HIGHLIGHTED ARTICLE | INVESTIGATION The Hidden Genomic and Transcriptomic Plasticity of Giant Marker Chromosomes in Cancer Gemma Macchia,*,1 Marco Severgnini,† Stefania Purgato,‡ Doron Tolomeo,* Hilen Casciaro,* Ingrid Cifola,† Alberto L’Abbate,* Anna Loverro,* Orazio Palumbo,§ Massimo Carella,§ Laurence Bianchini,** Giovanni Perini,‡ Gianluca De Bellis,† Fredrik Mertens,†† Mariano Rocchi,* and Clelia Tiziana Storlazzi* *Department of Biology, University of Bari Aldo Moro, 70125 Italy, †Institute for Biomedical Technologies, Consiglio Nazionale delle Ricerche, Segrate, 20090 Italy, ‡Department of Pharmacy and Biotechnology, University of Bologna, 40126 Italy, §Laboratorio di Genetica Medica, Istituto di Ricovero e Cura a Carattere Scientifico, Casa Sollievo della Sofferenza, San Giovanni Rotondo, 71013 Italy, **Laboratory of Solid Tumor Genetics, Université Côte d’Azur, Centre Nationnal de la Recherche Scientifique, Institute of Research on Cancer and Aging in Nice, 06108 France, and ††Department of Clinical Genetics, University and Regional Laboratories, Lund University, 22100 Sweden ORCID IDs: 0000-0002-4323-6393 (H.C.); 0000-0001-6583-3482 (O.P.) ABSTRACT Genome amplification in the form of rings or giant rod-shaped marker chromosomes (RGMs) is a common genetic alteration in soft tissue tumors. The mitotic stability of these structures is often rescued by perfectly functioning analphoid neocentromeres, which therefore significantly contribute to cancer progression. Here, we disentangled the genomic architecture of many neocentromeres stabilizing marker chromosomes in well-differentiated liposarcoma and lung sarcomatoid carcinoma samples. In cells carrying heavily rearranged RGMs, these structures were assembled as patchworks of multiple short amplified sequences, disclosing an extremely high level of complexity and definitely ruling out the existence of regions prone to neocentromere seeding. Moreover, by studying two well-differentiated liposarcoma samples derived from the onset and the recurrence of the same tumor, we documented an expansion of the neocentromeric domain that occurred during tumor progression, which reflects a strong selective pressure acting toward the improvement of the neocentromeric functionality in cancer. In lung sarcomatoid carcinoma cells we documented, extensive “centromere sliding” phenomena giving rise to multiple, closely mapping neocentromeric epialleles on sep- arate coexisting markers occur, likely due to the instability of neocentromeres arising in cancer cells. Finally, by investigating the transcriptional activity of neocentromeres, we came across a burst of chimeric transcripts, both by extremely complex genomic rearrangements, and cis/trans-splicing events. Post-transcriptional editing events have been reported to expand and variegate the genetic repertoire of higher eukaryotes, so they might have a determining role in cancer. The increased incidence of fusion transcripts, might act as a driving force for the genomic amplification process, together with the increased transcription of oncogenes. KEYWORDS neocentromere; fusion transcript; WDLPS; LSC; gene amplification ENOME amplification is a frequent genetic alteration in While double minutes and homogeneously-staining regions Gcancer, with variable cytogenetic manifestations includ- have been described in a variety of cancer types (Matsui et al. ing double minutes, homogeneously-staining regions, and/or 2013), RGMs are particularly common in soft tissue tumors, ring and giant rod-shaped marker chromosomes (RGMs) notably in well-differentiated liposarcomas (WDLPS), and (Matsui et al. 2013; L’Abbate et al. 2014; Nord et al. 2014). were shown to contain amplified sequences from several chromosomes (Nord et al. 2014). During tumor progression, Copyright © 2018 by the Genetics Society of America the ring chromosomes are frequently broken and resealed, or doi: https://doi.org/10.1534/genetics.117.300552 transformed into rod-shaped markers, capturing the telo- Manuscript received September 1, 2017; accepted for publication December 11, 2017; published Early Online December 26, 2017. meres from other chromosomes (Nord et al. 2014). This in- Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10. stability results in a highly complex internal structure of these 1534/genetics.117.300552/-/DC1. 