissues were analyzed from the TCGA ovarian cancer dataset. (BRCA1 Mutations, n = 53 patients, max = 7.385, min = 0.1097, median = 4.022, 25% percentile = 1.755, 75% percentile = 5.642; BRCA1 unaltered, n = 938 patients, max = 8.892, min = 0, median = 1.202, 25% percentile = 0.292, 75% percentile = 4.930;). The value of n indicates biologically independent samples. Data are presented as mean ± SD (e, f). Statistical significance was assessed by a two-tailed unpaired Student’s t-test (e–g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (e, f). Source data are provided as a Source data file. BRCA1 orchestrates EZH2-mediated H3K27Me3 deposition to CTL4 promoter It has been reported that BRCA1 can either stimulate or repress gene transcription16,17. To determine whether BRCA1 directly promotes CTL4 transcription, we transiently transfected a CTL4 promoter-driven luciferase reporter into UWB1.289/UWB-BRCA1 cells and KGN cells expressing control/BRCA1 sgRNA; however, we failed to observe an obvious difference in luciferase activity between BRCA1-proficient and deficient cells (Supplementary Fig. 4c), suggesting BRCA1 might not directly regulate CTL4 transcription. Then, we tested whether BRCA1 epigenetically affects CTL4 promoter activity, and the common epigenetic histone modifications, including H3K27Me3, H3K27Ac, H3K9Me3, H3K9Ac, H3K4Me3, were examined in UWB1.289 and UWB-BRCA1 cells. As shown in Fig. 4a, a decrease of H3K27Me3 and an increase of H3K27Ac were detected in BRCA1-deficient UWB1.289 cells. Similar results were also observed in KGN and Caov-3 cells expressing BRCA1 sgRNA (Fig. 4b and Supplementary Fig. 4d). H3K27Ac is well-defined as an active enhancer/promoter marker and is involved in activation of gene transcription26, while H3K27Me3 is associated with gene silencing27. To test whether these alterations of histone modifications could affect CTL4 expression, UWB1.289 and UWB-BRCA1 cells were treated with increasing concentrations of JQ1 (a selective inhibitor of BET bromodomain proteins that read H3K27Ac)28, tazemetostat (a histone methyltransferase inhibitor)29, or decitabine (a DNA methyltransferase inhibitor)30. Interestingly, we found that treatment of tazemetostat, but not JO1 or decitabine, reversed BRCA1-induced CTL4 downregulation (Fig. 4c). Tazemetostat treatment significantly increased CTL4 mRNA in UWB-BRCA1 cells, but not in BRCA1-deficient UWB1.289 cells (Supplementary Fig. 4e). ChIP-PCR analysis revealed that H3K27Me3 is mainly enriched at 0 to −200 bp upstream of CTL4 TSS in A2780 and UWB-BRCA1 cells (Supplementary Fig. 4f). These results suggest that BRCA1 represses CTL4 transcription by modulating H3K27me3 levels at its promoter.Fig. 4BRCA1 recruits EZH2 to the CTL4 promoter and promotes its H3K27Me3.ChIP-PCR analysis of indicated epigenetic histone modifications on CTL4 promoter in UWB1.289/UWB-BRCA1 cells (a) or KGN cells expressing control (ctrl) or BRCA1 sgRNA (sg)(b). The conventional PCR amplification (left) and its relative quantifications were determined (right) (n = 3). c UWB1.289 or UWB-BRCA1 cells were treated with increasing concentrations of JQ1, Tazemetostat, or Decitabine for 24 h, followed by western blot analysis of CTL4 protein level. d A2780 cells were transfected with the indicated plasmids, followed by immunoprecipitation (IP) analysis of EZH2/BRCA1 interaction. The samples derive from the same experiment but different gels: one for Myc-BRCA1, and another for GFP-EZH2. All gels were processed in parallel. e Co-IP analysis of endogenous EZH2/BRCA1 interaction in A2780 cells using anti-BRCA1 antibody. ChIP-PCR analysis of EZH2 occupancy at CTL4 promoter region in KGN cells expressing ctrl/BRCA1 sgRNA (f) or UWB1.289/UWB-BRCA1 cells (g), (n = 3). Ctrl or EZH2 shRNA were stably transfected into parental or BRCA1-KO KGN cells, followed by analysis of H3K27Me3 level in CTL4-promoter by ChIP-PCR (h), (n = 3), CTL4 protein level by western blot (i), and choline uptake (j), (n = 3). For the Western blot, the samples derive from the same experiment but from different gels: one for BRCA1 and EZH2, and another for CTL4 and β-actin. All gels were processed in parallel. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, b, f–h, j). Statistical significance was assessed by two-way ANOVA with Sidak’s multiple comparisons test (a, b, h, j) or ordinary one-way ANOVA with Tukey’s multiple comparisons test (f, g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a–i). Source data are provided as a Source data file. BRCA1 recruits EZH2 to the CTL4 promoter and promotes its H3K27Me3. ChIP-PCR analysis of indicated epigenetic histone modifications on CTL4 promoter in UWB1.289/UWB-BRCA1 cells (a) or KGN cells expressing control (ctrl) or BRCA1 sgRNA (sg)(b). The conventional PCR amplification (left) and its relative quantifications were determined (right) (n = 3). c UWB1.289 or UWB-BRCA1 cells were treated with increasing concentrations of JQ1, Tazemetostat, or Decitabine for 24 h, followed by western blot analysis of CTL4 protein level. d A2780 cells were transfected with the indicated plasmids, followed by immunoprecipitation (IP) analysis of EZH2/BRCA1 interaction. The samples derive from the same experiment but different gels: one for Myc-BRCA1, and another for GFP-EZH2. All gels were processed in parallel. e Co-IP analysis of endogenous EZH2/BRCA1 interaction in A2780 cells using anti-BRCA1 antibody. ChIP-PCR analysis of EZH2 occupancy at CTL4 promoter region in KGN cells expressing ctrl/BRCA1 sgRNA (f) or UWB1.289/UWB-BRCA1 cells (g), (n = 3). Ctrl or EZH2 shRNA were stably transfected into parental or BRCA1-KO KGN cells, followed by analysis of H3K27Me3 level in CTL4-promoter by ChIP-PCR (h), (n = 3), CTL4 protein level by western blot (i), and choline uptake (j), (n = 3). For the Western blot, the samples derive from the same experiment but from different gels: one for BRCA1 and EZH2, and another for CTL4 and β-actin. All gels were processed in parallel. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, b, f–h, j). Statistical significance was assessed by two-way ANOVA with Sidak’s multiple comparisons test (a, b, h, j) or ordinary one-way ANOVA with Tukey’s multiple comparisons test (f, g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a–i). Source data are provided as a Source data file. H3K27me3 histone modification is mainly catalyzed by histone methyltransferase EZH231, and BRCA1 could interact with EZH2 in ovarian cancer cells in both endogenous context and upon exogenous expression (Fig. 4d, e). Using myc-BRCA1 truncation mutants, we further identified EZH2 binding to the N-terminal aa.303-772 region of BRCA1 (Supplementary Fig. 4g). Meanwhile, aa.340-559 region of EZH2 is required for this interaction (Supplementary Fig. 4h). which is consistent with previous finding that BRCA1 directly binds at aa.341-559 region of EZH2 and disrupts EZH2 association with noncoding RNA (ncRNA) HOTAIR32. Thus, we hypothesized that BRCA1 might recruit EZH2 and deposit H3K27me3 at the CTL4 promoter. To test our hypothesis, we evaluated the occupancy of EZH2 at the CTL4 promoter in BRCA1 sgRNA-expressed KGN cells or UWB1.289/UWB-BRCA1 cells. As shown in Fig. 4f, g, depletion of BRCA1 resulted in a reduction of EZH2 occupancy at CTL4 promoter, while expression of BRCA1 significantly increased recruitment of EZH2 at CTL4 promoter. Then, we sought to determine whether BRCA1-deficiency caused a reduction of H3K27Me3 in CTL4 promoter attributes to a decrease of EZH2 abundance. As expected, depletion of EZH2 abolished the distinction of H3K27Me3 level (Fig. 4h), as well as CTL4 protein level (Fig. 4i), in BRCA1-depleted KGN cells. Consequently, the silence of EZH2 reversed the BRCA1-mediated suppression of choline uptake in KGN cells (Fig. 4j). CTL4-mediated choline transport confers metastasis potential to BRCA1-deficient cancer cells To explore the physiological consequences of elevated choline metabolism in BRCA1-deficient cancer cells, we first assessed the growth capacity of UWB1.289 and UWB-BRCA1 cells. Interestingly, no significant enhancement in clonogenic formation was observed in BRCA1-deficient UWB1.289 cells compared to UWB-BRCA1 cells (Supplementary Fig. 5a). Additionally, treatment with the choline uptake inhibitor hemicholinium-3 (HC-3) induced comparable levels of growth inhibition in both UWB1.289 and UWB-BRCA1 cells (Supplementary Fig. 5b). Interestingly, BRCA1-expression significantly inhibited cell invasion in UWB1.289 cells (Fig. 5a). Whereas, BRCA1-depletion remarkably increased cell invasion in A2780 or KGN cells (Supplementary Fig. 5c, d). Consistent with the enhanced invasive