Repression of apelin Furin cleavage sites provides antimetastatic strategy in colorectal cancer

Article Repression of apelin Furin cleavage sites provides antimetastatic strategy in colorectal cancer Béatrice Demoures1, Fabienne Soulet1, Jean Descarpentrie1, Isabel Galeano-Otero 1,JoséSanchez Collado1,MariaCasado 1,2, Tarik Smani 3,AlvaroGonzález 1,IsabelAlves 4, Fabrice Lalloué 5, Bernard Masri 6,EstelleRascol 4, Jean-William Dupuy 7,8,CyrilDourthe 1,8, Frédéric Saltel 1,8, Anne-Aurélie Raymond 1,8, Iker Badiola2, Serge Evrard1,9, Bruno Villoutreix 10,SimonPernot 1,9, Géraldine Siegfried 1✉& Abdel-Majid Khatib 1,9 ✉ Abstract The adipokine apelin has been directly implicated in various physiological processes during embryogenesis and human cancers. Nevertheless, the importance of the conversion of its precursor proapelin to mature apelin in tumorigenesis remains unknown. In this study, we identify Furin as the cellular proprotein convertase responsible for proapelin cleavage. We explore the therapeutic potential of targeting proapelin cleavage sites in metastatic colorectal cancer by introducing apelin-dm, a modified variant resulting from alteration in proapelin cleavage sites. Apelin-dm demonstrates efficacy in inhibiting tumor growth, promoting cell death, suppressing angiogenesis, and early colorectal liver metastasis events. Proteomic analysis reveals reciprocal regulation between apelin and apelin-dm on proteins associated with clinical outcomes in colon cancer patients. Apelin-dm emerges as a modulator of apelin receptor dynamics, influencing affinity, internalization, and repression of apelin signaling linked to various protein kinases. Pharmacokinetic and toxicity assessments confirm the specificity, safety, and stability of apelin-dm, as well as its facile hepatic metabolism. These findings position targeting proapelin cleavage as a promising therapeutic strategy against metastatic colorectal cancer, paving the way for further clinical exploration. Keywords Apelin; Furin; Colon Cancer; Liver Metastasis; Safety Pharmacology Subject Categories Cancer; Post-translational Modifications & Proteolysis; Vascular Biology & Angiogenesis https://doi.org/10.1038/s44321-025-00196-5 Received 8 February 2024; Revised 11 January 2025; Accepted 16 January 2025 Published online: 17 February 2025 Introduction Colorectal cancer (CRC) stands as the second leading cause of cancer-related mortality worldwide and ranks third in global cancer diagnoses (Morgan et al, 2023). Despite the efficacy of surgical resection and adjuvant treatments in early-stage disease, relapse and metastasis remain common. The prognosis of CRC is significantly influenced by local tumor staging, with liver metastasis being the predominant occurrence, followed by the lung. Approximately one in fiveCRCpatientspresentwithlivermetastasesat diagnosis, and up to half will develop liver metastasis during the disease course (Siriwardena et al, 2014). Early detection of colorectal liver metastasis (mCRC) is paramount for improving survival rates, enabling the selection of candidates for curative liver surgery and guiding treatment decisions. Apelin, a bioactive peptide initially discovered in the gastrointestinal tract, plays a key role in regulating various physiological functions through its interaction with the G-protein-coupled apelin receptor (Read et al, 2019). Apelin stands at the crossroads of physiological responses, exhibiting heightened expression under hypoxic conditions. Beyond its established roles in promoting epithelial cell growth, proliferation, and migration (BernierLatmani et al, 2022a;Wangetal,2004), apelin has emerged as a key player in the pathophysiology of metabolic and cardiovascular diseases (Palmer et al, 2022; Humphries and Wright, 2008;Adam et al, 2016). This peptide’s involvement in cancer is expansive, with documented expressions in various cancers. Apelin overexpression notably correlates with increased tumor growth and microvessel density across diverse cancer types (Bernier-Latmani et al, 2022a; Zuurbier et al, 2017;Lvetal,2022; Gourgue et al, 2020). In colorectal cancer, apelin concentration becomes a predictive indicator, influencing responses to bevacizumab therapy and affecting susceptibility to apoptosis-inducing agents (Zuurbier et al, 2017). 1University of Bordeaux, Bordeaux Institute of Oncology (BRIC)-UMR1312, Bordeaux, France. 2Department of Cell Biology and Histology, University of the Basque Country, B° Sarriena sn, 48940 Leioa, Spain. 3Institute of Biomedicine of Seville, University Hospital of Virgen del Rocío/University of Seville/CSIC, Avenida Manuel Siurot s/n, 41013 Seville, Spain. 4Univ. Bordeaux, CNRS, Bordeaux INP, CBMN, Bordeaux, France. 5EA3842CAPTuR, GEIST, Faculté de Médecine, Université de Limoges, 2 rue du Dr Marcland, 87025 Cedex Limoges, France. 6Institut Cochin, INSERM U1016, CNRS UMR 8104, Université Paris Cité, 75014 Paris, France. 7Bordeaux Protéome, F-33000 Bordeaux, France. 8Oncoprot Platform, TBM-Core US 005, Bordeaux, France. 9Institut Bergonié, Bordeaux, France. 10Université de Paris, Inserm UMR 1141, Robert-Debré Hospital, 75019 Paris, France. ✉E-mail: [email protected];[email protected] 1234567890();,: 504 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. The cDNA of apelin encodes a 77-amino-acid (aa). The apelin peptides isolated from different tissues and cell types suggest that they are processing products derived from the C-terminal portion of this peptide precursor. The main active forms of apelin are apelin-13 aa, apelin-17 aa, and apelin-36 aa (or proapelin). Apelin36 is directly generated from preproapelin (55 aa) by not wellknown proteases (s) (Lee et al, 2000; Masri et al, 2002). Subsequently, apelin-36 is rapidely cleaved within the two dibasic conserved sequence motifs RRK60F and RR64QR to generate apelin-17 and apelin-13, respectively (Figs. 1A and EV1), suggesting the involvement of proprotein convertases (PCs) in this process (Seidah and Prat, 2012; Siegfried et al, 2020; Scamuffa et al, 2008; Bontemps et al, 2007; Descarpentrie et al, 2022). This study reveals Furin as a unique cellular apelin precursor cleaving enzyme, uncovering the potential clinical significance of unprocessed proapelin, designated apelin-dm, obtained by Furin cleavage sites mutations. When expressed in colon cancer cells or administered pharmacologically to mice with colon cancer tumors, apelin-dm represses the malignant and metastatic phenotype of cancer cells. Apelin-dm altered apelin receptor internalization and signaling and directly competed with apelin and Elabela for apelin receptor binding and activation. Investigation of apln-dm peptide in a panel of Absorption, Distribution, Metabolism, Excretion (ADME), toxicity assays, and pharmacokinetic profiles demonstrated its good safety profile, specificity, and stability in mice and human serum. This highlights apelin-dm as a promising and safe therapeutic strategy for colorectal cancer and its associated metastatic liver events, paving the way for targeted clinical interventions. Results Apelin precursor is specifically cleaved by the proprotein convertase Furin To identify proprotein convertases (PCs) cleaving proapelin, coding region of proapelin was cloned from HUVEC and tagged with a V5 sequence for detection (Fig. 1A). Co-transfection of LoVo cells, aPC-deficient cell line (Takahashi et al, 1993)withapelinprecursor and various PC-encoding vectors revealed Furin’s exclusive role in reducing apelin-55 and apelin-36 levels while generating apelin-17 and apelin-13 (Fig. 1B). Co-transfection of CHO-FD11, cells with reduced Furin activity (Gordon et al, 1997;Sfaxietal,2014)with proapelin and Furin resulted in decreased apelin-17 and increased apelin-13 secretion (Fig. 1C). To assess the efficiency of PC inhibitors on proapelin processing by endogenous Furin, we treated HEK293A cells with the synthetic Furin inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethyl-ketone (CMK, 10 μM) (Seidah and Prat, 2012;Bontempsetal,2007)or expressed in these cells the Furin inhibitors, including the Furin prodoamine (P-p-Furin) (Scamuffa et al, 2014)andα1-PDX (Descarpentrie et al, 2022;Jeanetal,1998) (Fig. 1D). As demonstrated by Western blot analysis, transfection of HEK293A cells with a vector encoding proapelin (Control) resulted in ~95% processing. Treatment of cells with CMK or co-transfection with Furin inhibitors revealed that processing of proapelin is significantly blocked by CMK (~10% processing), α1-PDX (~50% procession) and P-p-Furin (~30% processing). In contrast, wildtype α1-antitrypsin failed to inhibit proapelin processing (Fig. 1D). In vitro enzymatic assays (Sfaxi et al, 2014; Basak et al, 2010;López et al, 2021)confirmed increased enzymatic activity with PCs expression and its reduction by Furin inhibitors expression (Appendix Fig. S1), supporting Furin as primary PC mediating proapelin cleavage, producing apelin-17 and apelin-13. Apelin, apelin receptor, and Furin expression is altered in colorectal cancer and colorectal liver metastases In a healthy colon, crypts contain stem cells responsible for the constant renewal and differentiation of the epithelial lining. While in cancerous tissue, these processes are disrupted, leading to uncontrolled cell proliferation, aberrant differentiation, and loss of normal crypt architecture (Humphries and Wright, 2008). Given the co-expression of apelin, apelin receptor and Furin, that we identified in colon tissues, colon carcinoma cells, and endothelial cells (Appendix Fig. S2), we sought to evaluate their expression and colocalization in both normal and colon cancer tissues. To do this, we first conducted immunohistochemical staining on a set of matched tissue sections, including normal human colon mucosa, primary colon tumors, and corresponding colorectal liver metastases from 35 patients. All these proteins were expressed in noncancerous colon tissues, primarily localized to the base of the colon crypts (Fig. 1E,F). The expression of apelin, apelin receptor and Furin in the colon crypts highlights their potential roles in maintaining the dynamic environment of the intestinal epithelium. In cancerous tissues, the loss of crypts was associated with a modified expression pattern of apelin (Fig. 1H), apelin receptor (Fig. 1G), and Furin (Fig. 1G,H), suggesting that these molecules may play roles in tumor progression and possibly in the tumor microenvironment. Quantitative analysis of immunohistochemistry staining in primary colorectal cancer (CRC) and metastatic colorectal cancer (mCRC) tumors, along with their respective normal colon tissues, revealed a consistent increase in the average staining levels of these proteins in the analyzed tumors (Appendix Fig. S3). Moderate correlation coefficients (ranging from 0.3 to 0.5) were observed between the expression levels of apelin and KI-67, apelin receptor and KI-67, as well as Furin and KI-67 in primary CRC (Fig. 1I) and their corresponding metastatic liver lesions (Fig. 1J). These positive correlations underscore the potential roles of apelin, apelin receptor, and Furin in CRC progression and metastasis. Although APLN mRNA was overexpressed in CRC tumors and their corresponding metastatic livers, the average tumoral/normal (T/N) ratio of apelin expression in hepatic metastases was almost similar to that of colon primary tumors (Fig. 1K). Furthermore, APLNR (Fig. 1L) and Furin (Fig. 1M) were overexpressed in metastatic tumors, with levels higher than those in matched healthy and primary colon tumors. Hypoxia is a common feature of the microenvironment in almost all solid tumors and is frequently associated with angiogenesis and cancer growth, including CRC and has been reported to induce APLN expression in endothelial cells (Andersen et al, 2011) and Furin during regenerative angiogenesis (Khatib et al, 2016). Analysis of Furin expression in endothelial cells cultured under normoxic (21% O 2 )andhypoxic(1%O 2 ) conditions revealed that hypoxia is also a strong inducer of Furin mRNA in endothelial cells, with increased expression observed at all tested hypoxia exposure times ranging from 4 to 24 h (Fig. 1N). Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 505 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. EMBO Molecular Medicine Béatrice Demoures et al 506 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Taken together, these findings indicate that the increased coexpression levels of apelin, the apelin receptor, and/or Furin are associated with CRC tumors, suggesting a potential role for these proteins in colon cancer progression and metastasis. Alteration of proapelin cleavage sites contributes to the reduction in colorectal tumor growth and liver metastasis To investigate the functional significance of apelin precursor processing by Furin in colon cancer, we mutated the proapelin cleavage sites from RR60KFRR64QR to SS60KFSS64QR, creating the apelin double mutant (apelin-dm) (Figs. 2AandEV1). Lentiviral vectors carrying mock, APLN, and APLN-DM were then introduced into colon cancer cell lines CT-26 and MC-38 (Appendix Fig. S4A). Despite Furin presence, the mutations prevent apelin-dm processing (Fig. 2A). Wild-type APLN expression in cancer cells led to twofold increased anchorage-independent growth, while APLN-DM reduced proliferation and showed >80% decrease (Appendix Fig. S4B, C). APLN expression prevented induced apoptosis, unlike APLN-DM (Appendix Fig. S4D). In vivo tumor growth assay in Rag2/γc mice showed apelin-increased growth, while apelin-dm significantly reduced it (Fig. 2B,C). Apelin-dm also inhibited tumor growth in syngeneic BALB/c and C57BL/6 mice (Fig. 2D,E),usingbothCT-26andMC-38cells.The expression of APLN-DM in these cells seems to affect their ability to mediate tumor growth with varying efficacy. Indeed, a greater inhibitory effect was observed in CT-26 cells compared to MC-38 cells, suggesting the potential implication of genetic background differences in the mice used and/or differences in the tumorigenic mediators secreted by CT-26 and MC-38 in their tumor microenvironments. In addition, MC-38 cells produce more apelin compared to CT-26 cells (Appendix Fig. S4A), suggesting the potential involvement of one or more of these differences in the lower anti-tumor efficacy of APLN-DM expression in MC-38 cells compared to CT-26 cells. Similarly, overexpression of APLN in MC-38 and CT-26 cells increased their ability to induce hepatic metastases and incidence, which is repressed by APLN-DM expression (Fig. 2F,G). Notably, APLN-DM-expressing cellderived tumors exhibited a significant decrease in KI-67+cells (Fig. 2H–J). Taken together, these findings suggest that apelin-dm represses tumor growth, angiogenesis, and liver metastasis that are otherwise promoted by apelin. Apelin-dm inversely regulates a protein subset induced by apelin, impacting clinical outcomes in colon cancer To identify the pathways mediated by apelin-dm involved in the repression of the malignant phenotype of colon cancer cells, we performed proteome analysis of cancer cells and the same cells expressing APLN and APLN-DM. Out of the 1825 cellular proteins identified, the expression of APLN and APLN-DM in cancer cells significantly increased the expression of only 14 and 153 proteins, and decreased 20 and 67 proteins, respectively (Fig. 2K). Further analysis revealed that the most significantly dysregulated proteins in response to APLN and APLN-DM expression in cancer cells were related to extracellular matrix regulation, metabolic activity, cytokine production, cell death, and growth (Dataset EV1). In cells expressing APLN-DM, the most induced proteins are singlestranded interacting protein 2 (Rbms2) and RNA-binding motif (RBP). Rbms2 is a tumor suppressor gene that inhibits cell proliferation by positively regulating the stability of P21, mediates cell apoptosis, and enhancing sensitivity to several cancer treatments (Xu et al, 2022;Sunetal,2018). The protein most repressed in apelin-dm-expressing cells was the oncogene GTPase NRas. Importantly, we found that 28 proteins were inversely regulated in cells expressing APLN and APLN-DM (Fig. 2L). Several of these proteins were linked to both good and poor prognosis in colon cancer patients. Indeed, Galectin-3-binding protein (LGALS3BP), gelsolin (Gsn), and Spartin (Spg20), which are associated with a favorable clinical outcome in colorectal carcinoma are downregulated in APLN-expressing cells and upregulated in APLN-DM-expressing cells. In contrast, ATPdependent RNA helicase (DHX36), Synembryn-A (Ric8a), Protein CYR61 (Cyr61), Tubulin alpha-1A (Tuba1a), and probable 28S rRNA (cytosine-C(5))-methyltransferase (NSUN5), which are associated with poor prognosis in colon cancer, are upregulated in APLN-expressing cells and downregulated in APLN-DMexpressing cells. Further analysis revealed that LGALS3BP Figure 1. Specific cleavage of apelin precursor (proapelin) by Furin and altered expression of apelin, Furin and apelin receptor in colon cancer and colorectal liver metastasis. (A) Schematic representation of the primary structure of the 77-amino acid (aa), 55 aa and 36 aa of human apelin precursors and the maturated apelin-17 and apelin-13 aa peptides. The signal peptide (SP), the cleavage site of apelin-55 to apelin-36, and the two proprotein convertase (PCs)-processing sites are indicated (arrows). (B) Proapelin processing by PCs was analyzed by immunoblotting in media of LoVo cells, a PCs activity-deficient cell line, that were co-transfected with empty vector (None), or vectors containing proapelin (Control) and indicated PCs cDNAs. (C) Proapelin processing by Furin in media and lysates from CHO-FD11 cells with reduced Furin activity co-transfected with empty vector (None), or vectors containing preproapelin and/or Furin constructs, was analyzed by western blotting. (D) Inhibition of proapelin processing was assessed by Western blot analysis in media of HEK293A cells co-transfected with empty vector (None), or vectors containing preproapelin (Control) and indicated Furin inhibitors, or treated with the PC inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethyl-ketone (CMK, 10μM). (B–D) The bars show the corresponding percentage of band intensities deduced from the ratio of apelin/(proapelin +apelin) and indicate the level of mature apelin accumulation (%). Data are representative of independent experiments (n=3), and all values represent the mean ± s.e.m. Pvalues were determined by a two-tailed Unpaired ttest. ns not significant. (E,F) Representative immunofluorescence images of colon cancer samples (CRC) and adjacent healthy tissues (Normal) from patients (n=35) that were stained with antiapelin (red) and anti-Furin (green) (E) or with anti-apelin receptor (red) and anti-Furin (green) (F). (G,H) Representative images of apelin (red) and KI-67 (green) staining (G) and apelin receptor (red) and Furin (green) (H) staining in 35 CRC, mCRC and corresponding normal tissues. (I,J) Plots of the correlation between KI-67 and apelin, KI-67 and apelin receptor, and KI-67 and Furin expression in CRC and mCRC that was determined by immunofluorescence following staining intensity analysis of (G,H) data, as measured on ImageJ and Spearman’s correlation analysis. (K–M) Relative expression of APLN (K), APLNR (L) and Furin (M) mRNA in human normal colon (N colon, n=49 patients), corresponding CRC (n=46 patients) and mCRC livers (n=30 patients). Significant differences P were determined by Unpaired ttests using mean ± s.e.m. values. (N) Kinetics of Furin mRNA accumulation in HUVEC cells cultured in normoxia (21% O 2 ) or hypoxia (1% O 2 ) for various time periods as indicated and data are representative of three independent experiments with values that represent the mean ± s.e.m. Significant differences Pwere determined by two-way ANOVA. Scale bar, 100 µm. Source data are available online for this figure. Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 507 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Tumor volume (mm3) Furin 6.5 1.06 Apelin-13 Apelin Apelin-dm 3.5 -+-+ Proapelin Pre-proapelin A CD 0 200 400 600 800 1000 1200 1400 0 9 12 14 16 19 21 Control APLN DM 0 500 1000 1500 2000 2500 3000 3500 0 6 8 11 13 15 18 20 Control APLN APLN DM CT-26/Rag2/γc Tumor volume (mm3) CT-26/Balb/c Control Apelin Apelin-dm Control Apelin-dm Days Days Tumor volume (mm3) P =0.0001 P =0.0005 P =0.0014 0 200 400 600 800 1000 1200 0 3 6 9 12 15 18 21 Control APLN DM MC38/ Black6 Control Apelin-dm Days E P =0.0015 Apelin Apelin-dm CT-26 CellsMC38 Cells Apelin Apelin-dmControl Control Metastasis No. Metastases Cell line incidence (%) per liver (range) CT-26 100 6-67 CT-26/Apelin 100 3-20 CT-26/Apelin-dm 25 1-12 MC-38 50 0-200 MC-38/Apelin 75 0-600 MC-38/Apelin-dm 50 0-20 F GH Apelin Vs Control Apelin-dm Vs Apelin Apelin-dm Vs Control I Apelin Control KI67 CD31 Apelin-dm L DHX36, N O % RSI P<0.0001 KI-67 P<0.0001 % RSI P<0.0001 P<0.0001 CD31 J Angiogenesis (CD31) Liver metastasis Proliferation (Ki67) Tumor grwoth Time post injection Tumor size Analysis B Tumor Liver Subcutaneous Intrasplenic Apelin Apelin-dm CT-26 MC-38 K5 4 3 2 1 0 Log10 fold change 2-4 -2 40 -Log10 p-value Apelin vs Control 0 2 4 Log10 fold change 2-4 -2 40 Apelin-dm vs Control -Log10 p-value M12 10 8 6 4 2 0 6 4 2 0 5 3 1 6 4 2 0 5 3 1 4 2 0 5 3 1 4 2 0 6 GsnAPLN APLNR Spg20 NSUN5 Normal Tumoral EMBO Molecular Medicine Béatrice Demoures et al 508 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. expression was most affected by APLN-DM (~9.5-fold increase, Dataset EV1). LGALS3BP has anti-tumor activity in colon cancer by suppressing Wnt signaling (Piccolo et al, 2015) and also plays a role in preventing and treating inflammatory diseases by suppressing TAK1-dependent NF-κB activation (Hong et al, 2019). Therefore, based on the ability of APLN-DM to regulate the expression of proteins involved in both good and poor prognosis of colon cancer patients, we used the web server GEPIA (http:// gepia.cancer-pku.cn/) (Tang et al, 2017) to analyze their expression in colon cancer patients. We found that Gsn and Spg20 mRNA expression was significantly lower, and NSUN5 mRNA expression was significantly higher in colon tumor tissues than in normal tissues (Fig. 2M). The same analysis also revealed upregulated expression of APLN and APLNR in colon cancer tumors (Fig. 2M). We next explored the correlation between the expression of APLN and APLNR with genes that are dysregulated by APLN or APLNDM expression in cancer cells, and involved in prognosis of colon cancer patients. Among the dysregulated molecules, we observed varying degrees of correlation with APLN and APLNR. Specifically, we found a low positive correlation between APLN and CYR61 (r= 0.16), a stronger correlation between APLNR and CYR61 (r= 0.68), a low correlation between APLNR and RIC8A (r=0.20), and a moderate correlation between APLNR and TUBA1A (r = 0.58) in tumor patients (Fig. 2N). Collectively, these results indicate that APLN-DM induces the expression of proteins associated with a favorable clinical outcome and limits the expression of proteins induced by APLN that are involved in the poor prognosis of colon cancer (Fig. 2O). Apelin-dm peptide represses apelin-mediated vascular network and tumor cells malignant phenotype We next synthesized the apelin-dm peptide to explore its therapeutic potential. In apelin-36, the sequence “RRKFRR”has been substituted with “SSKFSS”, where the R residues have been replaced by S. R is positively charged at neutral pH with approximate molecular (MW) of 174.2 Da. Whereas S is neutral with MW of 105.1 Da. This difference in molecular weight and charge properties likely contributes to the lower migration of apelin-dm peptide on SDS-PAGE than apelin-36 (Fig. 3A,B). First, the impact of varying concentrations of apelin-dm on in vitro angiogenesis was assessed through a tube-like structure formation assay (Appendix Fig. S5A,B). This assay, which evaluated both the number of junctions and tubule length, revealed that apelin-dm disrupted the formation of capillary-like structures in HUVECs in a concentration-dependent manner, with 100 nM being the lowest concentration that elicited a significant effect (Appendix Fig. S5A,B). Comparative analysis with Bevacizumab revealed similar inhibitory effects, with EC50 values of 0.130 μMfor Bevacizumab and 0.083 μM for apelin-dm, respectively (Appendix Fig. S5C). Previously, Bevacizumab was shown to exhibit inhibitory effects at concentrations around 100 nM, both in vitro and in CAM assays (Ljoki et al, 2022;Ademietal,2021), suggesting a similar anti-angiogenic effect for apelin-dm and Bevacizumab at this concentration. In the presence of apelin, HUVECs formed a network of capillary-like tubes, evidenced by a significant increase in the number of junctions and tube length, and this effect was counteracted by apelin-dm (Fig. EV2A,B). In addition, we found that apelin-dm inhibited both HUVEC and SMC proliferation and migration, especially those induced by apelin-13 or apelin-36 (Fig. EV2C–E). Real-time PCR showed the expression of apelin and its receptor in SMCs and HUVECs (Fig. EV2F). Analysis of the apelin receptor at the protein level revealed its high expression in HUVEC cells compared to CT-26 and MC-38 cells (Fig. EV2G). Further analysis revealed high expression of Furin in SMCs (5193 copies/cell) and HUVECs (6013 copies/cell), while it was lower in cancer cells, with 3667 and 3713 copies/cell for CT-26 and MC-38, respectively (Fig. 3C). Angiogenesis is a complex intercellular process involving the organization of various cells. To evaluate the effect of apelin-dm in the context of whole tissue, we performed the aortic ring assay (Fig. EV2H). While apelin induced sprouting of vessel-like structures in the aortic ring assay, apelin-dm repressed this effect. We examined whether apelin-dm could interfere with VEGFinduced angiogenic sprouting in endothelial cells. Our