Targeting protein hyper-SUMOylation halts cholangiocarcinoma progression by impairing cancer cell viability and tumor-stroma cross talk

Targeting protein hyper-SUMOylation halts cholangiocarcinoma progression by impairing cancer cell viability and tumor-stroma cross talk VISUAL ABSTRACT ORIGINAL ARTICLE Targeting protein hyper-SUMOylation halts cholangiocarcinoma progression by impairing cancer cell viability and tumor-stroma cross talk Paula Olaizola 1,2,3 |Irene Olaizola 1 |Marta Fernandez de Ara 1 | Ainhoa Lapitz 1,2,4 |Beatriz Val 1,2 |Laura Izquierdo-Sanchez 1,2 | Maite G. Fernandez-Barrena 2,5,6 |Laura Alvarez 5 |Colm J. O’Rourke 7 | Pui Y. Lee-Law 1,8 |Kyle Davies 3 |Andreea Gradinaru 3 | Raul Jimenez-Agüero 1 |Adelaida La Casta 1 |Ioana Riaño 1 | Rocio I.R. Macias 2,9 |Jose J. G. Marin 2,9 |Maria L. Martinez-Chantar 2,10 | Matias A. Avila 2,5,6 |Patricia Aspichueta 2,11,12 |Jesper B. Andersen 7 | Luke Boulter 3,13 |Luis Bujanda 1,2 |Pedro M. Rodrigues 1,2,14 | Maria J. Perugorria 1,2,15 |Jesus M. Banales 1,2,14,16 1 Department of Liver and Gastrointestinal Diseases, Biogipuzkoa Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/ EHU), Donostia-San Sebastian, Spain 2 National Institute for the Study of Liver and Gastrointestinal Diseases (CIBERehd, “Instituto de Salud Carlos III”), Spain 3 MRC Human Genetics Unit, Institute of Genetics and Cancer, University of Edinburgh, Edinburgh, UK 4 Department of Biochemistry and Molecular Biology, Faculty of Science and Technology, University of the Basque Country (UPV/EHU), Leioa, Spain 5 Hepatology Program, CIMA, University of Navarra, Pamplona, Spain 6 Instituto de Investigaciones Sanitarias de Navarra (IdiSNA), Pamplona, Spain 7 Department of Health and Medical Sciences, Biotech Research and Innovation Centre (BRIC), University of Copenhagen, Copenhagen, Denmark 8 Department of Gastroenterology & Hepatology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands 9 Experimenal Hepatology and Drug Targeting (HEVEPHARM) Group, Institute of Biomedical Research of Salamanca (IBSAL), University of Salamanca, Salamanca, Spain 10 Liver Disease Laboratory, CIC bioGUNE, Basque Research and Technology Alliance (BRTA), Bilbao, Spain 11 Department of Physiology, Faculty of Medicine and Nursing, University of the Basque Country (UPV/EHU), Leioa, Spain 12 Biocruces Bizkaia Health Research Institute, Cruces University Hospital, Barakaldo, Spain 13 Cancer Research UK Scotland Centre, Institute of Genetics and Cancer, Edinburgh, UK 14 IKERBASQUE, Basque Foundation for Science, Bilbao, Spain 15 Department of Medicine, Faculty of Medicine and Nursing, University of the Basque Country (UPV/EHU), Leioa, Spain 16 Department of Biochemistry and Genetics, School of Sciences, University of Navarra, Pamplona, Spain Paula Olaizola and Irene Olaizola contributed equally. Pedro M. Rodrigues, Maria J. Perugorria, and Jesus M. Banales shared senior authorship. Abbreviations: α-SMA, alpha-smooth muscle actin; CAFs, cancer-associated fibroblasts; CCA, cholangiocarcinoma; FAP1, fibroblast activation protein 1; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GO, gene ontology; HUVEC, human umbilical vein endothelial cells; iCCA, intrahepatic cholangiocarcinoma; IP, immunoprecipitation; NHC, normal human cholangiocytes; PTM, posttranslational modification; ROS, reactive oxygen species; SAE1, SUMO activating enzyme 1; SAMe, S-adenosyl-methionine; SUMO, small ubiquitin-related modifier; TIGER-LC, The Thailand Initiative in Genomics and Expression Research; TME, tumor microenvironment; UBC9, SUMO conjugating enzyme. Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal’s website, www.hepjournal.com. ------------------------------------------------------------------------------------------------------- This is an open access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 ( http://creativecommons. org/licenses/by-nc-nd/4.0/), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal. Copyright © 2025 The Author(s). Published by Wolters Kluwer Health, Inc. Received: 12 September 2024 | Accepted: 18 December 2024 DOI: 10.1097/HEP.0000000000001259 Hepatology. 2025;00:000–000. www.hepjournal.com | 1 Abstract Background and Aims: Cholangiocarcinoma (CCA) includes a diverse group of biliary malignancies with poor prognosis. Alterations in post-translational modifications contribute to disrupted protein dynamics, cellular disturbances, and disease. This study investigates the role of protein SUMOylation in cholangiocarcinogenesis and its potential as a therapeutic target. Approach and Results: Analysis of CCA tumors from 4 patient cohorts and CCA cell lines, revealed increased expression of the SUMOylation machinery genes SUMO activating enzyme 1 (SAE1) and the SUMO conjugating enzyme UBE2I, regardless of the tumor’s molecular profile, resulting in elevated levels of SUMO1-conjugated proteins. Higher SAE1 and UBE2I levels were both indicative of unfavorable clinical outcomes. Deregulated SUMOylated proteins in CCA, mostly linked to cell proliferation, survival, and homeostasis, were identified through immunoprecipitation and mass spectrometry. Genetic (UBE2I-knockdown) and pharmacological (ML792 and S-adenosyl-methionine) inhibition of SUMOylation effectively suppressed tumorigenesis in subcutaneous and oncogene-driven CCA models, reducing the presence of cancer-associated fibroblasts and increasing the recruitment of antitumor immune cells. In vitro, targeting SUMOylation induced CCA cell death and reduced cell proliferation, colony formation, and spheroid growth. Importantly, ML792 and S-adenosyl-methionine did not adversely affect normal human cholangiocytes. Moreover, co-culture of wild-type or UBE2I-knockdown CCA cells with cancerassociated fibroblasts revealed that depleting SUMOylation in CCA cells impaired cancer-associated fibroblast cell growth and altered their protein secretome, ultimately disrupting CCA growth through a regulatory feedback loop. Conclusions: Aberrant SUMOylation drives CCA progression by enhancing cell survival, proliferation, and shaping the tumor microenvironment. Targeting SUMOylation shows potential in inhibiting CCA growth, representing a promising therapeutic strategy. Keywords: biliary tract cancer, pathogenesis, posttranslational modifications, therapy, tumor microenvironment INTRODUCTION Cholangiocarcinoma (CCA) comprises a heterogeneous group of malignant tumors arising along the biliary tree, categorized anatomically as intrahepatic (iCCA), perihilar, and distal CCA.[1]Ranked as the second most common primary liver cancer, CCA contributes to approximately 3% of all gastrointestinal malignancies and is responsible for 2% of all cancer-related deaths.[1] Patients in the early stages of CCAs typically lack symptoms, leading to the majority of diagnoses occurring in advanced stages (70%), often when the cancer has already spread.