Advertisement

Treatment of HCC with Claudin-1 specific antibodies suppresses carcinogenic signaling and reprograms the tumor microenvironment

Open AccessPublished:October 26, 2022DOI:https://doi.org/10.1016/j.jhep.2022.10.011

      Highlights

      • CLDN1 is upregulated in HCC particularly at non-junctional localization and is associated with a tumor subtype, characterized by stemness and an immune-low tumor microenvironment.
      • Non-junctional CLDN1-targeting monoclonal antibodies suppress tumor growth and invasion in patient derived cell-based and complementary in vivo models.
      • Mechanistic studies suggest non-junctional CLDN1-targeting monoclonal antibodies to inhibit tumor stemness, metabolism and to reprogram the tumor immune microenvironment by perturbing interactions of CLDN1 with signaling proteins, including Notch ligand JAG1.
      • Robust pre-clinical proof-of-concept for CLDN1 mAbs as potential first in-class compound with a perspective to break the plateau of limited treatment response in advanced HCC

      Abstract

      Background and Aims

      Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death worldwide. Despite new treatment approvals, treatment response and prognosis of patients with advanced HCC remain poor. Claudin-1 (CLDN1) is a membrane protein expressed not only at tight junctions but also non-junctionally such as the basolateral membrane of the human hepatocyte. While CLDN1 within tight junctions is well characterized, the role of non-junctional CLDN1 and its role as a therapeutic target in HCC remains unexplored.

      Methods

      Using humanized monoclonal antibodies (mAbs) targeting specifically the extracellular loop of human non-junctional CLDN1 and a large series of patient-derived cell-based and animal model systems we aimed to investigate the role of CLDN1 as a therapeutic target for HCC.

      Results

      Targeting non-junctional CLDN1 markedly suppressed tumor growth and invasion in cell line-based models of HCC and patient-derived 3D ex vivo models. Moreover, the robust effect on tumor growth was confirmed in vivo in a large series of cell line-derived xenograft (CDX) and patient-derived xenograft (PDX) mouse models. Mechanistic studies including single-cell RNA sequencing of multicellular patient HCC tumorspheres suggested that CLDN1 regulates tumor stemness, metabolism, oncogenic signaling and perturbs the tumor immune microenvironment.

      Conclusions

      Our results provide the rationale for targeting CLDN1 in HCC and pave the way to novel therapeutic interventions with CLDN1 mAbs aimed at improving the limited efficacy of current therapies.

      Impact and Implications

      Hepatocellular carcinoma (HCC) is a cancer with high mortality and unsatisfactory treatment options. Here we identified the cell surface protein Claudin-1 as a target for treatment of advanced HCC. Monoclonal antibodies targeting Claudin-1 inhibit tumor growth in patient-derived ex vivo and in vivo models by modulating signaling, cell stemness and the tumor immune microenvironment. Given the differentiated mechanism of action, the identification of Claudin-1 as a novel therapeutic target for HCC provides an opportunity to break the plateau of limited treatment response. These results of this preclinical study pave the way for the clinical development of Claudin-1 specific antibodies for treatment of HCC in patients. It is therefore of key impact for physicians, scientists and drug developers in the field of liver cancer and GI oncology.

      Graphical abstract

      Keywords

      Conflict of interest statement

      Inserm, the University of Strasbourg and the Institut Hospitalo-Universitaire have filed patent applications for the use of anti-claudin-1 monoclonal antibodies for the treatment of liver disease, NASH and HCC (PCT/EP2016/055942, PCT/EP2017/056703) which have been licensed to Alentis Therapeutics Basel. T.F.B. is a founder, shareholder and advisor for Alentis. A.T., R.I., G.E., H.E.S. GT and M.M. are employees of Alentis Therapeutics. Y. H., T. S. and C. S. are shareholders of Alentis Therapeutics.

      Data availability statement

      Transcriptomic data acquired and reported in this paper has been deposited on GEO (GSE196393). All other data associated with this paper are available upon request to the corresponding author.

      Financial support statement

      The authors acknowledge the following financial support: European Research Council Grant ERC-AdG-2014 HEPCIR (T.F.B. and Y.H.); European Research Council Grant ERC-AdG-2020 FIBCAN (T.F.B. and Y.H.); European Research Council Grant ERC-PoC-2016 PRELICAN (T.F.B.); European Research Council Grant ERC-PoC-2018 HEPCAN (T.F.B.); European Research Council Consolidator grant HepatoMetabopath (M.H.); ARC Grant TheraHCC2.0 IHUARC IHU201301187 (T.F.B.); ANRS Grant ECTZ103701 (T.F.B.); SATT Conectus, University of Strasbourg (CANCLAU) (T.F.B.); French National Research Agency RHU DELIVER (ANR-21-RHUS-0001) and LABEX ANR-10-LABX-0028_HEPSYS (T.F.B.); Grand-Est Region (M.M. and N.A.); German Research Foundation (DFG) RO 5983/1-1 (N.R.). This work of the Interdisciplinary Thematic Institute IMCBio, as part of the ITI 2021-2028 program of the University of Strasbourg, CNRS and Inserm, was further supported by IdEx Unistra (ANR-10-IDEX-0002), and by SFRI-STRAT’US project (ANR 20-SFRI-0012) and EUR IMCBio (ANR-17-EURE-0023) under the framework of the French Investments for the Future Program and the France 2030 program.

      Author contributions

      T.F.B. initiated and led the study. N.R., S.C, F.H.D, E.C., C.S., L.M., J.L., B.N., M.B.Z., S. M., P.L. and T.F.B. designed experiments and analyzed data. N.R., S.C., C.T., F.H.T.D., F.D.Z., S.C.D., S.B., C.P., M.F., T.R., Z.N., A.L., M.A.O. and N.A. performed experiments and analyzed data. G. E. engineered humanized mAbs. M.M., R.M., L.M., M.F.V and N.Br. performed and analyzed animal experiments. M.M. and P.M. performed and analyzed PET-Scan experiments. R.I., T.S., G.E. and M.M. designed PDX mouse model studies. H.E.S. performed gene expression analyses. A.S., S.M. and G.T. designed or performed immunohistochemistry analyses. F.J. performed computational analyses. M.H., I.D., A.T., Y.H. and N.Ba. gave critical conceptual input. E.F. and P.P. recruited and prepared patient liver tissues for ex vivo experiments. N.R., E.C., S.C., M.M. and N.A. designed figures and tables. N.R., J.L. and T.F.B. wrote the manuscript. All authors read and approved the final manuscript to be submitted.