findings indicate that apelin-dm prevents sprouting in endothelial cells Figure 2. Alteration of proapelin cleavage sites represses colorectal tumor growth and colorectal liver metastasis. (A) Colon cancer cells CT-26 and MC-38 infected with lentivirus for stable expression of apelin and apelin-dm generated by mutation of the indicated PCs clevage sites RR60KFRR64 (arrows) and their processing were analyzed by Western blot analysis in the absence and presence of Furin (+/−). (B) Experimental design of in vivo experiments, indicating subcutaneous and intraspleenc injected mice with CT-26 and MC-38 cells expressing APLN and APLN-DM and analysis of developed tumors. (C–E), Tumor growth curves over time from representative experiments of subcutaneous inoculation of control CT-26 and MC-38 cells (10−6) or the same cells stably expressing APLN and APLN-DM into Rag2/γ c (C), Balb/c (D), or Black6 (E) mice (n=8 tumors per group/experiment, 3 experiments). (F) Representative images of livers from Black6 and BALB/c mice injected intrasplenically with control CT-26 and MC-38 cells or the same cells stably expressing APLN and APLN-DM (n=7 tumors per group/experiment, 3 experiments), respectively. (C–E) Error bars indicate s.e.m. center values indicate mean (Mann–Whitney Utest). (G) Incidence of liver nodules (%) and the number of metastases per liver (median) 15 days post-intrasplenic injection are shown. (H) Representative immunofluorescence images of liver tumor sections derived from control CT-26 cells or the same cells stably expressing APLN and APLN-DM that were stained with anti-Ki67 (proliferative index, red) and anti-CD31 (angiogenic index, green). Scale bar, 100 µm. (I,J) The percentage of relative staining intensity (% RSI) of KI67, and CD31 for control CT-26 tumors and CT-26 tumors expressing apelin and apelin-dm. Data are representative of three independent experiments as mean ± s.e.m. Pvalues by two-tailed Unpaired ttest. (K) Volcano plot of proteins in APLN and APLN-DM-expressing cells vs. Controls. Green dots indicate downregulated proteins; and red dots, upregulated proteins, Statistical significance was determined using Welch’sttest, with a Pvalue threshold of 0.05 (n=3). (L) Heatmaps plotting expression of the 28 proteins inversely regulated between APLN and APLN-DM-expressing cells. (M) Expression level of GSN, Spg20, NSUN5, APLN, APLNR in COAD (n=275) or non‐cancerous tissues (n=349) from Gene Expression Profiling Interactive Analysis (GEPIA) datasets (central band, boxes, and whiskers of the boxplot represent the median, first quartile, third quartile, minimum, and maximum values, respectively). Four-way ANOVA, using sex, age, ethnicity, and disease state, was used to calculate statistical significance. (N) Data were derived from GEPIA (Tang et al, 2017) indicating scatter plot graphs of the Spearman correlation analysis of CYR61 and APLN, CYR61 and APLNR, RIC8A and APLN, and RIC8A and APLNR in colon adenocarcinoma. (O) Schematic representation that summarizes APLN and APLN-DM regulated proteins associated with clinical outcome of colorectal cancer patients. Source data are available online for this figure. Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 509 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. 0 5 10 15 20 25 - - 15 5 10 A Apelin-dm: LVQPRGSRNGPGPWQGGSSKFSSQRPRLSHKGPMPF Apelin-36: LVQPRGSRNGPGPWQGGRRKFRRQRPRLSHKGPMPF Apelin-13: QRPRLSHKGPMPF Apelin-17: KFRRQRPRLSHKGPMPF B +- - - - - - - - -+-+- +-+- -+- +-+ -+- Apelin-13 Apelin-36 Apelin-dm 100 U 200 U 1234 5678 9 Apelin-36 Apelin-dm Apelin-17 Apelin-13 2 kDa Furin : Lane 0 2000 4000 6000 8000 Furin expression (copy number ) P = 0.0022 P < 0.0001 ns P < 0.0001 C ED -Apelin-dm - + + -+ -+ Apelin Neovascular branches (%) P = 0.0001 Apelin-dm --++ -+-+ Apelin P = 0.0007 P < 0.0001 0 100 200 300 G < 0.0001 <0.0001 < 0.0001 -+- + Vascular area (%) F Apelin-dm - + + -+-+ Apelin - ++ 50 60 70 80 90 100 110 Neovascular branches (%) Apelin-dm VEGF P <0.0001 P = 0.0030 -+ -++ - HI Apelin-dm -- + -+ + VEGF 0 100 200 300 400 0 100 200 300 400 0 20 40 60 80 100 024487296120 VEGF + Apelin-dm VEGF VEGF -++ + Apelin-dm -- (h) Migration (%) Confluence (%) JK -40 -40 p-ERK ERK VEGF -++ + Apelin-dm -- P=0.0219 P=0,0162 P-ERK activation (%) VEGF -+ + +-- Apelin-dm M N HT-29 A pelin-dm --++ -+-+Apelin L p<0.0001 Apelin-dm Colony (%) --++ -+-+Apelin P = 0.01 P = 0.001 P = 0.0097 O Apelin-dm Migration (%) --++ Apelin-13 -- -- Apelin-36 -+ -+ -- ++ -+ P CT-26 p<0.0001 p<0.0001 P < 0.0001 0 100 200 300 0 200 400 600 800 Q 020 40 60 80 100 0 5 10 15 20 25 p-pol s-pol 0200 400 600 800 0 5 10 15 20 p-pol s-pol S Apelin-dm (pM) Resonance position shifts (mdeg) Resonance position shifts (mdeg) TP = 0.0001 U V KD (pM) P = 0.0002 ns Resonance position shifts (mdeg) 0 50 100 150 Apelin-dm Apelin 80 100 120 140 160 180 Apelin-13 (pM) P = 0.0107 < 0.0001 < 0.0001 p<0.0001 R Computed Energy Score per Residue EMBO Molecular Medicine Béatrice Demoures et al 510 EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. caused by VEGF throughout the treatment periods (Fig. EV2J,K). Under these conditions, apelin-dm was comparable in effect to Bevacizumab. We also tested the anti-angiogenic effect of apelindm using the chick chorioallantoic membrane (CAM) assay. Apelin-dm treatment impaired both basal CAM angiogenesis and CAM angiogenesis induced by apelin. Examination of the CAM vasculature by isolectin B4 staining showed that while apelin induced vessel formation, apelin-dm mainly affected the formation of small vessels (Fig. 3D,E, yellow arrows, and (Fig. EV3A). In addition, we used the neonatal mouse retinal model as an in vivo assay to examine the effect of apelin-dm on neoangiogenesis. At birth, the retina is avascular, and a superficial vascular plexus grows from the center to the periphery during the first week after birth (Selvam et al, 2018). Isolectin B4 staining of the retinal vasculature at P5 revealed that apelin administration significantly increased vascular sprouting in mice compared to controls (Fig. 3F,G). In both untreated and apelin-treated mice, sprouting vessels were reduced by apelin-dm administration. In addition, apelin-dm hindered VEGF-induced endothelial cell sprouting, and disrupted angiogenesis in the CAM assay (Figs. 3H,I and EV3B). To investigate the mechanism by which apelin-dm represses VEGF-induced angiogenesis, we first analyzed its effect on HUVEC proliferation (Fig. 3J) and migration (Fig. 3K) in response to VEGF. As indicated, while VEGF significantly promoted HUVEC proliferation and migration, the presence of apelin-dm suppressed these effects. Furthermore, analysis of ERK activation, a signaling pathway mediated by VEGF and involved in HUVEC proliferation and migration, showed that while VEGF induced ERK activation, treatment with apelin-dm attenuated this effect (Fig. 3L,M). This suggests that apelin-dm interferes with the VEGF signaling pathway, which may explain the reduced angiogenesis mediated by VEGF in the presence of apelin-dm. In soft-agar assays, apelin-dm significantly inhibited the clonogenicity of colon cancer cells, countering the apelin-induced increase in colony formation (Fig. 3N,O). Apelin-13 and apelin-36 stimulated cancer cell migration, while apelin-dm mitigated this effect, suggesting interference with apelin’s autocrine action (Fig. 3P; Appendix Fig. S6). Indeed, all the analyzed colon cancer cells express both apelin and the apelin receptor (Appendix Fig. S2B). Computational docking analysis of apelin peptides and apelin-dm with apelin receptor: predicting binding modes and affinity Aligning apelin-13 with the experimental structure of the AMG3054 peptide mimetic in the presence of the active apelin receptor cryo-EM structure suggests that the C-terminal region of apelin-13 inserts into a deep pocket (Fig. 3Q). Additional CABS-flex simulations induced minor changes in the N-terminal region of apelin-13, while the C-terminus remained deeply anchored compared to the starting position of the peptide. Apelin-dm, apelin-36, and apelin-17, docked using CABS-dock, wrapped around the apelin receptor, showing additional favorable interactions in the peptide C-terminal region compared to apelin-13, particularly for apelin-dm and apelin-36. Models of the apelin receptor with apelin-dm exhibited higher total binding energy values than models with apelin-13 or apelin-17, indicating superior binding of apelin-dm. The predicted binding energy values between apelin-36 or apelin-dm and the receptor were similar (Fig. 3R). APOP identified allosteric sites, with the top site matching the peptide-binding site in the C-terminal region (allosteric site 1) (Fig. 3Q). Another allosteric site, accessible to Figure 3. Apelin-dm peptide represses apelin-mediated vascular network and tumor cells malignant phenotype: Apelin-dm and apelin competition for apelin receptor binding. (A) Amino acid sequences of apelin-13, apelin-17, apelin-36 and apelin-dm peptides. (B) SDS gel analysis of wild-type proapelin and apelin-dm peptide cleavage by Furin. Apelin-13 was added for comparaison. The difference in molecular weight and charge properties of the aa R and S in apelin-dm contributes to the lower migration of apelindm peptide than apelin-36. (C) Furin expression analysis expressed as copy number per cell in smooth muscle cells (SMC), endothelial cells (HUVEC), and cancer cells CT-26 and MC-38 using droplet digital polymerase chain reaction (ddPCR) assay. (D) Representative images of chick chorioallantoic membrane treated for 48 h with apelin (100 nM) or/and apelin-dm (100 nM). Vessel density quantification is presented in (E), with n=8 CAM per group. (F) Representative image of Isolectin B4 staining in retinal whole mounts from 100 nM apelin, apelin-dm or apelin and apelin-dm-treated mice on P5. (G) Quantification of retina vascularization, n=6 total animals per group, from three different experiments. Red arrows indicate the direction of vessel formation. (H) Representative images of CAM (n=8 per group) treated for 48 h with VEGF (30 ng/ml) and/or apelin-dm (100 nM), with small formed vessels indicated by arrowheads. (I) Small formed vessels density quantification in CAM (n=8 per group). (J) Cells were plated at low confluence for time-lapse phase-contrast video microscopy using an IncuCyte microscope, and cell proliferation in the presence of VEGF (30 ng/ml) or apelin-dm (100 nM) was monitored by automated confluence analysis at set intervals after plating (means, n=6 wells per group, three independent experiments). (K) Effect of apelin-dm (100 nM) on VEGF (30 ng/ml)-induced migration of HUVECs. (L,M) Effect of apelin-dm (100 nM) on VEGF (30 ng/ml)-induced ERK phosphorylation in HUVEC cells analyzed by immunoblotting (L). Quantification of ERK phosphorylation relative to control untreated HUVEC assigned 100% (M). (N) Anchorage-independent colony formation assay was performed on colon cancer cells HT-29 in the presence of apelin and/or apelin-dm. (O) The number of colony >100 µm were counted and the results were presented as the percentage of the developed colonies. (P) Effect of apelin peptides and/or apelin-dm on the migration of colon cancer cells CT-26. All data are representative of three independent experiments. All values represent the mean±s.e.m. Significant differences Pwere determined by two-way ANOVA. scales, 500 µm. (Q) Docking peptides into apelin receptor. A representative binding poses for apelin-dm (red) is shown, with the C-terminal region inserted in the main binding pocket (allosteric site 1), while the Nterminal region seemed to wrap around the receptor. The predicted structure of the apelin-36 and apelin receptor is highly similar and is also shown. The two Furin cleavage sites are shown. All apelin peptides are predicted to interact with another allosteric pocket (allosteric site 2) while only apelin-36 or apelin-dm is likely to contact yet another allosteric binding pocket (allosteric site 3). Colored spheres (yellow) show the predicted position of the membrane. Inset, apelin-17 and apelin-13 were also docked into the receptor and numerous contacts between these peptides and the protein are lost as compared to apelin-dm and apelin-36. (R) Apelin-apelin receptor docking scores per-residue and global binding energy scores. The per-residue decomposition of apelin-dm (or apelin36) docking score is shown (the per-residue scores between the two peptides are very similar). Each dot represents an amino acid, the color code is assigned according to the predicted binding score with apelin receptor for each residue. Only some residues are predicted to contribute significantly to the binding with the receptor (e.g., the C-term F). The approximate total interaction energy values for each peptide and the receptor are also shown. (S–U) Plasmon waveguide resonance (PWR). Binding curve for apelin (S), and apelin-dm (T) interaction with apelin receptor in HEK-apelin receptor. (U) Plots showing the results of KD values (n=3) corresponding to (S,T). (V) Conformational changes of apelin receptor in response to apelin and apelin-dm. Spectral changes induced by apelin and apelin-dm binding to apelin receptor with two polarizations (p-pol and s-pol) (n=3). P p-(perpendicular to the sensor surface) polarized light, S s-(parallel to the sensor surface). (U,V) Data are representative of three independent experiments and shown as the mean ± s.e.m. Significant differences Pdetermined by one-way ANOVA with Tukey’s multiple comparison tests. Source data are available online for this figure. Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 511 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. apelin-13, apelin-17, apelin-36, and apelin-dm, was identified and labeled allosteric site 2. A third allosteric pocket (allosteric site 3) can also interact with the N-terminal region of apelin-36 or apelin-dm (Fig. 3Q). The predicted binding energy scores favored apelin-dm or apelin-36 over apelin-13 or apelin-17 (Fig. 3R). Apelin-dm peptide-driven changes in apelin receptor conformation Apelin and apelin-dm’s effects on apelin receptor affinity and conformation were assessed using plasmon waveguide resonance (PWR) (Soulet et al, 2020). The sensor, pretreated with polylysine, captured cell fragments through electrostatic interactions. After capturing fragments, ligand incrementation-induced spectral changes tracked for both polarizations. Ligand affinity analyses showed apelin-dm’s highest affinity (KD ~30 pM) for cell membrane fragments, surpassing apelin (KD ~100 pM) (Fig. 3S–U). Spectral changes under ligand binding revealed distinct conformational changes induced by apelin-dm and apelin-13, with apelin-dm causing more substantial and anisotropic shifts (Fig. 3V). These differences indicated ligand-induced distinct receptor conformational states. Notably, using membranes from HEK cells lacking apelin receptor under the same conditions did not induce significant conformational changes in the receptor (Appendix Fig. S7). Apelin-dm exhibits high-affinity competition with apelin for apelin receptor binding The ability of apelin-dm to repress various biological functions of apelin suggests that apelin-dm probably acts as an inhibitor of endogenous and exogenous apelin in terms of receptor binding. Previously, apelin was reported to cause clathrin-mediated apelin receptor internalization (He et al, 2016;Reauxetal,2001;El Messari et al, 2004a) and translocation of β-arrestin to the cell surface, indicating translocation to the phosphorylated apelin receptor (Lee et al, 2010). After apelin-induced internalization, the apelin receptor can either be recycled to the cell surface or be degraded in lysosomes (Lee et al, 2010). Therefore, we compared the effects of apelin and apelin-dm on the cellular recycling of the apelin receptor. To facilitate the functional analysis of receptor internalization, we used U2OS cells stably co-expressing the human apelin receptor and β-arrestin 2-GFP. Stimulation of these cells with Tamara-apelin for 30 min induced receptor internalization, as visualized by the appearance of a punctuated pattern of red (Tamara-apelin) and green (apelin receptor/arrestin 2-GFP) fluorescence (Fig. 4A). In contrast, cells treated with apelin-dm showed reduced Tamara-apelin receptor internalization (Fig. 4A). The use of the same cells revealed that apelin-dm, like apelin, induces receptor internalization that colocalizes with clathrin (Fig. 4B). Similar to apelin, apelin-dm dose-dependently increases the Bioluminescence Resonance Energy Transfer (BRET) signal between the apelin receptor and Rab5 (Fig. EV4), confirming the ability of both apelin and apelin-dm to promote apelin receptor internalization. Using HEK293A cells stably expressing apelin receptor-EGFP fusion protein, we found that six hours after apelin13 washout, the apelin receptor seemed to be mainly returned to the cell surface. In contrast, in cells treated with apelin-dm, the intracellular vesicles of the apelin receptor remained internalized (Fig. EV5), suggesting delayed or blocked recycling of the apelin receptor in the presence of apelin-dm. To further validate that apelin-dm functions by directly competing with apelin for receptor binding, competitive binding assays were performed to verify the binding affinities of the radiolabeled peptide, [125I]-(Pyr1)-apelin13, in the presence of apelin-13 and apelin-dm. As shown in Fig. 4C, both peptides potently inhibited the specific binding of the [125I]-(Pyr1)-apelin-13 peptide to the apelin receptor in a concentration-dependent manner. However, the apelin-dm peptide showed an IC50 about 2.5 times smaller than that of apelin-13 (0.45 nM versus 1.13 nM), suggesting that lower concentrations of apelin-dm are required to compete with apelin-13 for apelin receptor binding. These results indicate that apelin-dm competes with apelin for apelin receptor binding and represses its activation and internalization. Apelin-dm modulates Elabela activity and its receptor binding Alongside apelin, Elabela can activate the apelin receptor, boasting a distinct amino acid sequence from apelin (Read et al, 2019). It colocalizes with apelin in endothelial cells and circulates in plasma. Elabela shares functional roles with apelin in angiogenesis and cancer (Yang et al, 2017;Souletetal,2020;Nysetal,2024). To investigate whether apelin-dm affects the interaction between Elabela and the apelin receptor, we conducted PWR and ligand affinity analyses. Our findings showed that apelin-dm competes with Elabela’s binding to the apelin receptor, as evidenced by an increased KD from 27 ± 1 to 56.6 ± 2.6 in the presence of apelin-dm (Fig. 4D). In addition, ERK activation analysis in HEK-apelin receptor cells revealed that apelin-dm inhibited Elabela-induced ERK activation (Fig. 4E). These results suggest that apelin-dm competes with apelin and Elabela for the apelin receptor and influences their functions. Apelin-dm targets apelin and non-apelin-relayed signaling pathways To evaluate the impact of apelin-dm on apelin receptor downstream signaling, we initially explored its potential effect on the ERK and AKT pathways, known to be activated by apelin (Masri et al, 2004). Our findings revealed that while apelin treatment activated ERK and AKT in HUVECs, SMCs, CT-26 cells, and HEK293 cells, apelin-dm significantly inhibited these processes (Fig. 4F,G; Appendix Fig. S8). For a more comprehensive assessment of the effect of apelin-dm on various kinase activities, we conducted a screening of diverse kinase activities using the PamGene technology (Versele et al, 2009) that showed substantial differences in kinase activity profiles in control and treated cells. Apelin treatment affected phosphorylation in 60 PTK and 58 STK peptides, whereas apelin-dm induced lower changes in phosphorylation in 15 PTK and 37 STK peptides (Fig. 4H–K; Appendix Fig. S9). In addition, 22 PTK and seven STK peptides exhibited significantly dysregulated phosphorylation between apelin and apelin-dm-treated cells. Apelin treatment indicated significantly higher activity for Src family kinases (Src, Lyn, Lck, HCK, BLK, Fyn) and insulin receptor family (InsR, IRR) among PTKs, as well as heightened activity in the ERK (ERK1 and ERK2) and CDK (CDK11, CDK9) families within STKs (Fig. 4L,M; Appendix Figs. S10 and 11). Conversely, in apelin-dm-treated cells, Src family kinases (Lyn, Lck, HCK, BLK) and CDK family (CDK11, EMBO Molecular Medicine Béatrice Demoures et al 512 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 519 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. for major competition on the apelin receptor to occur between mature apelin rather than apelin-36. The binding pocket crucial for interaction with the C-terminal region of apelin was predicted to be allosteric, corroborating with observed flexibility and structural changes in this region of the receptor. While apelin-13 was predicted to engage two allosteric sites, apelin-dm (or apelin-36) was anticipated to establish further contacts with an additional allosteric site situated on the receptor’s side. This suggests the possibility that the two peptides (e.g., apelin-13 and apelin-dm) induce slightly different structural changes upon binding. The precise 3D predictions of such changes remain challenging, acknowledging the inherent flexibility of the N-terminal of apelin-dm, yet the overall position of the peptides on the apelin receptor seems reasonable. It is important to note that the exact high atomic resolution of apelin-dm remains unknown. Utilizing a plasmon waveguide resonance (PWR) assay, we discerned that apelin-dm and apelin-13 elicited distinct conformational changes in the receptor, indicative of potential differences in apelin receptor affinity and in the likely interactions with the predicted allosteric sites for apelin-13 versus apelin-dm. Furthermore, our findings unveiled that, while apelin triggered the phosphorylation of various kinases associated with cell proliferation, survival, and migration encompassing members of the Src, InsR, ERK, and CDK family kinases apelin-dm exhibited a diminished effect. Notably, apelin-dm not only attenuated the stimulatory effects of apelin but also inhibited the ability of apelin to induce various kinase activities. This divergence in kinase activation profiles further emphasizes the distinct functional impact of apelin-dm compared to its processed peptide apelin. Protein half-life exhibits significant variability spanning several orders of magnitude (Toyama and Hetzer, 2013)withlonger-lived proteins typically serving as “housekeeping”proteins, while shorter-lived proteins often function as phosphorylation targets and signaling molecules. apelin, a key player in physiological processes, has a notably short in vivo half-life (<5 min) (Vickers et al, 2002), implicating its rapid turnover and degradation. Initial studies proposed the involvement of metalloproteases, specifically angiotensin-converting enzyme II (ACE2) and proline carboxy peptidase (PRCP), in apelin degradation, particularly targeting the Pro12-Phe13 residues (Kehoe et al, 2016). ACE2, in particular, efficiently hydrolyzes apelin-13, apelin-36, and apelin-36 (Vickers et al, 2002). Interestingly, the generated fragments from ACE2 cleavage did not induce apelin receptor internalization in vitro, suggesting that these cleavages may induce a distinct conformational state of the apelin receptor compared to full-length peptides (El Messari et al, 2004b). Another identified cleavage site between Leu5-Ser6 residues was attributed to neprilysin (NEP) (McKinnie et al, 2016). Our study revealed that modifying the cleavage site of the apelin precursor significantly increased its half-life. This implies that potential conformational changes introduced by cleavage site mutations may prevent the degradation of apelindsm by proteases such as ACE2 and NEP. Ligand–receptor dynamics often involve trafficking to lysosomes upon binding, where the ligand is released and degraded while the receptor undergoes recycling or complex degradation (Bonecchi et al, 2008). The extended half-life of certain receptors, previously associated with “slow”recycling between endosomes and the cell surface, has been reported for various receptors, including decoy receptors and membrane-localized vascular endothelial growth factor receptor-1 (mVEGFR1) (Boucher et al, 2017). The stability and trafficking of mVEGFR1 are regulated by the palmitoylating enzyme DHHC3 and depalmitoylating enzyme APT1. Impaired mVEGFR1 palmitoylation results in increased trafficking to lysosomes for degradation, disrupting vascular morphogenesis. Our findings highlight the impact of ligand half-life in delaying receptor recycling and trafficking upon binding, subsequently altering receptor function and leading to reduced angiogenesis and tumor progression. Our investigation into the pharmacokinetic properties of apelindm, which revealed ~50% plasma protein binding, presents several advantages that warrant consideration in therapeutic contexts. These binding properties strike a balance between a rapid onset of action and a prolonged duration of therapeutic effect. While a portion of the drug is bound to plasma proteins, ensuring a reservoir for sustained release, an equally substantial fraction remains unbound and pharmacologically active. This equilibrium allows for the swift attainment of therapeutic concentrations in the bloodstream while maintaining a steady supply of bioavailable drug molecules over an extended period. Consequently, apelin-dm may offer a favorable pharmacokinetic profile suitable for conditions requiring both immediate intervention and sustained therapeutic coverage. Highly protein-bound drugs may be susceptible to drug interactions and alterations in pharmacokinetics due to competition for plasma protein binding sites, potentially leading to suboptimal therapeutic outcomes or an increased risk of