|
Over the past ten years, automation of drug discovery has been synonymous with handling ever larger compound libraries and ever bigger sample arrays. This is linked with smaller and smaller assay volumes, more accurate liquid dispensing and more sensitive detectors to measure the ever-decreasing signal levels.
Miniaturizing automated assays has several advantages, particularly the opportunity for parallel sample processing to speed up throughput of screen nucleic acids, proteins, whole cells, or candidate compounds. And as a “free lunch,” this means only very small volumes are needed of the highly expensive enzymes, reagents, nucleic acids, cell lines and so on.
As we shall see, there is some doubt about whether massive parallelization of testing is really the key force driving automation today. But it’s still true that a technology principally developed for large-format assays — namely microfluidic dispensing — is regarded as the key for miniaturizing of automated assays in drug discovery.
“The single most important piece of automation we have installed in the last year is an integrated nanoliter dispensing system,” says Alex Szewczak, research fellow at Merck’s automated lead optimization department at the company’s Boston, Mass., research laboratories. “All our assays rely on it to dilute and dispense compounds.”
Dispensing Alternatives
For dispensing of bulk reagents in volumes above 50-100nl to the microlitre level, the industry has standardized on conventional microsolenoid-driven valves as “good enough” technology. But below that, in the tens of nanolitres region, three main types of dispensing are used: namely pin tools, piezo tip, and acoustic (ultrasonic). Each has its strengths and weaknesses, says Jason Johnson of automation firm Beckman Coulter which in September bought out the instrumentation product line of microfluidics pioneer Aurora Discovery.
The main pin tool company is V&P Scientific, which supplies the technology to several instrument companies. The technology is simple: “Just by dipping a solid pin into a liquid and removing it, you get a very uniform, precisely sized droplet of liquid that hangs on the tip of the pin,” says V&P’s CEO Patrick Cleveland. It can produce droplets of 2nl to 5[mu] to assay plates with an accuracy better than 5 per cent.
Pin tool technology has the advantage of higher spatial definition, says Johnson. It also allows you to do grid reformatting of your assay format. “On the downside it can suffer from cross-contamination and carryover, though those issues can be mitigated to some extent by correct wash protocols.”
Piezo technology — supplied by Perkin Elmer, GeSim, Beckman and Hummingbird — also allows grid reformatting. But it has the big plus that it can pick up and dispense droplets to as many places as the researcher wants, depending on the aspiration volume desired. “This allows you to build batches of multiple numbers of plates in one aspiration, as opposed to the pin tools technology where you are limited in the number of plates you can print and touch off,” Johnson notes. Piezo systems like the Beckman/Aurora BioRAPTR system can dispense volumes down to 100nl; the PicoRAPTR with an eight-tip piezo head can pipette out 1nl to 100µl.
The third method, acoustic dispensing, is supplied by LabCyte, Echo Technology, and EDC Biosystems, among others. It’s the newest and sexiest, giving the advantage of one-to-one transfers and non-contact (“touchless”) technology. It also has the bonus of being able to measure the water content of samples on the fly, which makes it useful for the centralized repositories that many drug discovery companies use to maintain their compounds. But it doesn’t offer the speed of the other technologies, making it less suitable for those companies that conduct huge numbers of screens on giant compound libraries.
Which technology a company goes with also depends on organizational factors. “There are some well-known differences in drug screening companies’ general philosophies as to how screening should be prosecuted and how compound formatting should be handled,” says Johnson. Merck, for example, uses a hub and spoke screening paradigm, with a central compound reformatting location that supplies daughter plates to most of the company’s screening laboratories. In other screening paradigms, the lab will apply the test compound on top of the cell or biochemical assays that would most likely favor pin tool technology. Other companies pre-spot their compound plates: when they have cell assays ready to run, they will dispense cells with piezo technology at each of their therapeutic sites or at a central HTS site.
Improving Cellular Assays
About 50 per cent of all drug discovery is now done with cell-based assays — up from an estimated 30 per cent in 2003 and still growing in importance. Yet miniaturization is still an acknowledged and trying problem for automated cell assays.
Cell waste is one of the big issues. “A lot of scientists are not doing cell based assays on high density plates precisely because of concerns about the automation, particularly how to dispense the cells,” says Johnson. It is very expensive to grow cell batches, and scientists can’t afford to waste even a few of them. This puts two constraints on automated cell assays.
