Establishment and characterization of NCC-DDLPS5-C1: a novel patient-derived cell line of dedifferentiated liposarcoma

Dedifferentiated liposarcoma (DDLPS) is a highly aggressive subtype of liposarcoma that is morphologically defined as a transition from a well-differentiated lipomatous component to a non-lipogenic one. Curative therapy for DDLPS is complete resection, and the benefits of current systemic chemotherapy remain marginal. Although DDLPS is molecularly characterized by co-amplification of MDM2 and CDK4 (12q14-15) and detailed genomic analyses have been conducted by multiple research groups, the effects of molecular targeted drugs are marginal, and novel therapeutic modalities are required. Although patient-derived cell lines are pivotal for cancer research, no DDLPS cell lines are currently available from public cell repositories. Accordingly, in this study, we established a novel DDLPS cell line, NCC-DDLPS5-C1, using surgically resected tumor tissues from a patient with DDLPS. NCC-DDLPS5-C1 cells exhibited typical gene amplification, overexpression of MDM2 and CDK4, and other DNA copy number alterations. The NCC-DDLPS5-C1 cells were capable of rapid cell proliferation, aggressive invasion, and spheroid formation, but not tumor formation in mice. We reported the utility of NCC-DDLPS5-C1 cells for a drug–response assay to detect anticancer drugs that significantly attenuated cell proliferation. Thus, we concluded that the NCC-DDLPS5-C1 cell line could be a useful resource for the study of DDLPS. Considering the diversity of disease in terms of clinical outcomes, continuous efforts are required to develop more patient-derived cancer models with different clinical and pathological backgrounds.


Introduction
Dedifferentiated liposarcoma (DDLPS) is a highly aggressive subtype of adipocytic sarcoma. Histologically, DDLPS is characterized by transient features from recurrent atypical lipomatous tumors or well-differentiated liposarcomas to highly cellular regions of high-grade pleomorphic spindle cell components [1]. The incidence of DDLPS is less than 0.1 per million individuals each year [2], and the disease occurs in approximately 10% of cases of intermediate (locally aggressive) well-differentiated liposarcoma [3]. Thus, DDLPS is considered a rare cancer. DDLPS often occurs in patients who are 50-60 years old, with the retroperitoneum and deep extremities being the typical site of origin [1]. DDLPS is particularly refractory to conventional chemotherapy, similar to well-differentiated liposarcomas [4,5]. Surgical excision is the first-line treatment for DDLPS, and even after complete resection for curative intent, local recurrence is observed in over 40% of cases with DDLPS, whereas distant metastasis is detected in 15-20% of cases. Accordingly, the prognosis of patients with DDLPS remains poor; the 5-year mortality rate is approximately 30%, and the mortality rate is higher after 10-20 years [1,4,6,7].
Patient-derived cell lines are pivotal bioresource for translational research and for novel therapies, because they allow the evaluation of novel anticancer drugs. Recently, a considerable amount of genomic and pharmacological data has been obtained and integrated into large-scale cell panels [15][16][17][18][19][20][21]. Analysis of these data could help to identify effective anticancer agents and their predictive biomarkers [22,23]. However, very few sarcoma-derived cell lines were included in these studies. For example, only 28 soft tissue sarcoma-derived cell lines were used in large-scale drug screening studies, accounting for only 2.2% of all 1300 cell lines examined [24]. According to a previous review of cell lines recorded in the cell line database, Cellosaurus [25], multiple cell lines have been reported only for a few types of sarcomas, and no DDLPS-derived cell lines are available from public cell repositories [26]. Thus, this paucity of adequate cell lines has hindered research on sarcomas.
To address these issues, we established sarcoma cell lines using surgically resected tumor tissues [27,28]. Previously, we reported four DDLPS cell lines [29][30][31][32]. However, because of the diversity of disease in terms of clinical outcomes, more cell lines from different patients with DDLPS are needed. Accordingly, in this study, we report the establishment of a novel DDLPS cell line, NCC-DDLPS5-C1, using a surgically resected specimen from a patient with DDLPS. We characterized NCC-DDLPS5-C1 cells and demonstrated their utility for drug screening.