capacity, BRCA1-deficient ovarian cancer cells expressed higher levels of mesenchymal markers N-cadherin and Vimentin, and reduced levels of epithelial marker E-cadherin33 (Supplementary Fig. 5e–g). Meanwhile, a higher occurrence of metastasis was observed in BRCA1-mutant ovarian cancer patients compared with patients carrying wild-type (WT) BRCA1 (Fig. 5b, Supplementary Data 5). These results demonstrate loss function of BRCA1 promotes cancer metastasis.Fig. 5CTL4-mediated choline uptake promotes metastasis in BRCA1-deficient ovarian cancer cells.a Matrigel-coated transwell analysis of invasion capacity in UWB1.289 and UWB-BRCA1 cells. Scale bar, 100 μm, (n = 3). b Peritoneal (upper left) and distal (lower left) metastasis in ovarian cancer patients carrying wild-type (WT) or mutant BRCA1 was shown by magnetic resonance (MR) and computed tomography (CT) imaging, respectively. Metastatic nodules were indicated by an arrow. The metastatic frequency was quantified (Right), (BRCA1 WT, n = 77 patients, metastasis = 34, no matastasis = 43; BRCA1 Mutant, n = 33 patients, metastasis = 25, no matastasis = 8). UWB-BRCA1 or UWB1.289 cells stably expressing ctrl or CTL4 shRNA, followed by transwell analysis of cell invasion (c) (n = 4) (Scale bar, 100 μm) or western blot analysis of mesenchymal marker vimentin and N-cadherin (d). The samples derive from the same experiment but from different gels: one for Vimentin and N-cadherin, a second for E-cadherin and β-actin, and a third for CTL4 and BRCA1. All gels were processed in parallel. e Ctrl or CTL4 shRNA were stably transfected into Ctrl or BRCA1-KO KGN cells, and cell migration was analyzed by microspheroid. Scale bar, 200 μm, (n = 3, Representative images from three independent experiments with similar results.) Ctrl or CTL4 shRNA were stably transfected into KGN cells with Ctrl or BRCA1-KO KGN cells, cells were orthotopically implanted into mice ovarian. The metastatic nodules(arrow) in the liver (f) (n = 5 mice each group) and pathological liver sections (g) (n = 3) were examined. Scale bar, 100 μm. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, c, e, f, g). Statistical significance was assessed by two -tailed unpaired Student’s t-test (a), Chi-square test (b), or ordinary one-way ANOVA with Tukey’s multiple comparisons test (c, e, f, g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (d). Source data are provided as a Source data file. CTL4-mediated choline uptake promotes metastasis in BRCA1-deficient ovarian cancer cells. a Matrigel-coated transwell analysis of invasion capacity in UWB1.289 and UWB-BRCA1 cells. Scale bar, 100 μm, (n = 3). b Peritoneal (upper left) and distal (lower left) metastasis in ovarian cancer patients carrying wild-type (WT) or mutant BRCA1 was shown by magnetic resonance (MR) and computed tomography (CT) imaging, respectively. Metastatic nodules were indicated by an arrow. The metastatic frequency was quantified (Right), (BRCA1 WT, n = 77 patients, metastasis = 34, no matastasis = 43; BRCA1 Mutant, n = 33 patients, metastasis = 25, no matastasis = 8). UWB-BRCA1 or UWB1.289 cells stably expressing ctrl or CTL4 shRNA, followed by transwell analysis of cell invasion (c) (n = 4) (Scale bar, 100 μm) or western blot analysis of mesenchymal marker vimentin and N-cadherin (d). The samples derive from the same experiment but from different gels: one for Vimentin and N-cadherin, a second for E-cadherin and β-actin, and a third for CTL4 and BRCA1. All gels were processed in parallel. e Ctrl or CTL4 shRNA were stably transfected into Ctrl or BRCA1-KO KGN cells, and cell migration was analyzed by microspheroid. Scale bar, 200 μm, (n = 3, Representative images from three independent experiments with similar results.) Ctrl or CTL4 shRNA were stably transfected into KGN cells with Ctrl or BRCA1-KO KGN cells, cells were orthotopically implanted into mice ovarian. The metastatic nodules(arrow) in the liver (f) (n = 5 mice each group) and pathological liver sections (g) (n = 3) were examined. Scale bar, 100 μm. The value of n indicates biologically independent samples. Data are presented as mean ± SD (a, c, e, f, g). Statistical significance was assessed by two -tailed unpaired Student’s t-test (a), Chi-square test (b), or ordinary one-way ANOVA with Tukey’s multiple comparisons test (c, e, f, g). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (d). Source data are provided as a Source data file. Since choline withdrawal significantly impaired cell invasion in KGN and A2780 cells (Supplementary Fig. 6a, b). We then tested whether the increased invasion in BRCA1-deficient cells results from enhanced choline metabolism. Intriguingly, we found that choline withdrawal abrogated BRCA1 deficiency-mediated increase of invasion in UWB1.289 cells (Supplementary Fig. 6c). Meanwhile, silence of CTL4 could also abolish increased cell invasion in UWB1.289 or BRCA1-depleted KGN and CAOV3 cells (Fig. 5c–e and Supplementary Fig. 6d, e). To further validate whether CTL4-mediated choline uptake is required for the high metastatic ability in BRCA1-deficient cells, a KGN and syngeneic ID8Trp53−/− HGOSC orthotopic ovarian cancer model that has metastasis potential was utilized34. Consistent with the in vitro transwell assay, silence of CTL4 completely reversed increased liver and gut metastasis in BRCA1-depleted KGN orthotopic ovarian model (Fig. 5f, g and Supplementary Fig. 6f). Similarly, CTL4 knockdown significantly reduced ascites accumulation and metastatic nodules in the peritoneum, mesentery, and omentum in the BRCA1-depleted ID8 Trp53−/− orthotopic model (Supplementary Fig. 7a, b and Supplementary Fig. 7d). Meanwhile, CTL4 knockdown mildly suppressed orthotopic tumor growth (Supplementary Fig. 6g and Supplementary Fig. 7c). Choline metabolite phosphocholine (Pho-Cho) promotes metastasis via activating FAM3C Next, we sought to dissect the underlying mechanism that choline metabolism promotes cancer metastasis. Choline is metabolized into betaine, acetylcholine, and phosphocholine (Pho-Cho) after uptake by cells2. We observed significant changes in Pho-Cho, acetylcholine (ACH), but not betaine levels in UWB1.289 cells following BRCA1 expression using a metabolomics assay (Supplementary Data 1). Exogenous addition of excessive acetylcholine failed to promote invasion in A2780 and KGN cells (Supplementary Fig. 8a). Treatment of acetylcholine inhibitor vesamicol did not also reserve the enhanced invasion in BRCA1-deficient KGN cells (Supplementary Fig. 8b). Additionally, we did not detect any changed in phosphorylation signaling between UWB1.289 cells and UWB-BRCA1 cells (Supplementary Fig. 8c). Interestingly, silencing the choline metabolic enzymes choline kinase alpha (CHKA) and sphingomyelin synthase 2 (SGMS2), but not phosphate cytidylyltransferase 1B (Pcyt1B), Choline Phosphotransferase 1 (CHPT1), or sphingomyelin synthase 1 (SGMS1), significantly impaired invasive ability of KGN cells (Supplementary Fig. 9a). However, choline withdrawal further inhibited invasion in SGMS2-depleted cells, but not in CHKA-depleted cells (Supplementary Fig. 9b), suggesting that CHKA-mediated regulation of cancer metastasis depends on choline metabolism. Additionally, CHKA depletion reversed the hyper-invasion mediated by BRCA1 deficiency in UWB1.289 cells (Supplementary Fig. 9c). CHKA catalyzes the conversion of choline to Pho-Cho, and exogenous supplementation of Pho-Cho enhanced invasion in BRCA1-proficient KGN cells and UWB-BRCA1 cells (Supplementary Fig. 9d, e). Notably, Pho-Cho supplementation rescued CHKA depletion- caused invasion impairment and CTL4 depletion-induced invasion suppression in BRCA1-deficient UWB1.289 cells (Supplementary Fig. 9f, g), indicating CHKA-catalyzed Pho-Cho promotes cancer metastasis. To investigate how Pho-Cho promotes cancer metastasis, the Pho-Cho-associated protein complex was purified and analyzed using mass spectrometry, revealing that the pro-metastatic protein FAM3C is bound with Pho-Cho (Fig. 6a, b). FAM3C, also named the interleukin-like EMT inducer (ILEI), is a secreting protein that promotes cancer metastasis35. Silencing of FAM3C diminished enhanced invasion in UWB1.289 cells (Supplementary Fig. 10a). Molecular docking analysis revealed that Pho-Cho strongly binds with human FAM3C; its predicted binding score (−6.30 Kcal/mol) is comparable to the known Pho-Cho receptor Leukocidin F(−5.97 Kcal/mol) and Cytolysin Sticholysin II (−5.73 Kcal/mol) (Fig. 6c and Supplementary Fig. 10b)36. Docking simulation shows that Pho-Cho interacts with Lys225, Gly187, Cys185, and Ser194 of FAM3C protein through hydrogen bond and salt bridge (Fig. 6c). The cellular