Ex VIVO method for predicting the response to sorafenib treatment of hepatocellular carcinoma in a subject

Abstract

The present invention relates to an ex vivo method for predicting the response to sorafenib treatment of hepatocellular carcinoma in a subject, comprising the steps of: 1) determining expression, in a liver tissue sample from the subject, of proteins associated with protein translation pathways, said sample containing pathological tissue and optionally non-pathological tissue, 2) detecting an increase of expression of proteins associated with protein translation pathways, in pathological tissue compared to non- pathological tissue from the liver tissue sample of the subject or from a standard non-pathological liver sample, wherein: If expression of proteins associated with protein translation pathways is increased in pathological tissue, the subject is unlikely to have a response to sorafenib treatment, If expression of proteins associated with protein translation pathways is not increased in pathological tissue, the subject is likely to have a response to sorafenib treatment

Description

[0001] EX VIVO METHOD FOR PREDICTING THE RESPONSE TO SORAFENIB TREATMENT OF HEPATOCELLULAR CARCINOMA IN A SUBJECT

[0002] Technical field

[0003] The present invention refers to an ex vivo method for predicting the response to sorafenib treatment of hepatocellular carcinoma in a subject.

[0004] Therefore, the present invention has utility in medical field, and more particularly in treatment response prediction field.

[0005] In the description below, the references into brackets ([ ]) refer to the listing of references situated at the end of the text.

[0006] Background of the Invention

[0007] Hepatocellular carcinoma (HCC) is a major public health issue and, as the most common primary liver tumor, its incidence reaches one million new cases per year worldwide. Alcohol consumption and the increase in obesity cases are believed to cause an increase in the incidence of HCC of 9% per year and new cases of HCC are now mostly caused by Metabolic- dysfunction Associated Fatty Liver Disease (MAFLD). HCC develops on an underlying cirrhotic liver in 80% cases, more rarely on a chronic liver disease without cirrhosis.

[0008] Although screening programs diagnose approximately 40% of HCCs at a curative stage, at least 50% of patients will be diagnosed at an intermediate or advanced stage. The prognosis remains unfavorable at these later stages due to extensive tumor burden, a high frequency of liver dysfunction, and deterioration of health status, which limit access to curative treatments. Thus, HCC is the second leading cause of cancer death worldwide.

[0009] In patients with advanced HCC (Barcelona Clinic Liver Cancer (BCLC) C) or with intermediate-stage (BCLC B) disease not eligible for, or progressing despite, locoregional therapies, systemic therapies are the gold standard of care.

[0010] Sorafenib (a multi-tyrosine kinase inhibitor, also called “TKI”) has been the standard treatment of care since 2007, based on slight improved overall survival (OS) in randomized controlled trials compared to supportive care. However, a few patients have an objective response. The management of advanced HCC has been recently disrupted with the development of new effective systemic treatments including immunotherapies, that improve both OS and progression free survival (PFS). Currently, the combination of atezolizumab (anti-programmed death-ligand 1 (PDL1)) and bevacizumab (anti-vascular endothelial growth factor (VEGF)) is now the first-line standard of care for advanced HCCs following the IMbrave150 study, increasing median OS to more than 19 months in patients with preserve liver function. Moreover, a combination of an anti-PDL1 (durvalumab) and an anti-cytotoxic T-lymphocyte antigen-4 (CTLA4, tremelimumab) has been approved last year after positive results of the Himalaya phase III trial, strengthening the therapeutic arsenal in first line setting.

[0011] In the era of sorafenib-exclusive therapy, many studies attempted to develop predictive tools for treatment response. Some results are encouraging: a greater delay in progression has been associated with higher pre-treatment phospho-ERK levels, high c-Met expression has predicted therapeutic efficacy, and a higher expression of phospho-c-Jun among non-responders has been found. However, all these biomarkers have never been validated in clinical practice.

[0012] Therefore, there is no precision medicine tool that can be used to validate the choice of sorafenib treatment in HCC. However, it is important to understand that in case of non-response to a first line of systemic treatment, it is estimated that less than 25% of patients can effectively receive a second line of treatment. In addition, the underlying hepatic reserve and the progression of oncological disease strain the prognosis of patients and limit access to later lines, but there is guideline for the optimal therapeutic sequence after first line treatment failure. Thus, a need exists of new tools for predicting the response to sorafenib treatment of hepatocellular carcinoma.

[0013] Description of the invention

[0014] After extensive research work, the Applicant has surprisingly shown that the level of expression of proteins associated with protein translation pathways is a significantly discriminant and robust tool for predicting response or lack of response of a patient to sorafenib treatment.

[0015] Therefore, the invention is based on using proteins associated with protein translation pathways level expression as a new tool to predict the response to sorafenib treatment in HCC.

[0016] One of the major advantages of the invention is that it makes it possible to direct patients towards another potentially more suitable treatment, such as, for example, other multikinase inhibitors, such as lenvatinib, or immunotherapy / anti-VEGF, if the subject is likely to have a response to sorafenib treatment.

[0017] Another advantage of the invention is that it can be performed using diagnostic biopsies made before treatment, especially by proteomic mass spectrometry analysis of the diagnostic biopsy. Therefore, only small amounts of tissue are needed for this analysis, thus not depleting the samples and it can be reused.

[0018] In addition, the Applicant developed a response prediction approach, notably proteomic profiling-based, for implementing the predicting process of the invention. Another advantage is that proteomic profiling combined with a predictive signature for response to treatment enable the most appropriate treatment to be prescribed directly to each patient without any loss of chance, and thus increase life expectancy in HCC.

[0019] Accordingly, in a first aspect, the present invention provides an ex vivo method for predicting the response to sorafenib treatment of hepatocellular carcinoma in a subject, comprising the steps of: 1) determining expression, in a liver tissue sample from the subject, of proteins associated with protein translation pathways, said sample containing pathological tissue and optionally non-pathological tissue,

[0020] 2) detecting an increase of expression of proteins associated with protein translation pathways, in pathological tissue compared to non- pathological tissue from the liver tissue sample of the subject or from a standard non-pathological liver sample, wherein:

[0021] If expression of proteins associated with protein translation pathways is increased in pathological tissue, the subject is unlikely to have a response to sorafenib treatment,

[0022] If expression of proteins associated with protein translation pathways is not increased in pathological tissue, the subject is likely to have a response to sorafenib treatment.

[0023] “Hepatocellular carcinoma” or HCC refers herein to any HCC, regardless of its stage of development. As an example, it may be an early- stage HCC, a mid-stage HCC or an advanced HCC, and a resectable or a non resectable HCC. The HCC may already have been treated with a first line therapy different than sorafenib, or not.

[0024] As used herein, a “subject” is a patient, animal, mammal, or human, suffering from HCC.

[0025] Sorafenib, also called “sorafenib tosylate” and sold under the brand name Nexavar®, is the molecule of formula:

[0026] Sorafenib has the CAS Number 284461-73-0. “Sorafenib treatment" refers herein to any treatment protocol involving sorafenib commonly used by the medical profession to treat cancers. The liver tissue sample may be a physical tissue sample. Therefore, the liver tissue sample may be any sample obtainable by known techniques, as example by at least one technique selected among macrodissection, laser microdissection of a formalin-fixed paraffin- embedded or frozen tissue, and biopsies, notably fresh biopsy. Any surface area of tissue sample may be used, as long as it contains pathological tissue and optionally non-pathological tissue.

[0027] “Pathological tissue” refers herein to liver tissue having HCC cancerous cells.

[0028] “Non-pathological tissue” refers herein to liver tissue that does not have HCC cancerous cells.

[0029] Advantageously, the liver tissue sample from the subject contains pathological and non-pathological tissue.

[0030] “Proteins associated with protein translation pathways" refers herein to all or part of the biological pathways involved in the process of protein translation, whether it be translation signaling, translation initiation and / or translation elongation, in cells of the liver tissue sample of a subject. Advantageously, protein translation pathways comprise translation signaling and eukaryotic initiation factor (elF) signaling. For example, translation signaling may comprise at least one protein chosen among PABPC1 (poly(A) binding protein cytoplasmic 1), RPL14 (ribosomal protein L14), RPL7A (ribosomal protein L7a), RPS2 (ribosomal protein S2) and RPS3 (ribosomal protein S3). For example, eiF signaling may comprise at least one protein chosen among eiF2, such as eiF2S3 (eukaryotic translation initiation factor 2 subunit gamma), eiF3, such as eiF3E (eukaryotic translation initiation factor 3 subunit E), eiF3F (eukaryotic translation initiation factor 3 subunit F) and eiF3l (eukaryotic translation initiation factor 3 subunit I), eiF4, such as eiF4A1 (eukaryotic translation initiation factor 4A1) and elFS3 (eukaryotic translation initiation factor 3).

[0031] Advantageously, expression of protein Cdc42 (cell division cycle 42) may also be tested. Therefore, steps 1) and 2) of the ex vivo method may also be performed for Cdc42, wherein: if expression of Cdc42 is increased in pathological tissue, the subject is likely to have a response to sorafenib treatment,

[0032] If expression of Cdc42 is not increased in pathological tissue, the subject is unlikely to have a response to sorafenib treatment.

[0033] Testing expression of proteins associated with Cdc42, in addition to proteins associated with protein translation pathways, may allow to increase the number of parameters in the prediction method and increase its computing power of the method.

[0034] In step 1) of the method, expression of proteins may be determined by any method known by the skilled person allowing to measure protein expression level in a tissue, or by determining relative protein abundance. It may be for example at least one method selected among proteomic profiling matching, immunohistochemistry, western blot, global or targeted mass spectrometry analysis, and reverse transcriptase PCR (polymerase chain reaction).