1Corresponding author: Department of Biology, University of Bari, 4 Via Orabona, 70125 markers, as well as in extensive heterogeneity with respect to Bari, Italy. E-mail: [email protected] size and number per cell (Garsed et al. 2014; Nord et al. Genetics, Vol. 208, 951–961 March 2018 951 2014). RGMs frequently lack functional centromeric alphoid identified in the primary cell cultures. This giant chromosome sequences and their mitotic stability is rescued by the emer- contained high-level amplification of chromosomal regions de- gence of perfectly functioning analphoid neocentromeres, riving from 10p to 12q and lacked a-satellite DNA (Pedeutour which might indirectly contribute to cancer progression et al. 2012). (Macchia et al. 2015). Nonetheless, there are few studies SNP array data addressing neocentromeres in cancer, probably because most of the technologies employed to study the tumor genotypes All cell lines were analyzed by the Affymetrix Genome-Wide are unable to unveil them. The occurrence of neocentromeres Human SNP Array 6.0 platform (Affymetrix, Santa Clara, CA), in cancer, therefore, could be more frequent than reported. as previously described (Storlazzi et al. 2010). Similarly, very little is known about the impact of neocentro- WGS meres on transcription, although centromeric satellite re- gions have been reported to produce noncoding transcripts WGS was carried out to disentangle the genomic architecture actively involved in the centromere assembly (Chan et al. of RGMs holding neocentromeres. Library preparations were 2012; Rosic et al. 2014; Quenet and Dalal 2015; McNulty performed using the TruSeqDNA Nano 350 bp protocol (Illu- et al. 2017). Also, genes within neocentromeres are still ac- mina, San Diego, CA). The sequencing data were acquired tively transcribed (Amor and Choo 2002; Wong et al. 2006). using the Illumina Xten at the New York Genome Center (New In line with these notions, the occurrence of neocentromeres York, NY), in a paired-end 150-cycle run (mean coverage in colon cancer cell lines was reported to correlate with large 403 per sample). Reads were aligned to the human reference DNase I hypersensitive sites, which are usually sites of active genome (GRCh37/hg19) using BWA-MEM (v.0.7.12) transcription or high nucleosome turnover (Athwal et al. (http://bio-bwa.sourceforge.net/; Li and Durbin 2009) and 2015). By combining chromatin immunoprecipitation (IP) PCR duplicates were removed using Picard (v.1.119) (http:// deep sequencing (ChIP-seq), whole-genome sequencing picard.sourceforge.net/). Candidate structural variations (WGS), immunofluorescence in situ hybridization (immu- (SVs) were identified using Delly (v. 0.5.9) and Crest no-FISH), whole-transcriptome sequencing (total RNA-seq), (v. 1.0) with default parameters (Wang et al. 2011; Rausch and other molecular analyses, we investigated in detail the et al. 2012). Copy number analysis was performed using genomic architecture of neocentromeres arising on RGMs, as Bayesian information criterion-sequation 0.7a (Xi et al. 2011), well as their contribution to transcription, in the lung sarco- and genomic intervals showing a log2 copyRatio . 0.5 and . matoid carcinoma (LSC) cell line 04T036 and in the three 2.5 were considered as amplified and highly amplified, respec- liposarcoma cell lines 93T449, 94T778, and 95T1000. Over- tively (Supplemental Material, File S1). all, our study uncovered the complex organization of neo- ChIP-seq centromeres in cancer, and shed light on the extraordinarily high genomic and transcriptomic plasticity associated with To determine the internal structure of the neocentromeres, RGMs in solid tumors. native ChIP-seq was performed as described (Wade et al. 2009). IP was run using a polyclonal antibody against the CENP-A (Trazzi et al. 2009). Both input and IP DNA frag- Materials and Methods ments were purified and processed using the TruSeq ChIP Library Preparation Kit (Illumina), and sequenced on the Tumor cell lines Illumina HiSequation 2500 at the Istituto di Genomica Appli- Four tumor cell lines (04T036, 93T449, 94T778, and cata (IGA) Technology Services facility (Udine, Italy) (single- 95T1000), kindly provided by The Centre Hospitalier Uni- end 100-cycle run, 140 M reads per sample). Raw reads were versitaire de Nice (France), were included in the study. aligned to the human reference genome (GRCh37/hg19) us- 04T036 was established from the LSC of a 50-year-old ing BWA-MEM (v 0.7.10). CENP-A-enriched regions corre- man. Cytogenetic and multicolor FISH analyses showed a sponding to putative neocentromeres were identified using near-triploid karyotype with numerous structural aberrations, the CNV (copy number variant)-seq tool, merging all over- four to six small RGMs containing chromosome 9 amplified lapping intervals (Xie and Tammi 2009). Selected regions sequences, and two RGMs containing chromosome 3 ampli- were then filtered to exclude alphoid sequences, weak en- fied sequences (Italiano et al. 2006). The 93T449 and 94T778 richments, and regions with read “spikes” piling-up in a single cell lines were obtained from a primary retroperitoneal position. Next, putative neocentromeric fragments were WDLPS at onset and at relapse,