January 1, 2013 (Vol. 33, No. 1)
Josh P. Roberts
There are different ways to query cells. Some of these use standard light microscopy observations of dye uptake or morphology, size, adherence, or mobility.
Others rely on fluorescence microscopy to track labeled antibodies or reporter-driven GFP expression, or on specialized spectrophotometers to examine enzyme-driven color or light intensity changes in the culture medium. And still others make use of nonoptical properties such as heat flow or electrical impedance to detect activity.
Especially in drug discovery, most such assays have been geared toward a particular reaction—whether it is recognition of a cell surface marker, triggering endocytosis, or exciting or inhibiting a phosphorylation cascade. Scientists have been working with very specific assays for very specific targets for years, explained Magnus Jansson, Ph.D., CSO of SymCel Sverige. “I think that the trend is turning back now toward a screening for more of a specific phenotypic effect.” He said there was a huge interest in label-free assays among the participants of the recent SMi Conference on “Cell Based Assays” in London, where he discussed SymCel’s pending entry into the field.
The company has developed a calorimetry-based assay system that can monitor the total metabolic activity by measuring the heat flow in and out of a cell culture well in real time, down to the microwatt level. “It’s not necessary to know anything about your system with regard to the receptor, nor what kind of assay you’re going to run to monitor that activity,” he pointed out.
Say you have a G-protein coupled receptor (GPCR) of unknown function—with this assay you can determine whether it effects a stimulatory or inhibitory response, or induces necrosis under your experimental conditions, Dr. Jansson proposed as an example. “You really don’t have to be that advanced or that long into your research of that novel target—you can pinpoint the result of your chemical, or try to elucidate a pathway without previous knowledge about it.”
SymCel’s calPlate assay is independent of cell type and media conditions. Three-dimensional cultures, adherent or suspension cells, bacteria, yeast, and even small parasites such as malaria can be queried. Likewise, clear, colored, and turbid media can be used. Because the assays require no labels they are noninvasive and nondestructive to the cells, inexpensive, simple, and quick to set up and use.
The calPlate, which is currently in beta testing at several European universities, charts a continuous readout of heat exchange throughout the assay, generally about three to six hours. Because it is based on a sealed system that needs to equilibrate with a heat sink, it is not designed to acquire rapid measurements. And due to lack of gas exchange there is an upper limit on the length of an assay as well—at least in aerobic cultures. After being removed from the instrument, cultures can be manipulated and returned for another run, however.
The calPlate uses a standardized 48-well plate format—with 32 of the stainless steel wells available for the user, the remainder used by the system for thermodynamic internal controls—making it amenable to automation. While this allows it to offer a higher throughput than existing calorimetry equipment, Dr. Jansson sees the system as having a role more in secondary screening and validation than high-throughput screening (HTS).
The Power of Three
When confronted with screening large panels of potential cancer drugs, HTS is definitely the way to go. Yet just about any single screen will fail to identify prospective candidates or tag those that are ineffective. “At the moment everyone just screens on 2D, so they’re just looking at cells stuck down on the plate, 48 hours later after putting a drug in, and seeing how many are dead,” noted Gareth Griffiths, Ph.D., science director at Imagen Biotech. “Of course, in the body, that’s not how it behaves at all—tumors are in 3D masses…and a drug that looks good on plastic doesn’t necessarily mean that it will be good in vivo.”
Imagen, a CRO, has developed an automation-friendly screen that combines three assays to provide complementary information not available from a single assay. The information garnered, he noted, affords “an extra dimension.”
Cultures are subjected to a standard 2D viability assay to determine the health of the cells two days after treatment with the test compound.
Similar assays are also performed on cultures created in hanging drops. The cells form 3D spheroids that do not come into contact with plastic. Some compounds are unable to penetrate the mass, and kill only cells exposed to the surface. “In the 2D assays it looks like it kills everything, but there of course it’s in contact with everything,” Dr. Griffiths explained. “So that gives you a bit more information about how it relates to perhaps an in vivo situation.”
The company also performs assays on cancer stem cell cultures. Single or a few cells are used to seed hydrogel-coated microtiter plates in a specialized nutrient-limited media to prevent adhesion and to insure that only the cells with stem-like properties will grow. Compounds are immediately added to the wells, and the cells are allowed to incubate for 14 days. The resultant spheres are fixed, stained, and analyzed.