[1]This delayed diagnosis, coupled with generally limited benefit from systemic therapies, contributes to unfavorable patient outcomes. Current potential curative strategies for CCA include surgical resection or liver transplantation; however, <30% of patients are eligible for these interventions, and the risk of tumor recurrence remains high.[2]For the remaining 70% of patients with unresectable cases, palliative treatment based on systemic therapy, including genomically guided targeted therapies,[3]remains the only feasible option, highlighting the urgent need for more effective treatment options. Several signaling pathways and secreted factors contribute to the development and growth of CCA Correspondence Jesus M. Banales, Department of Liver and Gastrointestinal Diseases, Biogipuzkoa Health Research Institute, Donostia University Hospital, University of the Basque Country (UPV/EHU), Donostia-San Sebastian 20014, Spain. Email: JESUSMARIA. [email protected] 2 | HEPATOLOGY cells.[1]Among these, posttranslational modifications (PTMs) have recently emerged as crucial in cancer. PTMs are dynamic and generally reversible processes wherein a functional group or protein binds to or is cleaved from a target protein. These chemical changes result in protein structure and function alterations, affecting their activity, turnover, localization, and interactions with other molecules. Different PTMs have been described and constitute crucial mechanisms for regulating protein abundance and function. In addition, PTMs enable cells to respond rapidly to changes in their environment, activating or inhibiting a plethora of biological processes, such as cell proliferation, survival, and signal transduction, without the need for new protein synthesis. We have previously identified protein NEDDylation, to be upregulated and hyperactivated in CCA, thereby promoting tumor growth. Of note, molecular or pharmacological inhibition of NEDDylation halted cholangiocarcinogenesis in experimental models[4]and a phase II clinical trial evaluating the potential pharmacological effect of NEDDylation inhibitors in patients with CCA is ongoing (NCT04175912). Among the different PTMs, protein SUMOylation has also garnered significant attention due to its ability to potentially modulate a wide range of cellular processes relevant to cancer progression, yet it remains poorly investigated.[5]Protein SUMOylation involves the covalent and reversible attachment of a small ubiquitin-like modifier (SUMO) to a lysine residue on a target protein. This process occurs through a three-step enzymatic cascade, which includes the heterodimer SUMO activating E1 enzyme (SAE1/UBA2), a sole SUMO conjugating E2 enzyme (UBC9), and specific E3 ligases. Mammals possess five SUMO paralogs (SUMO1–5), with SUMO 1, 2, and 3 being ubiquitously expressed, while SUMO 4 and 5 expression is restricted to specific tissues. Under physiological conditions, SUMO1 appears to be the primary substrate participating in cell cycle control, nuclear transport, or inflammation.[6]Aberrant protein SUMOylation has been linked to various liver pathologies, including polycystic liver disease,[7]alcohol-associated liver disease,[8]metabolic dysfunction–associated steatotic liver disease,[9]and HCC.[10,11]In the context of CCA, previous studies have suggested that protein SUMOylation contributes to tumor growth.[12,13]These observations highlight the importance of further investigating the precise molecular mechanisms through which SUMOylation contributes to cholangiocarcinogenesis, as well as assessing its potential as a therapeutic target in CCA. In this regard, natural compounds such as ginkolic acid and S-adenosylmethionine (SAMe) have been identified as inhibitors of protein SUMOylation. Additionally, highly selective small molecules, such as ML792, have been developed to specifically target and inhibit the activation of this pathway. Currently, selective compounds targeting the SUMOylation pathway are being investigated in clinical trials for solid tumors and lymphomas (NCT03648372, NCT04074330, and NCT04381650). Recent preclinical studies have also highlighted the immunomodulatory effects of SUMOylation inhibition, which can activate an antitumor immune response.[14] These effects have been further validated in clinical trials with TAK-981, a selective inhibitor of SAE1.[15,16] Building on these insights, our study aimed to investigate the role of protein SUMOylation in the pathogenesis of CCA, focusing not only on its effects in malignant cells but also on its impact on the tumor microenvironment (TME). This approach was intended to identify key SUMOylated proteins involved in CCA and to evaluate the therapeutic potential of modulating SUMOylation levels. METHODS Human samples Transcriptomic data from human CCA tumors and adjacent control tissues (ie, liver or normal bile duct), sourced from 3 distinct patient cohorts, were analyzed: Copenhagen (Denmark; dataset: GSE26566), The Cancer Genome Atlas, The Thailand Initiative in Genomics and Expression Research (TIGER-LC; dataset: GSE76311). Additionally, the San Sebastian cohort (JBA-TCG-2017-01) was used to validate the findings. In parallel, immunohistochemistry studies were carried out on paraffin-embedded iCCA tissues obtained from the Donostia University Hospital. The research protocols received approval by the Clinical Research Ethic Committee at the Biogipuzkoa Health Research Institute (BHRI, San Sebastian, Spain). Strict adherence to ethical standards was followed, and all participating patients provided written consent for the utilization of their tissue samples in biomedical research endeavors. Single-cell RNA-sequencing analysis of human iCCA tissue Single-cell transcriptome profiling of CCA tumors from the GSE151530 Gene Expression Omnibus dataset (PMID: 34216724) was analyzed as described in the Supplemental Material, http://links.lww.com/HEP/J723. Cell cultures Normal human cholangiocytes (NHC) were isolated from normal liver tissue and maintained using established methods.[4]Four distinct human CCA cell lines were used: HUCCT1 (iCCA), EGI1, TFK1, and WITT (all 3 extrahepatic CCA). Additionally, the commercially available human umbilical vein endothelial cells (HUVECs) were employed, following the culture instructions TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 3 provided by the suppliers. Furthermore, FAP + human cancer-associated fibroblasts (CAFs) were isolated from a surgically resected iCCA at the BHRI (San Sebastian, Spain), characterized and cultured as described.