toxicity. Conversely, drugs with minimal protein binding may exhibit rapid clearance and reduced bioavailability, necessitating frequent dosing regimens or higher doses to maintain therapeutic efficacy. Apelindm’s 50% plasma protein binding confers a degree of flexibility, striking a harmonious balance between efficacy and safety, and minimizing the likelihood of pharmacokinetic fluctuations that could compromise treatment efficacy. Previously, peptides with similar properties have been well tolerated with few side effects, suggesting that apelin-dm is likely to be translatable into clinical practice (Davenport et al, 2020). We elucidate the crucial role of apelin precursor cleavage by Furin and identify apelin-dm as a peptide with anti-tumorigenic and antimetastatic functions. Importantly, this peptide exhibits favorable pharmacological profiles. Overall, we present apelin-dm as a potent, selective, non-toxic inhibitor of the apelin receptor interaction, demonstrating efficacy in a mouse model of colon cancer and colorectal liver metastasis. Our data strongly support apelin-dm as a promising lead compound for further exploration in medicinal chemistry and drug development. Figure 8. ADME, pharmacokinetics, and toxicity analysis of apln-dm. Workflow (A) and analysis (B) of the Absorption, Distribution, Metabolism, Excretion (ADME), Pharmacokinetics (PK) and Toxicity of apelin-dm (details are provided in Appendix pharmacological study 1–4). MTD The maximum tolerated dose. EMBO Molecular Medicine Béatrice Demoures et al 520 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Methods Reagents and tools table Reagent/resource Reference or source Identifier or catalog number Experimental models Human Peripheral Blood Mononuclear Cells (hPBMCs) Institute Bergonié, Bordeaux or EFS, Bordeaux NA Colorectal cancer patients’sample Institute Bergonié, Bordeaux NA HT-29 cells (H. sapiens) ATCC HTB-38 LoVo cells (H. sapiens) ATCC CCL-229 HEK-293 cells (H. sapiens) ATCC CRL-1573 U2OS cells (H. sapiens) ATCC HTB-96 MDA-MB-231 cells (H. sapiens) ATCC HTB-26 MC-38 cells (M. musculus) Kerafast CVCL_B288 CT-26 cells (M. musculus) ATCC CRL-2638 CHO-FD11 cells (C. griseus) Gordon et al, 1997 NA C57BL6/J (M. musculus) Charles Rivers Laboratories N/A Rag2/γc(M. musculus) The Jackson Lab NA BALB/c (M. musculus) Charles Rivers Laboratories NA Recombinant DNA pIRES-APLN This study NA pIRES-APLN-V5 This study NA pIRESAPLN This study NA pIRES-apelin-dm This study NA APLNR-pEGFP-N1 This study NA Antibodies Anti-apelin Eurogentec NA Anti-apelin receptor Abcam ab84296 Anti-Ki67 Cell Signaling 9129T Anti-c-caspase 3 Cell Signaling 96617 Anti-CD31 BD PharmingenTM 553370 Anti-phospho ERK Cell Signaling 9106S Anti-ERK Cell Signaling 4695 Anti-β-actin Cell Signaling 4967 Anti–E-selectin Abcam ab2497 Antihuman Clathrin Abcam ab2731 Anti-V5 Invitrogen NA Secondary anti-rabbit Cell Signaling 7074 Secondary anti-mouse Cell Signaling 7076 Chemicals, enzymes, and other reagents Apelin peptides Clinisciences NA pERTKR-AMC Fluorogenic Peptide Substrate R&D system ES013 M-PER lysis buffer Thermo Fisher Scientific78503 Halt Protease and Phosphatase inhibitors Thermo Fisher Scientific87786 and 78420 STK reagent kit PamGene 32201 PTK reagent kit PamGene 32112 Bovine serum albumin (BSA) Euromedex 04-100-812-C [125I]-MIP-3b Perkin Elmer NA Reagent/resource Reference or source Identifier or catalog number MIP-3b R&D Systems 361MI FITC Annexin V Apoptosis Detection Kit I BD PharmingenTM NA Software Protein identification Proteome Discoverer 1.4 Mascot 2.5 NA Sequest HT NA Percolator algorithm NA Software Docking CABS-dock flexible docking engine Kurcinski et al, 2019 PULCHRA Rotkiewicz and Skolnick, 2008 MMTK http://dirac.cnrs-orleans.fr/MMTK.html pyDockEneRes Romero-Durana et al, 2020 PPM server Lomize et al, 2012 APOP Kumar et al, 2023 Fpocket Le Guilloux et al, 2009 UCSF ChimeraX Pettersen et al, 2021 PyMOL https://www.schrodinger.com ImageJ/Fiji https:// imagej.nih.gov/ij NA BioNavigator V63 PamGene BN63 Primers This study Mutant Sense Primer sequence 5’>3’ pIRES-APLN-M1 Sense 5’-GGCAGGGAGGTTCGAGTAAATTCCGCCGCCAGCGGCCC-3’ Antisense 5’-GGGCCGCTGGCGGCGGAATTTACTCGAACCTCCCTGCC-3’ pIRESAPLN-M2 Sense 5’-GGCAGGGAGGTCGGAGGAAATTCAGCAGCCAGCGGCCC-3’ Antisense 5’-GGGCCGCTGGCTGCTGAATTTCCTCCGACCTCCCTGCC-3’ pIRES-APLN-DM Sense 5’-GGCAGGGAGGTTCGAGTAAATTCAGCAGCCAGCGGCCC3’ Antisense 5’-GGGCCGCTGGCTGCTGAATTTACTCGAACCTCCCTGCC-3’ Species Gene Sens Primer sequence 5‘>3’ Human Furin Sense 5’-GCCCAGAATTGGACCACAGT-3’ Antisense 5’-TCCCGATGTCTTTGGGCTC-3’ Probe 5’-CAGCGGAAGTGCATCATCGACATCC-3’ APLN Commercial Hs00936329_m1 (Thermo Fisher Scientific) APLNR Sense 5’-TGACTTTGACCTCTTCCTCATGAAC-3’ Antisense 5’-GGGTTGAGGCAGCTGTTGAC-3’ Probe 5’-TCTTCCCCTACTGCACCTGCATCAGC-3’ GAPDH Sense 5’-CAAATTCCATGGCACCGTC-3’ Antisense 5’-CCCACTTGATTTTGGAGGGA-3’ Probe 5’-CCCATCACCATCTTCCAGGAGCGAG-3’ Mouse Furin Sense 5’-TGAGCCATTCGTATGGCTACG-3’ Antisense 5’-TGCGCACCTCTAGCCGTT-3’ Probe 5’-TGGTGGAACCCAAGGACATCGGC-3’ APLN Sense 5’-CCACTGATGTTGCCTCCAGAT-3’ Antisense 5’-TCACCAGGTAGCGCATGCT-3’ APLNR Sense 5’-GCATGCCTGGAAGGACTCTAA-3’ Antisense 5’-GGATTGGCTTGAACCTCAGAGTA-3’ HPRT1 Commercial Mm01545399_m1 (Thermo Fisher Scientific) Other BD AccuriTM C6 Cytometer Nikon C2si Eclipse Ti-S Tecan Infinite®F200 PRO, Tecan Group Ltd. Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 521 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Patient samples Samples from 35 colorectal cancer patients, along with corresponding liver metastatic tissues and adjacent normal regions, were collected from frozen and formalin-fixed paraffin-embedded (FFPE) tissues. The specimens were clinically and histopathologically diagnosed at Institut Bergonié, Bordeaux. The tissues, acquired after resection, were promptly placed on ice post-surgery and snap-frozen in liquid nitrogen for subsequent analysis. All human specimens were obtained following written informed consent approved by Bergonié Institute, Bordeaux, France. Patient consent forms for all samples were obtained at the time of tissue acquisition. Biopsies were de-identified. All the performed experiments are conformed to the principles set out in the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report. Mice The mouse strains used in this study included C57BL/6, BALB/c, and Rag2/γc, all obtained from the Jackson Laboratory. Both male and female mice, aged 6–8 weeks, were used in the experiments, with age matching applied wherever possible. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committees (IACUC) of the University of Bordeaux and were conducted under license V7#10362. Mice were maintained under specific pathogen-free (SPF) conditions, with a 12-h light/dark cycle, and provided with ad libitum access to standard chow and water. To ensure environmental enrichment and support animal welfare, cages were supplemented with chew sticks, handling tubes, and domed houses. All experiments complied with the ARRIVE guidelines, as applicable. Analysis of public datasets The comparison between colon tumor and normal tissues regarding GSN, Spg20, NSUN5, APLN, APLNR mRNA expression was performed using GEPIA (Tang et al, 2017). The correlation between CYR6, APLN, and APLNR mRNA expression, and RIC8A, APLN, and APLNR mRNA expression in colon adenocarcinoma (COAD) was also determined using GEPIA. Cloning and mutagenesis pIRES-APLN and pIRES-APLN-V5 constructs were generated by amplifying the coding region of human apelin by PCR using specific primers (Appendix Reagents and Tools Table). Mutagenesis was carried out by an oligonucleotide-directed mutagenesis system (Quick Change site-directed mutagenesis kit, Stratagene) according to the manufacturer’s recommendation using pIRES-APLN construct. The oligonucleotides used to generate the mutations for pIRES-APLNDM generation are listed in “Appendix Reagents and Tools Table“. APLNR cDNA from human colorectal cancer tissue was amplified using PCR and specific primers and the PCR product was subcloned into the pEGFP-N1 vector to generate the APLNR-pEGFP-N1 vector. All constructs were verified by sequencing. Cell transfection, transduction, and culture LoVo, HT-29, CT-26, MC-38, U2OS, and MDA-MB-231 cells were procured from ATCC (Manassas, VA, USA). CHO-FD11 cells with reduced Furin activity were characterized previously (Gordon et al, 1997). All cells were maintained in modified Eagle’smedium (MEM), Dulbecco’s modified Eagle’smedium(DMEM)orRPMI1640 with 10% fetal calf serum, 100 units/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-Glutamine (Dutscher), and incubated at 37 °C with 5% CO 2 . Authenticity and mycoplasma screening were routinely performed by PCR (every two weeks and before each animal experiment for mycoplasma screening). For hypoxic experiments, endothelial cells were serum-starved and placed in a sealed, humidified chamber maintained at 1% O 2 and 5% CO 2 for different time periods ranging from 4 to 24 h. Transfections used Effectene transfection reagent (Qiagen, Germany). Lentiviral vectors with wild-type APLN or APLN-DM cDNA were prepared using pIRES plasmids, cloned into a selfinactivating lentiviral vector with a tdTomato reporter gene (pRRLsin-MND-hPGK-tdTomato-WPRE) under myeloproliferative sarcoma virus enhancer control. All constructs were sequenced, and lentiviral vector production was conducted by the “Vect’UB” service platform at the TMB-Core of the University of Bordeaux. Ribonucleic acid (RNA) extraction, real-time PCR and droplet digital PCR (ddPCR) For real-time PCR, total RNA was extracted using an RNA isolation kit (Macherey-Nagel) including DNase treatment (Qiagen), according to the manufacturer’s instructions, and qPCR data were acquired with the StepOnePlusTM Real-Time PCR System (Applied Biosystems, Courtaboeuf, France), as previously described (Scamuffa et al, 2008). For ddPCR, cDNA was synthesized from 2μg of total RNA using Maxima Reverse Transcriptase (Fisher Scientific) and primed with oligo-dT primers (Fisher Scientific) and random primers (Fisher Scientific). PCR were performed in a final volume of 22 µl containing the required QX200 ddPCR EvaGreen Supermix (Bio-Rad) with a final concentration of 150 nM of each primer sets and 2 μl cDNA equivalent to 4 ng total RNA input. Each ddPCR assay mixture (22 µl) was loaded into a disposable droplet generator cartridge (Bio-Rad). Then, 70 ml of droplet generation oil (Bio-Rad) was distributed into each of the eight oil wells of the cartridge. The cartridge was then placed inside the QX200 droplet generator (Bio-Rad). When droplet generation was completed, the droplets were transferred to a 96-well semi-skirted ddPCR plate (Bio-Rad) heatsealed with foil in a PX1 PCR Plate Sealer (Bio-Rad) and amplified with a Mastercycler Nexus Gradient Thermal Cycler (Eppendorf). Thermal cycling conditions for EvaGreen assays were as follows: 95 °C for 5 min, followed by 45 cyclesof95°Cfor30sand61°Cfor1min,followedby4°C5min and a final inactivation step at 90 °C for 5 min. A no-template control and a negative control for each reverse transcription reaction were included in the assay. After thermal cycling the sealed plate was placed in the QX200 Droplet Reader (Bio-Rad) for data acquisition. The resulting data was analyzed using QuantaSoft Software (version 1.7; Bio-Rad). Production of apelin-13, apelin-36 and apelin-dm peptides Peptides apelin-36 (LVQPRGSRNGPGPWQGGRRKFRRQRPR LSHKGPMPF), apelin-13 (QRPRLSHKGPMPF) and apelin-dm (LVQPRGSRNGPGPWQGGSSKFSSQRPRLSHKGPMPF) were synthesized by Neo Biotech (Clinisciences, France) and EMBO Molecular Medicine Béatrice Demoures et al 522 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. resuspended in water or PBS to a 1–5 mM working solution. Peptides were added directly into the medium of the cells or injected into mice and fertilized eggs for the chick chorioallantoic membrane assay at the indicated concentrations. Immunoblotting Western blot analysis was conducted as previously described (Scamuffa et al, 2008), with modifications to detect various apelin isoforms. After their respective treatments, cells were lysed at 4 °C with lysis buffer (20 mM HEPES at pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 6 mM β-octylglucoside, 10 μg/mL aprotinin, 20 μM leupeptin, 1 mM NaF, 1 mM DTT, and 100 μMsodiumorthovanadate). Lysates were loaded on a 13.3% SDS-PAGE acrylamideglycerol gel, followed by transfer onto a nitrocellulose membrane (Protran 0.2 µm, Amersham). The membrane was fixed with 2.5% glutaraldehyde (Sigma G7651). Membranes were incubated with indicated primary antibodies that were revealed by HRPconjugated secondary antibodies (Amersham Pharmacia Biotech) and enhanced chemiluminescence (ECL Plus, Amersham Pharmacia Biotech) according to the manufacturer’s instructions. After imaging, band quantification was performed using ImageJ software (NIH), and protein levels were normalized to actin. Measurement of PCs activity The PCs activity in cells and media was assessed by the evaluation