August 1, 2011 (Vol. 31, No. 14)
Keeping Cells “Happy” and Healthy Helps Support Complex Multiplexed Assays
Maintaining cell viability throughout the cell culture process is virtually a nonissue in current drug discovery efforts. But with the move toward multiplexed assays, and industry’s desire for higher content without compromising throughput, it’s worthwhile to think of cell viability along a spectrum and to consider cell health and responsiveness in assay development, such that cells are appropriately responsive to what a given assay intends to assess.
Informa’s recent “Cell-based Assays” meeting in Berlin focused on “novel advances in assay development to improve lead discovery and increase correlation to in vivo and clinical results.” This is, of course, a key ongoing concern for drug discovery research—to ensure, as much as possible, that the investments into screening stand a good chance of eventual clinical success.
Marc Bickle, Ph.D., head of the HT technology development studio (HT-TDS) at Max Planck Institute of Molecular Cell Biology and Genetics, observed that basic culturing principles must be observed at all times.
Speaking about phenotypic profiling of primary human macrophages in automated microscopy screens during mycobacteria phagocytosis, he noted that, with macrophages, monitoring healthy morphology is crucial to successful experiments. Poor cell health can create false positives, given that mycobacteria highjack cellular metabolism for their own survival.
Dr. Bickle’s laboratory thus monitors cell health on many fronts. “For viability in highcontent screens,” he said, “images provide the easiest measure for detecting DAPI- or Hoescht-stained nuclei for simple counting” to monitor for polyploidy or micronuclei (indicative of carcinogenic or teratogenic effect or other disruptions of cell cycle) or measuring aberrant nuclear morphologies.
His talk also covered mycobacteria survival measurement using high-resolution imaging on a PerkinElmer Opera System and monitoring of intracellular survival of Mycobacterium bovis, with simultaneous measurement of cellular responses of human macrophages.
With macrophages, in particular, there is the technical challenge posed by the use of postmitotic cells. Given that they do not grow, change in the numbers of cells is not an assessment option, so it is essential to quantify morphology as a cell-health indicator using multiparametric image analysis.
Also, for assays to be meaningful, cells must retain their responsiveness after preparative manipulations. Dr. Bickle described Cellectricon’s Cellaxess HT® as a game changer for research in his laboratory.
Other systems his lab employed, namely electroporation in suspension or the use of shRNA, consistently caused cell death or macrophage activation, he reported. “With the Cellaxess system, the cells are perfectly happy,” requiring minimal manipulation with the automated platform that obviates both transfers to fresh plates and additional washings.
Dr. Bickle stressed the importance of extracting data correctly from assay readouts, noting that it is important either to analyze signaling pathways in conditions of cell health, or if stress conditions are being studied, to understand the parameters of the stress conditions. Further, he highlighted that common assessments of cell health, such as doubling time, are in essence physiologically removed from clinical pertinence.
“Very few cells in the body actually divide rapidly, so if you want to work with welldifferentiated cells expressing markers pertinent to a specific organ and function, you need to slow them down in the cell cycle—to take away the growth factors, for example, and provide a better matrix such as a collagen lattice or some kind of gel.”
HT-TDS is currently developing a toxicology assay that Dr. Bickle described as novel in that it assesses toxicity not via primary indicators such as cytochrome function but instead via secondary markers such as hepatocytic function. His team grows cells in a sandwich culture between two collagen layers to induce differentiation and then monitors for “caniculae formation, full polarization, and expression of normal hepatic markers.” As Dr. Bickle summarized, “If somebody is sick, he is not going to work. The same is true for cells.”
ZFN and Reporter Cell Lines
Abhishek Saharia, Ph.D., product manager for functional genomics at Sigma Life Science, presented data on the company’s CompoZr® Zinc Finger Nuclease (ZFN) technology.
The technology allows researchers to bypass the protein overexpression inherent in many assays that otherwise creates “a system that doesn’t truly resemble the patient,” Dr. Saharia said, given that such assays involve exogenous promoter systems.
The ZFN platform comprises engineered DNA-binding proteins that induce doublestrand breaks in DNA at user-specified sites, stimulating natural DNA-repair processes. The end result is “precisely targeted genomic edits” resulting in cell lines (including somatic) with targeted knockouts and integrations, Dr. Saharia explained.
Alternative systems may provide weaker data, he suggested, given that knockdown strategies are often partial and can lead to “incomplete effect.” Further, the ZFN platform enables the creation and assaying of robust human cell lines using any protein tag that has been endogenously incorporated into a specific locus of interest.
Dr. Saharia offered the example of use of the technology with the epidermal growth factor receptor (EGFR) protein, tagged with GFP, at the endogenous locus. “The advantage is endogenous levels of protein expression, with complete responsiveness based on endogenous promoter systems and regulatory elements.
“This puts researchers in a position to track protein localization and expression levels in a live cellular system that maintains all its natural responsive elements. Previously similar pathway responses could only be verified by endpoint experiments, either by Western blots or FACS analyses, but the ZFN system permits this visualization in a live cellular system.”
This kind of targeted disruption may prove fruitful for advancing personalized medicines, Dr. Saharia further suggested, given that it permits live visualized tracking of small molecules that disrupt specific signaling pathways such as activation of EGFR and can thus facilitate high-content screening.
Scaling HCS to HTS
Fabio Gasparri, Ph.D., principal research scientist at Nerviano Medical Sciences, stressed the importance of tending to basics. Nerviano conducts HCS assays to evaluate mechanisms of action of lead compounds affecting regulation of cell cycle or apoptosis.