All the assays are image-based, using a high-content screening system—essentially an automated fluorescence microscope. The algorithms for data analysis, though, “are the cleverest part of that,” Dr. Griffiths explained. “The software is able to take all the images and rapidly deduce how many cells are in each spheroid, for example, and how many of those cells are dying, what size the spheroid is, the texture, or morphology, things like that.”
First, the Normal
Mona Shehata, Ph.D., is interested in understanding breast cancer. To help her do so she studies normal, primary breast tissue.
John Stingl’s group at Cancer Research U.K. has been working on characterizing the main populations of breast cells—basal, luminal, and stromal. These facile distinctions do not tell the whole story, though. Some of the luminal cells express the estrogen receptor (ER+) and some are ER-, for example, and some can differentiate to synthesize milk, explained post-doc Dr. Shehata: “It stands to be a lot more complex that what was previously thought.”
She started with single-cell suspensions of cells of mouse or human breast tissue and depleted them of cells bearing endothelial- and hematopoietic markers (“I don’t want to look at nonspecific noise”) and sorted using flow cytometry based on expression of markers such as EpCAM and CD49f expression, known to distinguish basal from luminal cell populations. These, in turn, were screened for expression of about 300 cell-surface markers using BD Lyoplate™ panels, along with markers based on previous studies.
“By doing so we were able to identify further sub-populations within our general basal and luminal markers,” Dr. Shehata said. Among these are several cell-surface markers not previously known for breast. Subpopulations were subjected to various in vivo and in vitro assays to further determine their properties such as differentiation, proliferation, and gene expression. “We now think that there might be three or four different luminal cell types, each having a slightly different function.”
Now, she exclaimed, “we can start to do really cool experiments.”
Her lab is interested in tracing the lineages of the various breast cell populations: “What cell types give rise to daughter cells? Which type of daughter cell?” she asked. Cells can be separated based on their cell surface markers, tagged, and followed in repopulation studies, in mice, and in artificially created 3D culture “mini-breasts”.
Other members of her group are trying to re-create cancers in similar models using cells with oncogenes or tumor suppressors. There are five known molecular subtypes of breast cancer. It now becomes possible to ask whether the different subtypes of breast cancer arise from a single cell type, or whether they can be correlated with the different cell types found within the normal breast.
In Vitro to In Vivo
When all is said and done, an in vitro assay—no matter how sensitive, accurate, or carefully controlled—cannot always predict what happens in the animal model, let alone in a human. Similarly, results from animal testing are not easily translated in ways that can be corroborated by cost-effective, high-throughput in vitro assays.
Xavier Leroy, Ph.D., leads the research program in GPCR drug discovery at Actelion Pharmaceuticals, “from the beginning to entering Phase I.” He wanted the same readouts in 384-well plates as in mice, to move back and forth between in vivo and in vitro.
Actelion has been developing an assay based on splitting luciferase in two and fusing each part of the enzyme to proteins of interest. “As the two proteins interact together you can have a functional luciferase: add [the substrate] luciferin and you can detect light,” he explained.
The system “is universal. It has been in use for more than fifteen years, but not necessarily for drug discovery,” he pointed out. “I apply it for GPCR and to accessory proteins” such as beta arrestins, to screen potential agonists and antagonists.
The assay is inexpensive, straightforward, and fast, and works with both adherent and suspension cultures: just incubate the luciferin with the compound-treated cells for five minutes and read in a standard luminescence reader. There is no need to wash or lyse the cells. In addition, it can be read in real time, as opposed to the beta galactosidase assay on which it is based.
Yet the principal reason Dr. Leroy is adapting it to luciferase is “to have a system that can be studied in animal models,” which cannot be done with colorimetric assays, including fluorescence, he said. “Otherwise we never know if our drug actually will find our target in vivo, and at which levels, and whether it can block the target.”
Images presented by Dr. Leroy showed luciferase activity in subcutaneous tumors, recorded by an in vivo bioluminescence imaging system, after mice were treated with active compound but not vehicle.
Another advantage: “we don’t have to sacrifice the mice—we can reuse them,” he said.