of the enzymes ability to digest the universal PC substrate, the fluorogenic peptide pERTKR-MCA as described previously (Sfaxi et al, 2014; Scamuffa et al, 2014). Extracts were incubated with pERTKR-MCA (100 µM) during the indicated time periods in the presence of 25 mM Tris (pH 7.4), 25 mM methyl-ethane-sulfonic acid, and 2.5 mM CaCl 2 ,at37°C,andthefluorometric measurements were performed using a spectrofluorometer (Tecan Infinite® F200 PRO, Tecan Group Ltd, France). Internalization assay U2OS cells stably co-expressing human apelin receptor and βarrestin 2-GFP were seeded on poly D-lysine-coated glass slides in 12-well dishes (3 × 105cells/well). After 12 h, the medium was replaced with fresh medium containing apelin-(5-carboxytetramethylrhodamine: apelin-13-TAMRA (100 nM) (Proteogenix, Oberhausbergen, France), apelin-dm, or apelin-13-TAMRA and apelin-dm (100 nM) for 45 min. In another experiment, these cells were treated with apelin (100 nM) or apelin-dm (100 nM) and stained with mouse anti-human clathrin heavy chain antibody (Abcam). For washout experiments, cells were incubated for 30 min with apelin peptides (100 nM), replaced with serum-free medium, and fixed 2 h after washout. Analysis was performed using a Zeiss LSM-780 confocal microscope. Bioluminescence resonance energy transfer (BRET) measurement For the BRET assays, phenol red-free medium was removed from HEK293T cells transiently transfected with apelin receptor-Rluc and Rab5-YFP, and replaced with PBS containing calcium and magnesium. The assay was initiated by adding 10 µl of the cellpermeant Renilla luciferase substrate, coelenterazine h, to achieve afinal concentration of 5 µM. Five minutes later, apelin and apelin-dm were added to assess their activity. Plate readings were taken 15 min after substrate addition. BRET signals were collected using a Mithras LB940 instrument, which integrates signals sequentially in the 465–505 nm and 515–555 nm windows using appropriate bandpass filters, managed by MicroWin 2000 software. Net BRET signals were calculated by subtracting the BRET signal from cells expressing only Rluc-tagged apelin receptor from those co-expressing both Rluc-tagged apelin receptor and YFPtagged Rab5. Radioligand binding assay Radioligand binding assay (Euroscreen catalog: FAST-020B) was performed in 96-well plates (Master Block, Greiner, 786201) containing apelin receptor membrane extracts (2 μg protein/well), 0.02 nM [125I]-MIP-3b (Perkin Elmer, custom product) and apelin-13 or apelin-dm at various concentrations. Nonspecific binding was determined using a 200-fold excess of MIP-3b (R&D Systems, 361MI). Dose-response curves were established and the estimated IC 50 values were calculated for each compound. Soft agar, proliferation, and apoptosis assays Anchorage-independent colony formation assay was performed as previously described (Sfaxi et al, 2014; Scamuffa et al, 2014). Proliferation assays were performed using an IncuCyte live-cell microscopy incubator (Essen Bioscience) as previously described (Souletetal,2020). For the apoptosis assay, tumor cells were grown to 70% confluency, washed repeatedly to remove serum, and then incubated for the indicated time periods in serum-free media. Cells were washed and stained with FITC-labeled Annexin V using the FITC Annexin V Apoptosis Detection Kit I (BD PharmingenTM), according to the manufacturer’s instructions. Cells were analyzed by flow cytometry (BD AccuriTM C6 Cytometer), as previously described (Scamuffa et al, 2014). Tube-like formation assay HUVECs were seeded at a density of 25 × 103cells/well onto chamber slides (Labtek) previously coated with Geltrex (Gibco) at 37 °C for 30 min to allow polymerization. Appropriate mediumcontaining vehicles, apelin, apelin-dm or bevacizumab were added at various concentrations. The plates were examined for tube formation under an inverted light microscope at various time points. Photographs of the tubular network in the wells were captured using a digital camera attached to an inverted microscope. Aortic ring assay Six-week-old C57Bl/6 male mice were anesthetized with isoflurane for 5 min before being sacrificed. The descending aorta was isolated, cleared of adventitia, and placed in serum-free Opti-MEM (Gibco) supplemented with 1× antibiotic/antimycotic (Gibco, 100× stock solution). The aortas were sectioned into ∼25 rings of 0.5 mm (about 15–20 per aorta) thickness and placed in fresh serum-free Opti-MEM with 1× antibiotic/antimycotic for 1 h at 37 °C. The embedded rings Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 523 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. were incubated with apelin (100 nM), apelin-dm (100 nM), Bevacizumab (100 nM), and/or VEGF (30 ng/ml). For each ring, the microvessels emerging from the main ring and individual branches arising from them were quantified. Chick chorioallantoic membrane (CAM) angiogenesis assay Fertilized eggs were allowed to mature ex ovo. A small incision was made in the eggshell on day 3 of development, and seven days later, apelin and/or apelin-dm peptide was added to the CAM tissue (100 nM). In other experiments, 1 × 106control tumor cells or the same cells stably expressing wild-type APLN or APLN-DM were directly allowed to grow on the CAM tissue. Immunohistology in toto was performed for vessel staining. Images of CAM vessels were acquired using a Zeiss Axiophot microscope. Retina angiogenesis assay Postnatal Day 5 (P5) mice were intraperitoneally injected daily for 6 days with apelin (10 mg/kg), apelin-dm (10 mg/kg), or both (10 mg/kg each), via intraperitoneal (IP) injection. Following the treatment period, the mice were humanely euthanized, and their eyes were harvested. The eyes were immediately transferred to 4% paraformaldehyde (PFA) for fixation for 10–15 min, after which they were transferred to cold phosphate-buffered saline (PBS) on ice in a Petri dish for dissection. The retinas were carefully isolated and fixed in 4% PFA for 2 h at room temperature or overnight at 4 °C. After fixation, the retinas were blocked and permeabilized in a solution containing 1% bovine serum albumin (BSA) and 0.5% Triton X-100 overnight at 4 °C. Subsequently, the retinas were washed three times with a buffer solution composed of 0.5% Triton X-100, 1 mM CaCl 2 , 1 mM MgCl 2 ,and1mMMnCl 2 in PBS, pH 6.8. The retinas were then incubated with FITC-labeled Isolectin B4 (1:1000; Vector Laboratories) to label blood vessels. Following staining, the retinas were embedded in Tissue-Tek OCT Compound, and 10-μm cryosections were cut. Confocal immunofluorescence images were acquired with an inverted Nikon C2si Eclipse Ti-S microscope and analyzed using NIS-ElementsAR software (Nikon Instruments Europe B.V.). Maximal tolerated dose (MTD) and repeated dosing toxicity study Apelin-dm was administered at the maximum tolerated dose to six groups of BALB/c mice (n= 6/group; 3 males, 3 females). The control group received saline, while apelin-dm doses were 200, 500, 1000, 1500, and 2000 mg/kg. A single intraperitoneal injection on day 1 was followed by a 14-day observation period. Parameters monitored included mortality, morbidity, clinical signs, and body weight. At the study’s end, gross pathology and organ necropsy were conducted. For the repeated dosing toxicity study, 40 Balb/C mice (5 males, 5 females/ group) were injected 5 days/week for 2 weeks with saline or Apelin-dm (50, 150, 300 mg/kg). Evaluations encompassed clinical observations, body weight, food consumption, hematology, blood chemistry, gross pathology, selected organ necropsy, and weight, and histopathology examinations. Comprehensive study details are in Appendix pharmacological studies-1 and -2. The absorption, distribution, metabolism, excretion (ADME) and pharmacokinetics (PK) studies For in-solution property studies, plasma-binding protein assays were performed using the equilibrium dialysis technique, and an aqueous solubility test was performed using a shake flask system. For in vitro metabolism studies, the half-life and intrinsic clearance of the apelindm peptide were evaluated in human plasma and human and mouse liver microsomes, respectively. Cytochrome P450 (CYP) inhibition was assayed by fluorimetry using recombinant human CYP. Cardiac toxicity was assessed using the hERG-automated whole-cell patchclamp technique. For each assay, various reference compounds were used for comparison. Details of the complete study are provided in Appendix pharmacological studies-3 and -4. Preclinical efficacy experiments Tumor growth effects were assessed by injecting 1 × 106control or APLN/APLN-DM expressing CT-26/MC-38 cells subcutaneously into BALB/c, Rag2/γc, or C57BL/6 mice. Additional experiments involved injecting 1 × 106control CT-26/MC-38/MDA-MB231 cells into BALB/c, C57BL/6, or nude mice, respectively. Apelin-dm treatment (10 mg/kg, i.p.) or saline control commenced when tumors reached ∼60 mm3. LASER Doppler imaging assessed total blood flow, with FlowR®Hemodynamics visualization used for analysis. Results were expressed as % relative to pre-treatment tumor blood flow. Experimental liver metastases were induced as previously described (Sfaxi et al, 2014;Scamuffaetal,2008)by intrasplenic/portal injection of 1 × 106CT-26 or MC-38 cells, followed by randomization into treatment (10 mg/kg, i.p.) or saline groups after 5–6days. Immunostaining and confocal microscopy Sections derived from colon cancer and metastatic liver patients and their corresponding non-cancerous tissues, or from tumors induced in C57BL/6, BALB/c, or Rag2/γ c mice, were stained with unconjugated antibodies and fluorochrome-associated secondary antibodies. All samples were mounted with Prolong containing DAPI (Invitrogen), and confocal immunofluorescence images were taken using the inverted microscope Nikon C2si Eclipse Ti-S with NIS-ElementsAR software (Nikon Instruments Europe B.V.). Proteomic analysis Three independent biological replicates of total protein extracts from control CT-26 cells stably expressing APLN and APLN-DM were compared to control CT-26 cells by label-free protein quantification. Proteins were loaded on a 10% acrylamide SDSPAGE gel and visualized by Colloidal Blue staining. Migration was stopped when the samples had just entered the resolving gel and the unresolved region of the gel was cut into only one segment. Sample preparation and protein digestion by trypsin and NanoLC-MS/MS analysis were performed as previously described with modification (Henriet et al, 2017) and briefly detaile in supplementary Material and Methods. NanoLC-MS/MS analysis was performed using an Ultimate 3000 RSLC Nano-UPHLC system (Thermo Scientific, USA) coupled to a nanospray Q-Exactive hybrid quadruploleEMBO Molecular Medicine Béatrice Demoures et al 524 EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Orbitrap mass spectrometer (Thermo Scientific, USA). Each peptide extract was loaded onto a 300 µm ID × 5 mm PepMap C18 pre-column (Thermo Scientific, USA) at a flow rate of 20 µL/ min. After a 5 min desalting step, peptides were separated on a 75 µm ID × 25 cm C18 Acclaim PepMap®RSLC column (Thermo Scientific, USA) with a 4–40% linear gradient of solvent B (0.1% formic acid in 80% ACN) for 108 min. The separation flow rate was set at 300 nL/min. The mass spectrometer was operated in the positive ion mode at a 1.8 kV needle voltage. Data were acquired using Xcalibur 3.1 software in data-dependent mode. MS scans (m/ z350–1600) were recorded at a resolution of R = 70000 (at m/z 200) and an AGC target of 3 × 106ions was collected within 100 ms. Dynamic exclusion of selected precursors was set to 30 s and the top 12 ions were selected from fragmentation in the HCD mode. MS/MS scans with a target value of 1 × 105ions were collected with a maximum fill time of 100 ms and a resolution of R = 17,500. In addition, only +2and+3 charged ions were selected for fragmentation. Other settings were as follows: no sheath and auxiliary gas flow, heated capillary temperature, 200 °C; normalized HCD collision energy, 27%; and isolation width, 2 m/z.Protein identification was performed using Proteome Discoverer 1.4. Mascot 2.5 and Sequest HT algorithms were used for protein identification in batch mode by searching against a UniProt Mus musculus protein database (49 821 entries, release October 2016; https://www.uniprot.org/ website). The mass tolerances in MS and MS/MS were set to 10 ppm and 0.02 Da. Oxidation (M) and acetylation (K) were searched for dynamic modifications, and carbamidomethylation (C) was used as a static modification. Peptide validation was performed using Percolator algorithm (Käll et al, 2007)andonly“high confidence”peptides were retained corresponding to a 1% False Positive Rate at peptide level. Raw LC-MS/MS data were imported into Progenesis LC-MS (version 4.1; Nonlinear Dynamics, Waters) for feature detection, alignment, and quantification. All sample features were aligned according to retention times by manually inserting