[0035] Advantageously, in case of proteomic profiling matching, it may comprise determining abundance of proteins of the liver tissue sample from the subject, for example by carrying out label free proteomic analysis, or by a targeted quantification technique such as a method of liquid chromatography coupled with tandem mass spectrometry (LC-MS / MS), in discovery mode or with a targeted approach (eg. Multiple Reaction Monitoring (MRM), Parallel Reaction Monitoring (PRM)), in Data Dependant Acquisition (DDA) or Data-independent Acquisition (DIA).

[0036] In case of determining relative protein abundance, it may be the relative abundance of a protein of the above-mentioned pathways, i.e. protein translation pathways, and optionally Cdc42, of the non-pathological tissue over the corresponding protein in the pathological tissue. Alternatively, the comparison may be made between the pathological tissue of the subject and a standard non-pathological tissue of liver sample, used as a reference, for example an extract of normal liver tissue from one patient or a mixture of several patients. Alternatively, the comparison may be made with a reference, pre-determined, proteins of the above- mentioned pathways, i.e. protein translation pathways, and optionally Cdc42, which can be obtain from a non-pathological tissue liver sample from a third party, or from several non-pathological tissue liver samples in order to be statistically representative.

[0037] “Likely to have a response” to sorafenib treatment refers herein to a statistically significant level of tumor regression and / or decrease in alphafetoprotein (AFP) levels. It may be for example a tumor regression according to the RECIST criteria, which considers that tumor regression to be significant at at least 30% tumor mass and / or at at least 50% decrease in AFP. Alternatively, it may be for example a tumor regression and / or decrease in AFP levels, in the first two months after treatment initiation, of at least 5%, or at least 10%, or at least 20%, or at least 30%, or at least 40%, or at least 50%, or at least 60%, or at least 70%, or at least 80%, or at least 90%, or even 100%, or any number or range in between, compared with the value without sorafenib treatment.

[0038] “Unlikely to have a response” to sorafenib treatment refers herein to the other cases. It may be a tumor progression according to the RECIST criteria, which consider that a tumor progression is significant at 20% tumor progression. Alternatively, it may be a tumor progression to a similar extent to a case without treatment, for example a tumor progression of at least 5%, or at least 10%, or at least 20%, or at least 30%, or at least 40%, or more, within two months.

[0039] In another aspect, the present invention provides a method for choosing a treatment for a patient with HCC for an sorafenib treatment, comprising the steps of carrying out the ex vivo prediction method as defined above, and selecting the patient for whom expression of proteins associated with protein translation pathways is increased in pathological tissue.

[0040] In another aspect, the present invention provides a method of treating a patient with HCC, comprising the steps of: (i) selecting the patient for treatment for whom expression of proteins associated with protein translation pathways is increased in pathological tissue, and

[0041] (ii) administering to the patient an effective amount of sorafenib treatment.

[0042] This invention is further illustrated by the following examples with regard to the annexed drawings that should not be construed as limiting.

[0043] Brief description of the figures

[0044] - Figure 1 : represents principal Component Analysis (PCA) of the proteomic profiles of biopsies before treatment discriminating between responders and progressors under sorafenib, showing deregulated protein expressions between tumoral vs non-tumoral (T / NT) tissues in patients with an objective response (n=9) or with progression (n=15) under sorafenib. The dataset was reduced to a signature of 87 proteins with significantly different expression between patients with progressors (points) and responders (triangles).

[0045] - Figure 2: represents the details of T / NT ratios of proteins involved in protein translation, obtained by proteomic profile of biopsies before treatment discriminating between responders and progressors under sorafenib.

[0046] Examples

[0047] Example 1: Predicting response to sorafenib treatment in patients with advanced hepatocellular carcinoma

[0048] Materials and Methods Patients’ selection

[0049] We consecutively and prospectively included all patients treated with sorafenib, at the University Hospital of Bordeaux and the University Hospital of Angers for advanced -unresectable and / or metastatic, not eligible for or progressing despite loco-regional therapies- HCC.

[0050] For all patients, we retrospectively collected baseline characteristics before the beginning of the systemic treatment; including age, sex, BMI, performance status score (OMS), the type of underlying liver disease and cirrhosis, the CHILD-Pugh and ALBI (ALbumin-Blirubin) score.

[0051] Cirrhosis was assumed on clinical grounds (signs of hepatocellular insufficiency and / or portal hypertension), biological (stigmata of hepatocellular insufficiency and / or portal hypertension), radiological (hepatic dysmorphia, signs of portal hypertension), and confirmed from liver biopsies. The diagnosis of metabolic dysfunction-associated fatty liver disease (MAFLD) was made according to the recent international consensus published in 2021 (Eslam M et aL: "MAFLD: A Consensus- Driven Proposed Nomenclature for Metabolic Associated Fatty Liver Disease". Gastroenterology. 2020;158:1999-2014.e1 ([1])): in case of fatty liver highlighted by a hyperechoic aspect in ultrasound associated with obesity or overweight, type II diabetes, and in their absence, the combination of two metabolic risk factors: hypertriglyceridemia, low HDL (high density lipoprotein) - cholesterol (adjusted for sex), an increase in waist circumference (adjusted for ethnic origin) and high blood pressure.

