Support for lead optimisation and target validation

High-Content Analysis (HCA) provides a drug discovery tool capable of rapid screening of drug effects in pharmacologically relevant cell culture systems. Interest in HCA has been increasing during the past few years. This reflects the confidence that HCA-technology has established due to the stability and reliability offered to the drug discovery process. HCA offers the capability to support an experienced and open minded cell biologist in challenging the current limits of cell biology. HCA is a versatile tool providing statistically secured data of cellular and subcellular events, respectively.

The cell is the smallest functional unit of an organism and only in cellular systems pharmacologically and pathophysiologically relevant mechanisms can be studied efficiently. Therefore, a clear demand on cell culture systems exists during preclinical development in modern drug research. Cell culture systems offer a detailed insight into cellular systems, which is indispensable during actual modern compound development (Figure 1). This detailed insight into cellular systems can be analysed by modern HCA-systems, which enable the rapid automated acquisition and the subsequent statistically secured analysis of structural and metabolic changes within the cellular context. Such HCA systems offering automated and highly sophisticated, high- throughput image analysis can be applied within a HTS department for the support of rapid and effective drug screening. Cellular assays based on high-resolution microscopy have been limited to basic research approaches so far, but HCA is a tool that enables the rapid screening of drug effects in pharmacologically relevant cell culture systems and provides scientists with high-quality information about targets and potential candidates within the cellular context that is needed for effective drug research.

Using such a tool a drug researcher finds himself in the advantageous situation where he can gain information about the interaction of a compound within a biological system. Furthermore, he will be able to obtain information about unpredicted interactions within the cellular context. HCA-assays are applicable throughout nearly all stages of the drug discovery process. They can be applied to target validation, lead discovery and lead optimisation as well as to detailed functional studies before compounds are entering clinical development. HCA-assays have the potential to accelerate the drug discovery process by opening existing bottlenecks that can be found in previously cited stages of the drug discovery pipeline. They open these bottlenecks by performing higher-throughput cell based assays that provide multi-dimensional, multi-parametric, high quality data. HCA-assays offer great time-saving potential by allowing the establishment of multilinear designed assays. They can help to identify genes and proteins that are involved in diseases. And HCA-assays are also capable of qualifying the drugability and the therapeutic benefits of selected compounds. Importantly, HCA-assays can identify possible side effects of a drug candidate within the cellular context very early during the drug discovery process, offering potential for significant cost-saving.

High-Content Analysis at Schering AG

About four years ago a subcellular imaging project was started at Schering AG, which should evaluate and implement HCA. If applicable, HCA should be integrated into the lead discovery process as a novel application that offers highly sophisticated assays capable of analysing drug impact on the single cell level, as standard applications to improve the drug discovery process. HCA is an integral part of the drug finding process at Schering AG and our HCA-group offers a broad panel of HCA-applications, which can be used during Target Validation and Lead Optimisation.

The goal of the subcellular imaging project was the establishment of cellular in vitro screening procedures yielding a higher informational content. Cellular model systems were used with respect to scientific needs that had come from the various pharmacological indications of our facility. The use of microtiter-plate based high resolution and high-throughput microscopy allows the acquisition and analysis of multiple subcellular parameters on the single cell level in parallel. Image analysis routines can be individually modified and adapted to current experimental requisites and have been developed on an assay-by-assay basis. Such routines have the ability to distinguish between subpopulations in mixed cell culture samples. They provide statistically secured data, which have the potential of leading to reliable scientific conclusions. The intention was to establish a panel of highly sophisticated HCA-applications that should have the potential to yield an important quality improvement to the drug discovery and development process, according to the previously described approach.In order to implement HCA at Schering AG a specialised scientific lab group consisting of two technicians and one scientist was established. Molecular Devices’ Discovery 1 was the HCA-system used for this approach. The D1 system provides medium-throughput microtiter-plate acquisition capabilities and consists of an inverse microscope, based on standard fluorescence technology and capable of high-throughput, high-sophisticated image analysis. The D1 system comes with an image analysis suite called MetaMorph. This image analysis tool provides the HCA-user with a group of preprogrammed complex image analysis modules and furthermore allows the individual adaption and programming of image analysis routines by using the integrated graphical editor. This solution offers a great degree of freedom concerning a personalised image analyis approach. This is indispensible to meet the very varying requisites, which are coming from the various pharmacological indications. Initially, such a personalised HCA-approach requires intensive skill adaption training. This investment in time and specialist personnel pays off when HCA is confronted with novel and challenging applications. Having decided to go for HCA by this approach, we were able to establish a broad panel of applications and a broad panel of cell culture systems during the course of four years. These are currently available as established and reliable standard HCA tools. Our HCA applications are highly requested during the lead discovery process. During critical and difficult periods of a lead discovery project HCA-assays are highly recommended due to their reliability and better biological relevance compared to standard cellular assays. HCA assays provide better statistically secured data on the single cell level, yielding a better decision basis. At Schering AG 16 different HCA applications are in place and can be applied to 14 different cell systems. The volume of generated HCA-data is growing in an exponential way (Figure 2). This enormous increase goes hand-in-hand with increasing requests concerning the use of HCA-technology during Target Validation, which is driven by RNAi knock down studies that ask for phenotypical characterisation provided by HCA applications. Project-related functional HCA-assays are increasingly requested during Lead Optimisation, which are looking for a subcellular resolution to allow the analysis of protein-protein interactions on the subcellular level. This also contributes to the pre-described data explosion, due to the fact that it requires more images of higher resolution.

