Combining RNAi-based functional genomic screens with automated high-content image analysis of cells could become a new way for drug companies to identify novel targets, according to a recent study.
Earlier this month, researchers led by Marino Zerial, a director at the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden, Germany, and their collaborators published a paper in Nature using this approach to study the role of human kinases in endocytosis [Nature, Vol. 436, 7 July 2005, p. 78-86].
"The paper is phase number one to demonstrate that we could use functional genomics to interrogate novel pathways, such as in endocytosis, but also try to identify novel targets that could be utilized for chemical screens," Zerial told CBA News last week.
Zerial believes that large drug companies are becoming more and more aware of cell-based imaging screens, but have by and large been reluctant to adopt the technology.
"This is still an approach that relies on a lot of uncertainties," he said. "When you have a target, you know what you have, you screen against it. If you have instead a black box, you have no way to identify your target, and this is of course a problem. But by multiplexing technology, I think we can circumvent this problem," he said.
"These are examples [of what] you can only do when you do a multi-parameter type of screen where you look at multiple types of assays."
According to Ralph Garippa, a research leader with Roche Discovery Technologies in Nutley, NJ, a handful of drug companies, including Roche, are now starting to marry functional genomics and cellular image analysis. "I think the appealing part is the fact that one is able to do a gene knockdown but also get a functional correlate in a cell-based assay, and then to confirm that by visual characterization," he said. "Other techniques address parts, facets, aspects of it, but they don't address it as directly as the automated image analysis."
The authors of the Nature paper looked at the effect of silencing the entire human kinome — 590 human kinases, one at a time — in a human cell line on the ability of viruses and other molecules to enter the cell.
In a first step, they focused on viral infection, measuring recombinant viral proteins by automated confocal microscopy, using both Evotec's Opera instrument and Molecular Devices' Discovery-1 system. "These primary screens are relatively simple in readout, you just look for cells becoming green," said Lucas Pelkmans, a former postdoc in Zerial's group and the first author of the study. Pelkmans recently started his own research group at the ETH in Zurich, Switzerland.
But these screens were not able to distinguish between endocytosis and later events. To validate the role of genes that were identified in these screens, the researchers performed secondary assays, studying the uptake of fluorescent tracers by the cells. In those screens, they took high-resolution images of the cells to study organelles and different marker proteins in detail, using the Opera system, and looked for specific phenotypes.
What is new about these screens is that they were performed in high throughput, using fully automated imaging systems, and that they combined multiple assays, rather than just a single readout, according to Pelkmans. This allowed the researchers to find a group of genes that coordinates two different endocytosis pathways. "These are examples [of what] you can only do when you do a multi-parameter type of screen where you look at multiple types of assays," he said.
Except for one primary screen, which was performed at Cenix Bioscience, all screens were performed at the Max Planck Institute's own facility, called the High-Throughput Technology Development Studio (for more on TDS, see CBA News 7/4/2005 and 7/11/2005).
Next, the researchers want to expand their study from kinases to the entire genome, a project they are hoping to accomplish by the end of the year. They also plan to further study kinases that seem to play a role in endocytosis by using other technologies, such as proteomics.
The approach taken in this study might be of interest to drug companies for several reasons. First, pharma companies might be interested in using siRNA screens to study infection by medically relevant viruses, Pelkmans suggested.
In addition, using the detailed phenotypes, a gene-silencing screen could be combined with a small-molecule screen to identify chemical modulators of certain genes. "If you get a particular profile for a small molecule amongst all the phenotypes, and you also find a gene that gives exactly that profile, you have a high chance that the small molecule is actually targeting that gene or something in which that gene is also acting," said Pelkmans.
To combine high-content screening with functional genomics and chemical library screens, Zerial's group has already started collaborating with Herbert Waldmann from the Max Planck Institute in Dortmund, in a study that will use kinase and phosphatase inhibitors. Both groups are part of a large network of Max Planck researchers that will focus on such screens, Zerial said.
— Julia Karow ([email protected])