      Introduction

      Hepatocellular carcinoma (HCC) is a major public health burden and the fourth leading and fast rising cause of cancer-related death worldwide
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      . HCC typically develops on the background of advanced liver fibrosis caused by viral or metabolic injury. Irrespective of the etiology, hyperactivation of oncogenic signaling such as the Ras/Raf/MAPK, PI3K/AKT/–mTOR, Notch and Wnt/β-catenin pathways are common events involved in HCC initiation and progression. Moreover, the tumor microenvironment (TME) plays a key role for HCC outcome
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      .
      Current treatment options for advanced HCC are still unsatisfactory due to limited response rates
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      ,
      • Finn R.S.
      • Qin S.
      • Ikeda M.
      • Galle P.R.
      • Ducreux M.
      • Kim T.-Y.
      • et al.
      Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma.
      . Therapeutic resistance to current systemic therapies has been associated with tumor cell plasticity, such as epithelial-mesenchymal transition (EMT) and stemness, as well as an immune-exhausted or immune-excluded TME
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      ,
      • Calderaro J.
      • Ziol M.
      • Paradis V.
      • Zucman-Rossi J.
      Molecular and histological correlations in liver cancer.
      ,
      • Qin S.
      • Jiang J.
      • Lu Y.
      • Nice E.C.
      • Huang C.
      • Zhang J.
      • et al.
      Emerging role of tumor cell plasticity in modifying therapeutic response.
      . Persistence of pro-tumorigenic signals within the fibrotic niche on the other hand accounts for high risk of tumor recurrence following curative treatment approaches
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      . Collectively, there is an urgent unmet medical need for novel HCC therapeutics that address the drawbacks of drug resistance and tumor recurrence.
      Claudin-1 (CLDN1) is a transmembrane protein expressed in tight junctions (TJs) as well as outside of the TJs (non-junctional CLDN1=NJ-CLDN1), e.g. at the basolateral membrane of the human hepatocyte. Interestingly, NJ-CLDN1 serves as a cell entry factor and signal transducer of hepatitis C virus (HCV)
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      , a major cause of HCC worldwide. Using humanized monoclonal antibodies (mAb) that selectively targets NJ-CLDN1, we aimed to study the role of NJ-CLDN1 in tumor progression, plasticity, metabolism and signaling.

      Material and METHODS

      Human subjects and patient cohorts

      Human liver tissue samples for ex vivo perturbation studies were derived from patients with liver resections at the Pôle Hépato-digestif, Strasbourg University Hospitals (2014-2021). All patients provided a written informed consent within the ethical principles of the declaration of Helsinki, approved by the local and national ethics committee (comités de protection des personnes, protocol RIPH2, LivMod-IDRCB 2019-A00738-49, ClinicalTrial NCT04690972). Sampling of human liver tissue for immunostaining was conducted by Charles River Laboratories and US Biomax Inc. with informed consent of all donors and under HIPPA approved protocols. Demographic data and clinical characteristics of patients enrolled are summarized in Suppl. Table 1-3.

      Mouse models

      Hepa 1.6 mouse model: female 8-10 weeks old C57BL/6J mice were used for in vivo CLDN1 gain-of-function study using engineered Hepa 1.6 cell line. Experiments were performed at WuXi Apptec, China, in accordance with the local regulations. CDX mouse models: female and male 7-10 weeks old non-obese diabetic Rag1-/- IL2Rgc-/- (NRG) mice at the age of 7-10 weeks were used for all CDX mouse models. Experiments were performed at the animal facility of Inserm U1110 (approval number E67-482-7) according to local laws, ethics committee approval and authorization by the French Ministry of Research and Higher Education (APAFIS #10892-2017080511379629v3, #22327-2019100815074277v3 and #27709-2020101514256404v4). CDX mouse models with PET-Scan assessments was performed at the animal facility of Institut Pluridisciplinaire Hubert Curien according to local laws and ethics committee approval (APAFiS#26596-2020071609156326v3). PDX mouse model: female 7 weeks old BALB/c nude mice at the age of 7 weeks were used with experiments performed at Crown Bioscience, Inc. The protocols were reviewed and approved by the Institutional Animal Care and Use Committee of CrownBio prior to execution. Mice were randomly assigned to the study groups. For details on methodology, please see supplementary information.

      Results

      CLDN1 is overexpressed in HCC characterized by a progenitor phenotype and an “immune-low” type of TME

      To investigate the role of CLDN1 in HCC, we first analyzed CLDN1 mRNA and protein levels in HCC patients. Computational analysis of data retrieved from Genomic Data Commons Data Portal and the human protein atlas
      • Uhlén M.
      • Fagerberg L.
      • Hallström B.M.
      • Lindskog C.
      • Oksvold P.
      • Mardinoglu A.
      • et al.
      Proteomics. Tissue-based map of the human proteome.
      revealed that CLDN1 is the most highly expressed Claudin family member in HCC at mRNA (p< 0.0001, U-test, Fig. 1A) and protein levels (Suppl. Fig. 1A). Indeed, ∼75% of liver tumors show medium or high CLDN1 expression (Suppl. Fig. 1A). Moreover, CLDN1 is significantly up-regulated in pre-malignant dysplastic nodules of cirrhotic liver (GSE102383, p=0.03, U-test, Suppl. Fig. 1B), as well as HCC tissue compared to matched non-tumorous adjacent liver (GSE113996, p=0.02, Wilcoxon signed-rank test, Suppl. Fig. 1C). Immunostaining of CLDN1 in healthy and HCC tissue confirmed CLDN1 overexpression, particularly in tumors of high tumor grade (p< 0.0001, U-test, Fig. 1B-C; p=0.06, U-test, Suppl. Fig. 1D, Suppl. Table 3). Interestingly, CLDN1 showed aberrant non-junctional localization in HCC in contrast to primarily junctional localization in healthy liver tissue (Fig. 1D). The predominantly non-junctional localization of CLDN1 in HCC tissues was further confirmed by double immunohistochemistry showing absent co-localization of CLDN1 with TJ protein ZO-1 (Fig. 1E).
      Figure thumbnail gr1
      Fig. 1CLDN1 expression is upregulated in HCC and correlates with stemness and an immune-low tumor microenvironment. (A) Expression of claudin family members in the TCGA cohort (n=365 patients, p<0.0001, one-way ANOVA). (B-C) Immunostaining of CLDN1 in healthy (n=10) and HCC tissues (n=70) of different tumor grades. (B) Representative images, (C) quantification (p< 0.0001, U-test). Scale bars 100 μm. (D) Representative high magnification images indicate aberrant NJ localization of CLDN1 in HCC compared to primarily junctional localization in healthy tissue. Scale bars 50 μm. (E) Left panels: Immunostainings of CLDN1 and ZO-1 in consecutive sections of HCC liver tissues. Right panel: Co-staining of CLDN1 (brown) and ZO1 (violet). Scale bars 150 μm (F) CLDN1 expression on single cell level (GSE151530, n=14 different HCC tissues) shown as violin plots. (G) tSNE graphs of tumor cells included in GSE151530/CLDN1low or CLDN1high. (H) Enrichment of gene sets related to stemness and progenitor cells in CLDN1high tumor cells (GSE151530, FDR<0.05, Kolmogorov-Smirnov test). (I) Enrichment of gene sets related to an immune-enriched or immune-active tumor immune microenvironment in tumors with low CLDN1 expression. Bars indicate NES of significant enrichments (GSE151530, FDR<0.05, Kolmogorov-Smirnov test). (J) Enrichment of genes specific for Boyault Liver Cancer Subclass G1 in tumors with high CLDN1 expression (GSE20140, n=164 FDR<0.001, Kolmogorov-Smirnov-test). (K) CLDN1 expression in HCC tissue predicted to confer response (n=42) or resistance (n=98) to sorafenib treatment (GSE109211, p<0.0001, Student’s t-test). (L) CLDN1 mRNA expression in the non-cancerous tumor-microenvironment of HCC patients with (VM1, n=9) or without (VM0, n=11) venous metastasis (GSE5093, p<0.0001, Mann-Whitney test). (M) Recurrence-free survival in patients with high (50% above median, n=26) vs. low (50% below median, n=26) CLDN1 expression in tumor adjacent liver tissue (GSE76427, p=0.008, log rank test). Boxplots represent median (▬), 1st and 3rd quartile (bottom and top of the box) and single data points (●). Vertical bars show NES of significantly (FDR<0.05, Kolmogorov-Smirnov test) altered gene sets. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Abbreviations: FPKM= Fragments Per Kilobase Million; MH-HCC=hepatocyte mature HCC; HpSC-HCC=progenitor-like stem cell signature; VM0=HCC adjacent tissue without venous metastasis; VM1=HCC adjacent tissue with venous metastasis; ZO1=Zonula occludens-1.
      HCC is characterized by a strong intra- and inter-tumoral heterogeneity and various molecular phenotypes
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      . Characterization of CLDN1 expression in a large HCC single-cell RNA sequencing (scRNAseq) dataset revealed that tumor cells are the key cell population with robust CLDN1 expression (Fig. 1F, Suppl. Figure 2A-B). Gene set enrichment analysis (GSEA) revealed that highly CLDN1-expressing tumor cells (Fig. 1G, Suppl. Figure 2C-D) exhibited a progenitor or stem cell-like phenotype (FDR<0.05, Kolmogorov-Smirnov test, Fig. 1H). Similar results were obtained in various independent large HCC patient cohorts with available bulk transcriptomic data (GSE5975, p<0.0001, U-test Suppl. Fig. 2E, GSE112791, FDR<0.001, Kolmogorov-Smirnov test, Suppl. Fig. 2F).
      Suggesting a functional impact of CLDN1 expression on the tumor immune microenvironment (TIME), scRNAseq data from HCC tumors revealed paucity of immune cells and suppression of immune signatures in tumors with low CLDN1 expression in tumor cells (FDR< 0.05, Kolmogorov-Smirnov test, Fig. 1I, Suppl. Fig. 3A-B). Signatures of angiogenesis and desmoplastic reaction on the other hand were significantly induced (FDR< 0.01, Kolmogorov-Smirnov test, Fig. 1I). Similar results were obtained in an independent large patient cohort on bulk transcriptomic level (GSE112791, FDR=0.04, Kolmogorov-Smirnov test, Suppl. Fig. 3C). As expected, given the correlation with a progenitor phenotype and an immune-low or -excluded TME, HCCs with high CLDN1 expression showed enrichment for genes specific for the Boyault G1 subclass of HCC
      • Boyault S.
      • Rickman D.S.
      • de Reyniès A.
      • Balabaud C.
      • Rebouissou S.
      • Jeannot E.
      • et al.
      Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets.
      (FDR<0.001, Kolmogorov-Smirnov-test, Fig. 1J). While we did not detect an association of CLDN1 expression in HCC tumors with survival (Suppl. Fig. 3D), their association with transcriptomic signatures of sorafenib resistance
      • Pinyol R.
      • Montal R.
      • Bassaganyas L.
      • Sia D.
      • Takayama T.
      • Chau G.-Y.
      • et al.
      Molecular predictors of prevention of recurrence in HCC with sorafenib as adjuvant treatment and prognostic factors in the phase 3 STORM trial.
      support a prognostic function of CLDN1 (GSE109211, p<0.0001, Student’s t-test, Fig. 1K). Furthermore, high CLDN1 expression in HCC adjacent liver tissue was associated with a metastatic behavior of the corresponding tumor and with worse post-resection recurrence-free survival (GSE5093, p<0.0001, Mann-Whitney, Fig. 1L, GSE76427, p=0.008, log-rank test, Fig. 1M). Collectively, these data suggest that CLDN1 is a potential hallmark for tumor cell differentiation and TIME.

