Establishment of a novel experimental system using single cell-derived pleomorphic rhabdomyosarcoma cell lines expressing K-RasG12V and deficient in p53

Abstract

Pleomorphic rhabdomyosarcoma (PRMS) predominantly arises in adult skeletal musculature and is usually associated with poor prognosis. Thus, effective treatments must be developed. PRMS is a rare tumor; therefore, it is critical to develop an experimental system to understand the cellular and molecular mechanisms of PRMS. We previously demonstrated that PRMS develops after p53 gene deletion and oncogenic K-Ras expression in the skeletal muscle tissue. In that study, oncogenic K-Ras-expressing cells were diverse and the period until disease onset was difficult to control. In this study, we developed an experimental system to address this problem. Single cell-derived murine cell lines, designated as RMS310 and RMSg2, were established by limiting the dilution of cells from a lung metastatic tumor colony that were positive for various cancer stem cells and activated skeletal muscle-resident stem/progenitor cell marker genes by RT-PCR. All cell lines stably recapitulated the histological characteristics of human PRMS as bizarre giant cells, desmin-positive cells, and lung metastases in C57BL/6 mice. All subclones of the RMSg2 cells by the limiting dilution in vitro could seed PRMS subcutaneously, and as few as 500 RMSg2 cells were sufficient to form tumors. These results suggest that the RMSg2 cells are multipotent cancer cells that partially combine the properties of skeletal muscle-resident stem/progenitor cells and high tumorigenicity. Thus, our model system’s capacity to regenerate tumor tissue in vivo and maintain stable cells in vitro makes it useful for developing therapeutics to treat PRMS.

Introduction

Rhabdomyosarcoma (RMS) is an aggressive and relatively rare soft-tissue tumor. The World Health Organization classifies RMS into four subtypes, namely, pleomorphic, embryonal, spindle cell/sclerosing, and alveolar, each with differing histological, genetic, and clinical features [1]. Pleomorphic RMS (PRMS) is rare and arises predominantly in large adult muscles of the extremities, and it accounts for 3% of soft tissue sarcomas. PRMS is highly aggressive, usually metastasizes within five years of diagnosis, and presents 12.5 to 50% survival rates, resulting in 1-year to 20-month disease-free survival [2,3,4]. Therefore, affected patients and their families await the development of effective treatments for this disease.

Previously, we demonstrated the importance of RAS pathway activation and p53 deletion for PRMS development [5]. Recently, Chelsky et al. demonstrated that molecular profiling of PRMS resection cases involved alterations in the genes involved in the activation of the RAS/MAPK pathway and biallelic inactivation of TP53 [6], consistent with our study.

Traditional cancer therapies, such as radiotherapy and chemotherapy, successfully target proliferating cells to reduce the tumor burden; however, these therapies fail to prevent tumor progression. One of the reasons for this is the failure to eradicate cancer stem cells (CSCs), which belong to a small subset of cells within a tumor population and drive the progression and therapy resistance of cancer. Moreover, they can recapitulate the characteristics of the tumor from which they are derived [7,8,9]. Therapies targeting CSCs are being developed to improve cancer therapies and eliminate cancer from patients [10]. There are only a few reports on PRMS cell lines [11], and the CSCs of PRMS remain unclear, indicating that establishing PRMS tumorigenic cell lines containing CSCs is an urgent matter.

One week after electroporation for the induction of the oncogenic K-Ras gene, the site where PRMS was induced in muscle consisted of a proliferating population positive for the Sca-1 antigen, which is a stem cell marker. Sca-1 positive cells accounted for 6% of the total tumor cells [5], suggesting that CSCs originated from skeletal muscle-resident stem/progenitor cells [7, 12,13,14,15,16,17,18]. Several groups have used various biomarkers to identify CSCs in solid tumors, including breast, brain, thyroid, melanoma, colon, pancreatic, liver, prostate, lung, head and neck, ovarian, and stomach cancers [19]. In this study, we examined CSCs in PRMS using a gene expression microarray analysis and found that tumors expressed various specific genes in CSCs.

P53 knockout mice developed RMS between 4 and 10 weeks after oncogenic K-ras induction in adult skeletal muscle [5]. However, the timing of PRMS development of the mouse cannot be controlled. For this reason, it was challenging to obtain mice that developed symptoms simultaneously. We could easily obtain a model mouse that met the experimental conditions by establishing a PRMS cell line with tumorigenicity.

