Search

Supplementary Table S2: GO term enrichment analysis of differentially expressed genes between t1 and t0, overlapping in PDX 788, 926, 792

XLSX file 75K, This is an excel file containing 9 sheets, one for each of the 9 supplementary tables (S1 to S9) mentioned in the main text

Supplementary Table 1. Engraftment rates of lung cancer patient-derived xenografts vary according to histology and specimen type; Supplementary Table 3. TP53 mutations in primary lung cancer and corresponding PDX, as detected by exome sequencing; Supplementary Table 4. FGFR1 FISH scores are conserved for successive passages of squamous cell carcinoma PDXs 792, 406, 788, and 926; Supplementary Table 5. FGFR1 gene amplification does not always correlate with FGFR1 RNA expression in a tissue microarray patient cohort; Supplementary Table 6. FGFR1 gene amplification does not always correlate with FGFR1 RNA expression in squamous cell carcinoma patients from a tissue microarray cohort; Supplementary Table 7. FGFR1 gene amplification does not always correlate with RNA expression in a squamous cell carcinoma PDX cohort; Supplementary Material and Methods, and Supplementary Figure Legends

PDF file 234K, This is a file containing the list of supplementary tables, the list of supplementary figures, as well as the 7 supplementary figures (S1 to S7) and their correspondent legends, mentioned in the main text

We discovered a novel somatic gene fusion, CD74–NRG1, by transcriptome sequencing of 25 lung adenocarcinomas of never smokers. By screening 102 lung adenocarcinomas negative for known oncogenic alterations, we found four additional fusion-positive tumors, all of which were of the invasive mucinous subtype. Mechanistically, CD74–NRG1 leads to extracellular expression of the EGF-like domain of NRG1 III-β3, thereby providing the ligand for ERBB2–ERBB3 receptor complexes. Accordingly, ERBB2 and ERBB3 expression was high in the index case, and expression of phospho-ERBB3 was specifically found in tumors bearing the fusion (P < 0.0001). Ectopic expression of CD74–NRG1 in lung cancer cell lines expressing ERBB2 and ERBB3 activated ERBB3 and the PI3K–AKT pathway, and led to increased colony formation in soft agar. Thus, CD74–NRG1 gene fusions are activating genomic alterations in invasive mucinous adenocarcinomas and may offer a therapeutic opportunity for a lung tumor subtype with, so far, no effective treatment.Significance: CD74–NRG1 fusions may represent a therapeutic opportunity for invasive mucinous lung adenocarcinomas, a tumor with no effective treatment that frequently presents with multifocal unresectable disease. Cancer Discov; 4(4); 415–22. ©2014 AACR.This article is highlighted in the In This Issue feature, p. 377

Supplementary Fig. S1. Generation of patient-derived xenografts from surgery and biopsy specimens of non small cell lung cancer Supplementary Fig. S2. Squamous cell carcinoma patient-derived xenografts recapitulate the genomic phenotype of patients'' tumors Supplementary Fig. S3. Analysis of biomarkers of response to FGFR inhibition Supplementary Fig. S4. High FGFR1 RNA expression does not predict patient outcome Supplementary Fig. S5. PIK3CA amplification does not predict response to PI3K inhibitors Supplementary Fig. S6. FGFR targeted therapy in FGFR inhibitor resistant tumors does not increase cisplatin sensitivity