[0052] The diagnosis of alcohol-related liver disease (ALD) was made in the case of consumption of more than twenty-one standard doses of alcohol per week in men and fourteen standard doses of alcohol per week in women without liver steatosis.

[0053] We collected the following biological parameters : total bilirubin level, creatininemia, albuminemia, transaminases, alkaline phosphatase (PAL), gamma-glutamyl transferase (GGT), prothrombin rate (PT), alpha foetoprotein level (AFP), C-reactive protein level (CRP), platelets. Tumor characteristics were its stage at the beginning of the systemic treatment, the BCLC (Barcelona Clinic Liver Cancer (BCLC) C) classification assessed by cross-sectional imaging, the degree of differentiation of the tumor, the existence of micro and / or macrovascular invasion (in case of initial resection), the architecture of the tumor following WHO (World Health Organization) classification, the expression of Glutamine Synthetase (GS), Cytokeratin 19 (CK19), and the index of proliferation (KI67).

[0054] All patients underwent liver biopsies with sufficient tumoral and non- tumoral tissue (>1mm2) for subsequent proteomic analysis by MS before undergoing one of the systemic treatments.

[0055] The study was conducted closely with international guidance including the Good Clinical Practices and the Declaration of Helsinki. According to French national laws, the study was approved by a national Ethics Committee (CPP Est III, CER-BDX 2024-106) and Competent Authority (ANSM) has been informed. The processing and use of data collected as part of this study have been done in accordance with the General Data Protection Regulation (GDPR - EU 2016 / 679) and complies with Reference Methodology" (MR-004) of French Data Protection laws in force. All alive participants have been provided comprehensive information notice regarding the purpose of the study, possible risks and expected benefits of the study, and participant's rights (voluntary participation and freedom to decline participation in the trial).

[0056] Patients were included if they were 18 years or older, with a histologically confirmed and radiologically measured advanced HCC.

[0057] Exclusion criteria for the study were: the lack of biopsy from healthy liver and / or the tumor or insufficient quantity and / or quality of sample; an incomplete treatment regimen such as more than 15 days of interruption of sorafenib; another systemic treatment prior to introduction of sorafenib; a mixed component HCC such as hepatocholangiocarcinoma or neuroendocrine contingent; a pathological evidence obtained from a metastasis sample; an extra-hepatic recurrence. Follow-up

[0058] We analyzed responses and survivals of patients who underwent sorafenib regimen including OS, and PFS. The initial staging at diagnosis was based on computed tomography-scan (CT-scan) or magnetic resonance imaging (MRI). The duration of treatment was calculated from the day of the beginning of one of the regimens to the day of disease progression, a shift to other treatment regimens, or when the patient was lost to follow up or dead. We followed tumor response by either abdominal CT-scan or MRI, according to Response Evaluation Criteria in Solid Tumors (RECIST; version 1.1), and AFP levels. The response was determined according to radiological tumor assessment and / or AFP levels assessed in the first three months after treatment initiation. An objective response corresponded to a >30% tumor regression and / or a >50% decrease in AFP levels. A >20% tumor increase corresponded to a progression.

[0059] Laser capture microdissection and fixation reversion

[0060] One square millimeter of tumoral and non-tumoral tissues were microdissected with a PALM type 4 (Zeiss) laser micro-dissector from FFPE (formalin-fixed, paraffin-embedded) 5-pm-thick sections stained by hematoxylin. Each sample was done in 3 replicates on serial sections. The proteins were extracted, the fixation reversed and the proteins reduced and alkylated as previously described (Henriet E et aL: "Argininosuccinate synthase 1 (ASS1): A marker of unclassified hepatocellular adenoma and high bleeding risk". Hepatology. 2017;66:2016-2028 ([2])).

[0061] Samples preparation for proteomic analysis

[0062] Proteins were desalted and digested either in SDS-PAGE gel or a Single-pot, solid-phase-enhanced sample preparation (SP3) method (Henriet E et al. ([2]), Hughes CS et al.: "Single-pot, solid-phase-enhanced sample preparation for proteomics experiments". Nat Protoc. 2019; 14:68- 85 ([3])).

[0063] LC-MS / MS analysis and raw treatment mass spectrometry data.

[0064] NanoLC-MS / MS analysis were performed using different generation of mass spectrometers.

[0065] Statistics and bioinformatics analysis for proteomic data

[0066] The means of protein abundance or ratio between Responders and Progressors were compared using the Wilcoxon-Mann-Whitney U-test. Log-rank test was used to compare the survival between the biomarkerpositive and biomarker-negative patients. A p-value < 0.05 was considered statistically significant. The R version 4.3.3 was used to perform principal component analysis (PCA) with R FactorMineR and factoextra packages. Missing data were imputed using a R missForest package and the LOOCV with a R caret package.

