Genomic Decoding of Intrahepatic Cholangiocarcinoma Reveals Therapeutic Opportunities

Abstract

See “Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes,” by Sia D, Hoshida Y, Villanueva A, et al, on page 829. See “Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes,” by Sia D, Hoshida Y, Villanueva A, et al, on page 829. Cholangiocarcinoma is a devastating malignancy with increasing frequency worldwide.1El-Serag H.B. Rudolph K.L. Hepatocellular carcinoma: epidemiology and molecular carcinogenesis.Gastroenterology. 2007; 132: 2557-2576Abstract Full Text Full Text PDF PubMed Scopus (4467) Google Scholar This cancer type arises in the epithelium lining intrahepatic or extrahepatic biliary ducts. Intrahepatic cholangiocarcinoma (ICC) is classified as a peripheral-type tumor of the interlobular bile ducts, whereas tumors designated hilar are generally considered extrahepatic and originate from the main hepatic ducts or at the bifurcation of the common hepatic duct. ICC is often diagnosed at a late stage owing to the anatomical location, growth patterns and pathobiological heterogeneity. Therefore, the clinical management of the disease offers limited therapeutic options, that is, in the absence of advanced unresectable or metastatic disease, operative resection remains the only curative treatment for patients with ICC.2Blechacz B. Komuta M. Roskams T. et al.Clinical diagnosis and staging of cholangiocarcinoma.Nat Rev Gastroenterol Hepatol. 2011; 8: 512-522Crossref PubMed Scopus (459) Google Scholar However, intrahepatic recurrence is frequent after attempted curative resection and serves as a confounding variable.3Miwa S. Miyagawa S. Kobayashi A. et al.Predictive factors for intrahepatic cholangiocarcinoma recurrence in the liver following surgery.J Gastroenterol. 2006; 41: 893-900Crossref PubMed Scopus (122) Google Scholar Standard therapy administered to ICC patients is typically palliative, although a recent phase III clinical trial in patients with advanced biliary tract cancers reported a significant increase in the median overall survival by 11.7 months after combined chemotherapy of gemcitabine plus cisplatin versus gemcitabine alone.4Furuse J. Okusaka T. Bridgewater J. et al.Lessons from the comparison of two randomized clinical trials using gemcitabine and cisplatin for advanced biliary tract cancer.Crit Rev Oncol Hematol. 2011; 80: 31-39Crossref PubMed Scopus (33) Google Scholar, 5Valle J. Wasan H. Palmer D.H. et al.Cisplatin plus gemcitabine versus gemcitabine for biliary tract cancer.N Engl J Med. 2010; 362: 1273-1281Crossref PubMed Scopus (2713) Google Scholar Despite progress, discovery of new agents and actionable targets are urgently required. To achieve this goal in a rare neoplasm like ICC with diverse clinicopathology, heterogeneous underlying biology and diverse molecular landmarks, multicenter studies may be necessary to acquire sizable cohorts for in-depth genomic and clinical analyses. Although genomic studies of ICC are limited (reviewed in Andersen and Thorgeirsson6Andersen J.B. Thorgeirsson S.S. Genetic profiling of intrahepatic cholangiocarcinoma.Curr Opin Gastroenterol. 2012; 28: 266-272Crossref PubMed Scopus (80) Google Scholar), progress was recently made and several omics–based studies have become accessible detailing aspects of the cholangiocarcinoma genome landscape, such as transcriptomics,7Andersen J.B. Spee B. Blechacz B.R. et al.Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors.Gastroenterology. 2012; 142: 1021-1031Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar, 8Oishi N. Kumar M.R. Roessler S. et al.Transcriptomic profiling reveals hepatic stem-like gene signatures and interplay of miR-200c and epithelial-mesenchymal transition in intrahepatic cholangiocarcinoma.Hepatology. 2012; 56: 1792-1803Crossref PubMed Scopus (175) Google Scholar epigenomics,9Wang P. Dong Q. Zhang C. et al.Mutations in isocitrate dehydrogenase 1 and 2 occur frequently in intrahepatic cholangiocarcinomas and share hypermethylation targets with glioblastomas.Oncogene. 2012 Jul 23; ([Epub ahead of print])Google Scholar and whole-exome sequencing.10Ong C.K. Subimerb C. Pairojkul C. et al.Exome sequencing of liver fluke-associated cholangiocarcinoma.Nat Genet. 2012; 44: 690-693Crossref PubMed Scopus (369) Google Scholar In an earlier issue of Gastroenterology, an integrated genetic and genomic analysis by Andersen et al7Andersen J.B. Spee B. Blechacz B.R. et al.Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors.Gastroenterology. 2012; 142: 1021-1031Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar provided new insights into both the molecular pathogenesis and promising treatment options for peripheral and hilar cholangiocarcinoma patients. This was the first study of a large cohort of patients with cholangiocarcinoma intended for detailed genomic analysis. The central aim of this study was to identify previously unrecognized homogeneous subgroups of patients, and to provide potential novel therapeutic targets appropriate for these patients. In that study,7Andersen J.B. Spee B. Blechacz B.R. et al.Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors.Gastroenterology. 2012; 142: 1021-1031Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar 2 prognostic categories of patients distinguished by a 238-gene classifier including 36 predictive survival genes were described. Each prognostic class contained 2 subgroups (subgroup I, II and subgroup III, IV) characterized by distinct molecular traits. The unique patient subgroups were highlighted, based on BRAF/KRAS mutations, and co-activation of multiple oncogenic addiction pathways (eg, receptor tyrosine kinases [RTKs] and downstream RAS/RAF/MAPK signaling) improved the molecular classification and outcome prediction. Patients with the worst prognosis (subgroup III) were characterized by transcriptional enrichment of genes regulating inflammation and proteasome activities, suggesting a combination of proteasome (eg, bevacizumab) and tyrosine kinase inhibitors (eg, trastuzumab, lapatinib) or antiinflammatory drugs (eg, celecoxib) to present new therapeutic options for cholangiocarcinoma patients. Although recent molecular insights into ICC have improved our knowledge of the underlying pathobiology of this disease, understanding the genetic mechanisms involved in its development remains inadequate. There is an urgent need to undertake integrative, multidimensional studies with homogeneous patient cohorts to fuel translational genomics-based discoveries and implement these innovations in today's clinical decision making, to optimized therapy and future trials. This situation has been significantly improved by the paper of Sia et al in this issue of Gastroenterology.11Sia D. Hoshida Y. Villanueva A. et al.Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes.Gastroenterology. 2013; 144: 829-840Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar Sia et al performed gene expression, copy number (CN) and mutational analyses on archival samples obtained from 149 ICC patients diagnosed between 1995 and 2007. Their classification describes 2 broad molecular subclasses, the “proliferation” and “inflammation” classes. This classification was based on evaluation of functional characteristics and gene set enrichment of oncogenic signaling pathways. The patients in the proliferation class display aggressive disease with poor prognosis, namely, shorter survival and frequent recurrence characterized by molecular enrichment of oncogenic pathways (eg, RAS/RAF/MAPK, VEGF, and PDGF). The patients in the inflammation class have improved prognosis and are characterized by induction of immune-related signaling, dendritic cell and T helper 2-type activation of antiinflammatory interleukins (IL-4 and IL-10), establishing a central network of IL-10 and STAT3. It is not entirely a surprise that neither nuclear factor-κB activation nor the number of infiltrating immune cells could be distinguished within the classification. ICC is in general considered a stroma-rich and extremely aggressive cancer with an inflammatory and immune cell-controlled microenvironment regulating its growth, angiogenesis, invasion, and metastasis. In general, cholangiocarcinoma is a dismal disease; however, it is somewhat counterintuitive that patients in the inflammation class have an outcome better than those in the proliferation class, considering that tumor initiation and progression is tightly linked with chronic inflammation. The shift to a Th2 response and escape from the host anti-tumor immune response, particularly induction of M2 class tumor-associated macrophages by IL-4 and IL-10, is characteristic of tumor promotion and suppression of protective immune-related cells. Indeed, a strategy inhibiting the M2 macrophages could be beneficial for stimulating the innate immunity and target the tumor. However, our understanding of the microenvironment and crosstalk between the immune and cholangiocarcinoma cells is limited. In the present study by Sia et al,11Sia D. Hoshida Y. Villanueva A. et al.Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes.Gastroenterology. 2013; 144: 829-840Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar the authors used an interesting approach to establish the initial genomic classification applying an unsupervised (ie, balanced determination of unknown classes) non-negative matrix factorization clustering. This algorithm was developed as a class discovery tool to identify context-dependent molecular patterns and does not rely on distance computation or outcome parameters. The robustness of clustering and possible misclassification was then corrected using a leave-one-out cross-validation procedure to optimize the overall classification. The non-negative matrix factorization clustering demonstrates a robust intersample correlation when the rank K = 2 classes is assumed and the cophenetic distance (ie, a measure of the stability of the clusters) is greatest. However, the heatmap of the top 100 differentially expressed genes within the gene signature shows that the proliferation class extends beyond the clustering, which may explain the limitation in the prognostic classification with marginally significant overall survival (P < .048) and recurrence (P < .03). Regardless, development of the classification based on differences in the tumor molecular traits and optimizing its robustness by applying multiple class prediction models during leave-one-out cross-validation to minimize the size of the gene signature is arguably the correct path. A total of 1565 significant genes is an outsized signature that needs further attention because it is difficult to adapt for clinical practice. It may, however, be possible to optimize the gene signature using Cox proportional hazard modeling to predict the key survival genes. Next, Sia et al examined the landscape of chromosomal aberrations and select mutations (BRAF, EGFR, and KRAS) in a subset of patients. In accord with previous studies, mutations in KRAS were the most frequent variant. The majority of regional CN alterations showed no apparent accumulation in either class; however, the authors showed 1 interesting candidate with focal deletion containing the Hippo pathway gene SAV1, which is enriched in the proliferation class and significantly associated with reduced gene expression. Indeed, this may warrant a large translational study into the Hippo network in ICC; lately, this pathway has attracted attention in numerous genetic studies, detailing its involvement in stem cell regulation and development of primary liver cancers (reviewed in Avruch et al12Avruch J. Zhou D. Fitamant J. et al.Mst1/2 signalling to Yap: gatekeeper for liver size and tumour development.Br J Cancer. 2011; 104: 24-32Crossref PubMed Scopus (93) Google Scholar and Sheng et al13Zheng T. Wang J. Jiang H. et al.Hippo signaling in oval cells and hepatocarcinogenesis.Cancer Lett. 2011; 302: 91-99Crossref PubMed Scopus (35) Google Scholar). Refinement of the patient classification was attempted by data integration, suggesting a total of 6 patient groups. Whereas the non-negative matrix factorization clustering indicated the presence of 5–7 subgroups, the refined integrative classification is not apparent. Interestingly, a subgroup of ICCs in the proliferation class showed molecular similarities with several previously published hepatocellular carcinoma (HCC) signatures, all demonstrating an association with poor prognosis and a possible progenitor cell origin. Indeed, this similarity with poor outcome HCC was documented in several ICC studies,7Andersen J.B. Spee B. Blechacz B.R. et al.Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors.Gastroenterology. 2012; 142: 1021-1031Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar, 8Oishi N. Kumar M.R. Roessler S. et al.Transcriptomic profiling reveals hepatic stem-like gene signatures and interplay of miR-200c and epithelial-mesenchymal transition in intrahepatic cholangiocarcinoma.Hepatology. 2012; 56: 1792-1803Crossref PubMed Scopus (175) Google Scholar Also, genomic8Oishi N. Kumar M.R. Roessler S. et al.Transcriptomic profiling reveals hepatic stem-like gene signatures and interplay of miR-200c and epithelial-mesenchymal transition in intrahepatic cholangiocarcinoma.Hepatology. 