Analysis of Liver Cancer Cell Lines Identifies Agents With Likely Efficacy Against Hepatocellular Carcinoma and Markers of Response Stefano Caruso, Anna-Line Calatayud, Jill Pilet, Tiziana La Bella, Samia Rekik, Sandrine Imbeaud, Eric Letouzé, Léa Meunier, Quentin Bayard, Nataliya Rohr-Udilova, et al.
To cite this version:
Stefano Caruso, Anna-Line Calatayud, Jill Pilet, Tiziana La Bella, Samia Rekik, et al.. Analy- sis of Liver Cancer Cell Lines Identifies Agents With Likely Efficacy Against Hepatocellular Car- cinoma and Markers of Response. Gastroenterology, WB Saunders, 2019, 157 (3), pp.760-776. 10.1053/j.gastro.2019.05.001. hal-02291798
HAL Id: hal-02291798 https://hal.sorbonne-universite.fr/hal-02291798 Submitted on 19 Sep 2019
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Accepted Manuscript
Analysis of Liver Cancer Cell Lines Identifies Agents With Likely Efficacy Against Hepatocellular Carcinoma and Markers of Response
Stefano Caruso, Anna-Line Calatayud, Jill Pilet, Tiziana La Bella, Samia Rekik, Sandrine Imbeaud, Eric Letouzé, Léa Meunier, Quentin Bayard, Nataliya Rohr- Udilova, Camille Péneau, Bettina Grasl-Kraupp, Leanne de Koning, Bérengère Ouine, Paulette Bioulac-Sage, Gabrielle Couchy, Julien Calderaro, Jean-Charles Nault, Jessica Zucman-Rossi, Sandra Rebouissou
PII: S0016-5085(19)36771-X DOI: https://doi.org/10.1053/j.gastro.2019.05.001 Reference: YGAST 62635
To appear in: Gastroenterology Accepted Date: 1 May 2019
Please cite this article as: Caruso S, Calatayud A-L, Pilet J, La Bella T, Rekik S, Imbeaud S, Letouzé E, Meunier L, Bayard Q, Rohr-Udilova N, Péneau C, Grasl-Kraupp B, de Koning L, Ouine B, Bioulac- Sage P, Couchy G, Calderaro J, Nault J-C, Zucman-Rossi J, Rebouissou S, Analysis of Liver Cancer Cell Lines Identifies Agents With Likely Efficacy Against Hepatocellular Carcinoma and Markers of Response, Gastroenterology (2019), doi: https://doi.org/10.1053/j.gastro.2019.05.001.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Molecular characterization Transcriptome subgroups CL1 CL2 CL3 Mutations/CNA [WES] Mixed epithelial Mesenchymal- Hepatoblast-like Differentiation -mesenchymal like Transcriptome [RNA-seq] Liver TERT promoter, TP53, AXIN1, TSC1/TSC2 mut miRNome [miRNA seq] Cancer Main 11q13 amplification (FGF19 and CCND1) Cell Lines molecular Progenitor features: Wnt-TGFß signaling 126 proteins [RPPA] features IGF2, AFP, EPCAM+ Phospho-ß-catenin S675 CK19+ Drug/ MANUSCRIPTLinsitinib Biomarker Sorafenib + MK2206 Pharmacological profiling pairs Sorafenib + MEK1/2 inhibitors 1 n=34 Trametinib/FGF19 ampli FGFR4 inhibitors/FGF19 ampli AUC Dasatinib/CK19 0.5 ACCEPTED Alisertib/TP53 mut 31 drugs GI50 Rapamycin/TSC1-TSC2 mut Cellviability 0 approved/ Drug [C°] µM MET inhibitors: Cabozantinib/MET ampli clinical trials Alvespimycin,Tanespimicyn/NQO1 ACCEPTED MANUSCRIPT Analysis of Liver Cancer Cell Lines Identifies Agents With Likely Efficacy
Against Hepatocellular Carcinoma and Markers of Response
Short title : HCC cell lines to predict drug response
Authors
Stefano Caruso* 1,2 , Anna-Line Calatayud*1,2 , Jill Pilet*1,2 , Tiziana La Bella 1,2 , Samia
Rekik 1,2 , Sandrine Imbeaud 1,2 , Eric Letouzé 1,2 , Léa Meunier 1,2 , Quentin Bayard 1,2 , Nataliya
Rohr-Udilova 3, Camille Péneau 1,2 , Bettina Grasl-Kraupp 4, Leanne de Koning 5, Bérengère
Ouine 5, Paulette Bioulac-Sage 6,7, Gabrielle Couchy 1,2 , Julien Calderaro 8, Jean-Charles
Nault 1,2,9,10 , Jessica Zucman-Rossi §1,2,11#, Sandra Rebouissou §1,2#.
Institutions
1 MANUSCRIPT Centre de Recherche des Cordeliers, Sorbonne Universités, Inserm, UMRS-1138, F-75006 Paris, France;
2Functional Genomics of Solid Tumors, USPC, Université Paris Descartes, Université Paris
Diderot, Université Paris 13, Labex Immuno-Oncology, équipe labellisée Ligue Contre le
Cancer, F-75000 Paris, France;
3Division of Gastroenterology and Hepatology, Clinic of Internal, Medicine III, Medical
University of Vienna, Vienna, Austria;
4Department of ACCEPTEDMedicine I, Division: Institute of Cancer Research, Comprehensive Cancer
Center, Medical University of Vienna, Vienna, Austria;
5RPPA platform, Curie Institute, PSL research university, Paris, France;
6Bariton INSERM, UMR-1053, Bordeaux, France;
1 ACCEPTED MANUSCRIPT 7Department of Pathology, Pellegrin Hospital, Hospital of Bordeaux, Bordeaux, France;
8Anathomopathology Department, Henri Mondor Hospital, Créteil; University of Paris Est
Créteil, Inserm U955, Team 18, Mondor Institute of Biomedical Research, France;
9Liver unit, Jean Verdier Hospital, University Hospitals Paris-Seine-Saint-Denis, AP-HP,
Bondy, France;
10 Training and Research Unit of Health Medicine and Human Biology, University of Paris 13,
Community of Universities and Institutions Sorbonne Paris Cité, Paris, France;
11 European Hospital Georges Pompidou, AP-HP, F-75015, Paris, France
* These authors contributed equally to this work
§ These authors contributed equally to this work
Grant supports: This work was supported by Inserm,MANUSCRIPT the Association Française pour l’Etude du Foie (AFEF). The team is supported by the Ligue Nationale Contre le Cancer (Equipe
Labellisée), Labex OncoImmunology (Investissement d’Avenir), the Fondation Bettencourt-
Schueller (coup d’élan Award), the Ligue Contre le Cancer Comité de Paris (Duquesne award) and the Fondation pour la Recherche Médicale (Raymond Rosen award). SC is supported by a funding from Labex OncoImunology and CARPEM. ALC, LM and QB are supported by a fellowship from the Ministère de l'Education Nationale de la Recherche et de
Technologie (MENRT) and JP, SRek and CP received a fellowship from the Fondation pour la Recherche MédicaleACCEPTED (FRM), Association pour la Recherche sur le Cancer (ARC) and ANRS, respectively.
2 ACCEPTED MANUSCRIPT Abbreviations: LCCL: liver cancer cell lines, HCC: hepatocellular carcinoma, RT-PCR: reverse transcriptase polymerase chain reaction, RPPA: reverse phase protein array, WES: whole exome sequencing, CNA: copy number alterations, EN: elastic net, EMT: epithelial- mesenchymal transition, SD: standard deviation, TSG: tumor suppressor gene
#Corresponding Authors:
Sandra Rebouissou, PhD
Inserm UMRS-1138, Functional Genomics of Solid Tumors
27 rue Juliette Dodu
75010 Paris, France
TEL: +33 1 72 63 93 55
Email: [email protected] MANUSCRIPT Jessica Zucman-Rossi, MD, PhD
Inserm UMRS-1138, Functional Genomics of Solid Tumors
27 rue Juliette Dodu
75010 Paris, France
TEL: +33 1 72 63 93 58
Email: [email protected] ACCEPTED Disclosures: The authors disclose no conflicts.
3 ACCEPTED MANUSCRIPT Accession codes: The sequencing data reported in this study (WES, RNA-sequencing and miRNA-sequencing) have been deposited to the EGA (European Genome-phenome Archive) database and are available through the accession number [EGAS00001003536].
All the molecular and drug sensitivity data are available at: http://lccl.zucmanlab.com
Writing Assistance: no writing assistance
Author Contributions:
Study concept and design: SR, JZR
Acquisition of data: ALC, JP, SRek, TLB, CP, LDK, BO, GC, JCN
Analysis and interpretation of data: SC, SI, EL, LM, QB
Drafting of the manuscript: SC, ALC, SR, JZR
Collection of samples: NRU, BGK, PBS, JC, JCN
Study supervision: SR, JZR Revision of the manuscript and approval of the fina MANUSCRIPTl version of the manuscript: ALL Acknowledgments: We thank Philippe Merle and Fabien Zoulim for providing the HepaRG cell line. We would like also to thank Sabine Rajkumar, Caroline Lecerf, Floriane Bard and
Audrey Criqui for acquisition of RPPA data. We warmly thank Aurélien de Reyniès for providing the R-package for the transcriptomic prediction of Hoshida’s and Boyault’s HCC molecular subclasses. We are grateful to Frédéric Soysouvanh for help in drug screening. We thank Stephen Miller from Blueprint Medicines for providing us the FGFR4 inhibitor
BLU9931. We thank all the clinician surgeons and pathologists who have participated to this work: Jean Saric,ACCEPTED Christophe Laurent, Laurence Chiche, Brigitte Le Bail, Claire Castain
(CHU Bordeaux), Alexis Laurent, Daniel Cherqui, Daniel Azoulay (CHU Henri Mondor,
Créteil, APHP), Marianne Ziol, Nathalie Ganne-Carrié and Pierre Nahon (Jean Verdier
Hospital, Bondy, APHP). We also thank the Réseau national CRB Foie (BB-0033-0085), the tumor banks of CHU Bordeaux (BB-0033-00036), Jean Verdier Hospital (APHP) and CHU
4 ACCEPTED MANUSCRIPT Henri Mondor (APHP) for contributing to the tissue collection. We warmly thank Yannick
Ladeiro, Giuseppe Maltese and Codebase for their very important support during the website building process.
MANUSCRIPT
ACCEPTED
5 ACCEPTED MANUSCRIPT ABSTRACT
Background and aims: Hepatocellular carcinomas (HCCs) are heterogeneous aggressive tumors with low rates of response to treatment at advanced stages. We screened a large panel of liver cancer cell lines (LCCLs) to identify agents that might be effective against HCC and markers of therapeutic response.
Methods: We performed whole-exome RNA and microRNA sequencing and quantification of 126 proteins in 34 LCCLs. We screened 31 anti-cancer agents for their ability to decrease cell viability. We compared genetic, RNA, and protein profiles of LCCLs with those of primary HCC samples and searched for markers of response.
Results: The protein and RNA signatures of the LCCLs were similar to those of the proliferation class of HCC, which is the most aggressive tumor type. Cell lines with alterations in genes encoding members of the Ras–MAPK signaling pathway and that required FGF19 signaling via FGFR4 for survival MANUSCRIPT were more sensitive to trametinib than to FGFR4 inhibitors. Amplification of FGF19 resulted in increased activity of FGF19 only in tumor cells that kept a gene expression pattern of hepatocyte differentiation. We identified single agents and combinations of agents that reduced viability of cells with features of the progenitor subclass of HCC. LCCLs with inactivating mutations in TSC1 and TSC2 were sensitive to the mTOR inhibitor rapamycin, and cells with inactivating mutation in TP53 were sensitive to the AURKA inhibitor alisertib. Amplification of MET was associated with hypersensitivity to cabozantinib and the combination of sorafenib and inhibitors of MEK1 and
MEK2 had a synergisticACCEPTED anti-proliferative effect.
Conclusion: LCCLs can be screened for drugs and agents that might be effective for treatment of HCC. We identified genetic alterations and gene expression patterns associated with response to these agents. This information might be used to select patients for clinical trials.
6 ACCEPTED MANUSCRIPT Keywords: MEK inhibitor, liver tumor, biomarker, response to therapy
MANUSCRIPT
ACCEPTED
7 ACCEPTED MANUSCRIPT INTRODUCTION
Hepatocellular carcinoma (HCC) is an aggressive malignancy with few therapeutic options at advanced stages. Since 2008, sorafenib was the only systemic therapy approved in first-line for unresectable HCC.1 Recently, four new agents demonstrated significant survival advantage in phase 3 clinical trials in first-line (lenvatinib) 2 and second-line (regorafenib, cabozantinib and ramucirumab)3–5 and the FDA has granted an accelerated approval to nivolumab for HCC patients after sorafenib failure.6 However, all these compounds provide only limited benefit in terms of survival (2-3 months) with low response rates and high inter- individual variability. These observations could be related to the high degree of molecular complexity and diversity of HCC, which usually play a critical role in determining variable tumor response of patients to treatment. Biological markers are increasingly used in clinical oncology for diagnosis, prognosis, and therapeutic decision-making and have helped to improve patient outcomes in various cancers. MANUSCRIPT In HCC, there are currently no validated predictive molecular markers available for the most effective systemic treatments.