      Genetic driver mutations, TNF-α/NF-κB and hypoxia upregulate CLDN1 overexpression that accelerates tumor growth in vivo

      We next screened a large group of different cytokines, growth factors and conditions to identify molecular drivers of CLDN1 overexpression in HCC. We found hypoxia and TNF-α/NF-κB to strongly upregulate CLDN1 expression in human Huh7 HCC cell line (p=0.03, U-test, Fig. 2A-B). Moreover, assessment of CLDN1 expression dependent on genetic driver mutations in the TCGA liver cancer cohort revealed AXIN1 mutations to be associated with CLDN1 upregulation (p= 0.003, U-test, Suppl. Fig. 4), whereas CTNNB1 mutations were found to be associated with CLDN1 downregulation (p< 0.0001, U-test, Suppl. Fig. 4). This is in line with CLDN1 being associated with the G1 Boyault HCC subclassification that is linked to AXIN1 mutations
      • Boyault S.
      • Rickman D.S.
      • de Reyniès A.
      • Balabaud C.
      • Rebouissou S.
      • Jeannot E.
      • et al.
      Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets.
      . Finally, CLDN1 gain-of-function in mouse Hepa1.6 HCC cells, that do not endogenously express CLDN1 (Fig. 2C) led to accelerated tumor growth when subcutaneously injected into syngeneic C57BL/6J mice (p=0.008, U-test, Fig. 2D), validating the pro-tumorigenic phenotype of CLDN1 overexpression in vivo.
      Figure thumbnail gr2
      Fig. 2Regulation of CLDN1 expression by hypoxia and effect of CLDN1 overexpression on tumor growth in a syngeneic mouse model. (A, B) Upregulation of CLDN1 expression by hypoxia and TNFα (A) Left panel: representative immunoblots of HIF1α, CLDN1 and β-actin in Huh7 cells treated with cobalt chloride for 0, 4, 8 and 24h. Right panel: Normalized CLDN1 expression (n=3; p=0.03, U-test). (B) TNFα-NFκB. Left panel: Representative immunoblots of phospho-p65, p65, CLDN1 and β-actin in Huh7 cells treated with TNFα for 0, 4, 8 and 24h. Right panel: Normalized CLDN1 expression upon stimulation with TNF (n=3, p=0.03, U-test). (C, D) Tumor growth in a CLDN1 syngeneic mouse model engrafted with wildtype and CLDN1-overexpressing Hepa1.6 cells (GoF). (C) Expression of cell surface CLDN1 in wildtype or CLDN GoF Hepa1.6 cells. Shown are representative histograms of fluorescence intensity in Hepa1.6 cells stained with CLDN1 mAb or control. (D) Tumor volume in n=5 mice per group was monitored over 15 days (*p<0.05, **p<0.01, U-test). Bars show mean±SEM and single data points.