Multidrug resistance protein 1 (MRP1) is an ATP-dependent efflux pump that extrudes chemotherapeutic drugs from cancer cells, preventing their intracellular accumulation and thus their efficacy. MRP1 also regulates intracellular levels of glutathione (GSH), an antioxidant that when elevated in cancer provides resistance against chemotherapy- and radiotherapy-induced oxidative damage. The different binding sites for drugs and GSH on MRP1 has allowed the development of small molecule MRP1 modulators that block chemotherapy efflux whilst simultaneously enhancing GSH efflux to deplete GSH. As no therapeutic options are available to combat MRP1-mediated multidrug resistance, we have been investigating MRP1 modulators for their efficacy in sensitizing MRP1-expressing cancers to treatment. MRP1 immunohistochemistry staining in ovarian and non-small cell lung cancer (NSCLC) patient samples was assessed to determine the frequency of high MRP1 expression. High MRP1 staining was frequently observed in the subcellular compartment (37%, 47%) and, to a lesser extent, plasma membrane (7%, 16%) in ovarian and NSCLC patient tumour samples respectively. In NSCLC, subcellular MRP1 was associated with poorer overall survival in late stage disease (p=0.0095) and membrane MRP1 with poorer disease-free survival in Stage 4 cancer (p=0.01449). Analysis for ovarian cancer is ongoing. As MRP1 is therefore an attractive target for inhibition, we used membrane vesicle uptake assays to compare the ability of modulators to inhibit [3H]-labelled MRP1 substrate transport. 7-(difluoromethyl)-N-4-morpholinyl-5-phenylpyrazolo[1,5-a]pyrimidine-3-carboxamide (7914321) was identified as an analogue of the MRP1 inhibitor Reversan with improved potency. Interestingly, this inhibition was GSH dependent. Exploring further this relationship with GSH, the high MRP1-expressing A549 lung and SKOV3 ovarian cancer cell lines were treated with 7914321 alone or in combination with the GSH synthesis inhibitor buthionine sulfoximine (BSO), then GSH levels were determined by glutathione recycling assay. 7914321 stimulated GSH efflux in both cell lines in an MRP1-dependent manner and strongly synergised with BSO to cause near complete GSH depletion. To determine the potential of this synergistic GSH depletion to chemosensitize cells, the viability of A549 and SKOV3 cells treated with chemotherapy, 7914321, and BSO was assessed by clonogenic and resazurin-based cytotoxicity assays. 7914321/BSO significantly diminished clonogenicity in high but not low MRP1-expressing cells, and further sensitized cancer cell lines to MRP1-substrate chemotherapeutics compared to the single agents. Together, these findings provide preliminary evidence that 7914321 is a potent MRP1 modulator whose dual effects on MRP1 place it in a unique position to potentially enhance the therapeutic window to treat chemoresistant MRP1-overexpressing cancers. Citation Format: Kimberley M. Hanssen, Christine C. Gana, Madeleine S. Wheatley, Denise M. Yu, Claudia L. Flemming, Richard Young, Gavin M. Wright, Gwenaelle Conseil, Catherine J. Kennedy, Anna deFazio, Ben Solomon, Susan P. Cole, Murray D. Norris, Michelle Haber, Jamie I. Fletcher. Modulation of MRP1 activity reverses chemotherapy resistance in adult cancer cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3809.

Pulmonary grade-1 typical (TC) and grade-2 atypical (AC) carcinoids share molecular characteristics with grade-3 large-cell neuroendocrine carcinoma (LCNEC) despite the distinct clinical behaviors. Most carcinoids can be surgically resected, however, limited treatment options exist for metastatic disease, present in 10-23% of TC and 40-50% of AC. Comprehensive genomic studies could help identify better therapeutic opportunities, novel diagnostic markers, and provide insight on the mechanisms responsible for the increased aggressiveness of AC versus TC. Such studies are rare due to the limited availability of suitable material. We have established a multi-center collaboration that has given us access to a unique collection of samples. We have already characterized 40 TC and 60 LCNEC genomes/exomes, and 61 TC, 8 AC and 69 LCNEC trancriptomes (published data). In the present study, we have performed whole-exome and transcriptome sequencing on 20 AC patients. Methylation data from 850K Illumina arrays were also generated for these samples, and for a subset of 20 TC and 20 LCNEC previously mentioned. When comparing the mutational data on AC with that of TC and LCNEC, we have found that similar to TC, AC harbor recurrent alterations in chromatin remodeling genes (such as MEN1 and ARID1A). They also carry alterations in genes involved in other cancer-related pathways (based on STRING), such as cell motility and cell death explaining their more aggressive phenotype. Integrative clustering analysis (MOFA and iCLUSTER) based on expression and methylation data tends to classify carcinoids into four groups: groups 1 and 2 are mostly composed of females with TC, and differ by their age composition and smoking status (Fisher's exact test p=0.008 and 0.03, respectively). Groups 3 and 4 are mostly composed of males with AC (Fisher's exact test for tumor type p=8x10-5). When including the LCNEC data, the samples from group 3 cluster with LCNEC, suggesting that AC can display a variety of expression and methylation patterns that may be linked to aggressiveness. This result was supported by the better survival of groups 1 and 2 compared to groups 3 and 4 (log-rank p=0.02), for which survival was similar to that of patients with LCNEC. Here, we present for the first time: (i) a multi-omics study on AC; (ii) the methylome characterization of TC, AC, and LCNEC; and (iii) the results of a comparative analysis of TC, AC, and LCNEC based on their molecular characteristics. We have identified the genes and pathways that might explain the progression from low-grade TC to intermediate-grade AC. Our expression and methylation data also supports the existence of a “super-AC” group, which clusters with LCNEC. Finally, we have identified a panel of molecular alterations that may help pathologist distinguishing between these three entities. NL and NA contributed equally. LFC and MF jointly supervised this work. Citation Format: Noémie Leblay, Nicolas Alcala, David Hervás Marin, Tiffany M. Delhomme, Théo Giffon, Akram Ghantous, Amélie Chabrier, Cyrille Cuenin, Janine Altmueller, Geoffroy Durand, Catherine Voegele, Philippe Lorimier, Anne-Claire Toffart, Jules Derks, Odd Terje Brustugun, Joachim H. Clement, Joerg Saenger, John K. Field, Alex Soltermann, Gavin M. Wright, Luca Roz, Lucia Anna Muscarella, Paolo Graziano, Zdenko Herceg, Ernst-Jan Speel, Peter Nuernberg, James McKay, Nicolas Girard, Sylvie Lantuejoul, Juan Sandoval, Elisabeth Brambilla, Matthieu Foll, Lynnette Fernandez-Cuesta. Multi-omics comparative analyses of pulmonary typical carcinoids, atypical carcinoids, and large-cell neuroendocrine carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5358.