Human tumor-derived cell lines have been widely used as models for studying cancer biology. They are very useful to understand mechanisms that drive resistance and sensitivity to anti-cancer compounds, in particular when access to tissue samples is limited as in HCC for which non-invasive imaging has replaced biopsy for diagnosis. 7 Over the past decade several large-scale pharmacogenomics studies in cancer cell lines including NCI-60, CCLE (Cancer
Cell Line Encyclopedia) and GDSC (Genomics of Drug Sensitivity in Cancer) have proven their value for ACCEPTED biomarker discovery as well as to uncover mechanism of drug action and determine molecular contexts associated with specific tumor vulnerabilities.8–10 Although these programs have provided a wealth of publically available data for the scientific community, HCC cell lines were under-represented in these different datasets. Given the molecular heterogeneity of HCC, the analysis of a large panel of cell lines recapitulating HCC
8 ACCEPTED MANUSCRIPT diversity may be more informative and may help to better translate in vitro pharmacogenomics findings into clinical application.
The goal of this study was to identify molecular features that are predictive of drug sensitivity in HCC, focusing mainly on candidate drugs already approved or in clinical development and targeting key pathways involved in hepatocarcinogenesis. To this aim, we used a collection of 34 human liver cancer derived cell lines characterized extensively at the genomic, transcriptomic and protein level. Our purpose was to understand pharmacological responses in the light of molecular profiles across the whole panel of liver cancer cell lines in order to identify 1) new potential attractive drugs in the treatment of HCC 2) molecular markers predicting their sensitivity.
MANUSCRIPT
ACCEPTED
9 ACCEPTED MANUSCRIPT MATERIALS AND METHODS
Cell lines
The 34 Liver cancer cell lines (LCCL) were collected from public repositories or collaborations and grown in monolayers in quite similar conditions, at 37°C in a humidified
5% CO 2 incubator (Supplementary Table 1). Cell line identity was verified by whole exome sequencing (WES), BEL7402 and SMMC7721 cell lines were excluded because contaminated by HeLa cells as well as SK-HEP-1 that has an endothelial origin.11
Identification of putative somatic variants and copy number analysis
Identification of gene mutations and copy number alterations (CNA) was performed by WES
(Supplementary Material). TERT promoter and exon 1 of ARID1A were screened by Sanger sequencing because of low coverage in WES, as previously described.12
Selection of cancer driver genes in primary HCC tumors and comparison with LCCL We selected 72 HCC cancer driver genes from MANUSCRIPT 5 HCC publically available datasets to compare their alteration frequency between HCC and LCCL (Supplementary Table 2A). See the Supplementary Material for detailed description.
Mutational signature analysis
We used the Palimpsest R package to extract mutational signatures from WES data using the open-source R package on Github: https://github.com/FunGeST/Palimpsest. 13
RNA-sequencing
Total mRNA extractionACCEPTED was performed for the 34 LCCL using RNeasy mini kit (Qiagen) and quality was checked. 5 g of total RNA was used for sequencing (Supplementary Material).
10 ACCEPTED MANUSCRIPT Gene fusion prediction and HBV integration
Fusions detected by TopHat2 (--fusion-search --fusion-min-dist 2000 --fusion-anchor-length
13 --fusion-ignore-chromosomes chrM) were filtered using the TopHatFusion-post algorithm.
We kept only fusions validated by BLAST and with at least 10 split-reads or pairs of reads spanning the fusion event, and we removed fusions identified at least twice in a cohort of 36 normal liver samples. HBV insertions were screened as described in the Supplementary
Material.
Transcriptome analysis
Consensus clustering was performed with Bioconductor ConsensusClusterPlus package.
Principal component analysis using the first 3 components was also generated.
The Bioconductor DESeq2 package was used to detect differentially expressed genes between the LCCL transcriptomic subgroups. Detailed analysis as well as Hoshida’s and Boyault’s prediction subclasses are described in the SupplemeMANUSCRIPTntary Material. miRNA profiling
Total miRNA extraction from 34 LCCL was performed using TRIzol reagent (Thermo Fisher
Scientific, Carlsbad, CA, USA) according to manufacturers' recommendations. miRNA profiling was performed using 1 µg of total RNA according to the protocol described in the
Supplementary Material.
Reverse-phase-protein array
126 specific antibodies (Supplementary Table 3) were analyzed on the 34 LCCL according to the protocol describedACCEPTED in the Supplementary Material.
11 ACCEPTED MANUSCRIPT Single and combination drug screening
We analyzed 31 therapeutic compounds on the whole panel of 34 LCCL (Supplementary
Table 4). Drug screening was performed using the HP D300 digital dispenser (Tecan) to create dose-response curves as described in the Supplementary Material.
Identification of biomarkers related to drug sensitivity
To identify molecular features associated with drug response we performed elastic net regression as described in the Supplementary Material.
Statistical analysis
Statistical analysis and data visualization were performed using both R software version 3.5.1
(R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org) with
Bioconductor packages and PRISM7 software (GraphPad Software, Inc., La Jolla, CA).
Comparison of a continuous variable in 2 or more than 2 groups was performed using either parametric test (t-test or ANOVA) if the variable MANUSCRIPTwas normally distributed or non-parametric test (Mann-Whitney or Kruskal-Wallis test). Qualitative data were compared using the
Fisher’s exact test and Chi-squared test to compare binary and non-binary categorical variables, respectively. Correlation analysis between continuous variables was performed using Pearson r correlation when both variables were normally distributed with the assumptions of linearity and homoscedasticity or Spearman's rank-order correlation. All tests were two-tailed and P-value < 0.05 was considered as significant.
Tumors
The LICA-FR andACCEPTED TCGA cohorts of HCC patients including respectively 156 and 319 tumors cases were previously described.15,16
12 ACCEPTED MANUSCRIPT RESULTS
Liver cancer cell lines retain the genomic alterations identified in HCC tumors
We qualified and analyzed a total of 34 liver cancer cell lines (LCCL) including 32 derived from HCC and 2 from hepatoblastoma (HepG2 and Huh6) (Supplementary Table 1) that were compared to 821 HCC primary tumors including our HCC cohort (n=235), 17 and 2 independent public datasets from Korea (n=231), 18 and mixed Asian/European origin
(n=355).16 Male and female repartition was similar in LCCL and HCC tumors, while LCCL were enriched in HBV and HCV infections and patients were younger (median age 51 versus
60 years, P<0.0001) compared to HCC (Figure 1).
WES analysis identified a higher median mutation rate in LCCL compared to the 821 HCC primary tumors for both synonymous (median=81 versus 24) and non-synonymous mutations
(median=339 versus 68) (Figure 1, Supplementary Figure 1A and Supplementary Table 5) that could be partly caused by the lack of availabl MANUSCRIPTe matched constitutional DNA resulting in undetectable germline rare variants. Mutational signature analysis taking into account the type of substitution and the trinucleotide context revealed that most of the mutations identified in
LCCL were related to signature 1 and 5 both associated with age (Supplementary Figure 2).
Signature 24, related to aflatoxin B1 exposure, was identified in PLC/PRF5 showing the highest number of mutations (n=949) from an African patient with a typical TP53 R249 mutation (Supplementary Figure 2).
In LCCL, we identified mutations in 59 out of 72 cancer driver genes recurrently mutated in
HCC from 5 publicallyACCEPTED available datasets (Figure 1, Supplementary Tables 2A-B and
Supplementary Material). Mutational patterns of these genes were fairly similar in LCCL compared to HCC primary tumors (Spearman r=0.59 P<0.0001) (Figure 1). Also, 16 genes showed more frequent mutations in LCCL, including mutations in TP53 , AXIN1 and
FGF19 /CCND1 amplification known to be associated with clinically aggressive HCC
13 ACCEPTED MANUSCRIPT classified in the “proliferation class”.19 In contrast, CTNNB1 mutations belonging the “non- proliferation class” were less frequent in LCCL (15%) compared with primary tumors (29%).
CNA analysis identified recurrent homozygous deletions of CDKN2A /MTAP and AXIN1 , and focal amplification containing CCND1 and FGF19 in LCCL as in HCC (Supplementary
Figure 1B). As previously described, in primary HCC major associations between risk factors and gene mutation were found between alcohol intake and CTNNB1 and TERT promoter mutations and HBV infection and TP53 mutations. In LCCL, we confirmed the significant association between HBV infection and TP53 mutations (Supplementary Figure 3). In addition, 300 fusion transcripts were identified by RNA sequencing across the 34 LCCL.
Among them, 51 involved cancer driver genes related to HCC (n=11) or to other cancer types
(n=40) (Supplementary Table 6). We also detected 33 chimeric HBV-human fusion transcripts in 13 LCCL with recurrent insertions in TERT promoter (3/34), as previously reported in HCC primary tumors (Supplementary Table 7).16,20 Overall, our panel of LCCL shared the most common genetic alterations identifi MANUSCRIPTed in primary tumors recapitulating the genomic landscape of the most aggressive HCCs.
Transcriptomic identified 3 LCCL subgroups recapitulating the most aggressive subclasses of HCC
Unsupervised consensus classification of the RNA-sequencing data enabled to classify robustly 33/34 LCCL (except JHH1) defining 3 subgroups, CL1-3, further confirmed by principal componentACCEPTED analysis (Figure 2A). CL1 included the most differentiated LCCL with epithelial features and an “hepatoblast-like” appearance, characterized by the expression of hepatocyte and liver fetal/progenitor markers (Figure 2B-C and Supplementary Tables 8-9).
CL3 subgroup included less differentiated LCCL with more invasive, proliferative
“mesenchymal-like” profile, expressing higher level of stem cell markers and EMT/metastasis
14 ACCEPTED MANUSCRIPT genes and low levels of hepato-specific genes (Figure 2B-C and Supplementary Table 8-9).
CL2 subgroup displayed a mixed “epithelial-mesenchymal” pattern between CL1 and CL3 with an intermediate expression of hepato-specific genes and stem cell markers and a mild enrichment in TSC2 and NFE2L2 mutations (Figure 2B-C, Supplementary Table 8-9,
Supplementary Figure 3). The distribution of the other HCC driver genes was not different between the 3 subgroups (Supplementary Figure 4). Using a centroid-based prediction method, we showed that LCCL transcriptomic subgroups were closely similar to the most aggressive HCC primary tumors subclasses previously established by Boyault (G1-G3) and
Hoshida (S1-S2) corresponding to the “proliferation class” (Figure 2A). 21,22 CL1 was mainly similar to Boyault’s-G1 and Hoshida’s-S2 HCC subclasses whereas CL2 and CL3 subgroups corresponded to the Boyault’s-G3 and Hoshida’s-S1 subclasses (Figure 2A-B). However, the
“non-proliferative”, most differentiated and less aggressive HCC class (G4-G6 and S3) was not represented in our panel of LCCL. LCCL transcriptomic profiles showed remarkable stability compared with those from the external GDS MANUSCRIPTC dataset (Supplementary Figure 5A).
miRNome and protein expression are closely associated with LCCL transcriptomic classification
We analyzed miRNA expression profiles by miRNA sequencing and the expression of
126 candidate proteins by RPPA in the 34 LCCL. Unsupervised classification of miRNA and protein expression profiles were strongly associated with the transcriptomic CL1-CL3 subgroups (FigureACCEPTED 3A-B). Five miRNA showed a significant differential expression between the 3 LCCL transcriptomic subgroups including miR-1257 and two members of the miR-122, and miR-194 family (Figure 3A and Supplementary Table 10). Except for miR-1257 showing the highest expression in CL1 and the lowest expression in CL2 subgroup, expression of miR-
122-5p, miR-122-3p, miR-194-5p and miR-194-3p was closely related to the degree of LCCL
15 ACCEPTED MANUSCRIPT differentiation with the highest and lowest level in CL1 and CL3 subgroups, respectively
(Figure 3A and Supplementary Table 10). Accordingly, miR-122-5p was reported as liver- specific and represents the most abundant miRNA in mature hepatocytes. 23 In the same line, miR-194-5p was described as a liver epithelial cell marker and its downregulation increased
EMT and HCC metastasis in preclinical models.24
The 126 proteins analyzed by RPPA included 82 total proteins and 44 phospho-proteins involved in various signaling pathways (Supplementary Table 3 and Supplementary Table
11). Messenger RNA and their corresponding protein expressions were closely related together (Spearman’s r=0.25 versus random protein/mRNA pairs r=0.002; P<0.0001, Mann-
Whitney test), comparable to previous studies on tumor samples (mean Spearman’s r=0.3 25 )
(Supplementary Figure 6 and Supplementary Table 12). We identified 19 proteins significantly differentially expressed between LCCL transcriptomic subgroups (Figure 3B and
Supplementary Figure 7). Protein expression profiles reflected the differentiation state of MANUSCRIPT LCCL with proteins expressed in mature hepatocytes (ALDH1A1, E-cadherin, p190A, RSK2, HER3, FGFR4) that were dramatically downregulated in CL3 subgroup and expression of the hepatic progenitor marker cytokeratin 19 that was higher in CL1 and CL2 subgroups (Figure
3B). Protein expression identified a more pronounced activation of the TGFß (TGFß-I-III and phospho-SMAD2/3) and the non-canonical ß-catenin pathways (phosphoSer675) in CL2 and
CL3 subgroups (Figure 3B).
