Nanoliter and other types of liquid dispensers can speed testing, save money on compounds and reagents, and reduce error rates 

Patrick McGee 
Senior Editor 

click the image to enlarge

The inL10 was used to perform non-contact dispensing of 20 nL of protein solution with 0.9 mm spacing printed in a 96-well plate. (Source: Caliper)
It is no secret to those in drug discovery and development that bottlenecks in liquid handling can cost pharma and biotech companies time and money. The industry has looked to automated liquid handling as a solution to this problem and benchtop automated systems and sample dispensing systems have become much more common. Over the last two to three years, however, automated instrumentation capable of handling plates with 96, 384, or 1,536 wells and potentially even higher densities have entered labs. And, as the rate of throughput has increased, other tools like nanoliter dispensing have emerged that make it possible to greatly decrease the amount of reagents needed for assays.

"Advances in the nanodispensing field are probably the most innovative, the most exciting to the people in the labs. They like the idea that you'll be able to aspirate and dispense nanoliter quantities of things," says John Morin, PhD, director of high-throughput screening in the Screening Sciences Group at Wyeth Pharmaceuticals, Collegeville, Pa.

"The biggest practical advance that has had the biggest impact in our lives is the use of low-volume 384-well plates," he adds. Although it took time for these plates to be refined, Wyeth is now able to save money by reducing the volumes of their microassays to the approximate size seen in a 1,536-well plate while still retaining the automation available for handling 384-well pipetting. "That's been a big plus over the last few years."

David Mark, PhD, senior research director for Roche Discovery Technologies, Nutley, N.J., believes the most effective nanoliter dispensers on the market are made by Cartesian Dispensing Systems (acquired by Genomic Solutions in 2001), Ann Arbor, Mich.; Caliper (formerly Zymark) in Hopkinton, Mass.; and CyBio, in Woburn, Mass. "These, to us, are probably the most advanced and more unique and versatile systems of nanoliter pipetting or nanoliter dispensing," says Mark.

"The problem with nanoliter dispensing is that, because of the small volume, it has to be very accurate. The valves that are used for this work have to be very well engineered and deliver liquid very accurately and reproducibly. Some of the systems out there from other companies don't work as well as these three." Each company has a different approach to building the dispensing heads and has developed some way of measuring or calibrating them to ensure that each valve is dispensing the same volume, even at very low levels. Traditional 96- and 384-well pipetters, for example, cannot deliver accurately below one microliter, but the better nanoliter dispensers can deliver accurately down to the 10- to 50-nanoliter range, Mark says.

Roche is integrating a device called the NanoJet into its ultra high-throughput screening system and anticipates cutting reagent use by at least 30%. Roche worked with CyBio to develop the NanoJet, a noncontact dispenser that can deliver volumes ranging from 50 nanoliters to 20 microliters. The NanoJet features ceramic nozzles designed by Roche scientists in Switzerland that were licensed to CyBio. Mark says Roche decided to develop the system with CyBio after concluding that ones from Caliper and Cartesian Dispensing Systems did not meet their specifications.

"Although they are nanoliter dispensers, they function by using syringes that aspirate a certain amount of reagent and then dispense that amount. But the NanoJet is a flow-through type dispenser; it is a true dispenser. You don't need to aspirate, you just put reagents in a bottle that is linked up with this system and it will dispense through. It will automatically use pressure to push the liquid through, and the valves open and shut to dispense the liquid, so there's no aspirate step," Mark says.

"When you aspirate, you only aspirate a certain amount of liquid. So after you dispense what you've aspirated into the syringe, then you have to go back and aspirate some more and by doing that you have to calibrate your system again."

click the image to enlarge

The SelectT, an automated cell culture system for multiple cell lines and assay-ready plate production. (Source: The Automation Partnership)
Nanoliter functionality 
Another vendor driving innovation in nanoliter dispensing is Caliper. A few years ago Caliper developed a platform called the Sciclone ALH 3000, a standard volume liquid handler that can pipette accurately in 96-, 384- and 1,536-well formats down to 100 nanoliters. "[This product] is in place in a lot of drug discovery labs," says Mark Roskey, PhD, Caliper's vice president of marketing and applications. "But we had specific customers who said, 'Hey, can you take that same functionality down to lower volumes, below 100 nanoliters? What if we could pipette accurately in the range of 10 nanoliters?'"

Those questions led to the inL10, which was developed in collaboration with Johnson & Johnson, says Roskey. J&J wanted to accurately pipette 10 nanoliters to set up a drug-screening assay that was one microliter and had neat dispensing of a compound into the microliter so they could save dilution steps. Roskey says Caliper looked at available technologies, but it was obvious to them that standard air-based pipetters could not handle such low levels with the desired accuracy.

Caliper instead developed a 96-well pipetting system that utilizes microelectromechanical systems micro-flow meter devices with built-in temperature sensors. The sensors record the exact amount of liquid being dispensed in each channel and automatically compensate for changes in viscosity due to temperature fluctuations. "What this allows you to do is to very precisely control the timing of opening and closing of the valve based on this pressure-flow system.…Then the instrument reports back that pressure change. So it tells you, 'You asked for 10 nanoliters, I pipetted 9.9 nanoliters. Is this good enough or would you like me to try again to get 10 nanoliters?'"