HIC1 is a zinc-finger transcriptional repressor frequently targeted for deletion and/or epigenetic gene silencing in cancer. In cancer cell lines, re-expression of HIC1 results in the downregulation of target genes thought to promote context-dependent tumor growth, including Sirt1, Atoh1, Efna1 and components of the WNT and STAT3 signalling pathway. In mouse embryonic fibroblasts, we now show that deletion of a conditional loxP allele of Hic1 results in immediate growth arrest without significant overexpression of Hic1 target genes, followed by a prolonged period of cellular quiescence. After several weeks, we then observed the spontaneous emergence of rapidly growing clones that fulfil all the criteria for transformation, associated with aneuploidy and recurrent copy number alterations. By contrast, conditional deletion of p53 using the same experimental protocol resulted in prompt immortalisation without transformation, and a stable tetraploid karyotype resembling that seen in NIH-3T3 cells. These data suggest that inactivation of Hic1 in cells undergoing replication stress in tissue culture can induce complete transformation associated with marked genomic instability. To test this hypothesis in vivo, we employed the inhaled adenoviral Cre recombinase model in order to direct loxP recombination to the airway epithelium of adult mice. As previously described by other groups, Cre activation of a conditional oncogenic KRasG12D mutant resulted in the appearance of multiple lung adenomas within 6-8 weeks following viral inhalation. By contrast, mice carrying this KRasG12D allele, as well as homozygous for the conditional Hic1lox knockout allele, developed highly aggressive adenocarcinomas with prominent pleomorphic and micropapillary histology. These data suggest a novel tumour suppressor role for HIC1 in maintaining genomic stability in replicating cells. HIC1 is known to interact with components of the SWI/SNF and PRC2 complexes, which help maintain genomic integrity by re-establishing constitutive heterochromatin following DNA replication. We therefore speculate that loss of HIC1 function may phenocopy recently described mutations in components of these chromatin-remodelling complexes such as ARID1A and BRG1. Citation Format: Anette Szczepny, Lisa McKenzie, Samantha Jayesekara, Prue Russell, Gavin Wright, Stephen B. Baylin, Fernando J. Rossello, Jason E. Cain, David N. Watkins. Conditional deletion of the tumor suppressor Hic1 results in aneuploidy and single-step transformation. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-81. doi:10.1158/1538-7445.AM2014-LB-81

In nearly 50% of all lung adenocarcinomas the oncogenic driver is still unknown. We performed transcriptome sequencing on 25 pan-negative lung adenocarcinomas of never smokers leading to the identification of the novel CD74-NRG1 gene fusion. In an additional cohort of 102 adenocarcinomas we could identify and validate four additional cases bearing the CD74-NRG1 fusion gene - all of the invasive mucinous subtype. We then transduced NIH-3T3 cells with the fusion and found it localized on the cell surface in FACS experiments. As we found the identical breakpoint in CD74 in our fusion as in CD74-ROS1, we hypothesized that the functional active part needs to be NRG1. (Soda et al. 2012) The NRG1 isoform III-β3 that we found is generally not expressed in lung tissue - therefore the CD74 part leads to the expresseion of NRG1 III-β3 in the lung. As Neuregulins are known interaction partners of HER receptor family members we transduced the human adenocarcinoma cell lines H322 and H1568, that are HER receptor and KRAS wildtype, with the fusion and found phosphorylation of HER2 and HER3 in the transduced cells under starvation conditions compared to empty vector controls. (Hobbs et al. 2002) This holds true as well in the patient situation in which phospho-HER3 was only found in tumors expressing the fusion (p We could further show that the CD74-NRG1 fusion is associated with activation of the phosphoinsitide-3-kinase pathway activation and the mitogen-activated protein (MAP) kinase pathway . As functional consequence we could observe in several human adenocarcinoma cell lines increased colony forming ability in soft-agar assays showing the CD74-NRG1 fusion being a new oncogenic driver in lung adenocarcinoma. This finding may lead to new treatment options for patients harboring this fusion gene who do not benefit from any effective treatment option up to now. Citation Format: Dennis Plenker, Lynnette Fernandez-Cuesta, Hirotaka Osada, Ruping Sun, Marc Bos, Juliane Daßler, Gavin Wright, Elisabeth Brambilla, Reinhard Büttner, Sascha Ansen, Stefan Haas, Yasushi Yatabe, Roman K. Thomas. Functional characterization of recurrent CD74-NRG1 fusions in lung adenocarcinoma. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 441. doi:10.1158/1538-7445.AM2014-441

