The aim of this methodology is to identify cancer stem cells (CSC) in cancer cell lines and primary human tumor samples with the sphere-forming protocol, in a robust manner, using functional assays and phenotypic characterization with flow cytometry and Western blot.
Cancer stem cells (CSC) are a small population with self-renewal and plasticity which are responsible for tumorigenesis, resistance to treatment and recurrent disease. This population can be identified by surface markers, enzymatic activity and a functional profile. These approaches per se are limited, due to phenotypic heterogeneity and CSC plasticity. Here, we update the sphere-forming protocol to obtain CSC spheres from breast and gynecological cancers, assessing functional properties, CSC markers and protein expression. The spheres are obtained with single-cell seeding at low density in suspension culture, using a semi-solid methylcellulose medium to avoid migration and aggregates. This profitable protocol can be used in cancer cell lines but also in primary tumors. The tridimensional non-adherent suspension culture thought to mimic the tumor microenvironment, particularly the CSC-niche, is supplemented with epidermal growth factor and basic fibroblast growth factor to ensure CSC signaling. Aiming for robust identification of CSC, we propose a complementary approach, combining functional and phenotypic evaluation. Sphere-forming capacity, self-renewal and sphere projection area establish CSC functional properties. Additionally, characterization comprises flow cytometry evaluation of the markers, represented by CD44+/CD24– and CD133, and Western blot, considering ALDH. The presented protocol was also optimized for primary tumor samples, following a sample digestion procedure, useful for translational research.
Cancer populations are heterogeneous, with cells presenting different morphologies, proliferation and invasion capacity, due to differential gene expression. Among these cells, a minority population exists named cancer stem cells (CSC)1, which have the capacity for self-renewal, recapitulating the heterogeneity of the primary tumor niche and producing aberrantly differentiating progenitors that do not respond adequately to homeostatic controls2. CSC properties can be directly translated in clinical practice, given the association with events, such as tumorigenicity or resistance to chemotherapy3. The identification of CSC can lead to the development of targeted therapies that may include blockage of surface markers, promotion of CSC differentiation, blocking of CSC signaling pathway components, niche destruction, and epigenetic mechanisms4.
The isolation of CSC has been performed in cells lines and in samples of primary tumors5,6,7,8. The functional profile described for CSC includes clonogenic capacity, side population and tumorosphere formation9. The CD44high/CD24low phenotype has been consistently associated with breast CSC, which has proved to be tumorigenic in vivo and has been already associated with epithelial to mesenchymal transition5,10. High ALDH activity has also been associated with stemness and epithelial to mesenchymal transition (EMT) in several types of solid tumors11. ALDH expression has been associated with resistance to chemotherapy and to CSC phenotype in vitro12,13,14,15,16. Several other markers have been linked to CSC properties in different types of tumors, such as CD133, CD49f, ITGA6, CD1663,4 and others, as described in Table 1.
The tumorspheres consist of a three-dimensional model for the study and expansion of CSC. In this model, the cell suspensions from cell lines and from blood or tumor samples are cultivated in a medium supplemented with growth factors, namely epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF), without fetal bovine serum and in non-adherent conditions17. Inhibition of cell adhesion results in death by anoikis of differentiated cells18. Spheres are derived from the clonal growth of an isolated cell. For this purpose, the cells are distributed at low density to avoid cell fusion and aggregation19. Another strategy includes the use of semisolid methylcellulose20.
The sphere-forming protocol gained popularity in CSC isolation and expansion, due to time and cost and technical, profitable, and reproducible reasons21,22. Despite some reserves on the extent of which sphere formation reflects CSC, there is a propensity of stem cells to grow in non-adherent conditions with the characteristic phenotype, which resembles the native microenvironment21. None of the methods available for isolation of CSC from solid tumors has complete efficiency, highlighting the importance of developing more specific markers or combinations of methodologies and markers.