[17]Liver tissue was obtained following the guidelines approved by the Ethical Committee of Donostia University Hospital and after obtaining the necessary written informed consent from the patients. All the cells—NHC, CCA cell lines, HUVECs, and CAFs—underwent rigorous mycoplasma testing and consistently tested negative throughout the course of the experiments. Further details on cell culture and characterization can be found in the Supplemental Material, http://links.lww.com/HEP/J723. RNA isolation and gene expression Total RNA was extracted from human liver samples (ie, CCA tumors and surrounding normal liver tissue) as well as cultured cells, using TRI Reagent® (Sigma) following the manufacturer’s instructions. Subsequently, reverse transcription and quantitative realtime PCRs were performed as outlined in the Supplemental Data, http://links.lww.com/HEP/J723. The expression levels of glyceraldehyde-3-phosphate dehydrogenase (GAPDH)orribosome protein L22 (RPL22) were used as controls. Primer sequences (Sigma) are provided in Supplemental Table S1, http:// links.lww.com/HEP/J723. Histological analyses Hematoxylin and eosin staining was conducted to assess tissue morphology, as outlined.[7]Immunohistochemistry was employed to detect SUMO1 protein expression, including both free SUMO1 and SUMO1conjugated proteins, in paraffin-embedded sections of healthy liver tissue, normal surrounding tissue, and iCCA tissue from human samples, as described.[7]In the different mouse models of CCA, immunohistochemistry was used to detect CAF (ie, PDGFR, alpha smooth muscle actin (α-SMA), fibroblast activation protein 1 (FAP1), and COL1A1), immune (ie, CD3, CD4, CD8, F4/80, and NK1.1) and endothelial (ie, CD31) markers, as detailed.[7]Immunofluorescence was employed to detect proliferating cells, as detailed.[7]The antibodies employed are listed in Supplemental Table S2, http://links.lww.com/HEP/J723. Immunoblotting Changes in protein levels were assessed through immunoblotting, utilizing whole-cell lysates or cell lysates obtained via SUMO1-immunoprecipitation (IP) in both normal and tumor cholangiocytes, following the procedures outlined in the Supplemental Data, http:// links.lww.com/HEP/J723. The antibodies used for these analyses are listed in Supplemental Table S2, http:// links.lww.com/HEP/J723. Cell viability, proliferation, cell cycle, and apoptosis Viability, proliferation, cell cycle, and apoptosis assays were performed on both normal and tumor cholangiocytes after a 48-hour incubation time with varying doses of ML792 (ranging from 0.001 to 20 μM) or SAMe (ranging from 0.1 to 1 mM), as described in the Supplemental Material, http://links.lww.com/HEP/J723. Additionally, cell proliferation of CAF, HUVEC, and CCA cells was determined in co-culture assays using transwell assays. Hanging droplet spheroids Using the hanging droplet technique, 3D spheroids were generated and monitored over a 48-hour period in the presence or absence of ML792 (1 μM) or SAMe (0.3 mM). Besides, spheroid growth was assessed in control and UBE2I-knockdown CCA cells. Further details are outlined in the Supplemental Material, http://links.lww.com/HEP/J723. Colony formation The ability of CCA (EGI1) cells to form colonies was examined in vitro. For this, CCA cells were suspended in agar-coated 6-well plates and cultured with either ML792 (1 μM), SAMe (0.3 mM), or their respective vehicles (DMSO and H 2 O) for 3 weeks. Additionally, the colony formation ability of CCA (EGI1) control and SUMOylation-deficient CCA cells was also examined. Further culturing details can be found in the Supplemental Material, http://links.lww.com/HEP/J723. Cell migration In vitro assessment of cell migration was conducted using “wound-healing assays,”as detailed in the Supplemental Material, http://links.lww.com/HEP/J723. CCA cell migration was monitored in the presence or absence of ML792 (0.1 μM) and SAMe (0.3 mM) at intervals of approximately 18 hours. Images were captured using a Leica DM IL LED microscope equipped with a DFC 3000 G Leica digital camera. The wound-healing area was subsequently quantified using Adobe Photoshop CS5 (Version 12.1) software, based in Berkeley, CA. 4 | HEPATOLOGY CRISPR/Cas9 UBE2I was knocked down using the CRISPR/Cas9 technique in CCA cells, as detailed in the Supplemental Material, http://links.lww.com/HEP/J723. The guide RNA sequences employed are listed in Supplemental Table S3, http://links.lww.com/HEP/J723. Immunoprecipitation Whole-cell lysates (1 mg) from NHC and CCA cells (EGI1, TFK1, and WITT) were incubated with Dynabeads Protein G (Invitrogen), crosslinked to SUMO1 or IgG antibodies (Abcam). Proteins were subsequently eluted from the beads using 2% SDS, as described in the Supplemental Material, http://links.lww.com/HEP/ J723. Mass spectrometry–based high throughput proteomic analysis Quantitative shotgun proteomic analyses were conducted to compare SUMO1-IP samples, CRISPR control, and UBE2I knockdown CCA (EGI1) cells, as well as the cells and supernatants of CCA cells-CAF cocultures as described in the Supplemental Material, http://links.lww.com/HEP/J723. Experimental animal models of CCA Xenograft human CCA models in immunodeficient CD-1 nude mice were used to investigate the role of SUMOylation (UBE2I-knockdown, ML792, or SAMe) in cholangiocarcinogenesis. Additionally, an oncogenedriven CCA model (Kras G12D ,Tp53 KO ) in immunocompetent FVB/N mice was employed to evaluate the effects of ML792 and SAMe. Detailed information can be found in the Supplemental Material, http://links.lww. com/HEP/J723. Statistical analysis Statistical analyses were performed using GraphPad Prism software (Version 6.01, San Diego, CA). After normality assessment using the Shapiro-Wilk normality test, parametric Student ttest or nonparametric Mann-Whitney tests were applied to determine statistical significance between the 2 groups. For comparisons involving more than 2 groups, one-way ANOVA or Kruskal-Wallis tests were used. The data are presented as means ±SEM, and differences of p<0.05 were considered statistically significant. RESULTS The SUMOylation pathway is hyperactivated in CCA To investigate potential alterations in SUMOylation within CCA, we examined the expression levels of essential components of the SUMOylation pathway— including the activating enzymes SAE1 and UBA2, the conjugating enzyme UBE2I, the ligating enzyme PIAS1, and the ligands SUMO1-3—in resected human CCA tissues and control tissues across four distinct patient cohorts: Copenhagen, The Cancer Genome Atlas, TIGER-LC, and San Sebastian. Increased expression of SAE1,UBA2,UBE2I, and PIAS1 was observed in tumor samples across all independent cohorts, compared to controls consisting of surrounding liver tissue and normal bile ducts (Figure 1A, Supplemental Figure S1, http://links.lww.com/HEP/J724). On the other hand, the expression of SUMO1-3 ligands varied between the different cohorts (Supplemental Figure S2, http://links. lww.com/HEP/J724). Importantly, increased expression of SAE1 and UBE2I was evident in resected CCA tissues regardless of the underlying mutation status (IDH1,KRAS,TP53, or wild-type (WT) for these mutations, referred to as Udt) (Figure 1B) or their anatomical origin (Supplemental Figure S3, http://links. lww.com/HEP/J724), indicating that the upregulation of the SUMOylation machinery is a common feature in CCA. Of note, within the Copenhagen cohort, patients with high tumor levels of SAE1 