thermal shift assay (CETSA) confirmed the binding of Pho-Cho to wild-type (WT) FAM3C, but not to the K225A or C185A mutant-FAM3C (Fig. 6d and Supplementary Fig. 10c).Fig. 6Phosphocholine (Pho-Cho) binds and stabilizes FAM3C.A2780 cell lysates were incubated with Pho-Cho conjugated agarose beads or control beads, followed by Coomassie blue staining of the proteins associated with the beads (a) and western blot analysis of FAM3C/Pho-Cho interaction (b). c Molecular docking analysis shows the key amino acids on FAM3C that interact with Pho-Cho. d CETSA analysis of Pho-Cho/FAM3C association. The representative western blot of intracellular FAM3C and β-Actin level (upper) and quantification of FAM3C protein (lower) in the presence of DMSO or Pho-Cho at increasing temperature (n = 2, Representative images from three independent experiments with similar results.). e Western blot analysis of FAM3C protein level in UWB1.289 and UWB1.289 + BRCA1 (UWB-BRCA1) cells treated with or without 20 μM MG132 for 6 h. f Immunofluorescence (IF) analysis of FAM3C and GM130 protein in UWB1.289 and UWB-BRCA1 cells. Scale bar, 10 μm. g IF analysis of FAM3C and GM130 in UWB1.289 cells with or without choline withdrawal for 48 h. Scale bar, 10 μm. h UWB1.289/UWB-BRCA1 cells were co-transfected with flag-FAM3C and HA-Ubiquitin (Ub) plasmids, followed by analysis of flag-FAM3C polyubiquitination after treatment with 20 μM MG132 for 6 h. i UWB1.289 + BRCA1(UWB-BRCA1) cells were transfected with flag-FAM3C or HA-Ub plasmid as indicated. Twenty-four hours after transfection, cells were treated with or without 20 μM Pho-Cho by hypotonic shift as described in “Method,” followed by analysis of flag-FAM3C polyubiquitination. For (h, i), the samples derive from the same experiment but from different gels: one for HA-Ub, a second for Flag-FAM3C, and a third for BRCA1. All gels were processed in parallel. j IHC analysis of FAM3C levels in BRCA1 WT or mutant ovarian cancer tissues. Scale bar, 50 μm. (BRCA1 WT, n = 14 patients, BRCA1 Mutant, n = 14 patients). The value of n indicates biologically independent samples. Data are presented as mean ± SD (d, j). Statistical significance was assessed by a two-tailed unpaired Student’s t-test (j). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, b, e–i). Source data are provided as a Source data file. Phosphocholine (Pho-Cho) binds and stabilizes FAM3C. A2780 cell lysates were incubated with Pho-Cho conjugated agarose beads or control beads, followed by Coomassie blue staining of the proteins associated with the beads (a) and western blot analysis of FAM3C/Pho-Cho interaction (b). c Molecular docking analysis shows the key amino acids on FAM3C that interact with Pho-Cho. d CETSA analysis of Pho-Cho/FAM3C association. The representative western blot of intracellular FAM3C and β-Actin level (upper) and quantification of FAM3C protein (lower) in the presence of DMSO or Pho-Cho at increasing temperature (n = 2, Representative images from three independent experiments with similar results.). e Western blot analysis of FAM3C protein level in UWB1.289 and UWB1.289 + BRCA1 (UWB-BRCA1) cells treated with or without 20 μM MG132 for 6 h. f Immunofluorescence (IF) analysis of FAM3C and GM130 protein in UWB1.289 and UWB-BRCA1 cells. Scale bar, 10 μm. g IF analysis of FAM3C and GM130 in UWB1.289 cells with or without choline withdrawal for 48 h. Scale bar, 10 μm. h UWB1.289/UWB-BRCA1 cells were co-transfected with flag-FAM3C and HA-Ubiquitin (Ub) plasmids, followed by analysis of flag-FAM3C polyubiquitination after treatment with 20 μM MG132 for 6 h. i UWB1.289 + BRCA1(UWB-BRCA1) cells were transfected with flag-FAM3C or HA-Ub plasmid as indicated. Twenty-four hours after transfection, cells were treated with or without 20 μM Pho-Cho by hypotonic shift as described in “Method,” followed by analysis of flag-FAM3C polyubiquitination. For (h, i), the samples derive from the same experiment but from different gels: one for HA-Ub, a second for Flag-FAM3C, and a third for BRCA1. All gels were processed in parallel. j IHC analysis of FAM3C levels in BRCA1 WT or mutant ovarian cancer tissues. Scale bar, 50 μm. (BRCA1 WT, n = 14 patients, BRCA1 Mutant, n = 14 patients). The value of n indicates biologically independent samples. Data are presented as mean ± SD (d, j). Statistical significance was assessed by a two-tailed unpaired Student’s t-test (j). Confidence intervals are available in the Source data file. Representative images from three independent experiments with similar results (a, b, e–i). Source data are provided as a Source data file. Then, we aimed to explore how BRCA1-mediated regulation of Pho-Cho influences FAM3C. We detected significantly higher levels of FAM3C protein in BRCA1-deficient UWB1.289 cells compared to UWB1.289 + BRCA1(UWB-BRCA1) cells, while proteasome inhibitor MG132 treatment restored FAM3C protein levels in UWB1.289 + BRCA1(UWB-BRCA1) cells (Fig. 6e). FAM3C is processed in the Golgi apparatus before secretion. GM130 is a peripheral cytoplasmic protein strongly bound to the Golgi membrane and is recognized as a Golgi marker. To further validate BRCA1-mediated regulation of FAM3C protein expression, FAM3C and GM130 were co-stained in UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells. Consistently, BRCA1 expression significantly reduced FAM3C levels in UWB1.289 cells and induced a high degree of co-localization between FAM3C and GM130 (Fig. 6f), which indicates BRCA1-mediated downregulation of FAM3C results in its retention within the Golgi apparatus in UWB1.289 + BRCA1 (UWB-BRCA1) cells. Similarly, Choline withdrawal also remarkably reduced FAM3C expression in UWB1.289 cells (Fig. 6g). Pho-Cho replenishment also significantly increased FAM3C protein in UWB1.289 + BRCA1 (UWB-BRCA1) cells (Supplementary Fig. 10d). Additionally, exogenous addition of Pho-Cho raised FAM3C protein level, but not its mRNA level in KGN and A2780 cells (Supplementary Fig. 10e). We also failed to detect significant alteration in FAM3C mRNA between UWB1.289 and UWB1.289 + BRCA1(UWB-BRCA1) cells (Supplementary Fig. 10f), and Pho-Cho supplementation remarkably attenuated FAM3C degradation in KGN cells (Supplementary Fig. 10g). Additionally, depletion of CTL4 significantly downregulated FAM3C protein, but not its mRNA in UWB1.289 cells (Supplementary Fig. 10h). These results indicate that CTL4 mediated Pho-Cho production regulates FAM3C by enhancing its protein stability. As expected, a lower level of FAM3C polyubiquitination was observed in UWB1.289 cells compared with UWB1.289 + BRCA1(UWB-BRCA1) cells (Fig. 6h). Silence of CTL4 or CHKA dramatically increased FAM3C polyubiquitination in UWB1.289 cells, suggesting decreased FAM3C ubiquitination in UWB1.289 cells results from CTL4-mediated choline metabolism (Supplementary Fig. 11a, b). K97, K112, K156 and K225 were predicted as the potential ubiquitination sites on FAM3C (Supplementary Fig. 11c), and intracellular ubiquitination assay revealed that polyubiquitination of FAM3C mainly occurred on K112 and K225 (Supplementary Fig. 11d). Moreover, Pho-Cho supplementation significantly reduced polyubiquitination on WT FAM3C, but not on K225R mutant FAM3C in UWB1.289 + BRCA1 cells (Fig. 6i and Supplementary Fig. 11e). This finding aligns with the observation that Pho-Cho interact with and forms three hydrogen bond with K225, indicating that Pho-Cho’s association with FAM3C could disrupt the polyubiquitination linkage at K225. Consistent with the cellular observations, a significantly higher level of FAM3C was detected in BRCA1-mutant ovarian cancer tissues compared to ovarian cancer tissues with WT BRCA1 (Fig. 6j). Natural compound DT-13 targets CTL4 and inhibits metastasis in BRCA1-deficient ovarian cancer Next, we aimed to evaluate whether targeting CTL4-mediated choline uptake could effectively suppress metastasis in BRCA1-mutant ovarian cancer. Taking advantage of fluorescence-based propargyl choline uptake assay (Supplementary Fig. 2a), a library containing 1680 natural compounds was utilized to screen potential CTL4 inhibitor. Dozens of compounds were found to significantly reduce choline uptake, and 27 compounds were selected (Z-score < −2) for dose-response validation. Among them, 10 compounds (red dots) including Ginsenoside C-K(GC), Ginkgolic C15:1(GA), Valepotriate (Va), DT-13, Tenofovir alafenamide (TA), 6-Shogaol (Sh), Anacardic acid (AA), Proadifen hydrochloride (PH), Guanine-2′-deoxyriboside (DM), Laurocapram (La) exhibited dose-dependent inhibition of radioactive-based 99mTc-HYNIC-propargylcholine choline uptake (Fig. 7a, b, Supplementary Fig. 12a). Their does-response