This protocol describes a reproducible method for isolation of mouse rhabdomyosarcoma primary cells, tumorsphere formation and treatment, and allograft transplantation starting from tumorspheres cultures.
Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma in children. Although significant efforts have enabled the identification of common mutations associated with RMS and allowed discrimination of different RMS subtypes, major challenges still exist for the development of novel treatments to further improve prognosis. Although identified by the expression of myogenic markers, there is still significant controversy over whether RMS has myogenic or non-myogenic origins, as the cell of origin is still poorly understood. In the present study, a reliable method is provided for the tumorsphere assay for mouse RMS. The assay is based on functional properties of tumor cells and allows the identification of rare populations in the tumor with tumorigenic functions. Also described are procedures for testing recombinant proteins, integrating transfection protocols with the tumorsphere assay, and evaluating candidate genes involved in tumor development and growth. Described further is a procedure for allograft transplantation of tumorspheres into recipient mice to validate tumorigenic function in vivo. Overall, the described method allows reliable identification and testing of rare RMS tumorigenic populations that can be applied to RMS arising in different contexts. Finally, the protocol can be utilized as a platform for drug screening and future development of therapeutics.
Cancer is a heterogeneous disease; furthermore, the same type of tumor can present different genetic mutations in different patients, and within a patient a tumor is composed by multiple populations of cells1. Heterogeneity presents a challenge in the identification of cells responsible for initiating and propagating cancer, but their characterization is essential for the development of efficient treatments. The notion of tumor propagating cells (TPC), a rare population of cells that contribute to tumor development, has been previously extensively reviewed2. Despite the fact that TPCs have been characterized in multiple types of cancer, the identification of markers for their reliable isolation remains a challenge for several tumor types3,4,5,6,7,8,9. Thus, a method that does not rely on molecular markers but rather on TPC functional properties (high self-renewal and the ability to grow in low-attachment conditions), known as the tumorsphere formation assay, can be widely applied for the identification of TPCs from most tumors. Importantly, this assay can also be employed for expansion of TPCs and thus directly applied to cancer drug screening and studies on cancer resistance1,10.
Rhabdomyosarcoma (RMS) is a rare form of soft tissue sarcoma most common in young children11. Althoug RMS can be histologically identified through assessment of expression of myogenic markers, the RMS cell of origin has not been univocally characterized due to the multiple tumor subtypes and high heterogeneity of the tumor developmental stimuli. Indeed, recent studies have generated significant scientific discussion about whether RMS is of myogenic or non-myogenic origins, suggesting that RMS may derive from different cells types depending on the context12,13,14,15,16,17. Numerous studies on RMS cell lines have been performed employing the tumorsphere formation assay for the identification of pathways involved in tumor development and characterization of markers associated with highly self-renewal populations18,19,20,21.
However, despite the tumorsphere formation assay's potential for identifying RMS cells of origin, a reliable method that can be employed on primary RMS cells has not yet been described. In this context, a recent study from our group employed an optimized tumorsphere formation assay for the identification of the RMS cells of origin in a Duchenne muscular dystrophy (DMD) mouse model22. Multiple pre-tumorigenic cell types, isolated from muscle tissues, are tested for their ability to grow in low-attachment conditions, allowing the identification of muscle stem cells as cells of origin for RMS in dystrophic contexts. Described here is a reproducible and reliable protocol for the tumorsphere formation assay (Figure 1), which has been successfully employed for the identification of extremely rare cell populations that are responsible for mouse RMS development.
The housing, treatment, and sacrifice of mice were performed following the approved IACUC protocol of the Sanford Burnham Prebys Medical Discovery Institute.
1. Reagent preparation
2. Cell isolation and culture
3. Tumorsphere derivation
4. Tumorsphere treatment with recombinant proteins
5. Tumorsphere treatment with overexpression plasmids
6. Tumorsphere preparation for allograft transplantation
Tumorspheres detection
Cell isolation was optimized to obtain the maximum heterogeneity of cell populations present in the tumor tissue. First, since isolated tissues presented morphologically dissimilar areas, to enhance the chances of isolating uniform rare cell populations, sampling was performed from multiple areas of the tumor (Figure 1A, first panel on the left). Second, mechanical dissociation of the harvested samples was performed while maintaining homogeneity in the minced tissue size, despite different resistances that may have been present across the samples (Figure 1A, third and fourth panel from the left). Depending on the starting material (tumor aggressiveness, age and genotype of mouse, tumor location), recovery of the cells from the mechanical stress of the isolation process may vary, ranging from 3-7 days (Figure 1A, last panel on the right). To enhance cell survival and growth, media should be changed the day after isolation and then every 2 days. This will remove the debris and dead cells accumulated during the isolation process that might affect to cell viability. Tumorsphere formation assay should start after the cells have been passaged at least once to ensure optimal viability when placed in suspension cultures (Figure 1B, first panel on the left). In our hands, optimal results were obtained starting with cells from passage 2 (P2) (Figure 1C, first and second panel on the left). This specific passage was chosen after multiple testing. When P0 cells were plated in low-attachment conditions, they formed a low number of tumorspheres compared to later passages, possibly due to cellular debris still present after isolation. In later passages, different patterns of tumorsphere formation were observed and attributed to the selection that occurs after numerous passages in culture. When different cell lines are compared, it is suggested to start from the same passage.