As the introduction of the Cre recombinase gene is widespread in muscle cells, the cell line established from the primary tumor might be a mixture of tumor cells of various origins [5]. By obtaining a single-cell-derived cell line capable of forming PRMS, it is possible to obtain cells of a single origin that form PRMS.

To clarify and overcome these limitations, we attempted to establish PRMS cell lines by lung metastasis and limiting dilution.

Here, we established an experimental system that is useful for in vivo and in vitro research on PRMS using single-cell-derived PRMS cell lines with oncogenic K-Ras expression and p53 deficiency.

Materials and Methods

Animals

All animals were cared for according to the ethical guidelines established by the Institutional Animal Care and Use Committee of Mie University (Permit Number: 21-19, 2021-14). Ryr2^tm1Nobs mice (a mixed C57BL/6J and 129/SV background) [5] were backcrossed with C57BL/6J Jms Slc mice, obtained from Japan SLC Inc. (Shizuoka, Japan), over 12 generations.

Establishment of RMS3 cells from primary tumor

We isolated the primary tumor mass 10.5 weeks after K-RasG12V-induction in the femur muscle cells of an 8-week-old male mouse (Ryr2^{tm1Nobs/tm1Nobs}; p53^−/−) and cultured tumor-derived cells as described over ten passages [5] and named them RMS3.

Production and infection of GFP retrovirus

For GFP expression, we used a pMX-GFP retroviral vector [20]. For cell culture, basal medium (DMEM supplemented with 10% FCS, 100 µg/ml penicillin/streptomycin, and 2 µM L-glutamine) was used. An infectious viral solution was prepared according to a previously described protocol [21]. Briefly, 70% confluent PLAT-E cells were incubated with a precipitated FuGene6 transfection reagent (Promega, Madison, WI, USA) and saturated amounts of pMX-GFP plasmid DNA for 48 h. For GFP gene transduction, we added the virus solution to the 70% confluent cell culture of RMS310 for 48 h and then replaced it with the fresh basal medium.

Cloning cells via experimental metastatic colony formation and in vitro limiting dilution

We cloned RMS310 from an RMS3-derived lung metastatic tumor colony and serially cloned GFP-positive RMSg2 cells from an RMS310-derived lung metastatic tumor colony. We dispersed a monolayer of RMS3 or GFP virus-infected RMS310 cells with 0.05% trypsin-EDTA and injected 1 × 10⁴ single-cell suspensions in 200 µl of PBS into the orbital sinus of C57BL/6 mice. Two weeks after injection, we isolated and cultured cells from a single lung tumor mass, as described previously [5], and seeded them at 100 cells/dish in a 10 cm culture dish. Then, we picked single-cell colonies under a microscope and cultured them for ten passages.

Orbital sinus injection experimental metastasis and subcutaneous injection spontaneous metastasis model

We injected 1 × 10⁵ and 1 × 10⁶ RMSg2 cell suspensions in 200 µl PBS into the orbital sinus and subcutaneous tissue, respectively. Twelve days after injection, we evaluated metastatic lung colonies using a fluorescence micro-stereoscope (Nikon, Tokyo, Japan).

Histological analysis

Histological analysis was performed as previously described [22]. The following antibodies were used: rat anti-Sca-1 (1:50; BD Pharmingen, Boston, MA, USA, Cat No. 553333), rabbit anti-desmin (1:200; Santa Cruz Biotechnology, Dallas, TX, USA, Cat No sc-14026), HRP-conjugated goat anti-rat IgG (1:200; SouthernBiotech, Birmingham, AL, USA, Cat No. 3050–05), and HRP-conjugated goat anti-rabbit IgG (1:200; Santa Cruz Biotechnology, Cat No sc-2004).

Microarray analysis

Total RNA was extracted from the tumors as previously described [22]. We labeled cRNA and hybridized it with a microarray (Whole mouse Genome DNA 4 × 44 K, Agilent, Santa Clara, CA, USA) using Quick Amp Labeling and Gene Expression Hybridization- Kits (Agilent) according to the manufacturer’s recommendations.

RT-PCR analysis

Total RNA was extracted as previously described [22]. We synthesized cDNA using ReverTra Ace qPCR RT master mix with gDNA remover (Toyobo, Osaka, Japan). The PCR conditions were as follows: one cycle at 95°C for 2 min; 35 cycles at 98°C for 10 s, 60°C for 30 s, and 68°C for 1 min; and one cycle at 68°C for 5 min using KOD FX DNA polymerase (Toyobo). PCR products were electrophoresed and visualized by ethidium bromide staining. The primer sets are listed in Supplementary Table 1.