      CLDN1 mAb suppresses tumor growth and EMT in patient-derived ex vivo models

      Given the upregulation of NJ-CLDN1 in HCC (Fig. 1D-E), we studied its role as a therapeutic target for HCC using a fully humanized mAb directed against the first extracellular loop (EL) of CLDN1
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      . Flow cytometry revealed robustly enhanced binding of CLDN1 mAb to patient-derived tumor cells compared to matched adjacent liver non-tumoral cells (p=0.004, Wilcoxon signed-rank test, Suppl. Fig.5).
      Treatment of human hepatoma cell lines Huh7 and Hep3B with CLDN1 mAb significantly inhibited tumor sphere formation, growth and invasion (Suppl. Fig. 6A-C). We next assessed the effect of CLDN1 mAb on tumor growth in a fully patient-derived culture system, modeling tumor heterogeneity. Cultured as multicellular micro-tissues, primary HCC tumorspheres maintain original cell-cell contacts and recapitulate non-parenchymal cells of the TME, including T-cells, which are relevant for tumor progression and therapeutic resistance
      • Song Y.
      • Kim J.-S.
      • Kim S.-H.
      • Park Y.K.
      • Yu E.
      • Kim K.-H.
      • et al.
      Patient-derived multicellular tumor spheroids towards optimized treatment for patients with hepatocellular carcinoma.
      . CLDN1 mAb treatment markedly disrupted the architecture of HCC spheroids (Fig. 3A). Moreover, CLDN1 mAb showed a pronounced effect on sorafenib-resistant HCC spheroid cell viability
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      (p=0.003 and p=0.04, Student’s t-test, Fig. 3B). A subsequent screen in HCC spheroids derived from 15 different HCC patients (patients’ characteristics shown in Suppl. Table 1), corroborated the effects of CLDN1 mAb on tumor cell viability with superior response rates compared to sorafenib and nivolumab (47% vs. 33% and 15%, defined as a mean decrease in cell viability of >15%, Fig. 3C).
      Figure thumbnail gr3
      Fig. 3Treatment with CLDN1 mAb reduces viability in patient-derived ex vivo models of HCC. (A) Representative microscopic photos of tumor spheroids generated from HCC liver tissue treated with CLDN1 or control mAb on day 6 post-treatment are shown. Scale bars 200 μm. (B) Relative cell viability after 6 days of treatment with CLDN1 mAb or sorafenib compared to control mAb treated spheroids (n= 3 replicates per condition, p=0.003 and p=0.04, Student’s t-test). (C) Tumor spheroids (n=15 donors) were treated with CLDN1 mAb, control Ab, sorafenib or nivolumab. Heatmaps illustrate % cell viability compared to control mAb treated cells on day 6 using ATP quantification (n=15 different donors with at least duplicates per condition). (●). Bars show mean ±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Abbreviations: CAFs=Cancer-associated fibroblasts; ECM=Extracellular matrix.
      Collectively, these data indicate strong suppressive effects of CLDN1 mAb on HCC growth including sorafenib- and nivolumab-resistant tumors.

      CLDN1 mAb suppresses tumor growth in both cell line-derived xenograft (CDX) and patient-derived xenograft (PDX) mouse models

      We next assessed effect of the mAb on tumor growth in Huh7 and Hep3B cell line-derived xenograft (CDX) mouse models (Fig. 4A). CLDN1-specific mAb significantly reduced tumor growth in vivo (Fig. 4B, Suppl. Fig. 7A). CLDN1 mAb antagonized tumor cell proliferation as measured by the strong reduction in percentage of KI67+ cells (Fig. 4C). Assessment of caspase 3 cleavage and TUNEL suggested pro-apoptotic effects (p=0.002, U-test, Fig. 4D, Suppl. Fig. 7B).
      Figure thumbnail gr4
      Fig. 4CLDN1 mAb inhibits tumor growth in CDX and PDX mouse models of HCC. (A) Huh7 CDX mouse model study protocol (2 independent studies). (B) Tumor growth (n=5 mice per group, *p<0.05, **p<0.01, Mann-Whitney test). (C) Immunohistochemistry and quantification of KI67 expression in tumor tissues. Scale bars - 1 mm and 200 μm (**p<0.01, Mann-Whitney). (D) Tumor apoptosis. Left panel: representative immunoblots of cleaved caspase 3, caspase 3 and β-Actin. Right panel: quantification of normalized cleaved caspase/caspase 3 ratio in tumor tissues (two different lysates for n=5 mice; p=0.002, U-test). (E) PDX mouse model study protocol (6 independent studies with n=6 different HCC tumors). (F) The relative mean tumor volume (%) in CLDN1 mAb (n=3 mice per model) compared to corresponding vehicle control treated mice (n=2 mice per model) is shown. (G) Tumor growth for #LI6716 (n=3 CLDN1 mAb, n=2 control, p=0.004, Wilcoxon matched pairs test). (H) Tumor growth for #LI6280 (total 24 mice, n=8 per group treated with 10 mg/ml QIW CLDN1 mAb, 25 mg/ml BIW CLDN1 mAb or control, p=0.008 and p=0.004, Wilcoxon matched-pairs test). (I) Response prediction to CLDN1 mAb treatment in HCC spheroid and PDX models. Bars show NES of significantly (FDR<0.05, Kolmogorov Smirnov test) enriched gene sets. (J) Response prediction to CLDN1 mAb in HCC spheroids by transcriptomic signatures of fibroblasts/angiogenesis as well as an immune infiltrate (FDRs=0.03 and 0.01, Kolmogorov-Smirnov test). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
      We next used PDX mouse models that are widely used to partially recapitulate tumoral heterogeneity and predict clinical outcomes
      • Xu C.
      • Li X.
      • Liu P.
      • Li M.
      • Luo F.
      Patient-derived xenograft mouse models: A high fidelity tool for individualized medicine.
      to evaluate the CLDN1 mAb anti-tumoral efficacy. Following established tumor growth, mice from 6 different PDX mouse models (Suppl. Table 2; Fig. 4E) were randomized into groups receiving weekly i.p. injections of CLDN1 mAb (n=3 per model) or vehicle control (n=2 per model). Treatment with CLDN1 mAb suppressed tumor growth by 38.5% on average in 4 out of 6 PDX models within 28 days of treatment, a response rate superior to currently approved treatment in clinical practice
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      (Fig. 4F-G). The observed inhibition was confirmed in a second study with an increased number of mice per group showing highly significant, dose-dependent inhibition of HCC growth (Fig. 4H). Similar to CDX mice, CLDN1 mAb treatment induced apoptosis as shown by Caspase 3 cleavage (p= 0.01, U-test, Suppl. Fig. 7C). Body weight in CLDN1 mAb treated animals remained unaltered compared to the control group throughout the study (Suppl. Table 4).
      RNA-sequencing (RNA-seq) and gene set enrichment analysis (GSEA) performed on pre-treatment PDX HCC tissues revealed that Wnt/β-catenin, Hedgehog and EMT signatures were strongly enriched in responders, irrespectively of both tumor grade and histological features (Fig. 4I). Interestingly, treatment resistance was associated with MYC and oxidative stress signatures. Similar results were obtained in tumor samples, whose response to CLDN1 mAb therapy was evaluated in HCC spheroids (Fig. 4I), validating the predictive values of the PDX signatures. Interestingly, tumors responding to CLDN1 mAb showed enrichment of transcriptomic signatures predicting a fibrotic and immune-enriched microenvironment (FDR=0.03 and 0.01, Kolmogorov-Smirnov test, Fig. 4J). Collectively, these data provide robust preclinical proof-of-concept for the treatment of HCC with CLDN1 mAbs and suggest a functional role of the TIME in treatment response.

