Identification of Cancer Stem Cells Within Mixed Populations Using Live Cell RNA Detection

Despite the ubiquity of studies measuring mRNA and miRNA levels, many current techniques require cell lysis and are enormously disruptive to the cells being studied. Furthermore, amplification methods can create false positives or erroneously inflate differences. This article describes a new RNA detection technology that uses gold nanoparticles bound to sequence-specific oligonucleotide probes to enable target RNA levels to be determined within individual live cells. In the presence of their targets, the probes fluoresce and the cells can then be imaged and quantified using flow cytometry. Over time, the probes exit the cell through natural exocytosis, without adverse effects, thus enabling downstream assays on the same sample.

This technology permits labeling and sorting of cancer stem cells based on the expression of target mRNA transcripts from human tumor samples. The technology works across a wide range of solid tumor types and completely bypasses the use of surface markers and functional assays, which tend to be destructive and laborious and can vary from study to study. This approach greatly simplifies the process of identifying and isolating cancer stem cells while providing a better level of accuracy than was previously possible.

Cancer stem cells

In 1997, the first report of the presence of a stem-like cell being involved in cancer progress was reported in the scientific literature.1 In subsequent years, the cancer stem cell paradigm has contributed to our understanding of how cancer starts and progresses. All cancer starts from one cell; if this cell is a terminally differentiated tissue type cell, the tumor that is generated tends to be homogeneous and generally responds well to chemotherapy and radiation. If the first cell that undergoes the genetic mutation is an adult tissue stem cell, the tumor is more heterogeneous in makeup, with a small percentage of cells retaining a stem cell phenotype, and is more difficult to treat. A greater understanding of cancer stem cells may guide development of new chemotherapy agents. An important first step is to isolate cancer stem cells from the mixed population found in tumors.

Current methods to isolate cancer stem cells primarily involve the use of antibodies to label surface markers on the cells that are indicative of the stem cell phenotype. Different combinations of these markers are required for different types of tissue such as colon and pancreatic; currently, we lack a robust consensus marker for all types of tumor tissue.

Other methods are available to identify these stem cells. The side population assay or dye efflux assay, in which the cancer stem cells tend to pump out Hoechst dye, allows identification of a population that is functionally able to protect itself from a chemotherapeutic drug. Aldehyde dehydrogenase enzyme, highly overexpressed in stem cells, can also be used; the Aldefluor™ assay (Stem Cell Technologies, Vancouver, BC, Canada) develops a fluorescent signal proportional to the enzyme activity.

Once the cancer stem cells have been sorted, confirmation is necessary to ensure true cancer stem cells have been isolated. The gold standard for confirmation is in vivo tumorigenesis in which a very small number of the cancer stem cells, typically 50 to 100, are embedded in Matrigel® substrate (BD Biosciences, San Jose, CA) and injected into a mouse to determine if a tumor will develop. Invasion through Matrigel substrate using an in vitro invasion chamber can also be used to demonstrate the ability of cancer stem cells to secrete digestive enzymes such as metalloproteinases.

Another method to confirm isolation of cancer stem cells is use of a sphere culture assay in which a low attachment dish and special media conditions allow cancer stem cells to thrive in suspension. These cells will grow as small clusters or spheres, whereas the bulk tumor cells are unable to grow in an unattached fashion.

PCR, intracellular immunostaining, or Western blots are used to confirm expression of true stem cell transcription factors such as Nanog, Sox2, Oct3/4, and KLF-4. Unfortunately, these methods are destructive, requiring sacrifice of the isolated stem cells.

SmartFlare™ probes for live cell RNA detection (EMD Millipore, Temecula, CA) offer major advantages for identification of cancer stem cells by directly identifying the presence of the stem cell transcription factors such as Nanog, KLF-4, Oct3/4, and Sox2. Direct RNA detection and sorting of individual cells based on gene expression eliminates the need for antibodies to label surface markers, which can perturb the cancer stem cells due to the binding process.

Live cell RNA detection

Figure 1 describes the molecular mechanism of the SmartFlare RNA detection probe. The core of the probe is a gold nanoparticle with oligonucleotides conjugated to the surface. The longer strand (colored gray in Figure 1) is complementary to a target of interest inside the cell. Bound to that through normal base pairing is a shorter “flare” strand (colored purple), to which a fluorophore is bound; the fluorophore is in close proximity to the gold, which is a broad spectrum quencher. Target RNA will bind to the particle with a higher affinity than the flare strand; following binding of the target, the flare strand will leave the proximity of the gold and fluoresce.

 Figure 1 – Molecular mechanism of SmartFlare RNA detection probes.

The gold nanoparticles are added directly to cells in culture. The active endocytosis process allows the particles to enter the cell through natural cellular machinery and find the target inside the cell. If the target is present in the cell, it will bind to the particle and release the flare, and fluorescence is detected by microscopy or flow cytometry. The SmartFlare reagent is nontoxic and cells can be reused for additional assays once the SmartFlare work flow is completed.

