Profile: Miniaturization and integration: challenges and breakthroughs in microfluidics
Company history
As early as 1989 the potential for microfluidic systems to replace conventional ‘macro’ systems was recognized at Pharmacia Biotech (now Amersham Biosciences) in Uppsala, Sweden and research into microfluidics began. Progress made during the following ten years of research and development proved so successful that in the year 2000, the decision was taken to form an independent company, Gyros AB.Coming from such a strong biotechnology background and owning an extensive portfolio of over 40 patents and patent applications relating to microfludics, CD manufacture, system components, surface chemistry and specific application areas, it was no surprise when the company announced its intention to enter the biotechnology market with its own unique Gyros™ technology platform. The long term goal is to become the leading supplier of microlaboratories. In the short term the company has recently announced the first products developed using Gyros technology and offers Partnership Programs to clients who wish to co-develop systems for customized applications.
Experienced personnel
Developing commercial products from sound scientific ideas requires a broad range of skills and the personnel list at Gyros AB has been expanding to ensure that all requirements for a commercial success are in place. Senior management includes several individuals with over 20 years experience in research and development, sales and marketing and business development. They have held senior positions in both local and global, multi-national companies within the pharmaceutical, biotechnology, diagnostics and life science supply industry. Heading the Gyros team is Dr Maris Hartmanis, former Vice President, R&D for Amersham Biosciences, a scientist and manager with an extensive background in academia and biopharmaceutical research.Many key personnel from the original research and development group moved with Gyros AB, bringing with them a wealth of expertise in microfabrication and microfluidics, and the knowledge and experience needed to work with biomolecules and to develop commercial products. Dr Per Andersson, Director of Applied Research, has more than 10 years of hands-on experience with miniaturized separation and analysis systems, including several years research experience with chip-based microfluidics in the life sciences. As a post-doctoral research fellow with Professor Jed Harrison at the Department of Chemistry at University of Alberta, Canada, he developed cell based assays on a chip. He then joined Amersham Biosciences to become responsible for the scientific development of polymer chip technology based on the CD format and to supervise the development of nanoliter scale analytical systems based on centrifugal force. The Vice President Research, Mr Rolf Ehrnström, is a frequent scientific contributor at international conferences, a former Science Director for the Protein Science Area within Amersham Biosciences and coordinator/project manager of government-funded projects in Europe and the US.
Sound financial footing
The company has been on a sound financial footing from the very beginning. Gyros AB received the largest ever biotech-related investment in Sweden in 2000, and, in December 2001, the company was placed among the top 35 biotech ventures in the world after a successful second round financing that topped approximately 30 MUSD and became one of the largest ever biotech rounds in Europe.Technology development and milestones
The idea of microsystems is not a new one. In 1975 Jerman and Angel developed one of the first true microsystems, a gas chromatograph. The reason that this did not take over from the macro scale GC probably reflects the fact that the only real benefit was one of size.Today’s microfluidic systems have the capability to replace conventional ‘macro’ systems for many applications in the life sciences, but for this transition to occur there must be significant benefits to the user, driven strongly by the requirements of the application. The general advantages are clear: low volumes reduce consumption of samples and reagents and ultra small volumes can be handled more easily; the performance of many applications can be enhanced by increased surface∶volume ratios that facilitate high speed reactions or separations; processes can be run at scales more relevant to normal biological conditions; and throughput (defined as more samples per time unit) can be improved if large numbers of samples can be processed in parallel.
When looking at the world of drug discovery and development, one sees complex analytical problems being handled in high numbers, preferably in parallel, in order to reduce time to market and overall costs (Fig. 1) and an increasing demand to get more information from ever decreasing amounts of sample. In this field the benefits of miniaturization alone may not be enough to justify the adoption of microfluidic solutions. However, an ability to integrate multiple steps into a single, streamlined application and to run multiple samples in parallel would open up the added opportunities of doing more with less sample and improving performance, reproducibility and productivity. Fig. 2 illustrates one way of looking at the approaches to miniaturization that exist today. The real challenge for microfluidics is to fulfill the needs of the drug discovery market by offering a true paradigm shift that enables a large number of data points per sample to be acquired while still being able to process a large number of samples. The benefits of making a transition from macro to micro would be unquestionable.
 | ||
| Fig. 1 Complexity of the drug discovery process. Reproduced with kind permission: AKos Consulting & Solutions Gmbh. | ||
 | ||
| Fig. 2 Miniaturization today. | ||
CD microlaboratories—the Gyros approach
Traditional microfluidic devices tend to miniaturize existing ways of working. Fig. 3 shows a typical microfluidic device in which we see different ports for introducing samples, buffers and reagents, a reaction chamber followed by a separation bed and a detection area. | ||
| Fig. 3 Typical microfluidic device. Reproduced with kind permission from Chem. Eng. News, 1999, 77(8), 27–36. | ||
Simply miniaturizing a process has often been seen as a solution to the problem of running more samples per unit time. Over the years various engineering solutions have been applied to miniaturize applications. However, just scaling down a process can introduce new problems. Table 1 shows an example of how a true microfluidic solution may offer a better alternative to simple miniaturization.