effects on choline uptake were also evaluated using fluorescence-based propargyl choline incorporation and two compounds (TA, DM) failed to show dose-dependent inhibition in this assay (Supplementary Fig. 12b). Subsequent microscale thermophoresis (MST) binding assay revealed that 2 compounds including DT-13 (Kd = 0.64 ± 0.356 μM) and Valepotriate (Kd = 15.62 ± 3.954 μM) could directly bind to CTL4 (Supplementary Fig. 12c).Fig. 7Discovery of natural compound DT-13 as CTL4 inhibitor and potently suppresses metastasis in BRCA1-deficient ovarian cancer cells.a Screening the potential CTL4 inhibitors from a natural compound library through in situ fluorescence labeling of propargyl choline using KGN cells with CTL4 overexpression. Hemicholinium-3 (HC-3) and CTL4 siRNA were used as a negative control. b The inhibition curve of DT-13 on choline uptake in CTL4 overexpressed KGN cells (n = 3). c Chemical structure of DT-13. d Microscale thermophoresis (MST) analysis of DT-13 binding with CTL1-4 proteins (n = 3). e CETSA analysis of DT-13 association with full-length (FL) or indicated deletion mutant CTL4 proteins. f Molecular modeling of DT-13 in the binding pocket of CTL4. KGN cells were treated with increasing concentrations of DT-13 for 24 h, and the intracellular choline (g) and sphingomyelin (SM) (h) levels were measured and quantified (n = 3). i BRCA1-knockout (KO) KGN cells were treated with indicated concentrations of DT-13 for 24 h, followed by transwell analysis of cell invasion (n = 3). Scale bar, 100 μm. j BRCA1-KO KGN microspheroid were grown in the presence of indicated concentrations of DT-13, followed by analysis of cell invasion. Scale bar, 200 μm (n = 3, Representative images from three independent experiments with similar results.). k, l BRCA1-KO KGN cells were orthotopically implanted into mice ovarian, followed by treatment with DT-13, and the metastatic nodules in the liver were examined. The metastatic nodules(arrow) in the liver (k) (n = 5 mice each group) and pathological liver sections (l) (n = 3) were examined. Scale bar, 100 μm. The value of n indicates biologically independent samples (b, d, e, g–i, k, l). Representative images from three independent experiments with similar results (e). Data are presented as mean ± SD (b, d, e, g–l). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (g–l). Confidence intervals are available in the Source data file. Source data are provided as a Source data file. Discovery of natural compound DT-13 as CTL4 inhibitor and potently suppresses metastasis in BRCA1-deficient ovarian cancer cells. a Screening the potential CTL4 inhibitors from a natural compound library through in situ fluorescence labeling of propargyl choline using KGN cells with CTL4 overexpression. Hemicholinium-3 (HC-3) and CTL4 siRNA were used as a negative control. b The inhibition curve of DT-13 on choline uptake in CTL4 overexpressed KGN cells (n = 3). c Chemical structure of DT-13. d Microscale thermophoresis (MST) analysis of DT-13 binding with CTL1-4 proteins (n = 3). e CETSA analysis of DT-13 association with full-length (FL) or indicated deletion mutant CTL4 proteins. f Molecular modeling of DT-13 in the binding pocket of CTL4. KGN cells were treated with increasing concentrations of DT-13 for 24 h, and the intracellular choline (g) and sphingomyelin (SM) (h) levels were measured and quantified (n = 3). i BRCA1-knockout (KO) KGN cells were treated with indicated concentrations of DT-13 for 24 h, followed by transwell analysis of cell invasion (n = 3). Scale bar, 100 μm. j BRCA1-KO KGN microspheroid were grown in the presence of indicated concentrations of DT-13, followed by analysis of cell invasion. Scale bar, 200 μm (n = 3, Representative images from three independent experiments with similar results.). k, l BRCA1-KO KGN cells were orthotopically implanted into mice ovarian, followed by treatment with DT-13, and the metastatic nodules in the liver were examined. The metastatic nodules(arrow) in the liver (k) (n = 5 mice each group) and pathological liver sections (l) (n = 3) were examined. Scale bar, 100 μm. The value of n indicates biologically independent samples (b, d, e, g–i, k, l). Representative images from three independent experiments with similar results (e). Data are presented as mean ± SD (b, d, e, g–l). Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test (g–l). Confidence intervals are available in the Source data file. Source data are provided as a Source data file. DT-13 is a steroidal saponin isolated from Liriope muscari L. H. Bailey37. MST assay further showed that DT-13 binds to CTL4 with high affinity and weakly interacts with CTL2 (Fig. 7c, d and supplementary Fig. 13a). Two potential binding sites for DT-13 on the CTL4 protein were predicted, primarily comprising amino acids 550–661 and 365–457, respectively. (Supplementary Fig. 13b, c). The CETSA assay showed that DT-13 interacts with the full-length (FL) CTL4 protein and with CTL4 lacking amino acids 357–371 or 453–499, but not with CTL4 lacking amino acids 550–603 or 649–661 (Fig. 7e), indicating that DT-13 binds to a pocket within the CTL4 protein that comprises amino acids 550–661 (Fig. 7f). This pocket is also predicted to associate with choline during choline transport (Supplementary Fig. 13d), explaining the mechanism by which DT-13 inhibits CTL4-mediated choline uptake. Next, we aimed to evaluate whether the CTL4 inhibitor DT-13 could suppress choline metabolism in BRCA1-deficient ovarian cancer cells. As expected, DT-13 reduced choline and SM levels in UWB1.289 cells in a dose-dependent manner (Fig. 7g, h). Consequently, DT-13 exhibited significantly greater anti-invasive potency in BRCA1-KO KGN cells compared to Valepotriate (Fig. 7i, j and Supplementary Fig. 13e). Then, an orthotopic and metastatic cancer model using BRCA1-KO KGN cells was established to evaluate the therapeutic efficacy of DT-13 against metastasis in BRCA1-deficient ovarian cancers. Consistently, DT-13 treatment significantly reduced the number of liver and gut metastatic nodules originating from BRCA1-KO orthotopic tumors (Fig. 7k, l and Supplementary Fig. 14a, b). Moreover, DT-13 treatment at this effective dose did not cause notable body weight loss (Supplementary Fig. 14c), indicating that DT-13 is an effective and safe candidate for treating metastatic ovarian cancers with BRCA1 mutation. Discussion BRCA1 (Breast Cancer gene 1) has long been recognized for its pivotal role in maintaining genome stability through its involvement in homologous recombination-mediated DNA repair and DNA replication fork stabilization14. Mutations in the BRCA1 gene predispose individuals to hereditary breast and ovarian cancers, underscoring its importance in tumor suppression38. However, emerging evidences suggest that BRCA1 functions extend beyond its canonical roles in genome maintenance, contributing to tumor suppression through diverse mechanisms, including cell cycle progression, ferroptosis, mitochondrial function, and redox homeostasis13,39,40. Our present data showed that BRCA1 negatively regulates choline metabolism through suppressing choline uptake. BRCA1 deficiency leads to an increase in choline uptake, resulting in the elevation of choline-associated metabolites, including phosphocholine, in BRCA1-mutant ovarian cancer cells and tissues. These metabolic alterations contribute to the enhanced metastatic potential of BRCA1-mutant ovarian cancer cells. In line with our observation, clinical evidence showed that BRCA1-mutant ovarian or breast cancer patients were diagnosed with increased incidence of brain metastasis41. Previous studies have shown that the loss of BRCA1 promotes EMT and metastasis in mammary tumors23. While the role of BRCA1 in suppressing cancer metastasis is documented, the precise mechanisms by which it exerts this function remain incompletely understood. Here, our findings demonstrate that the loss of BRCA1 stimulates choline uptake and metabolism, driving metastasis in ovarian cancer cells. We further identified that the choline metabolite phosphocholine binds to the pro-metastatic protein FAM3C, preventing its polyubiquitination-mediated degradation. These results unveil a mechanism by which BRCA1 deficiency induces cancer metastasis, highlighting a critical link between altered choline metabolism and enhanced metastatic potential in BRCA1-mutant cancers. Enhanced choline metabolism plays a pivotal role in cell membrane synthesis and signaling, making it a recognized metabolic hallmark of tumor progression2. Consequently, several key enzymes in the choline metabolism pathway, such as choline transporters, choline kinase, and phosphocholine cytidylyltransferase, are frequently upregulated in various cancer tissues2. The aberrant expression of these choline metabolism-related enzymes is often driven by oncogenic signals, underscoring the intricate relationship between metabolic reprogramming and cancer development. Oncogenic signaling pathways play critical roles in regulating choline metabolism. For instance, oncogenic RAS controls the transcription of choline kinase and phosphocholine cytidylyltransferase through the activation of the MAPK pathway42. Additionally, inhibition of PI3K has been reported to block choline uptake in lung cancer cells, highlighting the involvement of PI3K in choline metabolism43. Consistent with these observations, our data demonstrate that the tumor suppressor BRCA1 negatively regulates choline metabolism by restraining choline uptake. This finding provides further insight into the metabolic reprogramming driven by oncogenic and tumor-suppressive signals in cancer progression. We show that BRCA1 directly binds to the promoter of the choline transporter CTL4 and epigenetically suppresses its transcription by recruiting the histone methyltransferase EZH2, thereby promoting H3K27me3 modification. The occupancy of EZH2 on chromatin is facilitated by its interaction with the noncoding RNA HOTAIR44. Notably, a previous study reported that BRCA1 disrupts the EZH2/HOTAIR association, leading to the re-targeting of EZH2 to chromatin32. Consistently, our findings demonstrate that BRCA1 directly interacts with EZH2 and facilitates its recruitment to the CTL4 promoter. The overexpression of CTL4 was validated in BRCA1-mutant ovarian cancer tissues, highlighting CTL4 as a feasible therapeutic target to inhibit ovarian cancer metastasis associated with BRCA1 mutations. Leveraging this insight, we identified the natural compound DT-13 as a potent CTL4 inhibitor, which efficiently suppresses BRCA1-deficient ovarian cancer invasion both in vitro and in vivo. Thus, our study identifies CTL4-mediated choline uptake as a potential target for treating metastatic ovarian cancer with BRCA1 mutations. Accumulating evidence showed that highly aggressive cancer cells undergo metabolism reprograming45. It is well-established that remodeling of glucose, lipid, and amino acids plays critical roles in cancer cell migration, invasion, and angiogenesis46–48. Higher levels of a number of key enzymes involved in those metabolism pathways have been observed in the metastatic lesions compared with primary tumors. Also, several oncometabolites, including D-2 hydroxyglutarate (D-2HG), succinate, and methylmalonic acid (MMA), could directly drive cancer metastasis via various molecular mechanisms49. Here, we identified phosphocholine as an oncometabolite that promotes cancer invasion through association with pro-metastatic protein FAM3C. Phosphocholine directly binds and prevents FAM3C degradation, therefore triggering cancer metastasis. In agreement with our findings, previous studies have demonstrated that inhibition of the phosphocholine producer choline kinase (ChoK) significantly suppresses invasion and migration in various cancer cells, reiterating that ChoK is a promising therapeutic target in cancer. Reflecting this potential, several ChoK inhibitors, including MN58b, RSM932A, CK37, and EB-3D, have been developed as anti-cancer agents and are currently undergoing pre-clinical and clinical evaluations50,51. Future investigations are warranted to determine whether these ChoK inhibitors could effectively target and mitigate metastasis in BRCA1-mutant cancers. In summary, we discovered that BRCA1-deficiency induces choline metabolism reprogramming in ovarian cancer cells, thereby promoting metastasis through the stabilization of the pro-metastatic protein FAM3C. Mechanistically, BRCA1 directly associates with and epigenetically suppresses choline transporter CTL4 expression via recruitment of the histone methyltransferase EZH2, leading to H3K27me3 modification at the CTL4 promoter. Furthermore, we identified a potent CTL4 inhibitor, DT-13, a natural compound that efficiently suppresses metastasis in BRCA1-deficient ovarian cancer cells. These findings provide further insights into the metabolic vulnerabilities of BRCA1-mutant cancers and establish CTL4 as a viable therapeutic target for metastatic ovarian cancers with BRCA1 mutations. Methods Reagents and antibodies Anti-CTL1(Cat# A104315,1:1000), anti-CTL3(Cat# A12820,1:1000), anti-CTL4(Cat# A10435,1:1000) and anti-CHT1(Cat# A8247,1:1000) were purchased from ABclonal technology. Anti-CTL2(Cat# ab133722,1:1000) was obtained from Abcam Biotechnology (Cambridge, UK). Anti-BRCA1(Cat# sc-6954,1:1000) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-β-Actin(Cat# 4967,1:2000), FAM3C (Cat# 15171,1:1000), anti-GM310 (Cat# 12480,1:3200), and anti-Myc (Cat# 2276,1:2000), anti-HA(Cat# 2367,1:2000), and anti-Flag(Cat# 2368,1:2000) were obtained from Cell Signaling Technology (Danvers, MA). Anti-Tubulin (Cat# 11224-1-AP,1:2000), anti-GFP(Cat# 66002-1-Ig,1:1000), anti-N-cadherin (Cat# 66219-1-Ig,1:1000), anti-E-cadherin (Cat# 60335-1-Ig,1:1000), and anti-Vimentin(Cat# 60330-1-Ig,1:1000) were purchased from Proteintech Group Inc. (Chicago, IL, USA). Liriope muscari baily saponins C (T-3809), Vesamicol(T24934), JQ1(T2110), Tazemetostat(T1788), Decitabine(T1508), Phosphorylcholine(T73602) were obtained from TargetMol (Shanghai, China). A natural compound library containing 1680 natural compounds was purchased from TargetMol (Shanghai, China). Cell culture and transfection UWB1.289, UWB-BRCA1, CAOV3, KGN, SKOV-3, and 293T cell lines were sourced from the American Type Culture Collection (ATCC). The A2780 line was purchased from Sigma-Aldrich (St. Louis, MO, USA), while the KGN line was additionally obtained from the RIKEN BioResource Research Center (RIKEN BRC). For culture, UWB1.289 and UWB-BRCA1 cells were maintained in a specialized formulation consisting of 50% RPMI 1640, 50% MEGM (Lonza, CC-3150), and 3% fetal bovine serum (FBS), supplemented with penicillin-streptomycin (100 U/mL, Gibco, USA). A2780, CAOV3, KGN, and SKOV-3 cells were grown in DMEM (KeyGEN, Nanjing, China) containing 10% FBS and penicillin-streptomycin. Mycoplasma testing was performed routinely with a commercial detection kit (D101-01, Vazyme, Nanjing, China). Transient transfections were performed using ExFect transfection reagent (Vazyme, NJ, China) following the manufacturer’s protocol. Choline uptake assay Cells grown on the coverslips were treated with 2 mM propargyl choline in Opti-MEM medium for 1 h, then fixed with 4% paraformaldehyde for 15 min, followed by washing with PBS. The sliders were incubated with permeabilization buffer (0.1% saponin, 3% BSA in PBS) for 15 min at room temperature. Then, sliders were incubated with 100 µl of reaction buffer (10 µM Alexa 488 Azide, 1 mM CuSO₄, 100 mM Tris, 50 mM ascorbic acid, 0.01% BSA, and 0.1% saponin in PBS) for 30 min in the dark. After washing with PBS, images were captured by a confocal microscope (Leica SP8, Leica Microsystems Inc). For screening the CTL4 inhibitor, KGN cells with CTL4 overexpression were treated with 10 μM compounds from the library for 2 h before the addition of propargyl choline, and the fluorescence intensity of Alexa 488 was measured with a SpectraMax iD5 plate reader and normalized to untreated cells. Orthotopic ovarian cancer metastasis tumor model Female nude mice aged 6–8 weeks were sourced from GemPharmaTech (Nanjing, China). All procedures involving animals received approval from the Ethics Committee of China Pharmaceutical University and were carried out in compliance with the university’s Animal Care and Use Committee guidelines. No animal exceeded the maximal permissible tumor size/burden during the study. Mice were maintained in a pathogen-free environment with a 12-h light/dark cycle; the ambient temperature was kept at approximately 22–25 °C with 45–50% relative humidity. For orthotopic injection, mice were anesthetized and placed prone under a stereomicroscope. A ~1 cm incision was made in the lower abdomen, approximately 1 cm from the spine, and the ovary was exposed using tweezers. A total of 5 × 10⁵ KGN cells (with or without BRCA1 depletion) suspended in 10 μL of a 1:1 mixture of PBS and Matrigel were introduced into the ovary using a 30-gauge needle. The ovary was then returned to its original position, and the incision was sutured. Mice were monitored daily for humane endpoints. Humane euthanasia was performed immediately when any of the following criteria were met: (1) ≥20% body weight loss from baseline; (2) severe clinical signs, including