Discrimination between tumorspheres and cell clusters is of fundamental importance for quantification of the assay. Figure 1C (last two panels on the right) clearly shows the morphological differences between a tumorsphere (left) and a cell cluster (right). Tumorspheres derive from a single cell that has a high ability to self-renew and grow in low-attachment conditions (both features of TPCs). Indeed, tumorsphere development indicates tumorigenicity potential. However, this assay is performed in vitro independently from the cues derived from the whole organism; thus, to validate cells' tumorigenic potential in vivo, allografts transplantation experiments should be performed (Figure 1D).
Validation of recombinant proteins treatment
Before setting up tumorspheres treatment, the optimal concentration at which the protein of interest triggers an effect on tumor cells should be determined. To do so, assessment of the level of expression of the protein's downstream target genes is necessary (Figure 2). A literature search on PubMed was performed before determining the target genes to test through qRT-PCR. In the event that the protein of interest has been shown to modulate different downstream pathways, selection of multiple genes associated with each of these pathways should be performed. For instance, Flt3l (Fms-like tyrosine kinase 3) has been shown to modulate STAT5 signaling in acute myeloid leukemia, inducing the downstream expression of p21, c-Myc, and CyclinD1 (Figure 2)27. Moreover, expression of Flt3l is required for dendritic cells differentiation mediating activation of STAT3 signaling pathway28. To assess STAT3 activity, Socs3 and CyclinD1 expression were checked (Figure 2). Analysis of the results showed a dose dependent effect of recombinant protein treatment on some of the tested genes (p21, CyclinD1, and Socs3), whereas others were not affected (c-Myc). It is important to determine the downstream genes responsive to the protein of interest in the specific tumor tested for reliable assessment of treatment effectiveness.
Optimization of the protocol for plasmid transfection
To establish an efficient protocol for plasmid transfection and further tumorsphere formation assay, the effects of transfection on adherent tumor cells were tested (Figure 3A). Transfection reagent treatment was performed following the manufacturer's protocol, and two different amounts of the reagent were tested. Efficiency was assessed employing a GFP reporter plasmid. In our hands, we observed a higher transfection efficiency using lower amounts of the reagent (Figure 3A). Indeed higher concentrations lead to increased cell death, starting at 48 h and becoming more evident after 72 h from the time of transfection. The same transfection protocol was not efficient in cells in suspension, as accumulation of dead cells and cellular debris became evident 24 h after the beginning of the treatment, indicating decreased cellular viability. To overcome this technical challenge, a two steps protocol was employed: perform transfection on adherent cells, detach them 24 h after the treatment, and plate them in suspension for 7 more days. Control tumorspheres were indeed expressing GFP at the end of the experiment (Figure 3B).
Figure 1: Tumor cell isolation, tumorsphere derivation, and transplantation. (A) Schematic representation of section 2 of the protocol (tumor cell isolation). Each key step of the protocol is summarized. (B) Schematic representation of section 3 of the protocol (tumorspheres derivation). Each key step of the protocol is summarized. (C) From left: bright-field image of isolated tumor cells at passage 2 (P2) at low magnification (scale bar = 50 µm) and high magnification (scale bar = 50 µm); bright-field image of a tumorsphere derived from tumor cells after 30 days in suspension culture (scale bar = 250 µm), and bright-field image of a cell cluster formed from tumor cells after 30 days in suspension culture (scale bar = 50 µm). (D) Schematic representation of section 6 of the protocol (tumorsphere allograft transplantation). Each key step of the protocol is summarized. Please click here to view a larger version of this figure.
Figure 2: Validation of downstream target genes for determination of recombinant protein concentration. qRT-PCR results for assessment of Flt3l treatment concentration. Two-way ANOVA was performed. Significance is shown for the comparison with non-treated control. (*p < 0.05; **p < 0.01; ***p < 0.001; n = 3). Please click here to view a larger version of this figure.