      CLDN1 mAb interferes with oncogenic signaling and suppresses cancer cell metabolism in vivo

      To further gain insights into the molecular mechanisms that underlie the anti-tumor growth effect of CLDN1 mAb, we next performed RNA-sequencing (RNA-seq) and GSEA analysis on the PDX responder #LI6716 (Fig. 4G). CLDN1 has previously been reported to regulate intracellular signaling cascades by forming multi-protein complexes at the membrane
      • Roehlen N.
      • Roca Suarez A.A.
      • el Saghire H.
      • Saviano A.
      • Schuster C.
      • Lupberger J.
      • et al.
      Tight Junction Proteins and the Biology of Hepatobiliary Disease.
      . Corroborating the role of NJ-CLDN1 as signaling hubs, mice treated with CLDN1 mAb showed strong suppression of several key oncogenic signaling pathways, with the strongest effects on TNF-α/NF-κB, TGF-β, IL-6/JAK/STAT3, KRAS and Wnt/β-catenin signaling (Fig. 5A). Validating the specificity of these results, suppression of canonical TNF-α and TGF-β signaling via p65 and SMAD2/3 was observed intermodally in CDX mice at the protein level (p=0.04 and 0.03, U test, Fig. 5B-C, Suppl. Fig. 8). Interestingly, transcriptomic analysis further revealed a strong suppression of hypoxia-related genes and a concomitant restoration of bile acid metabolism, glycolysis and cholesterol homeostasis in PDX mice treated with CLDN1 mAb (Fig. 5A). Metabolic reprogramming is a hallmark of carcinogenesis contributing to tumor progression and therapeutic resistance. The Warburg effect describes the increased uptake of glucose and conversion to lactate in proliferating tumor cells independent of hypoxic conditions
      • vander Heiden M.G.
      • Cantley L.C.
      • Thompson C.B.
      Understanding the Warburg effect: the metabolic requirements of cell proliferation.
      . Therefore, we evaluated the effect of CLDN1 mAb on cancer cell metabolism by 3’-deoxy-3’-[18F]-fluorothymidine ([18F]-FLT) and 2-deoxy-2-[18F]-fluoro-D-glucose ([18F]-FDG) PET Scan of CLDN1 mAb or control-treated CDX mice. 18FLT PET Scan of 5 representative CDX mice per group (Fig. 5D) showed reduced uptake of 18FLT in CLDN1 mAb- compared to control-treated animals (p=0.008, U-test, Fig. 5E left and middle panel). Moreover, the total lesion proliferation (TLP) was markedly smaller in CLDN1 mAb- compared to control-treated mice (p=0.03, U-test, Fig.5E, right panel). FDG PET Scan (Fig. 5F) on the other hand, revealed strongly reduced total lesion glycolysis (TLG) in Huh7 CDX mice treated by CLDN1 mAb (p= 0.02, U-test, Fig. 5G). In contrast, sorafenib treatment did not show any effect on TLG (Fig. 5G). We confirmed the role of CLDN1 in tumor cell metabolism by investigating the effect of CLDN1 mAb on the induction of a Warburg-like metabolic shift by HCV infection in the hepatoma cell line Huh7.5.1 cells. The flux of lactate and other metabolites was restored to the level of control cells upon CLDN1 mAb treatment (Fig. 5H and Suppl. Fig. 9A-B). Taken together, these data suggest that CLDN1 drives metabolic tumor reprogramming in different models, including HCC tumors in vivo.
      Figure thumbnail gr5
      Fig. 5Impact of CLDN1-specific mAb treatment on tumor signaling and metabolism. (A) Unbiased GSEA analysis of HALLMARK gene sets in tumor tissue of CLDN1 mAb compared to vehicle control-treated PDX mice (LI6716). Heatmaps show NES of significantly (FDR<0.05, Kolmogorov Smirnov test) altered gene sets. (B, C). Normalized ratio of phospho-p65/p65 and phospho-Smad2/3/Smad 2/3 in CDX mice (n=5 with two independent protein lysates for each mouse, p<0.05, U-test). (D) [18F]-FLT PET Scan study protocol and representative images showing [18F]-FLT uptake in CLDN1 mAb- or control-treated CDX mice. (E) Quantitative assessment of SUVmean (left panel), SUVmax (middle panel), and TLP (activity, right panel) in [18F]-FLT PET Scans (n=5 mice per group, p=0.008 and p=0.03, U-test). (F) [18F]-FDG PET Scan study protocol and representative images showing [18F]-FDG uptake in CLDN1 mAb, sorafenib- or control-treated CDX mice. (G) Quantitative assessment of TLG in [18F]-FDG PET Scans (n=4-5 mice per group, *p<0.05, Mann-Whitney test). (H) Metabolites from CLDN1 or control mAb-treated and HCV-infected or non-infected (Mock) Huh7.5.1diff cells analyzed by mass spectrometry showing liver cell lactate flux. Negative values indicate accumulation outside the cells (one representative experiment, performed in triplicate, **p<0.01, Student’s t-test). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Abbreviations: SUV=standardized uptake value; TLG=total lesion glycolysis; TLP=total lesion proliferation; [18F]-FLT= 3’-deoxy-3’-[18F]-fluorothymidine; [18F]-FDG= 2-deoxy-2-[18F]- fluoro- D-glucose.

      CLDN1 mAb alters the tumor cell plasticity by interfering with Notch signaling activation

      The association of CLDN1 overexpression with the HpSC-HCC molecular subtype suggests a functional role of CLDN1 with tumor cell plasticity. Indeed, gene sets related to liver cancer stemness and EMT were suppressed in CLDN1 mAb-treated PDX mice (FDR< 0.05 and FDR< 0.001, Kolmogorov-Smirnov-test, Fig. 6A). Supporting this concept, CDX mice treated with CLDN1 mAb showed a significantly suppressed expression of the stemness and EMT markers EPCAM and Fibronectin (FN1) (Fig. 6B). Linking the molecular effects on EMT with the observed inhibition of cell invasion (Suppl. Fig. 6C), CLDN1 mAb strongly suppressed expression of the matrix metalloproteinase MMP14 in cell-based models (p<0.001, U-test, Fig. 6C) and CDX mice (p=0.01, U-test, Fig. 6D, Suppl. Fig. 10). To validate the role of CLDN1 in cell plasticity we next assessed the effect of CLDN1 mAb treatment on EMT in patient-derived liver scaffold culture systems allowing assessment of cancer therapeutics in a three-dimensional growth microenvironment
      • Miyauchi Y.
      • Yasuchika K.
      • Fukumitsu K.
      • Ishii T.
      • Ogiso S.
      • Minami T.
      • et al.
      A novel three-dimensional culture system maintaining the physiological extracellular matrix of fibrotic model livers accelerates progression of hepatocellular carcinoma cells.
      . To study the effect of CLDN1 mAb on EMT
      • Radtke F.
      • MacDonald H.R.
      • Tacchini-Cottier F.
      Regulation of innate and adaptive immunity by Notch.
      repopulated liver scaffolds were treated with TGFβ (study protocol illustrated in Fig. 6E). CLDN1 mAb markedly suppressed EMT marker gene expression including Vimentin (VIM), FN1 and Snail Family Transcriptional Repressor 2 (SNAI2) (p=0.005, p=0.008 and p=0.005, Wilcoxon signed-rank test, Fig. 6F). Similar results were obtained in a complementary 3D model system, consisting of Huh7 cells co-cultured with primary cancer-associated fibroblasts in patient liver-derived fibrotic extracellular matrix hydrogel (Suppl. Fig. 11).
      Figure thumbnail gr6
      Fig. 6CLDN1 mAb modulates cancer cell plasticity and fate by interacting with Notch signaling. (A) Significant suppression of gene sets related to EMT and stemness in PDX mice treated with CLDN1 mAb (LI6716: FDR< 0.05 and FDR< 0.001, Kolmogorov-Smirnov-test). (B) Immunohistochemistry and quantitation of EPCAM and FN1 expression in CDX tumors (n=5). Scale bars - 1 mm and 200 μm (*p<0.05, U-test). (C) Inhibition of MMP14 expression in Huh7 spheroids. Left panel: representative immunoblots of MMP14 and β-Actin. Right panel: quantification of normalized MMP14 expression (n=3). (D) Quantification of normalized MMP14 expression in CDX HCCs (n=5, two independent protein lysates for immunoblots). (E) Patient-derived liver scaffold study protocol. (F) Normalized gene expression of EMT markers Vimentin (VIM), Fibronectin (FN1) and SNAI2 in CLDN1 or control mAb treated Huh7+CAF liver scaffolds (3-4 independent experiments in at least triplicates, p=0.005, p=0.008 and p=0.005, Wilcoxon signed-rank test). (G) Co-immunoprecipitation of Notch ligand Jagged 1 (JAG1) and CLDN1 in the cell membrane of Huh7 cells. The western blot shows JAG1 presence in CLDN1 eluate. (H) Inhibition of Notch signaling in Huh7 cells. Left panel: Representative immunoblots of cleaved Notch1 (Val1744), Notch1 and β-actin in Huh7 stimulated with Jagged-1 (JAG1). Right panel: Normalized ratio of cleaved Notch1/Notch1 protein expression (n=3, p=0.03, U-test). (I) Normalized ratio of cleaved Notch1/Notch1 in HCCs from CDX mice (5 mice per group and two independent lysates each), p=0.02, U-test). Bars show mean ±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.
      To unravel the molecular drivers of EMT and cell plasticity, we studied the effect of the CLDN1 mAb on Notch signaling - a key regulator of cell differentiation and stemness. Using cell membrane co-immunoprecipitation studies in Huh7 cells, we observed a direct interaction of CLDN1 with JAG1, the upstream inducer of canonical Notch signaling (Fig. 6G, Suppl. Fig. 13 A). The functional relevance of this interaction was confirmed by robust inhibition of Notch cleavage by CLDN1 mAb in HCC cell-based and CDX animal models (p=0.03, U-test, Fig. 6H and p=0.02, U-test, Fig. 6I, Suppl. Fig. 12).