hunched posture, ruffled fur, lethargy, or loss of response to external stimuli. Animals that did not reach these endpoints were euthanized at 5–6 weeks post-injection, after which livers, intestines, and ovaries were collected, photographed, fixed, and embedded in paraffin. For DT-13 treatment, oral administration of DT-13 was given daily starting from 5 days after cell inoculation. No animal exceeded the predefined humane endpoints during the study. Cellular thermal shift assay (CETSA) For CETSA, 293T cells were transfected with Flag-tagged FAM3C or CTL4 plasmids. After 48 h, cells were lysed in RIPA buffer. Lysates (100 μg protein) were treated with 20 μM phosphocholine, 20 μM DT‑13, or vehicle (0.1% DMSO) for 30 min at 22 °C. Following treatment, samples were heated across a temperature gradient and cleared by centrifugation (12,000 rpm, 10 min). After centrifugation, the cleared supernatants were denatured in SDS‑PAGE loading buffer by boiling for 5 min and then analyzed by Western blotting. Microscale thermophoresis (MST) binding assay MST binding assays were carried out on a NanoTemper Monolith NT.115 instrument following a previously described protocol52. Briefly, GFP-tagged CTL1-4 or GFP-only proteins were produced in 293T cells, which were then lysed using RIPA buffer. The cleared lysates harboring GFP-CTL1-4 or GFP were pre-incubated with serially diluted DT-13 for 30 min and subsequently filled into glass capillaries (MO-K022, NanoTemper Technologies). MST data acquisition was then conducted at 22.8 °C using medium MST power and 20% excitation power. Equilibrium dissociation constants (Kd) were derived from the mass-action model fitted with the NanoTemper software. Phosphorylcholine (Pho-Cho) pulldown assay Eight hundred micrograms of A2780 cell lysate in EBC buffer (50 mM Tris-HCl, pH = 7.6–8.0, 0.5% NP-40, 1 mM EDTA, 1 mM Na3VO4, 50 mM NaF, and 1 mM β-mercaptoethanol) were incubated with p-Aminophenyl phosphoryl choline agarose (Thermo, 20307) or plain agarose beads overnight at 4 °C with rotation. After extensive washing with PBS, the beads were boiled in loading buffer for 5 min. The proteins were resolved on SDS-PAGE and stained with Coomassie Brilliant Blue (Beyotime, P0017) and subsequently identified using mass spectrometry analysis by PTM- Biolabs Inc. (Hangzhou, China). Immunohistochemistry (IHC) staining assay Ovarian cancer tissues with wild-type (WT) and mutant BRCA1 were collected from the Nanjing Maternal and Child Health Hospital (Supplementary Data 4). The use of all human specimens was approved by the institutional review board of Nanjing Maternal and Child Health Hospital. All samples were obtained from female individuals, as ovarian cancer occurs exclusively in females. Informed consent for research participation was obtained from all participants(with no compensation provided). IHC was performed as previously described. Briefly, paraffin-embedded tissue sections were first deparaffinized with 100% xylene and rehydrated using a graded ethanol series (100%, 90%, 70%, 30%, and 0%). Endogenous peroxidase was then inactivated with 3% hydrogen peroxide, followed by antigen retrieval with citrate buffer. After washing with PBS, IHC staining was performed using the R.T.U. Vectastain Kit (Vector Laboratories, CA, USA) following the manufacturer’s protocol. Staining intensity was scored as follows: 1, no staining; 2, weak; 3, moderate; and 4, strong. The percentage of positive staining was scored as: 0 (<1%), 1 (1–25%), 2 (26–50%), 3 (51–75%), and 4 (>75%). The IHC score was calculated as the product of the intensity and percentage scores, ranging from 0 to 16. FAM3C ubiquitination assay Flag-FAM3C and HA-ubiquitin plasmids were co-transfected into the indicated cells. Forty-eight hours after transfection, cells were treated with 20 µM MG132 for 6 h before harvesting in EBC buffer (50 mM Tris-HCl, pH 7.6–8.0, 0.5% NP-40, 1 mM EDTA, 1 mM Na₃VO₄, 50 mM NaF, and 1 mM β-mercaptoethanol). After sonication, cell lysates were centrifuged at 12,000 rpm for 10 min, the supernatant was collected and incubated with anti-flag M2 agarose beads (Sigma) overnight at 4 °C with rotation. After extensive washing, 2 × loading buffer was added, and proteins were resolved on SDS-PAGE, followed by western blot analysis. Molecular docking Molecular docking simulations of phosphatidylcholine into human monomer FAM3C (5LC2) and mouse dimer FAM3C (5LC4) were performed using MOE2022.02 with default parameters. The CTL4 structure was generated using alpha-fold2 computational model, and the DT-13/CTL4 association was simulated using MOE2022.02 software and optimized with the MMFF94 force field. Generation of BRCA1-deficient cells Depletion of BRCA1 in ovarian cancer cells was achieved by lentivirus-mediated expression of shRNA or sgRNA. For the lentivirus production, pLKO.1 plasmid carrying BRCA1 shRNA or lentiCRISPR v2 plasmid carrying BRCA1 sgRNA was co-transfected with psPAX2 and pMD2.G into 293T cells, 48–72 h after transfection, lentivirus containing cell culture medium was collected, and added into ovarian cancer cells in presence of polybrene (6 μg/mL). After selection by puromycin, the depletion efficacy was validated by Western blot. The targeting sequences were listed in Supplementary Data 6. Microspheroid invasion assay 2 × 104 ovarian cancer cells in 15 µl DMEM/F12 medium were mixed with an equal volume of collagenase on ice, and then added to the center of an ultra-low attachment 24-well plate. The plate was inverted and incubated at 37 °C for 20 min before adding 1 ml of DMEM medium containing 10% FBS. Cells were grown for 10–14 days before image capture using leica ultra-high resolution reversed microscope DMi8. High-throughput screening of CTL4 inhibitor The Natural Product Library (L6000; TargetMol), comprising 1680 natural compounds, was utilized to screen for CLT4 inhibitors. Flag-CTL4-expressing KGN cells were plated at 10,000 cells/well in 96-well plates and cultured overnight. Cells were treated with compounds (10 μM) for 2 h, followed by incubation with 2 mM propargyl choline for 1 h. Cells were then fixed with 200 μL of 4% paraformaldehyde for 15 min at room temperature, and were subsequently permeabilized with 200 μL of permeabilization buffer (0.1% saponin, 0.01% BSA in PBS; 15 min). The click reaction was then performed in 100 μL of reaction mixture (10 μM Alexa FluorTM 488 azide, 1 mM CuSO4, 100 mM Tris, 50 mM ascorbic acid, 0.01% BSA, 0.1% saponin in PBS) for 30 min at room temperature in the dark. Following extensive PBS wash, fluorescence intensity was measured using a microplate reader(SpectraMax® iD5) (ex/em = 488/525). Z-scores (standard score) were calculated to normalize inter-experimental variability as previously descried53,54. Z-score is calculated using the formula: Z = (X−μ)/σ (X: the value detected by the microplate reader, μ: the mean of all values, and σ: the standard deviation of all value)53,54. RNA-seq RNA sequencing and analysis were performed by Novogene Biotechnology Inc. (Beijing, China) on an Illumina platform. Total RNA was extracted with Trizol and quality-controlled using the Agilent 5400 system. Total RNA (1 μg/sample) was used for mRNA library preparation. Strand-specific libraries were constructed with the NEBNext® Ultra™ RNA Library Prep Kit (Illumina®, NEB, USA) according to the manufacturer’s protocol, with unique index codes assigned to each sample. The library fragments were purified using the AMPure XP system (Beckman Coulter, Beverly, USA) to preferentially select cDNA fragments ranging from 150 to 200 bp in length. Library quality was assessed using the Agilent Bioanalyzer 2100 system, and subjected to paired-end sequencing (150 bp) on an Illumina NovaSeq 6000 platform (USA), generating approximately 12 Gb of raw data per sample. For bioinformatics analysis, Raw data (raw reads) of fastq format were first processed through in-house Perl scripts. In this step, clean data (clean reads) were obtained by trimming reads containing adapter, ploy-N or with low quality from raw data. Clean reads were then aligned to the human reference homo_sapiens_Ensembl_97 using HISAT2 (v2.2.1) with default parameters. The read numbers mapped to each gene were quantified as FPKM values using StringTie (1.3.3b). Differential gene expression analysis was performed using edgeR, where genes with |log2(fold change)| > 1 and adjusted p value (padj) < 0.05 (Benjamini-Hochberg correction) were considered statistically significant. The results were visualized as volcano plots using ggplot2 in R (v4.0.3). Read alignment distribution and coverage were further examined using the Integrative Genomics