May 15, 2011 (Vol. 31, No. 10)
Newfound Ability to Perform Plate-Based Assays in Primary Cells Just One of the Practices Changing HTS
Screening technologies to advance drug discovery took center stage at the SBS meeting in Orlando at the end of March. Key themes emerging from the scientific sessions and evident in the new technologies and products on display in the exhibit hall included the need to develop more biologically relevant screening strategies, made possible in part by the increasing availability of novel 3-D cell constructs that more closely resemble in vivo tissue environments.
Presenters shared their insights on recent advances and future directions in high-throughput screening (HTS) and assay development. Jonathan O’Connell, director of lead discovery and lead profiling at Bristol-Myers Squibb (BMS), traced the evolution of HTS at BMS and efforts to implement the most relevant bioassays.
As HTS evolved in the early 2000s, the company added advanced compound storage and automated cherry-picking capabilities. BMS currently operates seven integrated, automated HTS platforms. By 2004, assay miniaturization was well under way, and a 384-well format was standard fare, allowing for screening of >500,000 compounds per campaign. Since about 2008, 1,536-well assays have become routine, enabled by precise, low-volume acoustic dispensing technology. A primary screen will now typically exceed one million compounds, at a cost of about $0.025/well.
Secondary screening to assess compound selectivity and define mechanism of action commonly comprises at least four different types of assays and about 30,000 compounds. Improved integration of compound management/sampling and assay processing, with a more flexible infrastructure to support multiple assay technologies, is reducing cycle times and accelerating hit assessment and chemotype selection.
The current driver in drug discovery is the therapeutic target, as “a lot of the easier targets have been done,” said O’Connell. More complex targets require more sophisticated and more varied types of assays: for example, assays performed in primary cells, stem cells, or in 3-D cell matrices; biochemical assays using full-length proteins, in the presence of other proteins or co-factors; multiplexing of reagents and compounds; high-content assays that generate multiple data points per well; GPCR assays that screen compounds, agonists, and other modulators in parallel; and phenotypic assays that utilize hit deconvolution to yield the desired outcome.
O’Connell outlined three main challenges facing HTS. Noting that it is unrealistic to screen every compound in a library against every target, he encouraged companies to employ strategies such as CADD-based selection models to choose which compounds to screen. Additionally, better approaches are needed for screening ion-channel targets, to close the gap between conventional HTS assays and patch-clamp techniques.
Furthermore, hit-seeking activities require more relevant models—primary cells and stem cells—and the ability to design more sensitive, miniaturized high-content assays at high throughput. Technologies that will help companies meet these challenges include high-content microscopy, high-throughput flow cytometry, high-throughput real-time PCR to measure changes in endogenous gene expression, cost-effective label-free assay platforms, high-throughput mass spectrometry, and microfluidics technology.
Profiling Protein Interactions
Researchers at Merrimack Pharmaceuticals are exploring the signaling pathways activated by the fusion protein Etv6-NTRK3, formed from a combination of the Etv6 transcription factor and the tyrosine kinase domain of NTRK3. The fusion gene is present in a variety of tumor types. Tyrosine phosphorylation can lead to protein activation and subsequent downstream-signaling activity. Merrimack scientists designed a system to determine which tyrosine moities in Etv6-NTRK3 are phosphorylated and which are required for oncogenic transformation.
Jack Allen, a scientist at Merrimack, described the group’s use of tandem mass spectrometry to identify 17 sites of tyrosine phosphorylation on the fusion protein. They then screened each site using protein domain microarrays printed onto glass slides that represent nearly all of the human Src Homology 2 and phospho-tyrosine binding domains to pinpoint protein-interaction events that could trigger signaling. Allen reported on the use of site-directed mutagenesis in combination with phenotypic assays to identify a site on the Etv6 component of the protein that is required for cellular transformation in vitro and in vivo.
High-Throughput RT-PCR
Andrea Weston, senior research investigator at BMS, spoke about high-throughput screening with real-time PCR. In high-throughput screening, quantitative RT-PCR can be used to profile the effects of small molecules on gene expression in a real-life context. Although not a new technology, RT-PCR has not been readily amenable to high-throughput processes, but strategies to enable this are now available.
Challenges have included the laborious task of isolating RNA while preventing its degradation (cell lysis buffers now allow for qRT-PCR to be done directly in cell lysates); the two-step process and cDNA intermediate step required with traditional PCR methods (replaced by one-step PCR protocols that are readily automatable and require no mixing); the relatively high cost of PCR reagents and the time factor for thermal cycling (being overcome by miniaturization and the evolution to 1,536-well formats); and the fact that any error in liquid transfer will be amplified during the PCR reaction (a problem minimized by state-of-the-art low-volume, precise acoustic dispensing technologies).