Adipocyte differentiation assay in vitro

Cells were seeded at a density of 1 × 10⁵ cells/well in a 24-well plate and cultured in basal medium for one day. The medium was then replaced with a basal medium supplemented with 1 µg/ml insulin, 1 µM dexamethasone, and 0.5 mM 3-isobutyl-1-methylxanthine for two days. After that, we replaced the medium with 1 µg/ml insulin supplemented basal medium for five days and refreshed it every day. 3T3-L1 cells were obtained from the Japan collection of research bioresources (JCRB) cell bank (Osaka, Japan).

Oil Red O staining

Oil Red O staining of cultured cells was performed as previously described [23]. The tumor mass was isolated and frozen section slides were prepared as previously described [5]. The slides were washed with PBS and incubated in 60% isopropanol at room temperature for 1 min. The slides were then incubated in 0.18% Oil Red O/ 60% isopropanol at 37°C for 15 min and placed in 60% isopropanol for 2 min. They were then rinsed in distilled water and incubated in a hematoxylin solution for 5 min. After 2 min of washing in tap water, the cells were mounted using 50% glycerol/PBS and visualized under a microscope (Carl Zeiss Microscopy, Jena, Germany).

Statistical analysis

Non-paired Student’s t-test was used to analyze the significance of the differences between the two groups. Statistical significance was set at P<0.05.

Results

Establishment of single cell-derived PRMS cell lines

We established an RMS3 cell line from a primary tumor mass obtained from a K-RasG12V-induced mouse (^{RyR2tm1Nobs/tm1Nobs;p53−/−}) [5]. RMS3 cells were almost all positive for Sca-1 (Figs. 1A and B), suggesting that the Sca-1 positive population, which accounted for 6% of the tumor mass, could selectively proliferate and maintain stemness under culture conditions. Furthermore, the absence of desmin supported the undifferentiated state of RMS3 cells in the 2D culture (Figs. 1C and D). RMS3 cells show tumorigenicity and high lung metastasis potential upon transplantation into C57BL/6 mice. Tumors derived from RMS3 recapitulated the morphology and immunoreactivity indistinguishable from the original PRMS tissues (Figs. 1E, F, I–L and data not shown). We obtained two other lines with two consecutive clones, RMS310 and RMSg2, because RMS3 could be a mix of cells of multiple origins. We cloned RMS310 by limiting the dilution of cells derived from a single tumor cell-origin lung metastatic colony derived from RMS3. GFP-expressing virus-infected RMS310 cells produced GFP-positive metastatic lung colonies. From one of these colonies, we established RMSg2.

Fig. 1.

Histological and immunostaining analysis of pleomorphic rhabdomyosarcoma (PRMS) cell lines in vivo and in vitro. (A, B) Immunostaining of RMS3 cells cultured on a flat dish with anti-Sca-1 antibodies (A) and negative control without 1st antibodies (B). (C, D) Immunostaining of RMS3 cells cultured on a flat dish with anti-desmin antibodies (C) and negative control without 1st antibodies (D). (E–H) Hematoxylin and eosin-stained sections of tumor tissues from model mice (E) and subcutaneous transplantation of established PRMS cell lines: RMS3 (F), RMS310 (G), and RMSg2 (H). Arrow heads indicate bizarre giant cells. (I, J) Immunostaining of tumor tissues of model mice (I) and subcutaneous transplanted RMS3 cells (J) with anti-desmin antibodies counterstained with hematoxylin. (K, L) Immunostaining of tumor tissues of model mice (K) and subcutaneous transplanted RMS3 cells (L) with anti-Sca-1 antibodies counterstained with hematoxylin. (M) Whole-mount view of lungs from C57BL/6J mice at 12 days after orbital sinus injection of RMSg2 cells. Arrow heads indicate lung metastatic colonies. (N) Whole-mount view of GFP fluorescent lungs from C57BL/6J mice at 12 days after orbital sinus injection of RMSg2 cells. (O) Hematoxylin and eosin-stained section of lungs from C57BL/6J mice at 12 days after orbital sinus injection of RMSg2 cells. Arrow heads indicate bizarre giant cells. (P) Whole-mount view of lungs from C57BL/6J mice at 12 days after subcutaneous injection of RMSg2 cells. (Q) Whole-mount view of GFP fluorescent lungs from C57BL/6J mice at 12 days after subcutaneous injection of RMSg2 cells. (R) Hematoxylin and eosin-stained section of lungs from C57BL/6J mice at 12 days after subcutaneous injection of RMSg2 cells. Scale bar: 100 µm for A–L, 2.5 mm for M and P, 400 µm for N and Q, and 100 µm for O and R.