In this protocol, we detail the isolation of CSC with the sphere-forming protocol, with the principle of single-cell growth in non-adherent conditions and the capacity to produce a differentiated phenotype. A schematic representation of this procedure is represented in Figure 1. We also describe the characterization with surface markers and ALDH expression for CSC, both for breast and gynecological tumor cells lines and samples of primary tumors.
This protocol was performed complying with the ethical guidelines of the Coimbra Hospital and Universitary Center (CHUC) Tumor Bank, and was approved by CHUC's Ethics Committee for Health and by the Portuguese National Data Protection Commission.
1. Sphere-forming Protocol and Derived Adherent Populations from Continuous Cell Cultures
NOTE: Perform all procedures under strict sterile conditions.
2. Sphere-forming Protocol from Human Tumor Samples
NOTE: The use of human samples for research purposes must comply with each country's legislation, and to be approved by the Ethics Committee of the Institutions involved.
Figure 1: Obtaining cancer stem cells from human endometrial tumor samples (A) and breast and gynecological cancer cell lines (B). Human tumor samples are fragmented, enzymatically digested and plated in sphere culturing medium into poly-HEMA coated dishes. Cancer cell lines are detached, cell suspensions are counted, and single cells are distributed at low density into poly-HEMA coated plates under appropriate conditions. The spheres obtained, when placed under adherent culture conditions, produce derived adherent populations. Please click here to view a larger version of this figure.
3. Sphere-forming Capacity, Self-renewal, and Sphere Projection Area
NOTE: Sphere-forming capacity is the ability of a tumor cell population to produce spheres. Self-renewal is the ability of sphere cells to produce new colonies of spherical cells in suspension. The sphere projection area is representative of the area occupied by the sphere and therefore expressive of their size and the number of cell divisions undergone in a certain time period.
4. Cancer Stem Cell Marker Assessment with Flow Cytometry
NOTE: CD44+/CD24-/low phenotype was consistently associated with breast and gynecological cancer stem cells. The procedure described may be used to evaluate this and other cell surface markers.
5. Cancer Stem Cell Marker Assessment with Western Blot
NOTE: In addition to ALDH1 activity, high expression of this marker was consistently associated with breast and gynecological cancer stem cells13,14. The procedure described may be used to evaluate this and other cell markers.
The sphere-forming protocol allows spherical colonies to be obtained in suspension from several endometrial and breast cancer cell lines (Figure 2A) or after gentle enzymatic digestion of tissue from human tumor samples (Figure 2E). In both cases, a few days after plating, monoclonal spherical colonies in suspension are obtained. Both endometrial and breast cancer spheres give rise to a cell monolayer with similar morphology to the cell line of origin, 1 to 2 days after plating (Figure 2A).
Distinct lineage and tissue origins can be compared by the sphere-forming capacity, self-renewal and projection area. Representative results from breast cancer cell lines can be observed in the graphs in Figure 2B-D. The hormonal receptor-positive breast cancer MCF7 cells show higher sphere-forming capacity, self-renewal and projection area than the triple negative breast cancer cells HCC180614. For both cell lines, a small percentage of the cells plated (less than 3%) was able to produce spheres emphasizing cancer stem cells as a minority population within tumor cell heterogeneity. Cancer stem cells self-renewal was patented by a significantly different value of sphere self-renewal of the cell lines represented. Sphere projection area, as a rough measure of the spheres' dimension, correlates with the number of mitotic cycles and displays different time intervals for both lineages.
Whilst only a small proportion of cells is capable of forming tumorspheres in vitro and retaining self-renewal capacity carrying stem cells properties, several markers were associated with this phenotype.