Figure 3: Set up of tumorspheres transfection protocol. (A) Representative bright-field and fluorescent images of tumor cells treated with low (top) or high (bottom) concentration of transfection reagent (0.75 µL or 1.5 µL in 24 well plates) (scale bar = 50 µm). (B) Representative image of tumorspheres formed from GFP plasmid-treated cells, 7 days after plating the cells in suspension (scale bar = 50 µm). Please click here to view a larger version of this figure.
Multiple methods have been employed for isolation and characterization of TPCs from tumor heterogeneous cell populations: tumor clonogenic assays, FACS isolation, and tumorsphere formation assay. The tumor clonogenic assay was first described in 1971, used for stem cell studies, and only subsequently applied to cancer biology29,30. This method is based on the cancer stem cells intrinsic property to expand without constraints in soft gels cultures31. Since its development, this method has been widely used in cancer research for multiple purposes, including tumor cell heterogeneity studies, effect of hormonal treatment on cell growth, and tumor resistance studies31. To date, this assay is still employed for the identification of tumor initiating cells for multiple types of cancers32.
FACS isolation is based on prior knowledge of molecular markers present on the surface of cells on interest. It has been widely utilized for the isolation of TPCs from both liquid and solid tumors. For example, the first identification of human acute myeloid leukemia (AML) initiating cells was performed by Dr. Dick’s group by utilizing FACS fractionation and transplantation assays based on the knowledge of the markers present on bulk AML cells9. Utilizing a similar approach, Dr. Clarke’s group isolated breast cancer initiating cells3. Tumorsphere formation assay is a different approach used for the identification and study of TPCs. This method was first developed to identify cancer stem cells from brain tumors33. Interestingly, it was initially tested using the same conditions known to favor neural stem cells growth, thus favoring the self-renewal property also associated with TPCs33. Moreover, tumorsphere formation relies on the capacity of TPCs to growth in an anchorage-independent manner.
The three above described approaches can be used in parallel to enhance the probability to isolate and molecularly characterize TPCs populations, overcoming the limitations of each method. For instance, FACS, despite bringing the great advantage of isolating pure cell populations, strongly relies on the use of surface markers that are not yet known for all cancer types. Thus, its use is limited to the isolation of TPCs expressing known markers. Tumor clonogenic and tumorsphere formation assays are both based on cellular properties, known to be associated with TPCs. Both these methods can be employed as the first line of studies on new or not yet studied cancers. Moreover, these two methods can ensure the expansion and enrichment of the cell population of interest, facilitating the identification of molecular markers, and allowing for both cancer resistance and drug screening studies. In this context, tumorsphere formation assay is more advantageous, since these spheroid structures better recreate the environment present in tumor tissues (hypoxic areas in the center of the spheres)34. Indeed, it has been previously shown that 3D cultures are more reliable for predicting drug treatments outcome34. Tumorspheres can be recovered after culture, and used for allograft experiments: digestion of tumorspheres into single cells solutions and transplantation into recipient mice will allow for in vivo assessment of tumorigenic capacity of newly identified TPCs, compared to bulk tumor cells.
To successfully obtain a starting cellular material representative of tumor heterogeneity, it is critical to first perform random sampling of the tissue. RMS are characterized by fibrotic, fatty, or highly vascularized areas which are clearly distinguishable in isolated tumors, thus, to maintain this cellular diversity, collection of each area of the tissue will be required. Moreover, to enhance the chances of cells survival during the digestion process, mincing of the collected tissue needs to result in evenly sized pieces: smaller fragments are more likely to be overdigested inducing reduction of cells viability. This may be particularly tedious due to the diverse morphology and stiffness of RMS.
For quantification of the result of tumorsphere formation assay, it is of critical importance to distinguish between real tumorspheres and cellular clusters (Figure 1C, last two panels on the right). A tumorsphere is solid spheroid structure in which it is not possible to discriminate the cell composiiton; in contrast, in a cell cluster, single cells can be easily discriminated. Cell clusters may not assume a rounded shape and are significantly smaller when compared to tumorspheres, which range between 50-250 µm25.
To achieve tumorspheres transplantation, obtaining a uniform single cells solution is a key step. Indeed, given the tight structure and large size of a tumorsphere, digestion of the cells becomes a major limiting step in the preparation of a single cell suspension that is further employed for transplantation. Multiple cycles of enzymatic digestion combined with mechanical dissociation are thus necessary for complete dissociation of tumorspheres. To confirm the progress of dissociation and a successful outcome, the solution must be monitored under a bright-field microscope.