Utilisation of AFPs and Live Cell Imaging

AFP-(Auto-Fluorescent Protein)-technologies enable the analysis of subcellular protein distribution patterns, target translocation and protein-protein interactions during cell cycle progression within living cells. Thus, live cell imaging allows the tracking of dynamic processes in real time on single cell level and gives new perspectives to the investigation of intracellular pathways.

AFPs were established from native bioluminescence proteins. Such proteins can be found in certain kinds of organisms such as insects, jellyfishes, reef corals or fishes. Mutated forms of these proteins are commercially available and have the ability to absorb and emit light of specific wavelength depending on their molecular structure. The most popular and well investigated autofluorescent proteins are proteins from the GFP family (Green Fluorescent Protein) that have been extracted from Aequorea victoria jellyfish. Point mutations in the aminoacid structure of GFP lead to the blue and yellow shifted variants CFP (Cyan Fluorescent Protein) and YFP (Yellow Fluorescent Variant). Since also red and far red shifted autofluorescence proteins were established, the whole spectrum from 400nm to 750nm can be used for AFP applications.

Theoretically, up to four proteins, each tagged with one AFP of a different spectrum, can be visualised and analysed in a living cell at the same time. But the broad absorption and emission spectra of the AFPs lead to fluorescence crosstalk. Therefore, one is limited to a smaller set of simultaneously used AFP-tagged target proteins. The combination of optimal absorption and emission filter sets as well as optimal light sources can minimise these crosstalk problems and has to be calculated for each application. Otherwise, some applications in living cells call for crosstalk that exists between specific AFP combinations. For example, CFP and YFP enable the measurement of fluorescence resonance energy transfer (FRET), when labelled interacting partner proteins are coming in close spatial contact. The energy of the excited donor fluorophore CFP overlaps with the energy needed to excite the absorber fluorophore YFP and photoenergy is transferred. This results in a loss of fluorescence intensity of the donor CFP and to the emission of the acceptor YFP. FRET can be used for detection and subcellular quantification of direct protein-protein interactions in living cells since closed proximity of a few nanometers is essential for the emergence of an energy transfer.

The visualisation of target proteins in living cells enables one to analyse the specific functions of these proteins under well documented cell culture conditions. At first, the intracellular distribution and localisation of a distinct protein during cell cycle progression can be determined and analysed. Cell cycle dependent localisation processes of target proteins provide an insight into aspects of their native functionality. In addition, the dynamic of this distribution process within living cells and the resulting response to cellular perturbations like drug treatment can be analysed using FRAP-(fluorescence recovery after photobleaching)-technology. In FRAP experiments the fluorescence of an AFP within specific cellular compartments is irreversibly photobleached by an intensive laser flash. The kinetic of fluorescence redistribution into this area can be measured by fluorescence recovery. This method offers the possibility to study the mobility of the target protein and the involved transport processes within a living cell.

Live Cell Imaging at Schering AG

In our HCA-group live cell imaging experiments are deployed for target validation and lead optimisation within the drug discovery process. Using cells that have been transiently or stably transfected with AFP-tagged target proteins, such cellular assays shall shed light on the functional validity of target or a target oriented compound. For example, during target validation, a novel kinase target showed an anti-apoptotic impact on cells over-expressing this protein. In those cells the AFP-tagged kinase localised primarily in the nucleus but also in the cytoplasm during interphase (Figure 3). Time-lapse studies revealed that during cell cycle progression with the onset of metaphase, an accumulation of the AFP-tagged kinase in spindle pole regions, as well as in regions next to the condensed DNA, became observable. Driven by these results a suitable substrate protein involved in the kinase dependent anti-apoptotic effect was selected. The perturbation of the enzyme-substrate interaction after compound treatment revealed by AFP-colocalisation or superior by AFP-originated intracellular FRET signals will serve as functional read-out and analysis parameter. Furthermore, off target effects generated by the compound leading to cellular toxicity can be easily identified by strong phenotypical and morphological changes. During lead optimisation such HCA-models will be used increasingly for analysing efficacy and selectivity of new compounds on target proteins in living cells to substantially support the lead structure finding process.

Conclusions

Although fluorescence microscopy is not really a novel tool in scientific research, High-Content Analysis has continued to establish its place in pharma as well as academic institutes. This is because HCA combines in a powerful way the advantages of automated high-resolution fluorescence microscopy and stable, automated high-throughput image analysis, driven by modern computer technology. Hence, this novel cell analysis system leads to reliable, statistical secured data building the well-founded basis of solid scientific conclusions. In that way complex but intelligible HCA-systems can help cell biologists significantly during the analysis of signal transduction pathways.

By analysing target activity at the single cell level HCA provides functional data on the single cell level with high statistical relevance. It would be best if this comes along without forcing users to have extensive skills in scripting, but provides the degree of freedom in scripting that is needed scientifically. This is crucial for HCA to be recognised as a comprehensive and easily-accessible tool for real high-content and functional assays. It will almost certainly lead HCA to become an essential part of the drug discovery process.

Acknowledgements

We greatly acknowledge the work of Alexander Dimerski and Sebastian Schaefer whose practical contributions made this article possible. We would also like to thank our colleagues and mentors for fruitful discussions that furthered the progress of HCA at our facilities.