The flow cytometry protocol presented allows for versatile experimental approaches, considering surface antigens (see Table 1). Representative results, shown in Figure 3A-B, concern CD44/CD24 and CD133 membrane markers that have been proposed as corresponding to a more cancer stem cell-like phenotype. Analysis of spheres obtained from endometrial RL95-2 and ECC-1 cell lines allowed four populations to be identified (Figure 3A). Spheres obtained from endometrial RL95-2 comprised a CD44high/CD24– population three times larger than the parental cell line35. In the case of ECC-1 spheres, the CD44high/CD24– corresponds to the major population, which is also CD133 positive, while the CD44low/CD24–, CD44low/CD24+ and CD44–/CD24+ have negative or low CD133 expression.
Assessing surface and intracellular markers can also be performed by western blot after gentle sphere harvesting and careful protein sample preparation. Figure 3C shows typical results of ALDH change, a marker whose increased activity or augmented protein expression is associated with the cancer stem cell phenotype13,14 on spheres and derived adherent cells regarding the endometrial ECC1 cell line of origin.
Figure 2: Endometrial and breast cancer cells, spheres and derived adherent populations. (A). Representative images of endometrial (RL95-2 and ECC-1) and breast (MCF7 and HCC1806) cancer cell lines, respective endometrial (ES1) and breast (MS1) spheres and derived adherent populations (G1). Representative images of RL95-2, ECC-1, MCF7 and HCC1806 cancer cell lines were obtained at a magnification of 200x (Scale bar = 50 µm). Representative images of ES1 RL95-2 and ES1 ECC-1 were obtained at a magnification of 200x (Scale bar = 50 µm). Representative images of MS1 MCF7 and MS1 HCC1806 were obtained at a magnification of 200x (Scale bar = 100 µm). Representative images of G1 RL95-2 and G1 ECC-1 were obtained at a magnification of 200x (Scale bar = 50 µm). Representative images of G1 MCF7 and G1 HCC1806 were obtained at a magnification of 200x (Scale bar = 100 µm). (B-D). Sphere-forming capacity, self-renewal and sphere projection area of breast cancer spheres MCF7 and HCC1806. (E). Representative images of spheres obtained from human endometrial tumor samples. These images were captured at a magnification of 200x (Scale bar = 50 µm). Part of this figure has been modified from a previous publication with permission from the publisher14. Please click here to view a larger version of this figure.
Figure 3: Combined evaluation of cancer stem cells markers in endometrial cancer cells. (A). Representative plots of CD44/CD24 labelling of the RL95-2 cell line and of the RL95-2 sphere cells. (B). Representative histograms of CD133 labelling of sphere cells (ES1) obtained from RL95-2 and ECC-1 cell lines. Density represents a measure of the cell count. CD44+/CD24–, CD44low/CD24–, CD44low/CD24± and CD44–/CD24+ populations are painted in green, pink, blue and yellow, respectively. C. ALDH expression in ECC-1 cell line, spheres (ES1), and derived adherent population (G1). The immunoblot represents the ALDH and actin expression for the respective experimental conditions. ALDH expression was evaluated with the antibody ALDH1/2, which detects the isoforms ALDH1A1, ALDH1A2, ALDH1A3 and ALDH2 of mouse, rat and human origin. Part of this figure has been modified from previous publications with permission from the publishers13. Please click here to view a larger version of this figure.
Table 1: List of gynecological and breast cancer stem cells markers.