| Low density plates (96/384 wells) | High density plates (1536–9600 wells) | Microfluidic system |
|---|---|---|
| Open systems | Open systems | Closed systems |
| Liquid handling by displacement | Liquid handling by piezo/jet | Liquid handling integrated |
| Controllable evaporation | Evaporation constraints | Heterogenous or homogenous assays |
| Heterogenous assays | Homogenous assays |
A recent article (Microfluidics-downsizing large-scale biology, Nature, Biotechnology, 2001) highlighted some of the challenges that must be overcome in order to develop reliable and flexible microscale solutions that fulfill the needs of drug discovery applications:
1. Solve the problem of integrating elements such as pumps, valves and reservoirs for liquid handling.
2. Design efficient liquid pumping strategies that do not expose analytes to high electrical charge or overheating and that can be used in inorganic media.
3. Solve the problems arising from surface tension effects, air bubbles and evaporation in microchannels.
4. Develop a suitable technology to monitor liquid flow.
5. Solve the problems arising from small particles or precipitates clogging microchannels.
From the experience of scientists at Gyros, three other factors must also be considered to create commercially viable solutions.
1. Control of surface chemistry, for example to eliminate non-specific binding.
2. A straightforward interface between the macroworld and the microworld.
3. Reliable and reproducible manufacturing processes.
An additional factor in modern drug discovery is that it is often necessary, or at least preferable, to work at the nanoliter scale and here scaling laws become significant. Put simply, when working at the nanoliter scale surface tension becomes a more dominant force than gravity. This factor alone demands a reappraisal of how to miniaturize, but recent developments indicate that this phenomenon can be used to advantage to eliminate complex engineering solutions and facilitate a greater level of integration than is possible with a conventional microfluidic approach.
From the list of required improvements it can be seen that liquid handling is a key element. Table 2 illustrates the drawbacks and advantages of the many different approaches to handling liquid flow in microsystems. If we look only at the control of liquid flow through a traditional microfluidic system, the apparently simple action of adding an extra step into the microfluidic process to improve efficiency can result in an almost exponential increase in the complexity of the microfluidic design. Design complexity would be further compounded when trying to process hundreds of samples in parallel.
| Mechanism | Liquids pumped | Comments |
|---|---|---|
| Pressure | Aqueous and non-aqueous | Independent of solution composition |
| Dependent on viscosity, channel geometry | ||
| Electroosmosis | Mostly aqueous (pumping of non-aqueous liquids reported, but uncommon) | Very dependent on buffer composition (e.g. ionic strength, pH) |
| Requires high electric fields | ||
| Centrifugal force | Aqueous and non-aqueous | Independent of solution composition |
| Dependent on viscosity, density, channel geometry | ||
| Ultrasonic (travelling wave) | Aqueous and non-aqueous | Requires complex integrated structure |
| Electrohydrodynamic | Non-conductive liquids (non-aqueous) | Requires high electric fields |
| Capillary force (surface tension force) | Aqueous and non-aqueous | Passive: requires no external applied force |
| Once channel filled, liquid movement stops |
Application-specific CD microlaboratories
Using centrifugal force together with microfluidic engineering to control liquid flow, define volumes and perform other functions at the microscale was documented as early as 1973 (C. D. Scott and C. A. Burtis, Anal. Chem., 1973, 45(3), 327A–340A). To scale down further to micrometer channels, using nanoliter volumes, puts further demands on the design of the microfluidic system and how to control interfaces and surfaces.Gyros have combined the ingenuity of using centrifugal force together with the possibilities offered by modern engineering, materials and surface chemistry to develop a technology platform that enables those working in drug discovery to benefit not only from miniaturization, but also from an ability to integrate multiple steps into a streamlined process. Microlaboratories, in the form of a CD, can process hundreds of samples in parallel through application-specific microstructures, driven by centrifugal force as the CD spins. Inside each application-specific microstructure samples and reagents are driven through the application steps using a combination of precise microfabrication, controlled surface chemistry (for example to create hydrophobic breaks to stop liquid flow), and capillary or centrifugal force (for example to overcome hydrophobic breaks and move liquid further on through an application).
This novel solution appears to overcome most obstacles that will be met on the way to developing flexible, reliable microfluidic solutions. The control mechanisms employed eliminate the need for pumping devices, valves, tubing connectors or high voltage power supplies. Despite working in a nanoliter scale environment, evaporation problems are eliminated as samples can be processed within a sealed microlaboratory. Surfaces can be pretreated in specific areas of a microstructure, for example, to reduce non-specific binding or enhance cell growth.