      Treatment with CLDN1 mAb reprograms TIME in patient-derived HCC spheroids

      The TIME has been shown to play an important role in HCC progression and therapeutic resistance
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      . Given the association of CLDN1 expression with an “immune-low” TIME (Fig. 1) and prediction of response to CLDN1 mAb by a well-described signature of immune infiltration
      • Bagaev A.
      • Kotlov N.
      • Nomie K.
      • Svekolkin V.
      • Gafurov A.
      • Isaeva O.
      • et al.
      Conserved pan-cancer microenvironment subtypes predict response to immunotherapy.
      (Fig. 4), we next evaluated the effects of CLDN1 mAb on the TIME of patient-derived HCC tumor spheroids. HCC spheroids (#462, Suppl. Table 1) were treated with either CLDN1 mAb or an isotype control mAb for 24 h and then subjected to scRNA-seq (Fig. 7A). Unbiased sorting of viable cells allowed sequencing of the transcriptome of all major immune cell types including T cells, macrophages, monocytes and dendritic cells (Fig. 7B, Suppl. Fig 13A-G and Suppl. Table 5). As expected with the short treatment duration, we did not observe any significant difference in the number of any immune cell subsets (Suppl. Fig. 13A). However, in CLDN1 mAb-treated patient HCC spheroids, CD8+ T cells showed enrichment of genes associated with immune effector function and proliferation (Fig. 7C). In contrast, in two immunosuppressive Treg clusters, FOXP3 target genes were strongly down-regulated while genes associated with an immune effector function were up-regulated (Fig. 7D and E). Since CLDN1 expression was not detectable in T cells (Suppl. Figure 13D and F) these effects are most likely due to a direct impact of NJ-CLDN1 on tumor cells.
      Figure thumbnail gr7
      Fig. 7scRNAseq of patient-derived HCC spheroids suggests that CLDN1 mAb treatment modulates T cell effector activity. (A) Study protocol of scRNA-seq on sorted immune cells derived from CLDN1 or control mAb-treated HCC spheroids (n=10-20 spheroids per group). (B) 2D-visualization of single-cell transcriptomics of cells sorted from CLDN1 vs. control mAb-treated HCC spheroids using t-SNE maps. (C) Significant upregulation of gene sets related to immune effector processes and proliferation in CD8+ T cells derived from CLDN1 vs. control mAb-treated HCC spheroids. (D) Significant suppression of gene sets specific for Tregs and upregulation of markers of non-suppressive T cells in CD4+ CD25+ FOXP3high Tregs. (E) Significant suppression of gene sets specific for Tregs and upregulation of genes related to immune effector processes as well as naïve CD4+ T cells in CD4+ CD25+ CTLA4high FOXP3high Tregs. (F) Significantly suppressed signaling gene sets in tumor cells derived from CLDN1 vs. control mAb-treated HCC spheroids. (G) Suppression of gene sets related to stemness in tumor cells derived from CLDN1- vs. control mAb-treated HCC spheroids. (H) Enrichment of gene sets related to cytokine production in tumor cells derived from CLDN1 vs. control mAb-treated spheroids. Horizontal bars indicate NES; *FDR<0.05, **<0.01, ***<0.001, ****< 0.0001. Abbreviations: Tregs=regulatory T cells; t-SNE=t-distributed stochastic neighbor embedding.
      GSEA on HCC sorted tumor cells confirmed the downregulation of TGFβ-, KRAS, Wnt- and TNFα-signaling pathways whose activation has been associated with an immune TIME
      • Llovet J.M.
      • Castet F.
      • Heikenwalder M.
      • Maini M.K.
      • Mazzaferro V.
      • Pinato D.J.
      • et al.
      Immunotherapies for hepatocellular carcinoma.
      (Fig. 7F). Furthermore, gene sets associated with immunogenic cytokine activity were suppressed in tumor cells (Fig. 7H).
      Altogether, these data suggest that CLDN1 expressed on tumor cells may dictate an immunosuppressive TIME in HCC that can be reverted by CLDN1 mAb into an immunostimulatory TIME. These data thus provide a rationale for combining CLDN1 mAbs with immune checkpoint inhibitors in HCC.

      Discussion

      Applying patient-derived ex vivo and in vivo models and a highly specific mAb, we identified CLDN1 as a novel therapeutic target for treatment of HCC. CLDN1 mAb treatment inhibits tumor growth and its phenotype (Fig. 8) by (a) targeting non-junctional CLDN1 upregulated by TNFα and hypoxia (Fig. 1, Fig. 2); (b) suppression of cancer cell stemness and EMT, a hallmark of HCC tumors with high invasive capacity, therapeutic resistance and poor prognosis (Fig. 4); (c) reprogramming tumor metabolism, a feature of cancer cells determining cell survival, hyperplastic growth and evasion from immune responses (Fig. 5); (d) induction of apoptosis (Fig. 4, 6) and (e) alteration of the TIME with enhancement of anti-tumor activity (Fig. 7).
      Figure thumbnail gr8
      Fig. 8Mechanistic model for CLDN1-targeting therapies. Left: Up-regulation of CLDN1 expression by TNFα-NFκB signaling and hypoxia/HIF1α (, ). Anti-CLDN1 antibodies bind to non-junctional CLDN1 on the tumor cell surface. Middle: Anti-CLDN1 treatment interferes with tumor cell signaling regulating tumor stemness, differentiation, cell fate, proliferation and migration (Fig. 4, Fig. 5, Fig. 6) by perturbating the JAG-Notch pathway (). Other downstream targets may be mediated via non-canonical Notch signaling or by CLDN1 interaction/cross-talk with other adhesion molecules or RTKs
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      ,
      • Zona L.
      • Lupberger J.
      • Sidahmed-Adrar N.
      • Thumann C.
      • Harris H.J.
      • Barnes A.
      • et al.
      HRas signal transduction promotes hepatitis C virus cell entry by triggering assembly of the host tetraspanin receptor complex.
      . Right: High CLDN1 expression is associated with an immune-low or inactive TIME (, ). The change of the tumor cell phenotype by mAb treatment leads to activation of T-cells in the TIME likely by alterations in the tumor cell secretome and/or metalloproteinases (). Ab-mediated inhibition of liver fibrosis described previously

      El Saghire H, Saviano A, Roehlen N, Crouchet E, Duong FHT, Jühling F et al. Abstract of The International Liver Congress TM 2021 June 23-26, 2021. Abstract GS-2069.

      may further contribute to treatment response.
      Mechanistic studies indicated that inhibition of Notch signaling via direct interactions between CLDN1 and JAG1 likely plays a role for the observed effects on tumor cell fate (Fig. 6). Since we did not observe a significant induction of Notch effector genes in the PDX HCC transcriptome LI6716 (Fig. 5A), it is conceivable that Notch signaling is either an early event in the signal cascade or that perturbation of Notch signaling is heterogenous in individual tumors. The Notch pathway has been identified as a key regulator of cell differentiation, fate and survival and several functional and clinical studies have shown that the Notch pathway plays a role in the pathogenesis of HCC
      • Zhu C.
      • Ho Y.-J.
      • Salomao M.A.
      • Dapito D.H.
      • Bartolome A.
      • Schwabe R.F.
      • et al.
      Notch activity characterizes a common hepatocellular carcinoma subtype with unique molecular and clinicopathologic features.
      ,
      • Villanueva A.
      • Alsinet C.
      • Yanger K.
      • Hoshida Y.
      • Zong Y.
      • Toffanin S.
      • et al.
      Notch Signaling Is Activated in Human Hepatocellular Carcinoma and Induces Tumor Formation in Mice.
      . Since Notch signaling has also been identified as a regulator of innate and adaptive immune responses
      • Radtke F.
      • MacDonald H.R.
      • Tacchini-Cottier F.
      Regulation of innate and adaptive immunity by Notch.
      , its perturbation by CLDN1 mAb treatment may also contribute to the observed effects on the TIME (Fig. 7).
      The observed effects on NFkB, Wnt-β-catenin and P3IK signaling (Fig. 6A) may be mediated via non-canonical Notch signaling or by signal transduction of other cell membrane molecules shown to interact/cross-talk with CLDN1
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      ,
      • Zona L.
      • Lupberger J.
      • Sidahmed-Adrar N.
      • Thumann C.
      • Harris H.J.
      • Barnes A.
      • et al.
      HRas signal transduction promotes hepatitis C virus cell entry by triggering assembly of the host tetraspanin receptor complex.
      ,

      El Saghire H, Saviano A, Roehlen N, Crouchet E, Duong FHT, Jühling F et al. Abstract of The International Liver Congress TM 2021 June 23-26, 2021. Abstract GS-2069.

      . The CLDN1 mAb-induced alteration of tumor cell plasticity and its related immune modulatory effects highlight potential opportunities for combining CLDN1 mAb with immune-oncological approaches and multi-kinase inhibitors
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      .
      HCC arises almost exclusively in the context of liver fibrosis and chronic inflammation
      • Baglieri J.
      • Brenner D.A.
      • Kisseleva T.
      The Role of Fibrosis and Liver-Associated Fibroblasts in the Pathogenesis of Hepatocellular Carcinoma.
      . The stage of liver fibrosis hereby represents a key factor for patient outcome
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      . In addition to the tumor suppressive effects of CLDN1 mAb demonstrated in this study, we previously showed that CLDN1 targeting mAbs suppress liver fibrosis

      El Saghire H, Saviano A, Roehlen N, Crouchet E, Duong FHT, Jühling F et al. Abstract of The International Liver Congress TM 2021 June 23-26, 2021. Abstract GS-2069.

      . While HCC treatment strategies are frequently limited by the degree of liver cirrhosis
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      , the combined antifibrotic and tumor suppressive effects of CLDN1 mAbs provide a unique opportunity to target not only tumor growth but also fibrosis and de novo HCC development in the non-tumorous fibrotic microenvironment
      • Llovet J.M.
      • Kelley R.K.
      • Villanueva A.
      • Singal A.G.
      • Pikarsky E.
      • Roayaie S.
      • et al.
      Hepatocellular carcinoma.
      .
      Furthermore, this study provides prediction markers for patient selection to CLDN1 mAb therapies. Upregulation of EMT as well as signaling pathways implicated in stemness such as Wnt/β-catenin and Hedgehog signaling
      • Liu Y.-C.
      • Yeh C.-T.
      • Lin K.-H.
      Cancer Stem Cell Functions in Hepatocellular Carcinoma and Comprehensive Therapeutic Strategies.
      predicted response of HCC tumors to experimental treatment.
      Our data obtained here and in previous studies demonstrate that the administration of the antibody is safe without detectable adverse and off-target effects
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      ,
      • Colpitts C.C.
      • Tawar R.G.
      • Mailly L.
      • Thumann C.
      • Heydmann L.
      • Durand S.C.
      • et al.
      Humanisation of a claudin-1-specific monoclonal antibody for clinical prevention and cure of HCV infection without escape.
      . The absence of toxicity and off-target effects are likely due to the specific binding of the mAb to NJ-CLDN1 and not to TJ CLDN1
      • Mailly L.
      • Xiao F.
      • Lupberger J.
      • Wilson G.K.
      • Aubert P.
      • Duong F.H.T.
      • et al.
      Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
      ,

      El Saghire H, Saviano A, Roehlen N, Crouchet E, Duong FHT, Jühling F et al. Abstract of The International Liver Congress TM 2021 June 23-26, 2021. Abstract GS-2069.

      .
      Collectively, our data provide robust pre-clinical proof-of-concept for CLDN1 mAbs as potential first in-class compounds with a perspective to break the plateau of limited treatment response in advanced HCC, raising the outlook for patients with currently poor prognosis.

      Acknowledgments

      The authors thank the CRB (Centre de Ressources Biologiques-Biological Resource Centre) of the Strasbourg University Hospitals for the management of patient-derived liver tissue. The authors thank Dr. D. Root (Broad Institute of MIT and Harvard, Cambridge, MA) for providing expression plasmids for lentiviruses and sgRNAs for CLDN1 KO and Prof. Gerhard Cristofori (University of Basel) for the gift of Huh7 cells, Drs. F. Chisari (Scripps) and C. Rice (Rockefeller University) for providing Huh7.5.1 cells and Dr. R. Bartenschlager (University of Heidelberg) for providing HCV Jc1 strain, Thomas Cagarelli and Dr. Aurélie Bornand (University Hospital Geneve) for histopathological analyses and Dr. Olga Koutsopoulos (Inserm U1110, U Strasbourg) for helpful discussions, project management and support in fundraising. The authors thank the entire personnel of the CYRCe platform (IPHC, France) for their contribution to the PET experiments.

      Appendix A. Supplementary data

      Abbreviations:

      [18F]-FDG=
      2-deoxy-2-[18F]-fluoro-D-glucose
      [18F]-FLT=
      3’-deoxy-3’-[18F]-fluorothymidine
      CAFs=
      Cancer-associated fibroblasts
      CDX=
      cell line derived xenograft
      ECM=
      Extracellular matrix
      EMT=
      epithelial-mesenchymal transition
      FDR=
      False-discovery rate
      FPKM=
      Fragments Per Kilobase Million
      GoF=
      Gain-of-function
      GSEA=
      Gene set enrichment analysis
      HCC=
      Hepatocellular carcinoma
      HCV=
      Hepatitis C virus
      HpSC-HCC=
      progenitor-like stem cell signature
      mAb=
      monoclonal antibody
      MH-HCC=
      hepatocyte mature HCC
      NJ-CLDN1=
      non-junctional CLDN1
      PDX=
      Patient-derived xenograft
      RNA-seq=
      RNA-sequencing
      scRNAseq=
      singe cell RNA sequencing
      SNAI2=
      Snail Family Transcriptional Repressor 2
      TJs=
      Tight junctions
      TLG=
      total lesion glycolysis
      Tregs=
      regulatory T cells
      t-SNE=
      t-distributed stochastic neighbor embedding
      TIME=
      tumor immune microenvironment
      TME=
      tumor microenvironment
      VIM=
      Vimentin
      VM0=
      HCC adjacent tissue without venous metastasis
      VM1=
      HCC adjacent tissue with venous metastasis
      ZO1=
      Zonula occludens-1

      References

      Author names in bold designate shared co-first authorships
        • Llovet J.M.
        • Kelley R.K.
        • Villanueva A.
        • Singal A.G.
        • Pikarsky E.
        • Roayaie S.
        • et al.
        Hepatocellular carcinoma.
        Nat Rev Dis Primers. 2021; 7: 6
        • Finn R.S.
        • Qin S.
        • Ikeda M.
        • Galle P.R.
        • Ducreux M.
        • Kim T.-Y.
        • et al.
        Atezolizumab plus Bevacizumab in Unresectable Hepatocellular Carcinoma.
        N Engl J Med. 2020; 382: 1894-1905
        • Calderaro J.
        • Ziol M.
        • Paradis V.
        • Zucman-Rossi J.
        Molecular and histological correlations in liver cancer.
        J Hepatol. 2019; 71: 616-630
        • Qin S.
        • Jiang J.
        • Lu Y.
        • Nice E.C.
        • Huang C.
        • Zhang J.
        • et al.
        Emerging role of tumor cell plasticity in modifying therapeutic response.
        Signal Transduct Target Ther. 2020; 5: 228
        • Mailly L.
        • Xiao F.
        • Lupberger J.
        • Wilson G.K.
        • Aubert P.
        • Duong F.H.T.
        • et al.
        Clearance of persistent hepatitis C virus infection in humanized mice using a claudin-1-targeting monoclonal antibody.
        Nat Biotechnol. 2015; 33: 549-554
        • Uhlén M.
        • Fagerberg L.
        • Hallström B.M.
        • Lindskog C.
        • Oksvold P.
        • Mardinoglu A.
        • et al.
        Proteomics. Tissue-based map of the human proteome.
        Science. 2015; 3471260419
        • Boyault S.
        • Rickman D.S.
        • de Reyniès A.
        • Balabaud C.
        • Rebouissou S.
        • Jeannot E.
        • et al.
        Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets.
        Hepatology. 2007; 45: 42-52
        • Pinyol R.
        • Montal R.
        • Bassaganyas L.
        • Sia D.
        • Takayama T.
        • Chau G.-Y.
        • et al.
        Molecular predictors of prevention of recurrence in HCC with sorafenib as adjuvant treatment and prognostic factors in the phase 3 STORM trial.
        Gut. 2019; 68: 1065-1075
        • Song Y.
        • Kim J.-S.
        • Kim S.-H.
        • Park Y.K.
        • Yu E.
        • Kim K.-H.
        • et al.
        Patient-derived multicellular tumor spheroids towards optimized treatment for patients with hepatocellular carcinoma.
        J Exp Clin Cancer Res. 2018; 37: 109
        • Xu C.
        • Li X.
        • Liu P.
        • Li M.
        • Luo F.
        Patient-derived xenograft mouse models: A high fidelity tool for individualized medicine.
        Oncol Lett. 2019; 17: 3-10
        • Roehlen N.
        • Roca Suarez A.A.
        • el Saghire H.
        • Saviano A.
        • Schuster C.
        • Lupberger J.
        • et al.
        Tight Junction Proteins and the Biology of Hepatobiliary Disease.
        Int J Mol Sci. 2020; : 21
        • vander Heiden M.G.
        • Cantley L.C.
        • Thompson C.B.
        Understanding the Warburg effect: the metabolic requirements of cell proliferation.
        Science. 2009; 324: 1029-1033
        • Miyauchi Y.
        • Yasuchika K.
        • Fukumitsu K.
        • Ishii T.
        • Ogiso S.
        • Minami T.
        • et al.
        A novel three-dimensional culture system maintaining the physiological extracellular matrix of fibrotic model livers accelerates progression of hepatocellular carcinoma cells.
        Sci Rep. 2017; 7: 9827
        • Bagaev A.
        • Kotlov N.
        • Nomie K.
        • Svekolkin V.
        • Gafurov A.
        • Isaeva O.
        • et al.
        Conserved pan-cancer microenvironment subtypes predict response to immunotherapy.
        Cancer Cell. 2021; 39: 845-865.e7
        • Llovet J.M.
        • Castet F.
        • Heikenwalder M.
        • Maini M.K.
        • Mazzaferro V.
        • Pinato D.J.
        • et al.
        Immunotherapies for hepatocellular carcinoma.
        Nat Rev Clin Oncol. 2022; 19: 151-172
        • Zhu C.
        • Ho Y.-J.
        • Salomao M.A.
        • Dapito D.H.
        • Bartolome A.
        • Schwabe R.F.
        • et al.
        Notch activity characterizes a common hepatocellular carcinoma subtype with unique molecular and clinicopathologic features.
        J Hepatol. 2021; 74: 613-626
        • Villanueva A.
        • Alsinet C.
        • Yanger K.
        • Hoshida Y.
        • Zong Y.
        • Toffanin S.
        • et al.
        Notch Signaling Is Activated in Human Hepatocellular Carcinoma and Induces Tumor Formation in Mice.
        Gastroenterology. 2012; 143: 1660-1669.e7
        • Radtke F.
        • MacDonald H.R.
        • Tacchini-Cottier F.
        Regulation of innate and adaptive immunity by Notch.
        Nat Rev Immunol. 2013; 13: 427-437
        • Zona L.
        • Lupberger J.
        • Sidahmed-Adrar N.
        • Thumann C.
        • Harris H.J.
        • Barnes A.
        • et al.
        HRas signal transduction promotes hepatitis C virus cell entry by triggering assembly of the host tetraspanin receptor complex.
        Cell Host Microbe. 2013; 13: 302-313
      1. El Saghire H, Saviano A, Roehlen N, Crouchet E, Duong FHT, Jühling F et al. Abstract of The International Liver Congress TM 2021 June 23-26, 2021. Abstract GS-2069.

        • Baglieri J.
        • Brenner D.A.
        • Kisseleva T.
        The Role of Fibrosis and Liver-Associated Fibroblasts in the Pathogenesis of Hepatocellular Carcinoma.
        Int J Mol Sci. 2019; 20
        • Liu Y.-C.
        • Yeh C.-T.
        • Lin K.-H.
        Cancer Stem Cell Functions in Hepatocellular Carcinoma and Comprehensive Therapeutic Strategies.
        Cells. 2020; 9
        • Colpitts C.C.
        • Tawar R.G.
        • Mailly L.
        • Thumann C.
        • Heydmann L.
        • Durand S.C.
        • et al.
        Humanisation of a claudin-1-specific monoclonal antibody for clinical prevention and cure of HCV infection without escape.
        Gut. 2018; 67: 736-745