2012; 56: 1792-1803Crossref PubMed Scopus (175) Google Scholar, 14Seok J.Y. Na D.C. Woo H.G. et al.A fibrous stromal component in hepatocellular carcinoma reveals a cholangiocarcinoma-like gene expression trait and epithelial-mesenchymal transition.Hepatology. 2012; 55: 1776-1786Crossref PubMed Scopus (96) Google Scholar, 15Woo H.G. Park E.S. Cheon J.H. et al.Gene expression-based recurrence prediction of hepatitis B virus-related human hepatocellular carcinoma.Clin Cancer Res. 2008; 14: 2056-2064Crossref PubMed Scopus (135) Google Scholar and genetic16Cazals-Hatem D. Rebouissou S. Bioulac-Sage P. et al.Clinical and molecular analysis of combined hepatocellular-cholangiocarcinomas.J Hepatol. 2004; 41: 292-298Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar, 17Fujii H. Zhu X.G. Matsumoto T. et al.Genetic classification of combined hepatocellular-cholangiocarcinoma.Hum Pathol. 2000; 31: 1011-1107Crossref PubMed Scopus (85) Google Scholar, 18Imai Y. Oda H. Arai M. et al.Mutational analysis of the p53 and K-ras genes and allelotype study of the Rb-1 gene for investigating the pathogenesis of combined hepatocellular-cholangiocellular carcinomas.Jpn J Cancer Res. 1996; 87: 1056-1062Crossref PubMed Scopus (31) Google Scholar, 19Murata M. Miyoshi Y. Iwao K. et al.Combined hepatocellular/cholangiocellular carcinoma with sarcomatoid features: genetic analysis for histogenesis.Hepatol Res. 2001; 21: 220-227Crossref PubMed Scopus (28) Google Scholar analyses of mixed HCC-cholangiocarcinoma have demonstrated shared gene expression signatures and closely related genetic aberrations between ICC and HCC tumors, suggesting the provocative hypothesis that these may be derived from a common origin or precursor cell(s). This is compelling from a therapeutic standpoint and deserves serious attention in the future. Detailed analysis of HCC-cholangiocarcinoma may help to improve class-specific treatment decisions, particularly for tumors that may originate from common progenitors (ie, characterized by potential common “driver” genes or mutations). Notably, it may be interesting to test multikinase inhibitors (eg, sorafenib) in this specific patient group, particularly because it is an approved therapy with efficacy in HCC. The authors highlight the discussion on a “stratified” treatment strategy blocking specific oncogenic addiction loops versus the “class-specific” therapeutic approach, where treatment is based on classification and characterization of the homogeneous subclasses. In this context, we find it is encouraging to combine the 2 most extensive genomic studies on cholangiocarcinoma (ie, Andersen et al7Andersen J.B. Spee B. Blechacz B.R. et al.Genomic and genetic characterization of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors.Gastroenterology. 2012; 142: 1021-1031Abstract Full Text Full Text PDF PubMed Scopus (375) Google Scholar and Sia et al11Sia D. Hoshida Y. Villanueva A. et al.Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes.Gastroenterology. 2013; 144: 829-840Abstract Full Text Full Text PDF PubMed Scopus (384) Google Scholar; Figure 1) ; several common key aspects may stimulate future research directions. Stratification of class-specific risk groups among ICC patients may be essential for clinical success. RTKs, for example, enrichment of HER2 signaling in patients with poor outcome and association with EGFR and KRAS mutations was shown in both studies despite the lack of CN amplification. Tyrosine kinase inhibitors may not be effective as monotherapies against drug-resistant cholangiocarcinoma; however, consistent with the class-specific data and in combination with several compelling drug choices, both studies support the potential of targeting RAS/RAF/MAPK, such as the MEK1/2 inhibitor (selumetinib), which has been used against metastatic biliary cancers,20Bekaii-Saab T. Phelps M.A. Li X. et al.Multi-institutional phase II study of selumetinib in patients with metastatic biliary cancers.J Clin Oncol. 