Transcriptomic analysis and mutational status of IDH1 in paired primary-recurrent intrahepatic cholangiocarcinoma

Transcriptomic analysis and mutational status of IDH1 in paired primary-recurrent intrahepatic... Background: Effective target therapies for intrahepatic cholangiocarcinoma (ICC) have not been identified so far. One of the reasons may be the genetic evolution from primary (PR) to recurrent (REC) tumors. We aim to identify peculiar characteristics and to select potential targets specific for recurrent tumors. Eighteen ICC paired PR and REC tumors were collected from 5 Italian Centers. Eleven pairs were analyzed for gene expression profiling and 16 for mutational status of IDH1. For one pair, deep mutational analysis by Next Generation Sequencing was also carried out. An independent cohort of patients was used for validation. Results: Two class-paired comparison yielded 315 differentially expressed genes between REC and PR tumors. Up-regulated genes in RECs are involved in RNA/DNA processing, cell cycle, epithelial to mesenchymal transition (EMT), resistance to apoptosis, and cytoskeleton remodeling. Down-regulated genes participate to epithelial cell differentiation, proteolysis, apoptotic, immune response, and inflammatory processes. A 24 gene signature is able to discriminate RECs from PRs in an independent cohort; FANCG is statistically associated with survival in the chol-TCGA dataset. IDH1 wasmutated in theRECsoffive patients; 4ofthem displayed the mutation only in RECs. Deep sequencing performed in one patient confirmed the IDH1 mutation in REC. Conclusions: RECs are enriched for genes involved in EMT, resistance to apoptosis, and cytoskeleton remodeling. Key players of these pathways might be considered druggable targets in RECs. IDH1 is mutated in 30% of RECs, becoming both a marker of progression and a target for therapy. Keywords: Intrahepatic cholangiocarcinoma, Recurrence, IDH1 mutation, Microarray, Prognostic marker Background inflammation processes, such as cholangitis/primary Intrahepatic cholangiocarcinoma (ICC) is an aggressive sclerosing cholangitis (PSC), secondary biliary cirrhosis, malignancy arising from epithelial cells of the bile ducts choledocholithiasis, hepatolithiasis, cholecystitis, as well and is considered the second most common liver cancer as HCV and HBV infections promote ICC arising and type. Limited success in the clinical management and a progression [3–6]. Conventional chemotherapy, based persistent increase in the incidence world-wide have on combination of gemcitabine (GEM) and platinum made ICC one of the most lethal and fastest growing compounds, and radiotherapy, to date, are not effective malignancies. In the last three decades, a general incre- in improving long-term survival [7, 8]. Moreover, ment of ICC incidence was registered in the Western primary or acquired resistance is inevitable and no countries, and in particular in Italy [1, 2]. Chronic second-line chemotherapy has demonstrated efficacy. It is known that 5-years survival rate of ICC patients re- * Correspondence: caterina.peraldoneia@ircc.it; francesco.leone@ircc.it mains low, between 25 and 35% in most of the case G. Chiorino and F. Leone contributed equally to this work. 1 series. Literature data showed that ICC recurrences Medical Oncology Division, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. 142, km 3.95, 10060 Candiolo, Turin, Italy occur in about half of the patients after surgery with Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 2 of 10 curative intent, frequently during the first year and that IDH1 mutations promoted ICC by blocking hepato- usually in the liver [9]. cyte differentiation with an increased number of hepatic In the last years, different molecular studies were con- progenitors susceptible to other mutations [22]. Further, ducted using intensive high-throughput techniques (i.e. gene patients harboring IDH1 mutations had a distinct tran- expression and microRNA profiling, deep-sequencing), to scriptional signature enriched for hepatic stem cell broaden the knowledge on the biological aspects of ICC genes, identifying a particular subclass of ICC patients progression and to identify potential molecular targets. [23]. In preclinical models, IDH1 mutated cell lines were Genomic and molecular mechanisms involved in the onset, highly responsive to Src inhibitors, such as Dasatinib progression as well as in chemotherapy resistance in ICC and Saracatinib, suggesting potential targeted therapies are poorly documented. Preclinical investigations showed [24]. Recently, different preclinical studies aimed at the involvement of oncogenic pathways in cholangiocarci- studying the efficacy of IDH1 inhibitors have been per- nogenesis; among them, the overexpression of EGFR, formed. Further, phase I, II and III clinical trials were HER2, VEGFR and its ligand, MET, signaling pathways, planned and are ongoing to test the safety and efficacy of which cause a dysregulation of downstream effectors, such IDH1 inhibitors in different malignancies, such as glioma, as Ras/Raf/Mek/Erk and PI3K/Akt/PTEN axes [10–12]. cholangiocarcinoma, AML (NCT02074839, NCT02073994, Recently, using gene profiling techniques, Sia and collabora- NCT02719574, NCT02989857). tors [13]demonstrated thatICC couldbestratifiedonthe The identification of the peculiar molecular alterations bases of molecular characteristics, which correlate with dif- of recurrent lesions is required due to the high rate of ferent prognosis. One hundred and forty-nine ICC were local recurrence of this tumor [25]. To date, there are classified in two main classes according to their gene expres- only few data regarding the mechanisms involved in re- sion profiles; an inflammation class, associated to a “good” current disease. For this reason, in this work, we have prognosis, and a proliferation class, associated to a worse molecularly characterized paired primary/recurrent ICC prognosis [13]. In a work of Andersen and collaborators, tumors in order to provide a panel of markers (mutated 104 ICC samples were analyzed by gene expression profiling or deregulated genes) involved in the progression and the two prognostic groups were confirmed [14]. process of this subtype of tumors as well as of new suit- Recently, genomic analyses were conducted on primary ICC able targets for therapy. tumors by mutational profiling using different techniques, defining a broad range of mutations, according to the co- Methods horts analyzed. The most commonly observed alterations Patients were within TP53, KRAS, PI3K, BRAF, SMAD4, IDH1, Eighteen pairs of formalin fixed (ID #1-#18), paraffin em- IDH2, NRAS, ARID1A, PTEN, CDKN2A, CDK6, ERBB3, bedded (FFPE) primary and recurrent ICC tumors were MET, BRCA1, BRCA2, NF1, PTCH1,and TSC, with variable collected from 5 different Italian centers. The independent percentages due to the heterogeneity of the case stud- cohort is constituted by 13 fresh ICC tumors (ID#19-#31), ies [12, 15–17]. These data have been studied to plan 10 PRs (named CHC001-PR to CHC024-PR), 3 RECs tu- clinical trials aimed at inhibiting specific targets, alone or mors (CHC002-REC, CHC012-REC, and CHC017-REC) in combination with standard chemotherapy. However, and 7 ascites samples (ID #33-#38), named PARA-2 to the obtained results are modest and not of impact. PARA-11, (where cancer cells were isolated from ascites A key role in tumorigenesis seems to be played by mu- liquid obtained by paracentesis procedure), obtained from tant IDH1. IDH1 mutation causes an impaired produc- different patients. This cohort was analyzed separately for tion of α-KG in favour of the oncometabolite 2-HG [18]; gene expression profiling and mutational analysis of in particular, it acts as a competitor of α-KG, causing a IDH1, assuming that PARAs are progressive disease and hypermethylation of histones and of DNA and promot- consequently they could be assimilated to recurrences. ing epigenetic alterations, all phenomena typically found Additional file 1: Table S1 summarizes patient clinical and during progression and metastatic processes. The role of pathological characteristics, and the analyses performed. IDH1 as prognostic marker is controversial; literature The median age of patients is 65, ranging from 41 to 84; data demonstrated that IDH1 mutations correlated with 24 females and 14 males were analyzed. The two inde- good prognosis in brain tumors, such as glioma, glio- pendent cohorts were homogeneous in terms of gender blastoma and anaplastic astrocytoma [19]. On the con- (Fisher’s Exact test p-value = 0.7), age at diagnosis and trary, in acute myeloid leukemia (AML) and in ICC it time to recurrence (Student T-test p-values = 0.5 and 0.2, seems that the presence of mutations did not affect the respectively). overall survival (OS) and progression free survival (PFS) [12, 20]. It has been demonstrated that ICC patients are Nucleic acids extraction and quality control frequently mutated (about 25%) in IDH1 hot-spots [21]. Total RNA was extracted from FFPE tissues using the A recent work of Saha and collaborators demonstrated miRNeasy FFPE mini kit, following the manufacturer’s Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 3 of 10 instructions. Briefly, RNA quantity was evaluated by using the Benjamini-Hochberg correction [28]. TMev Nanodrop, while the quality was assessed by qRT-PCR (http://mev.tm4.org) and hclust R function were used to testing the Ct of two different amplicons of ACTB. Only perform hierarchical clustering of genes/samples using for 11 couples, the RNA had an acceptable quality to either selected genes or the global gene expression pro- perform further experiments. filing and to carry out principal component analysis DNA was extracted using Qiamp DNA FFPE kit, fol- (PCA). MetaCore version 6.29 (Thomson Reuters) was lowing manufacturer’s instructions. Briefly, tumor slides used for network and pathway maps analysis. GSEA was were stained with Hematoxylin and Eosin: tumor areas used to evaluate significant enrichment in predefined cu- were circled by a pathologist. Representative images of rated sets of genes from online pathway databases and H/E staining are shown in Additional file 2: Figure S1. publications in PubMed [29]. The weighted voting algo- Representative tumor areas were scraped, deparaffi- rithm and leave-one-out cross validation available within nized by xylene, rehydrated, subsequently treated with the SET tool (Signature Evaluation Tool) were used to proteinase K and then purified using columns. Total evaluate the discrimination power of our expression RNA of fresh frozen tissues and tumor cells obtained by signature on the validation set [30]. The SET algorithm paracentesis was extracted by Absolutely RNA miRNA allows re-evaluation and re-adjustment of the discrimin- kit (Agilent Technologies), while DNA was extracted by ation power of a given signature by selecting/de-select- QiAmp DNA mini kit (Qiagen), following manufac- ing genes repeatedly. Microarray data were deposited in turer’s protocols. Gene Expression Omnibus (GSE107102). cDNA mediated annealing, selection, extension and External dataset ligation (DASL) assay cBioportal was used to download expression profiles of Total RNA extracted from FFPE samples was retrotran- selected genes in the cholongiocarcinoma TCGA dataset scribed to cDNA using oligo-dT and random nanomer (33 samples), together with clinical information about