[0067] Integrative biology

[0068] Gene set enrichment analysis (GSEA) was performed against Ingenuity Pathways (I PA, Qiagen) database or Gene Ontology Database Cellular Component Database. Functional pathway and interaction analyze were carried out with the I PA platform.

[0069] Statistical analyses for clinical data

[0070] For the description of population characteristics, since the data are not normally distributed, the results are expressed in median (interquartile range) for quantitative data and in percentage for qualitative data. The comparison of quantitative variables was done using the Kruskal-Wallis test. The comparison of qualitative variables was done using the Fisher exact test as appropriate. A significance level of 5% was applied. Survival analyses were estimated using the Kaplan-Meier method. The logrank test allowed comparing survival curves established by the Kaplan- Meier method.

[0071] Univariate and multivariate survival analyses were performed using the Cox model. Variables that had a p-value unless that 0.20 in univariate analysis were integrated in the multivariate analysis.

[0072] Three dimensional cell culture

[0073] Huh7 cells and Hep3b cells were grown in Dulbecco’s Modified Eagle Medium (DMEM) containing 1g / L of glucose and 10% dialyzed FBS, and kept in 5% CO2 at 37°C. To remove excess lactate present in the FBS, FBS was dialyzed overnight through a 3.5kDa cut-off membrane (Thermo Fisher Scientific, #68035) in a 100-fold volume of PBS. Cells were seeded in Ultra-Low Attachment 96-well plates (Corning Costar #7007) for spheroid formation, and kept under a slight and constant agitation during all 3D culture.

[0074] One day after seeding, IACS-010759 (Abeam, #ab286961) or vehicle (DMSO 1%) was added to the culture medium. In order to obtain the same number of cells and the same spheroid size at day 3, 4000 cells per well were seeded for control conditions versus 8000 cells per well for IACS-010759 treated conditions.

[0075] At day 3 (after 48 h of treatment), peripheral blood mononuclear cells (PBMCs) from healthy donor blood were added to spheroid wells. PBMCs were obtained from buffy coats by the means of Ficoll density gradient centrifugation and frozen. The day of the experiment PBMCs were thawed and counted using Trypan blue to assess viability. PBMCs were seeded at a ratio of one tumor cell for ten PBMCs in spheroids wells in the same culture medium.

[0076] At day 5 (after 48 h of co-culture), spheroids were washed by pipetting up and down each spheroid to remove PBMCs that did not infiltrate. For flow cytometry, Huh7 spheroids were collected, dissociated with 0.25% trypsin and cells were stained with Propidium Iodide (Sigma, #P4864) to assess viability. For immunohistochemistry, Huh7 spheroids and Hep3b spheroids were collected and fixed overnight at 4°C in 4% paraformaldehyde.

[0077] Metabolism in vitro analyses

[0078] Oxygen consumption rate (OCR) assessment was done for Huh7 spheroids at day 3 (after 48 h of IACS-010759 or vehicle treatment). XFe 96-well microplates were coated with collagen type I solution according to Campioni et al protocol (Campioni G, Pasquale V, Busti S, Ducci G, Sacco E, Vanoni M. An Optimized Workflow for the Analysis of Metabolic Fluxes in Cancer Spheroids Using Seahorse Technology. Cells. 2022; 11 :866 ([4]). Four spheroids were plated in each well in 160 pl of Seahorse XF Base Medium (DMEM medium, pH 7.4, without Phenol red and glucose, supplemented with 2mM glutamine, 1 mM sodium pyruvate, NaCI 0.9% and FBS 10%) and prewarmed to 37°C. The cells were incubated in a CO2-free incubator at 37 °C for one hour. OCR was measured at baseline (basal OCR), and then successively after Oligomycin (inhibitor of ATP synthase), FCCP (mitochondrial uncoupling agent), Rotenone and Antimycin A (inhibitors of Complex I and Complex III respectively) injection. OCR was normalized on protein content.

[0079] For lactate quantification, spheroid supernatants were collected at day 3 (after 48 h of IACS-010759 or vehicle treatment). A sample of cell- free culture medium (low glucose DMEM with 10% dialyzed FBS) was used as the background control (basal lactate concentration in culture medium). Metabolic extractions on the spheroid supernatants (100 pl) were performed using the ethanol boiling method already described (Ceschin J, Saint-Marc C, Laporte J, Labriet A, Philippe C, Moenner M, et al. Identification of Yeast and Human 5-Aminoimidazole-4-carboxamide-1-p-d- ribofuranoside (AICAr) Transporters. J Biol Chem. 2014;289:16844-16854 ([5])). Metabolites were then separated by high performance ionic chromatography on an Integrion chromatography station (Thermo Electron) equipped with an AS11-HC 4p column (250 x 2 mm; Thermo Electron) and using the KOH gradient described (Pinson B, Moenner M, Saint-Marc C, Granger-Farbos A, Daignan-Fornier B. On-demand utilization of phosphoribosyl pyrophosphate by downstream anabolic pathways. J Biol Chem. 2023;299: 105011 ([6])). Lactate was detected by conductometry (Dionex Integrion conductivity detector, Thermo-Electron) and quantified by comparison with standard curve obtained with pure lactate (Sigma-Aldrich #71718). Lactate production after 3 days was determined by subtracting the lactate concentration in spheroid supernatants from that found in the cell-free culture medium and was normalized for each sample to the total protein amount in spheroids.

