Imaging-based screening methods applied to drug discovery – challenges and solutions

By PerkinElmer

In drug discovery, the trend continues to move away from target centric biochemical assays towards more phenotypic and “genomic” approaches to understanding a disease and finding ways to address it.

Why “High Content Screening”? Opportunities and Challenges

RNA interference (RNAi) allows high throughput genome-wide screening approaches for disease relevant targets in cellular assays. Pathway and phenotypic screening opens up a wealth of information which was not previously possible. The ability to investigate a complex biological response in a cell sample, such as, for example, a morphology change, provides the opportunity to gain new insights into the mode of action of a compound, even when the target is still unknown. It is also important to investigate compound action on the large number of known targets. Understanding the in vivo biology of those drug targets requires a large number of parameters to be monitored in parallel. Compound side effects must be detected and understood early in the drug discovery process in order to proceed only with those lead compounds which have a high probability of developing into a drug.

A common feature of all these approaches is that they rely on imaging based screening methods and the ability to make many diverse measurements on a cellular system using a suite of different assays. A multitude of readout parameters from each assay generate high information content. The methodology has therefore been summarized by the terms “High Content Analysis” (HCA) and “High Content Screening” (HCS).

HCS poses considerable technical and logistical challenges to drug discovery. Although high resolution imaging is already a commodity in basic cellular research there was and still is the difficulty of creating robust cellular imaging assays as well as dealing with the high volume complex data from image based screening. The various groups involved in a drug discovery process can be far apart. Pharmaceutical companies consolidate their screening facilities into a few locations worldwide and target identification labs can be in a different country. They may also cooperate with groups from Academia and with Biotechnology companies for specific projects at any point in the preclinical drug discovery chain; from target identification to early ADME/Toxicity studies. At the same time, the methods used in the various stages overlap more and more, and screening departments perform more and more downstream work. Hence the logistical challenges for data generation, data analysis and data interpretation vastly expand. Research groups need common tools to “understand each other” in order to move the projects ahead quickly because the timelines for each project tend to get shorter rather than longer and there no time can be wasted in order to align technology. The ideal tools show compatibility of hardware and software covering the whole range from research microscopy to high throughput imaging.

Instrumentation for HCA and HCS – needs and approaches

High Content Analysis starts with the first experiments in a biology laboratory, when a microscope is used to understand the cellular function of the molecules of interest and also to identify potential assay strategies for compound action. To aid this process, fluorescent proteins and dyes must be imaged in as much detail as possible.

We have chosen spinning disk imaging technology for high speed confocal measurements in our portfolio of microscopy and High Content Screening products. It enables high-speed live cell microscopy with minimized photobleaching and high throughput confocal screening with great resolution and background suppression.

High Content Screening itself has various instrument needs, depending on the application, throughput and step in the overall process. During the assay development stage, experiments are mainly performed in 96 well plates and with a few plates for each assay. Flexibility of fluorescence excitation and emission options for a lot of different dyes is essential as well as a high sensitivity for monitoring cellular receptors and signalling molecules at physiological expression levels. The High Content Screen applied in secondary screening or RNAi screening has to handle between 10,000 and 100,000 samples. 384 well plates dominate, but 1536 well plates are often a prerequisite when working at the higher throughput end. Sensitivity and resolution must still be high for many applications.

The Opera™ and Opera™ LX High Content Screening microplate readers from PerkinElmer Inc. fulfil the highest demands for High Content Screening, providing unrivalled resolution and speed. The use of a Nipkow spinning disk including a microlens enhanced pinhole array for confocal imaging is unique in high content screening. Lasers and a high power lamp cover the entire spectrum, from UV to near IR. Up to four CCD cameras take images in parallel detection channels. A wide range of objectives with different magnification can be used and a unique a utomated supply of immersion water enables the use of immersion objectives resulting in superior image quality and a significant reduction in exposure time compared to air objectives.

A wide variety of accessories can be added to the Opera, including plate handling, liquid handling, washing, centrifugation, incubation and plate storage. Our lengthy and successful experience in High Throughput Screening instrumentation means that we can create customized solutions for every need from our large portfolio of automation tools – from individual workstations up to full scale screening platforms.

Live cell microscopy applied to HCS

The ability to follow processes in live cells over time enables applications which cannot be addressed in a biochemical assay. In many cases, the time course of a reaction in a living cell is not known and must be investigated before it can be applied to High Content Screening. For example, detailed analysis of cell proliferation, differentiation and cytotoxicity needs instrumentation and software for assaying cell plates over extended time periods (hours and days).

During assay development, live cell imaging with environmental control can be advantageous because the time course of a cellular response must be followed in order to identify when the response is at maximum and when for the cell plates need to be fixed for analysis in a screen. The time course can be seconds (molecule trafficking), minutes (internalization, pathway activation), hours or even days (differentiation, cytotoxicity, genotoxicity).