Despite the multiple advantages provided by the tumorsphere formation assay, tumorspheres have been shown not to originate from every type of tumor or from all commercially available cell lines. In these cases, the assay cannot be employed as the standard for determining cell tumorigenicity and quantification of TPCs within a heterogeneous population. Another limitation of this assay is associated with the fact that different tumor types require different growing and dissociation conditions; thus, it requires time-consuming optimization and troubleshooting of both protocols for each tumor type or cell line. Moreover, fusion of multiple tumorspheres may occur in culture, making assessment of their sizes and numbers unclear.
Despite the fact that the tumorsphere formation assay has been previously utilized in RMS studies, it has been mainly applied to commercially available RMS cell lines to identify the molecular pathways involved in tumor formation and development18,20,21. Given the heterogeneous tissue composition, the numerous subtypes of tumors and diverse developmental contexts from which RMS originates, employment of RMS cell lines limits application of this assay for the identification of both cells of origin and developmental cues that lead to tumor development in vivo.
An attempt to develop an efficient protocol for tumorsphere derivation, starting from human sarcoma samples, has previously shown poor results. Indeed, tumorspheres were developed from only 10% of the samples35. Thus, there is a need for a reproducible and reliable assay for isolation of primary RMS cells and tumorsphere development. In response to this need, the described tumorsphere formation assay was optimized to be employed in primary cell cultures. The development of this protocol is the first step towards answering major questions in the field (i.e., how RMS cells of origin differ depending on environmental context). In more detail, described here is a reproducible protocol for the isolation of primary tumor cells from RMS tissues, formation of tumorspheres, tumorsphere treatment (both performed with recombinant proteins or overexpression plasmids), and allograft transplantation experiments.
In conclusion, the tumorsphere formation assay is a well-established and versatile method for the identification of TPCs in different types of tumors, enriching the cells with increased self-renewal capacity and the ability to grow in an anchorage-independent manner. This assay is based on functional characteristics of the cell populations being studied and not on the previous knowledge of molecular markers; thus, it can be applied as an exploratory tool for a wide range of tumors types. Moreover, the isolation of rare cell populations achieved with tumorsphere cultures makes this in vitro assay an ideal platform for cancer drug testing.
The authors have nothing to disclose.
This work was supported by the Ellison Medical Foundation grant AG-NS-0843-11, and the NIH Pilot Grant within the NCI Cancer Center Support Grant P30CA030199 to A.S.
Accutase cell dissociation reagent | Gibco | A1110501 | Detach adherent cells and dissociate tumorspheres |
Celigo | Nexcelom | Celigo | Microwell plate based image cytometer for adherent and suspension cells |
Collagenase, Type II | Life Technologies | 17101015 | Tissue digestion enzyme |
Dispase II, protease | Life Technologies | 17105041 | Tissue digestion enzyme |
DMEM high glucose media | Gibco | 11965092 | Component of tumor cells media |
DMEM/F12 Media | Gibco | 11320033 | Component of tumosphere media |
EDTA | ThermoFisher | S312500 | Component of FACS buffer |
EGF recombinant mouse protein | Gibco | PMG8041 | Component of tumosphere media |
FACSAria II Flow Cytometry | BD Biosciences | 650033 | Fluorescent activated cell sorter |
Fetal Bovine Serum | Omega Scientific | FB-11 | Component of tumor cells media |
Fluriso (Isofluornae) anesthetic agent | MWI Vet Supply | 502017 | Anesthetic reagent for animals |
FxCycle Violet Stain | Life Technologies | F10347 | Discriminate live and dead cells |
Goat Serum | Life Technologies | 16210072 | Component of FACS buffer |
Ham's F10 Media | Life Technologies | 11550043 | Component of FACS buffer |
Horse Serum | Life Technologies | 16050114 | Component of cell isolation media |
Lipofectamine 3000 transfection reagent | ThermoFisher | L3000015 | Transfection Reagent |
Matrigel membrane matrix | Corning | CB40234 | Provides support to trasplanted cells |
N-2 Supplemtns (100X) | Gibco | 17502048 | Component of tumosphere media |
Neomycin-Polymyxin B Sulfates-Bacitracin Zinc Ophthalmic Ointment | MWI Vet Supply | 701008 | Eyes ointment |
PBS | Gibco | 10010023 | Component of FACS buffer and used for washing cells |
pEGFP-C1 | Addgene | 6084-1 | GFP plasmid |
Penicillin – Streptomyocin | Life Technologies | 15140163 | Component of tumosphere and tumor cells media |
Recombinant Human βFGF-basic | Peprotech | 10018B | Component of tumosphere media |
Recombinant mouse Flt-3 Ligand Protein | R&D Systems | 427-FL-005 | Recombinant protein |
Trypan blue | ThermoFisher | 15250061 | Discriminate live and dead cells |