primers, byotinilated and bound to streptavidin particles. survival [31, 32]. mRNA Expression z-Scores (RNA Seq The reactions of labeling with the single color Cy3, V2 RSEM) of the selected genes were used to stratify pa- denaturation, and hybridization on Illumina Human tients in two groups according to expression medians. R Reference 8 BeadArrays were conducted according to survival package was applied to run survival analysis and the manufacturer’s instruction. The slides were washed generate Kaplan-Meier curves. and scanned using the Illumina BeadArray Reader (Illumina, San Diego, CA). Mutational analysis Quality control and quantification of extracted DNA were Gene expression analysis (GEP) by Agilent platform conducted by Bioanalyzer and Qubit, respectively. IDH1 For GEP analysis, Low Input Quick Amp Labeling Kit, exon 4 was amplified by nested PCR with relative specific one-color kit (Agilent Technologies) was used to amplify primers (IDH1 external primers: Forward 5’-TGAGCTCTA and label 100 ng of total RNA. Six hundred ng were TATGCCATCACTGCA-3′, Reverse 5’-CAATTTCAT hybridized on SurePrint G3 Human Gene Expression ACCTTGCTTAATGGG-3′; IDH1 internal primers: For- 8x60K v2 glass arrays. After arrays scansion, images were ward 5’-GCAGTTGTAGGTTATAACTATCC-3′;Reverse analyzed by the Feature Extraction Software from 5’-TGGGTGTAGATACCAAAAG-3′). The PCR products Agilent Technologies (version 10.7); raw data were then were purified using Wizard®SV Gel and PCR Clean-Up processed using the LIMMA (LInear Models for Micro- System (Promega, Milan, Italy) and sense and antisense array Analysis) package from Bioconductor [26]. sequences were obtained by using internal forward and re- verse primers, respectively. Sequencing was performed by Microarray data analysis BigDye Terminator Cycle sequence following the PE Ap- Raw data intensities were loaded into R statistical envir- plied Biosystem strategy and Applied Biosystems ABI onment. The normexp method was used for background PRISM3100 DNA Sequencer (Applied Biosystem, Forster correction with an offset of 50 and the quantile method City,CA).All mutationswere confirmed by two independ- for normalization. To remove batch effect between sub- ent PCR experiments. sets of experiments, the combat function was applied to the dataset [27]. Next generation sequencing (NGS) LIMMA was then used to identify differentially The Ion Torrent S5 platform was used to perform the expressed genes in recurrent vs primary ICC samples, NGS analysis of the Ampliseq CHPv2 which contains using a paired statistics for paired samples of the first the hot-spot mutations of 50 cancer related genes. tested cohort or unpaired statistics for the independent Briefly, DNA of patient #3, both PR and REC, was used one [26]; p-values were adjusted for multiple testing by to prepared libraries by Ion AmpliSeq Library kit 2.0 Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 4 of 10 (ThermoFisher Scientific Waltham, MA) according to these tumors. In particular, four PRs belonged to a dif- the manufacturer’s protocol. Libraries were then quanti- ferent branch arm. Principal component analysis (PCA) fied using Qubit (ThermoFisher Scientific Waltham, was able to clearly separate primary from recurrent sam- MA; 60 pM of each sample were run and sequenced for ples, as shown in Fig. 1b. Then, we refined the analysis, 2800 hot-spots of 50 among oncogenes and tumor sup- applying a two-class paired comparison and filtering pressors. Raw data were analyzed by Ion Reporter soft- data with a cut-off on logFC > 1 or < − 1 and an adjusted ware (ThermoFisher Scientific Waltham, MA) and p-value < 0.01, obtaining 315 significant deregulated filtered in p-value < 0.001 and coverage of > 400 were genes, of which 65 down- and 250 up-regulated in RECs accomplished. versus PRs (Additional file 3: Table S2). Figure 1c shows the heatmap of this genes signature. Results Using Metacore software, we performed an enrich- Transcriptomic analysis of paired PRs and RECs ICC ment analysis of pathway maps and process networks. To analyze sample distributions according to their tran- Additional file 4: Table S3 and Additional file 5: Table S4 scriptomic profiles, unsupervised hierarchical clustering describe the first 15 pathways and networks enriched in was applied to the global normalized intensity profiles. up-regulated genes. They were associated to Epithelial to As shown in Fig. 1a, PRs and RECs tumors had different Mesenchymal Transition (EMT) mediated by Rho alpha, distributions; for some patients, PRs and RECs had simi- PI3K and ILK mediated by TGF beta, cytoskeleton re- lar profiles, for others the RECs were very distant from modeling by GTPase, anti-apoptotic process mediated their PR tumors, underlining the high heterogeneity of by BAD phosphorylation, and in general cell cycle, Fig. 1 a Dendrogram obtained from unsupervised hierarchical clustering using Euclidean distance as similarity metrics and ward as linkage method. Each branch of the dendrogram is represented by the global gene expression profile of samples. The vertical axis indicates the Euclidean distance between samples/clusters. PR and REC of the same patient are represented with the same color. b Projection of principal components 1 and 3 after application of PCA on the paired samples cohort. Red squares correspond to primary samples while yellow squares to recurrences. c Unsupervised hierarchical clustering analysis of 315 significantly deregulated genes in REC vs PR tumors. A red-to-green gradient was used to indicate, for each gene, levels of up- or downregulation. The logFC values of the entire matrix used for hierarchical clustering are provided as Additional file 3:Table S2 Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 5 of 10 cytoskeleton remodeling and EMT networks. On the Further, we applied the 24 genes signature on the val- contrary, down-regulated genes affected in particular, idation dataset and we demonstrated that the expression PTEN signal transduction, immune response, apoptosis of 9 genes (NDST2, DAO, FANCG, CARD9, FOXJ2, mediated by p53, proteolysis related to cell cycle and SEMA3B, GDAP1L1, TRIOBP, PFKM) is able to distin- apoptosis, inflammation mediated by IL-6 signaling guish PRs from RECs (Fig. 3), with an error rate of 0.15 (Additional file 6: Table S5 and Additional file 7:Table S6). and p < 0.001. The weights of individual genes are re- The application of Gene Set Enrichment Analysis on ported in Additional file 11: Table S8. the complete list of pre-ranked genes was not very con- Finally, the expression profiles of the 9 genes and the clusive. However, pre-ranked GSEA on the 315 genes survival information about patients were downloaded signature showed a high number of overlapping genes from the cholangiocarcinoma TCGA dataset (n = 33) with the “Liver_cancer_UP” geneset described by Acevedo available through cBioportal (http://www.cbioportal.org). and collaborators (Enrichment score 0.38, p =0.009; Among the genes overexpressed in patients with worst Fisher’s exact test p-value = 0.002; Jaccard Index = 0.022) prognosis, the one mostly associated with survival is as shown in Additional file 8:Figure S2 [29]. FANCG, with Cox proportional hazard ratio = 3.242 (Fig. 4). Further, restricting the analysis with more stringent filters (logFC <−2or>2 and p < 0.01) we obtained a IDH1 is a potential marker of tumor progression gene signature of 24 deregulated genes, 10 down- and For 16 paired PR and REC ICC tumors, the mutational 14 up-regulated, able to separate RECs from PRs tu- analysis of IDH1 exon 4 was conducted. As shown in mors (Fig. 2). Additional file 12: Table S9, five patients (#3, #6, #11, #15 These genes could be grouped in four main functional and #18) harbored an IDH1 R132x mutation in the REC classes: regulation of apoptosis, (FOXJ2, SEMA3B, counterpart (31.3%); only in patient #6 (6.25%) the muta- CARD9), cellular migration and motility (CLDN23, tion was already present in PR. Figure 5 shows electrophe- TRIOBP), DNA-RNA processing (FANCG, UBLCP1), rograms of mutated samples compared to WT one. metabolic processes such as glycolysis and fatty acids In order to confirm the naïve IDH1 mutation identified metabolism (PECI, PFKM, NDST2, DAO). The transcripts by Sanger sequencing in REC tumor, patient #3 was ana- of this signature were investigated on an independent co- lyzed by NGS using the AmpliSeq technology on Ion Tor- hort of patients (10 RECs vs 11 PRs, IDs #19-#38, see rent device. In patient #3, the IDH1 mutation in codon 132 Additional file 1: Table S1 and Additional file 9: Table S7); was confirmed in REC tumor, with a 17.37% of frequency as shown in Additional file 10: Figure S3, expression fold (p = 0.0001) and PR counterpart resulted WT. Overall, the changes of 17 out of 24 genes were concordant in the two mutational profile is partially overlapping between PR and cohorts, with 8 genes differentially expressed in a statisti- REC, even if a higher number of missense mutations was cally significant manner in the independent cohort as well. identified in PR. Additional file 13: Table S10 summarizes Fig. 2 Unsupervised clustering analysis of 24 significantly deregulated genes in 13 REC tumors compared to 11 PR tumors. Tmev software was used, with Euclidean distance as similarity metrics and complete linkage as linkage method. Log Intensities of each gene were standardized by median centering and dividing by standard deviation. Red/green rectangles indicate expression higher/lower than the median, respectively Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 6 of 10 Fig. 3 Predictive role of 9 out 24 genes of the signature. The expression of these genes is able to clearly separate PRs (square) and RECs (circle) in the validation cohort of patients. Signal to noise scores provided by SET are shown for each gene in Additional file 11: Table S8 the mutational pattern (exonic missense and synonymous, from their primary tumors (PR) by the transcriptomic intronic) of patient #3. Among them, KDR, APC, RB1, enrichment in genes involved in proliferation, motility TP53, IDH1 are already described [33]. and migration, apoptosis resistance, and epithelial-to mesenchymal transition. Further, mutated IDH1 hotspot Discussion (codon 132) emerged in about 30% of REC tumors sug- In this work, we demonstrated that recurrent intrahepa- gesting that it could be a putative marker of progression. tic cholangiocarcinomas (REC ICC) are distinguishable Results obtained by the comparison of matched REC/PR ICC suggest that recurrent lesions could be molecularly different from their primary tumors due to a clonal se- lection toward drug resistant and more malignant tumor cells in RECs. Mutational analysis by Sanger revealed that 4 patients harbored IDH1 R132x mutations in RECs, but not in the PRs; NGS analysis performed in a PR-REC pair showed that IDH1 mutation was gained in REC, confirming our direct sequencing results. How- ever, other missense mutations found in PR were lost, suggesting that a clonal sieving occurred in REC. As a matter of fact, Sanger sequencing has limitations; small DNA fragments could be analyzed with a single reaction and some mutations may re- sult undetectable, due to the low clonal representa- tion. On the contrary, NGS is more sensitive and covers broad range spectra of mutations with a low amount of DNA and should be preferred as a screening method. However, our data supports the heterogeneous nature of this tumor type, both intra- and inter-patients. Many factors, including treatment, Fig. 4 Kaplan-Meier curves for 33 patients from the TCGA could concur in the evolution of the tumor. For the cholangiocarcinoma external dataset, with survival information available. Patients are divided in two groups according to FANCG choice of second-line treatments it should be consid- expression. Red curve: FANCG expression higher than the median. ered that responsive tumor clones are inhibited by Black curve: FANCG expression lower than the median. Log-rank test previous therapy and concurrently disease progression is p-value = 0.0544. Cox proportional hazard ratio = 3.242 sustained by chemotherapy-resistant clones. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 7 of 10 Fig. 5 Representative electropherograms of mutated samples in the hot-spot codon 132 of IDH1 According to microarray data, PRs and RECs of the mesenchymal markers and reduction of epithelial same patient could have very similar or distinct profiles. markers, are crucial events during tumor progression to- This finding is enhanced in particular in those patients ward a more aggressive, proliferating and drug resistant for whom we had two RECs; in one case, the two RECs phenotypes [38, 39]. We showed that a 24 genes signa- are comparable and the expression profile is far from ture is able to distinguish RECs from PRs in the main their PR, but for the other patient, one REC is near to cohort of analysis. The trend of these genes is partially PR while the other one displays a different pattern of ex- confirmed on an independent cohort of patients consti- pression. Further, the complexity of these tumors is tuted by fresh frozen PR or REC ICC tumors and tumor highlighted not only by the heterogeneity identified be- cells obtained from ascites liquid of ICC patients in pro- tween RECs and their PRs, but also within different bi- gressive disease (PARAs). The two cohorts are small, opsies of the same lesion [34]. Globally, we found an which could limit the robustness of our results, but they increased expression of genes involved in nucleic acids are homogeneous in terms of baseline characteristics processing, transcription of RNA and non-coding RNA, (sex, age at diagnosis) and time to recurrence. They are and angiogenesis regulation, with a concomitant mainly composed of T2 stage tumors, with ascites in- down-regulation of genes related to epithelial cell differ- cluded in the validation cohort only, which might be po- entiation. The down-regulation of epithelial cell differen- tentially confounding. However, in the latter cohort the tiation genes suggested that the epithelial-like phenotype expression of 9 of the signature genes is able to discrim- is switched towards a mesenchymal-like. This data is inate RECs from PRs in a statistically significant manner. confirmed by pathways and maps analyses, with an en- Moreover, these molecules might represent novel richment of up-regulated genes involved in EMT, in par- therapeutic targets. Namely, CARD9 is a marker of ticular induced by TGF-beta. In agreement with several tumor progression and poor prognosis in hepatocarci- evidences in other types of cancer [35–37], these data noma (HCC), B cell lymphoma, and clear cell renal confirmed that EMT is one of the crucial steps of recur- carcinoma [40–42] and promotes metastatization acti- rence and drug resistance. We also described an vating metastasis-associated macrophages (MAM) [43]. up-regulation of genes involved in cytoskeleton remodel- Besides, the higher expression of FANCG suggested that ing and cell cycle regulation in REC tumors. Many stud- REC tumors displayed increased DNA damage and ies described the close interconnection among these activated DNA repair. The downregulation of FANCG events; the remodeling of extracellular matrix and was associated with effective treatment with GEM and reorganization of cytoskeleton, along with expression of radiolabeled Trastuzumab in tumor xenograft model of Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 8 of 10 disseminated intraperitoneal disease (46). Moreover, we an independent cohort of patients. The same trend was found in 17 out found an up-regulation of PIK4CA which is involved in of 24 genes. * indicates statistically significant results. Y axis: log fold change expression obtained in the two independent cohorts. (TIF 162 kb) proliferation and chemoresistance in other tumors, such Additional file 11: Table S8. Weights of the 9 genes signature. (XLSX 9 kb) as medulloblastoma [44]. The same trend of expression Additional file 12: Table S9. Sanger sequencing of IDH1 exon 4. (DOCX was revealed for TRIOBP, already described in pancreatic 15 kb) cancer where is involved in cell motility and migration Additional file 13: Table S10. NGS sequencing of patient #3. (XLSX through cytoskeleton remodeling [45, 46]. Interest- 15 kb) ingly, the emergence of IDH1 in about 30% of REC patients suggested that it could be considered not Abbreviations only a marker of progression but also a potential tar- 2-HG: 2-hydroxyglutarate; AML: Acute myeloid leukemia; DASL: cDNA mediated annealing, selection, extension and ligation; EMT: Epithelial to get for tailored therapy. In fact, it has been demon- mesenchymal transition; FDR: False discovery rate; FFPE: Formalin fixed strated that IDH1 mutation promoted sensitivity to paraffin embedded; GEM: Gemcitabine; GEP: Gene expression analysis; the multitarget inhibitor Dasatinib [24]. Moreover, GSEA: Gene set enrichment analysis; HBV: Hepatitis B virus; HCV: Hepatitis C virus; ICC: Intrahepatic cholangiocarcinoma; LIMMA: Linear models for mutated-IDH1 targeting agents are now under clinical microarray analysis; MAM: Metastasis-associated macrophages; NGS: Next investigation in different solid tumors, including ICC. generation sequencing; PFS: Progression free survival; PR: Primary tumor; As an example, BAY1436032 targeting the hot-spot PSC: Primary sclerosing cholangitis; REC: Recurrent tumor; SET: Signature evaluation tool; TCGA: The Cancer Genome Atlas; WT: Wild-type; α-KG: Alpha mutation R132x is now tested in patients with ad- ketoglutarate vanced solid tumors enrolled in one open-label, non-randomized, multicenter phase I-II clinical trial Acknowledgements (NCT02746081). Of main interest, the effect of A special acknowledgement is given to an Italian association for the study of ICC, created in memory of Annalisa Pozzi. AG-120 is now compared to placebo in the phase III, multicenter, randomized double-blind ClarIDHy trial Funding on non-resectable/metastatic cholangiocarcinoma. “Associazione Italiana Ricerca sul Cancro–AIRC 5X1000 2010-Ministry of Health, FPO”. Project n°16:30 “Identificazione di nuove vie di trasduzione del segnale intra- cellulare sensibili ai farmaci nel colangiocarcinoma intraepatico (ICC)”.Fondazione Conclusions Piemontese per la Ricerca sul Cancro (FPRC) “Identification of new druggable In conclusion, our data on transcriptomic and mutational pathways in intrahepatic cholangiocarcinoma” 5 per Mille 2010 Ministero della status of REC ICC suggest that a personalized approach, Salute. Università di Torino anno 2014 –: Fondo per la ricerca locale (Linea B), LEOF_RIC_LOC_14_01 project title: “Transcriptomic and genetic analysis of in which tumor molecular/genetic characterization are paired primary and recurrent intrahepatic cholangiocarcinoma”.CPN andDS: followed from diagnosis to disease progression, is advis- FPRC 5 × 1000 Ministero della Salute 2012. GC is supported by a grant of Com- able not only for prognostic purposes, but also to identify pagnia di San Paolo; AS is supported by Associazione Italiana per la Ricerca sul Cancro grant AIRC 5 × 1000 n. 12182. the emergence of other druggable targets after first-line treatment failure. Availability of data and materials Microarray data were deposited in Gene Expression Omnibus (GSE107102). Additional files Authors’ contributions All authors read and approved the final version of the manuscript. CPN Additional file 1: Table S1. Clinical pathological characteristics of ICC conceived the work, performed the experiments an wrote the article draft; patients. (DOCX 19 kb) PO performed microarray data analysis; GC performed validation experiments and wrote the article draft; YP performed NGS analyses and wrote the article Additional file 2: Figure S1. Representative images of H/E staining of draft, DS revised the manuscript and performed statistical analyses, LDC and PRs (A and C) and their RECs counterparts (B and D). (TIF 793 kb) EM performed DASL analysis, DR, AS, AMDR, AG, FC, CR, PI recruited patients Additional file 3: Table S2. List of significant deregulated genes in and biological material for this work and revised the manuscript; MA and FL RECs versus PRs. (XLSX 39 kb) conceived and supervised the work and revised the manuscript; GC conceived Additional file 4: Table S3. Pathway maps obtained with Metacore the work, performed data analysis and revised the manuscript. analyzing only up-regulated genes. (DOCX 16 kb) Additional file 5: Table S4. Process networks obtained with Metacore Ethics approval and consent to participate analyzing only up-regulated genes. (DOCX 15 kb) Biological material was obtained, in accordance with the Declaration of Helsinki from patients who have signed the informed consent. The study Additional file 6: Table S5. Pathway maps obtained with Metacore was approved by each local institutional review boards (Comitato Etico analyzing only down-regulated genes. (DOCX 15 kb) “IRCCS Candiolo”; “Comitato Etico AORN Cardarelli-Santobono”;Comitato Additional file 7: Table S6. Process networks obtained with Metacore Etico “Policlinico Gemelli”; “Comitato Etico per la sperimentazione clinica analyzing only down-regulated genes. (DOCX 15 kb) delleprovincediVerona eRovigo”; “Comitato Etico indipendente, IRCSS Additional file 8: Figure S2. The GSEA dataset of Acevedo et al., [47] Istituto Clinico Humanitas”. The coordinator center approved the entire on liver cancer was found enriched for up-regulated genes (Enrichment study according to the PROFILING protocol (“Studio prospettico per la score 0.38; p = 0.009, pre-ranked GSEA analysis). (TIF 174 kb) determinazione del profilo molecolare di resistenza alle terapie target in pazienti con malattie neoplastiche”, 001-IRCC-00IIS-10, version 6.1, FPO-IRCCS, Additional file 9: Table S7. List of deregulated genes of the l’Istituto di Ricovero e Cura a Carattere Scientifico Candiolo (TO)). indepentent cohort of patients. (XLSX 95 kb) Additional file 10: Figure S3. Comparison between the expression Competing interests values of selected genes obtained by DASL array and GEP performed on The authors declare that they have no competing interests. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 9 of 10 Publisher’sNote 14. Andersen JB, Spee B, Blechacz BR, Avital I, Komuta M, Barbour A, Conner EA, Springer Nature remains neutral with regard to jurisdictional claims in Gillen MC, Roskams T, Roberts LR, et al. Genomic and genetic characterization published maps and institutional affiliations. of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology. 2012;142:1021–31. e1015 Author details 15. Chong DQ, Zhu AX. The landscape of targeted therapies for cholangiocarcinoma: Medical Oncology Division, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. current status and emerging targets. Oncotarget. 2016;19:46750–767. 