Another feature is that each channel offers independent pipetting control and liquid level detection. This allows any combination of channels to be used at any time, and each channel can pipette its own fluid volume. The inL10 also has 109 microprocessors built into its 96 channels. "That's where the 'i' in inL10 comes from, because there's a certain amount of intelligence built in," Roskey says.

click the image to enlarge

Wyeth's automated sample storage and retrieval system, which is used to manage the company's liquid compound store. (Source: Wyeth)
Strategic alliance 
Wyeth has concluded a strategic alliance with The Automation Partnership (TAP), Cambridge, UK, which has helped the company outfit new labs that they will move into within the next year. "They're very good at manufacturing liquid handling devices and we're going to be utilizing nanoliter dispensing in our new systems," Morin says. He added that the "biggest innovation" occurring at Wyeth is its purchase of SelecT, one of TAP's automated cell culture systems. SelecT mechanizes all liquid handling involved in culturing cells and was designed by TAP with five pharmaceutical companiesGlaxoSmithKline, Merck, Pfizer, Bristol-Myers Squibb, and AstraZeneca.

SelecT is essentially a giant laminar hood that uses robotic arms to do the pipetting, Morin says. "It's probably the most impressive single piece of automation that we have right now. There's a huge impact in decreasing not just the amount of work that goes into maintaining cells and plating them to get them ready for screening, but also, it basically removes a lot of the human error from that process."

At Abbott Laboratories, Abbott Park, Ill., researchers are enhancing methods to distribute compounds across the organization using more advanced liquid handling tools, says David Burns, PhD, head of the company's high-throughput screening group. He adds that there is a great deal of work going on at Abbott and other companies to integrate liquid handling into the early stages preceding HTS. "I think it's really a push toward very rapid cherry picking and getting a small volume of material for retesting or for creating a small, specialized sub-library for HTS screening, and that's more or less in terms of the liquid handling distribution point."

They have already selected a vendor and are in the process of developing a very high-throughput liquid store for all of Abbott Park that they hope to commission this year, he says. "With the new mini-tube stores we'll be able to cherry pick for a sub-library screen for HTS and provide SAR compounds for therapeutic areas. Hopefully, it will really become the workhorse in terms of compound distribution."

Abbott's compound preparation area is primarily responsible for delivering compounds either in microtiter plates or on high-density cards, and they rely heavily on automated liquid handlers and workstations. For standard plate screening they have a Sciclone ALH 3000 which allows them to distribute one to five microliters of volume for the microtiter plates. "We also have a Sciclone inL10 on order that will allow us to distribute much smaller volumes for IC50 determinations," Burns says.

Old School: Building a Better Mousetrap 
Much time and money is being spent to automate liquid handling in pharmaceutical drug discovery and development, but some companies are taking a different approach by simply building a better mousetrap. nAscent BioSciences Inc., Cambridge, Mass., recently introduced PocketTips. The company touts the disposable pipette tips, which are compatible with existing liquid handling equipment, as the first capable of delivering nanoliter amounts of compounds for drug discovery assays.

"This, I think, is one of the more revolutionary devices out there," says John Morin, PhD, director of high-throughput screening in the Screening Sciences Group at Wyeth Pharmaceuticals, Collegeville, Pa. "At the same time that you see all of this high-end technology marching forward that enables us to dispense nanoliter amounts from multiple wells simultaneously, you also have this pipette tip, which is a disposable consumable engineered so precisely that you're able to achieve nanoliter pipetting with a hand-held pipette in a laboratory."

PocketTips were designed to overcome weaknesses inherent in current methods for liquid transfer, including poor performance in the sub-microliter range, cross-contamination of biological targets, or poor reliability. Capillary metering of compounds and integral dilution in the tip prior to transfer to the assay site ensure precise dispensing, the company says. PocketTips are designed to work with all pipetters and are able to manage sub-microliter fluid volumes. Last July, nAscent delivered the first PocketTips, which are capable of dispensing 100 nanoliters of compound, to companies like Amgen, AstraZeneca, Aventis, Merck, Novartis, Pfizer, and Wyeth. The company will also introduce two new versions for the 96- and 384-well channel models of Beckman Coulter's BioMek FX liquid handler.
Separating the vendors 
While there are a number of liquid handling tools available to researchers, Morin believes the most important thing separating vendors is the volumes their devices can handle accurately and precisely. "I think that the general quality of the assay results that we get from our high-throughput screens has improved as we've moved to high-end liquid handling devices. Your margin of error when you're trying to detect a real hita valid hit from a statistical fluctuationis small. So the further you can drive those statistical fluctuations down, the better off you are."

But while liquid handling has made great advances over the last several years, the technology still continues to evolve. "I think liquid handling is the final frontier for automation. It is the single most difficult aspect of putting together an automated system," says Morin. He believes that the biggest shortcoming of the present technology is that it all depends on physical manipulation to move the fluid from one point to another.

Morin says he is particularly excited about a technology being developed by Labcyte Inc. (formerly Picoliter Inc.), Sunnyvale, Calif., which uses acoustic waves to manipulate droplets as small as a picoliter (See related story on acoustic technology on page 46). "There are a few major players that are going to invest in it heavily, and we'll all be on the sidelines watching. In a year or so, we'll be seriously considering whether we should be entering that field also."

Although liquid handling is being automated rapidly, it would be a mistake to do away with traditional technologies like 384-well pipetters, says Roche's Mark. "You need a combination of both. Together, you have a very powerful system. You can have the nanoliter dispensers working together with a traditional 384-well pipetter and you can set up any type of assay. It has the flexibility and the ability for you to do miniaturized assays and save on reagent usage."

Burns believes there is growing interest in noncontact nanoliter dispensing and that the smaller volumes needed are enabling researchers to save reagents and compounds while at the same time enabling 1,536-well assays. "I also think that a lot of the newer hardware and automation are enabling one to be more careful about quality control and make sure that you are actually dispensing your samples into the microtiter plates, for example."