Viewer (IGV, v2.8.9). Chromatin immunoprecipitation (ChIP) assay ChIPs were performed using SimpleChIP plus sonication chromatin IP kit (CST, #56383) according to the manufacturer’s instructions. Briefly, cells were cross-linked with 1% formaldehyde for 10 min at room temperature, followed by quenching with glycine solution (CST, #7005). Cells were then lysed sequentially using cell lysis buffer (CST, #96529) and nuclear lysis buffer (CST, #28778). Chromatin was sheared by sonication to generate DNA fragments ranging from 300–500 bp, as verified by agarose gel electrophoresis. Sheared chromatin was immunoprecipitated overnight at 4 °C with rotation using 5 μl of anti-BRCA1 antibody (or anti-mouse IgG as a negative control) coupled to protein G magnetic beads. Beads were washed three times with ChIP wash buffer (CST, #7008), and bound complexes were eluted with ChIP elution buffer (CST, #7009). Crosslinks were reversed by incubation with 5 M NaCl and proteinase K at 65 °C for 6 h. DNA was purified using spin columns and eluted in 30 μL of elution buffer. For ChIP-seq, the DNA was sequenced and analyzed by Novogene Biotechnology Inc. (Beijing, China). Briefly, immunoprecipitated DNA was quantified using a Qubit® 3.0 Fluorometer with the Qubit® DNA Assay Kit (Thermo Fisher Scientific), and purity was assessed by NanoPhotometer® spectrophotometry (IMPLEN), respectively. ChIP-seq libraries were prepared by Novogene Corporation (Beijing, China) and sequenced on an Illumina platform (150 bp paired-end). And library quality was verified using an Agilent Bioanalyzer 2100 system. After that raw sequencing data (FASTQ format) were quality-filtered using fastp (v0.19.11) to remove reads containing adapters, reads containing ploy-N and low-quality reads. Then, clean reads were aligned to the human reference genome (GRCh38/hg38) using BWA-MEM (v0.7.12) with default parameters. Peak calling was performed using MACS2 (v2.1.0) with a q-value cutoff of 0.05. IgG controls were used to filter nonspecific binding. Significant peaks were annotated to genomic features (promoters, exons, enhancers) using ChIPseeker (v1.26.0), with gene ontology (GO) and KEGG pathway enrichment analyses conducted via the GOseq R package and KOBAS software. Statistical significance for enriched motifs and pathways was set at *p* < 0.05 (Fisher’s exact test with Benjamini-Hochberg correction). Uptake of radiolabeled 99mTc-HYNIC-Propargylcholine Radiolabeling of 99mTc-HYNIC-Propargylcholine was performed at the Department of Nuclear Medicine, Mianyang Central Hospital (Sichuan, China) as previously described55. The brief synthesis method was following: HYNIC-NHS (2 mg, 2 eq), 5-azidopentan-1-amine (3 mg, 1 eq) were dissolved in 0.15 ml DMSO, followed by the addition of 0.5 µl DIEA and reacted in a metal bath at 40 °C for 1 h. After isolation and purification, the product and propargylcholine were dissolved in an aqueous solution containing CuSO4 and L-ascorbic acid and reacted for 1 h to synthesize HYNIC-Propargylcholine that was purified by preparative HPLC and confirmed by mass spectrometry. For radiolabeling, 5 µg HYNIC-Propargylcholine, 10 µl TPPTS aqueous solution(100 mg/ml), and 10 µl tricine aqueous solution(130 mg/ml) were added to 100 µl saline and mixed thoroughly. Subsequently, 5 mCi pertechnetate eluate was added, and was reacted at 100 °C for 30 min. Radiochemical purity (RCP) was determined by HPLC, and analytical HPLC was employed to analyze the radiochemical purity. For uptake measurement of 99mTc-HYNIC-Propargylcholine, 104 of KGN-Flag-CTL4 cells were seeded in 96-well plates and incubated overnight before compound treatment. Then, cells were treated with serial concentrations of the tested compound for 2 h before adding 99mTc-HYNIC-Propargylcholine (final dose: 1 µCi) and incubated at 37 °C for another 1 h. After incubation, the cells were washed twice with PBS before being lysed by 100 µl NaOH solution (1 M), and the CPM (Counts Per Minute) were counted using a 2470 Wizard gamma counter (Revvity, Shanghai, China). Measurement of intracellular choline and phosphocholine by LC-MS/MS 3×107cells were lysed in 800ul cold methanol/acetonitrile/water solution (2:2:1, v/v/v) and the mixture was vortexed, sonicated (30 s, 4 °C), and centrifuged (14,000 × g, 20 min, 4 °C). The resulting supernatant containing metabolites was collected for LC-MS/MS analysis. For tissue samples, 100 mg tumor tissues were homogenized in 1 mL of methanol/acetonitrile/water solution using a cryogenic homogenizer (60 Hz, 2 min), followed by centrifugation and the supernatant were subjected to LC-MS/MS analysis using the Triple Quad 5500+ (AB SCIEX Manufacturing). The samples (2 μL) were injected automatically (4 °C) and separated using a Nexera LC-40 UHPLC system (Shimadzu) equipped with an ODS-C18 column (4.6 μm, 250 mm; 40 °C) at 0.5 mL/min; The mobile phases consisted of 10 mM ammonium formate (A) and acetonitrile (B), with the elution system of 30% A,70% B (choline) or 75% A, 25% B (phosphocholine). Samples were analyzed on an AB Sciex Qtrap 5500+ mass spectrometer in ESI positive/negative modes. MRM transitions and retention times were: m/z 104.1 → 45.1 (3.73 min) for choline, and m/z 184.1 → 125.1 (3.96 min) for phosphocholine. Raw data were processed using XCMS for retention time alignment and peak area extraction. Quantification was performed by linear regression of standard curves. Immunofluorescence Cells cultured on glass coverslips were treated with 4% paraformaldehyde for 15 min. Following three washes with PBS, the cells were simultaneously permeabilized and blocked using a blocking buffer (5% donkey serum, 0.3% Triton X-100 in PBS) for 1 h. Primary antibodies were then applied in an antibody dilution buffer (0.3% Triton X-100, 1% BSA in PBS) and incubated at 4 °C overnight. After removal of the primary antibody solution and three PBS rinses, Alexa Fluor 488 or Alexa Fluor 555-conjugated secondary antibody conjugates were added for 1 h in the dark at ambient temperature. Following extensive washing with PBS, the coverslips were mounted using an antifade reagent containing DAPI. Confocal microscopy was carried out on a Leica SP8 instrument (Leica Microsystems Inc). Intracellular sphingomyelin (SM) measurement Sphingomyelin levels in ovarian cancer cells or tumors (Supplementary Data 3) were measured by the sphingomyelin quantification colorimetric Assay Kit (ab287856, Abcam) according to the manufacturer’s instructions. All samples were obtained from female individuals, as ovarian cancer occurs exclusively in females. Informed consent for research participation was obtained from all participants(with no compensation provided). Tumor tissues or cells were first lysed in SM assay buffer and centrifuged at 10,000 × g for 5 min. 20 μl of the resulting supernatants were mixed with an equal volume of SM assay buffer and boiled at 70 °C until cloudy, and centrifuged at 10,000 × g for 2 min. Subsequently, 5 μl of supernatant was mixed with 45 μl of SM assay buffer, followed by the addition of 50 μl reaction mix and incubation at 37 °C for 1 h. After reaction, the absorbance at 570 nm was measured using the SpectraMax iD5. Intracellular choline measurement Intracellular choline levels were measured using the choline quantitation kit (#40007, ATT Bioquest). Tumor tissues or cells were lysed in RIPA buffer (10 mM Tris-HCl, pH 8.0, 140 mM NaCl, 1 mM EDTA, 1% NP-40, 0.1% Sodium deoxycholate, 0.1%SDS) and centrifuged at 12,000 × g for 15 min. 50 μl of supernatant was mixed with an equal volume of choline assay working solution and incubated at room temperature for 15 min. The fluorescence intensity at 540/590 nm was subsequently measured using the SpectraMax iD5. Cell invasion assay Cells (5–10 × 104) suspended in 150 µL of serum-free medium were placed into the upper chamber of a Transwell insert that had been pre-coated with Matrigel (BD Biosciences, CA, USA) and assembled in a 24-well plate. The lower chamber received 700 µL of complete culture medium supplemented with 10% FBS. The plate was then placed at 37 °C for a period of 12–24 h. Cells that had migrated to the underside of the membrane were fixed and stained with 0.5% crystal violet for 30 min, followed by manual counting. Dual-Luciferase reporter assay 1–1000 bp upstream of CTL4 TSS (promoter region) was cloned into pGL3-basic vector between KpnI and XhoI. Then, the pGL3-basic carrying CTL4 promoter was co-transfected with the pRL Renilla luciferase control into UWB1.289/UWB BRCA1 cells or control/BRCA1 sgRNA-expressing KGN cells. Luciferase activity was measured at 48 h after transfection using the dual-luciferase reporter