142, km 3.95, 10060 Candiolo, Turin, Italy. Cancer Genomics Lab, Fondazione 16. Nakamura H, Arai Y, Totoki Y, Shirota T, Elzawahry A, Kato M, Hama N, Edo ed Elvo Tempia Valenta, Biella, Italy. Department of Oncology, University Hosoda F, Urushidate T, Ohashi S, et al. Genomic spectra of biliary tract of Turin, Torino, Italy. Functional Genomics and Bioinformatics, Department cancer. Nat Genet. 2015;47:1003–10. of Applied Research and Technology Development, Fondazione IRCCS 17. Simbolo M, Fassan M, Ruzzenente A, Mafficini A, Wood LD, Corbo V, Melisi Istituto Nazionale dei Tumori, Milan, Italy. Division of Hepatobilio-Pancreatic D, Malleo G, Vicentini C, Malpeli G, et al. Multigene mutational profiling of and Colorectal Surgery, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. 142, cholangiocarcinomas identifies actionable molecular subgroups. Oncotarget. km 3.95, Candiolo, Italy. ARC-Net Research Centre and Department of 2014;5:2839–52. Diagnostics and Public Health – Section of Pathology, University and 18. Garber K. Oncometabolite? IDH1 discoveries raise possibility of new metabolism Hospital Trust of Verona, Verona, Italy. Hepatobiliary Surgery Unit, Gemelli targets in brain cancers and leukemia. J Natl Cancer Inst. 2010;102:926–8. Hospital, Catholic University of the Sacred Heart, Rome, Italy. Department of 19. Sturm D, Witt H, Hovestadt V, Khuong-Quang DA, Jones DT, Konermann C, Health’s Sciences and Medicine “V. Tiberio”, University of Molise, Pfaff E, Tönjes M, Sill M, Bender S, et al. Hotspot mutations in H3F3A and Campobasso, Italy. Hepatobiliary and Liver Transplant Unit, Cardarelli IDH1 define distinct epigenetic and biological subgroups of glioblastoma. Hospital, Naples, Italy. Center for Autoimmune Liver Diseases, Humanitas Cancer Cell. 2012;22:425–37. Clinical and Research Center, Rozzano, Italy. Department of Experimental 20. Goyal L, Govindan A, Sheth RA, Nardi V, Blaszkowsky LS, Faris JE, Clark JW, and Clinical Medicine, University of Firenze, Florence, Italy. UOC di Ryan DP, Kwak EL, Allen JN, et al. Prognosis and Clinicopathologic features Gastroenterologia, Azienda Ospedaliera San Gerardo, Monza, Italy. of patients with advanced stage Isocitrate dehydrogenase (IDH) mutant and IDH wild-type intrahepatic cholangiocarcinoma. Oncologist. 2015;20:1019–27. Received: 20 November 2017 Accepted: 25 May 2018 21. Borger DR, Tanabe KK, Fan KC, Lopez HU, Fantin VR, Straley KS, Schenkein DP, Hezel AF, Ancukiewicz M, Liebman HM, et al. Frequent mutation of isocitrate dehydrogenase (IDH)1 and IDH2 in cholangiocarcinoma identified through broad-based tumor genotyping. Oncologist. 2012;17:72–9. References 22. Saha SK, Parachoniak CA, Bardeesy N. IDH mutations in liver cell plasticity 1. Bragazzi M, Cardinale V, Carpino G, Venere R, Semeraro R, Gentile R, Gaudio and biliary cancer. Cell Cycle. 2014;13:3176–82. E, Alvaro D. Cholangiocarcinoma: epidemiology and risk factors. Transl 23. Saha SK, Parachoniak CA, Ghanta KS, Fitamant J, Ross KN, Najem MS, Gastrointest Cancer. 2012;1:21–32. Gurumurthy S, Akbay EA, Sia D, Cornella H, et al. Mutant IDH inhibits 2. Alvaro D, Crocetti E, Ferretti S, Bragazzi MC, Capocaccia R, Committee AC. HNF-4α to block hepatocyte differentiation and promote biliary cancer. Descriptive epidemiology of cholangiocarcinoma in Italy. Dig Liver Dis. Nature. 2014;513:110–4. 2010;42:490–5. 24. Saha SK, Gordan JD, Kleinstiver BP, Vu P, Najem MS, Yeo JC, Shi L, Kato Y, Levin 3. Kim HJ, Kim JS, Joo MK, Lee BJ, Kim JH, Yeon JE, Park JJ, Byun KS, Bak YT. RS, Webber JT, et al. Isocitrate dehydrogenase mutations confer Dasatinib Hepatolithiasis and intrahepatic cholangiocarcinoma: a review. World J hypersensitivity and SRC dependence in intrahepatic cholangiocarcinoma. Gastroenterol. 2015;21:13418–31. Cancer Discov. 2016;6:727–39. 4. Chaiteerakij R, Juran BD, Aboelsoud MM, Harmsen WS, Moser CD, Giama 25. Ercolani G, Vetrone G, Grazi GL, Aramaki O, Cescon M, Ravaioli M, Serra C, NH, Allotey LK, Mettler TA, Baichoo E, Zhang X, et al. Association between Brandi G, Pinna AD. Intrahepatic cholangiocarcinoma: primary liver resection variants in inflammation and cancer-associated genes and risk and survival and aggressive multimodal treatment of recurrence significantly prolong of cholangiocarcinoma. Cancer Med. 2015;4:1599–602. survival. Ann Surg. 2010;252:107–14. 5. Tao LY, He XD, Xiu DR. Hepatitis B virus is associated with the clinical features 26. Smyth GK. Linear models and empirical bayes methods for assessing and survival rate of patients with intrahepatic cholangiocarcinoma. Clin Res differential expression in microarray experiments. Stat Appl Genet Mol Hepatol Gastroenterol. 2016;40:682–87. Biol. 2004;3:Article3. 6. Wang Z, Sheng YY, Dong QZ, Qin LX. Hepatitis B virus and hepatitis C virus 27. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression play different prognostic roles in intrahepatic cholangiocarcinoma: a meta- data using empirical Bayes methods. Biostatistics. 2007;8:118–27. analysis. World J Gastroenterol. 2016;22:3038–51. 28. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and 7. Ulahannan SV, Rahma OE, Duffy AG, Makarova-Rusher OV, Kurtoglu M, powerful approach to multiple testing. Royal Statistical Society. 1995;57:12. Liewehr DJ, Steinberg SM, Greten TF. Identification of active chemotherapy 29. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, regimens in advanced biliary tract carcinoma: a review of chemotherapy trials Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set in the past two decades. Hepat Oncol. 2015;2:39–50. enrichment analysis: a knowledge-based approach for interpreting genome- 8. Yang R, Wang B, Chen YJ, Li HB, Hu JB, Zou SQ. Efficacy of gemcitabine plus wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–50. platinum agents for biliary tract cancers: a meta-analysis. Anti-Cancer Drugs. 30. Jen CH, Yang TP, Tung CY, Su SH, Lin CH, Hsu MT, Wang HW. Signature 2013;24:871–7. evaluation tool (SET): a java-based tool to evaluate and visualize the sample 9. Park HM, Yun SP, Lee EC, Lee SD, Han SS, Kim SH, Park SJ. Outcomes for discrimination abilities of gene expression signatures. BMC Bioinformatics. patients with recurrent intrahepatic cholangiocarcinoma after surgery. Ann 2008;9:58. Surg Oncol. 2016;23:4392–400. 31. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, Sun Y, 10. Yoshikawa D, Ojima H, Iwasaki M, Hiraoka N, Kosuge T, Kasai S, Hirohashi S, Jacobsen A, Sinha R, Larsson E, et al. Integrative analysis of complex Shibata T. Clinicopathological and prognostic significance of EGFR, VEGF, cancer genomics and clinical profiles using the cBioPortal. Sci Signal. and HER2 expression in cholangiocarcinoma. Br J Cancer. 2008;98:418–25. 2013;6:pl1. 11. Nakazawa K, Dobashi Y, Suzuki S, Fujii H, Takeda Y, Ooi A. Amplification and 32. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, Jacobsen A, overexpression of c-erbB-2, epidermal growth factor receptor, and c-met in Byrne CJ, Heuer ML, Larsson E, et al. The cBio cancer genomics portal: an biliary tract cancers. J Pathol. 2005;206:356–65. open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401–4. 12. Zhu AX, Borger DR, Kim Y, Cosgrove D, Ejaz A, Alexandrescu S, Groeschl RT, Deshpande V, Lindberg JM, Ferrone C, et al. Genomic profiling of intrahepatic 33. Lee CH, Wang HE, Seo SY, Kim SH, Kim IH, Kim SW, Lee ST, Kim DG, Han MK, cholangiocarcinoma: refining prognosis and identifying therapeutic targets. Lee SO. Cancer related gene alterations can be detected with next-generation Ann Surg Oncol. 2014;21:3827–34. sequencing analysis of bile in diffusely infiltrating type cholangiocarcinoma. 13. Sia D, Hoshida Y, Villanueva A, Roayaie S, Ferrer J, Tabak B, Peix J, Sole M, Exp Mol Pathol. 2016;101:150–6. Tovar V, Alsinet C, et al. Integrative molecular analysis of intrahepatic 34. Brandi G, Farioli A, Astolfi A, Biasco G, Tavolari S. Genetic heterogeneity in cholangiocarcinoma reveals 2 classes that have different outcomes. cholangiocarcinoma: a major challenge for targeted therapies. Oncotarget. Gastroenterology. 2013;144:829–40. 2015;6:14744–53. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 10 of 10 35. Iwahashi S, Shimada M, Utsunomiya T, Imura S, Morine Y, Ikemoto T, Takasu C, Saito Y, Yamada S. Epithelial-mesenchymal transition-related genes are linked to aggressive local recurrence of hepatocellular carcinoma after radiofrequency ablation. Cancer Lett. 2016;375:47–50. 36. Suresh R, Ali S, Ahmad A, Philip PA, Sarkar FH. The role of Cancer stem cells in recurrent and drug-resistant lung Cancer. Adv Exp Med Biol. 2016;890:57–74. 37. Kubelt C, Hattermann K, Sebens S, Mehdorn HM, Held-Feindt J. Epithelial-to- mesenchymal transition in paired human primary and recurrent glioblastomas. Int J Oncol. 2015;46:2515–25. 38. Chou YS, Yang MH. Epithelial-mesenchymal transition-related factors in solid tumor and hematological malignancy. J Chin Med Assoc. 2015;78:438–45. 39. Savagner P. Epithelial-mesenchymal transitions: from cell plasticity to concept elasticity. Curr Top Dev Biol. 2015;112:273–300. 40. Zekri AR, El-Kassas M, Saad Y, Bahnassy A, El-Din HK, Darweesh SK, Abdel Hafez H, Esmat G. Caspase recruitment domains. New potential markers for diagnosis of hepatocellular carcinoma associated with HCV in Egyptian patients. Ann Hepatol. 2013;12:774–81. 41. Nakamura S, Matsumoto T, Yada S, Hirahashi M, Suekane H, Yao T, Goda K, Iida M. Overexpression of caspase recruitment domain (CARD) membrane- associated guanylate kinase 1 (CARMA1) and CARD9 in primary gastric B- cell lymphoma. Cancer. 2005;104:1885–93. 42. Tan W, Hildebrandt MA, Pu X, Huang M, Lin J, Matin SF, Tamboli P, Wood CG, Wu X. Role of inflammatory related gene expression in clear cell renal cell carcinoma development and clinical outcomes. J Urol. 2011;186:2071–7. 43. Yang M, Shao JH, Miao YJ, Cui W, Qi YF, Han JH, Lin X, Du J. Tumor cell- activated CARD9 signaling contributes to metastasis-associated macrophage polarization. Cell Death Differ. 2014;21:1290–302. 44. Guerreiro AS, Fattet S, Kulesza DW, Atamer A, Elsing AN, Shalaby T, Jackson SP, Schoenwaelder SM, Grotzer MA, Delattre O, Arcaro A. A sensitized RNA interference screen identifies a novel role for the PI3K p110γ isoform in medulloblastoma cell proliferation and chemoresistance. Mol Cancer Res. 2011;9:925–35. 45. Bao J, Wang S, Gunther LK, Kitajiri S, Li C, Sakamoto T. The actin-bundling protein TRIOBP-4 and -5 promotes the motility of pancreatic cancer cells. Cancer Lett. 2015;356:367–73. 46. Seipel K, O'Brien SP, Iannotti E, Medley QG, Streuli M. Tara, a novel F-actin binding protein, associates with the trio guanine nucleotide exchange factor and regulates actin cytoskeletal organization. J Cell Sci. 2001;114:389–99. 47. Acevedo LG, Bieda M, Green R, Farnham PJ. Analysis of the mechanisms mediating tumor-specific changes in gene expression in human liver tumors. Cancer Res. 2008;68:2641–51. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Genomics Springer Journals

Transcriptomic analysis and mutational status of IDH1 in paired primary-recurrent intrahepatic cholangiocarcinoma

Free
10 pages

Loading next page...
 
/lp/springer_journal/transcriptomic-analysis-and-mutational-status-of-idh1-in-paired-9Rx02j3gbi
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Life Sciences; Life Sciences, general; Microarrays; Proteomics; Animal Genetics and Genomics; Microbial Genetics and Genomics; Plant Genetics and Genomics
eISSN
1471-2164
D.O.I.