Patient history
The patient was a 77-year-old man with DDLPS. He reported a growing mass in the left inguinal region. He was diagnosed with inguinal hernia at a previous hospital; however, the mass was found to be a fatty tissue during surgery, and marginal resection was performed to remove the lipoma. Five months after surgery, a mass recurred in the same region, and further inspection was performed. Magnetic resonance imaging revealed a delineated mass (Fig. 1a, b), and the patient was referred to the National Cancer Center Hospital (Tokyo, Japan) based on suspicion of liposarcoma. Needle biopsy indicated DDLPS, and wide resection, including left testiclectomy, was performed without any preoperative chemotherapy or radiotherapy. Microscopic observations revealed that the tumor showed dense proliferation of mitotically active spindle cells (Fig. 1c). The tumor cells were immunohistochemically positive for MDM2 (Fig. 1d). A well-differentiated liposarcoma component was focally observed in the periphery (Fig. 1e), confirming the diagnosis of dedifferentiated liposarcoma. Part of the resected tumor was used to establish the cell line described in this study. The ethical committee of the National Cancer Center approved the use of clinical materials for this study, and written informed consent was obtained from the donor patient.

Histological analysis
For histopathological observations, we prepared paraffinembedded tumor tissues and sectioned them into 4-μm-thick sections. The sections were then deparaffinized and stained

Immunohistochemistry
We conducted immunohistochemical analyses are described in our previous reports [29][30][31][32] using the 4-μm-thick sections prepared as described above. Before the antibody reaction, the sections were exposed to 3% hydrogen peroxide for 15 min and then subjected to heat-induced epitope retrieval. The sections were incubated with primary antibodies against MDM2 (IF2, 1:100; Zymed Laboratories, San Francisco, CA, USA) and labeled with peroxidase (EnVision system; Dako, Santa Clara, CA, USA).

Authentication and quality control of the established cell line
Authentication was achieved by examining short tandem repeats (STRs) at 10 loci using the GenePrint 10 system (Promega, Madison, WI, USA) according to our previous studies [29][30][31][32]. We confirmed the absence of mycoplasma contamination by examining the DNA present in the cells, as previously reported for DDLPS [29][30][31][32].

Genetic analysis
To observe DNA copy number alterations, we conducted single nucleotide polymorphism (SNP) array genotyping using the Infinium OmniExpressExome-8 v1.4 BeadChip (Illumina, San Diego, CA, USA) following the manufacturer's instructions and the procedure described in our previous studies [29][30][31][32]. The SNP array data were analyzed using R version 4.0.3 (R Foundation for Statistical Computing, http:// www.R-proje ct. org) and DNAcopy package version 1.64.0 (Bioconductor, https:// bioco nduct or. org/). When the chromosome regions had copy numbers greater than 3 or less than 1, they were considered amplifications and losses, respectively. Genes with copy number alterations were annotated using the biomaRt package version 2.46.0 (Bioconductor) and "Cancer Gene Census" in the Catalogue Of Somatic Mutations In Cancer database (GRCh 37 v91), as previously reported [29][30][31][32].

Cell proliferation assay
An xCELLigence RTCA-DP device (Agilent, Santa Clara, CA, USA) was employed to monitor cell proliferation in real time. Electronic microtiter plates were seeded with 5.0 × 10 3 cells (E-Plate; Agilent), and proliferation was measured for 170 h using the xCELLigence system according to the manufacturer's instructions. Cell density measurements were performed in quadruplicate with programmed signal detection every 20 min. Data acquisition and analyses were performed using RTCA software (version 1.2, OLS).

