Technological improvements in automation for chemistry are fueling the integration of chemical synthesis into the discovery process, making it possible to purify, analyze and screen lead compounds faster and more efficiently. Dave Parry, CEO of Cyclofluidic in Hertfordshire, UK, offers insights into some recent innovations in this arena.
"Instead of working with a single molecule at a time, discretely hand made by a scientist in a lab, it's now possible to do an extensive range of chemical processes in an automated, hands-off manner," says Parry, chair of the SLAS2014 Advances in Integrated Chemical Synthesis scientific session. "A number of individual improvements to the whole range of processes and technologies involved in discovery have taken place over the past few years. Now is a good time to bring those together to demonstrate the power of a more convergent approach."
"In its simplest form, drug discovery involves designing a compound by a medicinal chemist, synthesis and screening in one or more assays that generate knowledge about the compound, which is then passed back to the medicinal chemist, who thinks about it for a while, consults with his colleagues and suggests, 'Ah, I think I'll make this molecule next,'" says Parry. This iterative process generates various structure-activity relationships (correlations of the molecular structure of a compound with its biological activities) and molecular profiles not only of where a potential drug might act in the body, but also how the body will deal with the molecule.
Along the way, a medicinal chemist sees a particular compound multiple times, assessing each time how the compound is interacting with a target. "He or she may say, 'I want to improve the affinity, so I'll try increasing the polarity in this direction and see if I can pick up a stronger interaction,'" Parry explains. "They'll then design the next molecule and either make it themselves or coordinate with an internal or external group to have the molecule made. They'll characterize it using a range of analytical techniques to make sure they've got the right molecule and not something similar or, indeed, very different. It will then be submitted for assay(s) and the results will be fed back to them. Based on the assay feedback, they'll know if they were right or wrong about increasing the polarity. At that point, they may try putting a bit more polarity on the molecule to see if they can increase the affinity for the target even more. Or adding polarity might not have worked at all, so they need to provide less polarity, and so on, over and over."
The bottom line, says Parry, is that lead optimization "is still very much an empirical process in which you get information and go through multiple rounds of designing, purifying, analyzing and screening. And it's made considerably more challenging by the fact that, generally, the chemist is optimizing not just against a single target, but against multiple targets." Although the actual synthesis time might involve as little as an hour or two, the process generally takes at least a day or two if a lab makes a compound on site. "If the design process is outsourced, which is increasingly common, the amount of time needed can increase considerably," he notes. "Because it's not unusual for a lab to work on parallel chemistry and request dozens or molecules at a time, the delay before researchers actually have molecules in their hands and ready to submit for assay can in some instances extend to many weeks."
That time consuming and costly bottleneck is beginning to be breached, however. New tools for chemical synthesis and new approaches to drug discovery as a whole are "radically" speeding up the optimization process, Parry says.
Improvements in robotics and automation "make it easier and feasible" to integrate chemical synthesis more seamlessly into the discovery process, Parry observes. At Cyclofluidic, for example, "We use a combination of microfluidics and liquid handling automation to join processes together." Computational power also is increasing, and algorithms are improving, making it possible to run useful predictive calculations in minutes rather than hours, he adds.
On the chemistry side, techniques have improved to the point that "you can work with larger quantities of leads that are suitable for biological assay, and be very confident that you can purify and analyze them multiple times in a fully automated manner," Parry says. He points to Chemspeed Technologies, headquartered in Switzerland, as an example of a company focused on chemistry optimization techniques using novel hardware and software. "They've integrated some clever automation into the phase in which you have interesting molecules and know you are likely to need larger amounts of them," he says. The company's workstations, which will be described in the SLAS2014 Advances in Integrated Chemical Synthesis session by Eva Lu Ping Wu, are designed to improve both throughput/output and the quality of drug development in areas such as sample preparation, synthesis, and pre-formulation and formulation studies.
While some groups are focused on improving specific steps in the discovery and/or development phases, Parry and others are taking a holistic view of the discovery process and asking what they can do to make it "more continuous," Parry says. An example is IRBM Science Park, based in Italy. "IRBM is asking, 'Can we look at the synthesis step much more with screening in mind—that is, think about the chemistry with the biology in mind—and develop an approach that enables us to integrate the two?" IRBM's Alberto Bresciani, lab head – screening technologies, will describe the company's new synthesis-to-assay concept, which is based on the use of DMSO as a synthesis solvent, and present a case study in which a pre-lead series is identified by focused screening, and advanced using this approach.
Parry's group asked a somewhat broader question. "We looked at the entire process of design through to acquisition of biological data, and asked, 'Can we do this in a way that enables us to complete the whole process in a matter of a few hours, rather than in days or weeks?'" Part of the solution involves the company's microfluidics platform, which runs a number of assays in parallel, each generating biological data within a couple of hours. "To shorten the time between the generation of that data and designing the next molecule, we integrated the wet science—from reagents for the synthesis all the way through to getting a biochemical assay point—with computational chemistry methods. So now we have a loop that takes about 90 minutes, is fully automated, and runs 24/7, and at any point in time we can take the biological data as it comes out and use that to update a computational model. This lets us speed up the discovery process and take a serial, iterative approach, using at every iteration the data available to us, hence stepping efficiently through the emerging SAR landscape."
A key enabler of the integration of chemical synthesis into the rest of the discovery process is collaboration among disciplines, Parry acknowledges. "Everyone is very much in the lab together. We and others who are working on this kind of integration don't employ the concept of a 'chemistry group' and a 'biology group.' If people are working in silos in different departments and different parts of the organization, it's very difficult to get these processes together," he explains. Parry views the output of this collaboration as "disruptive" because "it results in a very significant change in the speed with which you can solve problems and generate quality data."
Scientists looking to work in these collaborative environments "need to be good at their individual discipline and also have an appreciation of the other disciplines involved in the process they're working in," Parry adds. "And—no surprise here—they need to be highly computer literate because there's a huge amount of software that goes into running this equipment."
Another technology that has the potential to overhaul the discovery process in years to come is 3D printing. "We now have the ability to effectively print materials we could not have printed previously, raising the possibility of a merging between synthesis chemists and materials chemists," says Parry. An example is the work of Lee Cronin, Regius Chair of Chemistry at the University of Glasgow, and colleagues. "Cronin is asking, 'Can we effectively print bespoke reactors that enable us to do things we cannot do with the current armory of the medicinal chemist?'" The answer is yes. Cronin's group has created 3D-printed "reactionware"—vessels for chemical reactions that make the vessel itself part of the chemical reaction. In the SLAS2014 session, Cronin's colleague Victor Sans will discuss the application of reactionware to the development of "digital" synthesis techniques, as well as for the discovery of new classes of molecules and biological-synthetic devices for medical applications.
In a recent TED talk, Cronin expressed the desire to 3D print a "universal chemistry set," with beakers, test tubes and other lab equipment on one side and molecules on the other side, and then put the two together.
"What Sans will talk about is certainly a long way ahead of what we and IRBM are doing, and a long way ahead of what goes on in a traditional organization," Parry says. "It's a glimpse of what can be done now, and how that is likely to change things in the near future—and it's also a bit visionary, as well."
Parry will moderate the Advances in Integrated Chemical Synthesis scientific session at SLAS2014, Monday, January 20, from 10:30-12:30 at the San Diego Convention Center.
January 6, 2014