up to 50 landmarks, followed by automatic alignment to maximally overlay all two-dimensional (m/z and retention time) feature maps. Singly charged ions and ions with more than six charge states were excluded from the analysis. All remaining features were used to calculate a normalization factor for each sample, which was corrected for experimental variation. Peptide identifications (with FDR < 1%; see above) were imported into Progenesis software. Only nonconflicting features and unique peptides were considered for quantification at the protein level. Quantitative data were considered for proteins, quantified by a minimum of two unique peptides, a fold change above 2, and a statistical Pvalue (Welch’s ttest) lower than 0.05. PamGene kinase assay Following treatment with APLN an APLN-DM, the cells were washed and lysed in M-PER lysis buffer (Thermo Fisher Scientific, 78503) supplemented with Halt Protease and Phosphatase inhibitors (87786 and 78420, Thermo Fisher Scientific), following PamGene’s instructions. The lysates were centrifuged at 4 °C for 15 min at 15,000 × g, and the supernatant was stored at –80 °C until further analysis. For the PamGene kinase assay, Each PamChip® contains four identical porous arrays spotted with distinct 13amino acid-long peptide substrates with phosphosites (196 for the protein tyrosine kinase (PTK) and 144 for the serine–threonine kinase (STK) arrays). Phosphorylation of phosphosites is then used to predict one or multiple upstream kinases (protein tyrosine kinases for the PTK PamChip®and serine–threonine kinases for the STK PamChip®). Therefore, STK and PTK microarray assays were performed according to the manufacturer’s instructions (PTK assay on PamStation 12 User Manual, Version 3.0; Serine Threonine Kinase assay on PamStation 12 User Manual, Version 5.1) (ref). In brief, cell lysates were mixed with reaction solutions generated with the STK reagent kit (PamGene, 32201) or PTK reagent kit (PamGene, 32112) and ampuwa ultrapure water (Fresenius Kabi) and placed on STK PamChips (PamGene, 32501) and PTK PamChips (PamGene, 32508), respectively. Triplicates of the cell lysates derived from the apelinand apelindm-treated cells were processed within the same assay run to allow for correct comparisons. Fluorescently labeled anti-phosphoantibodies were used to detect the phosphorylation activity of kinases present in the sample during and after the pumping of lysates through the 3-dimensional surface of the array. During the assay, the samples were pumped through the porous membrane, and images of each array were taken at several exposure times using a camera in the workstation. Images were later used by BioNavigator®software to calculate the signal values for each phosphosite. The data workflow consisting of image quantification, quality control, statistical analysis, visualization, and interpretation was performed using BioNavigator®software. Analysis of PamGene was performed using the BioNavigator v63 software (BN63; PamGene®International, The Netherlands). After excluding several peptides (undetectable or without kinetics (PTK) and quality control (QC)), 125 of 196 PTK substrates and 120 of 144 STK substrates on the PamChip®kinome array were included in the final analysis. For data normalization, the ComBat batch-effect correction method (Johnson et al, 2007) was used after the Log 2 transformation of the S100 signals. The ratios of the normalized signals of the treated samples and controls were used to calculate the log-fold change for each peptide. For generating peptide phosphorylation heatmaps and compare the phosphorylation levels across all samples, normalized Log 2 S100 signals were used. Statistical significance was tested using Unpaired ttests, and the results are represented by volcano plots generated using BN63 (BN63; PamGene®International, The Netherlands). Peptides with a Pvalue < 0.05 were considered a significant change in the degree of phosphorylation of a peptide in the two groups. Docking computations and structural analyses The crystal structure (PDB 5VBL) of apelin receptor in complex with apelin peptide mimetics (AMG3054) (Ma et al, 2017; Yue et al, 2022) was used for the docking of apelin-13 and apelin-dm and the generation of the apelin receptor-apelin-13 or apelin receptorapelin-dm complex. The ten best apelin receptor-apelin-dm models were generated using the standalone version of the CABS-dock flexible docking engine (Kurcinski et al, 2019). The flexibility of the docked apelin receptor-apelin-dm peptide models and the apelin receptor-apelin-13 transposed model were investigated using the standalone version of CABS-flex (Kurcinski et al, 2019). The CABS Python engine tools use a medium-resolution coarse-grained representation of a protein structure (i.e., not all atoms are present but only the C-alpha and beta atoms, while a pseudo-atom is placed Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 525 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. in the center of mass of the remaining side chain atoms) to accelerate the simulation (~3–4 orders of magnitude faster than allatom molecular dynamics simulation). The full atom structural models were reconstructed using the Protein Chain Reconstruction Algorithm (PULCHRA) (Rotkiewicz and Skolnick, 2008). Additional short energy minimization of the peptides and modeled receptor-peptide complexes was performed with the molecular modeling toolkit MMTK (http://dirac.cnrs-orleans.fr/MMTK.html) to refine the geometry of the system. The per-residue decomposition (of the various energy-minimized apelin receptor-apelin-13 or apelin-dm structural models) of docking energy was computed using pyDockEneRes (Romero-Durana et al, 2020). The computed binding energy score obtained from pyDockEneRes was a linear combination of three energetic terms: electrostatic, desolvation, and van der Waals. The orientation of the receptor structure with respect to the lipid bilayer was predicted using the PPM server (Lomize et al, 2012). The search for allosteric pockets on the receptor was performed using the open-source code, APOP (Kumar et al, 2023). This algorithm predicts binding pockets (via the use of Voronoi tesselation and alpha shapes) and local hydrophobic density using Fpocket (Le Guilloux et al, 2009) and perturbs the predicted pockets by stiffening pairwise interactions in the elastic network of the residues lining the investigated pocket to emulate ligand binding. Structural analysis was performed using UCSF ChimeraX (Pettersen et al, 2021)andthefigure was prepared using PyMOL (https://www.schrodinger.com). Plasmon waveguide resonance (PWR) PWR was used to follow apelin receptor conformational changes upon apelin and apelin-dm addition to HEK-apelin receptor cell membrane fragments overexpressing apelin receptor immobilized on the sensor surface. PWR measurements were performed in a homemade instrument functioning at a fixed wavelength of 632 nm and variable incident angle with an angular resolution of about 0.5 millidegrees, as previously described (Soulet et al, 2020). The polarization angle of the incident light was 45° to allow both p-polarized (parallel to the incident light and perpendicular to the sensor surface) and s-polarized (perpendicular to the incident light and parallel to the sensor surface) light resonances to be obtained within a single angular scan. The sensor consists of a BK-7 prism coated with silver and silica to support the waveguide. All the measurements were performed at 22 °C. Statistical analysis Unless otherwise indicated, data are presented as mean ± s.e.m. A twotailed ttest and oneor two-way ANOVA with Tukey’smultiple comparisons test were used to analyze the data. Pearson’scoefficient was calculated to determine the correlation between normally distributed protein expression in human tumors. Differences between mice groups were analyzed using the Mann–Whitney test in GraphPad Prism (GraphPad Software). No inclusion or exclusion criteria were applied to the human samples or animals included in the analysis, and the statistical significance level is illustrated with Pvalues. Blinding and randomization were applied when pertinent. Specifically for experiments involving animal models and patient-derived samples, that were randomly assigned to experimental groups to minimize selection bias. Statistical Pwas set than 0.05. The relative staining intensity (RSI) was calculated using ImageJ by measuring the integrated density of staining (I stain) within manually selected ROIs, subtracting the background staining intensity (I bg) and normalizing by the area of the ROI (A stain). The Relative Staining Intensity (RSI) was calculated using the following formula: RSI = Istain −Ibg/Astain. Data availability The sequencing data were deposited in the ProteomeXchange consortium via the PRIDE (Perez-Riverol et al, 2019)partner repository. The proteomic data from this publication have been deposited to the Proteomic analysis of Control, APLN, APLN-DMexpressing cells database https://www.ebi.ac.uk/pride/login and assigned the identifier PXD045256. The reviewer account details are Username: [email protected] and Password: XiQWy1dj. The source data of this paper are collected in the following database record: biostudies:S-SCDT-10_1038-S44321-025-00196-5. Expanded view data, supplementary information, appendices are available for this paper at https://doi.org/10.1038/s44321-025-00196-5. Peer review information A peer review file is available at https://doi.org/10.1038/s44321-025-00196-5 The paper explained Problem Colorectal cancer stands as the second leading cause of cancer-related mortality worldwide and ranks third in global cancer diagnoses. Despite the efficacy of surgical resection and adjuvant treatments in early-stage disease, relapse and metastasis remain common and the underlying mechanisms involved in these processes remained elusive. Thus, the discovery of a novel efficient regulator may help elucidate the potential mechanisms of the colorectal liver metastasis pathway and provide potential therapeutic strategies for colorectal liver metastasis treatment. Results We report apelin-specific cleavage and activation by Furin as a key player in colon cancer and early events of colon cancer and liver cells interaction leading to metastasis. Alteration of apelin cleavage sites generates apelin-dm peptide that effectively represses the malignant and metastatic phenotype of cancer cells and angiogenesis through modulation of apelin receptor dynamics, affinity, internalization, and diverse apelin signaling pathways. Pharmacokinetic and toxicity assessments confirm the specificity, safety, and stability of apelin-dm. Impact Through multi-level studies of cell culture, pathological specimens, and animal models, we identify the importance of Furin in apelin activity in colon cancer and liver metastasis and apelin-dm peptide as a potential therapeutic strategy, paving the way for further clinical exploration. EMBO Molecular Medicine Béatrice Demoures et al 526 EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. References Adam F, Khatib A-M, Lopez JJ, Vatier C, Turpin S, Muscat A, Soulet F, Aries A, Jardin I, Bobe R et al (2016) Apelin: an antithrombotic factor that inhibits platelet function. Blood 127:908–920 Ademi H, Shinde DA, Gassmann M, Gerst D, Chaachouay H, Vogel J, Gorr TA (2021) Targeting neovascularization and respiration of tumor grafts grown on chick embryo chorioallantoic membranes. PLoS ONE 16:e0251765 Andersen CU, Hilberg O, Mellemkjær S, Nielsen-Kudsk JE, Simonsen U (2011) Apelin and pulmonary hypertension. Pulm Circ 1:334–346 Basak A, Chen A, Scamuffa N, Mohottalage D, Basak S, Khatib A-M (2010) Blockade of furin activity and furin-induced tumor cells malignant phenotypes by the chemically synthesized human furin prodomain. Curr Med Chem 17:2214–2221 Bernier-Latmani J, Cisarovsky C, Mahfoud S, Ragusa S, Dupanloup I, Barras D, Renevey F, Nassiri S, Anderle P, Squadrito ML et al (2022a) Apelin-driven endothelial cell migration sustains intestinal progenitor cells and tumor growth. Nat Cardiovasc Res 1:476–490 Berta J, Kenessey I, Dobos J, Tovari J, Klepetko W, Jan Ankersmit H, Hegedus B, Renyi-Vamos F, Varga J, Lorincz Z et al (2010) Apelin expression in human non-small cell lung cancer: role in angiogenesis and prognosis. J Thorac Oncol 5:1120–1129 Bonecchi R, Borroni EM, Anselmo A, Doni A, Savino B, Mirolo M, Fabbri M, Jala VR, Haribabu B, Mantovani A et al (2008) Regulation of D6 chemokine scavenging activity by ligandand Rab11-dependent surface up-regulation. Blood 112:493–503 Bontemps Y, Scamuffa N, Calvo F, Khatib A-M (2007) Potential opportunity in the development of new therapeutic agents based on endogenous and exogenous inhibitors of the proprotein convertases. Med Res Rev 27:631–648 Boucher JM, Clark RP, Chong DC, Citrin KM, Wylie LA, Bautch VL (2017) Dynamic alterations in decoy VEGF receptor-1 stability regulate angiogenesis. Nat Commun 8:15699 