Invasion assay using a real-time cell analyzer
Invasion potential was evaluated using a real-time cell analyzer (xCELLigence; Agilent, Santa Clara, CA, USA) according to the manufacturer's instructions and the procedure described in our previous studies [29][30][31][32]. Briefly, the established cells and MG63 osteosarcoma cells [24] (JCRB; Ibaraki Osaka, Japan) were seeded into upper chambers coated with Matrigel at a density of 4 × 10 4 cells/chamber. The tissue culture medium, which was used to maintain the cells, was poured to the lower chamber. Thereafter, the cells on the Matrigel-coated membrane migrated to the bottom chamber and adhered to the electronic sensors on the underside of the membrane. The attached cells influenced the electrical impedance of the electronic sensors, and the invasion capability of the cells was estimated based on the positive correlation between the impedance and the number of cells. Impedance was monitored and plotted as a function of time after seeding.

Assessment of tumorigenicity in nude mice
Animal experiments were performed in compliance with the guidelines of the Institute for Laboratory Animal Research, National Cancer Center Research Institute. Female BALB/c nude mice from CLEA Japan, Inc. (Tokyo, Japan) were housed under specific pathogen-free conditions, with food and water available ad libitum under a 12-h light/dark cycle. We mixed 100 μL cells with Matrigel (BD Biosciences) and subcutaneously injected the mixture into mice (1 × 10 6 cells). Subsequently, tumor size was measured weekly using digital calipers, and tumor volume was calculated using the following formula: volume = (length × width 2 )/2. After 2 weeks, the tumors were resected and stained with H&E.

Screening for the antiproliferative effects of anticancer reagents
The inhibitory effects of anticancer agents on cells proliferation were monitored as previously reported [29][30][31][32]. A list of anticancer agents is provided in Supplementary Table 1. We performed concentration-response experiments to validate the available hits in the pilot screening and determined the sample concentration required to inhibit cell growth by 50% in comparison with the growth of control cells (IC 50 ) from concentration-response curves.

Authentication of the established cell line
A cell line was established from the tumor tissue of a patient with DDLPS; the cell line was designated NCC-DDLPS5-C1. The STRs of the cells and their original tumors were analyzed. We found that the STR allele pattern of NCC-DDLPS5-C1 was almost identical to that of the corresponding original tumor tissue (Table 1), and the STR pattern of NCC-DDLPS5-C1 cells did not match that of any cell line in the public cell banks. Thus, we concluded that NCC-DDLPS5-C1 was a novel DDLPS cell line. We did not detect the DNA sequence unique to mycoplasma in the tissue culture medium of NCC-DDLPS5-C1 cells (data not shown).

Characterization of the cell line
We performed SNP array analysis of multiple allelic duplications in NCC-DDLPS5-C1 cells (Fig. 2, Supplementary Table 2, 3 and 4). Characteristic amplification of the 12q13-15 region, which included MDM2 and CDK4, was observed in this study. Western blot analysis confirmed the overexpression of MDM2 and CDK4 in NCC-DDLPS5-C1 cells (Fig. 3a-c). The original images are displayed in Supplementary Fig. 2. These findings demonstrated that the NCC-DDLPS5-C1 cell line maintained the characteristics of DDLPS.

Tumorigenesis in nude mice
NCC-DDLPS5-C1 cells inoculated into BALB/c nude mice did not form tumors.  (2) 8, 9 8, 9 Sensitivity to anticancer drugs The inhibitory effects of 214 anticancer agents on NCC-DDLPS5-C1 cell proliferation were examined at a fixed concentration of 10 µM (Supplementary Table 5). The anticancer agents showing marked inhibitory effects on proliferation were further examined, and their IC 50 values were calculated ( Table 2). Growth curves for two anticancer drugs, bortezomib and romidepsin, are shown in Fig. 5.