The screen itself is typically run as an endpoint assay using fixed cell samples; when using fluorescently labelled antibodies there is usually no other option. However it can be advantageous to use live cells in a screen because this can make an assay protocol very simple and avoids many washing steps, which is critical for setting up a fast and robust high content screen. The advancements in fluorescent proteins and live cell compatible fluorescent probes have made such protocols feasible. In some cases live cells can be used in a screen without controlling sample environment. Running a screen at room temperature slows down transport processes in a cell giving a larger time window for screening. However, in many cases the live cell samples have to be kept under environmental control for reproducible results. Maintaining live cells at or near physiological conditions is a challenging but essential requirement for the success of many screening programs.

The Opera is available with an integrated environmental control which offers the best possible solution for fully automated high speed and high resolution cellular screening using live cells. Significant barriers to high throughput live cell imaging have been overcome with the addition of precision environmental controls. The entire plate receptacle, scan table and objective area are heated and temperature controlled, eliminating any temperature gradients within the plate that might arise from contact with instrument components. Sterile filtered air with controlled levels of temperature, CO 2 and humidity are circulated in the plate area. This high level of environmental control within narrow boundaries creates incubator conditions for screens in which an individual plate is in the Opera for several hours.

In a typical live cell screen, many plates have to be measured once or even multiple times. This can best be achieved by controlling the environment on the plate reader and transferring the plates to and from an incubator with robotic automation. For repeated measurements of plates over hours to days, we offer a complete Live Cell Workstation including an external incubator and plate handling, available with a plate capacity of several tens to one hundred or even more. This set-up is also suitable for screens in which the cellular response must be monitored over several days. High throughput live cell screening can be achieved using our scaleable cell :: explorer screening platform. Technical solutions for controlling incubation temperature, CO 2 and humidity as well as an automated lid handling will ensure ideal environmental conditions.

Image analysis for screening

The requirements for image analysis in High Content Screening are as demanding as the requirements for instrument hardware. Multi-step image segmentation, object identification and quantification procedures require time consuming computing but very fast feedback from image analysis is necessary for quality control. Results must be available immediately after the assay has run and on-line analysis is essential for an efficient screening process. High throughput image analysis therefore relies on software with fast data processing capability. Our Acapella™ image analysis software is designed for speed and runs even very complex algorithms as fully automated, on-line processes during image data acquisition. Acapella is optimized for multi-processor computers and is therefore fully scaleable from image analysis development up to high throughput image based screening.

Image analysis software for HCS applications should ideally include a large toolbox of image analysis modules and a wide range of ready-made applications which are easy to understand and to apply. The ability to adjust imaging algorithms to applications and cell types as well as the option to combine algorithms for multiplexed readout is also crucial. The powerful and intuitive new release of Acapella addresses these requirements and provides tools for every user, from the beginner to the expert. The software includes the Acapella Script Collection, a large library of ready-made image analysis solutions for frequently used HCS assays such as translocation, receptor activation/endocytosis, signal molecule recruitment, cell cycle, cell viability, and many more. Generic applications for multi-parameter analysis of specific features such as cell and nuclear morphology and spot detection are also included, making development of a new assay very easy. The solutions can be modified and combined as required as they contain modular building blocks for cellular object detection. Creating new algorithms based on these building blocks is very efficient.

How to manage data?

Image based screening poses specific challenges to data storage and management. Users are routinely collecting Terabytes (TB) of data and need to be able to effectively store, recall and analyse the data coming from various High Content Screening instruments as well as research microscopes and other image creating instruments. An industry standard image data management system scalable and cross-platform compatible is required for High Content Screening.

Columbus database products from PerkinElmer address the major issues of High Content data management. Complex HCS multi-channel images can be stored and accessed by multiple users and a very wide range of file formats from different sources is accepted. Columbus utilizes the OMERO server, developed by the Open Microscopy Environment, a multi-site collaborative effort among academic laboratories and a number of commercial entities creating open tools to support data management for biological light microscopy. As such, it is the ideal interface between various instruments and image analysis software available. Images are automatically transferred from, for example, one or more Opera systems and can then be viewed and explored by any number of users via their personal workstations. Each user has defined security settings with read/write/delete privileges for their own data and that of other users. In contrast to other HCS data management systems the metadata detailing how the images were acquired, which algorithm and input parameters were used to analyze the images, the output parameters and numerical results are all stored with the image data. This means that the complete history of any experiment can be reviewed when examining the image data, so that a better understanding of the results can be obtained.

High speed Acapella image analysis is fully integrated in the Columbus Conductor product. Image data sets can be analyzed even in batch mode with Acapella running as a 64-bit multi-threading server application.