Lung neuroendocrine tumors (LNETs) comprise small-cell lung cancer (SCLC), large-cell neuroendocrine tumors (LCNEC), and pulmonary carcinoids (PCA), and account for 25% of all lung cancer cases. The low 5-year survival rate of the highly aggressive LNETs (SCLC and LCNEC) combined with the lack of an effective treatment, suggest that understanding how these tumors arise and identifying therapeutic targets are unmet needs. We performed genome/exome, and transcriptome sequencing of 29 SCLC (Ref.1), 60 LCNEC, and 45 PCA to better understand their molecular origin and identify altered genes that may offer therapeutic opportunities. In contrast to SCLC and LCNEC, we found that RB1 and TP53 mutations were rare events in PCA suggesting that PCA are not early progenitor lesions of SCLC or LCNEC, but arise through independent mechanisms. Moreover, GSEA analysis showed that genes of the RB1 pathway were downregulated in SCLC but not in PCA. Our data also show that inactivation of chromatin-remodeling genes, specifically genes involved in histone methylation and subunits of the SWI/SNF complex, is sufficient to drive transformation in PCA. In a preliminary analysis of 15 LCNEC (Ref.2) we observed a predominance of mutations typical of SCLC, such as RB1, TP53, and CREBBP/EP300. In this larger series, we additionally found samples with mutations frequent in adenocarcinoma (AD) or squamous (SQ) lung cancer. We could then distinguish two well-defined groups of LCNEC: a SCLC-like group, carrying MYCL1 amplifications and mutations in both RB1 and TP53 genes; and an AD/SQ-like group, harbouring CDKN2A deletions, TTF1 amplifications, and frequent mutations in KEAP1 and STK11. Interestingly, RB1, STK11, and KEAP1 mutations happened in an almost mutually exclusive way. These data suggest that LCNEC might represent an evolutionary trunk that can branch to SCLC or AD/SQ, and also that LNETs and non-LNETs are not completely different entities. This is already suggested by the fact that one of the resistance mechanisms of EGFR-mutant adenocarcinomas to tyrosine-kinase inhibitors is through trans-differentiation to SCLC (Ref.3). Finally, we also identified new targetable driver genes in SCLC and LCNEC: FGFR1 amplifications were observed in 6% and 18% of the cases respectively; PTEN mutations were identified in 10% of the SCLC cases; and interestingly, one of the LCNEC samples (belonging to the SCLC-like group) harboured an activating RFWD2-NTRK1 fusion gene suggesting that fusions affecting NTRK1 may not only be a targetable opportunity for AD (Ref.4) but also for LCNEC and, based on the molecular similarities, also SCLC. (1) Peifer and Fernandez-Cuesta et al. Nat Genetics 2012 (2) The Clinical Lung Cancer Genome Project (CLCGP) and Network Genomic Medicine (NGM) Sci Transl Med 2013 (3) Sequist et al., Sci Transl Med 2011 (4) Vaishnavi et al. Nat Medicine 2013 Citation Format: Lynnette Fernandez-Cuesta, Martin Peifer, Xin Lu, Danila Seidel, Thomas Zander, Frauke Leenders, Luka Ozretić, Odd-Terje Brustugun, John K. Field, Gavin Wright, Benjamin Solomon, Reinhard Buettner, Christian Brambilla, Elisabeth Brambilla, Roman K. Thomas. Cross-entity mutation analysis of lung neuroendocrine tumors sheds light into their molecular origin and identifies new therapeutic targets. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1531. doi:10.1158/1538-7445.AM2014-1531