exhibited increased tumor recurrence after surgical resection compared to patients with low SAE1 levels (Figure 1C), and poorly differentiated tumors were characterized by higher UBE2I expression compared to well and moderately differentiated ones (Figure 1D). Immunostaining confirmed enrichment of SUMO1 (ie, a combination of free SUMO1 ligand and SUMO1-conjugated proteins) in the CCA tumors of patients compared to the biliary epithelium of both non-tumor surrounding liver samples and normal human liver tissues (Figure 2A). To functionally investigate the role of SUMOylation and confirm that changes in this PTM are specifically present in cancer cells within the tumor, we used four distinct human CCA cell lines (EGI1, HUCCT1, TFK1, and WITT). Compared to primary cultured NHC, SAE1 and UBE2I were overexpressed in all CCA cell lines (Figure 2B). Consistent with these observations in patient tissues, the expression of SUMO1-3 ligands varied more between CCA cell lines and NHC (Supplemental Figure S4A, http://links.lww.com/HEP/J724). In accordance with increased expression of SUMOylation components, we identified an increase in the levels of SUMO1-conjugated proteins across all CCA cell lines compared to NHC by immunoblotting (Figure 2C), and a prominent decrease of free SUMO1 levels (Figure 2C), indicating increased SUMOylation activity. TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 5 0Copenhagen 200 400 SAE1 mRNA expression (%) 600 0Copenhagen 200 100 300 400 UBE2I mRNA expression (%) 500 0 NBD SL KRASmut TP53mut UdtIDHmut 200 100 SAE1 mRNA expression (%) 300 0 NBD SL KRASmut TP53mut UdtIDHmut 200 100 UBE2I mRNA expression (%) 500 400 300 0 Well Moderate Poor UBE2I mRNA expression (%) 1500 1000 500 0 050 Time (months) Tumor differentiation p<0.05 100 Low SAE1 High SAE1 150 50 Recurrence-free survival (%) 100 (A) (B) (C) (D) **** *** **** **** **** * ** ** ** ** * *** *** * ** *** *** ** 0TCGA 200 100 300 400 Normal bile ducts (NBD) Surrounding liver (SL) CCA 0TIGER-LC 120 80 20 100 140 160 0San Sebastian 200 100 300 500 400 0TCGA 200 100 300 400 0TIGER-LC 80 20 100 120 0San Sebastian 200 100 300 500 400 FIGURE 1 Expression levels of the SUMOylation pathway enzymes in human CCA tissue. (A) SAE1 and UBE2I mRNA expression in CCA tumors compared to NBDs and/or SL human tissue from the Copenhagen, TCGA, TIGER-LC, and San Sebastian cohorts of patients. (B) SAE1 and UBE2I mRNA expression in IDH1,KRAS,TP53 mutant CCA tissues or wild-type status for all 3 genes (Udt) (n =104) compared to SL human tissue (n =132) and to NBD (n =6) (Copenhagen cohort). (C) Kaplan-Meier analysis of SAE1 mRNA expression in CCA tumors (Copenhagen cohort). (D) UBE2I mRNA expression in CCA tumors from the Copenhagen cohort grouped by tumor differentiation grade. ∗ p<0.05; ∗∗ p<0.01; ∗∗∗ p<0.001 (Kruskal-Wallis, Mann-Whitney, and Student ttests). Abbreviations: CCA, cholangiocarcinoma; iCCA, intrahepatic cholangiocarcinoma; NBDs, normal bile ducts; SL, surrounding liver; TCGA, The Cancer Genome Atlas; TIGER-LC, The Thailand Initiative in Genomics and Expression Research. 6 | HEPATOLOGY (A) (B) (D) (E) (F) (C) Non-tumor surroundingHealthy Human liver SUMO1 CCA 0 200 400 SAE1 mRNA expression (%) 800 600 ***** 0 10 20 SUMOylated proteins (fold change) 40 30 *** ** * * *** 0 EGI1 HUCCT1 TFK1 WITT 100 NHC CCA NHC SUMO1 β-actin CCA NHC CCA EGl1 SUMOylated proteins HUCCT1TFK1 WITT EGl1 Input IP Ctrl IgG Input IP Ctrl IgG Input IP Ctrl IgG Input IP Ctrl IgG TFK1 WITT 200 UBE2I mRNA expression (%) 400 300 NFĸB signaling 1 10 100 Survival Cell proliferation Response to ER stress DNA replication and cell cycle Translation Upregulated Protein folding * ** *** 0 EGI1 NHC HUCCT1 TFK1 WITT 0.5 Free SUMO1 (fold change) 1.5 1.0 SUMO1 SUMOylated proteins 0 log2 (CCA vs. NHC) -5 5 2 -log10 (p value) -log10 (p value) p=0.001 p=0.01 p=0.05 6 4 0 FIGURE 2 Abnormal expression of the SUMOylation pathway components in human CCA cells in vitro. (A) SUMO1 expression in human healthy, non-tumor SL and iCCA. Scale bar 50 μm. (B) SAE1 and UBE2I mRNA expression in NHC and CCA cell lines (n =6). (C) Representative immunoblot and quantification of SUMO1 and SUMO1-conjugated proteins in NHC and CCA cells (n =6). (D) Representative immunoblot of SUMO1-IP in NHC and CCA cells (n =3). (E) Volcano plot of all identified SUMO1-IP proteins (n =309) by mass spectrometry comparing fold enrichment in EGI1 CCA cells to NHC. (F) Proteomic analyses of significant identified proteins (n =78) between CCA cells (EGI1) and NHC by GO. ∗ p<0.05; ∗∗ p<0.01; ∗∗∗ p<0.001 (Student or Mann-Whitney ttests). Abbreviations: CCA, cholangiocarcinoma; Ctrl, control; ER, endoplasmic reticulum; GO, gene ontology; IP, immunoprecipitation; NHC, normal human cholangiocytes; SL, surrounding liver; SUMO, small ubiquitin-related modifier. TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 7 Elevated levels of SUMOylated proteins contribute to the progression of CCA To identify the SUMOylation targets that could be involved in CCA growth, we conducted IP of SUMO1conjugated proteins from both CCA cell lines and NHC, followed by proteomic analysis by mass spectrometry. Immunoblotting confirmed the enrichment of SUMO1-IP proteins in CCA cell cultures compared to controls (inputs or control IgG) (Figure 2D). Shotgun proteomic analysis revealed a total of 128, 30, and 14 differentially SUMOylated proteins between CCA cell lines (EGI1, TFK1, and WITT) and NHC, respectively. Among these, the levels of 3 proteins (RANGAP1, PRDX1, and DDX5), all known SUMOylation targets, were increased in all CCA cell lines. Notably, the majority of the differentially expressed SUMO1-conjugated proteins identified were found to be increased (highlighted in red) [99.8%, 90%, and 64% in EGI1, TFK1, and WITT, respectively] compared to NHC (Figure 2E representing results for the EGI1 cell line, Supplemental Figure S4B, C, http://links.lww.com/HEP/J724). Gene ontology analysis of the upregulated SUMOylated proteins in CCA (EGI1) cells highlighted their involvement in various cellular processes, including protein translation (eg, ABCF1, RS27, and RL23A) and folding (eg, HSPB1, TRAP1, and TCPH), DNA replication and cell cycle (eg, CDK1, DDX5, MCM6, and MCM7), DNA repair (eg, SSRP1 and RUBV1), and response to stress (eg, PRDX1 and GSTP1) (Figure 2F). Interestingly, PIGR, a potential diagnostic marker for CCA,[18]displayed increased SUMOylation in CCA cells compared to NHC. Additionally, the STAT3 transcription factor exhibited predominant SUMOylation in CCA (EGI1) cells compared to NHC. It is worth noting that IL-6 and Wnt3a ligands, previously associated with cholangiocarcinogenesis,[1]were discovered to stimulate protein SUMOylation in CCA cells (Supplemental Figure S4D, E, http://links.lww.com/HEP/J724). CRISPR/Cas9-mediated UBE2I knockdown demonstrates the key role of SUMOylation in CCA progression both in vitro and in vivo Stable CRISPR/Cas9-mediated UBE2I-knockdown was performed in CCA (EGI1) cells to further prove the impact of SUMOylation in cholangiocarcinogenesis. Immunoblot analysis confirmed effective UBE2I silencing, as UBC9 protein levels were reduced by 65%–80% in 2 distinct CRISPR/Cas9-UBE2I CCA cell clones compared to CRISPR/Cas9 control cells (Figure 3A). This reduction in UBC9 protein expression was consistently reflected at the mRNA level, as UBE2Iknockdown CCA cells exhibited minimal UBE2I mRNA expression (Supplemental Figure S5A, http://links.lww. com/HEP/J724), accompanied by decreased levels of protein SUMOylation (Supplemental Figure S5B, http:// links.lww.com/HEP/J724). UBC9 depletion led to a reduction in cell viability, demonstrating an antiproliferative effect when compared to WT (CRISPR/Cas9 control) CCA cells (Figure 3B). This effect was mediated by inducing cell cycle arrest at the G 1 phase (Figure 3B). Additionally, UBE2I-knockdown CCA cells showed decreased baseline rates of apoptotic cell death (Figure 3C) but a diminished ability to form 3D hanging-spheroids or anchorage-independent colonies compared to WT CCA cells (Figure 3D). Moreover, CRISPR/Cas9-UBE2I CCA cells injected subcutaneously in immunodeficient mice exhibited reduced tumorigenic (Figure 3E) and proliferative (Supplemental Figure S6A, http://links.lww.com/HEP/J724) capacities compared to WT (CRISPR/Cas9 control) cells. Tumors derived from UBE2I-knockdown CCA cells also showed fewer α-SMA + and FAP + fibroblasts compared to those generated with WT CCA cells. However, no changes in tumor vasculature were observed (Supplemental Figure S6B, http://links.lww.com/HEP/J724). Importantly, these findings were consistently confirmed in 2 different UBE2I-knockdown clones (Figure 3 and Supplemental Figure S6, http://links.lww.com/HEP/J724). Shotgun proteomic analyses identified a total of 1572 and 1127 differentially expressed proteins between CRISPR/Cas9-UBE2I clones #1 and #2 in comparison to WT CCA cells (Supplemental Figure S5C, http://links. lww.com/HEP/J724). Among these proteins, 902 (57.4%) were upregulated (highlighted in red), while 670 (42.6%) were downregulated (indicated in blue) in CRISPR/Cas9-UBE2I #1 CCA cells in contrast to WT CCA cells. Similarly, 683 proteins (60.6%) were found to be upregulated (red), and 444 (39.4%) were downregulated (blue) in CRISPR/Cas9-UBE2I #2 CCA cells relative to WT CCA cells (Supplemental Figure S5C, http://links.lww.com/HEP/J724). Of note, both CRISPR/ Cas9-UBE2I clones shared 371 upregulated proteins and 233 downregulated proteins. Gene ontology analysis revealed that the commonly upregulated proteins in CRISPR/Cas9-UBE2I CCA clones participate in various biological processes such as protein folding (eg, FKBP5, FKBP9, and PPIB), proteasomal degradation (eg, BAG6, PSMA1, and SGT1), protein stabilization (eg, CDC37, HSPA1B, and CALR), antigen processing and presentation (eg, HLA-A, ERAP2, B2M), as well as fibroblast apoptosis (eg, BAK1 and CFDP1) (Supplemental Figure S5D, http://links.lww.com/HEP/J724). In contrast, the commonly downregulated proteins are associated with DNA replication (eg, MCM2, MCM6, and RFA1), cell proliferation (eg, SRC, AKT, and XRCC5), cell migration (eg, EGFR, CLDN1, and CARMIL1), redox homeostasis (eg, PRDX1, GSH1, and APEX1), angiogenesis (eg, ENG, AAMP, and ADAM9), TGFβsignaling (eg, TGFBR2, SMAD4, and FAK1), as well as PDGFR signaling (eg, BCAR 8 | HEPATOLOGY The SUMOylation status of CCA cells influences their interplay with the microenvironment Differential expression of proteins involved in angiogenesis, antigen processing, and presentation, as well as fibroblast-associated pathways (ie, TGFβand PDGFR signaling) in UBE2I-knockdown CCA cells, suggests the involvement of UBC9 and protein SUMOylation in shaping the interaction between transformed cholangiocytes and the TME. Additionally, treatment with SUMOylation inhibitors induced changes in CAFs, endothelial cells, and immune cell populations in both the xenograft and oncogene-driven CCA mouse models. This prompted us to investigate whether protein SUMOylation in CCA cells influences their interaction with various components of the TME. Co-culture in vitro experiments were conducted between CCA cells and previously isolated and characterized human CCA-derived CAFs or human endothelial cells (HUVECs) (Figure 8 and Supplemental Figure S12, http://links.lww.com/HEP/J724). Co-culturing WT CCA cells with CAFs increased CAF proliferation, an effect that was significantly attenuated when UBE2I-- knockdown CCA cells were used (Figure 8B). Proteomic analysis identified a total of 4179 proteins, of which 213 were differentially expressed between CAFs co-cultured with WT CCA cells and those co-cultured with SUMOylation-impaired CCA cells (Figure 8C). Among the 111 downregulated proteins in CAFs co-cultured with UBE2I CRISPR/Cas9 CCA cells (depicted in blue), we identified proteins involved in collagen synthesis and matrix remodeling (eg, COL1A1, COL1A2, COL3A1, and COL6A2), cell cycle regulation and proliferation (eg, NUCKS1, EB3, and STK3), and cell metabolism and energy production (eg, ACLY and CRAT). Additionally, among the 102 upregulated proteins in CAFs co-cultured with SUMOylation-impaired CCA cells (depicted in red), we identified SMAD2, a protein known to inhibit cell proliferation and protect against TGFβ-mediated fibrosis by counteracting TGFβ/Smad3 signaling.[19]Furthermore, the secretome of these CAFs was analyzed, revealing a total of 1222 proteins, 186 of which were differentially expressed between CAFs exposed to UBE2I-deficient CCA cells and those co-cultured with WT CCA cells (Figure 8D). Notably, upregulated proteins (depicted in red) were associated with biological processes such as signal transduction (eg, MET, which activates RAS-ERK, PI3 kinase-AKT, or PLC gamma-PKC), antigen processing and presentation (eg, B2M, a component of the class I MHC), and chemotaxis (eg, S100A4, MK). In contrast, downregulated proteins in the secretome of CAFs cocultured with UBE2I-deficient CCA cells (depicted in blue) were involved in processes such as angiogenesis (eg, TGFBI, MYDGF, and SERPINE1), cell migration (eg, MMP1, MMP2, LAMC1, LAMC2, and LAMB3), and extracellular matrix (eg, COL18A1, Perlecan, and FN1) (Figure 8D). On the other hand, co-culturing WT CCA cells with CAFs (Figure 8E) promoted increased cancer cell proliferation (Figure 8F). However, in UBE2I-knockdown CCA cells incubated with CAFs, this increased tumor cell proliferation was not observed (Figure 8F), highlighting the critical role of the SUMOylation status in cancer cells for responding to CAF-derived signals, as well as the necessity of reciprocal communication between cancer and stromal cells during tumors growth. To further investigate this interaction, we conducted proteomic analyses of WT or UBE2I-CRISPR/Cas9 CCA cells, as well as their secretomes, after incubation with CAFs (Figures 8G, H). In total, 1609 proteins were identified, of which 574 were differentially secreted