assay System (Promega, USA) following the manufacturer’s instructions. All experiments were conducted in triplicate to ensure reproducibility. Cell cycle progression analysis Cells were fixed overnight in 70% ethanol at 4 °C, washed with PBS, and stained with RNase A(100 μg/mL) and PI (50 μg/mL) for 30 min at room temperature in the dark, followed by flow cytometry analysis. The data were analyzed with FlowJo. Live single cells were gated by FSC-A/SSC-A and FSC-A/FSC-H, the G1, S, and G2/M phases were distinguished based on DNA content (PI fluorescence intensity), as shown in Supplementary Fig. 15. Untargeted metabolomics Metabolomic profiling was performed by Applied Protein Technology Inc. (Shanghai, China). Cell samples (1 × 107 cells per sample) were extracted with cold methanol/acetonitrile/H2O (2:2:1, v/v/v) followed by centrifugation (14,000 × g, 4 °C, 20 min). The resulting supernatant was lyophilized using a vacuum centrifuge. For LC-MS analysis, dried samples were reconstituted in 100 μL of acetonitrile/water (1:1, v/v). Quality control (QC) samples were generated by combining 10 μL aliquots from each sample. Chromatographic separation was performed on a UHPLC system (1290 Infinity LC, Agilent Technologies) equipped with a 2.1 mm × 100 mm ACQUITY UPLC BEH Amide 1.7 μm column (Waters, Ireland) coupled to a quadrupole time-of-flight mass spectrometer (AB Sciex TripleTOF 6600). The mobile phase, applied in both ESI positive and negative modes, consisted of solvent A (water containing 25 mM ammonium acetate and 25 mM ammonium hydroxide) and solvent B (acetonitrile). The gradient elution program was as follows: 95% B held for 0.5 min, linearly decreased to 65% B over 6.5 min, then reduced to 40% B in 1 min and held for 1 min, followed by an increase to 95% B in 0.1 min, with a 3 min re-equilibration period. The ESI source conditions were set as follows: Ion Source Gas1 (Gas1) at 60 psi, Ion Source Gas2 (Gas2) at 60 psi, curtain gas (CUR) at 30 psi, source temperature at 600 °C, and IonSpray Voltage Floating (ISVF) at ±5500 V. For full MS scanning, the m/z range was set to 60–1000 Da with an accumulation time of 0.20 s/spectrum. For auto MS/MS acquisition, the m/z range was 25–1000 Da with an accumulation time of 0.05 s/spectrum for product ion scans. MS/MS acquisition was performed using information-dependent acquisition (IDA) in high-sensitivity mode. Other parameters included: collision energy (CE) fixed at 35 V with a spread of ±15 eV; declustering potential (DP) set to 60 V (positive) and −60 V (negative); exclusion of isotopes within 4 Da; and a maximum of 10 candidate ions monitored per cycle. Prior to multivariate statistical analysis, the data were preprocessed using sum normalization and Pareto scaling. Metabolite identification was performed by matching accurate mass (mass error <10 ppm) using XCMS software and comparing MS/MS spectra against an in-house database of authentic standards. Metabolites with variable importance in projection (VIP) scores >1 were statistically evaluated using Student’s t-test, with p < 0.05 considered statistically significant. CCK-8 assay After transfection, cells were seeded in 96-well plates at a density of 5–10 × 103cells per well in 200 µL of culture medium. After cell attachment, the plates were further incubated at 37 °C for 48 h. Subsequently, 20 µL of CCK-8 solution (Beyotime, Shanghai, China) was added to each well, followed by incubation for 1–4 h before measuring the absorbance at 450 nm. siRNA-mediated knockdown BRCA1 or CTL4 siRNA were obtained from RiboBio (Guangzhou, China) and its targeting sequences were listed in Supplementary Data 6. siRNA transfection was performed using riboFECT™ CP Transfection Kit (Guangzhou, China) following the manufacturer’s instructions. Briefly, cells were grown in 6-well plates until 30–50% confluency and culture medium was changed with serum- and antibiotic-free medium before transfection. Then the transfection complex was prepared by mixing 5 μL of 20 μM siRNA with 120 μL diluted 1× riboFECT™ CP Buffer, followed by adding 12 μL of riboFECT™ CP Reagent. After incubation at room temperature for 15 min, the mixture was added dropwise to the cells, and knockdown efficiency was validated by western blot at 48–72 h after transfection. Phosphorylation array assay The Proteome Profiler Human Phospho-Kinase Array Kit (ARY003C) was obtained from R&D Systems. Cells were collected and lysed using RIPA lysis buffer. The supernatant was obtained after centrifugation. The phosphorylation array membranes were incubated with the lysate overnight. After washing with the wash buffer, the membranes were incubated with the detection antibody for 2 h. After three washes, the membranes were incubated with streptavidin-HRP for 30 min. The membranes were treated with Chemi Reagent Mix, and images were captured using the Tanon 4600 imaging system. Hypotonic shift addition of phosphorylcholine into cells Addition of hydrophilic metabolites into living cells using hypotonic shift as previously descried. Cells were quickly rinsed in KHB buffer(10 mM Hepes and 30 mM KCl) and then incubated with 20 μM phosphorylcholine in KHB buffer at 37 °C for 10 min. The KHB buffer was then replaced with warm medium for 15 min. After 15 min, the completed medium was replaced again, and the cells were cultured for 16 h before subsequent experimental analysis. Colony formation assay Around 1 × 103 of ovarian cancer cells were seeded into six-well plates with three technical replicates. The cells were cultured with or without treatment for 10–14 days to allow colony formation. Colonies were then stained with 0.5% crystal violet in 25% methanol for 30 min, followed by thorough washing to remove excess stain. Then, colonies were counted and normalized to untreated cells. HGSOC orthotopic ovarian cancer metastasis tumor model Cell line generation: Briefly, ID8 cells were transiently transfected with a pSpCas9(BB)-2A-Puro vector harboring Trp53-targeting sgRNA. After removing the non-transfected cells by puromycin selection, cells were plated onto a 96-well plate at 1 cell per well, and the depletion of p53 was confirmed by western blot after clone expansion. Then, ID8 Trp53−/−cells were stably transfected with lentiCRISPR V2 expressing Brca1 sgRNA through puromycin selection, and depletion of BRCA1 was confirmed by western blot. To silence CTL4 in ID8 Trp53−/− cells expressing BRCA1 sgRNA, lentivirus bearing Slc44a4 shRNA in pLKO.1 hygromycin vector was used to infect the cells, and the knockdown efficiency of CTL4 was validated by western blot after hygromycin selection. Orthotopic ovarian cancer model: Female C57BL/6 mice (6–8 weeks old, GemPharmaTech) were housed under specific pathogen-free conditions. All procedures were approved by the Institutional Animal Care and Use Committee. Mice were anaesthetized with isoflurane. A 1 cm incision was made in the lower abdomen, and the left ovary was exteriorized. Cells (5 × 10⁵ in 10 μL of a 1:1 mixture of PBS and Matrigel, Corning) were injected into the ovarian bursa using a 30-gauge needle. The ovary was repositioned and the incision sutured. Mice were euthanized either upon reaching predefined humane endpoints or at 10 weeks post-implantation. Humane euthanasia was performed when any of the following criteria were met: (1) ≥20% body weight loss from baseline; (2) rapid body weight gain (≥10% within 24–48 h) accompanied by pronounced abdominal distension; (3) respiratory distress (dyspnea or tachypnea), lethargy, or loss of response to external stimuli. For animals that survived to the experimental endpoint, ascites volume was measured, metastatic nodules were counted, and primary ovarian tumors were harvested and weighed at 10 weeks. No animal exceeded these predefined humane endpoints during the study. Statistics and reproducibility All statistical analyses were performed using Prism software (GraphPad Prism 8). Data are presented as mean ± SD or mean ± SEM as indicated in the figure legends. Statistical significance was assessed using a two-tailed unpaired Student’s t-test, ordinary one-way ANOVA with Tukey’s multiple comparisons test, or two-way ANOVA with Sidak’s multiple comparisons test, as specified in individual figures. Statistical significance is indicated in the figures. All experiments were biologically independent and repeated at least three times. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 Supplementary Data 7 Reporting summary Transparent Peer Review file Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 Supplementary Data 7 Reporting summary Transparent Peer Review file Source data Source data Source data