[0080] Flow cytometry

[0081] Spheroids of Huh7 cells were incubated for 48 h in Ultra-Low Attachment 96-well plates with CellTrace Violet-labeled PBMCs at a ratio 1 :10 in complete medium (low glucose DMEM with 10% dialyzed FBS). Briefly, 1 million PBMCs were incubated with 1 pM CellTrace Violet dye (Invitrogen, #C34557) for 20 minutes at 37°C, the cells were then incubated with complete medium for 5 minutes at room temperature and centrifuged before seeding with spheroids. Spheroids were collected at day 5 (after 48 h of co-culture), washed and dissociated with 0.25% trypsin. Viable cells were analyzed by side scatter and propidium iodide (PI) negative staining. Analysis was performed on BD FACSCanto TM II. Immune infiltration was assessed by quantifying the ratio of viable CellTrace Violet positive PBMCs to total viable cells.

[0082] Immunohistochemistry

[0083] Spheroids were collected at day 5 (after 48 h of co-culture), washed and fixed overnight at 4°C in 4% paraformaldehyde (PFA) before immunohistochemistry analysis of immune infiltration. Spheroids were embedded in histogel, dehydrated and embedded in paraffin. Samples were cut into 3pm-sections and heated at 60°C for at least 2 h. Paraffin removal and rehydration were done using Dako PT-link at pH 6 for 20 min. Slides were incubated with H2O2 for 10 min at room temperature to prevent endogenous peroxidase activity. Immunohistochemistry was performed using CD45 antibody (mouse monoclonal antibody, 1:400, Dako Agilent #M070101-2). Secondary antibody was FLEX anti-HRP antibody and the staining procedure was performed using DAB (EnVision Flex, Dako Agilent). All images were obtained using a microscope Nikon DS-Ri2 and analyzed using QuPath vO.5.1 software.

[0084] Statistical analyses for in vitro experiments

[0085] Each value is presented as mean + standard error of the mean (SEM). All experiments were conducted in triplicate or more. The Student’s t-test was employed for comparisons between two groups. For comparisons between multiple groups, one-way analysis of variance (ANOVA) was used. A p-value of less than 0.05 was considered as significant. Statistical analyses were performed using GraphPad Prism 10.2.2 (GraphPad Software Inc).

[0086] Results

[0087] Patient selection

[0088] We compared the tumoral proteomic profile of 9 patients with objective responses under sorafenib with 7 patients who progressed under sorafenib (Table 1). Most characteristics at the beginning of the treatment were similar between the two groups. There were more ALD patients in the responders’ group (66.7% vs none in the progressors’ group, p=0.42), they had more frequently cirrhosis (77.8%% vs 33.9%, p=0.3). Progressors had a worse performance status score (OMS = 1 for 28.6% of them vs none in the responder’s group). 2 patients who did not respond to sorafenib had a massive trabecular contingent (28.6% vs none in the responders’; p=0.41).

[0089] 2 patients in the responder’s group have an ALD and HCV. 3 patients who did not respond to sorafenib had been previously treated by surgery (N=2) or radiofrequency (N=1) for local HOC.

[0090] Table 1 : Clinical characteristics of patients treated with sorafenib selected for proteomic analysis. Data are numbers of patients (%) or medians (IQRs). AFP: alpha-foetoprotein; ALBI: Albumin-Bilirubin; ALD: alcohol-related liver disease; BCLC: the Barcelona Clinic Liver Cancer classification; BMI, body mass index; HBV: hepatitis B virus; HCV: hepatitis C virus; IQR, interquartile range. MAFLD: metabolic-associated fatty liver disease; OMS: performance status score.

[0091] Clinical outcomes

[0092] Median OS was numerically higher in responders to sorafenib than in progressors (median OS of 12.2 (9.97-NA) months vs 4 (3.07-NA) months respectively) without reaching significance (p=0.21).

[0093] Progression free survival was significantly higher in responders to sorafenib than in progressors (median PFS not reached (6.0-NA) vs 1.5 (NA-NA) months respectively, p=0.00018).

