**1. Introduction**

Chondrosarcoma is a malignant cartilage-forming tumor that represents, with approximately 25% of the cases, the second most common bone sarcoma. The most common subtype, representing 80% of the total, is the primary or conventional central chondrosarcoma which is characterized by the pathological formation of hyaline cartilage within the medullar cavity [1]. Other less common subtypes include the periosteal chondrosarcoma, which occurs in the surface of the bone, and secondary chondrosarcoma, which arises in benign lesions such as enchondroma, frequently associated to Ollier disease or Maffucci syndrome, or osteochondroma [1]. The dedifferentiated chondrosarcoma is a distinct variety of chondrosarcoma, accounting for approximately a 10% of the cases. This subtype is characterized by the presence of a low-grade well-differentiated cartilaginous tumor juxtaposed to a high-grade anaplastic sarcoma [1].

Within the complex cytogenetic scenario characteristic of chondrosarcomas, a few chromosomal alterations and gene mutations are frequently found. Thus, mutations in Isocitrate Dehydrogenase-1 *(IDH1*) and -2 (*IDH2*) are found in 87% of enchondromas, up to 70% of conventional central chondrosarcomas and 54% of dedifferentiated chondrosarcomas and may drive sarcomagenic processes [2–4]. In addition, mutations in the major cartilage collagen (*COL2A1*) or Tumor Supressor P53 (*TP53*) genes have been found in approximately 35% and 20% of the chondrosarcomas respectively [5,6]. Other frequent copy number variations, such as the amplification of the 12q13 region, containing Mouse Double Minute 2 homolog (*MDM2*) and the cyclin-dependent kinase 4 genes, and the deletion of the 9p21 region, which includes the Cyclin Dependent Kinase Inhibitor 2A (*CDKN2A*) locus, contribute to the deregulation of the p53 and Retinoblastoma (RB) pathways [5,7].

Wide surgical resection is the mainstay therapeutic option for localized chondrosarcomas. However, these tumors often show high local recurrence and metastatic potential. Furthermore, chondrosarcomas are inherently resistant to conventional chemo and radiotherapy and nowadays there are no effective treatments available for metastatic or inoperable tumors [8–12]. Proposed mechanisms of chemoresistance include the role of the cartilaginous extracellular matrix as a barrier for drug diffusion, the overexpression of members of the adenosine triphosphate (ATP) binding cassette transmembrane family of efflux pumps and antiapoptotic proteins, and the presence of a high percentage of low proliferating/quiescent cells [9,11,12]. Interestingly, some of these features are related to those described for tumor cell subsets presenting stem cell properties (cancer stem cells, CSC). These CSC subpopulations have been characterized in several subtypes of sarcomas and associated to the expression/activity of pluripotency factors, like Sex Determining Region Y-Box 2 (SOX2), stem cell markers, like Aldehyde Dehydrogenase 1 Family Member A1 (ALDH1), or to the ability to grow as floating clonal spheres (tumorspheres) [13,14]. CSC subpopulations have been barely characterized in chondrosarcoma and therefore new models amenable for the study of these subpopulations are needed to find vulnerabilities against these drug-resistant subpopulations.

Despite their failure to completely reproduce the genetic and microenvironmental conditions of tumors, cell lines are still indispensable models to investigate altered mechanisms in cancer, to study CSC subpopulations, and to serve as drug-screening platforms in a high-throughput and logistically simple and rapid way [15–17]. To our knowledge, 14 tumor-derived chondrosarcoma lines, corresponding to seven conventional, five dedifferentiated, and two secondary chondrosarcomas, have been reported so far [18–27]. In this work, we succeeded in establishing three cell lines from two secondary (CDS06 and CDS11) and one dedifferentiated (CDS-17) chondrosarcoma as well as another cell line derived from a CDS-17 xenograft (T-CDS17). We studied their tumorigenic and invasive potential, characterized the present CSC subpopulations, and analyzed the most relevant chondrosarcoma-related genetic alterations both in the cell lines and their original tumors. Furthermore, using a whole exome sequencing (WES) approach in CDS17 and T-CDS17 cells and their matched patient samples we were able to track the genomic adaptation of tumor cells to in vitro and in vivo growth.

#### **2. Experimental Section**

#### *2.1. Establishment of Cell Lines*

Human samples and data from donors included in this study were provided by the Principado de Asturias BioBank (PT17/0015/0023) integrated in the Spanish National Biobanks Network upon obtaining of written informed consent from patients. All experimental protocols have been performed in accordance with institutional review board guidelines and were approved by the Institutional Ethics Committee of the University Central Hospital of Asturias (approval number: 45/16). This study was performed in accordance with the Declaration of Helsinki.

Tumor samples were subjected to mechanical disaggregation followed by an enzymatic dissociation using MACS® Tissue Dissociation Kit and the GentleMACS Dissociator system (Miltenyi Biotec, Bergisch Gladbach, Germany). At the end of the incubation, culture medium (DMEM-Dulbecco's Modified Eagle Medium supplemented with 10% FBS, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomicin) was added and the cell suspension was filtered to remove clusters. Tumor cells were collected by centrifugation, resuspended in fresh culture medium and seeded in culture flasks. As an alternative protocol to derived cell lines, some fresh tumor specimens were cut into several small fragments, transferred to dry 25 cm2 culture flasks, covered with a drop of medium and incubated until outgrowth of tumor cells was observed. Cell cultures derived by both methods were subcultured when they reached 80–90% confluence. As a procedure to select tumoral cells and rid of stromal cells, we performed soft agar colony formation assays using the CytoSelectTM 96-Well Cell Transformation Assay Kit (Cell Biolabs Inc, San Francisco, CA, USA). Cells able to form colonies under these anchorage-independent growth conditions are supposed to be transformed. These colonies were recovered, left to attach to plastic substrate and grow in culture medium as normal adherent cultures. Short Tandem Repeat (STR) analyses were performed to compare the identity of cell lines with matched tumor sample (Supplemental Information).