by control versus UBE2I-deficient CCA cells (Figure 8H). Consistent with the proteomic data from UBE2I CRISPR/Cas9 CCA cells, their secretory profile revealed an upregulation of proteins (depicted in red) involved in antigen processing and presentation, inflammatory responses, and chemotaxis (Figure 8H). Angiogenic proteins were also found downregulated in SUMOylation-deficient CCA cells (Supplemental Figure S5D, http://links.lww.com/HEP/J724). In line with this, co-culturing control CCA cells with endothelial cells stimulated endothelial cell proliferation (Supplemental Figure S12, http://links.lww.com/HEP/J724). However, UBE2I-knockdown CCA cells failed to increase the proliferative capacity of endothelial cells (Supplemental Figure S12, http://links.lww.com/HEP/J724) and were also refractory to the proliferation-enhancing signals provided by endothelial cells (Supplemental Figure S12, http://links.lww.com/HEP/J724). DISCUSSION CCA is a highly aggressive malignancy affecting bile duct epithelial cells. Currently, effective therapies for CCA are lacking, with surgical resection remaining the sole potential curative option, despite high rates of disease recurrence. In light of this challenging scenario, it is crucial to comprehensively unravel the intricate molecular mechanisms driving CCA progression to identify novel therapeutic targets. In this study, we demonstrate that dysregulated protein SUMOylation contributes to the pathogenesis of CCA. Our data reveal elevated expression of SAE1 and UBE2I, core enzymes involved in the activation and conjugation steps of the SUMOylation pathway, respectively, in human CCA, leading to the SUMO1-conjugation of key oncogenic proteins in tumor epithelial cells. Previous research has suggested a role for SUMOylation in cholangiocarcinogenesis,[12,13]and our findings, derived from four independent international cohorts, support these observations. We show that UBE2I and SAE1 levels are elevated in biliary tumors, not only in comparison with surrounding tissue (predominantly TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 15 2 *** *** *** * *** * 150 CAF Biological processes Upregulated Phagocytosis Signal transduction Antigen presentation Receptor signaling Chemotaxis 01234 Secretome (CAF-UBE2IKD CCA vs. CAF-WT CCA) CAF proteome -log10 (p value) 111 NUCKS WT or UBE2lKD CCA cells WT or UBE2lKD CCA cells CAFs CAFs secretome secretome cell proteome cell proteome SMAD2 p=0.001 p=0.01 p=0.05 COL1A1 4 2 -4 -2 0 2 4 log2 (CAF-UBE2IKD CCA vs. CAF-WT CCA) 102 Proliferation (% to control media) 100 50 0 Co-culture –WT#1#2 UBE2lKD (A) (B) 150 CCA Ctrl #1 #2 Proliferation (% to Ctrl CCA cells) 100 50 0 CAF co-culture –+–+–+ (C) CAF secretome -log10 (p value) -log10 (p value) B2MG S100A4 p=0.001 p=0.01 p=0.05 FN1 MMP1 6 8 4 -4 -2 0 2 4 log2 (CAF-UBE2IKD CCA vs. CAF-WT CCA) (D) (E) (F) 107 79 CCA secretome -log10 (p value) HLAC p = 0.001 p = 0.01 p = 0.05 MCM7 8 10 4 2 6 -4 -2 0 2 4 log2 (UBE2IKD CCA-CAF vs. Ctrl CCA-CAF) (H) 299 275 CCA proteome -log10 (p value) p=0.01 p=0.001 p=0.05 AKT MCM2 10 15 -5 0 5 log2 (UBE2IKD CCA-CAF vs. Ctrl CCA-CAF) (G) 1438 1438 Downregulated Response to ER stress Angiogenesis Cell migration ECM organization ERK1/2 signaling 01234 -log10 (p value) Biological processes Upregulated Phagocytosis Cell-cell adhesion Inflammatory response Chemotaxis Apoptosis Antigen presentation 01234 Secretome (UBE2IKD CCA-CAF vs. WT CCA-CAF) -log10 (p value) Downregulated Translation Glycolysis Cell division RNA splicing 02 64810 -log10 (p value) 2 FIGURE 8 Differential protein expression in the secretome and cells in the co-culture of CAFs with UBE2I-deficient or WT CCA cells. (A) Schematic representation of CAF co-culture experimental design with WT or UBE2I-knockdown CCA cells. (B) Proliferation of CAFs co-cultured with WT or UBE2I-knockdown CCA cells (n =3). Volcano plot of all identified proteins by mass spectrometry, comparing fold enrichment in (C) CAFs and in (D) the secretome of CAFs co-cultured with UBE2I-knockdown (n =6) or with WT CCA cells (n =6), with the respective GO analysis of differentially expressed proteins. (E) Schematic representation of CAFs co-culture experimental design with WT or UBE2I-knockdown CCA cells. (F) Proliferation of WT and UBE2I-knockdown CCA clones in the presence or absence of CAFs (n =3). Volcano plot of all identified proteins by mass spectrometry, comparing fold enrichment in (G) UBE2I-knockdown CCA cells (n =5) to WT CCA cells (n =6) and (H) the secretome of UBE2I-knockdown CCA cells or WT CCA cells co-cultured with CAFs (n =4), and the respective GO analysis of differentially expressed proteins. ∗ p<0.05, ∗∗∗ p<0.001 (1-way ANOVA tests). Abbreviations: CAFs, cancer-associated fibroblasts; CCA, cholangiocarcinoma; GO, gene ontology; WT, wild type. 16 | HEPATOLOGY hepatocytes) but also with normal bile ducts. For the first time, we also demonstrate the upregulation of these enzymes in human CCA cells, compared to NHCs, accompanied by a global increase in SUMOylated protein levels in CCA. Furthermore, our results show that elevated SAE1 and UBE2I expression is a common feature of CCA tissues, regardless of the underlying driver mutations or anatomical origin. This strongly suggests that increased protein SUMOylation is a general event in cholangiocarcinogenesis, independent of the tumor’s genetic background. Interestingly, poorly differentiated CCA tumors, which are often associated with unfavorable outcomes, exhibited higher expression of the SUMOylation pathway conjugating enzyme UBE2I. Meanwhile, tumors with higher expression of the activator SAE1 were also linked to worse recurrence-free survival. These associations with differentiation and survival outcomes highlight the potential prognostic and clinical relevance of our findings. Through unbiased proteomic analysis of SUMO1conjugated proteins, we identified several SUMOylation targets with increased abundance in CCA, including CDK1, a key regulator of the cell cycle known to undergo posttranscriptional modification via NEDDylation. The potential impact of CDK1 SUMOylation on cell cycle regulation and tumor progression remains unexplored. However, SUMOylation inhibition reduces CDK1 expression and induces cell cycle arrest, implicating CDK1 SUMOylation in controlling cell cycle dynamics and tumor development. Notably, previous studies have reported increased nuclear accumulation of p27kip1 in human CCA cells following SUMOylation inhibition, which was linked to cell cycle arrest and enhanced chemosensitivity, further supporting the relevance of protein SUMOylation in CCA.