An automated microscope-based screening and computer-assisted image analysis approach enables measurements of multiple features at the cellular level at the same time, giving rise to the term high-content screening (HCS).
“Another distinct advantage of HCS,” comments O. Joseph Trask, head of cellular imaging core, The Hamner Institutes for Health Sciences, “is the ability to quantitatively measure phenotypic changes at the single-cell level following compound challenge. However, assay development still remains a challenge.”
In collaboration with Tim Wiltshire, Ph.D., at the University of North Carolina, the Hamner team led by Russell Thomas, Ph.D., used the HCS platform to identify potential cellular toxicity pathways. Primary embryo fibroblasts isolated from 32 genetically characterized inbred mouse strains were exposed in concentration response to a diverse set of environmental and pharmaceutical agents. Comparison of the cross-strain toxicity response pointed to candidate genes.
“To enable a true comparison between cells of different genetic origins, we had to figure out how to plate all 32 cell lines on the same 384-well plate and to make them grow to the same steady state,” says Dr. Wiltshire. This was achieved by a creative adaptation of the Thermo Fisher Multidrop instrument. The instrument’s eight input tube lines, typically delivering cells and media from a single vessel, were modified to uptake cells from eight different vessels, each containing a unique cell line. Four of these modules in tandem produced a 384-well plate with 12 wells per cell line.
The test compounds had to be added in small volumes to improve solubility in growth media and reduce toxicity of organic solvents. The team modified an existing Biomek pin tool liquid handler to deliver compounds in the 200 nanoliter range directly from the compound library. “This HCS screen produced a particularly complex set of data including dose-response measurements of four multiplexed cellular responses at two exposure time points in 32 different cell lines, requiring the development of new informatics tools,” says Dr. Thomas. The team has identified several interesting pathways and is working on a detailed biological evaluation of the key genetic targets.
Protein-Protein Interaction Targets
“High-content screening is, indeed, a powerful tool for studying the effects of compounds in living cells,” agrees Paul Johnston, Ph.D., research associate professor, School of Pharmacy, University of Pittsburgh.
“Introduction of high-throughput, high-resolution multichannel imaging devices integrated with powerful image analysis algorithms/modules made many complex analyses, such as protein-protein interactions, possible.” Interaction between proteins is essential for all cellular functions, and thus represents a novel class of therapeutic targets. Historically, protein-protein interactions (PPIs) have been difficult to tap for drug discovery.
“We now know that protein-binding interfaces are composed of discreet ‘hot spots’ of contact rather than one contiguous surface,” continues Dr. Johnston. “We can combine the power of structure-based drug design with high-content screening to improve success rate of finding small molecule PPI-disruptors.”
Dr. Johnston’s team implemented a semi-automated process to screen for disruptors of interaction between androgen receptor (AR) and TIF2 transcription initiation factor, which is thought to be important in prostate cancer. The process utilizes a positional biosensor method, originally commercialized by Cellumen, now a division of Apredica, a Cyprotex company.
Once set up, the pipeline can generate up to fifty 384 plates per day using standard liquid-handling and culture-dispensing robots. “As Dr. Hillson correctly noted, intricacy of frequent re-programming of laboratory robots is an impediment to a complete automation of our workflow,” says Dr. Johnston. “However, the process is readily scalable.”
Recombinant TIF2 encodes a localization signal directing it into the nucleolus, a compartment of a nucleus. AR can be stimulated to translocate into the nuclear compartment by treating cells with dihydrotestosterone. Co-localization of the proteins in the nucleolus is readily detected by imaging of fluorescent reporters tagged to each of the protein interaction partners. Initial screening of 1,700 compounds from the LOPAC set and NIH Clinical Collection library revealed several hits that either prevented the formation of AR-TIF2 complexes or disrupted already formed interactions. The team is gearing up to screen a much larger (>200,000) compound library.
“Chemspeed introduced numerous disruptive technologies in laboratory automation workflows,” says Michael Schneider, senior vp, business development. “Our Object Oriented Workflow Design is the foundation of all our product lines. By combining independent, functional Shuttles (sample carrier) and Objects (processing workstations) we created a new workflow paradigm that allows to process samples in parallel and sequential fashion. This flexible design supports practically unlimited variations of workflows.”
Each object is an independent “machine”, capable of receiving inputs, processing data, and sending outputs to other objects. Chemspeed’s modular objects are connected by a production-proven track system, carrying up to 100 shuttles. Each shuttle carries its own processes and interacts differently with “objects” along the track.
“This makes the workflow more robust,” continues Schneider. “If one shuttle fails for whatever reason, the others continue on.” The key to designing object-oriented workflows is to carefully identify all the required process steps, to connect them in a logical fashion, and to establish input and output parameters.
Another enabling key element of Chemspeed’s platforms is the flexible exchangeability of the more than 40 robotic tool features. Akin to a swiss-army knife, the unique robotic arm is able to accommodate substance handling, action, or analysis, including gravimetric dispensing, homogenization, ultrasonic dispersion, capping, and many others. These tools can be seamlessly exchanged during the run, providing additional flexibility to workflow design.
Pharmaceutical compound management fits perfectly in such workflow. The processes may include compound retrieval, removal of screw caps, adding of solvents, microdispensing, barcoding, quality assurance analysis, and return to storage. Chemspeed perfected gravimetric dispensing of solids, liquids, and viscous liquids.
Upcoming SWILE dispensing technology is specifically positioned for compound management. Chemspeed’s SWILE supports the “many-to-many” dispense mode and is capable of gravimetrically dispensing sub milligram quantities of compounds with a wide range of consistency, from viscous oils via waxy components to solids. Thanks to an automatically exchanged disposable all glass dispense “tip”, cross-contamination is completely eliminated.