We also identified two major networks of protein co-regulation in the whole panel of LCCL with proteins involvedACCEPTED in DNA repair, cell cycle, apoptosis and in the PI3K/AKT/mTOR pathway (Figure 3C). Finally, investigating relationship between protein expression and the mutational status of the HCC driver genes mutated in LCCL yielded 268 significant associations (Supplementary Table 13). As expected, cyclin D1 was overexpressed in LCCL harboring co-amplification of CCND1 and FGF19 . AXIN1 protein was overexpressed in the
16 ACCEPTED MANUSCRIPT CTNNB1 mutated LCCL. In addition, consistent with their inactivation, mutations in the tumor suppressor genes (TSG) TSC2 and AXIN1 were associated with the downregulation of the corresponding proteins, we also found a known association between inactivating mutations of KEAP1 and overexpression of NQO1 protein (Figure 4D).
Drug screening and molecular features associated with drug sensitivity
In our panel of 34 LCCL we screened 31 drugs including compounds approved or in clinical development in HCC or other cancers and targeting key pathways of liver tumorigenesis (Figure 4A and Supplementary Table 4 and 14). We also analyzed four drug combinations with sorafenib including the AKT inhibitor MK-2206, the HDAC inhibitor resminostat and the two MEK1/2 inhibitors trametinib and refametinib.
We showed that the most potent drugs were those that target general processes, such as proteasome, mitosis or protein folding (Figure 4A). Of note inhibitors targeting both PI3K/mTOR or mTOR alone were among the 7MANUSCRIPT most effective drugs with a median AUC close to doxorubicin, a chemotherapeutic agent used to treat HCC by transarterial chemoembolization (Figure 4A). Surprisingly, sorafenib, lenvatinib, cabozantinib and regorafenib that are multikinase inhibitors used in first or second line systemic treatment of
HCC, showed mild efficiency and were ranked at the 26th , 25 th , 31 th and 34 th position, respectively (Figure 4A). Remarkably, the MEK1/2 inhibitor trametinib showed the highest variable responses within the 34 LCCL. The combination of sorafenib either with the AKT inhibitor MK-2206 or the anti-HDAC resminostat produced an additive effect in inhibiting cell viability, asACCEPTED indicated by a median combination index (CI) of 1.1 and 1.2, respectively
(Figure 4D and Supplementary Table 14). Strikingly, sorafenib showed synergistic effect when combined with MEK1/2 inhibitors in around 60-70% of the cell lines (Figure 4D), that was more pronouced with trametinib (Figure 4A). Overall, there was a good correlation between AUC and GI50 values (Supplementary Figure 8) and, our drug sensitivity profiles
17 ACCEPTED MANUSCRIPT were well correlated with those from the external GDSC LCCL dataset (Supplementary
Figure 5B).
Unsupervised hierarchical clustering of drug responses on the whole series of 34 LCCL showed common sensitivity profiles for drugs with similar mechanism of action and identified two main subgroups of LCCL associated with transcriptomic subgroups (P<0.009) (Figure
4B). Accordingly, the global drug response rate was higher in the most differentiated CL1 subgroup, compared with CL2 and CL3 subgroups that were less differentiated and resistant to most of the analyzed compounds (Figure 4C). Of note, cabozantinib, targeting multiple kinases including c-MET showed a pharmacological profile close to the two selective MET inhibitors (PHA-665752 Spearman’s r=0.38, P=0.02 and JNJ-38877605 Spearman’s r=0.39,
P=0.02) (Figure 4B).
We identified 8 drugs and 3 combinations showing different sensitivity pattern according to the transcriptomic subgroups (Figure 4D). Among them, 6 drugs including cabozantinib MANUSCRIPT (multi-kinase inhibitor), linsitinib (anti-IGF1R), alvespimycin (anti-HSP90), JNJ-38877605 (anti-MET), nutlin 3 (anti-MDM2) and the combination sorafenib-MK2206, showed a higher sensitivity specifically in the CL1 subgroup. Of note, for the combination sorafenib-MK2206
(anti-AKT) the median CI was 0.9 in CL1 compared to 1.2 for CL2 and CL3 subgroups indicating a synergistic effect of the combination in at least half of the CL1 cell lines
(Supplementary Table 14). Refametinib (anti-MEK) and tanespimycin (anti-HSP90) were more efficient in CL1 compared with CL3 subgroup, while trametinib (anti-MEK1/2) showed higher efficiencyACCEPTED both in CL1 and CL2 subgroups. Combination of sorafenib either with trametinib or refametinib was more potent both in CL1 and CL2 subgroups. We also identified 312 drug-protein predictive pairs (Figure 4E and Supplementary Table 15). Among them, we validated the association between the high expression level of NQO1 and the high sensitivity to the two HSP90 inhibitors, alvespimycin and tanespimycin. 9,10,26 Expression of
18 ACCEPTED MANUSCRIPT cytokeratin 19 was strongly associated with higher sensitivity to dasatinib (src-inhibitor), in agreement with the higher dasatinib vulnerability in a “progenitor-like” subtype of LCCL.27
In our study, cytokeratin 19 expression was also associated with higher response to trametinib
(anti-MEK1/2) and navitoclax (anti Bcl-2, Bcl-XL, and Bcl-w).
We also identified 143 significant associations between genetic alterations and drug sensitivity (Supplementary Tables 16 and 17) among which, TSC1 /TSC2 inactivating mutations were linked with a higher sensitivity to the mTOR inhibitor rapamycin (Figure 4F).
In addition, we confirmed in our large panel of LCCL, the hypersensitivity to the AURKA inhibitor Alisertib in the TP53 -mutated cell lines (Figure 4F).28 Interestingly, the only MET - amplified cell line (MHCC97H, Figure 4D) was highly sensitivity to the two selective MET inhibitors (PHA-665752 and JNJ-38877605) as well as cabozantinib (Figure 4D and
Supplementary Table 14).
In order to explore among all the molecular features (genomic alterations, miRNA and mRNA expression), those that were the most associated wiMANUSCRIPTth drug response, we used elastic net (EN) regression. This analysis yielded a huge number of molecular markers linked to drug response with a median of 95 associated features per drug (min: 0, max: 139), when using an EN score
≥0.7 (Supplementary Figure 9 and Supplementary Table 18-19) and uncovered strong associations in particular, between sensitivity to the MEK1/2 inhibitors trametinib and refametinib and expression of HSD17B7 (see next paragraph). We could not identify strong predictors for lenvatinib and regorafenib because of their poor in vitro efficiency. However, elastic net analysisACCEPTED identified high mRNA level of PRMT5 and GPS1 as the best predictors of sorafenib sensitivity and STEAP2 and STEAP1 expression as strong predictors of cabozantinib sensitivity. In contrast, resistance to the two drugs was predicted by overexpression of UBE2H and SLC25A29 for sorafenib and NBAS and LAPTMA4 for cabozantinib (Supplementary Table 19). Among the different features analyzed, mRNA
19 ACCEPTED MANUSCRIPT expression showed the best predictive value, while miRNA expression and genomic alterations were poorly predictive of drug response, even when adjusting on the number of tested features (Supplementary Figure 9 and Supplementary Table 18).
Sensitivity to the MEK inhibitor trametinib is related to RAS-MAPK genomic alterations and cell differentiation
We focused our analysis on the MEK inhibitor trametinib, that showed a bimodal sensitivity pattern, with a group of highly sensitive and a group of resistant LCCL (Figure
5A). We identified oncogenic alterations known to activate the RAS-MAPK pathway in half of the LCCL (Figure 5A). All these genomic alterations were mutually exclusive, they affected 4 oncogenes (FGF19 and MET amplifications, NRAS Q61L mutation and ERBB4 fusion), and 2 TSG (RPS6KA3 and NF1 ) (Figure 5A-B and Supplementary Figure 10). We identified an enrichment of the RAS-MAPK pathwayMANUSCRIPT genomic alterations in the most sensitive LCCL to trametinib (10/14; 71%) compared to those that were resistant (7/20; 35%) while it did not reach statistical significance (P=0.08, Fisher’s exact test) (Figure 5A).
Intriguingly, among the 11 FGF19 -amplified cell lines, nearly half of them (5/11) were not sensitive to trametinib and they showed highly variable levels of FGF19 mRNA not explained by FGF19 copy number (Figure 5A and Supplementary Figure 10A). Strikingly, expression of FGF19 in LCCL was both dependent on gene amplification and on the transcriptomic subgroup with the highest expression in the most differentiated CL1 subgroup and the lowest expressionACCEPTED in the poorly differentiated CL3 subgroup both in amplified and non-amplified LCCL (Figure 5A). In contrast, CCND1 that was invariably co-amplified with
FGF19 , was overexpressed in all the amplified LCCL whatever the transcriptomic subgroup
(Supplementary Figure 10A). Moreover, the 3 other key components of the FGF19/FGFR4 pathway including FXR (encoded by NR1H4 ) known to transactivate FGF19 were co-
20 ACCEPTED MANUSCRIPT regulated together with FGF19 in LCCL (Figure 5C) . Accordingly, within the FGF19 - amplified cell lines, we identified a strong correlation between trametinib sensitivity and
FGF19 mRNA level (Figure 5D). A similar association was observed with the two FGFR4 inhibitors BLU-9931 and H3B-6527, which reinforce the link between trametinib sensitivity,
FGF19 amplification and expression (Figure 5D). Altogether, these results revealed that
FGF19 amplification solely is not sufficient to predict sensitivity to both trametinib and
FGFR4 inhibitors but the expression of a full pathway modulated by the context of differentiation is required.
Then, we searched for robust predictors of trametinib response among all the molecular features analyzed using elastic net regression. We identified 5 genes (HSD17B7 , RORC ,
MRPS14 , SERINC2 , LAD1) with a high mRNA expression associated with higher trametinib sensitivity, named “trametinib 5 gene-score” to predict accurately the response (Figure 5E and
Supplementary Table 19). We validated these results in the GDSC external dataset MANUSCRIPT (Supplementary Figure 11). In primary HCC, both in TCGA (n=319) and LICA-FR (n=156) datasets, the G3 and S1 transcriptomic subclasses expressed the lowest level of the signature
(Figure 6A and Supplementary Figure 12). In tumors, we also showed that the mean expression of the FGF19/FGFR4 pathway varied according to the transcriptomic subclasses of HCC in the two datasets with the highest expression in the G1 and S2 subclasses and the lowest expression in the G3 and S1 subclasses. Interestingly, expression of the “trametinib 5 gene-score” signature was highly correlated with the mean expression of the FGF19/FGFR4 pathway in bothACCEPTED datasets (Spearman’s r=0.4, P<0.0001 for TCGA cohort, Spearman’s r=0.33, P<0.0001 for LICA-FR cohort) (Figure 6A and Supplementary Figure 12).
In HCC primary tumors, 24 focal amplifications of the 11q13.3 region were identified in the
TCGA series, including 21 cases with both FGF19 and CCND1 amplification and 3 cases with CCND1 amplification alone (Figure 6B). As observed in LCCL, FGF19 mRNA
21 ACCEPTED MANUSCRIPT expression in tumors was related to both the gene amplification and the transcriptomic subclass, with no FGF19 overexpression when amplifications occurred in the G3 and S1 subclasses, showing lower expression of NR1H4 (Figure 6A-B). By contrast, as in cell lines,
CCND1 was invariably overexpressed in the amplified tumors whatever the transcriptomic subclass suggesting that only FGF19 expression was sensitive to the differentiation context.
Overall, G1/S2 subclasses of HCC may be the best candidate for a trametinib or anti-FGFR4 therapy while G3/S1 subclasses are unlikely to respond.
MANUSCRIPT
ACCEPTED
22 ACCEPTED MANUSCRIPT DISCUSSION
In the present study, we showed that our panel of LCCL recapitulates the diversity of the most aggressive “proliferation class” of HCC both at the genomic and transcriptomic levels and this is a unique tool to translate our understanding of liver cancer development into therapeutics (summarized in Figure 7).
By combining genomic, transcriptomic and protein profile analysis in our panel of 34 LCCL, we identified strong similarities with the established HCC molecular subclasses, suggesting that LCCL are representative models of primary tumors. We identified three robust transcriptomic subgroups of LCCL driven by the differentiation state and sharing features similar to those described in HCC tumors. The CL1 “hepatoblast-like” subgroup of cell lines expresses hepato-specific genes and fetal/progenitor markers, it corresponds to the
“progenitor subclass” of HCC.19 CL2 “mixed epithelial-mesenchymal” and CL3 “mesenchymal-like” subgroups were less differentiatMANUSCRIPTed with an activation of the TGFß and non-canonical ß-catenin pathways, they were more similar to the “Wnt-TGFß”22 and
Boyault’s G3 subclasses of HCC 21 (Figure 7). The high expression of EMT-metastasis genes is another feature of the CL2 and CL3 subgroups of LCCL that may result from the aberrant activation of the TGFß pathway.29 However, consistent with the underrepresentation of
CTNNB1 mutations in LCCL, the less aggressive “non-proliferation class” of HCC was not represented in our panel of LCCL and of note even those cell lines mutated for CTNNB1 did not properly reflect the CTNNB1 mutated subclass of primary HCC as they were all classified in the proliferationACCEPTED class and showed TP53 mutations, representing only a small and atypical fraction of primary HCC mutated for CTNNB1 (Supplementary Figure 13). These observations may be explained by the selection of the most aggressive phenotypes during cell line establishment that favors cells with the best survival capabilities. Additional LCCL are needed to represent the “non-proliferation class”, which should be feasible with the
23 ACCEPTED MANUSCRIPT development of new cell culture techniques such as those using Rho-kinases inhibitors.
Our findings provided novel insights regarding the crucial interplay between the differentiation context, the genetic alterations and drug response in HCC (summarized in
Figure 7). Strikingly, the global drug response rate among LCCL was related to the transcriptomic subgroup and the cell differentiation state with the most differentiated CL1 subgroup showing the highest drug sensitivity.