SmartFlare probes offer a number of advantages over current RNA detection methods. Quantitative PCR is an endpoint determination and allows a quantitative determination of RNA expression level within cells, but results are reported for the whole population. Transfection reagents can be used for live cell probes but can decrease the viability of the cell and also can change gene expression. Fluorescence in situ hybridization is also an endpoint read.

Identification and sorting of cancer stem cells

Samples for the experiments described below were prepared from fresh tumor isolates, generally processed within one hour of obtaining them from surgery. After developing single cell suspensions, cells were incubated with SmartFlare probes overnight at 37 ºC in a 5% CO2 environment. Following incubation, flow cytometry was used to analyze and sort the cancer stem cells.

Figure 2 – Ovarian metastatic tumor cells analyzed using SmartFlare probes

Figure 2 shows representative data using SmartFlare probes on an ovarian metastatic tumor. Panels A and B represent the actual SmartFlare product. Panel A is the scramble control nanoparticle that was used for background gating. Panel B shows a 16% population above background fluorescence, indicating that those cells are expressing the Nanog mRNA transcripts, a stem cell marker.

The Nanog+ cells are then subgated after further surface staining for other cancer stem cell markers. Panel C shows 74% CXCR4 positivity, and Panel D shows 84% MDR1 positivity. Panels E and F represent the Aldefluor assay, which has its own negative control. Diethylaminobenzaldehyde (DEAB) is an inhibitory chemical used to block the aldehyde dehydrogenase activity; the negative gate was set on that. These cells show a 70% positivity for Aldefluor. Panel G represents the CD24/44 assay, which indicates stem cells showing low CD24 and very high levels of CD44 in the fourth quadrant at the lower right, which is 87%.

Figure 3 – Multiplexed SmartFlare assay using Nanog and KLF-4 probes.

One of the advantages of the SmartFlare system is the ability to combine two probes for different mRNA transcripts. Figure 3 shows the use of both Nanog and KLF-4 SmartFlare probes with primary ovarian tumor cells. Panel A is the scramble control, and Panel B is the dual positives. The upper right quadrant in Panel B shows 6% of the cells dual-staining for both Nanog and KLF-4. Similar to Figure 2, additional immunostaining was used to further analyze the cells. Panel C shows the CD24/44 assay, Panel D shows CXCR4 at 70%, Panel E shows MDR1 at 95%, and Panels F and G show results of the Aldefluor assay with only 42% positivity in this case.

Invasion assays

Because the SmartFlare work flow is nondestructive to cells, cell sorting can be performed following assessment of gene expression. Panel A in Figure 4 shows representative histograms from pre- and postsorting of the negative and positive Nanog populations. Cells from the negative and positive fractions were studied in a Matrigel invasion assay utilizing eight-micron pore FluoroBlok™ inserts (BD Biosciences). A rigorous assay was set up so that the cells had to move through a fairly thick layer of Matrigel substrate. Panel B shows representative images of the negative and positive FluoroBlok chambers; of a total of 2500 cells seeded, very few cells in the negative control were invasive, while an average of 600 Nanog+ cells came through after a 24-hour migration period (cells stained with Vybrant DyeCycle violet [Life Technologies, Grand Island, NY] for visualization).

Among several tumor types studied (both primary and metastatic sites), the variation in Nanog positivity ranged from as low as 3.4% to as high as 19% with an average of about 10%. For the concomitant staining with the other stem cell markers, Aldefluor performs rather poorly, only 67% of the time on average showing dual positivity with the Nanog-positive cells. The CD24/44 staining improves quite a bit (78.8% dual positivity)—80.7% for CXCR4 dual positivity, and 87.2% for the MDR1, indicating probably the best correlation with the Nanog positivity.

Figure 4 – Invasion assay results.

In summary, SmartFlare probes may be a more reliable method to isolate cancer stem cells. The probes work across the entire range of tumor types, eliminating the need for very specific cocktails for different kinds of tumors such as colon or pancreatic. SmartFlare technology also presents a very simplified, nondestructive work flow with minimal handling of the cells that would further cause damage to them. Expression levels can be assessed on a cell-by-cell basis, versus the entire population, and the cells remain fully functional after RNA detection and available for other assays.


  1. Bonnet, D.; Dick, J.E. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nature Med. Jul 1997, 3(7), 730–7.

Don Weldon is R&D Manager, EMD Millipore, 28820 Single Oak Dr., Temecula, CA 92590, U.S.A.; tel.: 951-514-4566; e-mail: . Steve McClellan is Manger, Research Operations, and Chief, Flow Cytometry Core Laboratory, Mitchell Cancer Institute, University of South Alabama, Mobile, AL, U.S.A.; e-mail: