By John S. MacNeil
For cell-based assay technology, screening potential drug compounds for activity in disease-related cell lines has traditionally served as its raison d’etre. Like many technologies with applications in drug discovery and development, the advent of automated and multiplexed cell biology assays in the early to mid-’90s was met with optimistic pronouncements of how cell-based assay systems would kick-start the commercialization of new drugs.
Cell-based assays are still a stalwart of pharmaceutical industry laboratories, but more recently the technology has taken on a new challenge within pharma — and academia as well: helping to determine the effects of RNA interference experiments. By combining the power of RNAi to accurately knock down specific gene sequences with the throughput and insight obtained from cell-based assay systems, pharma researchers are learning more about potential drug targets faster.
The combination shouldn’t be altogether surprising. Typically, pharma researchers studying disease pathways identify one or more proteins that, in theory, when removed from the system should short-circuit progression of the disease. To prove their hypothesis, pharma scientists could try to design an antibody for each protein target, and investigate whether inhibiting the protein slows or eradicates the disease. But a faster and easier (and therefore cheaper) approach to removing the protein from the system is to destroy the mRNA encoding the protein of interest.
Enter RNAi. By designing a short-interfering RNAi reagent (siRNA) or short-hairpin RNAi reagent (shRNA) that corresponds to the target gene sequence, pharma researchers can rapidly determine whether the hypothetical drug target is worth the expense of designing a drug compound. Of course, things are not always that easy. RNAi reagents are known to have off-target effects, which can complicate the experiment because the researcher may not be able to identify the off-target effects without comprehensively examining the cell system — a tedious endeavor by traditional cell biology.
That’s where cell-based assay systems come in. The approach, also known as high-content screening, is designed to automate and multiplex one or more biochemical experiments across multiple samples — typically arrayed in 96- or 384-well plates. Imaging robotics and algorithms for digitizing the image data then allow scientists to quantify and compare the results of the assay across the samples arrayed in the wells.
Separate But Equal
In practice, combining RNAi experiments and cell-based assay technology can take different forms in a pharmaceutical company laboratory. At the Novartis Institute for Biomedical Research in Cambridge, Mass., scientists have built on the experience of their colleagues using cell-based assay technology to screen for compound activity, says Craig Mickanin, a senior scientist at NIBR. Although his efforts to direct target identification and validation experiments that rely on RNAi and cell-based assay systems are completely separate from the compound screening operation, he says the two groups communicate closely and complement each other by sharing their experiences with the technology.
In fact, because Mickanin’s group is attempting to identify the most promising drug targets with the help of RNAi experiments, his scientists’ efforts should in theory lead to fewer compound screening experiments based on high-throughput cell-based assays, Mickanin says. This is because high-throughput compound screening is often employed as a way to determine the effect of a potential drug on a cell system even when the target is unknown, he says. If the target were already identified and validated, Novartis researchers could proceed immediately to directed biochemical experiments aimed at designing a small molecule tailor-made for the target, he says.
The strategy for devising an RNAi/cell-based assay experiment at NIBR typically involves working closely with colleagues focused on specific diseases to modify and optimize the mammalian cellular system representing the disease, Mickanin says. Choosing between shRNA and siRNA reagents often involves a trade-off between the prolonged effects of shRNA reagents, given their ability to stably integrate into the host-cell genome, and the cost savings and greater availability associated with using siRNA reagents, given their relative maturity in the marketplace, he adds.
Optimizing transfection reagents for delivering the RNAi reagents into the cell system presents another challenge, Mickanin says. There are a number of commercially available systems for transfection, but when his group is confronted with a new cell line, NIBR scientists must spend time tinkering with the experiment parameters to find the optimal protocol for successfully transfecting the RNAi reagents. Likewise, miniaturizing the experiment prior to implementing it on a 96- or 384-well plate system and developing a robust readout protocol are additional obstacles his group must overcome.
Mickanin says his group utilizes robotic platforms and imaging systems from a variety of manufacturers, and that his scientists have taken advantage of the technological advances that vendors have achieved in recent years with respect to the ability of imaging systems to extract and quantify useful biological information.
“Screening with siRNAs is essentially performing high-throughput cell biology, and one of the most powerful tools in cell biology for years has been microscopy. There existed a limitation prior to the commercialization of this technology to be able to apply microscopy, but to a larger number of samples, and in a smaller, miniaturized format,” says Mickanin. “So a number of different high-content imaging platforms have come out in recent years, and I think that high-content imaging is particularly powerful in analyzing the results of screens using genomic reagents, such as siRNAs or cDNAs.”
At Wyeth Pharmaceuticals, scientists employ RNAi experiments to identify and validate drug targets in a number of cell systems, including those reflective of inflammation, oncology, and cardiovascular disease, says Christopher Miller, associate director for applied genomics and biological technologies at Wyeth. Once researchers in Miller’s group choose the cell line that most closely approximates the disease of interest, the next task is to select the RNAi reagents best suited to knocking out the targets under investigation, and then choose or design a cellular assay that produces a readout conducive to confocal microscopy.
Often the readout can be quite simple. In oncology studies, for example, Miller says cell death, or apoptosis, is an easily identifiable parameter for automated imaging systems. Alternatively, Wyeth researchers can design the assay to measure more complex changes in the cell, such as variations in the shape and texture of cell membranes, or the level and distribution of specific fluorescent-tagged molecules. In terms of hardware, Miller’s group at Wyeth uses cell-based assay systems manufactured by Cellomics — specifically, the ArrayScan VTI.
Designing an effective RNAi/cell-based assay experiment also hinges largely on the informatics associated with the imaging system, Miller says. Algorithms for quantifying visual data — whether developed by outside vendors or in-house — require optimization to ensure they’re producing accurate numbers, as well as statistical software to allow scientists to trawl through data collected from past cell-based assay experiments, he says.
Mickanin at NIBR seconds Miller’s emphasis on informatics as one of the keys to successfully implementing an RNAi/cell-based assay experiment. With the power of high-content imaging to pick up minute changes in cell behavior, scientists must rely on algorithms to quantify these variations, and place them in the proper context, he says. And given the ready availability of open-source imaging software, in conjunction with vendors’ efforts to improve the quality of their own algorithms, Mickanin sees the future of cell-based assay technology in a rosy light: “I don’t think there are many cellular phenotypes that we can observe that can’t be broken into numbers and quantified.”