Cell differentiation also interferes with specific signaling pathways. FGF19 amplification is a very appealing therapeutic target with the development of selective inhibitors of its receptor,
FGFR4.30,31 However, FGF19 amplifications were not always associated with its overexpression both in LCCL and HCC tumors. We showed that only LCCL expressing a hepatocyte differentiation program with conserved expression of the FXR transcription factor and of the receptor complex, including FGFR4 and KLB, were sensitive to the FGFR4 inhibitors, which extends other previous findings. 30–32 Remarkably, this group of LCCL was also highly sensitive to inhibition of MEK1/2, an MANUSCRIPT effector downstream FGF19/FGFR4, with trametinib, which is already approved in BRAF -mutated melanoma.33 In LCCL, trametinib has demonstrated higher potency than the anti-FGFR4 BLU-9931 thereby representing a new attractive drug for targeting HCC addicted to the FGF19/FGFR4 pathway. Accumulating evidence indicate that the differentiation context plays a determinant role in treatment response, in particular it was shown that a therapy targeting a specific mutation will not necessarily have the same efficacy in tumors sharing the same mutation but arising in
34 different tissue types.ACCEPTED In this line, a recent report demonstrated that gene expression and the tissue of origin predicted much better drug sensitivity in pan-cancer cell line analysis than genetic alterations. 35 Here, we showed that this concept could be also generalized to tumors from the same organ. Interestingly, trametinib is also efficient in LCCL harboring other alterations in the RAS-MAPK pathway such as RPS6KA3 inactivation or MET amplification.
24 ACCEPTED MANUSCRIPT However, MET amplifications, even if they are rare events, are more efficiently targeted by specific MET inhibitors including cabozantinib and these findings could be translated in the clinics as we confirmed recently MET oncogenic addiction in a patient with an advanced
HCC amplified for MET, who achieved a complete tumor response after treatment by teponinib, a specific Met inhibitor.36 Our study has also highlighted a synergistic effect of sorafenib in combination with MEK1/2 inhibitors with higher sensitivity in the CL1 and CL2 subgroups of LCCL, in lines with recent studies in HCC preclinical models and in HCC patients.37,38
Our screening also identified potential new attractive drugs already approved in the clinics in specific molecular contexts. Our results showed responses in LCCL harboring inactivating mutations in TSC1 or TSC2 treated by an mTOR inhibitor, suggesting that the 7% of HCC demonstrating the same alterations could benefit from rapamycin or alternative inhibitors, in line with previous reports (Figure 7). 16,18,39–41 Dasatinib also showed enhanced efficiency in LCCL expressing high levels of cytokeratin 19.27 MANUSCRIPT We recently reported a specific enrichment of immunohistochemical expression of CK19 in the “progenitor subclass” of HCC which may represent a good candidate for dasatinib therapy. 41 Additionally, we identified other potential drugs that may specifically target the “progenitor subclass” of HCC such as linsitinib, an inhibitor of IGF1R, or sorafenib in combination with the anti-AKT MK-2204. Indeed, linsitinib hypersensitivity in the CL1 subgroup of LCCL, that recapitulates the “progenitor subclass” of HCC, is consistent with the strong overexpression of IGF2. Accordingly, a recent work in transgenicACCEPTED mice demonstrated the pro-oncogenic role of IGF2 in the liver and showed that blocking IGF2 by an antibody efficiently impaired growth of liver tumor cells overexpressing IGF2, in vitro and in vivo .42 In the present work, we also enlighten the specific vulnerability of TP53 -mutated LCCL to the AURKA inhibitor alisertib corroborating a recent study in mice.28 Thus, alisertib may represent a new therapeutic opportunity for P53 mutated
25 ACCEPTED MANUSCRIPT HCC patients, as TP53 is the most frequently mutated TSG in HCC and, until now, was considered to be undruggable. The lack of biomarker-driven clinical trials may partly explain why some drugs that appear to be effective in vitro such as rapamycin, tivantinib or dasatinib failed in patients. Unexpectedly, gold standard therapies for advanced HCC including sorafenib, lenvatinib, cabozantinib and regorafenib showed poor efficiency in LCCL suggesting that they have only limited anti-proliferative effect on liver tumor cells but more likely target tumor microenvironment. Accordingly, all the effective drugs in HCC share an anti-angiogenic activity that could not be explored in vitro and represents a limitation in the use of cellular models.
In conclusion, our work showed that LCCL represent a valuable and powerful resource for drug-biomarker discovery that may be useful to guide future clinical trials.
Moreover, this study provides a comprehensive molecular characterization of the most widely used LCCL that are freely accessible on our website: http://lccl.zucmanlab.com MANUSCRIPT
ACCEPTED
26 ACCEPTED MANUSCRIPT REFERENCES
1. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular carcinoma. N Engl J Med 2008;359:378–390.
2. Kudo M, Finn RS, Qin S, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial. Lancet 2018;391:1163–1173.
3. Bruix J, Qin S, Merle P, et al. Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo- controlled, phase 3 trial. Lancet 2017;389:56–66.
4. Zhu AX, Kang Y-K, Yen C-J, et al. Ramucirumab after sorafenib in patients with advanced hepatocellular carcinoma and increased α-fetoprotein concentrations (REACH-2): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 2019.
5. Abou-Alfa GK, Meyer T, Cheng A-L, et al. Cabozantinib in Patients with Advanced and Progressing Hepatocellular Carcinoma. N Engl MANUSCRIPT J Med 2018;379:54–63. 6. El-Khoueiry AB, Sangro B, Yau T, et al. Nivolumab in patients with advanced hepatocellular carcinoma (CheckMate 040): an open-label, non-comparative, phase 1/2 dose escalation and expansion trial. Lancet 2017;389:2492–2502.
7. Bruix J, Reig M, Rimola J, et al. Clinical decision making and research in hepatocellular carcinoma: pivotal role of imaging techniques. Hepatology 2011;54:2238–
2244. 8. ShoemakerACCEPTED RH. The NCI60 human tumour cell line anticancer drug screen. Nat Rev Cancer 2006;6:813–823.
9. Barretina J, Caponigro G, Stransky N , et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 2012;483:603–607.
10. Garnett MJ, Edelman EJ, Heidorn SJ , et al. Systematic identification of genomic
27 ACCEPTED MANUSCRIPT markers of drug sensitivity in cancer cells. Nature 2012;483:570–575.
11. Rebouissou S, Zucman-Rossi J, Moreau R, et al. Note of caution: Contaminations of hepatocellular cell lines. J Hepatol 2017;67:896–897.
12. Nault JC, Mallet M, Pilati C, et al. High frequency of telomerase reverse-transcriptase promoter somatic mutations in hepatocellular carcinoma and preneoplastic lesions. Nat
Commun 2013;4:2218.
13. Shinde J, Bayard Q, Imbeaud S, et al. Palimpsest: an R package for studying mutational and structural variant signatures along clonal evolution in cancer. Bioinformatics
2018;34:3380–3381.
14. Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47.
15. Bayard Q, Meunier L, Peneau C, et al. Cyclin A2/E1 activation defines a hepatocellular carcinoma subclass with a rearrangement signature of replication stress. Nat Commun 2018;9:5235. MANUSCRIPT 16. Cancer Genome Atlas Research Network. Electronic address: [email protected],
Cancer Genome Atlas Research Network. Comprehensive and Integrative Genomic
Characterization of Hepatocellular Carcinoma. Cell 2017;169:1327-1341.e23.
17. Schulze K, Imbeaud S, Letouzé E , et al. Exome sequencing of hepatocellular carcinomas identifies new mutational signatures and potential therapeutic targets. Nat Genet
2015;47:505–511. 18. Ahn S-M,ACCEPTED Jang SJ, Shim JH, et al. Genomic portrait of resectable hepatocellular carcinomas: implications of RB1 and FGF19 aberrations for patient stratification. Hepatology
2014;60:1972–1982.
19. Llovet JM, Montal R, Sia D, et al. Molecular therapies and precision medicine for hepatocellular carcinoma. Nat Rev Clin Oncol 2018;15:599–616.
28 ACCEPTED MANUSCRIPT 20. Zhao L-H, Liu X, Yan H-X, et al. Genomic and oncogenic preference of HBV integration in hepatocellular carcinoma. Nat Commun 2016;7:12992.
21. Boyault S, Rickman DS, Reyniès A de, et al. Transcriptome classification of HCC is related to gene alterations and to new therapeutic targets. Hepatology 2007;45:42–52.
22. Hoshida Y, Nijman SMB, Kobayashi M, et al. Integrative transcriptome analysis reveals common molecular subclasses of human hepatocellular carcinoma. Cancer Res
2009;69:7385–7392.
23. Deng X-G, Qiu R-L, Wu Y-H, et al. Overexpression of miR-122 promotes the hepatic differentiation and maturation of mouse ESCs through a miR-122/FoxA1/HNF4a-positive feedback loop. Liver Int 2014;34:281–295.
24. Meng Z, Fu X, Chen X, et al. miR-194 is a marker of hepatic epithelial cells and suppresses metastasis of liver cancer cells in mice. Hepatology 2010;52:2148–2157.
25. Akbani R, Ng PKS, Werner HMJ , et al. A pan-cancer proteomic perspective on The Cancer Genome Atlas. Nat Commun 2014;5:3887. MANUSCRIPT 26. Kelland LR, Sharp SY, Rogers PM, et al. DT-Diaphorase expression and tumor cell sensitivity to 17-allylamino, 17-demethoxygeldanamycin, an inhibitor of heat shock protein
90. J Natl Cancer Inst 1999;91:1940–1949.
27. Finn RS, Aleshin A, Dering J, et al. Molecular subtype and response to dasatinib, an
Src/Abl small molecule kinase inhibitor, in hepatocellular carcinoma cell lines in vitro.
Hepatology 2013;57:1838–1846. 28. Dauch D,ACCEPTED Rudalska R, Cossa G, et al. A MYC-aurora kinase A protein complex represents an actionable drug target in p53-altered liver cancer. Nat Med 2016;22:744–753.
29. Zavadil J, Böttinger EP. TGF-beta and epithelial-to-mesenchymal transitions.
Oncogene 2005;24:5764–5774.
30. Joshi JJ, Coffey H, Corcoran E, et al. H3B-6527 Is a Potent and Selective Inhibitor of
29 ACCEPTED MANUSCRIPT FGFR4 in FGF19-Driven Hepatocellular Carcinoma. Cancer Res 2017;77:6999–7013.
31. Hagel M, Miduturu C, Sheets M, et al. First Selective Small Molecule Inhibitor of
FGFR4 for the Treatment of Hepatocellular Carcinomas with an Activated FGFR4 Signaling
Pathway. Cancer Discov 2015;5:424–437.
32. Gao Q, Wang Z-C, Duan M, et al. Cell Culture System for Analysis of Genetic
Heterogeneity Within Hepatocellular Carcinomas and Response to Pharmacologic Agents.
Gastroenterology 2017;152:232-242.e4.
33. Flaherty KT, Robert C , Hersey P, et al. Improved survival with MEK inhibition in
BRAF-mutated melanoma. N Engl J Med 2012;367:107–114.
34. Prahallad A, Sun C, Huang S , et al. Unresponsiveness of colon cancer to
BRAF(V600E) inhibition through feedback activation of EGFR. Nature 2012;483:100–103.
35. Iorio F, Knijnenburg TA, Vis DJ, et al. A Landscape of Pharmacogenomic Interactions in Cancer. Cell 2016;166:740–754. 36. Nault JC, Martin Y, Caruso S, et al. Clinical MANUSCRIPT impact of genomic diversity from early to advanced hepatocellular carcinoma. Hepatology 2019:In Press.
37. Wang C, Jin H, Gao D, et al. Phospho-ERK is a biomarker of response to a synthetic lethal drug combination of sorafenib and MEK inhibition in liver cancer. J Hepatol
2018;69:1057–1065.
38. Lim HY, Merle P, Weiss KH, et al. Phase II Studies with Refametinib or Refametinib plus Sorafenib in Patients with RAS-Mutated Hepatocellular Carcinoma. Clin Cancer Res 2018;24:4650–4661.ACCEPTED 39. Schulze K, Imbeaud S, Letouzé E, et al. Exome sequencing of hepatocellular carcinomas identifies new mutational signatures and potential therapeutic targets. Nat Genet
2015;47:505–511.