Buccafusca G, Proserpio I, Tralongo AC, Rametta Giuliano S, Tralongo P (2019) Early colorectal cancer: diagnosis, treatment and survivorship care. Crit Rev Oncol/Hematol 136:20–30 Cazzato RL, Buy X, Alberti N, Fonck M, Grasso RF, Palussière J (2015) Flat-panel cone-beam CT-guided radiofrequency ablation of very small (≤1.5 cm) liver tumors: technical note on a preliminary experience. Cardiovasc Interv Radiol 38:206–212 Davenport AP, Scully CCG, de Graaf C, Brown AJH, Maguire JJ (2020) Advances in therapeutic peptides targeting G protein-coupled receptors. Nat Rev Drug Discov 19:389–413 del Toro R, Prahst C, Mathivet T, Siegfried G, Kaminker JS, Larrivee B, Breant C, Duarte A, Takakura N, Fukamizu A et al (2010) Identification and functional analysis of endothelial tip cell-enriched genes. Blood 116:4025–4033 Descarpentrie J, Araúzo-Bravo MJ, He Z, François A, González Á, GarciaGallastegi P, Badiola I, Evrard S, Pernot S, Creemers JWM et al (2022) Role of furin in colon cancer stem cells malignant phenotype and expression of LGR5 and NANOG in KRAS and BRAF-mutated colon tumors. Cancers 14:1195 El Messari S, Iturrioz X, Fassot C, De Mota N, Roesch D, Llorens-Cortes C (2004a) Functional dissociation of apelin receptor signaling and endocytosis: implications for the effects of apelin on arterial blood pressure. J Neurochem 90:1290–1301 El Messari S, Iturrioz X, Fassot C, De Mota N, Roesch D, Llorens-Cortes C (2004b) Functional dissociation of apelin receptor signaling and endocytosis: implications for the effects of apelin on arterial blood pressure. J Neurochem 90:1290–1301 Falcão-Pires I, Ladeiras-Lopes R, Leite-Moreira AF (2010) The apelinergic system: a promising therapeutic target. Expert Opin Ther Targets 14:633–645 Feng M, Yao G, Yu H, Qing Y, Wang K (2016) Tumor apelin, not serum apelin, is associated with the clinical features and prognosis of gastric cancer. BMC Cancer 16:794 Gordon VM, Benz R, Fujii K, Leppla SH, Tweten RK (1997) Clostridium septicum alpha-toxin is proteolytically activated by furin. Infect Immun 65:4130–4134 Gourgue F, Mignion L, Van Hul M, Dehaen N, Bastien E, Payen V, Leroy B, Joudiou N, Vertommen D, Bouzin C et al (2020) Obesity and triple-negative-breastcancer: is apelin a new key target? J Cell Mol Med 24:10233–10244 Henriet E, Abou Hammoud A, Dupuy JW, Dartigues B, Ezzoukry Z, Dugot-Senant N, Leste-Lasserre T, Pallares-Lupon N, Nikolski M, Le Bail B et al (2017) Argininosuccinate synthase 1 (ASS1): A marker of unclassified hepatocellular adenoma and high bleeding risk. Hepatology 66:2016–2028 He L, Chen L, Li L (2016) The mechanosensitive APJ internalization via clathrinmediated endocytosis: A new molecular mechanism of cardiac hypertrophy. Med Hypotheses 90:6–10 Hong C-S, Park M-R, Sun E-G, Choi W, Hwang J-E, Bae W-K, Rhee JH, Cho S-H, Chung I-J (2019) Gal-3BP negatively regulates NF-κB signaling by inhibiting the activation of TAK1. Front Immunol 10:1760 Humphries A, Wright NA (2008) Colonic crypt organization and tumorigenesis. Nat Rev Cancer 8:415–424 Jean F, Stella K, Thomas L, Liu G, Xiang Y, Reason AJ, Thomas G (1998) alpha1Antitrypsin Portland, a bioengineered serpin highly selective for furin: application as an antipathogenic agent. Proc Natl Acad Sci USA 95:7293–7298 Johnson WE, Li C, Rabinovic A (2007) Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8:118–127 Käll L, Canterbury JD, Weston J, Noble WS, MacCoss MJ (2007) Semisupervised learning for peptide identification from shotgun proteomics datasets. Nat Methods 4:923–925 Kehoe K, Van Elzen R, Verkerk R, Sim Y, Van der Veken P, Lambeir A-M, De Meester I (2016) Prolyl carboxypeptidase purified from human placenta: its characterization and identification as an apelin-cleaving enzyme. Biochim Biophys Acta 1864:1481–1488 Khatib A-M, Auguste P, Fallavollita L, Wang N, Samani A, Kontogiannea M, Meterissian S, Brodt P (2005) Characterization of the host proinflammatory response to tumor cells during the initial stages of liver metastasis. Am J Pathol 167:749–759 Khatib AM, Kontogiannea M, Fallavollita L, Jamison B, Meterissian S, Brodt P (1999) Rapid induction of cytokine and E-selectin expression in the liver in response to metastatic tumor cells. Cancer Res 59:1356–1361 Khatib A-M, Lahlil R, Hagedorn M, Delomenie C, Christophe O, Denis C, Siegfried G (2016) Biological outcome and mapping of total factor cascades in response to HIF induction during regenerative angiogenesis. Oncotarget 7:12102–12120 Kojima Y, Quertermous T (2008) Apelin-APJ signaling in retinal angiogenesis. Arterioscler Thromb Vasc Biol 28:1687–1688 Kumar A, Kaynak BT, Dorman KS, Doruker P, Jernigan RL (2023) Predicting allosteric pockets in protein biological assemblages. Bioinformatics 39:btad275 Kurcinski M, Pawel Ciemny M, Oleniecki T, Kuriata A, Badaczewska-Dawid AE, Kolinski A, Kmiecik S (2019) CABS-dock standalone: a toolbox for flexible protein-peptide docking. Bioinformatics 35:4170–4172 Le Guilloux V, Schmidtke P, Tuffery P (2009) Fpocket: an open source platform for ligand pocket detection. BMC Bioinforma 10:168 Lee DK, Cheng R, Nguyen T, Fan T, Kariyawasam AP, Liu Y, Osmond DH, George SR, O’Dowd BF (2000) Characterization of apelin, the ligand for the APJ receptor. J Neurochem 74:34–41 Lee DK, Ferguson SSG, George SR, O’Dowd BF (2010) The fate of the internalized apelin receptor is determined by different isoforms of apelin mediating differential interaction with beta-arrestin. Biochem Biophys Res Commun 395:185–189 Béatrice Demoures et al EMBO Molecular Medicine © The Author(s) EMBO Molecular Medicine Volume 17 | March 2025 | 504–534 527 Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206. Ljoki A, Aslam T, Friis T, Ohm RG, Houen G (2022) In vitro angiogenesis inhibition and endothelial cell growth and morphology. Int J Mol Sci 23:4277 Lomize MA, Pogozheva ID, Joo H, Mosberg HI, Lomize AL (2012) OPM database and PPM web server: resources for positioning of proteins in membranes. Nucleic Acids Res 40:D370–376 López JJ, Siegfried G, Cantonero C, Soulet F, Descarpentrie J, Smani T, Badiola I, Pernot S, Evrard S, Rosado JA et al (2021) Furin prodomain ppFurin enhances Ca2+entry through Orai and TRPC6 channels’activation in breast cancer cells. Cancers 13:1670 Lv S, An Y, Dong H, Xie L, Zheng H, Cheng X, Zhang L, Teng T, Wang Q, Yan Z et al (2022) High APLN expression predicts poor prognosis for glioma patients. Oxid Med Cell Longev 2022:8393336 Ma Y, Yue Y, Ma Y, Zhang Q, Zhou Q, Song Y, Shen Y, Li X, Ma X, Li C et al (2017) Structural basis for apelin control of the human apelin receptor. Structure 25:858–866.e4 Masri B, Lahlou H, Mazarguil H, Knibiehler B, Audigier Y (2002) Apelin (65-77) activates extracellular signal-regulated kinases via a PTX-sensitive G protein. Biochem Biophys Res Commun 290:539–545 Masri B, Morin N, Cornu M, Knibiehler B, Audigier Y (2004) Apelin (65-77) activates p70 S6 kinase and is mitogenic for umbilical endothelial cells. FASEB J 18:1909–1911 McKinnie SMK, Fischer C, Tran KMH, Wang W, Mosquera F, Oudit GY, Vederas JC (2016) The metalloprotease neprilysin degrades and inactivates apelin peptides. Chembiochem 17:1495–1498 Morgan E, Arnold M, Gini A, Lorenzoni V, Cabasag CJ, Laversanne M, Vignat J, Ferlay J, Murphy N, Bray F (2023) Global burden of colorectal cancer in 2020 and 2040: incidence and mortality estimates from GLOBOCAN. Gut 72:338–344 Nys N, Khatib A-M, Siegfried G (2024) Apela promotes blood vessel regeneration and remodeling in zebrafish. Sci Rep 14:3718 Palmer ES, Irwin N, O’Harte FP (2022) Potential therapeutic role for apelin and related peptides in diabetes: an update. Clin Med Insights Endocrinol Diab 15:11795514221074679 Perez-Riverol Y, Csordas A, Bai J, Bernal-Llinares M, Hewapathirana S, Kundu DJ, Inuganti A, Griss J, Mayer G, Eisenacher M et al (2019) The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res 47:D442–D450 Pettersen EF, Goddard TD, Huang CC, Meng EC, Couch GS, Croll TI, Morris JH, Ferrin TE (2021) UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci 30:70–82 Picault F-X, Chaves-Almagro C, Projetti F, Prats H, Masri B, Audigier Y (2014a) Tumour co-expression of apelin and its receptor is the basis of an autocrine loop involved in the growth of colon adenocarcinomas. Eur J Cancer 50:663–674 Piccolo E, Tinari N, D’Addario D, Rossi C, Iacobelli V, La Sorda R, Lattanzio R, D’Egidio M, Di Risio A, Piantelli M et al (2015) Prognostic relevance of LGALS3BP in human colorectal carcinoma. J Transl Med 13:248 Pitkin SL, Maguire JJ, Kuc RE, Davenport AP (2010) Modulation of the apelin/APJ system in heart failure and atherosclerosis in man. Br J Pharm 160:1785–1795 Read C, Nyimanu D, Williams TL, Huggins DJ, Sulentic P, Macrae RGC, Yang P, Glen RC, Maguire JJ, Davenport AP (2019) International Union of Basic and Clinical Pharmacology. CVII. Structure and pharmacology of the apelin receptor with a recommendation that elabela/toddler is a second endogenous peptide ligand. Pharm Rev 71:467–502 Reaux A, De Mota N, Skultetyova I, Lenkei Z, El Messari S, Gallatz K, Corvol P, Palkovits M, Llorens-Cortès C (2001) Physiological role of a novel neuropeptide, apelin, and its receptor in the rat brain. J Neurochem 77:1085–1096 Romero-Durana M, Jiménez-García B, Fernández-Recio J (2020) pyDockEneRes: per-residue decomposition of protein-protein docking energy. Bioinformatics 36:2284–2285 Rotkiewicz P, Skolnick J (2008) Fast procedure for reconstruction of full-atom protein models from reduced representations. J Comput Chem 29:1460–1465 Scamuffa N, Sfaxi F, Ma J, Lalou C, Seidah N, Calvo F, Khatib A-M (2014) Prodomain of the proprotein convertase subtilisin/kexin Furin (ppFurin) protects from tumor progression and metastasis. Carcinogenesis 35:528–536 Scamuffa N, Siegfried G, Bontemps Y, Ma L, Basak A, Cherel G, Calvo F, Seidah NG, Khatib A-M (2008) Selective inhibition of proprotein convertases represses the metastatic potential of human colorectal tumor cells. J Clin Investig 118:352–363 Seidah NG, Prat A (2012) The biology and therapeutic targeting of the proprotein convertases. Nat Rev Drug Discov 11:367–383 Selvam S, Kumar T, Fruttiger M (2018) Retinal vasculature development in health and disease. Prog Retinal Eye Res 63:1–19 Sfaxi F, Scamuffa N, Lalou C, Ma J, Metrakos P, Siegfried G, Ragg H, Bikfalvi A, Calvo F, Khatib A-M (2014) Repression of liver colorectal metastasis by the serpin Spn4A a naturally occurring inhibitor of the constitutive secretory proprotein convertases. Oncotarget 5:4195–4210 Siegfried G, Descarpentrie J, Evrard S, Khatib A-M (2020) Proprotein convertases: key players in inflammation-related malignancies and metastasis. Cancer Lett 473:50–61 Siriwardena AK, Mason JM, Mullamitha S, Hancock HC, Jegatheeswaran S (2014) Management of colorectal cancer presenting with synchronous liver metastases. Nat Rev Clin Oncol 11:446–459 Soulet F, Bodineau C, Hooks KB, Descarpentrie J, Alves I, Dubreuil M, Mouchard A, Eugenie M, Hoepffner J-L, López JJ et al (2020) ELA/APELA precursor cleaved by furin displays tumor suppressor function in renal cell carcinoma through mTORC1 activation. JCI Insight 5:e129070. Sun X, Hu Y, Wu J, Shi L, Zhu L, Xi P-W, Wei J-F, Ding Q (2018) RBMS2 inhibits the proliferation by stabilizing P21 mRNA in breast cancer. J Exp Clin Cancer Res 37:298 Takahashi S, Kasai K, Hatsuzawa K, Kitamura N, Misumi Y, Ikehara Y, Murakami K, Nakayama K (1993) A mutation of furin causes the lack of precursorprocessing activity in human colon carcinoma LoVo cells. Biochem Biophys Res Commun 195:1019–1026 Takano S, Uchida K, Inoue G, Matsumoto T, Aikawa J, Iwase D, Mukai M, Miyagi M, Takaso M (2018) Vascular endothelial growth factor expression and their action in the synovial membranes of patients with painful knee osteoarthritis. BMC Musculoskelet Disord 19:204 Tang Z, Li C, Kang B, Gao G, Li C, Zhang Z (2017) GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res 45:W98–W102 Toyama BH, Hetzer MW (2013) Protein homeostasis: live long, won’t prosper. Nat Rev Mol Cell Biol 14:55–61 Versele M, Talloen W, Rockx C, Geerts T, Janssen B, Lavrijssen T, King P, Göhlmann HWH, Page M, Perera T (2009) Response prediction to a multitargeted kinase inhibitor in cancer cell lines and xenograft tumors using high-content tyrosine peptide arrays with a kinetic readout. Mol Cancer Ther 8:1846–1855 Vickers C, Hales P, Kaushik V, Dick L, Gavin J, Tang J, Godbout K, Parsons T, Baronas E, Hsieh F et al (2002) Hydrolysis of biological peptides by human angiotensin-converting enzyme-related carboxypeptidase. J Biol Chem 277:14838–14843 Wang G, Anini Y, Wei W, Qi X, OCarroll A-M, Mochizuki T, Wang H-Q, Hellmich MR, Englander EW, Greeley GH (2004) Apelin, a new enteric peptide: localization in the gastrointestinal tract, ontogeny, and stimulation of gastric EMBO Molecular Medicine Béatrice Demoures et al 528 EMBO Molecular Medicine Volume 17 | March 2025 | 504 –534 © The Author(s) Downloaded from https://www.embopress.org on June 18, 2025 from IP 193.147.173.206.