Recapitulation of PRMS in mice by transplantation of cell lines with CSC features

Tumors derived from RMS310 and RMSg2 also recapitulated the morphology and immunoreactivity indistinguishable from the original PRMS tissues (Figs. 1E–H and data not shown), suggesting that RMS3 contained PRMS CSCs and that RMS310 and RMSg2 originated from a single cell. These results showed that the cell lines maintained their original CSC features even after serial cloning. To evaluate the self-renewal capacity of RMSg2 cells, we calculated the plating efficiency of a single RMSg2 cell and tumorigenicity of the generated subclones by limiting dilution. Ninety percent (n=44/49) of the single-cell cultures proliferated. Ten randomly selected subclones formed 100% of the tumorigenic cells by subcutaneous transplantation (Table 1). Furthermore, we demonstrated that as few as 500 RMSg2 cells formed tumors and the histology of the tumors was identical to that of the primary tumor (Table 1 and data not shown). Therefore, we concluded that the RMSg2 cell line is a CSC cell line of PRMS.

Table 1. CSCs feature and tumorigenicity of RMSg2 cells.

Self-renewal capacity of RMSg2 cells	Plating efficiency	89.8% (44/49)
	Tumorigenic efficiency of subclones	100% (30/30)
Tumor formation ability of RMSg2 cells	injected cell number	5000	1000	500	250	100
	Tumors/injections	3/4	1/4	1/4	0/4	0/4

Experimental and spontaneous metastatic ability of RMSg2 cells

We tested the ability of RMSg2 cells to induce experimental and spontaneous lung metastases. RMSg2 cells injected in the orbital sinus (n=5) and subcutaneously (n=3) generated 131.8 ± 94.7 and 7.3 ± 2.3 lung metastatic colonies, respectively. The size of the colonies from the subcutaneous tumors (317 ± 178 µm; n=14) was smaller than that of the tumor cells injected into the orbital sinus (584 ± 355 µm; n=64) (P=0.0076) (Figs. 1M–R). This suggests spontaneous lung metastases occurred in the transplanted region after tumor growth. The histology of the metastatic colonies was indistinguishable from that of the primary tumor (Figs. 1A, O, and R). Thus, our system can be useful for studying metastases, which is a symptom of PRMS.

Expression of CSC and skeletal muscle-resident stem/progenitor cell marker genes in the tumor cell lines of the system

Next, we surveyed the cell marker genes of CSCs [19], skeletal muscle-resident stem/progenitor cells, FAPs, PICs, and SCs [12, 15,16,17,18] by mRNA microarray analysis of tumors derived from RMS3 (n=2; accession No. GSE203596). The results showed that most CSC marker genes were positive except CD133, as shown in Supplementary Tables 2 and 3. Furthermore, quiescent or activated skeletal muscle-resident stem/progenitor cells were positive for all marker genes except Pax7 and Myf5. These results indicate that cells expressing these markers were present in the tumor. We then confirmed their expression in the three PRMS cell lines using RT-PCR (Fig. 2 and Table 2). All CSC markers except for EpCAM were present in all cell lines. Tumor cell lines were also positive for all FAP markers. Except for Pax7, Myf5, MyoD and PW1, PIC and SC markers were present in all tumor cell lines. Therefore, the tumors of our mouse model were derived from cells with CSCs and activated skeletal muscle-resident stem/progenitor-like features in terms of gene expression.

Fig. 2.

RT-PCR analyses of established pleomorphic rhabdomyosarcoma (PRMS) cell lines. (A) RT-PCR analysis of cancer stem cell marker gene expression in 2D cultured RMS3, RMS310, and RMSg2 cells. Positive control of EpCAM and CD133 was used for colon and kidney cells of 8-week-old mice, respectively. (B) RT-PCR analysis of myogenic regulatory factor, SC, FAP, and PIC marker gene expression in 2D cultured RMS3, RMS310, and RMSg2 cells. Positive controls included embryos at 13.5 days post coitum (d.p.c.). 3, RMS3; 310, RMS310; g2, RMSg2; C, colon; K, kidney.