40. Ho DWH, Chan LK , Chiu YT, et al. TSC1/2 mutations define a molecular subset of
30 ACCEPTED MANUSCRIPT HCC with aggressive behaviour and treatment implication. Gut 2017;66:1496–1506.
41. Calderaro J, Couchy G, Imbeaud S, et al. Histological subtypes of hepatocellular carcinoma are related to gene mutations and molecular tumour classification. J Hepatol
2017;67:727–738.
42. Martinez-Quetglas I, Pinyol R, Dauch D, et al. IGF2 Is Up-regulated by Epigenetic
Mechanisms in Hepatocellular Carcinomas and Is an Actionable Oncogene Product in
Experimental Models. Gastroenterology 2016;151:1192–1205.
Author names in bold designate shared co-first authorship
MANUSCRIPT
ACCEPTED
31 ACCEPTED MANUSCRIPT FIGURES LEGENDS
Figure 1. Mutational landscape of driver genes in 34 LCCL compared with HCC primary tumors. Top panel shows total (putative somatic) mutation rate for each cell line.
The heatmap represents mutations and CNA of the 59 HCC-associated genes mutated in
LCCL. On the left, histograms showing comparison of gene alteration frequency between
LCCL and HCC (Fisher’s exact test, P-value: * <0.05; ** <0.001; *** <0.0001 and
Spearman’s correlation). In the bottom, pie charts and histograms comparing the distribution of viral infection, gender (Fisher’s exact test) and age (Mann-Whitney test) between HCC and
LCCL.
Figure 2. Transcriptomic analysis of 34 LCCL. A) Consensus clustering and principal component analysis (PCA) of LCCL mRNA expression profiles for the optimal number of clusters at k=3 and association with Hoshida’s and Boyault’s HCC transcriptomic subclasses (Chi-square test). B) Gene set enrichment analysis MANUSCRIPT (GSEA) showing the four main categories of gene sets enriched in each group of LCCL (FDR<0.01; normalized enrichment score
(ES)>4 (red bars) in at least one LCCL group) (see also Supplementary Table 8). C)
Expression pattern of genes related to liver differentiation, EMT-metastasis and proliferation differentially expressed between the 3 subgroups of LCCL (ANOVA test) (see also
Supplementary Table 9). Horizontal line represents the median.
Figure 3. miRNAACCEPTED and protein profiles analysis in 34 LCCL. A) Left panel, consensus clustering of LCCL miRNA expression profiles (at k=3) and association with transcriptomic subgroups (Chi-square test). Right panel, number of miRNA differentially expressed between each transcriptomic subgroup of LCCL, median expression (horizontal line) per subgroup is shown for the 5 differentially expressed miRNAs between the 3 LCCL subgroups and
32 ACCEPTED MANUSCRIPT expression per LCCL is represented on the heatmap (see also Supplementary Table 10). B)
Left panel, hierarchical clustering of LCCL protein profiles and association with transcriptomic subgroups (Chi-square test). Right panel, boxplots of the 19 proteins differentially expressed between LCCL transcriptomic subgroups (Mann-Whitney test) (see also Supplementary Table 11 and Supplementary Figure 5). C) Protein interaction network showing the most significant positive (red lines) and negative (blue dashed lines) correlations between protein pairs across LCCL. D) Volcano plot comparing protein expression according to the mutational status of HCC driver genes. Blue and red points represent protein-gene mutation interaction, and indicate proteins underexpressed and overexpressed in LCCL when the gene (in italics) is mutated, respectively. Only interactions with a P-value <0.01 and a
Log 2 ratio>1 are shown (see also Supplementary Table 13).
Figure 4. Drug responses and associated molecular features in 34 LCCL. A) Top panel, pathways and biological processes targeted by theMANUSCRIPT panel of tested drugs. Below, boxplots showing for each drug the distribution of sensitivity values across LCCL. + indicates the mean. Bold: drugs approved in HCC. B) Hierarchical clustering showing patterns of sensitivity to 31 drugs and 4 combinations in LCCL and association with transcriptomic subgroups (Chi-square test). AUC values in rows were centered and scaled (z-score). At the bottom, boxplots comparing Spearman’s rank correlation coefficients (left) and P-values
(right) between AUC of drug pairs with identical or different molecular targets. C) Violin plots representingACCEPTED the distribution of AUC values for the whole panel of drugs according to LCCL transcriptomic subgroup. Tin black line: SD; black dot: median. D) Left panel, boxplots for the 11 drugs/combinations with different sensitivity profiles between LCCL transcriptomic subgroups (Red dot: MHCC97H, MET -amplified cell line). Right panel, the
Combination Index distribution is shown for the four drug combinations (see also
33 ACCEPTED MANUSCRIPT Supplementary Table 14). E) Volcano plot of Spearman’s correlations and significance between drug sensitivity (AUC) and protein expression. Blue and Red dots indicate respectively the most significant negative and positive correlations ( P-value<0.05, and
Spearman’s r>0.5) (see also Supplementary Table 15). F) Sensitivity to rapamycin and alisertib and mutational status of TSC1/TSC2 and TP53 genes (see also Supplementary Table
16).
Statistical difference between groups was determined by a Mann-Whitney test in panels B, C,
D and F.
Figure 5. Alteration of the RAS-MAPK pathway and trametinib sensitivity in 34 LCCL.
A) Trametinib sensitivity, genomic alterations of the RAS-MAPK and their consequence on mRNA and protein expression in LCCL (see also Supplementary Figure 6). Difference in
FGF19 mRNA according to the gene amplification status and the transcriptomic subgroups was assessed by a two-way ANOVA. B) Schematic MANUSCRIPT representation of oncogenic alterations of the RAS-MAPK pathway identified in LCCL and drugs analyzed in the present study (H3B-
6527 was evaluated in another study 30 ). C) mRNA expression of the key components of the
FGF19/FGFR4 pathway, difference between groups was assessed using an ANOVA test and correlations between each component of the pathway by a Pearson’s test. D) Correlation between sensitivity of drugs targeting MEK1/2 or FGFR4 and FGF19 mRNA expression in
FGF19 amplified LCCL (Spearman’s test). E) Volcano plot and heatmap showing the 10-top mRNA predictiveACCEPTED of trametinib response identified by EN regression (see also Supplementary Table 19). On the right below, correlation between trametinib response and the mean expression of the 5 mRNA (green) overexpressed in sensitive LCCL (“trametinib 5-gene score”) (Spearman’s test).
Variance stabilized values were used for mRNA level except for MET and ERBB4 .
34 ACCEPTED MANUSCRIPT Figure 6. Transcriptional expression of the “trametinib 5-gene score” and the key components of the FGF19/FGFR4 pathway in HCC according to the transcriptomic classifications . A) Boxplots showing expression of the “trametinib 5-gene score” and the mean mRNA level of the FGF19/FGFR4 pathway according to Boyault’s and Hoshida’s transcriptomic classifications; below, mean expression level +/- SD is shown for each component of the FGF19/FGFR4 pathway. The color scale below each boxplot indicates mRNA level per sample (Red: high expression, Blue: low expression), ranked in each transcriptomic subclass by the “trametinib 5-gene score”. Differences between transcriptomic subclasses were assessed using ANOVA. Variance stabilized values were used for mRNA level. B) On the top: schematic representation of the 11q13.3 genomic region containing
CCND1 and FGF19 . At the bottom, Tukey boxplots showing mRNA expression of CCND1 and FGF19 according to their amplification status and HCC transcriptomic subclass.
Figure 7. Summary of HCC molecular classesMANUSCRIPT previously established and their corresponding LCCL subgroups with the main drug/biomarker pairs associations identified in the present study. HCC molecular classes and associated features were extracted from previous reports.19,41 Ampl: amplification; mut: mutation
ACCEPTED
35 Figure 1
ACCEPTED MANUSCRIPT34 Liver cancer cell lines
10001000 SynonymousSynonymous 800800 Non-synonymousNon-synonymous 600600 400 Mutations 400 # # Mutations 200 # Mutations 200 Nb Nb HCC primary Liver cancer 0 0 Histological diagnosis tumors (n=821) cell lines (n=34) Gender
Tumors Cell lines Ethnicity
50 0 50 100 100 Alteration frequency (%) HBV
100 50 0 50 100 Comut Plot HCV
Promoter
Promoter TP53TP53 (85%) (85%)
CNV alteration
CNV alteration*** TERTTERT (71%) (71%)