10.1186/s12864-018-4829-0
Publisher site
See Article on Publisher Site

Abstract

Background: Effective target therapies for intrahepatic cholangiocarcinoma (ICC) have not been identified so far. One of the reasons may be the genetic evolution from primary (PR) to recurrent (REC) tumors. We aim to identify peculiar characteristics and to select potential targets specific for recurrent tumors. Eighteen ICC paired PR and REC tumors were collected from 5 Italian Centers. Eleven pairs were analyzed for gene expression profiling and 16 for mutational status of IDH1. For one pair, deep mutational analysis by Next Generation Sequencing was also carried out. An independent cohort of patients was used for validation. Results: Two class-paired comparison yielded 315 differentially expressed genes between REC and PR tumors. Up-regulated genes in RECs are involved in RNA/DNA processing, cell cycle, epithelial to mesenchymal transition (EMT), resistance to apoptosis, and cytoskeleton remodeling. Down-regulated genes participate to epithelial cell differentiation, proteolysis, apoptotic, immune response, and inflammatory processes. A 24 gene signature is able to discriminate RECs from PRs in an independent cohort; FANCG is statistically associated with survival in the chol-TCGA dataset. IDH1 wasmutated in theRECsoffive patients; 4ofthem displayed the mutation only in RECs. Deep sequencing performed in one patient confirmed the IDH1 mutation in REC. Conclusions: RECs are enriched for genes involved in EMT, resistance to apoptosis, and cytoskeleton remodeling. Key players of these pathways might be considered druggable targets in RECs. IDH1 is mutated in 30% of RECs, becoming both a marker of progression and a target for therapy. Keywords: Intrahepatic cholangiocarcinoma, Recurrence, IDH1 mutation, Microarray, Prognostic marker Background inflammation processes, such as cholangitis/primary Intrahepatic cholangiocarcinoma (ICC) is an aggressive sclerosing cholangitis (PSC), secondary biliary cirrhosis, malignancy arising from epithelial cells of the bile ducts choledocholithiasis, hepatolithiasis, cholecystitis, as well and is considered the second most common liver cancer as HCV and HBV infections promote ICC arising and type. Limited success in the clinical management and a progression [3–6]. Conventional chemotherapy, based persistent increase in the incidence world-wide have on combination of gemcitabine (GEM) and platinum made ICC one of the most lethal and fastest growing compounds, and radiotherapy, to date, are not effective malignancies. In the last three decades, a general incre- in improving long-term survival [7, 8]. Moreover, ment of ICC incidence was registered in the Western primary or acquired resistance is inevitable and no countries, and in particular in Italy [1, 2]. Chronic second-line chemotherapy has demonstrated efficacy. It is known that 5-years survival rate of ICC patients re- * Correspondence: caterina.peraldoneia@ircc.it; francesco.leone@ircc.it mains low, between 25 and 35% in most of the case G. Chiorino and F. Leone contributed equally to this work. 1 series. Literature data showed that ICC recurrences Medical Oncology Division, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. 142, km 3.95, 10060 Candiolo, Turin, Italy occur in about half of the patients after surgery with Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 2 of 10 curative intent, frequently during the first year and that IDH1 mutations promoted ICC by blocking hepato- usually in the liver [9]. cyte differentiation with an increased number of hepatic In the last years, different molecular studies were con- progenitors susceptible to other mutations [22]. Further, ducted using intensive high-throughput techniques (i.e. gene patients harboring IDH1 mutations had a distinct tran- expression and microRNA profiling, deep-sequencing), to scriptional signature enriched for hepatic stem cell broaden the knowledge on the biological aspects of ICC genes, identifying a particular subclass of ICC patients progression and to identify potential molecular targets. [23]. In preclinical models, IDH1 mutated cell lines were Genomic and molecular mechanisms involved in the onset, highly responsive to Src inhibitors, such as Dasatinib progression as well as in chemotherapy resistance in ICC and Saracatinib, suggesting potential targeted therapies are poorly documented. Preclinical investigations showed [24]. Recently, different preclinical studies aimed at the involvement of oncogenic pathways in cholangiocarci- studying the efficacy of IDH1 inhibitors have been per- nogenesis; among them, the overexpression of EGFR, formed. Further, phase I, II and III clinical trials were HER2, VEGFR and its ligand, MET, signaling pathways, planned and are ongoing to test the safety and efficacy of which cause a dysregulation of downstream effectors, such IDH1 inhibitors in different malignancies, such as glioma, as Ras/Raf/Mek/Erk and PI3K/Akt/PTEN axes [10–12]. cholangiocarcinoma, AML (NCT02074839, NCT02073994, Recently, using gene profiling techniques, Sia and collabora- NCT02719574, NCT02989857). tors [13]demonstrated thatICC couldbestratifiedonthe The identification of the peculiar molecular alterations bases of molecular characteristics, which correlate with dif- of recurrent lesions is required due to the high rate of ferent prognosis. One hundred and forty-nine ICC were local recurrence of this tumor [25]. To date, there are classified in two main classes according to their gene expres- only few data regarding the mechanisms involved in re- sion profiles; an inflammation class, associated to a “good” current disease. For this reason, in this work, we have prognosis, and a proliferation class, associated to a worse molecularly characterized paired primary/recurrent ICC prognosis [13]. In a work of Andersen and collaborators, tumors in order to provide a panel of markers (mutated 104 ICC samples were analyzed by gene expression profiling or deregulated genes) involved in the progression and the two prognostic groups were confirmed [14]. process of this subtype of tumors as well as of new suit- Recently, genomic analyses were conducted on primary ICC able targets for therapy. tumors by mutational profiling using different techniques, defining a broad range of mutations, according to the co- Methods horts analyzed. The most commonly observed alterations Patients were within TP53, KRAS, PI3K, BRAF, SMAD4, IDH1, Eighteen pairs of formalin fixed (ID #1-#18), paraffin em- IDH2, NRAS, ARID1A, PTEN, CDKN2A, CDK6, ERBB3, bedded (FFPE) primary and recurrent ICC tumors were MET, BRCA1, BRCA2, NF1, PTCH1,and TSC, with variable collected from 5 different Italian centers. The independent percentages due to the heterogeneity of the case stud- cohort is constituted by 13 fresh ICC tumors (ID#19-#31), ies [12, 15–17]. These data have been studied to plan 10 PRs (named CHC001-PR to CHC024-PR), 3 RECs tu- clinical trials aimed at inhibiting specific targets, alone or mors (CHC002-REC, CHC012-REC, and CHC017-REC) in combination with standard chemotherapy. However, and 7 ascites samples (ID #33-#38), named PARA-2 to the obtained results are modest and not of impact. PARA-11, (where cancer cells were isolated from ascites A key role in tumorigenesis seems to be played by mu- liquid obtained by paracentesis procedure), obtained from tant IDH1. IDH1 mutation causes an impaired produc- different patients. This cohort was analyzed separately for tion of α-KG in favour of the oncometabolite 2-HG [18]; gene expression profiling and mutational analysis of in particular, it acts as a competitor of α-KG, causing a IDH1, assuming that PARAs are progressive disease and hypermethylation of histones and of DNA and promot- consequently they could be assimilated to recurrences. ing epigenetic alterations, all phenomena typically found Additional file 1: Table S1 summarizes patient clinical and during progression and metastatic processes. The role of pathological characteristics, and the analyses performed. IDH1 as prognostic marker is controversial; literature The median age of patients is 65, ranging from 41 to 84; data demonstrated that IDH1 mutations correlated with 24 females and 14 males were analyzed. The two inde- good prognosis in brain tumors, such as glioma, glio- pendent cohorts were homogeneous in terms of gender blastoma and anaplastic astrocytoma [19]. On the con- (Fisher’s Exact test p-value = 0.7), age at diagnosis and trary, in acute myeloid leukemia (AML) and in ICC it time to recurrence (Student T-test p-values = 0.5 and 0.2, seems that the presence of mutations did not affect the respectively). overall survival (OS) and progression free survival (PFS) [12, 20]. It has been demonstrated that ICC patients are Nucleic acids extraction and quality control frequently mutated (about 25%) in IDH1 hot-spots [21]. Total RNA was extracted from FFPE tissues using the A recent work of Saha and collaborators demonstrated miRNeasy FFPE mini kit, following the manufacturer’s Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 3 of 10 instructions. Briefly, RNA quantity was evaluated by using the Benjamini-Hochberg correction [28]. TMev Nanodrop, while the quality was assessed by qRT-PCR (http://mev.tm4.org) and hclust R function were used to testing the Ct of two different amplicons of ACTB. Only perform hierarchical clustering of genes/samples using for 11 couples, the RNA had an acceptable quality to either selected genes or the global gene expression pro- perform further experiments. filing and to carry out principal component analysis DNA was extracted using Qiamp DNA FFPE kit, fol- (PCA). MetaCore version 6.29 (Thomson Reuters) was lowing manufacturer’s instructions. Briefly, tumor slides used for network and pathway maps analysis. GSEA was were stained with Hematoxylin and Eosin: tumor areas used to evaluate significant enrichment in predefined cu- were circled by a pathologist. Representative images of rated sets of genes from online pathway databases and H/E staining are shown in Additional file 2: Figure S1. publications in PubMed [29]. The weighted voting algo- Representative tumor areas were scraped, deparaffi- rithm and leave-one-out cross validation available within nized by xylene, rehydrated, subsequently treated with the SET tool (Signature Evaluation Tool) were used to proteinase K and then purified using columns. Total evaluate the discrimination power of our expression RNA of fresh frozen tissues and tumor cells obtained by signature on the validation set [30]. The SET algorithm paracentesis was extracted by Absolutely RNA miRNA allows re-evaluation and re-adjustment of the discrimin- kit (Agilent Technologies), while DNA was extracted by ation power of a given signature by selecting/de-select- QiAmp DNA mini kit (Qiagen), following manufac- ing genes repeatedly. Microarray data were deposited in turer’s protocols. Gene Expression Omnibus (GSE107102). cDNA mediated annealing, selection, extension and External dataset ligation (DASL) assay cBioportal was used to download expression profiles of Total RNA extracted from FFPE samples was retrotran- selected genes in the cholongiocarcinoma TCGA dataset scribed to cDNA using oligo-dT and random nanomer (33 samples), together with clinical information about primers, byotinilated