[12,13] Additionally, we observed elevated levels of SUMOylated PRDX1 in CCA cells compared to NHC. PRDX1 is known to protect cells against oxidative stress by detoxifying peroxides. Pharmacological inhibition of SUMOylation resulted in increased ROS levels, likely due to disrupted SUMOylation of PRDX1 and other proteins, which may contribute to the enhanced apoptotic cell death observed upon treatment. We also found that SUMOylated STAT3 and SUMOylated β-catenin were upregulated in CCA cells compared to NHC. Although previous research has indicated the possibility of SUMOylation at lysine 451 of STAT3,[20] its relevance in CCA has yet to be elucidated. CCAs often arise in the context of inflammation, where proinflammatory and regulatory molecules such as IL-6 and Wnt ligands promote tumorigenesis by activating the JAK/STAT3 and Wnt/β-catenin signaling pathways. Interestingly, we have observed an increase of SUMO1-conjugated proteins in CCA cells following stimulation with IL-6 and Wnt3a, suggesting a potential role for SUMOylation in the activation of these pathways. After establishing that differentially SUMOylated proteins regulate CCA cell survival and proliferation, we further investigated the relevance of the SUMOylation machinery in CCA cells through genetic and pharmacological inhibition. Disrupting UBE2I in CCA cells using CRISPR/Cas9 led to strong antitumor effects, both in vitro and in vivo, confirming previous findings from acute silencing of this gene.[12]In fact, SUMOylation has been suggested to play a key role in CCA, as Circ-RAPGEF5 was shown to promote cholangiocarcinogenesis by stabilizing SAE1.[13]Next, we tested the therapeutic potential of two SUMOylation pathway inhibitors, ML792 and SAMe. Both compounds not only reduced SUMOylated protein levels in CCA cells but also induced apoptosis and inhibited proliferation by causing cell cycle arrest at the G2/M phase, along with a reduction in colony formation. Importantly, neither ML792 nor SAMe caused harmful effects on NHCs, demonstrating their selective toxicity toward CCA cells due to their upregulated SUMOylation machinery. These effects collectively contributed to a substantial decrease in CCA tumorigenicity, with both ML792 and SAMe effectively halting tumor growth in various in vivo contexts. Additionally, targeting SUMOylation influenced the TME, reducing the levels of CAFs and endothelial cells and promoting an antitumor immune response. This highlights the potential of SUMOylation inhibition not only to target malignant cells but also to disrupt other key components of the TME. In this regard, other inhibitors of SAE1, such as ML93 and TAK-981, have been developed. These small-molecule inhibitors have shown robust and selective activity in targeting SAE1 and the SUMOylation pathway in both in vitro and in vivo models. Notably, ongoing clinical trials are currently exploring the potential of TAK-981 in various advanced solid tumors or cases of relapsed/ refractory liquid tumors (NCT04776018, NCT04065555, NCT04074330, NCT04381650, and NCT03648372) and some of these are exploring the potential combination of TAK-981 with immunotherapy. Herein, our findings support the potential of targeting SUMOylation as an immune-oncology strategy for CCA tumors.[14–16] Targeted therapies have transformed personalized medicine for CCA. However, only a small proportion of patients exhibit actionable genetic alterations, and even then, their responses are typically partial and modest. Our data reveal a global increase in protein SUMOylation in human CCA, which may carry significant prognostic and therapeutic implications. Since elevated protein SUMOylation appears to be a common and widespread event during cholangiocarcinogenesis, targeting this pathway presents a promising therapeutic approach for all CCA patients, regardless of their specific genetic alterations. Furthermore, the therapeutic potential of inhibiting the SUMOylation pathway goes beyond merely reducing the malignancy of transformed cholangiocytes. It also plays a critical role in disrupting the recruitment of stromal cells and the modulation of TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 17 the TME. Thus, inhibiting protein SUMOylation holds great promise as a novel treatment strategy for CCA, offering the potential to significantly improve patient outcomes and overall well-being. AUTHOR CONTRIBUTIONS Paula Olaizola, Irene Olaizola, Ainhoa Lapitz, Beatriz Val, Maite G. Fernandez-Barrena, Colm J. O’Rourke, Pui Y. Lee-Law, Rocio I. R. Macias, Jose J. G. Marin, Maria L. Martinez-Chantar, Matias A. Avila, Patricia Aspichueta, Jesper B. Andersen, Luke Boulter, Luis Bujanda, Pedro M. Rodrigues, Maria J. Perugorria, Jesus M. Banales, and Ioana Riaño: study concept and design, analysis and interpretation of data, drafting of the manuscript. Paula Olaizola, Irene Olaizola, Marta Fernandez de Ara, Ainhoa Lapitz, Beatriz Val, Laura Izquierdo-Sanchez, Maite G. Fernandez-Barrena, Laura Alvarez, Colm J. O’Rourke, Pui Y. Lee-Law, Kyle Davies, Andreea Gradinaru, Raul Jimenez-Agüero, Ainhoa Lapitz, Pedro M. Rodrigues, and Jesus M. Banales: data acquisition. Paula Olaizola, Irene Olaizola, Ainhoa Lapitz, Beatriz Val, Colm J. O’Rourke, Pedro M. Rodrigues, Jesus M. Banales: statistical analysis. Jose J. G. Marin, Rocio I. R. Macias, Maria L. Martinez-Chantar, Matias A. Avila, Luis Bujanda, Maria J. Perugorria, and Jesus M. Banales: obtained funding. FUNDING INFORMATION Spanish Carlos III Health Institute (ISCIII) [Jesus M. Banales (FIS PI21/00922, PI18/01075 and Miguel Servet CPII19/00008); Maria J. Perugorria (FIS, PI17/00022, PI20/00186); Pedro M. Rodrigues (Sara Borrell CD19/ 00254); Jesus M. Banales and Patricia Aspichueta (PMP21/00080)] cofinanced by “Fondo Europeo de Desarrollo Regional”(FEDER); Spanish Ministry of Economy and Competitiveness (MINECO: “Ramón y Cajal”Program RYC-2015-17755 to Maria J. Perugorria); CIBERehd (ISCIII): Jesus M. Banales, Pedro M. Rodrigues, Maria J. Perugorria, Maria L. MartinezChantar, Matias A. Avila, Maite G. Fernandez-Barrena, Jose J. G. Marin, Patricia Aspichueta, Paula Olaizola, Laura Izquierdo-Sanchez, Ainhoa Lapitz, and Luke Boulter, Spain; “Diputación Foral Gipuzkoa”(2020CIEN-000067-01 and 2021-CIEN-000029-04-01 to Pedro M. Rodrigues), Spain; Department of Health of the Basque Country (2019111024 to Maria J. Perugorria, 2017111010 to Jesus M. Banales, 2020111077 and 2021111021 to Jesus M. Banales and Patricia Aspichueta), “Euskadi RIS3”(2019222054, 2020333010, 2021333003 to Jesus M. Banales), BIOEF (Basque Foundation for Innovation and Health Research: EiTB Maratoia BIO15/CA/016/BD to Jesus M. Banales, Maria L. Martinez-Chantar, MA), and Elkartek (KK-2020/00008 to Jesus M. Banales and Maria L. Martinez-Chantar), Spain; La Caixa Scientific Foundation (Jesus M. Banales and Maria L. Martinez-Chantar HR17-00601), Spain; “Fundación Científica de la Asociación Española Contra el Cáncer”(AECC Scientific Foundation: “Rare cancers call 2017”to Jesus M. Banales, Maria L. MartinezChantar, Matias A. Avila, Jose J. G. Marin, “AECC Lab call 2023”to Pedro M. Rodrigues, and AECC 2024 to Jesus M. Banales), Spain; AMMF-The Cholangiocarcinoma Charity (EU/2019/AMMFt/001, to Jesus M. Banales and Pedro M. Rodrigues), United Kingdom, Fondo de Investigaciones Sanitarias, Instituto de