Mutations/Indels Mutations/Indels AXIN1AXIN1 (38%) (38%) *** FGF19FGF19 (32%) (32%) *** CCND1CCND1 (32%) (32%) *** KMT2DKMT2D (24%) (24%) ** ARID1AARID1A (21%) (21%) * CDKN2ACDKN2A (18%) (18%) * TSC2TSC2 (18%) (18%) * AKAP9AKAP9 (18%) (18%) * ATMATM (18%) (18%) * APOBAPOB (15%) (15%) KEAP1KEAP1 (15%) (15%) * CTNNB1CTNNB1 (15%) (15%) NCOR1NCOR1 (12%) (12%) * NF1NF1 (12%) (12%) * TSC1TSC1 (12%) (12%) * MTAPMTAP (12%) (12%) FLT4FLT4 (12%) (12%) * FAT4FAT4 (12%) (12%) NFE2L2NFE2L2 (9%) (9%) LRP1BLRP1B (9%) (9%) CDKN1ACDKN1A (9%) (9%) ARID2ARID2 (9%) (9%) ZNF521ZNF521 (9%) (9%) ZFHX3ZFHX3 (9%) (9%) PREX2PREX2 (9%) (9%) NUP98NUP98 (9%) (9%) Spearman r=0.59 * MUC1MUC1 (9%) (9%) P<0.0001 EYSEYS (9%) (9%) ALBALB (9%) (9%) SETD2SETD2 (6%) (6%) RPS6KA3RPS6KA3 (6%) (6%) PDE4DIPPDE4DIP (6%) (6%) NBEANBEA (6%) (6%) JAK3JAK3 (6%) (6%) FOXP1FOXP1 ( 6%)(6%) CREBBPCREBBP (6 (6%)%) ARAR ( 6(6%)%) ACVR2AACVR2A (6 (6%)%) CDK6CDK6 (3%) (3%) ZNRF3ZNRF3 (3%) (3%) TBL1XR1TBL1XR1 (3 (3%)%) RB1RB1 (3 (3%)%) PTPRBPTPRB (3 (3%)%) PTENPTEN (3 (3%)%) NRASNRAS (3 (3%)%) NCOR2NCOR2 (3 (3%)%) MYCMYC (3 (3%)%) METMET (3 (3%)%) Mutations/Indels LZTR1LZTR1 (3 (3%)%) HNRNPA2B1HNRNPA2B1 (3 (3%)%) MANUSCRIPT CNA HNF1AHNF1A (3 (3%)%) G6PC (3%) G6PC (3%) Promoter COL3A1COL3A1 (3 (3%)%) CLTCCLTC (3 (3%)%) CD274CD274 (3 (3%)%) BRD4BRD4 (3 (3%)%) Mutations/Indels Mutations/Indels BAP1BAP1 (3 (3%)%) CNV alteration CNV alteration HBV pos 3 B1 Li7 Promoter HBVPromoter pos HBV neg - HLF B1 HLE Li7 1.1 1.2 Huh7 Huh1 Huh6 JHH5 JHH6 JHH7 JHH4 JHH2 JHH1 38% - - HLF HCC-3 HBV neg Hep3B HLE HBV P=0.07 HepG2 SNU368 SNU423 SNU354 SNU475 SNU878 SNU387 SNU761 SNU886 SNU739 SNU398 Mahlavu SNU449 SNU182 55% Huh7 Huh1 Huh6 HCC-1.1 HCC-1.2 HepaRG JHH5 JHH6 JHH7 JHH4 JHH2 JHH1
PLC/PRF5
MHCC97H HCC Hep3B
100 50 0 50 100 HepG2 Tumors Cell lines Mahlavu SNU368 SNU423 SNU354 SNU475 SNU878 SNU387 SNU761 SNU886 SNU739 SNU398 SNU449 SNU182 HCC HCC Sample HepaRG PLC/PRF5 n=802 n=33 HCV pos MHCC97H Truncating Missense Homozygous_Deletion HCV neg Promoter In-Frame Focal_Amplification 17% HCV pos HCV33% neg Clinical annotations: Mutation type: HCV P=0.07 Hepatocellular Carcinoma Truncating Hepatoblastoma n=802 n=21 Promoter Male Male Male Missense Female Female Female In-frame Asian Gender P=0.22 Homozygous deletion 85% European 74% Focal amplification African-American n=820 n=33 African Positive 125 12 Negative 100 10 75 8 6 50 4 25
# patients 2 # patients ACCEPTED Nb patients 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Age (years) [Bin center] Figure 2
A Consensus clustering of 34 LCCL (2000 mRNA) ACCEPTEDConsensus matrix MANUSCRIPT k=3 B GSEA analysis Color Key
consensus matrix k=3 CL1 CL2 CL3 Liver differentiation EMT/metastasis -1 -0.5 0 0.5 1 Value Proliferation Hoshida’s class 1 Normalized ES 2 3 2.3 (min) ; 8.9 (max) GENE SET LCCL subgroupGroup CL1 CL2 CL3 Frequency of HSIAO_LIVER_SPECIFIC_GENES Hoshida PHoshida<0.0001 Classification S2 S1 HALLMARK_XENOBIOTIC_METABOLISM Boyault P=0.01 co-classification 1 CAIRO_LIVER_DEVELOPMENT_DN G1 0.9 OHGUCHI_LIVER_HNF4A_TARGETS_DN G2 0.8 SU_LIVER 0.7 G3 0.6 HALLMARK_BILE_ACID_METABOLISM 0.5 KEGG_METABOLISM_OF_XENOBIOTICS_BY_CYTOCHROME_P450 0.4 0.3 CAIRO_HEPATOBLASTOMA_DN 0.2 YAMASHITA_LIVER_CANCER_STEM_CELL_DN 0.1 REACTOME_PHASE1_FUNCTIONALIZATION_OF_COMPOUNDS Z-score 0 KEGG_DRUG_METABOLISM_CYTOCHROME_P450 1 KEGG_COMPLEMENT_AND_COAGULATION_CASCADES 0.5 SERVITJA_LIVER_HNF1A_TARGETS_DN 0 HALLMARK_E2F_TARGETS REACTOME_DNA_REPLICATION -0.5 Transcriptomic Group REACTOME_MITOTIC_M_M_G1_PHASES -1 PCA MARKEY_RB1_ACUTE_LOF_UP ZHOU_CELL_CYCLE_GENES_IN_IR_RESPONSE_6HR CHANG_CYCLING_GENES ● ZHOU_CELL_CYCLE_GENES_IN_IR_RESPONSE_24HR WHITFIELD_CELL_CYCLE_LITERATURE
30 30 ● ● ● HALLMARK_G2M_CHECKPOINT ● ● GEORGES_CELL_CYCLE_MIR192_TARGETS 20 20 ● ● ● ● SARRIO_EPITHELIAL_MESENCHYMAL_TRANSITION_DN ● ● 10 10 ● ● HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION ● ● ● ● ● ● 0 0 ONDER_CDH1_TARGETS_2_UP ● PC3 PC3 ● SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP
10 ● ● -10− ● 40 ● CHANG_CORE_SERUM_RESPONSE_UP ● 40 PC2 ● ● 30
20 ● − 30 KEGG_FOCAL_ADHESION -20 ● 20 ● ● 10 20 ● SARRIO_EPITHELIAL_MESENCHYMAL_TRANSITION_UP 30 10 -30− 0 −10 0 WINNEPENNINCKX_MELANOMA_METASTASIS_UP B1 Li7
40 -10 HLF HLE −20
Huh1 Huh6 Huh7 -40 − JHH5 JHH7 JHH2 JHH1 JHH4 JHH6 HOSHIDA_LIVER_CANCER_SUBCLASS_S1 HCC-3 Hep3B HepG2 3 -20
SNU368 SNU878 SNU761 SNU354 SNU886 Mahlavu SNU387 SNU423 SNU475 SNU739 SNU182 SNU398 SNU449 −60 −40 −20 0 20 40 HCC-1.2 HCC-1.1 HepaRG -
B1 -60 20 PLC/PRF5 -40 0 40 MHCC97H Li7 -20 1.2 1.1
- - HOSHIDA_LIVER_CANCER_SUBCLASS_S2 HLF HLE
Huh1 Huh6 Huh7 PC1 JHH5 JHH7 JHH2 JHH1 JHH4 JHH6 PC1 HCC Hep3B HepG2 SNU368 SNU878 SNU761 SNU354 SNU886 Mahlavu SNU387 SNU423 SNU475 SNU739 SNU182 SNU398 SNU449 HCC HepaRG HCC PLC/PRF5 MHCC97H C Hepatocyte Progenitor Stem cell markers markers markers 1.2 3 1.1 - - -
PLC/PRF5 B1 HCC Huh1 Huh6 JHH5 HepG2 Hep3B HCC Huh7 JHH7 Li7 SNU878 SNU761 HepaRG JHH2 SNU354 MHCC97H SNU368 SNU886 JHH6 SNU398 SNU449 SNU423 Mahlavu JHH4 SNU182 SNU475 SNU387 HLE HLF HCC SNU739 JHH1 Hepatocyte Liver progenitor Stem cell CL1 CL2 CL3 P<0.0001 P<0.0001 P<0.0001 CPS1 14 14 12 Z-score CDH1 CEBPA 3.08 ALB 12 12 10 FGFR4 KLB 10 10 8 -0.1 CYP3A7 HNF1A 8 8 APOB SERPINA1 6 -3.18 HNF4A 6 MANUSCRIPT6 PROX1 4 AFP 4 4 EPCAM EMT/metastasis Proliferation IGF2 [variancestabilized values] mRNA [RNA-seq] mRNA [RNA-seq] 2 mRNA [RNA-seq] 2 SALL4 2 PROM1 markers markers
LGR5 level GPC3 EMT/metastasis Proliferation LIN28B CD47 P<0.0001 P=0.008 LCCL subgroup CD44 14 16 KDR IL6 CL1 (n=11) NCAM2 12 15 CL2 (n=9) THY1 CL3 (n=13) KIT 10
SNAI2 markersmRNA 14 ITGB3 TWIST1 8 ZEB2 13 Unclassified (JHH1) PLAUR 6 VIM MKI67 subclass 4 12 mRNA [RNA-seq] Mesenchymal mRNA [RNA-seq] Epithelial 2 11 Mean Pluripotency/proliferation
ACCEPTED Figure 3
A Consensus clustering of 34 LCCL (250 miRNA) miRNA differentially expressed (FDR<0.05) Color Key LCCL transcriptomic group ACCEPTED MANUSCRIPTCL1 vs CL2 CL2 vs CL3 miR-122-5p miR-122-3p miR-1257 CL1 Unclassified 20 15 CL2 10 -1 -0.5 0 0.5 1 Value 10 CL3 15 8 7 col1 5 P=0.001 105 16 10 6 col2 Z-score 0 4 1 5 miRNA expression miRNA expression 0.5 5 miRNA expression -5 2 0 101 24 miR-194-5p miR-194-3p -0.5 20 10 -1 hsa-miR-122-5p 15 consensus matrix k=3 Consensus matrix k=3 105 hsa-miR-122-3p 5 consensus matrix k=3 hsa-miR-194-5p 10 1 2 hsa-miR-194-3p 0 13 2 5 3 CL1 vs CL3 hsa-miR-1257 miRNA expression 0 miRNA expression -5
Frequency of CL1 CL2 CL3 co-classification Z-score hsa-miR-122-5p 1 3 hsa-miR-122-3p 2 0.5 1 hsa-miR-194-5p 0 hsa-miR-194-3p 0 -1 hsa-miR-1257 3 - B1 Li7 B1 1.1 1.2 Li7 - - HLF HLE HLF HLE Huh1 Huh6 Huh7 JHH4 JHH2 JHH6 JHH1 JHH7 JHH5 Huh7 Huh6 Huh1 JHH7 JHH5 JHH2 JHH6 JHH4 JHH1 HCC Hep3B HepG2 HCC-3 Hep3B HepG2 Mahlavu SNU475 SNU739 SNU182 SNU398 SNU761 SNU423 SNU449 SNU387 SNU886 SNU354 SNU878 SNU368 HCC HepaRG HCC SNU354 SNU368 SNU761 SNU878 SNU886 SNU423 SNU182 SNU449 SNU739 SNU398 SNU475 Mahlavu SNU387 B1 HCC-1.2 HCC-1.1 HepaRG Li7 HLF HLE Huh1 Huh6 Huh7 JHH4 JHH2 JHH6 JHH1 JHH7 JHH5 HCC-3 Hep3B HepG2 PLC/PRF5 MHCC97H PLC/PRF5 Mahlavu SNU475 SNU739 SNU182 SNU398 SNU761 SNU423 SNU449 SNU387 SNU886 SNU354 SNU878 SNU368 HCC-1.1 HCC-1.2 HepaRG MHCC97H PLC/PRF5 MHCC97H
Hierarchical clustering of 34 LCCL (126 proteins) B LCCL transcriptomic group Cytokeratin-19 ALDH1A1 E-cadherin p190A PKA-C-alpha RSK2 FGFR4 P-H2AX Color Key 1 1 2 1 3 CL1 Unclassified 4 P=0.03 P=0.001 P<0.0001 P=0.02 P=0.02 P=0.02 P<0.0001 P=0.007 P=0.02 P=0.01 P=0.007 P=0.02 P=0.02 2 P=0.01 P=0.01 -1 0 1 CL2 3 1 2 P=0.01 3 Value 0 CL3 0 0 1 2 0 1 P=0.004 0 2 P.MSK1.Ser360 -1 PDGFR.alpha 1 RSK2 PTEN -1 P.EGFR.Tyr1173 -1 0 P.Stat1.Tyr701 -1 p190A -1 P.TSC2.Ser939 0 P.AKT.Ser473 TSC2Z-score 1 Chk2 -2 PD.L1 P.HER2.Tyr1221.1222 -2 -1 FGFR4 1 -2 p21.Waf1.Cip1 -1 -2 BAD IGF1.Receptor.beta HIF.1.alpha P.Chk2.Thr68 0.5 -2 P.S6.Ribosomal.Protein.Ser235.236 KI67 -2 -3 -3 -2 -3 0 NF.kappa.B.p65 TGF.beta.I.III H2AX DNA.PK 0 P.DNAPK.Ser2612 TRAIL Cdk1.Cdc2 Histone.H3R2me2s.symmetric P-SMAD2/3 HER4.ErbB4 p27.Kip1 -0.5 Cytokeratin7 P.Cdc6.Ser106 H2AX HER3 HER2 Bcl-xL JunB p21-Waf1-Cip P.c.Jun.Ser63 -Ser465/423-467/425
ERCC1 [RPPA] Beclin.1 -1 Bax 1 2 P.ATM.Ser.Thr.ATR.S.T.QG 2 P=0.002 P=0.02 P=0.003 P.ATM.ser1981 P=0.0007 P<0.0001 P=0.01 Cyclin.B1 1 P=0.01 EGFR P=0.01 P.STAT3.Tyr705_1438Missing values 1 P=0.04 3 P=0.0005 PCNA P=0.04 1.0 C.reactive.protein P.p42.44.MAPK.Thr202.Tyr204 2 P.GSK3.beta.Ser9 P=0.004 P.mTOR.Ser2448 P.p70.S6.kinase.Thr389 P.beta.catenin.Ser675 0 SUV39H1 P=0.02 DDR1 2 Cyclin.D1 0 0 1 PKM2 P.PRAS40.Thr246 0.5 JunB P.SMAD2.Ser465.467_SMAD3.Ser423.425 0 XIAP P.NF.kappa.B.p65.Ser536 P.IRS1.Tyr632 c.Jun -1 1 Histone.H3.Acetyl.K9 P.Histone.H3.Ser10 Cdc25C.5H9 -1 Cdk2 Cdk5 0.0 -1 P.FANCD2.Ser222 0 Cleaved.Notch1.Val1744 -2 hTERT Histone.H3.trimethylated.K9.H3K9me3 P.p27.KIP1.Thr198 0 P.PDK1.Ser241 HER2 -2 P.HER3.Tyr1289 -2 Trf1 Trf2 -0.5 CREB P.4E.BP1.Thr70 -4 Chk1 -2 BAP1 P.p38.MAPK.Thr180.Tyr182 P.Fyn.Tyr528.Src.Tyr530 XRCC1 P.PTEN.ser380.Thr382.383
3 P.Shc.Tyr239.240
- SFRP1 P.Rb.Thr356 B1 Li7 1.1 1.2 P.beta.catenin.Ser552_1245 P-ß-catenin-Ser675 PKM2 P.Cyclin.B1.Ser133
- - EGFR TGFß-I-III HLF
HLE P.calnexin.S583
P.MCM2.Ser108 Proteinexpression Log
Huh6 Huh1 Huh7 Aurora.A JHH6 JHH2 JHH4 JHH7 JHH1 JHH5 4 FANCD2 2 NQO1 P=0.03 HCC Hep3B P<0.0001 HepG2 FAK P=0.002 X53BP1 P=0.02
SNU475 SNU739 Mahlavu SNU182 SNU423 SNU761 SNU449 SNU387 SNU354 SNU368 SNU878 SNU886 SNU398 AMPK.alpha HCC HepaRG HCC P=0.03 Ku80 1 P=0.02 1 HER3 P=0.02 PKA.C.alpha Differentiation Apoptosis-Autophagy PLC/PRF5
MHCC97H ALDH.1A1 P=0.004 AXIN1 BCL2 1 3 Hsp90.beta ATM Bcl.xL Adhesion-Cytoskeleton Cell cycle P.H2AX Survivin Hsp90.alpha Cytokeratin.19 0 Beta.catenin Other TGFß P.p70.S6.kinase.Thr421.Ser424 Protein-protein interactions (34 LCCL) P.MEK1.2.Ser217.221 0 MANUSCRIPT0 C Glutamine.synthetase 2 E.cadherin EpCAM MAPK Wnt Vimentin FGF19 LCK MET P−Shc−Tyr239−240 P.HER4.Tyr1162 P.MET.Tyr1234.1235 -1 Tyrosine kinase Metabolism P.Rb.Ser807.811 -1 RSPO2 AFP p190B 1 -1 Cdk1−Cdc2 P−PTEN−ser380−Thr382−383 P−ATM−Ser−Thr−ATR−LEF1S−T−QG DNA repair B1 Li7 HLF HLE P−PDK1−Ser241 IGF1−Receptor beta
Huh6 Huh1 Huh7 -2 -2
JHH6 CREBJHH2 JHH4 JHH7 JHH1 JHH5 HCC-3 Hep3B P−MSK1−Ser360 HepG2 SNU475 SNU739 Mahlavu SNU182 SNU423 SNU761 SNU449 SNU387 SNU354 SNU368 SNU878 SNU886 SNU398 HCC-1.1 HCC-1.2 HepaRG Bax PLC/PRF5 MHCC97H ERCC1 P−4E−BP1−Thr70 P−Rb−Thr356 P−ATM−ser1981
Trf2 P−p27−KIP1−Thr198 c−Jun P−c−Jun−Ser63 D Cdk5 P−Cdc6−Ser106 Association protein expression-gene alteration (34 LCCL) Beclin−1 53BP1 PD−L1 P−FANCD2−Ser222 Cytokeratin−7 H3K9me3 AXIN1_CTNNB1 P−Fyn−Tyr528_Src−Tyr530 Histone−H3−Acetyl−K9 P<0.01 4 Cyclin−B1 XRCC1 Chk1 P−IRS1−Tyr632 P−Histone−H3−Ser10 HER3 TSC2_TSC2 Cyclin-D1_FGF19 AFP Cyclin-D1_CCND1 P−p38−MAPK−Thr180−Tyr182 Ku80 P-PDK1-Ser241_NFE2L2 FANCD2 P−EGFR−Tyr1173 AXIN1_AXIN1 JunB_FLT4 Cdc25C Total protein Chk2 P−AKT−Ser473 P−S6−Ribosomal−Protein−Ser235−236 3 DDR1_NUP98 NQO1_KEAP1 NF-kappa-B-p65_SETD2 Cdk5_ZNF521 Phospho-protein Cleaved−Notch1−Val1744 Cytokeratin-19_SETD2 P−Cyclin−B1−Ser133 P−p42−44−MAPK−Thr202−Tyr204 IGF1-Receptor-beta_AR PDGFR-alpha_TSC2 Aurora−A -value BAP1 P Cytokeratin-7_TP53 P-MET-Tyr1234-1235_MUC1 Spearman r correlation Hsp90−beta P−GSK3−beta−Ser9 ATM RSK2_PREX2 (FDR<0.05): p27−Kip1 Cdk2 10 P-PTEN-ser380-Thr382-383_NBEA P-HER4-Tyr1162_ALB P−p70−S6−kinase−Thr389 BCL2 NF-kappa-B-p65_JAK3 FGF19_MTAP P-MET-Tyr1234-1235_FLT4 0.54 (min) P−TSC2−Ser939 2 HER2 P-PTEN-ser380-Thr382-383_TSC1 NQO1_TERT 0.83 (max) HER4 P−H2AX - Log PD-L1_ZNF521 P−mTOR−Ser2448 AXIN1 -0.55 (min) Survivin P−STAT3−Tyr705_1438 P-SMAD2-Ser465-467_SMAD3-Ser423-425_PREX2 -0.72 (max) C−reactive protein Adhesion-Cytoskeleton Tyrosine kinase MET P−NF−kappa−B−p65−Ser536 P−DNAPK−Ser2612 PKA−C−alpha 1 Differentiation PI3K/AKT/mTOR P−p70−S6−kinase−Thr421−Ser424 Bcl−xL Hsp90−alpha Angiogenesis MAPK Beta−catenin TGF−beta−I−III Apoptosis-Autophagy JAK/STAT PCNA XIAP NQO1 Vimentin P>0.01 Cell cycle Wnt LCK P−MET−Tyr1234−1235 Chromatin NFkB E−cadherin FGFR4 EpCAM 0 DNA repair TGFß ACCEPTEDDDR1 -8 -6 -4 -3 -2 -1 0 1 2 3 4 6 Telomere Oxidative stress hTERT P−beta−catenin−Ser675 Other P−MEK1−2−Ser217−221 Log2 ratio of protein expression (Mutated vs non-mutated) Figure 4
A B Mitosis ProteinACCEPTED degradation MANUSCRIPTZ-score [AUC] 1 1 11 5 Tyrosine kinases DNA methylation resistant LCCL transcriptomic group 1 1 PI3K-AKT-mTOR signaling DNA replication 4 CL1 Unclassified 1 31 MAPK/ERK signaling Wnt/ß-catenin signaling 1 2 CL2 drugs 6 Chaperone protein folding Apoptosis 0 CL3 2 Epigenetic erasers Insulin growth factor signaling 2 -2 3 MET signaling P53 signaling 2 3 -4 P = 0.009 FGFR4/FGF19 signaling Drug combination sensitive MKMK-2206-2206
1.2 Sorafenib_MKSorafenib+MK-2206-2206 1.2 LinsitinibLinsitinib
] Sorafenib_ResminostatSorafenib+Resminostat
1.0 ResminostatResminostat 1.0 SorafenibSorafenib EntinostatEntinostat
0.8 PHAPHA-665752-665752 0.8 JNJJNJ-38877605-38877605 CabozantinibCabozantinib
0.6 0.6 DecitabineDecitabine RegorafenibRegorafenib ICGICG-001-001
0.4 Tanespimycin Sensitivity[AUC 0.4Drug sensitivity (AUC) Tanespimycin AlvespimycinAlvespimycin CD532CD532
0.2 0.2 TrametinibTrametinib Drug SelumetinibSelumetinib Sorafenib_RefametinibSorafenib+Refametinib
0.0 0.0 Sorafenib_TrametinibSorafenib+Trametinib RefametinibRefametinib DasatinibDasatinib BortezomibBortezomib
CD532 Nutlin3 AlisertibAlisertib Alisertib Brivanib ICG-001 Linsitinib TivantinibTivantinib MK-2206 Dasatinib Tivantinib Paclitaxel Entinostat BLU-9931 Sorafenib Navitoclax Trametinib Decitabine VinblastineVinblastine Vinblastine Lenvatinib Rapamycin Bortezomib Selumetinib Doxorubicin Refametinib
Resminostat PaclitaxelPaclitaxel PHA-665752 Regorafenib Alvespimycin Tanespimycin PF-04691502
Cabozantinib BrivanibBrivanib JNJ-38877605 BLUBLU-9931-9931 DoxorubicinDoxorubicin LenvatinibLenvatinib Sorafenib_MK-2206 NutlinNutlin-3-3 Sorafenib_Trametinib
Sorafenib_Refametinib PFPF-04691502-04691502 Sorafenib_Resminostat RapamycinRapamycin NavitoclaxNavitoclax P<0.0001 B1 C Li7 HLF HLE Huh6 Huh1 Huh7 JHH5 JHH4 JHH7 JHH6 P=0.0001 P=0.66 JHH1 JHH2 HCC-3 Hep3B 1.2 HepG2 1.2 ● ● resistant SNU761 SNU878 SNU398 SNU182 SNU739 SNU423 Mahlavu SNU387 SNU886 ● ●●● SNU368 SNU449 SNU354 SNU475 HCC-1.1 HCC-1.2 ●●●● HepaRG
●● ●●●●●●● ●●●●● ●●●●● ●●●●●●●●●●●●● ●●●●●●●●●●●●●●●● ●●●●●●●●●●●●● PLC/PRF5
●●●●●●●●●●●●●●●●●●●●●●●●●●● MHCC97H ●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●●●●● P<0.0001 P<0.0001 ●●●●●●●●●●●●●●●●●●●●●● ● ● ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● 0.9 ●●●●●●●●●●●●●●●●●●● 0.9 ●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●● 1.0 ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ● ●●●●●●●●●●●●●●●●●●●●●●● 0 ●●●●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●● ●●●●●●●●●●●●●● 10 ●●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●● ●●●●●●●●●●●●
●●●●●●●●● ●●●●●●●●●●●●●●●●●●●● ●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●●●●●●● -2 ●●●●●●●●●●●●●●●●●● ●●●● ●●●●●●●●●●
+ 4 combo)4 + ●●●●●●●●●● ●●●●●●●●●●● 10 ●●●●●●●●●●●●● 0.5 ●●●●●●●●●● ●●●●●●●● ●●●●●●
AUC ●●●●●●●●●●●● ●●●●● 0.6 0.6 ●●●●●●● ●●●●●●●●●●● ●●●●●●●●●●● -4 ●●●●●●
●● 10 ●●●●●●● ●●●
●● ●●●● -value] ●●●● ●●●●● ●●●● Drug AUC Drug
●●● ●●●● P -6 ●●●●●
0.0 [ ●●●●●●● ●●●●●●● 10 ●● ●●●●● ●●●● ●●● ●●●●● [Spearman r] ●●●●● -8 ●●●●● ●●
0.3 ● 10 ●● 0.3 ●● ● ●● (31drugs Drug AUC correlation
Drug AUC correlation -0.5
● ● sensitive Same Different Same Different CL1CL1 CL2CL2 CL3CL3 target target target target (n=11)(N=11) (n=9)(N=9) (n=13)(N=13) (n=11) (n=454) (n=11) (n=454)
D Trametinib Refametinib Cabozantinib Tanespimycin Alvespimycin Linsitinib Sorafenib_MK-2206 Antagonism 1.4 20 1.4 P=0.008 P=0.02 P=0.06 P=0.01 P=0.01 P=0.01 Median=1.1 Additive 1.2 1.2 1.2 15 P=0.8 P=0.002 P=0.1 P=0.3 P=0.003 P=0.5 1.2 P=0.008 P=1 10 1.2 P=0.34 1.1 P=0.02 P=0.09 P=0.9 1.2 Synergy 1.0 5
1.0 1.0 1.0 1.0 % Frequency 0 MANUSCRIPT1.0 0.5 1.0 1.5 0.8 0.8 0.9 0.8 32% Bin Center 0.8 AUC AUC AUC AUC AUC AUC 0.8 0.6 0.8 Sorafenib_Resminostat 0.6 0.6 25 0.6 0.7 20 Median=1.2 Antagonism 0.4 0.4 0.4 15 0.6 0.6 10 Additive 5 0.2 0.4 0.5 0.2 0.2 Synergy
% Frequency 0 0.0 0.5 1.0 1.5 2.0 2.5 Sorafenib_MK-2206 Nutlin-3 Drugsensitivity [AUC] Sorafenib_Trametinib Sorafenib_Refametinib JNJ-38877605 29% Bin Center 1.1 1.2 1.1 P=0.01 1.1 P=0.006 P=0.01 P=0.01 P=0.05 P=0.04 1.2 Sorafenib_Refametinib P=0.56 P=0.02 P=0.003 P=0.4 P=0.04 P=0.6 P=0.04 P=0.3 20 1.0 1.0 1.1 Median=0.8 0.9 P=0.37 Frequency% 15 0.9 1.0 10 Antagonism 0.9 5 0.7 1.0 Additive 0.8 % Frequency 0 AUC AUC AUC AUC AUC 0.0 0.5 1.0 1.5 2.0 2.5 0.8 0.8 Synergy 0.7 70%Bin Center 0.5 0.9 0.7 0.6 0.6 Sorafenib_Trametinib 0.3 20 0.5 0.6 0.8 15 Median=0.7 10 LCCL transcriptomic group CL1 (n=11) CL2 (n=9) CL3 (n=13) 5 Antagonism % Frequency 0 0.0 0.5 1.0 1.5 2.0 2.5 Additive Association drug sensitivity-protein expression (34 LCCL) 58%Bin Center E F Synergy no difference RESISTENT SENSITIVEno difference RESISTENTRESISTANT SENSITIVE Rapamycin Bin center [Combination Index] 1.0 P=0.02 CI<1 Synergism
Dasatinib-Cytokeratin.19 Dasatinib-Cytokeratin.19 1
4 4 0.4 Dasatinib-P.Rb.Thr356 Dasatinib-P.Rb.Thr356 M NM Alvespimycin-NQO1 Alvespimycin-NQO1 Sorafenib-PKM2 Sorafenib-PKM2 CD532-H2AX CD532-H2AXLinsitinib-EGFR Linsitinib-EGFR TSC1/TSC2 value Trametinib-Cytokeratin.19- Trametinib-Cytokeratin.19JNJ-38877605-P.beta.catenin.Ser675 Navitoclax-P.PDK1.Ser241JNJ-38877605-P.beta.catenin.Ser675 Navitoclax-P.PDK1.Ser241 3 Tanespimycin-NQO1P 3 Tanespimycin-NQO1Tanespimycin-TSC2 Tivantinib-P.PDK1.Ser241Tanespimycin-TSC2 Tivantinib-P.PDK1.Ser241 Refametinib-HER2 Bortezomib-p190BRefametinib-HER2 Bortezomib-p190B Decitabine-P.Histone.H3.Ser10 Decitabine-P.Histone.H3.Ser10 Vinblastine-BAD Vinblastine-BAD Alisertib Sorafenib-Survivin Alisertib-P.beta.catenin.Ser552_1245Sorafenib-Survivin Alisertib-P.beta.catenin.Ser552_1245 Tanespimycin-ALDH1A1 Tanespimycin-ALDH1A1 PHA-665752-P.beta.catenin.Ser675 PHA-665752-P.beta.catenin.Ser675 P=0.02 Navitoclax-Cytokeratin.19 Paclitaxel-P.PRAS40.Thr246Navitoclax-Cytokeratin.19 Paclitaxel-P.PRAS40.Thr246 Drugsensitivity [AUC] 1.0 Dasatinib-Cdk1.Cdc2 Log10 Dasatinib-Cdk1.Cdc2 Refametinib-Chk1 - Refametinib-Chk1 Decitabine-P.p38.MAPK.Thr180.Tyr182 Decitabine-P.p38.MAPK.Thr180.Tyr182 0.9 2 2 P<0.05 0.8 P>0.05 0.7 -log10 adjusted p-value adjusted -log10 -log10 adjusted p-value adjusted -log10 0.6 1 1 M NM TP53
0 0 -1.0 -0.5 -1.0 0.0 -0.5 0.5 0.0 1.0 0.5 1.0 Correlation Spearman’scoefficient rankCorrelationcorrelation coefficientcoefficient Figure 5
A Trametinib sensitivity (34 LCCL)ACCEPTED MANUSCRIPTB JNJ-3887760 FGF19 Sensitive: Resistant: PHA-665752 NRG GI50 [0.008-10 µM] Density 2 gi 0 5 10 Trametinib0 5 GI5010 1 Trametinib GI50
0 ERBB4 MET KLB 0 P P P -1 P SHC FGFR4
score of AUCscore of GRB2 Density -
Z -2 name SOS 0.0 0.5 1.0 BLU-9931 -2 0.0 Trametinib0.5 AUC1.0 -3 Trametinib AUC H3B-6527 transcriptomic transcriptomic Unsupervised Unclassified B1 Li7 NF1 RAS HLF HLE Class CL2 CL1 CL2 CL1 CL1 CL1 CL2 CL2 CL3 CL2 CL2 CL3 CL1 CL1 CL1 CL3 CL1 CL2 CL1 CL3 CL1 CL3 CL2 CL2 CL3 CL3 CL3 CL3 CL3 CL3 CL1 CL3 CL3 Huh7 Huh6 Huh1 JHH1 JHH5 JHH7 JHH2 JHH4 JHH6 HCC-3 Hep3B HepG2 SNU878 SNU761 SNU886 SNU398 SNU423 SNU354 SNU387 SNU739 SNU368 Mahlavu SNU182 SNU475 SNU449 HCC-1.2 HCC-1.1 HepaRG B1 Li7 Sorafenib RAF PLC/PRF5 MHCC97H Genomic alteration: HLF MAPK MAPK HLE
alteration Huh7 Huh6 Huh1 JHH1 JHH5 JHH7 JHH2 JHH4 JHH6 RAS-MAPK pathway HCC-3 Hep3B Trametinib FGF19 HepG2 FGF19 MEK
MET SNU878 SNU761 SNU886 SNU398 SNU423 SNU354 SNU387 SNU739 SNU368 Mahlavu SNU182 SNU475 SNU449 Refametinib HCC-1.2 HCC-1.1 HepaRG MET NRAS Selumetinib PLC/PRF5
MHCC97H NRAS ERBB4 RPS6KA3 ERBB4 ERK RPS6KA3
NF1 RPS6KA3 NF1
MET ERBB4 RSK2 4 C Amplification FGFR4 KLB NR1H4 500 3 2 FGF19 MHCC97H JHH1 2 Homozygous deletion P<0.0001 P<0.0001 P<0.0001 2 14 14 0 Fusion 14 0.63 0.56 0.4 100 12 12 12 1 Missense 10 10 50 -2 10 KLB 0.81 FGFR4 Hep3B 8 8 0 0 -4 8 Protein level Log 6 6 mRNAlevel [fpkm] mRNAlevel [fpkm] KLB mRNA
-50 HCC-3 mRNAlevel 0.64 0.65 6 NR1H4 mRNA -1 -6 FGFR4 mRNA 4 4 4 2 2 Amplification no yes P<0.05 LCCL transcriptomic subgroup: NR1H4 P<0.001 FGF19 expression CL1 CL2 CL3 P<0.0001 15.0 FGF19 amplified z-score FGF19 non-amplified 12.5 2 FGF19 mRNA 1 FGFR4 mRNA P<0.0001 (Ampl.) 0 KLB mRNA 10.0 -1 NR1H4 mRNA P<0.001 (Groups) -2 B1 Li7 HLF 7.5 HLE Huh7 Huh1 Huh6 JHH7 JHH5 JHH1 JHH4 JHH2 JHH6 mRNAlevel HCC-3 Hep3B HepG2 SNU878 SNU761 SNU886 SNU368 SNU354 SNU449 SNU475 SNU739 SNU398 SNU182 Mahlavu SNU387 SNU423 HCC-1.2 HCC-1.1 HepaRG
5.0 PLC/PRF5 MHCC97H
2.5 CL1 CL2 CL3 D FGF19 amplified LCCL (n=11)
FGF19 mRNA [variance satbilized values] satbilized FGF19 [variance mRNA LCCL transcriptomic subgroup Trametinib (anti-MEK) BLU-9931 (anti-FGFR4) H3B-6527 (anti-FGFR4) 12 r= -0.66 ; P=0.04 1.0 1.0 10 CL1 Cell differentiation MANUSCRIPT8 CL2 0.8 0.8 6 CL3 AUC GI50 0.6 0.6 4 2 H3B-6527 GI50 BLU-9931 AUC Trametinib AUC Trametinib r= -0.62 ; P=0.04 r= -0.58 ; P=0.06 0.4 0.4 0 2 4 6 8 10121416 2 4 6 8 10121416 2 4 6 8 10121416 mRNA FGF19 mRNA FGF19FGF19 mRNA FGF19 mRNA Sensitive miRNA E Resistant Mutation CTPS1 (0.99) Z-score RORC KCMF1TUFT1 CTPS1 C1orf122 (0.98) 1 RORC MRPS14 KCMF1 HSD17B7 LAD1CTPS1TUFT1 KCMF1 (0.96) 0.5 1.001.00 MRPS14HSD17B7 LAD1 C1orf122 C1orf122 TUFT1 (0.96) 0 SERINC2SERINC2 ADPRHL2 ADPRHL2 (0.95) ADPRHL2 -0.5 HSD17B7 (0.98) -1 0.750.75 RORC (0.98) MRPS14 (0.96) SERINC2 (0.94) LAD1 (0.92) B1 Li7 EN Score HLF 0.500.50 HLE Huh7 Huh6 Huh1 JHH1 JHH5 JHH7 JHH2 JHH4 JHH6 HCC-3 Hep3B HepG2 SNU878 SNU761 SNU886 SNU398 SNU423 SNU354 SNU387 SNU739 SNU368 Mahlavu SNU182 SNU475 SNU449 HCC-1.2 HCC-1.1 HepaRG PLC/PRF5 MHCC97H 12 1.2 0.250.25 1.0 10 Bootstrapfrequency (Score) 0.8 8 0.6 Trametinib Trametinib 6 0.000.00 0.4 r= -0.64; P<0.0001 Trametinib AUC Trametinib Frequency Bootstrap (Score) Bootstrap Frequency Frequency Bootstrap (Score) Bootstrap Frequency
0.2 5-gene score mRNA -0.02-0.010 -0.005-0.01 0.000 0.000.005 0.0100.01 Liver cancer cell lines 0.3 0.6 0.9 1.2 Trametinib AUC AverageAverageAverageACCEPTED Coefficients Coefficients LCCL transcriptomic subgroup: CL1 CL2 CL3 Unclassified Figure 6
HCC TCGA series (n=319) A Boyault’s transcriptomic classificationACCEPTED MANUSCRIPTHoshida’s transcriptomic classification n= 57 34 27 98 25 78 102 74 143 G1 G2 G3 G4 G5 G6 S1 S2 S3 13 13
gene 12 12 - 5 11 11 10 10
scoremRNA 9 9 Trametinib P=0.018 P<0.0001 8 8 Trametinib signature mRNA signature Trametinib Trametinib signature mRNA StabVar z-score mRNA signature Trametinib Mean 11.4 11.1 11 11.1 11.4 11.2 11 11.3 11.3 13 13 12 12 11 11 mRNA 10 10
FGF19/FGFR4 9 9 8
pathway 8 P=0.0002 P<0.0001
Mean 7 7 Mean Mean FGF19/KLB/FGFR4/NR1H4 FGF19/KLB/FGFR4/NR1H Mean 11 10.9 10.4 10.6 10.6 10.7 10.5 11 10.7 9 9
FGF19/FGFR4/KLB/NR1H4 pathway mRNA P=0.08 8 8
mRNA 7 7 mRNA FGF19 6 6 P=1 5 FGF19 FGF19 mRNA 15 14 14 13 13 12 mRNA FGFR4 P<0.0001 12 P<0.0001 11 11 FGFR4 mRNA FGFR4 mRNA 13 13 12 12 11 11 KLB
mRNA 10 10 P=0.0005 P<0.0001 9 9 KLB mRNA KLB mRNA 12 12 11 11 mRNA NR1H4 NR1H4 10 P<0.0001 10 P<0.0001 NR1H4 mRNA NR1H4 mRNA q13.3 B MANUSCRIPT Chromosome 11 69 460 kb 69 480 kb 69 500 kb 69 520 kb
CCND1 ORAV1 FGF19 Focal amplifications n=21 in HCC (TCGA series) n=3 CCND1 expression FGF19 expression FGF19 amplified 18 CCND1 amplified 18 CCND1 non-amplified FGF19 non-amplified 16 16 14 14 12 values] 12 10 10 8 6
stablilized 8 G1 G2 G3 G4 G5 G6 G1 G2 G3 G4 G5 G6 (n=55/4) (n=31/3) (n=23/4) (n=93/7) (n=23/2) (n=74/4) (n=55/3) (n=31/3) (n=23/4) (n=93/5) (n=23/2) (n=74/4) 18 18 [variance 16 16 14 level 14 12 12 ACCEPTED10 mRNA 10 8 6 8 S1 S2 S3 S1 S2 S3 (n=95/7) (n=66/8) (n=134/9) (n=95/7) (n=67/7) (n=136/7) Transcriptomic subgroups (TCGA HCC series n=319) Figure 7
HCC primaryACCEPTEDtumors classes MANUSCRIPT previously established Molecular Proliferation class ~50% Non-proliferation class ~50% subclasses Lee, 2004 Cluster A Cluster B Boyault, 2007 G1 G2 G3 G5 G6 G4 Chiang, 2008 Proliferation CTNNB1 Poly 7 Interferon Hoshida, 2009 S2 “progenitor” S1 “TGFß-Wnt” S3 TCGA, 2017 iCluster 1 iCluster 3 iCluster 2
TP53 mut ; 11q13 amplification (FGF19 and CCND1) CTNNB1 mut Chr 7 AXIN1 mut TERT promoter mut ampl RPS6KA3 mut TSC1/TSC2 mut PI3K-AKT-mTOR, RAS-MAPK, MET, E2F1 signaling Main molecular features KRT19+ (mRNA) Phospho-ERK Wnt-TGFß signaling Progenitor features: IGF2+, AFP+, EPCAM+ CK19+ (IHC) HBV HCV-alcohol Clinical features Poor differentiation-vascular invasion Moderate-well differentiation Worse outcome Better outcome
Liver cancer cell lines subgroups identified in the present study Transcriptomic CL3 subgroups CL1 CL2 No representative models HepaRG, Li7, HCC-1.1, SNU182, JHH2, SNU354, HepG2, Hep3B, B1, HCC- SNU387, SNU398, SNU368, 1.2, HCC-3, Huh7, SNU423, SNU449, SNU761, Cell lines PLC/PRF5, Huh6, Huh1, SNU475, Mahlavu, SNU878, JHH5, JHH7 HLE, HLF, JHH4, SNU886, JHH6, SNU739 MHCC97H Hepatoblast Mixed epithelial MesenchymalMANUSCRIPT- -like -mesenchymal like Differentiation
TERT promoter, TP53, AXIN1, TSC1/TSC2 mut 11q13 amplification (FGF19 and CCND1) Main molecular features Progenitor features: Wnt-TGFß signaling IGF2+, AFP+, EPCAM+ Phospho-ß-catenin S675 CK19+ Linsitinib Sorafenib + MK2206 Sorafenib + Trametinib or Refametinib * ** Trametinib/FGF19 ampl Drug/ BLU-9931, H3B-6527/FGF19 ampl * Biomarker pairs Dasatinib/CK19 * ACCEPTEDAlisertib/TP53 mut * Rapamycin/TSC1-TSC2 mut * JNJ-38877605, PHA-665752, Cabozantinib/MET ampl ** Alvespimycin,Tanespimicyn/NQO1 * * With functional validation in experimental models ** With « proof of concept » validation in HCC clinical trial