and bound to streptavidin particles. survival [31, 32]. mRNA Expression z-Scores (RNA Seq The reactions of labeling with the single color Cy3, V2 RSEM) of the selected genes were used to stratify pa- denaturation, and hybridization on Illumina Human tients in two groups according to expression medians. R Reference 8 BeadArrays were conducted according to survival package was applied to run survival analysis and the manufacturer’s instruction. The slides were washed generate Kaplan-Meier curves. and scanned using the Illumina BeadArray Reader (Illumina, San Diego, CA). Mutational analysis Quality control and quantification of extracted DNA were Gene expression analysis (GEP) by Agilent platform conducted by Bioanalyzer and Qubit, respectively. IDH1 For GEP analysis, Low Input Quick Amp Labeling Kit, exon 4 was amplified by nested PCR with relative specific one-color kit (Agilent Technologies) was used to amplify primers (IDH1 external primers: Forward 5’-TGAGCTCTA and label 100 ng of total RNA. Six hundred ng were TATGCCATCACTGCA-3′, Reverse 5’-CAATTTCAT hybridized on SurePrint G3 Human Gene Expression ACCTTGCTTAATGGG-3′; IDH1 internal primers: For- 8x60K v2 glass arrays. After arrays scansion, images were ward 5’-GCAGTTGTAGGTTATAACTATCC-3′;Reverse analyzed by the Feature Extraction Software from 5’-TGGGTGTAGATACCAAAAG-3′). The PCR products Agilent Technologies (version 10.7); raw data were then were purified using Wizard®SV Gel and PCR Clean-Up processed using the LIMMA (LInear Models for Micro- System (Promega, Milan, Italy) and sense and antisense array Analysis) package from Bioconductor [26]. sequences were obtained by using internal forward and re- verse primers, respectively. Sequencing was performed by Microarray data analysis BigDye Terminator Cycle sequence following the PE Ap- Raw data intensities were loaded into R statistical envir- plied Biosystem strategy and Applied Biosystems ABI onment. The normexp method was used for background PRISM3100 DNA Sequencer (Applied Biosystem, Forster correction with an offset of 50 and the quantile method City,CA).All mutationswere confirmed by two independ- for normalization. To remove batch effect between sub- ent PCR experiments. sets of experiments, the combat function was applied to the dataset [27]. Next generation sequencing (NGS) LIMMA was then used to identify differentially The Ion Torrent S5 platform was used to perform the expressed genes in recurrent vs primary ICC samples, NGS analysis of the Ampliseq CHPv2 which contains using a paired statistics for paired samples of the first the hot-spot mutations of 50 cancer related genes. tested cohort or unpaired statistics for the independent Briefly, DNA of patient #3, both PR and REC, was used one [26]; p-values were adjusted for multiple testing by to prepared libraries by Ion AmpliSeq Library kit 2.0 Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 4 of 10 (ThermoFisher Scientific Waltham, MA) according to these tumors. In particular, four PRs belonged to a dif- the manufacturer’s protocol. Libraries were then quanti- ferent branch arm. Principal component analysis (PCA) fied using Qubit (ThermoFisher Scientific Waltham, was able to clearly separate primary from recurrent sam- MA; 60 pM of each sample were run and sequenced for ples, as shown in Fig. 1b. Then, we refined the analysis, 2800 hot-spots of 50 among oncogenes and tumor sup- applying a two-class paired comparison and filtering pressors. Raw data were analyzed by Ion Reporter soft- data with a cut-off on logFC > 1 or < − 1 and an adjusted ware (ThermoFisher Scientific Waltham, MA) and p-value < 0.01, obtaining 315 significant deregulated filtered in p-value < 0.001 and coverage of > 400 were genes, of which 65 down- and 250 up-regulated in RECs accomplished. versus PRs (Additional file 3: Table S2). Figure 1c shows the heatmap of this genes signature. Results Using Metacore software, we performed an enrich- Transcriptomic analysis of paired PRs and RECs ICC ment analysis of pathway maps and process networks. To analyze sample distributions according to their tran- Additional file 4: Table S3 and Additional file 5: Table S4 scriptomic profiles, unsupervised hierarchical clustering describe the first 15 pathways and networks enriched in was applied to the global normalized intensity profiles. up-regulated genes. They were associated to Epithelial to As shown in Fig. 1a, PRs and RECs tumors had different Mesenchymal Transition (EMT) mediated by Rho alpha, distributions; for some patients, PRs and RECs had simi- PI3K and ILK mediated by TGF beta, cytoskeleton re- lar profiles, for others the RECs were very distant from modeling by GTPase, anti-apoptotic process mediated their PR tumors, underlining the high heterogeneity of by BAD phosphorylation, and in general cell cycle, Fig. 1 a Dendrogram obtained from unsupervised hierarchical clustering using Euclidean distance as similarity metrics and ward as linkage method. Each branch of the dendrogram is represented by the global gene expression profile of samples. The vertical axis indicates the Euclidean distance between samples/clusters. PR and REC of the same patient are represented with the same color. b Projection of principal components 1 and 3 after application of PCA on the paired samples cohort. Red squares correspond to primary samples while yellow squares to recurrences. c Unsupervised hierarchical clustering analysis of 315 significantly deregulated genes in REC vs PR tumors. A red-to-green gradient was used to indicate, for each gene, levels of up- or downregulation. The logFC values of the entire matrix used for hierarchical clustering are provided as Additional file 3:Table S2 Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 5 of 10 cytoskeleton remodeling and EMT networks. On the Further, we applied the 24 genes signature on the val- contrary, down-regulated genes affected in particular, idation dataset and we demonstrated that the expression PTEN signal transduction, immune response, apoptosis of 9 genes (NDST2, DAO, FANCG, CARD9, FOXJ2, mediated by p53, proteolysis related to cell cycle and SEMA3B, GDAP1L1, TRIOBP, PFKM) is able to distin- apoptosis, inflammation mediated by IL-6 signaling guish PRs from RECs (Fig. 3), with an error rate of 0.15 (Additional file 6: Table S5 and Additional file 7:Table S6). and p < 0.001. The weights of individual genes are re- The application of Gene Set Enrichment Analysis on ported in Additional file 11: Table S8. the complete list of pre-ranked genes was not very con- Finally, the expression profiles of the 9 genes and the clusive. However, pre-ranked GSEA on the 315 genes survival information about patients were downloaded signature showed a high number of overlapping genes from the cholangiocarcinoma TCGA dataset (n = 33) with the “Liver_cancer_UP” geneset described by Acevedo available through cBioportal (http://www.cbioportal.org). and collaborators (Enrichment score 0.38, p =0.009; Among the genes overexpressed in patients with worst Fisher’s exact test p-value = 0.002; Jaccard Index = 0.022) prognosis, the one mostly associated with survival is as shown in Additional file 8:Figure S2 [29]. FANCG, with Cox proportional hazard ratio = 3.242 (Fig. 4). Further, restricting the analysis with more stringent filters (logFC <−2or>2 and p < 0.01) we obtained a IDH1 is a potential marker of tumor progression gene signature of 24 deregulated genes, 10 down- and For 16 paired PR and REC ICC tumors, the mutational 14 up-regulated, able to separate RECs from PRs tu- analysis of IDH1 exon 4 was conducted. As shown in mors (Fig. 2). Additional file 12: Table S9, five patients (#3, #6, #11, #15 These genes could be grouped in four main functional and #18) harbored an IDH1 R132x mutation in the REC classes: regulation of apoptosis, (FOXJ2, SEMA3B, counterpart (31.3%); only in patient #6 (6.25%) the muta- CARD9), cellular migration and motility (CLDN23, tion was already present in PR. Figure 5 shows electrophe- TRIOBP), DNA-RNA processing (FANCG, UBLCP1), rograms of mutated samples compared to WT one. metabolic processes such as glycolysis and fatty acids In order to confirm the naïve IDH1 mutation identified metabolism (PECI, PFKM, NDST2, DAO). The transcripts by Sanger sequencing in REC tumor, patient #3 was ana- of this signature were investigated on an independent co- lyzed by NGS using the AmpliSeq technology on Ion Tor- hort of patients (10 RECs vs 11 PRs, IDs #19-#38, see rent device. In patient #3, the IDH1 mutation in codon 132 Additional file 1: Table S1 and Additional file 9: Table S7); was confirmed in REC tumor, with a 17.37% of frequency as shown in Additional file 10: Figure S3, expression fold (p = 0.0001) and PR counterpart resulted WT. Overall, the changes of 17 out of 24 genes were concordant in the two mutational profile is partially overlapping between PR and cohorts, with 8 genes differentially expressed in a statisti- REC, even if a higher number of missense mutations was cally significant manner in the independent cohort as well. identified in PR. Additional file 13: Table S10 summarizes Fig. 2 Unsupervised clustering analysis of 24 significantly deregulated genes in 13 REC tumors compared to 11 PR tumors. Tmev software was used, with Euclidean distance as similarity metrics and complete linkage as linkage method. Log Intensities of each gene were standardized by median centering and dividing by standard deviation. Red/green rectangles indicate expression higher/lower than the median, respectively Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 6 of 10 Fig. 3 Predictive role of 9 out 24 genes of the signature. The expression of these genes is able to clearly separate PRs (square) and RECs (circle) in the validation cohort of patients. Signal to noise scores provided by SET are shown for each gene in Additional file 11: Table S8 the mutational pattern (exonic missense and synonymous, from their primary tumors (PR) by the transcriptomic intronic) of patient #3. Among them, KDR, APC, RB1, enrichment in genes involved in proliferation, motility TP53, IDH1 are already described [33]. and migration, apoptosis resistance, and epithelial-to mesenchymal transition. Further, mutated IDH1 hotspot Discussion (codon 132) emerged in about 30% of REC tumors sug- In this work, we demonstrated that recurrent intrahepa- gesting that it could be a putative marker of progression. tic cholangiocarcinomas (REC ICC) are distinguishable Results obtained by the comparison of matched REC/PR ICC suggest that recurrent lesions could be molecularly different from their primary tumors due to a clonal se- lection toward drug resistant and more malignant tumor cells in RECs. Mutational analysis by Sanger revealed that 4 patients harbored IDH1 R132x mutations in RECs, but not in the PRs; NGS analysis performed in a PR-REC pair showed that IDH1 mutation was gained in REC, confirming our direct sequencing results. How- ever, other missense mutations found in PR were lost, suggesting that a clonal sieving occurred in REC. As a matter of fact, Sanger sequencing has limitations; small DNA fragments could be analyzed with a single reaction and some mutations may re- sult undetectable, due to the low clonal representa- tion. On the contrary, NGS is more sensitive and covers broad range spectra of mutations with a low amount of DNA and should be preferred as a screening method. However, our data supports the heterogeneous nature of this tumor type, both intra- and inter-patients. Many factors, including treatment, Fig. 4 Kaplan-Meier curves for 33 patients from the TCGA could concur in the evolution of the tumor. For the cholangiocarcinoma external dataset, with survival information available. Patients are divided in two groups according to FANCG choice of second-line treatments it should be consid- expression. Red curve: FANCG expression higher than the median. ered that responsive tumor clones are inhibited by Black curve: FANCG expression lower than the median. Log-rank test previous therapy and concurrently disease progression is p-value = 0.0544. Cox proportional hazard ratio = 3.242 sustained by chemotherapy-resistant clones. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 7 of 10 Fig. 5 Representative electropherograms of mutated samples in the hot-spot codon 132 of IDH1 According to microarray data, PRs and RECs of the mesenchymal markers and reduction of epithelial same patient could have very similar or distinct profiles. markers, are crucial events during tumor progression to- This finding is enhanced in particular in those patients ward a more aggressive, proliferating and drug resistant for whom we had two RECs; in one case, the two RECs phenotypes [38, 39]. We showed that a 24 genes signa- are comparable and the expression profile is far from ture is able to distinguish RECs from PRs in the main their PR, but for the other patient, one REC is near to cohort of analysis. The trend of these genes is partially PR while the other one displays a different pattern of ex- confirmed on an independent cohort of patients consti- pression. Further, the complexity of these tumors is tuted by fresh frozen PR or REC ICC tumors and tumor highlighted not only by the heterogeneity identified be- cells obtained from ascites liquid of ICC patients in pro- tween RECs and their PRs, but also within different bi- gressive disease (PARAs). The two cohorts are small, opsies of the same lesion [34]. Globally, we found an which could limit the robustness of our results, but they increased expression of genes involved in nucleic acids are homogeneous in terms of baseline characteristics processing, transcription of RNA and non-coding RNA, (sex, age at diagnosis) and time to recurrence. They are and angiogenesis regulation, with a concomitant mainly composed of T2 stage tumors, with ascites in- down-regulation of genes related to epithelial cell differ- cluded in the validation cohort only, which might be po- entiation. The down-regulation of epithelial cell differen- tentially confounding. However, in the latter cohort the tiation genes suggested that the epithelial-like phenotype expression of 9 of the signature genes is able to discrim- is switched towards a mesenchymal-like. This data is inate RECs from PRs in a statistically significant manner. confirmed by pathways and maps analyses, with an en- Moreover, these molecules might represent novel richment of up-regulated genes involved in EMT, in par- therapeutic targets. Namely, CARD9 is a marker of ticular induced by TGF-beta. In agreement with several tumor progression and poor prognosis in hepatocarci- evidences in other types of cancer [35–37], these data noma (HCC), B cell lymphoma, and clear cell renal confirmed that EMT is one of the crucial steps of recur- carcinoma [40–42] and promotes metastatization acti- rence and drug resistance. We also described an vating metastasis-associated macrophages (MAM) [43]. up-regulation of genes involved in cytoskeleton remodel- Besides, the higher expression of FANCG suggested that ing and cell cycle regulation in REC tumors. Many stud- REC tumors displayed increased DNA damage and ies described the close interconnection among these activated DNA repair. The downregulation of FANCG events; the remodeling of extracellular matrix and was associated with effective treatment with GEM and reorganization of cytoskeleton, along with expression of radiolabeled Trastuzumab in tumor xenograft model of Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 8 of 10 disseminated intraperitoneal disease (46). Moreover, we an independent cohort of patients. The same trend was found in 17 out found an up-regulation of PIK4CA which is involved in of 24 genes. * indicates statistically significant results. Y axis: log fold change expression obtained in the two independent cohorts. (TIF 162 kb) proliferation and chemoresistance in other tumors, such Additional file 11: Table S8. Weights of the 9 genes signature. (XLSX 9 kb) as medulloblastoma [44]. The same trend of expression Additional file 12: Table S9. Sanger sequencing of IDH1 exon 4. (DOCX was revealed for TRIOBP, already described in pancreatic 15 kb) cancer where is involved in cell motility and migration Additional file 13: Table S10. NGS sequencing of patient #3. (XLSX through cytoskeleton remodeling [45, 46]. Interest- 15 kb) ingly, the emergence of IDH1 in about 30% of REC patients suggested that it could be considered not Abbreviations only a marker of progression but also a potential tar- 2-HG: 2-hydroxyglutarate; AML: Acute myeloid leukemia; DASL: cDNA mediated annealing, selection, extension and ligation; EMT: Epithelial to get for tailored therapy. In fact, it has been demon- mesenchymal transition; FDR: False discovery rate; FFPE: Formalin fixed strated that IDH1 mutation promoted sensitivity to paraffin embedded; GEM: Gemcitabine; GEP: Gene expression analysis; the multitarget inhibitor Dasatinib [24]. Moreover, GSEA: Gene set enrichment analysis; HBV: Hepatitis B virus; HCV: Hepatitis C virus; ICC: Intrahepatic cholangiocarcinoma; LIMMA: Linear models for mutated-IDH1 targeting agents are now under clinical microarray analysis; MAM: Metastasis-associated macrophages; NGS: Next investigation in different solid tumors, including ICC. generation sequencing; PFS: Progression free survival; PR: Primary tumor; As an example, BAY1436032 targeting the hot-spot PSC: Primary sclerosing cholangitis; REC: Recurrent tumor; SET: Signature evaluation tool; TCGA: The Cancer Genome Atlas; WT: Wild-type; α-KG: Alpha mutation R132x is now tested in patients with ad- ketoglutarate vanced solid tumors enrolled in one open-label, non-randomized, multicenter phase I-II clinical trial Acknowledgements (NCT02746081). Of main interest, the effect of A special acknowledgement is given to an Italian association for the study of ICC, created in memory of Annalisa Pozzi. AG-120 is now compared to placebo in the phase III, multicenter, randomized double-blind ClarIDHy trial Funding on non-resectable/metastatic cholangiocarcinoma. “Associazione Italiana Ricerca sul Cancro–AIRC 5X1000 2010-Ministry of Health, FPO”. Project n°16:30 “Identificazione di nuove vie di trasduzione del segnale intra- cellulare sensibili ai farmaci nel colangiocarcinoma intraepatico (ICC)”.Fondazione Conclusions Piemontese per la Ricerca sul Cancro (FPRC) “Identification of new druggable In conclusion, our data on transcriptomic and mutational pathways in intrahepatic cholangiocarcinoma” 5 per Mille 2010 Ministero della status of REC ICC suggest that a personalized approach, Salute. Università di Torino anno 2014 –: Fondo per la ricerca locale (Linea B), LEOF_RIC_LOC_14_01 project title: “Transcriptomic and genetic analysis of in which tumor molecular/genetic characterization are paired primary and recurrent intrahepatic cholangiocarcinoma”.CPN andDS: followed from diagnosis to disease progression, is advis- FPRC 5 × 1000 Ministero della Salute 2012. GC is supported by a grant of Com- able not only for prognostic purposes, but also to identify pagnia di San Paolo; AS is supported by Associazione Italiana per la Ricerca sul Cancro grant AIRC 5 × 1000 n. 12182. the emergence of other druggable targets after first-line treatment failure. Availability of data and materials Microarray data were deposited in Gene Expression Omnibus (GSE107102). Additional files Authors’ contributions All authors read and approved the final version of the manuscript. CPN Additional file 1: Table S1. Clinical pathological characteristics of ICC conceived the work, performed the experiments an wrote the article draft; patients. (DOCX 19 kb) PO performed microarray data analysis; GC performed validation experiments and wrote the article draft; YP performed NGS analyses and wrote the article Additional file 2: Figure S1. Representative images of H/E staining of draft, DS revised the manuscript and performed statistical analyses, LDC and PRs (A and C) and their RECs counterparts (B and D). (TIF 793 kb) EM performed DASL analysis, DR, AS, AMDR, AG, FC, CR, PI recruited patients Additional file 3: Table S2. List of significant deregulated genes in and biological material for this work and revised the manuscript; MA and FL RECs versus PRs. (XLSX 39 kb) conceived and supervised the work and revised the manuscript; GC conceived Additional file 4: Table S3. Pathway maps obtained with Metacore the work, performed data analysis and revised the manuscript. analyzing only up-regulated genes. (DOCX 16 kb) Additional file 5: Table S4. Process networks obtained with Metacore Ethics approval and consent to participate analyzing only up-regulated genes. (DOCX 15 kb) Biological material was obtained, in accordance with the Declaration of Helsinki from patients who have signed the informed consent. The study Additional file 6: Table S5. Pathway maps obtained with Metacore was approved by each local institutional review boards (Comitato Etico analyzing only down-regulated genes. (DOCX 15 kb) “IRCCS Candiolo”; “Comitato Etico AORN Cardarelli-Santobono”;Comitato Additional file 7: Table S6. Process networks obtained with Metacore Etico “Policlinico Gemelli”; “Comitato Etico per la sperimentazione clinica analyzing only down-regulated genes. (DOCX 15 kb) delleprovincediVerona eRovigo”; “Comitato Etico indipendente, IRCSS Additional file 8: Figure S2. The GSEA dataset of Acevedo et al., [47] Istituto Clinico Humanitas”. The coordinator center approved the entire on liver cancer was found enriched for up-regulated genes (Enrichment study according to the PROFILING protocol (“Studio prospettico per la score 0.38; p = 0.009, pre-ranked GSEA analysis). (TIF 174 kb) determinazione del profilo molecolare di resistenza alle terapie target in pazienti con malattie neoplastiche”, 001-IRCC-00IIS-10, version 6.1, FPO-IRCCS, Additional file 9: Table S7. List of deregulated genes of the l’Istituto di Ricovero e Cura a Carattere Scientifico Candiolo (TO)). indepentent cohort of patients. (XLSX 95 kb) Additional file 10: Figure S3. Comparison between the expression Competing interests values of selected genes obtained by DASL array and GEP performed on The authors declare that they have no competing interests. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 9 of 10 Publisher’sNote 14. Andersen JB, Spee B, Blechacz BR, Avital I, Komuta M, Barbour A, Conner EA, Springer Nature remains neutral with regard to jurisdictional claims in Gillen MC, Roskams T, Roberts LR, et al. Genomic and genetic characterization published maps and institutional affiliations. of cholangiocarcinoma identifies therapeutic targets for tyrosine kinase inhibitors. Gastroenterology. 