“First you have to minimize the dead volume by keeping pipe volumes small, taking out manifolds, shortening the fluid pathways as much as you can,” says Johnson. Second is the problem of cell mortality. “It’s very important to keep almost all of them alive during dispensing, because if a few die then they will all start signaling each other to die through apoptosis,” he says. Microplates themselves have to be made of cell-friendly materials to reduce unwanted binding — one strategy is to use low surface energy plastics like cyclo-olefins for the plate. Alternatively, plain polystyrene plates can now be chemically treated so as to create a potential in the well where the cells will sit on the bottom and grow up.
“The problems are largely solvable today, people who want to do cell based assays in high density format can do it now, although the demands on technology are constantly rising,” says Johnson. “There are already people running 1-2µl assays at 3456 well format, so 1536 can certainly be done.”
But Marc Feiglin, chief technology officer for life science research at automation leader Tecan, argues that miniaturization of cell assays has its limits. “After a certain point, how few cells can you work with before your data is not statistically significant?” he asks. “A single cell doesn’t usually tell you much, so in general when we take measurements we look at cell populations. That’s one reason why, in very high throughput screening, people
are still mainly using fluorescence-based [non-cellular] assays.”
Feiglin is also doubtful about the prospects for high content screening (HCS), which is essentially, he says, an attempt to automate confocal cell microscopy by putting it into microplate format. Tecan itself, he admits, hasn’t worked out how to play the HCS market, given that some competitor companies have made massive investments and, with the possible exception of Cellomics and its giant partner GE, seen little back in revenues. Feiglin sees further consolidation in the market, following the [December 1, 2006] announcement by Evotec that its HCS/HTS division is to be sold to rival PerkinElmer. (See sidebar, “PerkinElmer Shores Up Cellular Technology Offerings,” page 26.)
384 or 1536? And as for 3456 ...
As for the wider issue as to whether there is a general move to 1536 and larger automated assay formats, opinion is divided. Beckman’s Johnson insists that many companies are now comfortable with 1536-well formats as the quality of automation, homogeneous assays, evaporation control techniques, and the accurate specification of microplates themselves have all improved.
“Just a year or two ago many automation providers under-specified the precision of their staging technology and some of the plate suppliers shipped plates that weren’t the right geometry,” says Johnson. That, he claims, has been largely taken care of by new ANSI standards for geometry, height, footprint, and well field — although he adds, it would be nice if the specs were even tighter. The ANSI standard now includes 1536 well plates as as well as 96, showing that this end of the market is maturing.
Some companies, like Merck, have even committed to 3456 technology at microliter volumes, and are sticking to it come hell or high water. “Merck is absolutely committed to both 1536 and 3456 high density formats,” says Szewczak. The company’s primary HTS site — its automated biotechnology center in North Wales, PA — has two 3456-well ultra-high throughput screening platforms as well as a 1536 capability. The Boston lab, which is more focused on supporting med-chem lead optimization after an HTS is completed, uses 384-well as the workhorse format for optimization assays, but also has 1536-well capability.
“Merck has spent tens of millions of dollars on 3456-well technology, they can’t possibly write that off,” notes Feiglin. “But in general we [Tecan] aren’t seeing many companies using even 1536-well plates, let alone 3456.”
He points to the changing paradigm of drug discovery — the gradual
acceptance that massive parallel screening of giant libraries is maybe not the panacea it was thought to be.
Ten years ago there was a major
investment in HTS, with companies like Aurora and Evotec selling fully automated systems. “Pharma became enamoured with the technology, but these investments didn’t significantly change drug discovery output in terms of really new clinical leads,” says Feiglin. Moreover, the costs of the tests, of running all the reagents, were too high. “As a result we’ll never see that level of investment again,” he predicts.
Now drug discovery companies are being more selective. Many of the large pharma companies’ centralized screening facilities have been scaled back or even closed. As a result, any move to 1536 or higher has stalled, Feiglin thinks. “Most of the samples are still coming in as 96 or 384 format; reformatting them takes time and maybe takes away any benefit to going to 1536.” Just about every company he knows is sticking to 384 as a workhorse technology.
“They’d rather see a validated proven solution at 384 or even 96-well format, rather than go through the whole process of validating it at 1536,” he says. “Especially with the bubbling and evaporation difficulties you encounter at 1536 and higher formats.”
That sort of equipment doesn’t need a huge central HTS; it can be distributed down to lab level.
Feiglin sees cell culture and preparation as a more promising area for expansion: “The biggest issue in cell-based assays today is producing cells on a small scale. There is tremendous interest in automation for that.” Merck’s Szewczak is also seeing a trend to increasing automation of cell assays and cell manipulations such as culturing and genetic transformations. His lab at Merck is currently installing some automated cell culture systems that he regards as vitally important to its lead optimization work — almost as important as the local integrated compound storage and retrieval system now nearing completion.
But the real future of automation could be in genomics, not HTS. Feiglin says figures show a third of liquid handlers are now sold for genomics applications — gene expression profiling, sequencing, DNA arrays. “Whereas ten years ago I would say 60 per cent of liquid handlers were sold for HTS, genomics is now bigger than HTS,” he says. “And there, the automation trend is away from higher throughput towards quality and ‘walk-away’ time — where I need to process only six or eight samples, but to do it well and reproducibly, and to walk away from it while it’s happening.”
PerkinElmer Shores Up Cellular Technology Offerings
By Laurie Sullivan, senior technology editor, Pharma DD
With its pending acquisition of Evotec Technologies, PerkinElmer is poised to fortify its standing in the cell analysis instrumentation market. The deal also brings about a union between two automation powerhouses. PerkinElmer announced in November it would pay approximately $30 million in cash for the Hamburg, Germany-based supplier of tools and technologies for cellular research. It’s one of the the latest in a string of acquisitions designed to complement its existing portfolios in various areas of the life sciences.
“The acquisition will immediately strengthen our position in cellular analysis and high-content screening by extending our current cellular science offering,” says Mary Duseau, global sales leader, molecular medicine, PerkinElmer Life and Analytical Sciences. “From our perspective, the acquisition gives us the scale, penetration, and brand recognition in high-content screening as well as in cell sorting and selection technology.”
Commenting on the deal, Carsten Claussen, CEO of Evotec Technologies, said in a statement that the company looked forward to becoming the Center for Cellular Sciences in Hamburg within the PerkinElmer group. According to Duseau, PerkinElmer is similarly enthusiastic. “Evotec Technologies [has the] capability to be our center of excellence for advanced cellular technology and tools,” says Duseau. “The longer-term benefit is that strategically, it will provide extensions into that space.”
Notes Quintin J. Lai, an analyst with Robert W. Baird & Co.: “We view the Evotec Technologies acquisition as an example of PerkinElmer’s continuing strategy to enhance growth in key segments. We think this acquisition will expand PerkinElmer’s high-content screening product offering to pharma, biotech, and academic customers.” Baird increased its 2007 revenue estimate by approximately $16 million to reflect the acquisition and predicted that Evotec would have a neutral impact on 2007 earnings.
Evotec Technologies, which generated sales of €17.0 million in 2005, is a majority owned subsidiary of Evotec AG. Evotec Technologies provides systems for confocal imaging, cell handling, ultra-high-throughput screening, as well as image capture and cellular analysis software. Included in its product portfolio is the Opera HCS platform, which is a tool for high-content analysis combining confocal imaging with the throughput that is required for primary and secondary screening.
Evotec AG’s CEO Joern Aldag expressed in a statement that PerkinElmer is best positioned to leverage Evotec Technologies’ full potential. PerkinElmer reported revenues of $1.5 billion in 2005 and is a global presence, with 8,000 employees serving customers in more than 125 countries.
“One of the primary benefits of the acquisition, both for Evotec Technologies’ customers and PerkinElmer’s customers, is that PerkinElmer can bring immediate global distribution capabilities, service and support, and complementary product lines,” says
Duseau.
The deal is expected to close in early 2007, capping off a year that saw a flurry of buyout activity. PerkinElmer made five acquisitions in 2006, in each case broadening its offerings in a particular sector. (See Table.)
Just three weeks after announcing its agreement with Evotec, PerkinElmer said it planned to acquire Belgium-based Euroscreen Products S.A., a majority owned subsidiary of Euroscreen S.A. This acquisition, also expected to close in early 2007, would build upon the Evotec purchase and further strengthen PerkinElmer’s position in cellular screening.
Under the planned transaction, Euroscreen Products would transfer its portfolio of GPCR screening tools and its exclusive global license to aequorin technology to PerkinElmer. Aequorin technology generates large, luminescent signals that result in higher-quality data and fewer false positives in high-throughput and ultra-high-throughput screening environments.
|
