We carried out several studies with the aim of deepening global understanding of the pathogenesis of HCC, a cancer with an incidence of 750,000 new cases yearly that ranks as the 3rd cause of global cancer-related death. Sorafenib is the sole systemic therapy for this disease, and thus, identifying novel drivers and targets for therapies remains a relevant unmet medical need. In parallel, we studied rare liver tumors that lack clear molecular understanding and effective targeted therapies such as ICC and FLC.
A. - Genomic studies and molecular classification of HCC
Following our previous research on molecular classification of HCC (Hoshida, NEJM 2008; Chiang, Cancer Res 2008; Hoshida, Can Res, 2009) we generated #1) A 186-gene signature from adjacent cirrhotic tissue with prognostic capacity (Hoshida et al., Gastroenterology, 2013). This gene signature identifies a subgroup of hepatitis C-related early-stage cirrhotic patients at high-risk of HCC development, which might be used for selection in chemo preventive studies. #2) A refined prognostic signature from tumoral tissue. After providing and integrative analysis of clinical, pathological and genomic signatures to predict outcome in HCC patients (Villanueva et al., Gastroenterology 2011), a 5-gene signature (HN1, RAN, RAMP3, KRT19, and TAF9) associated with disease-specific survival was generated and validated in four different Eastern and Western cohorts (Nault et al., Gastroenterology, 2013). #3) microRNAs-based molecular classification of HCC. We provided a miRNA-based classification of HCC according to miRNA expression profiles (Toffanin, Gastroenterology. 2011). We demonstrated also that miR-517a is an oncomiR that promotes tumor progression. Our finding offered the rationale for developing therapies targeting miR-517a for patients with HCC. In addition, by using a novel platform interrogating 2332 snoRNAs we identified aberrant overexpression of the locus DLK1-DI03 which included 45 miRNAs and 41 snoRNAs. The relevance of this locus was demonstrated in an animal model generated for that purpose which selectively overexpressed the locus and induced development of HCC at 6 mo. These findings demonstrated the accuracy of in vivo gene targeting in modeling human cancer and suggested future applications in studying tumors in diverse animal species. (Wang, PNAS, 2012).
B.- Oncogenic drivers and signaling pathways in HCC
The European project HEPTROMIC (www.heptromic.eu) based on cancer genomics was awarded with a European Commission grant (FP-7 Health grant No. 259744-2). The project was coordinated by Dr. Llovet and includes the participation of 6 research centers [M. Esteller (Spain), J. Zucman-Rossi (France), V. Mazzaferro (Italy), L. Zender (Germany), T. Golub (Harvard, USA)] and 2 small-medium enterprises that are involved in the transfer of knowledge to industry (Diagenode, Belgium and TCLand, France). Several articles were published as a result of this project: Vettert, Hepatology 2012; Keng, Hepatology 2013; Nault, Gastroenterology 2013; Nault, Nat. Commun. 2013. Among those, though, the most relevant define novel mutations identified by whole exome sequencing associated with epidemiological events such as HCV, alcohol and HBV infection (Nault, Nature Genetics 2015; sub). In addition by exploring the whole methylome we identified a) Methylome-based 35 gene signature with prognostic implications. This discovery has been protected with a patent filled on 2014: Methods for the prognosis of Hepatocellular Carcinoma (European Application Number: EP14382535.4), and b) identification of novel tumor suppressors by exploring the whole methylome (Villanueva, Hepatology 2015).