Marker | Stem cell origin | References |
CD24 | Ovarian cancer | 60 |
CD29 | Breast cancer | 61 |
CD44 | Ovarian cancer | 62,63 |
CD44/CD24-/low | Breast cancer | 13,5,3,64 |
CD44+/CD24-/low/ESA | Breast cancer | 65 |
CD44/ALDH1+/hi | Breast cancer | 66 |
CD44/CD24-/low/ABCG2 | Breast cancer | 67 |
CD44/CD24-/low/ALDH1 | Breast cancer | 43,68,69 |
CD44/CD24-/low/EpCAM | Breast cancer | 5 |
CD44/CD24-/low/SSEA-3 | Breast cancer | 70 |
CD44/CD49f/CD133/2 | Breast cancer | 71 |
CD44/CD133/ALDH1+/hi | Breast cancer | 69 |
CD44/CD117 | Ovarian cancer | 7 |
CD44/MyD88 | Ovarian cancer | 72,73 |
CD44/E-cadherin–/CD34– | Ovarian cancer | 74,75 |
CD44/CD24/Epcam | Ovarian cancer | 76,77 |
CD44/CD24– | Ovarian cancer | 78,79 |
CD44/CD166 | Ovarian cancer | 80 |
CD44/CD24 | Cervical cancer | 81 |
CD49f | Breast cancer | 4 |
Cervical cancer | 82 | |
CD117 or c-Kit | Endometrial cancer | 83 |
Ovarian cancer | 62,84 | |
CD133 | Breast cancer | 4,85 |
Ovarian cancer | 62,86 | |
Endometrial cancer | 13,87,88,89 | |
Cervical cancer | 82 | |
CD133hi/CXCR4hi/ALDH1hi | Breast cancer | 90 |
CD133/ALDH1 | Breast cancer | 91 |
Ovarian cancer | 60,92 | |
CD133/CXCR4 | Endometrial cancer | 93 |
ABCG2 | Breast cancer | 65 |
Cervical cancer | 82,81 | |
ALDH-1 | Breast cancer | 4 |
Endometrial cancer | 94,95 | |
Cervical cancer | 82 | |
CXCR4 or CD184 | Breast cancer | 96 |
EpCAM/CD49f | Breast cancer | 97 |
EpCAMhi/PROCRhi/SSEA-3 | Breast cancer | 70 |
GD2/GD3/GD3Shi | Breast cancer | 98 |
ITGA6 | Breast cancer | 4 |
PROCR | Breast cancer | 43 |
NOTE: This table provides a list of surface epitopes expressed by various gynaecological and breast cancer stem cells; most of these markers are also expressed by a range of other tissues. This list does not aim to include all markers reported.
Table 2: List of tubes to be included in a typical flow cytometry experiment to evaluate the CD24/CD44 phenotype. The table shows a minimal set of sample tubes required for a co-staining experiment, including necessary controls.
Tube | Condition | Antigen-fluorophore |
1 | Unstained cells | none |
2 | Single stained CD44 | CD44-PE |
3 | Single stained CD24 | CD24-APC |
4 | Double-stained CD44/CD24 | CD44-PE and CD24-APC |
NOTE: This experiment can be performed adding annexin V-FICT to tube 4 and adding the respective control tube in order to gate the annexin V negative cells and exclude eventual cells in apoptosis.
This protocol details an approach to obtain tumorspheres from cancer cell lines and primary human samples. Tumorspheres are enriched in a sub-population with stem cell-like properties36. This enrichment in CSC is dependent on viability in an anchorage-free environment while differentiated cells are reliant on adhesion to a substrate37. As primary plating of tumor cells in a low adherence environment that imposes suspension does not ensure enrichment in CSC per se, we provide strategies to evaluate self-renewal (sphere-forming capacity and self-renewal), differentiation capacity (derived adherent populations), and phenotype of CSC (with flow cytometry and/or western blot). Cancer stem cells can be identified via several broadly described phenotypic markers (see Table 1).
As human tumor primary cultures are often challenging to establish and to maintain in culture, the sphere-forming protocol might provide a tool for handling these samples. The enzymatic digestion procedure suggested provided single-cell suspensions from endometrial tissue samples38. The sphere-forming protocol provides significant numbers of CSC, which are difficult to obtain by other means. The tridimensional model might be more efficient at mimicking the in vivo situation, namely the physiological microenvironment and tumor heterogeneity, than conventional monolayer cell cultures.
