October 15, 2007 (Vol. 27, No. 18)

Susan Aldridge, Ph.D.

Technology Complements HCS and HCA to Yield Better Accuracy Early in Development

High-content screening (HCS) and high-content analysis (HCA) are important elements of drug discovery and development. There is no hard and fast dividing line between the two, but HCS is generally higher-throughput, while HCA has an emphasis on gaining the maximum information from an assay.

Both HCS and HCA generally involve a combination of automated high-resolution quantitative image analysis combined with an informatics element that is applied to the output of cell-based assays or to preclinical work.

Image analysis inevitably plays a key role in extracting information from a compound–cell/tissue interaction in either an assay or an animal model. It can therefore make a vital contribution to reducing the high rate of attrition currently being experienced by the pharmaceutical industry. In recent years, automation of image analysis has begun to make a major contribution.

The Cambridge Healthtech “High-Content Analysis Europe” conference, held in Prague in September, looked at examples of case histories and technologies that have recently been brought to bear in HCS and HCA. Further discussion on the role of image analysis in HCS/HCA took place at Informa Life Sciences’ “High Content Analysis” conference held in Vienna earlier this year.

Quicker Assay Development

The large-scale adoption of these approaches depends very much upon efficient assay development. A major factor for this is optimizing the image analysis side to create both robust and relevant measures of the biology involved, according to Mark Collins, Ph.D., marketing manager cellular imaging Thermo Fisher Scientific (www.thermo.com). He says that the approach taken to rapid assay development by subsidiary Cellomics can help overcome bottlenecks in HCS/HCA.

When a high-content assay is carried out, an image is taken of the cells and the associated biological events. This has to be quantitatively measured with image analysis. “Many surveys have shown the importance of rapid optimization of an assay once its biology is working,” said Dr. Collins. “So turning the biology into meaningful numbers is the real bottleneck.”

Thermo Fisher Scientific is good at solving this challenge with its software, Dr. Collins added. The company offers rapid assay development with the Cellomics® intelligent acQuisition (iQ) system. This is an iterative data-driven approach to creating an assay.

“When building an assay with Cellomics iQ, you would be looking at the image while the system calculates the numbers, whereas with the old approach you take the picture and then do the calculations off-line,” explained Dr. Collins. In other words, he noted, it is rather like the difference between a digital and a film camera. With the latter instrument you only view the data once the film is developed. With the Cellomics approach once the scientist knows they have a good picture, it can be optimized to deliver the right data straight away.

“It is iterative and the user does not need to be an image analysis expert to tune the system to get the best outcome,” Dr. Collins added.

An advantage of iQ is that it can be used in a rare-event assay for genotoxicity as it assures that sufficient evaluations are done in the assay to satisfy FDA guidelines, Thermo Fisher Scientific noted. Another application is in cell cycle analysis, where iQ can be used to maintain an assay window even when the dose of the compound causes cell death. The system ensures enough images are collected because it carries out the acquisition and analysis at the same time.

In short, iQ changes the emphasis in an assay from time-to-data to time-to-decision. “This is what makes the productivity of the Cellomics approach greater than that of other platforms,” asserted Dr. Collins.

Leveraging Use Earlier

Dr. Collins believes that Cellomics’ method for rapid assay development can help address the well known crisis of productivity in the pharmaceutical industry. Too many compounds fail at a late stage because they are not effective with respect to their target, the ADME is wrong, or the toxicology models are incorrect.

“High-content techniques, which place everything into the context of the cell, really play into these three reasons for failure,” Dr. Collins explained. “A lot of data from customers now shows that high content can produce better predictive toxicology before clinical trials.” That is why HCS/HCA is now being adopted widely throughout the life science industry.

“We are trying to turn high content into something that is like a home theater in a box—a total solution,” Dr. Collins stated. Future developments will include 3-D analysis approaches and additional data analysis tools, which the company will provide through its collaborations with Spotfire (www.spotfire.com) and GeneData (www.genedata.com).