[0094] Specificity of the tumoral proteomic profile to predict treatment response

[0095] We compared the proteomic profiles of responders with progressors under sorafenib treatment. We identified and quantified 3153 proteins per patient (with >2 specific peptides per proteins), allowing us to extract 87 proteins differentially deregulated significantly between responders and progressors. By performing a PCA to this significant dataset, we were also able to differentiate patients according to their response to the TKI thanks to their proteomic profile. (Figure 1). Throught a heat map of this dataset, we confirmed that proteomic profiles were distinct between patients who had achieved an objective response and those who had progressed. We performed biological pathways enrichment analysis by GSEA (Gene set enrichment analysis). In sorafenib progressors, as already described in the literature, cytochrome activities were significantly reduced (Tian L-Y, Smit DJ, Jucker M. The Role of PI3K / AKT / mTOR Signaling in Hepatocellular Carcinoma Metabolism. Int J Mol Sci. 2023;24:2652 ([7]), Zhang Y, Feng J, Mi Y, Fan W, Qin R, Mei Y, et al. Epigenetic Activation of Cytochrome P450 1A2 Sensitizes Hepatocellular Carcinoma Cells to Sorafenib. Drug Metab Dispos. 2024;52:555-564 ([8])). Our proteomic analysis further revealed that proteins associated with protein translation pathways (especially translation, eukaryotic initiation factor (elF) signaling) and mTOR signaling were significantly increased on average by 40%. The only exception, Cdc42, is here significantly overexpressed in responding patients (T / NT median ratio = 1.95) and is not modified (T / NT median ratio = 1.02) in patients progressing under sorafenib. (Figure 2). These data demonstrate the potential and specificity of proteomic profiling for predicting response to different treatments, and also for identifying pathways involved in treatment resistance.

[0096] Discussion

[0097] The lack of predictive markers for response to treatments in advanced HCC still remains a big challenge, and we need to strive for precision medicine to determine the optimal treatment and therapeutic sequence, especially in the context of new available combinations. We demonstrated for the first time that the information required for this predictive strategy was available in the proteome of advanced diagnostic HCC biopsies, using laser capture and high-resolution mass spectrometry on small formalin-fixed paraffin-embedded (FFPE) liver samples. We found distinct proteomic profiles associated with efficacy to sorafenib, compared to non-responders.

[0098] Regarding the proteomic profile linked to sorafenib, we identified biological pathways and serve as a positive control for our approach. Indeed, the PI3K / AKT / mTOR signaling pathway regulates crucial cellular processes in the physiological setting such as proliferation, survival, metabolism, motility, angiogenesis and HOC is frequently associated with mTOR signaling alterations (Tian L-Y et al. ([7])). HCCs with upregulated PI3K / AKT signaling seem to be more aggressive, and at risk for earlier recurrence. EIFs pathways are involved in the initiation step of protein translation and eiF3 (a 13-subunit complex) particularly interacts with PI3K / AKT / mTOR pathway by the phosphatidylinositol-3-kinases, initiating translation (Smith MD et al.: "Assembly of elF3 Mediated by Mutually Dependent Subunit Insertion. Structure. 2016;24:886-896 ([9]), Gufler S et al.: "The translational Bridge between Inflammation and Hepatocarcinogenesis. Cells. 2022; 11:533 (

[0010] )). Increased expression of certain subunits of elF3 has been associated with proliferation, invasion and tumorigenicity in HCC (Gomes-Duarte A et al.: "elF3: a factor for human health and disease". RNA Biol. 2017;15:26-34 (

[0011] ), Zhu Q et al.: "Elevated expression of eukaryotic translation initiation factor 3H is associated with proliferation, invasion and tumorigenicity in human hepatocellular carcinoma". Oncotarget. 2016;7:49888-49901 (

[0012] )). In the context of this study, increased expression of translational actors in patients progressing under sorafenib could mean a reprogramming of the tumor proteome, with neoexpression or increased expression of proteins involved in treatment resistance. Characterizing this translational reprogramming could reveal new pharmacological targets to counter sorafenib resistance.

[0099] Current efforts are focused on identifying non-invasive response predictive markers, such as artificial intelligence, blood biomarkers, circulating DNA, circulating cells, exosomes). Response prediction strategies that are non-invasive are very indirect and are not yet available for HCC patients. Furthermore, the identification of mechanisms of resistance to treatments is not known. Nevertheless, today we have easy access to a large amount of direct information contained in liver biopsies. The proteome is the end product of a multitude of gene expression regulation, transcriptional, translation and post-translational processes. It reflects cellular functionalities and consequently is the most representative of the biological functions involved in tumor phenotype. Analysis of the tumor and adjacent hepatic proteome allows us to access valuable and personalized information on expression deregulation in the tumor and in the tumor microenvironment.

[0100] So far, our methodology demonstrated that we could correctly differentiate response groups and reveal the biological pathways involved in resistance to sorafenib treatment, and paves the way for precision medicine in cases of advanced HCC.