Using high density plate formats and acoustic dispensing, reaction volumes have decreased to as low as 0.5 µL and 1,536 different experiments can be done in a single run. Weston presented data comparing RT-PCR performed in cell lysates with amplification of isolated RNA and described the results as “encouraging”—though not as clean as with isolated RNA—adding, “you do lose dynamic range.”
She described acoustic dispensing technology from Labcyte (Echo® liquid handlers and software) and EDC Biosystems (ATS Acoustic Transfer System), as well as Beckman Coulter’s microfluidic system (BioRAPTR® FRD®), each with advantages and disadvantages such as fixed versus variable droplet sizes, greater or lesser dead volume, and software-defined or user-calibrated adjustments to compensate for variability in the meniscus.
Russell presented preliminary work using acoustic transfer of cell lysates from 1,536-well plates, with early feasibility testing yielding “promising” results with both the Labcyte and EDC instruments. As acoustic transfer drives down volumes, the cost of doing RT-PCR decreases. Weston hopes that competition will push the cost down to no more than $0.10/well. BMS is working with Roche to develop a fully automated PCR process based on the LightCycler® 1536 instrument.
Ion-Channel Targets
Ion channels are increasingly common targets for drug discovery, but traditional plate-based screening methods, such as FRET and calcium flux assays, provide only indirect measurements of ion conductance and may not correlate closely with in vivo electrophysiology. While manual patch-clamp techniques remain the gold standard for screening ion channels, automated electrophysiology has not yet achieved sufficiently high throughput for use in large-scale screening and typically requires secondary assays to confirm hits and eliminate artifacts.
There remains a “gap between manual patch clamp and high throughput,” noted Juna Kammonen, senior scientist at Pfizer. A shift to plate-based screening will require assays that can be performed in smaller numbers of primary cells. Kammonen described his group’s adoption of two new strategies that have enabled higher-throughput whole cell electrophysiology assays.
They are using the FluxOR™ thallium flux assay from Invitrogen (a Life Technologies) to measure ion flow across an ion channel, together with Genedata’s Screener® Kinetics Analyzer software to enable multiparametric and time-series data analysis to support decision making on which compounds to take forward.
Pfizer is also working to achieve higher-throughput implementation of the IonWorks Quattro automated patch-clamp system (Molecular Devices) and improved correlation with PatchXpress® to yield greater confidence in the data generated. Furthermore, the company is using Port-a-Patch (Nanion Technologies) technology in automated plate-based applications for biophysical characterization of new ion-channel cell lines.
Experimentation with the Fluxion Biosciences IonFlux automated patch-clamp system is yielding promising results, and is an example of how technology is “starting to bridge the gap between plate-based assays and electrophysiology.” Kammonen reported that the instrument’s microfluidics technology provides fast, precise perfusion. A single instrument can perform a dose-response assay on 32 compounds in about 15 minutes, and multiple units can be stacked on a robotic platform to increase throughput.
He noted, in particular, the design of the system, with 12 interconnected wells per experimental zone: eight compound wells, two cell groups, and two cell suspension wells that provide built-in redundancy. This 12-well pattern is repeated in 32 experimental zones on a 384-well plate; screening of eight compounds with two recordings generated for each yields 512 data points/plate. The system’s fast liquid exchange rate allows for detection of rapid responses and screening of fast-gated channels such as GABA.
Label-Free Screening
“Creating a holistic screening strategy for label-free technology in a plate based screening group” was the topic of a talk by Rachel Russell, associate director, Pfizer Global R&D, in which she detailed how label-free technology in assay development and screening is facilitating the use of primary cells in place of engineered cell lines for primary screening applications.
“Know what you are asking in the screen,” when you perform reagent validation and assay design using a label-free platform, Russell cautioned. She posed key questions, such as whether frozen primary cells are suitable for use in a label-free system, and whether a native cell grown in isolation on a plastic plate will behave as it would in the body, or whether soft substrates would promote more natural cell growth and differentiation and yield more relevant pharmacological and biological outcomes.
Russell mentioned several innovative technologies that are contributing to the successful development of label-free cell-based assays, including the CellKey® 96-well bioimpedance-based label-free detection system (Molecular Devices), the BIND® SCANNER plate-based instrument using photonic crystal technology (SRU Biosystems), the PathHunter® cell-based assay systems (DiscoveRx), and the Epic® optical biosensor-based screening platform (Corning).
“Label-free is part of the way forward,” she concluded, noting that it provides a method for measuring physiological responses across multiple pathways in parallel.