At Boehringer Ingelheim (www.boehringer-ingelheim.com) HCS is used to study the activation and internalization of chemokine receptors, according to Ralf Heilker, Ph.D., senior scientist, lead discovery. Optical developments have enabled rapid and automated processing for a large number of microtiter plates in this context. Sophisticated object recognition algorithms allow Boehringer Ingelheim’s HCS systems to carry out automated image analysis on an industrial scale.

Meanwhile a flexible programming interface enables Bayer Schering Pharma (www.bayerscheringpharma.co.uk) to establish specific image-analysis routines, reports Philip Denner, Ph.D., lead discovery scientist. These are used in the study of kinase-dependent substrate phosphorylation in the nucleus of eukaryotic cells and subsequent substrate translocation into the cytosol. This technique has been applied in secondary screening to determine the impact of kinase-specific inhibitors within a cellular context.

Along with image analysis, new methods of automated immunohistochemistry (IHC) can add speed and quality to preclinical work, Elke Persohn, Ph.D., head cellular imaging at Novartis (www.novartis.com), said at the Informa meeting. Novartis uses Ventana Medical Systems’ (www.ventanamed.com) immunostainer for IHC and Definiens’ (www.definiens.com) analyst software for image analysis. The software is based upon cognition network technology, basically using semantic networks of objects and their mutual relationships.

“These methods allow automatic detection and evaluation of sections from parts of organs that we are interested in, such as intestinal crypts or mucosa,” Dr. Persohn says.

In one experiment using this approach, the proliferation index of jejunum stained with PCNA and BrDU, standard stains for identifying dividing cells, was measured. The proliferation index is a way of detecting whether a test compound has the ability to increase cell proliferation in comparison to the corresponding control. This could indicate whether the test compound is a potential cancer risk and ought not to be further progressed to clinical development.

The company used the Definiens software for automatic measurement of the area of 20 crypts in the jejunum, counting the PCNA and BrdU positive and negative cells within. Previously, this would have been done manually under the microscope, which is time-consuming, tedious, and less accurate. In this manual process there is also no control over knowing if a particular cell has been counted or not.

In a second experiment, slides of four colon and four cecum sections from rats were scanned with a Carl Zeiss Microimaging Mirax scanner. The resulting images were automatically analyzed with the Definiens software. Novartis researchers measured the total mucosal area and counted BrdU positive and negative cells, leading to a calculation of proliferation index to see if there was a treatment-related increase in mucosal area and cell proliferation.

“These approaches make a difference in that we can not only work faster but we can also evaluate more tissue and more sections, which give us more data for more robust assessments of compounds in preclinical development,” concluded Dr. Persohn.

Image analysis is also useful in the study of phospholipidosis, as discussed by Ed Ainscow, principal scientist, Advanced Science & Technology Laboratory, at AstraZeneca (www.astrazeneca.com). Phospholipidosis is a drug-induced aberrant accumulation of phospholipid into lysosomal-derived multilamellar vesicles in hepatocytes and other cells types. A synthetic fluorescent phospholipid has been used to model this process in primary rat hepatocytes using automated image acquisition and analysis.

Finally, Petra Perner, Ph.D., director, Institute of Computer Vision and Applied Computer Sciences, at IbaI Leipzig (www.ibai-institut.de), talked about its flexible image analysis and interpretation system called Cell-Interpret. The instrument automatically interprets images in the way a human expert would. The system has been trained for pattern analysis on the well-established human cell line Hep-2.

Susan Aldridge, Ph.D, is a freelance science and medical writer specializing in biotechnology, pharmaceuticals, chemistry, medicine, and health. E-mail: [email protected].

Previous articleResearchers Find Oncogene Mutations Are Diverse and Vary in Each Individual
Next articleCharley’s Fund Grants AVI $2.45M for Duchenne Muscular Dystrophy R&D