Salud Carlos III, Spain, co-funded by the European Regional Development Fund/European Social Fund, “Investing in your future”(PI22/00526 to Jose J. G. Marin, and PI23/ 00681 for Rocio I.R. Macias); “Junta de Castilla y Leon” (SA113P23) (to Jose J. G. Marin). Paula Olaizola received funds from the Basque Government (PRE_2016_1_0269, POS_2022_1_0044), the Spanish Association for the Study of the Liver (Asociación Española para el Estudio del Hígado, Juan Rodés) and UKRI (EP/Y028546/1), PYL-L from the European Association for the Study of the Liver (EASL; Sheila Sherlock Award 2017), IO from Basque Government (PRE_2019_1_0197) and MINECO (FPU 19/03327), Ainhoa Lapitz from the Basque Government (PRE_2017_1_ 0345). The funding sources had no involvement in study design, data collection and analysis, decision to publish, or article preparation. This article is based upon work from the COST Actions CA18122 European Cholangiocarcinoma Network (Euro-Cholangio-Net) and CA22125 Precision Medicine in Biliary Tract Cancer (Precision-BTC-Network) supported by COST (European Cooperation in Science and Technology: http://www.cost.eu). ACKNOWLEDGMENTS The authors thank Gnosis S.p.A. for supplying the SAMe for the in vitro and in vivo experiments and Takeda Pharmaceutical Company Limited for supplying the ML792 for the in vitro experiments. The authors thank the Proteomics Platform at CIC bioGUNE for the services. CONFLICTS OF INTEREST The authors have no conflicts to report. ORCID Paula Olaizola https://orcid.org/0000–0002–1358– 5429 Irene Olaizola https://orcid.org/0000–0002–5634– 8909 Ainhoa Lapitz https://orcid.org/0000–0002–7928– 5760 Maite G. Fernandez-Barrena https://orcid.org/0000– 0003–0375–6236 Colm J. O’Rourke https://orcid.org/0000–0002–2224– 2663 Pui Y. Lee-Law https://orcid.org/0000–0001–7639– 9765 18 | HEPATOLOGY Andreea Gradinaru https://orcid.org/0009–0002– 0451–3615 Ioana Riaño https://orcid.org/0000–0001–8410–2706 Rocio I.R. Macias https://orcid.org/0000–0002–4748– 0326 Jose J. G. Marin https://orcid.org/0000–0003–1186– 6849 Maria L. Martinez-Chantar https://orcid.org/0000– 0002–6446–9911 Matias A. Avila https://orcid.org/0000–0001–6570– 3557 Patricia Aspichueta https://orcid.org/0000–0002– 3553–1755 Jesper B. Andersen https://orcid.org/0000–0003– 1760–5244 Luke Boulter https://orcid.org/0000–0002–7954–6705 Luis Bujanda https://orcid.org/0000–0002–4353– 9968 Pedro M. Rodrigues https://orcid.org/0000–0001– 6193–7436 Jesus M. Banales https://orcid.org/0000–0002–5224– 2373 REFERENCES Bold references with shared authorship. 1. Banales JM, Marin JJG, Lamarca A, Rodrigues PM, Khan SA, Roberts LR, et al. Cholangiocarcinoma 2020: The next horizon in mechanisms and management. Nat Rev Gastroenterol Hepatol. 2020;17:557–88. 2. Izquierdo-Sanchez L, Lamarca A, La Casta A, Buettner S, Utpatel K, Klümpen HJ, et al. Cholangiocarcinoma landscape in Europe: Diagnostic, prognostic and therapeutic insights from the ENSCCA Registry. J Hepatol. 2022;76:1109–21. 3. Vogel A, Bridgewater J, Edeline J, Kelley RK, Klümpen HJ, Malka D, et al. Biliary tract cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2023;34:127–40. 4. Olaizola P, Lee-Law PY, Fernandez-Barrena MG, Alvarez L, Cadamuro M, Azkargorta M, et al. Targeting NAE1-mediated protein hyper-NEDDylation halts cholangiocarcinogenesis and impacts on tumor-stroma crosstalk in experimental models. J Hepatol. 2022;77:177–90. 5. Seeler JS, Dejean A. SUMO and the robustness of cancer. Nature Reviews Cancer. 2017;17:184–97. 6. Saitoh H, Hinchey J. Functional heterogeneity of small ubiquitinrelated protein modifiers SUMO-1 versus SUMO-2/3. J Biol Chem. 2000;275:6252–8. 7. Lee-Law PY,Olaizola P, Caballero-Camino FJ, IzquierdoSanchez L, Rodrigues PM, Santos-Laso A, et al. Targeting UBC9-mediated protein hyper-SUMOylation in cystic cholangiocytes halts polycystic liver disease in experimental models. J Hepatol. 2021;74:394–406. 8. Floris A, Chandla S, Lim Y, Barbier-Torres L, Seth K, Khangholi A, et al. Sumoylation of methionine adenosyltransferase alpha 1 promotes mitochondrial dysfunction in alcohol-associated liver disease. Hepatology. 2024;80:102–18. 9. Stein S,Lemos V, Xu P, Demagny H, Wang X, Ryu D, et al. Impaired SUMOylation of nuclear receptor LRH-1 promotes nonalcoholic fatty liver disease. J Clin Invest. 2017;127: 583–92. 10. Tomasi ML, Tomasi I, Ramani K, Pascale RM, Xu J, Giordano P, et al. S-adenosyl methionine regulates ubiquitin-conjugating enzyme 9 protein expression and sumoylation in murine liver and human cancers. Hepatology. 2012;56:982–93. 11. Zubiete-Franco I, García-Rodríguez JL, Lopitz-Otsoa F, SerranoMacia M, Simon J, Fernández-Tussy P, et al. SUMOylation regulates LKB1 localization and its oncogenic activity in liver cancer. EBioMedicine. 2019;40:406–21. 12. Huang J, Tan X, Liu Y, Jiang K, Luo J. Knockdown of UBE2I inhibits tumorigenesis and enhances chemosensitivity of cholangiocarcinoma via modulating p27kip1 nuclear export. Mol Carcinog. 2023;62:700–15. 13. Zheng J,Wang Y,Tao L, Cai J, Shen Z, Liu Y, et al. CircRAPGEF5 promotes intrahepatic cholangiocarcinoma progression by stabilizing SAE1 to facilitate SUMOylation. J Exp Clin Cancer Res. 2023;42:239. 14. Chen Y. A new immuno-oncology target —SUMOylation. Trends Cancer. 2023;9:606–8. 15. Goel S, Ulahannan SV, Olszanski AJ, LoRusso P, Sanborn RE, Sharma S, et al. A phase 1b, multicenter, dose-escalation study of subasumstat (TAK-981) in combination with pembrolizumab in patients (pts) with advanced solid tumors. J Clin Oncol. 2022;40: 2506. 16. Dudek A, Juric D, Dowlati A, Vaishampayan U, Assad H, Rodón J, et al. 476 First-in-human phase 1/2 study of the first-in-class SUMO-activating enzyme inhibitor TAK-981 in patients with advanced or metastatic solid tumors or relapsed/refractory lymphoma: Phase 1 results. J Immunother Cancer. 2021;9: A505–6. 17. Holt AP, Haughton EL, Lalor PF, Filer A, Buckley CD, Adams DH. Liver myofibroblasts regulate infiltration and positioning of lymphocytes in human liver. Gastroenterology. 2009;136: 705–14. 18. Lapitz A, Azkargorta M, Milkiewicz P, Olaizola P, Zhuravleva E, Grimsrud MM, et al. Liquid biopsy-based protein biomarkers for risk prediction, early diagnosis, and prognostication of cholangiocarcinoma. J Hepatol. 2023;79:93–108. 19. Meng XM, Huang XR, Chung ACK, Qin W, Shao X, Igarashi P, et al. Smad2 protects against TGF-β/Smad3-mediated renal fibrosis. J Am Soc Nephrol. 2010;21:1477. 20. Zhou Z, Wang M, Li J, Xiao M, Chin YE, Cheng J, et al. SUMOylation and SENP3 regulate STAT3 activation in head and neck cancer. Oncogene. 2016;35:5826–38. How to cite this article: Olaizola I, Lapitz A, Val B, Izquierdo-Sanchez L, Fernandez-Barrena MG, O’Rourke CJ, et al. Targeting protein hyperSUMOylation halts cholangiocarcinoma progression by impairing cancer cell viability and tumorstroma crosstalk. Hepatology. 2025;■■:■■–■■.https://doi.org/10.1097/ HEP.0000000000001259 TARGETING PROTEIN HYPER-SUMOYLATION HALTS CHOLANGIOCARCINOMA | 19