Cytostatic or cytotoxic effects are determined through specific marker assessments. For confounding compound classes, such as those inducing polynuclear formation or that increase cell volume, which in turn can affect monoparametric readouts such as ATP content, his lab applies alternative methods such as cell count to better determine the real effects of lead compounds.
Currently, the lab employs time-lapse microscopy among its tools. “The time variable is crucial for better evaluation and good characterization of lead compounds. Focusing only on endpoint assays has some limitations; for instance, they may not clearly distinguish between cytostatic and cytotoxic effects.”
Dr. Gasparri is looking toward adoption of technologies that will further strengthen the robustness of Nerviano’s R&D efforts. He said image-based technologies, like TAP Biosystems’ cell- IQ® platform or impedence-based technology such as Roche’s xCELLigence system for monitoring real-time cell proliferation in plates, “could be useful to specifically monitor cytostatic or cytotoxic effects and improve early characterization studies.”
Dr. Gasparri’s lab currently conducts highcontent mechanism-based assays using the Thermo Fisher Scientific Cellomics ArrayScan HCS reader.
“In particular, we monitor the regulation of the cell cycle in response to targets we’ve identified at Nerviano related to cell-cycle control and to signal transduction. We employ assays based on immunofluorescence or reporter cell lines stably transfected with GFP proteins. Recently, we’ve been performing primary HC screening assays and are integrating these to analyze multiparametric data arising across different HC primary screens.”
Ultimately, this is all intended to increase the predictive contribution of earlier in vitro data gathering toward in vivo outcomes. After all, he insisted, when discussing in vivo work, “we are talking about models, so even as we try to increase the relevance of our models, they remain models.”
Stem Cell-Based Assays
Davide Danovi, M.D., Ph.D., research associate in Steven Pollard’s laboratory at the University College of London’s Cancer Institute, described approaches to carry out live image-based chemical screens using glioma stem cells.
Pollard’s laboratory has developed a method to isolate cells from normal brain tissue and also from brain tumor (glioblastoma) samples. The cell culture protocol essentially involves defined medium with growth factors and the use of laminin, which avoids cell suspension (and thus the formation of neurospheres).
“Neurospheres can be used for different applications, but we argue that if you want to screen for drugs to cure glioma, you need to test drugs on the relevant cell population, which is the stem cell-like population,” Dr. Danovi explained.
The lab then applies image-based assays to quantitate morphological changes and assess proliferative behaviors, “in order to be able to isolate drug leads that specifically affect a particular subtype of glioma, targeting glioma stem cells versus normal neural stem cells.”
Modules from various sources are used, including CyBio’s Cybi®-Selma liquid-handling device, an IncuCyte image-based platform from Essen BioScience, and Cell Profiler freeware from the Broad Institute.
In-house bioinformatics efforts led by Amos Folarin allow “proper image analysis, so we can track cells, morphological changes, and mitoses, for example. Our approach uses proliferation as a readout to score compounds that are cytotoxic or cytostatic.” Among cytostatic compounds, those that induce terminal differentiation of glioma stem cells are considered for potential glioblastoma therapies.
Reading Cells in Complex Assays
Eberhard Krausz, Ph.D., director for assay development and target validation at Janssen Research & Development (a division of Janssen Pharmaceutica), discussed two research tracks, one focused on protein-protein interactions and the other on the use of a 3-D co-culture assay for tumor microenvironment assessments.
The kind of assay approach he described as “more accurate and increasingly important” is to monitor cell viability concurrently (not serially or in parallel) with target binding events.
For instance, several current approaches for protein-protein interactions “show directly in cells how compounds penetrate, find the target, and interact with other targets,” he said. One particular approach Dr. Krausz presented is a high-content screen that anchors one of the two proteins under study, either in the nucleus or other specific locations such as at a particular DNA site.
Simultaneous assessment of “stressinduced phenotypical change such as morphological changes give an idea that the cell is either under stress or apoptotic. In particular HCS allows multiparametric data analysis to extract the direct target information as well as information on any kind of influence of ‘cellular happiness’ that could indicate off-target effects that would create potential problems later in follow up and development of any compound that comes out of such screen.”
Janssen R&D has also developed a 3-D tumor-growth assay that “has certain advantages over 2-D classical cell-proliferation assays,” Dr. Krausz described. “We know from gene-expression studies that tumor cells under assay exhibit changed behavior in ways similar to tumors in human patients. Ideally then you would want to have tumor cells in co-culture that mimic the in vivo tumor microenvironment, which would then include normal, nontumor cells such as fibroblasts and other cell types.
“Typically there are interactions between tumor cells and nontumor cells. Disruption of these interactions could allow us to exploit new options for targeted therapies that far exceed typical antiproliferative strategies,” he said.
“A very important point with this assay is that we can discriminate between simple cytotoxic compounds and those that show promising potential mechanisms of action through our ability to count individual cells or colonies and discriminate according to colony size. Essentially we seed tumor cells along with primary mesenchymal stem cells in a matrix, treat them, incubate, and then count colonies.
“Individual tumor cells will either divide to form minitumors or remain in the singlecell status. With noncytotoxic compounds, the total number of colonies does not change because the cells experience only growth inhibition,” which Dr. Krausz described as a particular strength of this assay approach.
Dr. Krausz sees developments in live-cell assays, such as 3-D assays and co-culture, as having a “bright future on many fronts,” not only for disease areas but also in ADMET. Ideally, he posited, their utility would come with a drastic reduction in the number of compounds actually being screened but in more complex assay platforms.