[0101] Reference List Eslam M, Sanyal AJ, George J, Sanyal A, Neuschwander-Tetri B, Tiribelli C, et al. MAFLD: A Consensus-Driven Proposed Nomenclature for Metabolic Associated Fatty Liver Disease. Gastroenterology. 2020; 158: 1999-2014.e1. Henriet E, Abou Hammoud A, Dupuy J-W, Dartigues B, Ezzoukry Z, Dugot-Senant N, et al. Argininosuccinate synthase 1 (ASS1): A marker of unclassified hepatocellular adenoma and high bleeding risk. Hepatology. 2017;66:2016-2028. Hughes CS, Moggridge S, Muller T, Sorensen PH, Morin GB, Krijgsveld J. Single-pot, solid-phase-enhanced sample preparation for proteomics experiments. Nat Protoc. 2019;14:68-85. Campioni G, Pasquale V, Busti S, Ducci G, Sacco E, Vanoni M. An Optimized Workflow for the Analysis of Metabolic Fluxes in Cancer Spheroids Using Seahorse Technology. Cells. 2022;11 :866. Ceschin J, Saint-Marc C, Laporte J, Labriet A, Philippe C, Moenner M, et al. Identification of Yeast and Human 5-Aminoimidazole-4- carboxamide-1-p-d-ribofuranoside (AICAr) Transporters. J Biol Chem. 2014;289:16844-16854. Pinson B, Moenner M, Saint-Marc C, Granger-Farbos A, Daignan- Fornier B. On-demand utilization of phosphoribosyl pyrophosphate by downstream anabolic pathways. J Biol Chem. 2023;299: 105011. Tian L-Y, Smit DJ, Jucker M. The Role of PI3K / AKT / mTOR Signaling in Hepatocellular Carcinoma Metabolism. Int J Mol Sci. 2023;24:2652. Zhang Y, Feng J, Mi Y, Fan W, Qin R, Mei Y, et al. Epigenetic Activation of Cytochrome P450 1A2 Sensitizes Hepatocellular Carcinoma Cells to Sorafenib. Drug Metab Dispos. 2024;52:555- 564. Smith MD, Arake-Tacca L, Nitido A, Montabana E, Park A, Cate JH. Assembly of elF3 Mediated by Mutually Dependent Subunit Insertion. Structure. 2016;24:886-896. Gufler S, Seeboeck R, Schatz C, Haybaeck J. The Translational Bridge between Inflammation and Hepatocarcinogenesis. Cells.

[0102] 2022; 11 :533. Gomes-Duarte A, Lacerda R, Menezes J, Romao L. elF3: a factor for human health and disease. RNA Biol. 2017;15:26-34. Zhu Q, Qiao G-L, Zeng X-C, Li Y, Yan J-J, Duan R, et al. Elevated expression of eukaryotic translation initiation factor 3H is associated with proliferation, invasion and tumorigenicity in human hepatocellular carcinoma. Oncotarget. 2016;7:49888-49901.

Claims

CLAIMS1. An ex vivo method for predicting the response to sorafenib treatment of hepatocellular carcinoma in a subject, comprising the steps of:1) determining expression, in a liver tissue sample from the subject, of proteins associated with protein translation pathways, said sample containing pathological tissue and optionally non-pathological tissue,2) detecting an increase of expression of proteins associated with protein translation pathways, in pathological tissue compared to non- pathological tissue from the liver tissue sample of the subject or from a standard non-pathological liver sample, wherein:If expression of proteins associated with protein translation pathways is increased in pathological tissue, the subject is unlikely to have a response to sorafenib treatment,If expression of proteins associated with protein translation pathways is not increased in pathological tissue, the subject is likely to have a response to sorafenib treatment.

2. An ex vivo method according to claim 1 , wherein protein translation pathways comprise translation signaling and eukaryotic initiation factor (elF) signaling.

3. An ex vivo method according to claim 2, wherein translation signaling comprises at least one protein chosen among PABPC1 , RPL14, RPL7A, RPS2 and RPS3.

4. An ex vivo method according to claim 2, wherein eiF signaling comprises at least one protein chosen among eiF2, such as eiF2S3, eiF3, such as eiF3E, eiF3F and eiF3l, eiF4, such as eiF4A1 and EIFS35. An ex vivo method according to any one of the preceding claims, further performing steps 1) and 2) for Cdc42, wherein: if expression of Cdc42 is increased in pathological tissue, the subject is likely to have a response to sorafenib treatment,If expression of Cdc42 is not increased in pathological tissue, the subject is unlikely to have a response to sorafenib treatment.

6. An ex vivo method according to any one of the preceding claims, wherein protein expression is determined by at least one method selected among proteomic profiling matching, immunohistochemistry, western blot, global or targeted mass spectrometry analysis, and reverse transcriptase PCR.

7. An ex vivo method according to claim 6, wherein proteomic profiling matching comprises determining abundance of proteins of the liver tissue sample from the subject.

8. An ex vivo method according to claim 7, wherein determining the abundance of proteins is carried out by label free proteomic analysis, or by a targeted quantification technique.

9. An ex vivo method according to claim 8, wherein the protein abundance is the relative abundance of a protein of the non-pathological tissue over the corresponding protein in the pathological tissue.

10. An ex vivo method according to any one of the preceding claims, wherein said hepatocellular carcinoma is resectable or non resectable hepatocellular carcinoma.

11. An ex vivo method according to any one of the preceding claims, wherein samples are obtainable by at least one technique selected amongmacrodissection, laser microdissection of a formalin-fixed paraffin- embedded or frozen tissue, and fresh biopsy.

Images