The certainty about the monoclonal origin of tumorspheres is a critical step of this protocol. Minimizing aggregation, which tends to occur in suspension cultures, and a thorough optimization of seeding densities to distribute single-cell suspensions are crucial24. Other authors suggested the plating of a single cell per well39,40. To avoid this laborious procedure, we overcame this issue by ensuring a single-cell suspension is plated in low density in a methylcellulose-enriched medium. Due to its water holding and viscosity enhancing properties23, methylcellulose provides a semi-solid medium that avoids migration and aggregation, ensuring the monoclonality of the spheres obtained21. The number of days in culture is another aspect which is dependent on optimization, as the number of days necessary to obtain spheres with diameters superior to 40 µm is dependent on each cell type doubling time24. The low or serum-free medium is another characteristic of the protocol, as FBS-containing medium is relevant for differentiated cell-growth in adherent conditions41, as in the parental cell lines and in the derived adherent cells. The protocol depends on the maintenance of a steady concentration of the specific growth factors. EGF signaling plays an important role in the maintenance of pluripotency pathways while bFGF acts as a mitogen contributing to the generation of spheres42,43.
The sphere-forming protocol, associated with appropriate techniques, provides the means to expand, isolate, and evaluate specific populations of CSC21. Several authors have pointed to its utility in assessing stem cell gene expression44,45,46,47 and stemness in tumours47,48, to study epithelial-mesenchymal transition44,49 and tumorigenesis45,48, to evaluate the effect of new therapies21,50 and drug resistance44,51, and to establish cultures from primary samples21,45,46. However, it is important to keep in mind that it is a sensitive experiment, highly dependent on adequate culture conditions. Additionally, the spheres present cellular heterogeneity due to CSC asymmetric division52 and do not represent a good model of the complexity of cancer stem cell formation and maintenance in the in vivo niche46.
Besides the sphere-forming protocol, other functional assays have been used for the detection of CSC. In vivo tumorigenicity entails the inoculation of low cell numbers in immunocompromised mice to obtain tumours36,53. This depends on the availability of proper conditions to perform animal studies, and due to the non-species-specific microenvironment, the recovery of living cells might be challenging. A colony forming unit assay, evaluating cell ability to generate colonies after they are plated at low density52, provides low cell numbers. Side-population relies on fluorescence-activated cell sorting (FACS) to isolate a group of cells with the ability to extrude the Hoescht 33342 stain. This sensitive method relies on the expression of ATP binding cassette protein (ABC) transporters, responsible for drug efflux54. Nevertheless, the side-population is associated with some disadvantages, namely, non-specificity for some phenotypes of CSC and the characteristics of the dye, which is toxic and largely influenced by experimental conditions (temperature, concentration)54,55. ALDEFLUOR is another flow cytometry-based assay for the identification of cells with intracellular ALDH activity. The main issue is the lack of reproducibility between studies that seem to be highly influenced by the culture conditions54.
The sphere-forming protocol is often combined with phenotypic analysis, as we proposed here, emphasizing the utility of complementary methods to identify CSC13,14. We recommended CSC enrichment via the sphere-forming protocol and further confirmation of stemness via assessment of biochemical markers by flow cytometry and western blot. Flow cytometry studies identified heterogeneous populations within the spheres. In fact, there is an enrichment in CSC in the studies shown, represented in this protocol by the CD44high/CD24low cells. Due to CSC asymmetric self-renewal24, other cell phenotypes were also identified. In the case of CD133, representative results showed the population with higher stemness to be positive in the case of the ECC-1 cell line, but negative in RL95-2 spheres. This points to the lack of specificity of some CSC markers described, which are not unique to these cells and might vary with the plasticity of the phenotype, and to the importance of using a combination of strategies to confirm stemness.
Western blot is an alternative methodology that might be useful in certain cases. For instance, while ALDH activity is broadly used, it is now known that multiple isoforms contribute to ALDEFLUOR metabolization54. Thus, specific antigen-antibody methods might be more reliable and we already showed the association between ALDH protein expression and stemness13,14.