Following our previous publications aiming to dissect the role of signaling cascades in the pathogenesis of HCC (mTOR, Ras signaling), we have been exploring additional pathways such as 1) IGF signaling: Our integrative genomic analysis showed activation of IGF signaling in 20% of HCCs, particularly in tumors from the Proliferation subclass. Effective blockage of IGF signaling with A12 monoclonal antibodies supported the rationale for testing this therapy in clinical trials (Tovar, J Hepatol. 2010). 2) Wnt signaling Analysis of the Wnt pathway alterations at the transcriptome, immunohistochemical and DNA level in 642 HCCs revealed that Wnt pathway is activated in 49% of patients. We identified two subclasses of tumors with Wnt activation: a) the CTNNB1 subclass, characterized by mutations and nuclear translocation of β-catenin and by transcriptional activation of Wnt-related liver-specific genes, and b) the Wnt-TGFβ subclass, with deregulation of genes that typically activate the Wnt pathway and with WT β–catenin. (Lachenmayer, Clin Cancer Res. 2012). 3) Notch signaling. We generated transgenic mice expressing Notch in hepatoblasts and cholangiocytes, under the control of the albumin promoter and the enhancer of alpha-fetoprotein. These mice developed HCC with a penetrance of 100% at 12 months. Its transcriptome analysis allowed us to generate a gene signature that detects Notch activation in 30% of patients with HCC. In addition, we showed that blocking Notch activation in vitro using either a) inhibitors of γ-secretase activity or b) a dominant negative tool, significantly reduced cell proliferation. (Villanueva et al., Gastroenterology, 2012). 4) Tumor suppressors and HDAC inhibitors. We identify alterations in the expression and gene copy in histone deacetylases HDAC3 and HDAC5 in 334 HCCs. Treatment of several cell lines and HCC xenograft models with a pan HDAC inhibitor (panobinostat) alone or in combination with sorafenib, caused potent anti-tumor effects, affecting their survival (Lachenmayer, J Hepatol., 2012). In a parallel study by using whole methylome analysis and experimental functional studies we identified SMPD3 as a novel tumor suppressor in HCC (Revill, Gastroenterology 2013).
C.- Genomic studies and molecular classification of intrahepatic cholangiocarcinoma
Only 30% of ICCs are amenable for curative resection, while there is no standard of care for systemic treatment. We performed an integrative genomic analysis (transcriptome, SNP array and mutations) in 149 ICC samples and used next-generation sequencing for discovery of novel oncogenic drivers. The study supported by a grant from AECC (2011-2016) led to three main publications A) Molecular classification of ICC (Sia et al. Gastroenterology, 2013). We described a molecular classification of ICC defining two molecular classes: the inflammation class, characterized by the activation of pathways involved in inflammation, cytokines activation and STAT-3 activation, and a proliferation class, characterized by the activation of pathways involved in proliferation and survival (Ras pathway, mTOR pathway, MET pathway). We found that aggressive ICCs with poor prognosis are enriched in the ICC proliferation class. B) Role of IDH mutations in ICC development and progression (Saha, Nature 2014). IDH mutations are present in 17% of ICC cases. By using GEMM we showed that IDH mutations in hepatoblast prevent hepatocyte differentiation and induce cholangiocyte proliferation. Albeit as standalone genomic hit IDHmut do not induce ICC, on a background of RAS mutations, they increase oncogenesis providing the rationale for exploring IDH-targeted therapies for this orphan disease, and c) Identification of a novel oncogenic fusion FGFR2-PPHLN1 and ARAF mutations (Sia, Nature Commun 2014, in press). By applying RNA-seq and exome sequencing we discovered a gene translocation leading to an oncogenic fusion including FGFR2 with intact kinase domain. This fusion is present in 16% of cases and has transforming properties and responds to specific inhibitors. As a result of this finding we have been involved in the design of a phase II study aiming to test specific inhibitors in patients with ICC and FGFR2 molecular aberrations. In addition we have protected the IP of this discovery with a patent: Methods for Diagnosing and Treating Intrahepatic Cholangiocarcinoma (Application Number: US 61/876,451). Overall we described a novel landscape of 9 actionable mutations in ICC, which involves 70% of these patients.
D. - Molecular classification of fibrolamellar hepatocellular carcinoma .
Fibrolamellar carcinoma (FLC) accounts for less than 1% of liver cancers, and unlike HCC it develops in non-cirrhotic livers of children/young adults with unknown etiologic factors. Treatment options are limited to surgical intervention. We have provided an integrative genomic analysis from a large series of FLC patients revealing a molecular classification (3 subclasses), description of mutational landscape (DNAJB1-PRKACA fusion 79%, BRAC2) and proposing a prognostic signature (Cornellà, Gastroenterology 2014).
Please visit the Llovet Lab.