2011; 29: 2357-2363Crossref PubMed Scopus (243) Google Scholar or targeting RAS/PI3K/AKT/mTOR signaling (eg, MK2206 or RAD00121Thomas H.E. Mercer C.A. Carnevalli L.S. et al.mTOR inhibitors synergize on regression, reversal of gene expression, and autophagy in hepatocellular carcinoma.Sci Transl Med. 2012; 4: 139ra84Crossref PubMed Scopus (79) Google Scholar). Also, the authors outline the interesting and persuasive option of targeting the JAK–STAT pathway, taking advantage of the novel STAT3 and JAK1–JAK2 inhibitors. JAK2 inhibitors such as AZD1480 have been shown to potently and concomitantly block JAK–STAT signaling22Hedvat M. Huszar D. Herrmann A. et al.The JAK2 inhibitor AZD1480 potently blocks Stat3 signaling and oncogenesis in solid tumors.Cancer Cell. 2009; 16: 487-497Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar and the phosphorylation of Jak2, Stat3, MAPK, and Fgfr3.23Scuto A. Krejci P. Popplewell L. et al.The novel JAK inhibitor AZD1480 blocks STAT3 and FGFR3 signaling, resulting in suppression of human myeloma cell growth and survival.Leukemia. 2011; 25: 538-550Crossref PubMed Scopus (102) Google Scholar Causative mutations are especially compelling drug targets both for discovery and clinical success. The authors highlight the need to further study the efficacy of drugs like PLX-4032 targeting BRAFV600E. However, BRAF mutations are rare (2%–4%) and scattered across the classification. Whole-exome sequencing is a more direct approach for the discovery of driver mutations and exome CN alterations to study the clinical usefulness of oncogenic addiction loops. Targeting EGFR is an attractive strategy; however, in both studies KRAS mutations, a predictor of ineffective anti-EGFR antibody therapies, are associated with the same patient groups. Targeting oncogenic drivers, for example, mutations in the IDH1/IDH2 genes,9Wang P. Dong Q. Zhang C. et al.Mutations in isocitrate dehydrogenase 1 and 2 occur frequently in intrahepatic cholangiocarcinomas and share hypermethylation targets with glioblastomas.Oncogene. 2012 Jul 23; ([Epub ahead of print])Google Scholar, 24Borger D.R. Tanabe K.K. Fan K.C. et al.Frequent mutation of isocitrate dehydrogenase (IDH)1 and IDH2 in cholangiocarcinoma identified through broad-based tumor genotyping.Oncologist. 2012; 17: 72-79Crossref PubMed Scopus (558) Google Scholar is compelling and novel IDH-targeted therapies are eagerly anticipated. Alternatively, the IDH genes influence epigenetic control, suggesting a potential benefit of DNA methyltransferase inhibitors.25Andersen J.B. Factor V.M. Marquardt J.U. et al.An integrated genomic and epigenomic approach predicts therapeutic response to zebularine in human liver cancer.Sci Transl Med. 2010; 2: 54ra77Crossref PubMed Scopus (98) Google Scholar Unfortunately, IDH mutations were not examined in this cohort. Notably, epigenetic alterations may explain the observed enrichment of RTK and MET signaling in ICCs without CN amplification. Last, targeting MET (eg, trivantinib) has to be regarded as a promising strategy for ICCs with c-Met overexpression and deregulated MET signaling. In conclusion, whether preferring a “class-specific” or a “stratified” therapeutic strategy, continued implementation of integrative–omics is urgently needed. Detailed information from multiple data layers obtained from rigorously selected patient cohorts together with comprehensive clinicopathologic information can, as elegantly demonstrated in this issue of Gastroenterology, advance our understanding of the molecular pathogenesis of ICC and improve the outcome of future clinical trials. Integrative Molecular Analysis of Intrahepatic Cholangiocarcinoma Reveals 2 Classes That Have Different OutcomesGastroenterologyVol. 144Issue 4PreviewCholangiocarcinoma, the second most common liver cancer, can be classified as intrahepatic cholangiocarcinoma (ICC) or extrahepatic cholangiocarcinoma. We performed an integrative genomic analysis of ICC samples from a large series of patients. Full-Text PDF Covering the CoverGastroenterologyVol. 144Issue 4PreviewAchalasia has been classically characterized by the absence of peristalsis and impaired relaxation of the lower esophageal sphincter and treatment rests on relieving the functional obstruction at the gastroesophageal junction with endoscopic pneumatic dilation (PD), laparoscopic Heller myotomy (LHM), or botulinum toxin injection. The European achalasia trial compared PD and LHM in 201 patients with achalasia and demonstrated similar efficacy of these 2 treatments. Recently, Pandolfini et al described 3 subtypes of achalasia based on high-resolution manometry and response to therapy: Type I (classic), achalasia with minimal esophageal pressurization; type II, achalasia with esophageal compression, and most likely to respond to therapy; and type III, achalasia with spasm. Full-Text PDF