Sphere-forming capacity, self-renewal and derived adherent populations represent the capacity of CSC to indefinitely divide and produce a differentiated progeny, which clinically translates to events such as relapse, metastization and resistance to treatment9. Drug resistance in CSC can be explained by overexpression of multidrug resistance (MDR) membrane proteins, ALDH expression involved in detoxification mechanisms, DNA repair mechanisms, protection against reactive oxygen species and resistance to apoptosis56. CSC have the capacity to be quiescent due to their plasticity and this has emerged as a mechanism of drug resistance. This population can be spared from chemo- and radiotherapy due to cell-cycle arrested differentiated cells57. Spheres are a tumor population with reported resistance to cytostatic drugs used in conventional treatment and have also been a focus for combination with targeted therapies54,58,59. The sensitivity of spheres can be tested for cytostatics used in breast and endometrial cancers. In addition, the isolation of CSC from a tumor sample can be a platform for the clinical application of therapy specific to each tumor, predicting resistance and consequent recurrent disease.
The authors have nothing to disclose.
This study was funded by the Portuguese Society of Gynecology through the 2016 Research Prize and by CIMAGO. CNC.IBILI is supported through the Foundation for Science and Technology, Portugal (UID/NEU/04539/2013), and co-funded by FEDER-COMPETE (POCI-01-0145-FEDER-007440). The Coimbra Hospital and Universitary Center (CHUC) Tumor Bank, approved by CHUC’s Ethics Committee for Health and by the Portuguese National Data Protection Commission, was the source of endometrial samples of patients followed at the institution’s Gynecology Service. Figure 1 was produced using Servier Medical Art, available from www.servier.com.
Absolute ethanol | Merck Millipore | 100983 | |
Accutase | Gibco | A1110501 | StemPro Accutas Cell Dissociation Reagent |
ALDH antibody | Santa Cruz Biotechnology | SC166362 | |
Annexin V FITC | BD Biosciences | 556547 | |
Antibiotic antimycotic solution | Sigma | A5955 | |
BCA assay | Thermo Scientific | 23225 | Pierce BCA Protein Assay Kit |
Bovine serum albumin | Sigma | A9418 | |
CD133 antibody | Miteny Biotec | 293C3-APC | Allophycocyanin (APC) |
CD24 antibody | BD Biosciences | 658331 | Allophycocyanin-H7 (APC-H7) |
CD44 antibody | Biolegend | 103020 | Pacific Blue (PB) |
Cell strainer | BD Falcon | 352340 | 40 µM |
Collagenase, type IV | Gibco | 17104-019 | |
cOmplete Mini | Roche | 118 361 700 0 | |
Dithiothreitol | Sigma | 43815 | |
DMEM-F12 | Sigma | D8900 | |
DNAse I | Roche | 11284932001 | |
ECC-1 | ATCC | CRL-2923 | Human endometrium adenocarcinoma cell line |
Epidermal growth factor | Sigma | E9644 | |
Fibroblast growth factor basic | Sigma | F0291 | |
Haemocytometer | VWR | HERE1080339 | |
HCC1806 | ATCC | CRL-2335 | Human mammary squamous cell carcinoma cell line |
Insulin, transferrin, selenium Solution | Gibco | 41400045 | |
MCF7 | ATCC | HTB-22 | Human mammary adenocarcinoma cell line |
Methylcellulose | AlfaAesar | 45490 | |
NaCl | JMGS | 37040005002212 | |
Poly(2-hydroxyethyl-methacrylate | Sigma | P3932 | |
Putrescine | Sigma | P7505 | |
RL95-2 | ATCC | CRL-1671 | Human endometrium carcinoma cell line |
Sodium deoxycholic acid | JMS | EINECS 206-132-7 | |
Sodium dodecyl sulfate | Sigma | 436143 | |
Tris | JMGS | 20360000BP152112 | |
Triton-X 100 | Merck | 108603 | |
Trypan blue | Sigma | T8154 | |
Trypsin-EDTA | Sigma | T4049 | |
��-actin antibody | Sigma | A5316 |