Table 2. RT-PCR analysis of cancer stem cells (CSCs)-, FAPs-, PICs-, and SCs- marker gene expression in PRMS cells.

CSCs	+	CD44	CD24	ALDH	CD166	CD34	CD38	CD49f	CD90
		CD87	ABCG2	ABCB5	CD20	Itga2	Itgb1
	−	CD133	EpCAM
FAPs	+	Sca1	PDGFRa	CD34	Adam12	Tcf4,	Ors1*	PPARγ*	C/EBPα*
PICs	+	Sca1	PDGFRa	CD34	Pax3*
	−	PW1	Pax7*	MyoD*
SCs	+	Pax3	Cav1	Vcam1	CD34	Myf6*
	−	PW1	Pax7	Myf5*	MyoD*

+, positive signal; −, no signal; *, marker genes of activated state.

Differentiation of RMSg2 cells into adipogenic cells in vivo and in vitro

As shown in Fig. 3A, RMSg2 cells generated PRMS-containing adipogenic cells after subcutaneous injection. As shown in Fig. 3B, RMSg2 cells differentiated into adipogenic cells in vitro under specific conditions. 3T3-L1 cells, which were used as the positive control, presented over 90% differentiation into adipogenic cells under the same conditions (Fig. 3C). Furthermore, PRMS cell lines expressed active FAP marker genes (Fig. 2B). These results demonstrated that PRMS cell lines can initiate adipogenesis under appropriate conditions.

Fig. 3.

Adipogenic differentiation of the pleomorphic rhabdomyosarcoma (PRMS) cell lines. (A) Fluorescent GFP images and Oil Red O staining analysis of sections of subcutaneously transplanted RMSg2 tumor tissues. (B) Oil Red O staining analysis of adipogenic differentiated g2 cells. (C) Oil Red O staining analysis of adipogenic differentiated 3T3-L1 cells. Scale bar: 200 µm for A.

Discussion

RMSs are classified histologically into mainly three types: embryonic (ERMS) and alveolar (ARMS) in early childhood and pleomorphic (PRMS), predominantly in adults over 40 years of age. PRMS occurs primarily in the large muscles of the extremities. They are predominantly chemoresistant and have poor survival [24]. Furthermore, they are rare, with limited data for the development of treatment. Therefore, research into therapeutics for PRMS requires the development of animal models. RMS is highly aggressive and metastasizes. Previously, Almacellas-Rabaiget et al. and Codenotti et al. reported a spontaneous lung metastasis model by xenograft of human ARMS cell lines and an experimental lung metastasis model by xenograft of human ERMS cell lines in immunodeficient mice, respectively [25, 26]. In this study, we established an experimental system that can recapitulate lung metastasis of PRMS for the first time. Furthermore, the GFP-expressing RMSg2 cell line could be used in both spontaneous and experimental metastasis models in immunologically normal mice.

The established cell lines expressed the characteristic marker genes, ABCG2, ABCB5, and ALDH. Both ABCG2 and ABCG5 belong to the ATP-binding cassette (ABC) transporter family. ABCG2 and ABCB5 impart resistance to 5-fluorouracil, mephedrone, cisplatin in liver CSCs and resistance to doxorubicin in circulating melanoma cells [27, 28]. ALDH is responsible for resistance to several chemotherapeutic agents in adenocarcinoma and breast cancer [29, 30]. Therefore, these ABC transporters and ALDH may be involved in chemotherapy resistance of PRMS.

Our PRMS cell lines displayed an adipocyte lineage similar to that of Sonic Hedgehog-driven ERMS [31] and expressed the gene set of cell markers closely to that of FAPs. Thus, PRMS CSCs may be transformed into FAP lineage cells that acquire CSC features through oncogenic K-Ras signaling and p53 deficiency. Alternatively, the reported stem cell lineage infidelity [32] may result in an unidentified tumor-origin cell ectopically expressing the FAP marker gene set. However, whether the FAPs, PICs, and SCs of p53 knockout mice can form PRMS by the expression of oncogenic K-Ras remains to be further investigated.

The results presented here indicate that our established model system is a helpful tool for studying the features of PRMS tumors.

Competing Interests

The authors declare no conflicting interests.