2012;142:1021–31. e1015 Author details 15. Chong DQ, Zhu AX. The landscape of targeted therapies for cholangiocarcinoma: Medical Oncology Division, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. current status and emerging targets. Oncotarget. 2016;19:46750–767. 142, km 3.95, 10060 Candiolo, Turin, Italy. Cancer Genomics Lab, Fondazione 16. Nakamura H, Arai Y, Totoki Y, Shirota T, Elzawahry A, Kato M, Hama N, Edo ed Elvo Tempia Valenta, Biella, Italy. Department of Oncology, University Hosoda F, Urushidate T, Ohashi S, et al. Genomic spectra of biliary tract of Turin, Torino, Italy. Functional Genomics and Bioinformatics, Department cancer. Nat Genet. 2015;47:1003–10. of Applied Research and Technology Development, Fondazione IRCCS 17. Simbolo M, Fassan M, Ruzzenente A, Mafficini A, Wood LD, Corbo V, Melisi Istituto Nazionale dei Tumori, Milan, Italy. Division of Hepatobilio-Pancreatic D, Malleo G, Vicentini C, Malpeli G, et al. Multigene mutational profiling of and Colorectal Surgery, Candiolo Cancer Institute - FPO, IRCCS, Str. Prov. 142, cholangiocarcinomas identifies actionable molecular subgroups. Oncotarget. km 3.95, Candiolo, Italy. ARC-Net Research Centre and Department of 2014;5:2839–52. Diagnostics and Public Health – Section of Pathology, University and 18. Garber K. Oncometabolite? IDH1 discoveries raise possibility of new metabolism Hospital Trust of Verona, Verona, Italy. Hepatobiliary Surgery Unit, Gemelli targets in brain cancers and leukemia. J Natl Cancer Inst. 2010;102:926–8. Hospital, Catholic University of the Sacred Heart, Rome, Italy. Department of 19. Sturm D, Witt H, Hovestadt V, Khuong-Quang DA, Jones DT, Konermann C, Health’s Sciences and Medicine “V. Tiberio”, University of Molise, Pfaff E, Tönjes M, Sill M, Bender S, et al. Hotspot mutations in H3F3A and Campobasso, Italy. Hepatobiliary and Liver Transplant Unit, Cardarelli IDH1 define distinct epigenetic and biological subgroups of glioblastoma. Hospital, Naples, Italy. Center for Autoimmune Liver Diseases, Humanitas Cancer Cell. 2012;22:425–37. Clinical and Research Center, Rozzano, Italy. Department of Experimental 20. Goyal L, Govindan A, Sheth RA, Nardi V, Blaszkowsky LS, Faris JE, Clark JW, and Clinical Medicine, University of Firenze, Florence, Italy. UOC di Ryan DP, Kwak EL, Allen JN, et al. Prognosis and Clinicopathologic features Gastroenterologia, Azienda Ospedaliera San Gerardo, Monza, Italy. of patients with advanced stage Isocitrate dehydrogenase (IDH) mutant and IDH wild-type intrahepatic cholangiocarcinoma. Oncologist. 2015;20:1019–27. Received: 20 November 2017 Accepted: 25 May 2018 21. Borger DR, Tanabe KK, Fan KC, Lopez HU, Fantin VR, Straley KS, Schenkein DP, Hezel AF, Ancukiewicz M, Liebman HM, et al. Frequent mutation of isocitrate dehydrogenase (IDH)1 and IDH2 in cholangiocarcinoma identified through broad-based tumor genotyping. Oncologist. 2012;17:72–9. References 22. Saha SK, Parachoniak CA, Bardeesy N. IDH mutations in liver cell plasticity 1. Bragazzi M, Cardinale V, Carpino G, Venere R, Semeraro R, Gentile R, Gaudio and biliary cancer. Cell Cycle. 2014;13:3176–82. E, Alvaro D. Cholangiocarcinoma: epidemiology and risk factors. Transl 23. Saha SK, Parachoniak CA, Ghanta KS, Fitamant J, Ross KN, Najem MS, Gastrointest Cancer. 2012;1:21–32. Gurumurthy S, Akbay EA, Sia D, Cornella H, et al. Mutant IDH inhibits 2. Alvaro D, Crocetti E, Ferretti S, Bragazzi MC, Capocaccia R, Committee AC. HNF-4α to block hepatocyte differentiation and promote biliary cancer. Descriptive epidemiology of cholangiocarcinoma in Italy. Dig Liver Dis. Nature. 2014;513:110–4. 2010;42:490–5. 24. Saha SK, Gordan JD, Kleinstiver BP, Vu P, Najem MS, Yeo JC, Shi L, Kato Y, Levin 3. Kim HJ, Kim JS, Joo MK, Lee BJ, Kim JH, Yeon JE, Park JJ, Byun KS, Bak YT. RS, Webber JT, et al. Isocitrate dehydrogenase mutations confer Dasatinib Hepatolithiasis and intrahepatic cholangiocarcinoma: a review. World J hypersensitivity and SRC dependence in intrahepatic cholangiocarcinoma. Gastroenterol. 2015;21:13418–31. Cancer Discov. 2016;6:727–39. 4. Chaiteerakij R, Juran BD, Aboelsoud MM, Harmsen WS, Moser CD, Giama 25. Ercolani G, Vetrone G, Grazi GL, Aramaki O, Cescon M, Ravaioli M, Serra C, NH, Allotey LK, Mettler TA, Baichoo E, Zhang X, et al. Association between Brandi G, Pinna AD. Intrahepatic cholangiocarcinoma: primary liver resection variants in inflammation and cancer-associated genes and risk and survival and aggressive multimodal treatment of recurrence significantly prolong of cholangiocarcinoma. Cancer Med. 2015;4:1599–602. survival. Ann Surg. 2010;252:107–14. 5. Tao LY, He XD, Xiu DR. Hepatitis B virus is associated with the clinical features 26. Smyth GK. Linear models and empirical bayes methods for assessing and survival rate of patients with intrahepatic cholangiocarcinoma. Clin Res differential expression in microarray experiments. Stat Appl Genet Mol Hepatol Gastroenterol. 2016;40:682–87. Biol. 2004;3:Article3. 6. Wang Z, Sheng YY, Dong QZ, Qin LX. Hepatitis B virus and hepatitis C virus 27. Johnson WE, Li C, Rabinovic A. Adjusting batch effects in microarray expression play different prognostic roles in intrahepatic cholangiocarcinoma: a meta- data using empirical Bayes methods. Biostatistics. 2007;8:118–27. analysis. World J Gastroenterol. 2016;22:3038–51. 28. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and 7. Ulahannan SV, Rahma OE, Duffy AG, Makarova-Rusher OV, Kurtoglu M, powerful approach to multiple testing. Royal Statistical Society. 1995;57:12. Liewehr DJ, Steinberg SM, Greten TF. Identification of active chemotherapy 29. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, regimens in advanced biliary tract carcinoma: a review of chemotherapy trials Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP. Gene set in the past two decades. Hepat Oncol. 2015;2:39–50. enrichment analysis: a knowledge-based approach for interpreting genome- 8. Yang R, Wang B, Chen YJ, Li HB, Hu JB, Zou SQ. Efficacy of gemcitabine plus wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–50. platinum agents for biliary tract cancers: a meta-analysis. Anti-Cancer Drugs. 30. Jen CH, Yang TP, Tung CY, Su SH, Lin CH, Hsu MT, Wang HW. Signature 2013;24:871–7. evaluation tool (SET): a java-based tool to evaluate and visualize the sample 9. Park HM, Yun SP, Lee EC, Lee SD, Han SS, Kim SH, Park SJ. Outcomes for discrimination abilities of gene expression signatures. BMC Bioinformatics. patients with recurrent intrahepatic cholangiocarcinoma after surgery. Ann 2008;9:58. Surg Oncol. 2016;23:4392–400. 31. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, Sun Y, 10. Yoshikawa D, Ojima H, Iwasaki M, Hiraoka N, Kosuge T, Kasai S, Hirohashi S, Jacobsen A, Sinha R, Larsson E, et al. Integrative analysis of complex Shibata T. Clinicopathological and prognostic significance of EGFR, VEGF, cancer genomics and clinical profiles using the cBioPortal. Sci Signal. and HER2 expression in cholangiocarcinoma. Br J Cancer. 2008;98:418–25. 2013;6:pl1. 11. Nakazawa K, Dobashi Y, Suzuki S, Fujii H, Takeda Y, Ooi A. Amplification and 32. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, Jacobsen A, overexpression of c-erbB-2, epidermal growth factor receptor, and c-met in Byrne CJ, Heuer ML, Larsson E, et al. The cBio cancer genomics portal: an biliary tract cancers. J Pathol. 2005;206:356–65. open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2:401–4. 12. Zhu AX, Borger DR, Kim Y, Cosgrove D, Ejaz A, Alexandrescu S, Groeschl RT, Deshpande V, Lindberg JM, Ferrone C, et al. Genomic profiling of intrahepatic 33. Lee CH, Wang HE, Seo SY, Kim SH, Kim IH, Kim SW, Lee ST, Kim DG, Han MK, cholangiocarcinoma: refining prognosis and identifying therapeutic targets. Lee SO. Cancer related gene alterations can be detected with next-generation Ann Surg Oncol. 2014;21:3827–34. sequencing analysis of bile in diffusely infiltrating type cholangiocarcinoma. 13. Sia D, Hoshida Y, Villanueva A, Roayaie S, Ferrer J, Tabak B, Peix J, Sole M, Exp Mol Pathol. 2016;101:150–6. Tovar V, Alsinet C, et al. Integrative molecular analysis of intrahepatic 34. Brandi G, Farioli A, Astolfi A, Biasco G, Tavolari S. Genetic heterogeneity in cholangiocarcinoma reveals 2 classes that have different outcomes. cholangiocarcinoma: a major challenge for targeted therapies. Oncotarget. Gastroenterology. 2013;144:829–40. 2015;6:14744–53. Peraldo-Neia et al. BMC Genomics (2018) 19:440 Page 10 of 10 35. Iwahashi S, Shimada M, Utsunomiya T, Imura S, Morine Y, Ikemoto T, Takasu C, Saito Y, Yamada S. Epithelial-mesenchymal transition-related genes are linked to aggressive local recurrence of hepatocellular carcinoma after radiofrequency ablation. Cancer Lett. 2016;375:47–50. 36. Suresh R, Ali S, Ahmad A, Philip PA, Sarkar FH. The role of Cancer stem cells in recurrent and drug-resistant lung Cancer. Adv Exp Med Biol. 2016;890:57–74. 37. Kubelt C, Hattermann K, Sebens S, Mehdorn HM, Held-Feindt J. Epithelial-to- mesenchymal transition in paired human primary and recurrent glioblastomas. Int J Oncol. 2015;46:2515–25. 38. Chou YS, Yang MH. Epithelial-mesenchymal transition-related factors in solid tumor and hematological malignancy. J Chin Med Assoc. 2015;78:438–45. 39. Savagner P. Epithelial-mesenchymal transitions: from cell plasticity to concept elasticity. Curr Top Dev Biol. 2015;112:273–300. 40. Zekri AR, El-Kassas M, Saad Y, Bahnassy A, El-Din HK, Darweesh SK, Abdel Hafez H, Esmat G. Caspase recruitment domains. New potential markers for diagnosis of hepatocellular carcinoma associated with HCV in Egyptian patients. Ann Hepatol. 2013;12:774–81. 41. Nakamura S, Matsumoto T, Yada S, Hirahashi M, Suekane H, Yao T, Goda K, Iida M. Overexpression of caspase recruitment domain (CARD) membrane- associated guanylate kinase 1 (CARMA1) and CARD9 in primary gastric B- cell lymphoma. Cancer. 2005;104:1885–93. 42. Tan W, Hildebrandt MA, Pu X, Huang M, Lin J, Matin SF, Tamboli P, Wood CG, Wu X. Role of inflammatory related gene expression in clear cell renal cell carcinoma development and clinical outcomes. J Urol. 2011;186:2071–7. 43. Yang M, Shao JH, Miao YJ, Cui W, Qi YF, Han JH, Lin X, Du J. Tumor cell- activated CARD9 signaling contributes to metastasis-associated macrophage polarization. Cell Death Differ. 2014;21:1290–302. 44. Guerreiro AS, Fattet S, Kulesza DW, Atamer A, Elsing AN, Shalaby T, Jackson SP, Schoenwaelder SM, Grotzer MA, Delattre O, Arcaro A. A sensitized RNA interference screen identifies a novel role for the PI3K p110γ isoform in medulloblastoma cell proliferation and chemoresistance. Mol Cancer Res. 2011;9:925–35. 45. Bao J, Wang S, Gunther LK, Kitajiri S, Li C, Sakamoto T. The actin-bundling protein TRIOBP-4 and -5 promotes the motility of pancreatic cancer cells. Cancer Lett. 2015;356:367–73. 46. Seipel K, O'Brien SP, Iannotti E, Medley QG, Streuli M. Tara, a novel F-actin binding protein, associates with the trio guanine nucleotide exchange factor and regulates actin cytoskeletal organization. J Cell Sci. 2001;114:389–99. 47. Acevedo LG, Bieda M, Green R, Farnham PJ. Analysis of the mechanisms mediating tumor-specific changes in gene expression in human liver tumors. Cancer Res. 2008;68:2641–51.

Journal

BMC GenomicsSpringer Journals

Published: Jun 5, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off