First application
The advantages of working at microscale must be clear for any application in order to justify the transition from ‘macro’ to ‘micro’ scale. By considering conventional working procedures and pinpointing areas in which miniaturization and integration can be of significant benefit, CD microlaboratories can be developed to enhance the performance of critical applications.A review of processes used in proteomic investigations highlighted that existing sample preparation techniques prior to analysis by MALDI mass spectrometry are not well suited to processing hundreds or thousands of samples per day and, at the same time, achieving the high sensitivity needed to identify even low abundant proteins. Consequently, sample preparation frequently becomes a rate-limiting step. Large numbers of samples must be concentrated, desalted and crystallized with matrix. User intervention steps are often required, risking loss of sample or contamination, reducing sensitivity in the final analysis and so increasing the chance that low abundant proteins cannot be identified.
It appeared that the sample preparation process, crucial to successful protein identification, could be significantly improved if several individual steps could be miniaturized, optimized and integrated using the Gyros technology platform. Fig. 4 illustrates the first application-specific CD microlaboratory from Gyros. Gyrolab MALDI is designed for automated preparation of samples prior to analysis by MALDI mass spectrometry. Protein digests are concentrated, desalted, mixed with matrix and crystallized in MALDI target areas on the CD. Up to 96 samples are processed in parallel under identical conditions. The CD is transferred to a MALDI mass spectrometer for analysis and identification without the need to transfer samples to a separate target plate.
 | ||
| Fig. 4 96 application-specific microstructures are connected to a common distribution channel for simultaneous application of wash or eluent solutions. | ||
Each step in the process is optimized to enhance sample recovery and concentration, from precise volume definition within the CD through to crystallization (Fig. 5).
 | ||
| Fig. 5 Each microstructure is comprised of optimized microfabricated units that perform specific steps in the integrated sample preparation process, for example volume definition, chromatographic purification and a MALDI target area. | ||
In these experiments, using Bruker Biflex III or Bruker Autoflex MALDI instruments, sensitivity levels in the attomole to low femtomole range, close to the detection limits of the mass spectrometers, were achieved (Fig. 6).
 | ||
| Fig. 6 Spectrum of a BSA digest, equivalent to only 400 attomoles, analyzed on a MALDI target in a CD microlaboratory. Mass accuracy <100 ppm. Resolution >6000. Ten peaks were identified as belonging to BSA by a database search of all available mammalian proteins. | ||
Fig. 7 shows the high level of reproducibility of the mass spectra obtained when identical samples have been processed in parallel within the CD microlaboratory.
 | ||
| Fig. 7 High reproducibility as samples are processed under identical conditions. Results show spectra from eight samples run in parallel through 8 microstructures and analyzed by MALDI mass spectrometry. | ||
Workstation platform
CD microlaboratories are run in a robotic workstation (Fig. 8) that acts as the macro to micro interface. With so much functionality within a CD microlaboratory the workstation becomes essentially a high precision tool for transferring liquids into the CD and monitoring results. | ||
| Fig. 8 Gyrolab Workstation. Six microplates and 5 CD microlaboratories are stored within the system. | ||
CDs are moved to the spinning station for processing. A high precision robot transfers samples and reagents from microplates to the inlets of the CD where capillary force draws the liquids into the microstructures. Multiple needles on the robotic arm apply both samples and solutions (Fig. 9).
 | ||
| Fig. 9 Precision transfer into the CD microlaboratory. | ||
The needles are rinsed thoroughly at a wash station between every step.
High precision mechanics and frequent position calibration by an XYZ robot ensure 100% reliability in position accuracy of the robotic arm.
Up to 5 CDs can be processed in a single run under the control of application-specific software. The workstation can be readily adapted for different applications by modifying the control software and adding detectors specific for each application.
Conclusion: realizing the potential of miniaturization and integration
The advantages of working at microscale must be clear for any application in order to justify the transition from ‘macro’ to ‘micro’ scale. The true potential of microfluidic solutions will be realized by taking a new look at conventional working procedures, pinpointing areas in which miniaturization and integration can be of significant benefit and developing solutions that enhance both performance and productivity. For example, earlier steps such as enzyme digestion could be integrated into the sample preparation process for MALDI mass spectrometry and other applications such as immunoassays, enzyme assays and cell-based assays are already being tested using Gyros technology. The development of platforms such as that from Gyros AB offers a significant step towards new ways of working in drug discovery and development.Rolf Ehrnström
Gyros AB, Uppsala, Sweden
| This journal is © The Royal Society of Chemistry 2002 |