Discussion
DDLPS is a highly aggressive type of liposarcoma that is refractory to conventional chemotherapy and radiotherapy. Recurrence and metastasis are often observed, even after complete surgical resection with curative intent, and the prognosis remains poor. Therefore, novel therapeutic strategies are required to treat DDLPS. Advancements in modern technology have enabled large-scale genomic studies combined with drug screening using a number of cell lines, and molecular targeted drugs, such as MDM2 inhibitors, have been developed in recent years. To facilitate the development of novel therapies, we established a patient-derived cancer model in DDLPS. NCC-DDLPS5-C1 cells showed typical amplification of MDM2 and CDK4, and their overexpression was also confirmed at the protein level. Therefore, NCC-DDLPS5-C1 cells could be suitable for investigating molecular mechanism of tumorigenesis. Because NCC-DDLPS5-C1 cells showed rapid proliferation and aggressive invasion ability, Single nucleotide polymorphism (SNP) array was used for allele-specific copy number analysis, and DNA copy number variations were determined for A the original DDLPS tumor and B NCC-DDLPS5-C1 cells. Note that the major alterations were consistent between the original tumor and the tumor cells, which were maintained under tissue culture conditions. The X-and Y-axes indicate the chromosome number and the log ratio of copy number, respectively. The passage number of NCC-DDLPS5-C1 was 21 in this experiment We found that NCC-DDLPS5-C1 cells were markedly sensitive to two anticancer drugs, bortezomib and romidepsin. We previously established four DDLPS cells lines, i.e., NCC-DDLPS1-C1, NCC-DDLPS2-C1, NCC-DDLPS3-C1, and NCC-DDLPS4-C1, which also exhibited sensitivity to romidepsin [29][30][31][32]. The drugs which showed the sensitivity to DDLPS cell lines are summarized in Supplementary  Table 6, suggesting that the cell lines exhibited the similar but distinct response to drug treatments. The overall features of drug response should be integrated with the genetic and proteomic alterations, and such studies should be performed using numerous numbers of cell lines from DDLPS with different clinical and pathological backgrounds. The growthsuppressive effects of romidepsin in DDLPS have not been examined in clinical trials; therefore, its clinical utility in the treatment of DDLPS should be established in the future studies.
Considering the diversity of disease in terms of patient outcomes, additional cell lines from patients with different backgrounds should be established. However, it is unclear how many cell lines are required for the development of treatments for DDLPS. Phenotypic variations will be observed when additional cell lines are established and characterized, and we will need to have more cell lines until when the variations do not increase. In general, several dozens of patients should be included in clinical trials, and a similar number of cell lines may be required to obtain conclusive results. This study has five limitations that should be mentioned. First, we established a cell line from a single patient with DDLPS, and we need to obtain more cell lines so that we can obtain the data with more conclusive. Second, we did not clone the cells, and multiple different types of tumor cells should be included in NCC-DDLPS5-C1. Although NCC-DDLPS5-C1 may maintain the heterogeneity of tumor cells, which is observed in the original tumor tissues, the characters will be changed after the long passaging. Third, the NCC-DDLPS5-C1 cells did not form tumor when they were inoculated in the mice. As DDLPS is a highly aggressive sarcoma, NCC-DDLPS5-C1 may have a potential to survive and proliferate in the mice. We may need to challenge the different types of immune-deficient mice to promote the formation of tumor of NCC-DDLPS5-C1 in the mice. Fourth, it is not clear from which portion of tumor tissue the NCC-DDLPS5-C1 was established. DDLPS is defined as a transition from a welldifferentiated lipomatous component to a non-lipogenic one, which included morphologically different types of tumor cells. The identification of precious histological origin of established cell lines is generally difficult, and to integrate the genomic and proteomic data with the characters of cell lines, it is crucial to identify the origin of cell line. Lastly, different in vitro models, such as organoids, should be developed, and complimentary used, so that we will have more reliable patient-derived cancer models.

Ethics approval
The ethical committee of the National Cancer Center approved the use of clinical materials for this study (approval number 2004-050).
Informed consent Written informed consent for publication was provided by the patient.